[meta] Apply clippy changes

Causes linker errors for unknown reasons.

.clang-format

@@ -0,0 +1,32 @@
+BasedOnStyle: LLVM
+
+Language: Cpp
+Standard: Cpp11
+
+AccessModifierOffset: -1
+AlignEscapedNewlines: Left
+AlwaysBreakAfterReturnType: None
+AlwaysBreakTemplateDeclarations: Yes
+AllowAllParametersOfDeclarationOnNextLine: false
+AllowShortFunctionsOnASingleLine: Inline
+BinPackParameters: false
+BreakBeforeBinaryOperators: NonAssignment
+BreakBeforeTernaryOperators: true
+BreakConstructorInitializers: AfterColon
+BreakInheritanceList: AfterColon
+ColumnLimit: 120
+ConstructorInitializerAllOnOneLineOrOnePerLine: true
+ContinuationIndentWidth: 4
+DerivePointerAlignment: false
+IndentCaseLabels: true
+IndentPPDirectives: None
+IndentWidth: 4
+MaxEmptyLinesToKeep: 1
+PointerAlignment: Left
+ReflowComments: true
+SortIncludes: false
+SortUsingDeclarations: true
+SpaceAfterTemplateKeyword: false
+SpacesBeforeTrailingComments: 2
+TabWidth: 4
+UseTab: Never

.gitignore

@@ -1,3 +1,4 @@
 __pycache__
 /target
+/nac3standalone/demo/linalg/target
 nix/windows/msys2

.pre-commit-config.yaml

@@ -1,24 +1,24 @@
 # See https://pre-commit.com for more information
 # See https://pre-commit.com/hooks.html for more hooks
 
-default_stages: [commit]
+default_stages: [pre-commit]
 
 repos:
   - repo: local
     hooks:
       - id: nac3-cargo-fmt
         name: nac3 cargo format
-        entry: cargo
+        entry: nix
         language: system
         types: [file, rust]
         pass_filenames: false
         description: Runs cargo fmt on the codebase.
-        args: [fmt]
+        args: [develop, -c, cargo, fmt, --all]
       - id: nac3-cargo-clippy
         name: nac3 cargo clippy
-        entry: cargo
+        entry: nix
         language: system
         types: [file, rust]
         pass_filenames: false
         description: Runs cargo clippy on the codebase.
-        args: [clippy, --tests]
+        args: [develop, -c, cargo, clippy, --tests]

Cargo.toml

@@ -4,6 +4,7 @@ members = [
   "nac3ast",
   "nac3parser",
   "nac3core",
+  "nac3core/nac3core_derive",
   "nac3standalone",
   "nac3artiq",
   "runkernel",

flake.lock

@@ -2,11 +2,11 @@
   "nodes": {
     "nixpkgs": {
       "locked": {
-        "lastModified": 1718530797,
-        "narHash": "sha256-pup6cYwtgvzDpvpSCFh1TEUjw2zkNpk8iolbKnyFmmU=",
+        "lastModified": 1731319897,
+        "narHash": "sha256-PbABj4tnbWFMfBp6OcUK5iGy1QY+/Z96ZcLpooIbuEI=",
         "owner": "NixOS",
         "repo": "nixpkgs",
-        "rev": "b60ebf54c15553b393d144357375ea956f89e9a9",
+        "rev": "dc460ec76cbff0e66e269457d7b728432263166c",
         "type": "github"
       },
       "original": {

flake.nix

@@ -6,6 +6,7 @@
   outputs = { self, nixpkgs }:
     let
       pkgs = import nixpkgs { system = "x86_64-linux"; };
+      pkgs32 = import nixpkgs { system = "i686-linux"; };
     in rec {
       packages.x86_64-linux = rec {
         llvm-nac3 = pkgs.callPackage ./nix/llvm {};
@@ -15,6 +16,22 @@
           ln -s ${pkgs.llvmPackages_14.clang-unwrapped}/bin/clang $out/bin/clang-irrt
           ln -s ${pkgs.llvmPackages_14.llvm.out}/bin/llvm-as $out/bin/llvm-as-irrt
           '';
+        demo-linalg-stub = pkgs.rustPlatform.buildRustPackage {
+          name = "demo-linalg-stub";
+          src = ./nac3standalone/demo/linalg;
+          cargoLock = {
+            lockFile = ./nac3standalone/demo/linalg/Cargo.lock;
+          };
+          doCheck = false;
+        };
+        demo-linalg-stub32 = pkgs32.rustPlatform.buildRustPackage {
+          name = "demo-linalg-stub32";
+          src = ./nac3standalone/demo/linalg;
+          cargoLock = {
+            lockFile = ./nac3standalone/demo/linalg/Cargo.lock;
+          };
+          doCheck = false;
+        };
         nac3artiq = pkgs.python3Packages.toPythonModule (
           pkgs.rustPlatform.buildRustPackage rec {
             name = "nac3artiq";
@@ -24,7 +41,7 @@
               lockFile = ./Cargo.lock;
             };
             passthru.cargoLock = cargoLock;
-            nativeBuildInputs = [ pkgs.python3 pkgs.llvmPackages_14.clang llvm-tools-irrt pkgs.llvmPackages_14.llvm.out llvm-nac3 ];
+            nativeBuildInputs = [ pkgs.python3 (pkgs.wrapClangMulti pkgs.llvmPackages_14.clang) llvm-tools-irrt pkgs.llvmPackages_14.llvm.out llvm-nac3 ];
             buildInputs = [ pkgs.python3 llvm-nac3 ];
             checkInputs = [ (pkgs.python3.withPackages(ps: [ ps.numpy ps.scipy ])) ];
             checkPhase =
@@ -32,7 +49,9 @@
               echo "Checking nac3standalone demos..."
               pushd nac3standalone/demo
               patchShebangs .
-              ./check_demos.sh
+              export DEMO_LINALG_STUB=${demo-linalg-stub}/lib/liblinalg.a
+              export DEMO_LINALG_STUB32=${demo-linalg-stub32}/lib/liblinalg.a
+              ./check_demos.sh -i686
               popd
               echo "Running Cargo tests..."
               cargoCheckHook
@@ -88,18 +107,18 @@
             (pkgs.fetchFromGitHub {
               owner = "m-labs";
               repo = "sipyco";
-              rev = "939f84f9b5eef7efbf7423c735d1834783b6140e";
-              sha256 = "sha256-15Nun4EY35j+6SPZkjzZtyH/ncxLS60KuGJjFh5kSTc=";
+              rev = "094a6cd63ffa980ef63698920170e50dc9ba77fd";
+              sha256 = "sha256-PPnAyDedUQ7Og/Cby9x5OT9wMkNGTP8GS53V6N/dk4w=";
             })
             (pkgs.fetchFromGitHub {
               owner = "m-labs";
               repo = "artiq";
-              rev = "923ca3377d42c815f979983134ec549dc39d3ca0";
-              sha256 = "sha256-oJoEeNEeNFSUyh6jXG8Tzp6qHVikeHS0CzfE+mODPgw=";
+              rev = "28c9de3e251daa89a8c9fd79d5ab64a3ec03bac6";
+              sha256 = "sha256-vAvpbHc5B+1wtG8zqN7j9dQE1ON+i22v+uqA+tw6Gak=";
             })
           ];
           buildInputs = [
-            (python3-mimalloc.withPackages(ps: [ ps.numpy ps.scipy ps.jsonschema ps.lmdb nac3artiq-instrumented ]))
+            (python3-mimalloc.withPackages(ps: [ ps.numpy ps.scipy ps.jsonschema ps.lmdb ps.platformdirs nac3artiq-instrumented ]))
             pkgs.llvmPackages_14.llvm.out
           ];
           phases = [ "buildPhase" "installPhase" ];
@@ -149,7 +168,7 @@
         buildInputs = with pkgs; [
           # build dependencies
           packages.x86_64-linux.llvm-nac3
-          llvmPackages_14.clang llvmPackages_14.llvm.out  # for running nac3standalone demos
+          (pkgs.wrapClangMulti llvmPackages_14.clang) llvmPackages_14.llvm.out  # for running nac3standalone demos
           packages.x86_64-linux.llvm-tools-irrt
           cargo
           rustc
@@ -162,6 +181,11 @@
           pre-commit
           rustfmt
         ];
+        shellHook =
+          ''
+          export DEMO_LINALG_STUB=${packages.x86_64-linux.demo-linalg-stub}/lib/liblinalg.a
+          export DEMO_LINALG_STUB32=${packages.x86_64-linux.demo-linalg-stub32}/lib/liblinalg.a
+          '';
       };
       devShells.x86_64-linux.msys2 = pkgs.mkShell {
         name = "nac3-dev-shell-msys2";

nac3artiq/Cargo.toml

@@ -12,15 +12,10 @@ crate-type = ["cdylib"]
 itertools = "0.13"
 pyo3 = { version = "0.21", features = ["extension-module", "gil-refs"] }
 parking_lot = "0.12"
-tempfile = "3.10"
-nac3parser = { path = "../nac3parser" }
+tempfile = "3.13"
 nac3core = { path = "../nac3core" }
 nac3ld = { path = "../nac3ld" }
 
-[dependencies.inkwell]
-version = "0.4"
-default-features = false
-features = ["llvm14-0", "target-x86", "target-arm", "target-riscv", "no-libffi-linking"]
-
 [features]
 init-llvm-profile = []
+no-escape-analysis = ["nac3core/no-escape-analysis"]

nac3artiq/demo/list_cmp.py

@@ -0,0 +1,24 @@
+from min_artiq import *
+from numpy import int32
+
+
+@nac3
+class EmptyList:
+    core: KernelInvariant[Core]
+
+    def __init__(self):
+        self.core = Core()
+
+    @rpc
+    def get_empty(self) -> list[int32]:
+        return []
+
+    @kernel
+    def run(self):
+        a: list[int32] = self.get_empty()
+        if a != []:
+            raise ValueError
+
+
+if __name__ == "__main__":
+    EmptyList().run()

nac3artiq/demo/min_artiq.py

@@ -112,10 +112,15 @@ def extern(function):
     register_function(function)
     return function
 
-def rpc(function):
-    """Decorates a function declaration defined by the core device runtime."""
-    register_function(function)
-    return function
+
+def rpc(arg=None, flags={}):
+    """Decorates a function or method to be executed on the host interpreter."""
+    if arg is None:
+        def inner_decorator(function):
+            return rpc(function, flags)
+        return inner_decorator
+    register_function(arg)
+    return arg
 
 def kernel(function_or_method):
     """Decorates a function or method to be executed on the core device."""
@@ -201,7 +206,7 @@ class Core:
         embedding = EmbeddingMap()
 
         if allow_registration:
-            compiler.analyze(registered_functions, registered_classes)
+            compiler.analyze(registered_functions, registered_classes, set())
             allow_registration = False
 
         if hasattr(method, "__self__"):

nac3artiq/demo/str_abi.py

@@ -0,0 +1,26 @@
+from min_artiq import *
+from numpy import ndarray, zeros as np_zeros
+
+
+@nac3
+class StrFail:
+    core: KernelInvariant[Core]
+
+    def __init__(self):
+        self.core = Core()
+
+    @kernel
+    def hello(self, arg: str):
+        pass
+
+    @kernel
+    def consume_ndarray(self, arg: ndarray[str, 1]):
+        pass
+
+    def run(self):
+        self.hello("world")
+        self.consume_ndarray(np_zeros([10], dtype=str))
+
+
+if __name__ == "__main__":
+    StrFail().run()

nac3artiq/src/lib.rs

@@ -1,10 +1,4 @@
-#![deny(
-    future_incompatible,
-    let_underscore,
-    nonstandard_style,
-    rust_2024_compatibility,
-    clippy::all
-)]
+#![deny(future_incompatible, let_underscore, nonstandard_style, clippy::all)]
 #![warn(clippy::pedantic)]
 #![allow(
     unsafe_op_in_unsafe_fn,
@@ -16,63 +10,65 @@
     clippy::wildcard_imports
 )]
 
-use std::collections::{HashMap, HashSet};
-use std::fs;
-use std::io::Write;
-use std::process::Command;
-use std::rc::Rc;
-use std::sync::Arc;
+use std::{
+    collections::{HashMap, HashSet},
+    fs,
+    io::Write,
+    process::Command,
+    rc::Rc,
+    sync::Arc,
+};
 
-use inkwell::{
+use itertools::Itertools;
+use parking_lot::{Mutex, RwLock};
+use pyo3::{
+    create_exception, exceptions,
+    prelude::*,
+    types::{PyBytes, PyDict, PyNone, PySet},
+};
+use tempfile::{self, TempDir};
+
+use nac3core::{
+    codegen::{
+        concrete_type::ConcreteTypeStore, gen_func_impl, irrt::load_irrt, CodeGenLLVMOptions,
+        CodeGenTargetMachineOptions, CodeGenTask, CodeGenerator, WithCall, WorkerRegistry,
+    },
+    inkwell::{
+        context::Context,
         memory_buffer::MemoryBuffer,
-    module::{Linkage, Module},
+        module::{FlagBehavior, Linkage, Module},
         passes::PassBuilderOptions,
         support::is_multithreaded,
         targets::*,
         OptimizationLevel,
-};
-use itertools::Itertools;
-use nac3core::codegen::{gen_func_impl, CodeGenLLVMOptions, CodeGenTargetMachineOptions};
-use nac3core::toplevel::builtins::get_exn_constructor;
-use nac3core::typecheck::typedef::{TypeEnum, Unifier, VarMap};
-use nac3parser::{
-    ast::{ExprKind, Stmt, StmtKind, StrRef},
+    },
+    nac3parser::{
+        ast::{Constant, ExprKind, Located, Stmt, StmtKind, StrRef},
         parser::parse_program,
-};
-use pyo3::create_exception;
-use pyo3::prelude::*;
-use pyo3::{exceptions, types::PyBytes, types::PyDict, types::PySet};
-
-use parking_lot::{Mutex, RwLock};
-
-use nac3core::{
-    codegen::irrt::load_irrt,
-    codegen::{concrete_type::ConcreteTypeStore, CodeGenTask, WithCall, WorkerRegistry},
+    },
     symbol_resolver::SymbolResolver,
     toplevel::{
-        composer::{ComposerConfig, TopLevelComposer},
+        builtins::get_exn_constructor,
+        composer::{BuiltinFuncCreator, BuiltinFuncSpec, ComposerConfig, TopLevelComposer},
         DefinitionId, GenCall, TopLevelDef,
     },
-    typecheck::typedef::{FunSignature, FuncArg},
-    typecheck::{type_inferencer::PrimitiveStore, typedef::Type},
+    typecheck::{
+        type_inferencer::PrimitiveStore,
+        typedef::{into_var_map, FunSignature, FuncArg, Type, TypeEnum, Unifier, VarMap},
+    },
 };
-
 use nac3ld::Linker;
 
-use tempfile::{self, TempDir};
-
-use crate::codegen::attributes_writeback;
-use crate::{
-    codegen::{rpc_codegen_callback, ArtiqCodeGenerator},
-    symbol_resolver::{DeferredEvaluationStore, InnerResolver, PythonHelper, Resolver},
+use codegen::{
+    attributes_writeback, gen_core_log, gen_rtio_log, rpc_codegen_callback, ArtiqCodeGenerator,
 };
+use symbol_resolver::{DeferredEvaluationStore, InnerResolver, PythonHelper, Resolver};
+use timeline::TimeFns;
 
 mod codegen;
 mod symbol_resolver;
 mod timeline;
 
-use timeline::TimeFns;
-
 #[derive(PartialEq, Clone, Copy)]
 enum Isa {
     Host,
@@ -126,7 +122,7 @@ struct Nac3 {
     isa: Isa,
     time_fns: &'static (dyn TimeFns + Sync),
     primitive: PrimitiveStore,
-    builtins: Vec<(StrRef, FunSignature, Arc<GenCall>)>,
+    builtins: Vec<BuiltinFuncSpec>,
     pyid_to_def: Arc<RwLock<HashMap<u64, DefinitionId>>>,
     primitive_ids: PrimitivePythonId,
     working_directory: TempDir,
@@ -146,14 +142,32 @@ impl Nac3 {
         module: &PyObject,
         registered_class_ids: &HashSet<u64>,
     ) -> PyResult<()> {
-        let (module_name, source_file) = Python::with_gil(|py| -> PyResult<(String, String)> {
+        let (module_name, source_file, source) =
+            Python::with_gil(|py| -> PyResult<(String, String, String)> {
                 let module: &PyAny = module.extract(py)?;
-            Ok((module.getattr("__name__")?.extract()?, module.getattr("__file__")?.extract()?))
+                let source_file = module.getattr("__file__");
+                let (source_file, source) = if let Ok(source_file) = source_file {
+                    let source_file = source_file.extract()?;
+                    (
+                        source_file,
+                        fs::read_to_string(source_file).map_err(|e| {
+                            exceptions::PyIOError::new_err(format!(
+                                "failed to read input file: {e}"
+                            ))
+                        })?,
+                    )
+                } else {
+                    // kernels submitted by content have no file
+                    // but still can provide source by StringLoader
+                    let get_src_fn = module
+                        .getattr("__loader__")?
+                        .extract::<PyObject>()?
+                        .getattr(py, "get_source")?;
+                    ("<expcontent>", get_src_fn.call1(py, (PyNone::get(py),))?.extract(py)?)
+                };
+                Ok((module.getattr("__name__")?.extract()?, source_file.to_string(), source))
             })?;
 
-        let source = fs::read_to_string(&source_file).map_err(|e| {
-            exceptions::PyIOError::new_err(format!("failed to read input file: {e}"))
-        })?;
         let parser_result = parse_program(&source, source_file.into())
             .map_err(|e| exceptions::PySyntaxError::new_err(format!("parse error: {e}")))?;
 
@@ -193,10 +207,8 @@ impl Nac3 {
                     body.retain(|stmt| {
                         if let StmtKind::FunctionDef { ref decorator_list, .. } = stmt.node {
                             decorator_list.iter().any(|decorator| {
-                                if let ExprKind::Name { id, .. } = decorator.node {
-                                    id.to_string() == "kernel"
-                                        || id.to_string() == "portable"
-                                        || id.to_string() == "rpc"
+                                if let Some(id) = decorator_id_string(decorator) {
+                                    id == "kernel" || id == "portable" || id == "rpc"
                                 } else {
                                     false
                                 }
@@ -209,9 +221,8 @@ impl Nac3 {
                 }
                 StmtKind::FunctionDef { ref decorator_list, .. } => {
                     decorator_list.iter().any(|decorator| {
-                        if let ExprKind::Name { id, .. } = decorator.node {
-                            let id = id.to_string();
-                            id == "extern" || id == "portable" || id == "kernel" || id == "rpc"
+                        if let Some(id) = decorator_id_string(decorator) {
+                            id == "extern" || id == "kernel" || id == "portable" || id == "rpc"
                         } else {
                             false
                         }
@@ -264,7 +275,7 @@ impl Nac3 {
                     arg_names.len(),
                 ));
             }
-            for (i, FuncArg { ty, default_value, name }) in args.iter().enumerate() {
+            for (i, FuncArg { ty, default_value, name, .. }) in args.iter().enumerate() {
                 let in_name = match arg_names.get(i) {
                     Some(n) => n,
                     None if default_value.is_none() => {
@@ -300,6 +311,64 @@ impl Nac3 {
         None
     }
 
+    /// Returns a [`Vec`] of builtins that needs to be initialized during method compilation time.
+    fn get_lateinit_builtins() -> Vec<Box<BuiltinFuncCreator>> {
+        vec![
+            Box::new(|primitives, unifier| {
+                let arg_ty = unifier.get_fresh_var(Some("T".into()), None);
+
+                (
+                    "core_log".into(),
+                    FunSignature {
+                        args: vec![FuncArg {
+                            name: "arg".into(),
+                            ty: arg_ty.ty,
+                            default_value: None,
+                            is_vararg: false,
+                        }],
+                        ret: primitives.none,
+                        vars: into_var_map([arg_ty]),
+                    },
+                    Arc::new(GenCall::new(Box::new(move |ctx, obj, fun, args, generator| {
+                        gen_core_log(ctx, &obj, fun, &args, generator)?;
+
+                        Ok(None)
+                    }))),
+                )
+            }),
+            Box::new(|primitives, unifier| {
+                let arg_ty = unifier.get_fresh_var(Some("T".into()), None);
+
+                (
+                    "rtio_log".into(),
+                    FunSignature {
+                        args: vec![
+                            FuncArg {
+                                name: "channel".into(),
+                                ty: primitives.str,
+                                default_value: None,
+                                is_vararg: false,
+                            },
+                            FuncArg {
+                                name: "arg".into(),
+                                ty: arg_ty.ty,
+                                default_value: None,
+                                is_vararg: false,
+                            },
+                        ],
+                        ret: primitives.none,
+                        vars: into_var_map([arg_ty]),
+                    },
+                    Arc::new(GenCall::new(Box::new(move |ctx, obj, fun, args, generator| {
+                        gen_rtio_log(ctx, &obj, fun, &args, generator)?;
+
+                        Ok(None)
+                    }))),
+                )
+            }),
+        ]
+    }
+
     fn compile_method<T>(
         &self,
         obj: &PyAny,
@@ -312,6 +381,7 @@ impl Nac3 {
         let size_t = self.isa.get_size_type();
         let (mut composer, mut builtins_def, mut builtins_ty) = TopLevelComposer::new(
             self.builtins.clone(),
+            Self::get_lateinit_builtins(),
             ComposerConfig { kernel_ann: Some("Kernel"), kernel_invariant_ann: "KernelInvariant" },
             size_t,
         );
@@ -388,7 +458,6 @@ impl Nac3 {
                         pyid_to_type: pyid_to_type.clone(),
                         primitive_ids: self.primitive_ids.clone(),
                         global_value_ids: global_value_ids.clone(),
-                        class_names: Mutex::default(),
                         name_to_pyid: name_to_pyid.clone(),
                         module: module.clone(),
                         id_to_pyval: RwLock::default(),
@@ -419,9 +488,25 @@ impl Nac3 {
 
             match &stmt.node {
                 StmtKind::FunctionDef { decorator_list, .. } => {
-                    if decorator_list.iter().any(|decorator| matches!(decorator.node, ExprKind::Name { id, .. } if id == "rpc".into())) {
-                        store_fun.call1(py, (def_id.0.into_py(py), module.getattr(py, name.to_string().as_str()).unwrap())).unwrap();
-                        rpc_ids.push((None, def_id));
+                    if decorator_list
+                        .iter()
+                        .any(|decorator| decorator_id_string(decorator) == Some("rpc".to_string()))
+                    {
+                        store_fun
+                            .call1(
+                                py,
+                                (
+                                    def_id.0.into_py(py),
+                                    module.getattr(py, name.to_string().as_str()).unwrap(),
+                                ),
+                            )
+                            .unwrap();
+                        let is_async = decorator_list.iter().any(|decorator| {
+                            decorator_get_flags(decorator)
+                                .iter()
+                                .any(|constant| *constant == Constant::Str("async".into()))
+                        });
+                        rpc_ids.push((None, def_id, is_async));
                     }
                 }
                 StmtKind::ClassDef { name, body, .. } => {
@@ -429,19 +514,26 @@ impl Nac3 {
                     let class_obj = module.getattr(py, class_name.as_str()).unwrap();
                     for stmt in body {
                         if let StmtKind::FunctionDef { name, decorator_list, .. } = &stmt.node {
-                            if decorator_list.iter().any(|decorator| matches!(decorator.node, ExprKind::Name { id, .. } if id == "rpc".into())) {
+                            if decorator_list.iter().any(|decorator| {
+                                decorator_id_string(decorator) == Some("rpc".to_string())
+                            }) {
+                                let is_async = decorator_list.iter().any(|decorator| {
+                                    decorator_get_flags(decorator)
+                                        .iter()
+                                        .any(|constant| *constant == Constant::Str("async".into()))
+                                });
                                 if name == &"__init__".into() {
                                     return Err(CompileError::new_err(format!(
                                         "compilation failed\n----------\nThe constructor of class {} should not be decorated with rpc decorator (at {})",
                                         class_name, stmt.location
                                     )));
                                 }
-                                rpc_ids.push((Some((class_obj.clone(), *name)), def_id));
+                                rpc_ids.push((Some((class_obj.clone(), *name)), def_id, is_async));
                             }
                         }
                     }
                 }
-                _ => ()
+                _ => (),
             }
 
             let id = *name_to_pyid.get(&name).unwrap();
@@ -480,7 +572,6 @@ impl Nac3 {
             pyid_to_type: pyid_to_type.clone(),
             primitive_ids: self.primitive_ids.clone(),
             global_value_ids: global_value_ids.clone(),
-            class_names: Mutex::default(),
             id_to_pyval: RwLock::default(),
             id_to_primitive: RwLock::default(),
             field_to_val: RwLock::default(),
@@ -497,6 +588,10 @@ impl Nac3 {
             .register_top_level(synthesized.pop().unwrap(), Some(resolver.clone()), "", false)
             .unwrap();
 
+        // Process IRRT
+        let context = Context::create();
+        let irrt = load_irrt(&context, resolver.as_ref());
+
         let fun_signature =
             FunSignature { args: vec![], ret: self.primitive.none, vars: VarMap::new() };
         let mut store = ConcreteTypeStore::new();
@@ -534,13 +629,12 @@ impl Nac3 {
         let top_level = Arc::new(composer.make_top_level_context());
 
         {
-            let rpc_codegen = rpc_codegen_callback();
             let defs = top_level.definitions.read();
-            for (class_data, id) in &rpc_ids {
+            for (class_data, id, is_async) in &rpc_ids {
                 let mut def = defs[id.0].write();
                 match &mut *def {
                     TopLevelDef::Function { codegen_callback, .. } => {
-                        *codegen_callback = Some(rpc_codegen.clone());
+                        *codegen_callback = Some(rpc_codegen_callback(*is_async));
                     }
                     TopLevelDef::Class { methods, .. } => {
                         let (class_def, method_name) = class_data.as_ref().unwrap();
@@ -551,7 +645,7 @@ impl Nac3 {
                             if let TopLevelDef::Function { codegen_callback, .. } =
                                 &mut *defs[id.0].write()
                             {
-                                *codegen_callback = Some(rpc_codegen.clone());
+                                *codegen_callback = Some(rpc_codegen_callback(*is_async));
                                 store_fun
                                     .call1(
                                         py,
@@ -566,6 +660,11 @@ impl Nac3 {
                             }
                         }
                     }
+                    TopLevelDef::Variable { .. } => {
+                        return Err(CompileError::new_err(String::from(
+                            "Unsupported @rpc annotation on global variable",
+                        )))
+                    }
                 }
             }
         }
@@ -586,33 +685,12 @@ impl Nac3 {
         let task = CodeGenTask {
             subst: Vec::default(),
             symbol_name: "__modinit__".to_string(),
-            body: instance.body,
-            signature,
-            resolver: resolver.clone(),
-            store,
-            unifier_index: instance.unifier_id,
-            calls: instance.calls,
-            id: 0,
-        };
-
-        let mut store = ConcreteTypeStore::new();
-        let mut cache = HashMap::new();
-        let signature = store.from_signature(
-            &mut composer.unifier,
-            &self.primitive,
-            &fun_signature,
-            &mut cache,
-        );
-        let signature = store.add_cty(signature);
-        let attributes_writeback_task = CodeGenTask {
-            subst: Vec::default(),
-            symbol_name: "attributes_writeback".to_string(),
             body: Arc::new(Vec::default()),
             signature,
             resolver,
             store,
             unifier_index: instance.unifier_id,
-            calls: Arc::new(HashMap::default()),
+            calls: instance.calls,
             id: 0,
         };
 
@@ -625,7 +703,9 @@ impl Nac3 {
             let buffer = buffer.as_slice().into();
             membuffer.lock().push(buffer);
         })));
-        let size_t = if self.isa == Isa::Host { 64 } else { 32 };
+        let size_t = context
+            .ptr_sized_int_type(&self.get_llvm_target_machine().get_target_data(), None)
+            .get_bit_width();
         let num_threads = if is_multithreaded() { 4 } else { 1 };
         let thread_names: Vec<String> = (0..num_threads).map(|_| "main".to_string()).collect();
         let threads: Vec<_> = thread_names
@@ -634,16 +714,27 @@ impl Nac3 {
             .collect();
 
         let membuffer = membuffers.clone();
+        let mut has_return = false;
         py.allow_threads(|| {
             let (registry, handles) =
                 WorkerRegistry::create_workers(threads, top_level.clone(), &self.llvm_options, &f);
-            registry.add_task(task);
-            registry.wait_tasks_complete(handles);
 
-            let mut generator =
-                ArtiqCodeGenerator::new("attributes_writeback".to_string(), size_t, self.time_fns);
-            let context = inkwell::context::Context::create();
-            let module = context.create_module("attributes_writeback");
+            let mut generator = ArtiqCodeGenerator::new("main".to_string(), size_t, self.time_fns);
+            let context = Context::create();
+            let module = context.create_module("main");
+            let target_machine = self.llvm_options.create_target_machine().unwrap();
+            module.set_data_layout(&target_machine.get_target_data().get_data_layout());
+            module.set_triple(&target_machine.get_triple());
+            module.add_basic_value_flag(
+                "Debug Info Version",
+                FlagBehavior::Warning,
+                context.i32_type().const_int(3, false),
+            );
+            module.add_basic_value_flag(
+                "Dwarf Version",
+                FlagBehavior::Warning,
+                context.i32_type().const_int(4, false),
+            );
             let builder = context.create_builder();
             let (_, module, _) = gen_func_impl(
                 &context,
@@ -651,9 +742,27 @@ impl Nac3 {
                 &registry,
                 builder,
                 module,
-                attributes_writeback_task,
+                task,
                 |generator, ctx| {
-                    attributes_writeback(ctx, generator, inner_resolver.as_ref(), &host_attributes)
+                    assert_eq!(instance.body.len(), 1, "toplevel module should have 1 statement");
+                    let StmtKind::Expr { value: ref expr, .. } = instance.body[0].node else {
+                        unreachable!("toplevel statement must be an expression")
+                    };
+                    let ExprKind::Call { .. } = expr.node else {
+                        unreachable!("toplevel expression must be a function call")
+                    };
+
+                    let return_obj =
+                        generator.gen_expr(ctx, expr)?.map(|value| (expr.custom.unwrap(), value));
+                    has_return = return_obj.is_some();
+                    registry.wait_tasks_complete(handles);
+                    attributes_writeback(
+                        ctx,
+                        generator,
+                        inner_resolver.as_ref(),
+                        &host_attributes,
+                        return_obj,
+                    )
                 },
             )
             .unwrap();
@@ -662,37 +771,24 @@ impl Nac3 {
             membuffer.lock().push(buffer);
         });
 
-        let context = inkwell::context::Context::create();
+        embedding_map.setattr("expects_return", has_return).unwrap();
+
+        // Link all modules into `main`.
         let buffers = membuffers.lock();
         let main = context
-            .create_module_from_ir(MemoryBuffer::create_from_memory_range(&buffers[0], "main"))
+            .create_module_from_ir(MemoryBuffer::create_from_memory_range(
+                buffers.last().unwrap(),
+                "main",
+            ))
             .unwrap();
-        for buffer in buffers.iter().skip(1) {
+        for buffer in buffers.iter().rev().skip(1) {
             let other = context
                 .create_module_from_ir(MemoryBuffer::create_from_memory_range(buffer, "main"))
                 .unwrap();
 
             main.link_in_module(other).map_err(|err| CompileError::new_err(err.to_string()))?;
         }
-        let builder = context.create_builder();
-        let modinit_return = main
-            .get_function("__modinit__")
-            .unwrap()
-            .get_last_basic_block()
-            .unwrap()
-            .get_terminator()
-            .unwrap();
-        builder.position_before(&modinit_return);
-        builder
-            .build_call(
-                main.get_function("attributes_writeback").unwrap(),
-                &[],
-                "attributes_writeback",
-            )
-            .unwrap();
-
-        main.link_in_module(load_irrt(&context))
-            .map_err(|err| CompileError::new_err(err.to_string()))?;
+        main.link_in_module(irrt).map_err(|err| CompileError::new_err(err.to_string()))?;
 
         let mut function_iter = main.get_first_function();
         while let Some(func) = function_iter {
@@ -778,6 +874,41 @@ impl Nac3 {
     }
 }
 
+/// Retrieves the Name.id from a decorator, supports decorators with arguments.
+fn decorator_id_string(decorator: &Located<ExprKind>) -> Option<String> {
+    if let ExprKind::Name { id, .. } = decorator.node {
+        // Bare decorator
+        return Some(id.to_string());
+    } else if let ExprKind::Call { func, .. } = &decorator.node {
+        // Decorators that are calls (e.g. "@rpc()") have Call for the node,
+        // need to extract the id from within.
+        if let ExprKind::Name { id, .. } = func.node {
+            return Some(id.to_string());
+        }
+    }
+    None
+}
+
+/// Retrieves flags from a decorator, if any.
+fn decorator_get_flags(decorator: &Located<ExprKind>) -> Vec<Constant> {
+    let mut flags = vec![];
+    if let ExprKind::Call { keywords, .. } = &decorator.node {
+        for keyword in keywords {
+            if keyword.node.arg != Some("flags".into()) {
+                continue;
+            }
+            if let ExprKind::Set { elts } = &keyword.node.value.node {
+                for elt in elts {
+                    if let ExprKind::Constant { value, .. } = &elt.node {
+                        flags.push(value.clone());
+                    }
+                }
+            }
+        }
+    }
+    flags
+}
+
 fn link_with_lld(elf_filename: String, obj_filename: String) -> PyResult<()> {
     let linker_args = vec![
         "-shared".to_string(),
@@ -847,7 +978,7 @@ impl Nac3 {
             Isa::RiscV32IMA => &timeline::NOW_PINNING_TIME_FNS,
             Isa::CortexA9 | Isa::Host => &timeline::EXTERN_TIME_FNS,
         };
-        let primitive: PrimitiveStore = TopLevelComposer::make_primitives(isa.get_size_type()).0;
+        let (primitive, _) = TopLevelComposer::make_primitives(isa.get_size_type());
         let builtins = vec![
             (
                 "now_mu".into(),
@@ -863,6 +994,7 @@ impl Nac3 {
                         name: "t".into(),
                         ty: primitive.int64,
                         default_value: None,
+                        is_vararg: false,
                     }],
                     ret: primitive.none,
                     vars: VarMap::new(),
@@ -882,6 +1014,7 @@ impl Nac3 {
                         name: "dt".into(),
                         ty: primitive.int64,
                         default_value: None,
+                        is_vararg: false,
                     }],
                     ret: primitive.none,
                     vars: VarMap::new(),
@@ -957,7 +1090,12 @@ impl Nac3 {
         })
     }
 
-    fn analyze(&mut self, functions: &PySet, classes: &PySet) -> PyResult<()> {
+    fn analyze(
+        &mut self,
+        functions: &PySet,
+        classes: &PySet,
+        content_modules: &PySet,
+    ) -> PyResult<()> {
         let (modules, class_ids) =
             Python::with_gil(|py| -> PyResult<(HashMap<u64, PyObject>, HashSet<u64>)> {
                 let mut modules: HashMap<u64, PyObject> = HashMap::new();
@@ -967,14 +1105,22 @@ impl Nac3 {
                 let getmodule_fn = PyModule::import(py, "inspect")?.getattr("getmodule")?;
 
                 for function in functions {
-                    let module = getmodule_fn.call1((function,))?.extract()?;
+                    let module: PyObject = getmodule_fn.call1((function,))?.extract()?;
+                    if !module.is_none(py) {
                         modules.insert(id_fn.call1((&module,))?.extract()?, module);
                     }
+                }
                 for class in classes {
-                    let module = getmodule_fn.call1((class,))?.extract()?;
+                    let module: PyObject = getmodule_fn.call1((class,))?.extract()?;
+                    if !module.is_none(py) {
                         modules.insert(id_fn.call1((&module,))?.extract()?, module);
+                    }
                     class_ids.insert(id_fn.call1((class,))?.extract()?);
                 }
+                for module in content_modules {
+                    let module: PyObject = module.extract()?;
+                    modules.insert(id_fn.call1((&module,))?.extract()?, module);
+                }
                 Ok((modules, class_ids))
             })?;
 

nac3artiq/src/symbol_resolver.rs

@@ -1,14 +1,30 @@
-use inkwell::{
+use std::{
+    collections::{HashMap, HashSet},
+    sync::{
+        atomic::{AtomicBool, Ordering::Relaxed},
+        Arc,
+    },
+};
+
+use itertools::Itertools;
+use parking_lot::RwLock;
+use pyo3::{
+    types::{PyDict, PyTuple},
+    PyAny, PyObject, PyResult, Python,
+};
+
+use nac3core::{
+    codegen::{
+        types::{NDArrayType, ProxyType},
+        CodeGenContext, CodeGenerator,
+    },
+    inkwell::{
+        module::Linkage,
         types::{BasicType, BasicTypeEnum},
         values::BasicValueEnum,
         AddressSpace,
-};
-use itertools::Itertools;
-use nac3core::{
-    codegen::{
-        classes::{NDArrayType, ProxyType},
-        CodeGenContext, CodeGenerator,
     },
+    nac3parser::ast::{self, StrRef},
     symbol_resolver::{StaticValue, SymbolResolver, SymbolValue, ValueEnum},
     toplevel::{
         helper::PrimDef,
@@ -20,21 +36,8 @@ use nac3core::{
         typedef::{into_var_map, iter_type_vars, Type, TypeEnum, TypeVar, Unifier, VarMap},
     },
 };
-use nac3parser::ast::{self, StrRef};
-use parking_lot::{Mutex, RwLock};
-use pyo3::{
-    types::{PyDict, PyTuple},
-    PyAny, PyObject, PyResult, Python,
-};
-use std::{
-    collections::{HashMap, HashSet},
-    sync::{
-        atomic::{AtomicBool, Ordering::Relaxed},
-        Arc,
-    },
-};
 
-use crate::PrimitivePythonId;
+use super::PrimitivePythonId;
 
 pub enum PrimitiveValue {
     I32(i32),
@@ -79,7 +82,6 @@ pub struct InnerResolver {
     pub id_to_primitive: RwLock<HashMap<u64, PrimitiveValue>>,
     pub field_to_val: RwLock<HashMap<ResolverField, Option<PyFieldHandle>>>,
     pub global_value_ids: Arc<RwLock<HashMap<u64, PyObject>>>,
-    pub class_names: Mutex<HashMap<StrRef, Type>>,
     pub pyid_to_def: Arc<RwLock<HashMap<u64, DefinitionId>>>,
     pub pyid_to_type: Arc<RwLock<HashMap<u64, Type>>>,
     pub primitive_ids: PrimitivePythonId,
@@ -133,6 +135,8 @@ impl StaticValue for PythonValue {
                         format!("{}_const", self.id).as_str(),
                     );
                     global.set_constant(true);
+                    // Set linkage of global to private to avoid name collisions
+                    global.set_linkage(Linkage::Private);
                     global.set_initializer(&ctx.ctx.const_struct(
                         &[ctx.ctx.i32_type().const_int(u64::from(id), false).into()],
                         false,
@@ -163,7 +167,7 @@ impl StaticValue for PythonValue {
                 PrimitiveValue::Bool(val) => {
                     ctx.ctx.i8_type().const_int(u64::from(*val), false).into()
                 }
-                PrimitiveValue::Str(val) => ctx.ctx.const_string(val.as_bytes(), true).into(),
+                PrimitiveValue::Str(val) => ctx.gen_string(generator, val).into(),
             });
         }
         if let Some(global) = ctx.module.get_global(&self.id.to_string()) {
@@ -329,8 +333,19 @@ impl InnerResolver {
             Ok(Ok((primitives.exception, true)))
         } else if ty_id == self.primitive_ids.list {
             // do not handle type var param and concrete check here
+            let list_tvar = if let TypeEnum::TObj { obj_id, params, .. } =
+                &*unifier.get_ty_immutable(primitives.list)
+            {
+                assert_eq!(*obj_id, PrimDef::List.id());
+                iter_type_vars(params).nth(0).unwrap()
+            } else {
+                unreachable!()
+            };
+
             let var = unifier.get_dummy_var().ty;
-            let list = unifier.add_ty(TypeEnum::TList { ty: var });
+            let list = unifier
+                .subst(primitives.list, &into_var_map([TypeVar { id: list_tvar.id, ty: var }]))
+                .unwrap();
             Ok(Ok((list, false)))
         } else if ty_id == self.primitive_ids.ndarray {
             // do not handle type var param and concrete check here
@@ -340,7 +355,7 @@ impl InnerResolver {
             Ok(Ok((ndarray, false)))
         } else if ty_id == self.primitive_ids.tuple {
             // do not handle type var param and concrete check here
-            Ok(Ok((unifier.add_ty(TypeEnum::TTuple { ty: vec![] }), false)))
+            Ok(Ok((unifier.add_ty(TypeEnum::TTuple { ty: vec![], is_vararg_ctx: false }), false)))
         } else if ty_id == self.primitive_ids.option {
             Ok(Ok((primitives.option, false)))
         } else if ty_id == self.primitive_ids.none {
@@ -460,7 +475,7 @@ impl InnerResolver {
                 };
 
             match &*unifier.get_ty(origin_ty) {
-                TypeEnum::TList { .. } => {
+                TypeEnum::TObj { obj_id, .. } if *obj_id == PrimDef::List.id() => {
                     if args.len() == 1 {
                         let ty = match self.get_pyty_obj_type(
                             py,
@@ -477,7 +492,21 @@ impl InnerResolver {
                                 "type list should take concrete parameters in typevar range".into(),
                             ));
                         }
-                        Ok(Ok((unifier.add_ty(TypeEnum::TList { ty: ty.0 }), true)))
+                        let list_tvar = if let TypeEnum::TObj { obj_id, params, .. } =
+                            &*unifier.get_ty_immutable(primitives.list)
+                        {
+                            assert_eq!(*obj_id, PrimDef::List.id());
+                            iter_type_vars(params).nth(0).unwrap()
+                        } else {
+                            unreachable!()
+                        };
+                        let list = unifier
+                            .subst(
+                                primitives.list,
+                                &into_var_map([TypeVar { id: list_tvar.id, ty: ty.0 }]),
+                            )
+                            .unwrap();
+                        Ok(Ok((list, true)))
                     } else {
                         return Ok(Err(format!(
                             "type list needs exactly 1 type parameters, found {}",
@@ -530,7 +559,10 @@ impl InnerResolver {
                             Err(err) => return Ok(Err(err)),
                             _ => return Ok(Err("tuple type needs at least 1 type parameters".to_string()))
                         };
-                    Ok(Ok((unifier.add_ty(TypeEnum::TTuple { ty: args }), true)))
+                    Ok(Ok((
+                        unifier.add_ty(TypeEnum::TTuple { ty: args, is_vararg_ctx: false }),
+                        true,
+                    )))
                 }
                 TypeEnum::TObj { params, obj_id, .. } => {
                     let subst = {
@@ -693,11 +725,12 @@ impl InnerResolver {
         };
         match (&*unifier.get_ty(extracted_ty), inst_check) {
             // do the instantiation for these four types
-            (TypeEnum::TList { ty }, false) => {
+            (TypeEnum::TObj { obj_id, params, .. }, false) if *obj_id == PrimDef::List.id() => {
+                let ty = iter_type_vars(params).nth(0).unwrap().ty;
                 let len: usize = self.helper.len_fn.call1(py, (obj,))?.extract(py)?;
                 if len == 0 {
                     assert!(matches!(
-                        &*unifier.get_ty(*ty),
+                        &*unifier.get_ty(ty),
                         TypeEnum::TVar { fields: None, range, .. }
                             if range.is_empty()
                     ));
@@ -706,8 +739,25 @@ impl InnerResolver {
                     let actual_ty =
                         self.get_list_elem_type(py, obj, len, unifier, defs, primitives)?;
                     match actual_ty {
-                        Ok(t) => match unifier.unify(*ty, t) {
-                            Ok(()) => Ok(Ok(unifier.add_ty(TypeEnum::TList { ty: *ty }))),
+                        Ok(t) => match unifier.unify(ty, t) {
+                            Ok(()) => {
+                                let list_tvar = if let TypeEnum::TObj { obj_id, params, .. } =
+                                    &*unifier.get_ty_immutable(primitives.list)
+                                {
+                                    assert_eq!(*obj_id, PrimDef::List.id());
+                                    iter_type_vars(params).nth(0).unwrap()
+                                } else {
+                                    unreachable!()
+                                };
+                                let list = unifier
+                                    .subst(
+                                        primitives.list,
+                                        &into_var_map([TypeVar { id: list_tvar.id, ty }]),
+                                    )
+                                    .unwrap();
+                                Ok(Ok(list))
+                            }
+
                             Err(e) => Ok(Err(format!(
                                 "type error ({}) for the list",
                                 e.to_display(unifier)
@@ -754,7 +804,9 @@ impl InnerResolver {
                     .map(|elem| self.get_obj_type(py, elem, unifier, defs, primitives))
                     .collect();
                 let types = types?;
-                Ok(types.map(|types| unifier.add_ty(TypeEnum::TTuple { ty: types })))
+                Ok(types.map(|types| {
+                    unifier.add_ty(TypeEnum::TTuple { ty: types, is_vararg_ctx: false })
+                }))
             }
             // special handling for option type since its class member layout in python side
             // is special and cannot be mapped directly to a nac3 type as below
@@ -929,7 +981,7 @@ impl InnerResolver {
         } else if ty_id == self.primitive_ids.string || ty_id == self.primitive_ids.np_str_ {
             let val: String = obj.extract().unwrap();
             self.id_to_primitive.write().insert(id, PrimitiveValue::Str(val.clone()));
-            Ok(Some(ctx.ctx.const_string(val.as_bytes(), true).into()))
+            Ok(Some(ctx.gen_string(generator, val).into()))
         } else if ty_id == self.primitive_ids.float || ty_id == self.primitive_ids.float64 {
             let val: f64 = obj.extract().unwrap();
             self.id_to_primitive.write().insert(id, PrimitiveValue::F64(val));
@@ -942,15 +994,21 @@ impl InnerResolver {
             }
 
             let len: usize = self.helper.len_fn.call1(py, (obj,))?.extract(py)?;
-            let elem_ty = if let TypeEnum::TList { ty } =
-                ctx.unifier.get_ty_immutable(expected_ty).as_ref()
-            {
-                *ty
-            } else {
-                unreachable!("must be list")
+            let elem_ty = match ctx.unifier.get_ty_immutable(expected_ty).as_ref() {
+                TypeEnum::TObj { obj_id, params, .. } if *obj_id == PrimDef::List.id() => {
+                    iter_type_vars(params).nth(0).unwrap().ty
+                }
+                _ => unreachable!("must be list"),
             };
-            let ty = ctx.get_llvm_type(generator, elem_ty);
             let size_t = generator.get_size_type(ctx.ctx);
+            let ty = if len == 0
+                && matches!(&*ctx.unifier.get_ty_immutable(elem_ty), TypeEnum::TVar { .. })
+            {
+                // The default type for zero-length lists of unknown element type is size_t
+                size_t.into()
+            } else {
+                ctx.get_llvm_type(generator, elem_ty)
+            };
             let arr_ty = ctx
                 .ctx
                 .struct_type(&[ty.ptr_type(AddressSpace::default()).into(), size_t.into()], false);
@@ -1038,7 +1096,7 @@ impl InnerResolver {
                 if self.global_value_ids.read().contains_key(&id) {
                     let global = ctx.module.get_global(&id_str).unwrap_or_else(|| {
                         ctx.module.add_global(
-                            ndarray_llvm_ty.as_underlying_type(),
+                            ndarray_llvm_ty.as_base_type().get_element_type().into_struct_type(),
                             Some(AddressSpace::default()),
                             &id_str,
                         )
@@ -1132,7 +1190,11 @@ impl InnerResolver {
             data_global.set_initializer(&data);
 
             // create a global for the ndarray object and initialize it
-            let value = ndarray_llvm_ty.as_underlying_type().const_named_struct(&[
+            let value = ndarray_llvm_ty
+                .as_base_type()
+                .get_element_type()
+                .into_struct_type()
+                .const_named_struct(&[
                     llvm_usize.const_int(ndarray_ndims, false).into(),
                     shape_global
                         .as_pointer_value()
@@ -1145,7 +1207,7 @@ impl InnerResolver {
                 ]);
 
             let ndarray = ctx.module.add_global(
-                ndarray_llvm_ty.as_underlying_type(),
+                ndarray_llvm_ty.as_base_type().get_element_type().into_struct_type(),
                 Some(AddressSpace::default()),
                 &id_str,
             );
@@ -1154,7 +1216,9 @@ impl InnerResolver {
             Ok(Some(ndarray.as_pointer_value().into()))
         } else if ty_id == self.primitive_ids.tuple {
             let expected_ty_enum = ctx.unifier.get_ty_immutable(expected_ty);
-            let TypeEnum::TTuple { ty } = expected_ty_enum.as_ref() else { unreachable!() };
+            let TypeEnum::TTuple { ty, is_vararg_ctx: false } = expected_ty_enum.as_ref() else {
+                unreachable!()
+            };
 
             let tup_tys = ty.iter();
             let elements: &PyTuple = obj.downcast()?;
@@ -1410,6 +1474,7 @@ impl SymbolResolver for Resolver {
         &self,
         id: StrRef,
         _: &mut CodeGenContext<'ctx, '_>,
+        _: &mut dyn CodeGenerator,
     ) -> Option<ValueEnum<'ctx>> {
         let sym_value = {
             let id_to_val = self.0.id_to_pyval.read();

nac3artiq/src/timeline.rs

@@ -1,9 +1,12 @@
-use inkwell::{
+use itertools::Either;
+
+use nac3core::{
+    codegen::CodeGenContext,
+    inkwell::{
         values::{BasicValueEnum, CallSiteValue},
         AddressSpace, AtomicOrdering,
+    },
 };
-use itertools::Either;
-use nac3core::codegen::CodeGenContext;
 
 /// Functions for manipulating the timeline.
 pub trait TimeFns {
@@ -31,7 +34,7 @@ impl TimeFns for NowPinningTimeFns64 {
             .unwrap_or_else(|| ctx.module.add_global(i64_type, None, "now"));
         let now_hiptr = ctx
             .builder
-            .build_bitcast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
+            .build_bit_cast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
             .map(BasicValueEnum::into_pointer_value)
             .unwrap();
 
@@ -80,7 +83,7 @@ impl TimeFns for NowPinningTimeFns64 {
             .unwrap_or_else(|| ctx.module.add_global(i64_type, None, "now"));
         let now_hiptr = ctx
             .builder
-            .build_bitcast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
+            .build_bit_cast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
             .map(BasicValueEnum::into_pointer_value)
             .unwrap();
 
@@ -109,7 +112,7 @@ impl TimeFns for NowPinningTimeFns64 {
             .unwrap_or_else(|| ctx.module.add_global(i64_type, None, "now"));
         let now_hiptr = ctx
             .builder
-            .build_bitcast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
+            .build_bit_cast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
             .map(BasicValueEnum::into_pointer_value)
             .unwrap();
 
@@ -207,7 +210,7 @@ impl TimeFns for NowPinningTimeFns {
             .unwrap_or_else(|| ctx.module.add_global(i64_type, None, "now"));
         let now_hiptr = ctx
             .builder
-            .build_bitcast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
+            .build_bit_cast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
             .map(BasicValueEnum::into_pointer_value)
             .unwrap();
 
@@ -258,7 +261,7 @@ impl TimeFns for NowPinningTimeFns {
         let time_lo = ctx.builder.build_int_truncate(time, i32_type, "time.lo").unwrap();
         let now_hiptr = ctx
             .builder
-            .build_bitcast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
+            .build_bit_cast(now, i32_type.ptr_type(AddressSpace::default()), "now.hi.addr")
             .map(BasicValueEnum::into_pointer_value)
             .unwrap();
 

nac3ast/Cargo.toml

@@ -10,7 +10,6 @@ constant-optimization = ["fold"]
 fold = []
 
 [dependencies]
-lazy_static = "1.4"
 parking_lot = "0.12"
 string-interner = "0.17"
 fxhash = "0.2"

nac3ast/src/ast_gen.rs

@@ -5,14 +5,12 @@ pub use crate::location::Location;
 
 use fxhash::FxBuildHasher;
 use parking_lot::{Mutex, MutexGuard};
-use std::{cell::RefCell, collections::HashMap, fmt};
+use std::{cell::RefCell, collections::HashMap, fmt, sync::LazyLock};
 use string_interner::{symbol::SymbolU32, DefaultBackend, StringInterner};
 
 pub type Interner = StringInterner<DefaultBackend, FxBuildHasher>;
-lazy_static! {
-    static ref INTERNER: Mutex<Interner> =
-        Mutex::new(StringInterner::with_hasher(FxBuildHasher::default()));
-}
+static INTERNER: LazyLock<Mutex<Interner>> =
+    LazyLock::new(|| Mutex::new(StringInterner::with_hasher(FxBuildHasher::default())));
 
 thread_local! {
     static LOCAL_INTERNER: RefCell<HashMap<String, StrRef>> = RefCell::default();

nac3ast/src/lib.rs

@@ -1,10 +1,4 @@
-#![deny(
-    future_incompatible,
-    let_underscore,
-    nonstandard_style,
-    rust_2024_compatibility,
-    clippy::all
-)]
+#![deny(future_incompatible, let_underscore, nonstandard_style, clippy::all)]
 #![warn(clippy::pedantic)]
 #![allow(
     clippy::missing_errors_doc,
@@ -14,9 +8,6 @@
     clippy::wildcard_imports
 )]
 
-#[macro_use]
-extern crate lazy_static;
-
 mod ast_gen;
 mod constant;
 #[cfg(feature = "fold")]

nac3core/Cargo.toml

@@ -4,20 +4,26 @@ version = "0.1.0"
 authors = ["M-Labs"]
 edition = "2021"
 
+[features]
+default = ["derive"]
+derive = ["dep:nac3core_derive"]
+no-escape-analysis = []
+
 [dependencies]
 itertools = "0.13"
 crossbeam = "0.8"
-indexmap = "2.2"
+indexmap = "2.6"
 parking_lot = "0.12"
-rayon = "1.8"
+rayon = "1.10"
+nac3core_derive = { path = "nac3core_derive", optional = true }
 nac3parser = { path = "../nac3parser" }
-strum = "0.26.2"
-strum_macros = "0.26.4"
+strum = "0.26"
+strum_macros = "0.26"
 
 [dependencies.inkwell]
-version = "0.4"
+version = "0.5"
 default-features = false
-features = ["llvm14-0", "target-x86", "target-arm", "target-riscv", "no-libffi-linking"]
+features = ["llvm14-0-prefer-dynamic", "target-x86", "target-arm", "target-riscv", "no-libffi-linking"]
 
 [dev-dependencies]
 test-case = "1.2.0"

nac3core/build.rs

@@ -1,4 +1,3 @@
-use regex::Regex;
 use std::{
     env,
     fs::File,
@@ -7,40 +6,58 @@ use std::{
     process::{Command, Stdio},
 };
 
+use regex::Regex;
+
 fn main() {
-    const FILE: &str = "src/codegen/irrt/irrt.c";
+    let out_dir = env::var("OUT_DIR").unwrap();
+    let out_dir = Path::new(&out_dir);
+    let irrt_dir = Path::new("irrt");
+
+    let irrt_cpp_path = irrt_dir.join("irrt.cpp");
 
     /*
      * HACK: Sadly, clang doesn't let us emit generic LLVM bitcode.
      * Compiling for WASM32 and filtering the output with regex is the closest we can get.
      */
-    let flags: &[&str] = &[
+    let mut flags: Vec<&str> = vec![
         "--target=wasm32",
-        FILE,
+        "-x",
+        "c++",
+        "-std=c++20",
         "-fno-discard-value-names",
-        match env::var("PROFILE").as_deref() {
-            Ok("debug") => "-O0",
-            Ok("release") => "-O3",
-            flavor => panic!("Unknown or missing build flavor {flavor:?}"),
-        },
+        "-fno-exceptions",
+        "-fno-rtti",
         "-emit-llvm",
         "-S",
         "-Wall",
         "-Wextra",
         "-o",
         "-",
+        "-I",
+        irrt_dir.to_str().unwrap(),
+        irrt_cpp_path.to_str().unwrap(),
     ];
 
-    println!("cargo:rerun-if-changed={FILE}");
-    let out_dir = env::var("OUT_DIR").unwrap();
-    let out_path = Path::new(&out_dir);
+    match env::var("PROFILE").as_deref() {
+        Ok("debug") => {
+            flags.push("-O0");
+            flags.push("-DIRRT_DEBUG_ASSERT");
+        }
+        Ok("release") => {
+            flags.push("-O3");
+        }
+        flavor => panic!("Unknown or missing build flavor {flavor:?}"),
+    }
 
+    // Tell Cargo to rerun if any file under `irrt_dir` (recursive) changes
+    println!("cargo:rerun-if-changed={}", irrt_dir.to_str().unwrap());
+
+    // Compile IRRT and capture the LLVM IR output
     let output = Command::new("clang-irrt")
         .args(flags)
         .output()
-        .map(|o| {
+        .inspect(|o| {
             assert!(o.status.success(), "{}", std::str::from_utf8(&o.stderr).unwrap());
-            o
         })
         .unwrap();
 
@@ -48,7 +65,17 @@ fn main() {
     let output = std::str::from_utf8(&output.stdout).unwrap().replace("\r\n", "\n");
     let mut filtered_output = String::with_capacity(output.len());
 
-    let regex_filter = Regex::new(r"(?ms:^define.*?\}$)|(?m:^declare.*?$)").unwrap();
+    // Filter out irrelevant IR
+    //
+    // Regex:
+    // - `(?ms:^define.*?\}$)` captures LLVM `define` blocks
+    // - `(?m:^declare.*?$)` captures LLVM `declare` lines
+    // - `(?m:^%.+?=\s*type\s*\{.+?\}$)` captures LLVM `type` declarations
+    // - `(?m:^@.+?=.+$)` captures global constants
+    let regex_filter = Regex::new(
+        r"(?ms:^define.*?\}$)|(?m:^declare.*?$)|(?m:^%.+?=\s*type\s*\{.+?\}$)|(?m:^@.+?=.+$)",
+    )
+    .unwrap();
     for f in regex_filter.captures_iter(&output) {
         assert_eq!(f.len(), 1);
         filtered_output.push_str(&f[0]);
@@ -59,18 +86,22 @@ fn main() {
         .unwrap()
         .replace_all(&filtered_output, "");
 
-    println!("cargo:rerun-if-env-changed=DEBUG_DUMP_IRRT");
-    if env::var("DEBUG_DUMP_IRRT").is_ok() {
-        let mut file = File::create(out_path.join("irrt.ll")).unwrap();
+    // For debugging
+    // Doing `DEBUG_DUMP_IRRT=1 cargo build -p nac3core` dumps the LLVM IR generated
+    const DEBUG_DUMP_IRRT: &str = "DEBUG_DUMP_IRRT";
+    println!("cargo:rerun-if-env-changed={DEBUG_DUMP_IRRT}");
+    if env::var(DEBUG_DUMP_IRRT).is_ok() {
+        let mut file = File::create(out_dir.join("irrt.ll")).unwrap();
         file.write_all(output.as_bytes()).unwrap();
-        let mut file = File::create(out_path.join("irrt-filtered.ll")).unwrap();
+
+        let mut file = File::create(out_dir.join("irrt-filtered.ll")).unwrap();
         file.write_all(filtered_output.as_bytes()).unwrap();
     }
 
     let mut llvm_as = Command::new("llvm-as-irrt")
         .stdin(Stdio::piped())
         .arg("-o")
-        .arg(out_path.join("irrt.bc"))
+        .arg(out_dir.join("irrt.bc"))
         .spawn()
         .unwrap();
     llvm_as.stdin.as_mut().unwrap().write_all(filtered_output.as_bytes()).unwrap();

nac3core/irrt/irrt.cpp

@@ -0,0 +1,5 @@
+#include "irrt/exception.hpp"
+#include "irrt/list.hpp"
+#include "irrt/math.hpp"
+#include "irrt/ndarray.hpp"
+#include "irrt/slice.hpp"

nac3core/irrt/irrt/cslice.hpp

@@ -0,0 +1,9 @@
+#pragma once
+
+#include "irrt/int_types.hpp"
+
+template<typename SizeT>
+struct CSlice {
+    void* base;
+    SizeT len;
+};

nac3core/irrt/irrt/debug.hpp

@@ -0,0 +1,25 @@
+#pragma once
+
+// Set in nac3core/build.rs
+#ifdef IRRT_DEBUG_ASSERT
+#define IRRT_DEBUG_ASSERT_BOOL true
+#else
+#define IRRT_DEBUG_ASSERT_BOOL false
+#endif
+
+#define raise_debug_assert(SizeT, msg, param1, param2, param3) \
+    raise_exception(SizeT, EXN_ASSERTION_ERROR, "IRRT debug assert failed: " msg, param1, param2, param3)
+
+#define debug_assert_eq(SizeT, lhs, rhs)                                           \
+    if constexpr (IRRT_DEBUG_ASSERT_BOOL) {                                        \
+        if ((lhs) != (rhs)) {                                                      \
+            raise_debug_assert(SizeT, "LHS = {0}. RHS = {1}", lhs, rhs, NO_PARAM); \
+        }                                                                          \
+    }
+
+#define debug_assert(SizeT, expr)                                                  \
+    if constexpr (IRRT_DEBUG_ASSERT_BOOL) {                                        \
+        if (!(expr)) {                                                             \
+            raise_debug_assert(SizeT, "Got false.", NO_PARAM, NO_PARAM, NO_PARAM); \
+        }                                                                          \
+    }

nac3core/irrt/irrt/exception.hpp

@@ -0,0 +1,85 @@
+#pragma once
+
+#include "irrt/cslice.hpp"
+#include "irrt/int_types.hpp"
+
+/**
+ * @brief The int type of ARTIQ exception IDs.
+ */
+using ExceptionId = int32_t;
+
+/*
+ * Set of exceptions C++ IRRT can use.
+ * Must be synchronized with `setup_irrt_exceptions` in `nac3core/src/codegen/irrt/mod.rs`.
+ */
+extern "C" {
+ExceptionId EXN_INDEX_ERROR;
+ExceptionId EXN_VALUE_ERROR;
+ExceptionId EXN_ASSERTION_ERROR;
+ExceptionId EXN_TYPE_ERROR;
+}
+
+/**
+ * @brief Extern function to `__nac3_raise`
+ *
+ * The parameter `err` could be `Exception<int32_t>` or `Exception<int64_t>`. The caller
+ * must make sure to pass `Exception`s with the correct `SizeT` depending on the `size_t` of the runtime.
+ */
+extern "C" void __nac3_raise(void* err);
+
+namespace {
+/**
+ * @brief NAC3's Exception struct
+ */
+template<typename SizeT>
+struct Exception {
+    ExceptionId id;
+    CSlice<SizeT> filename;
+    int32_t line;
+    int32_t column;
+    CSlice<SizeT> function;
+    CSlice<SizeT> msg;
+    int64_t params[3];
+};
+
+constexpr int64_t NO_PARAM = 0;
+
+template<typename SizeT>
+void _raise_exception_helper(ExceptionId id,
+                             const char* filename,
+                             int32_t line,
+                             const char* function,
+                             const char* msg,
+                             int64_t param0,
+                             int64_t param1,
+                             int64_t param2) {
+    Exception<SizeT> e = {
+        .id = id,
+        .filename = {.base = reinterpret_cast<void*>(const_cast<char*>(filename)),
+                     .len = static_cast<SizeT>(__builtin_strlen(filename))},
+        .line = line,
+        .column = 0,
+        .function = {.base = reinterpret_cast<void*>(const_cast<char*>(function)),
+                     .len = static_cast<SizeT>(__builtin_strlen(function))},
+        .msg = {.base = reinterpret_cast<void*>(const_cast<char*>(msg)),
+                .len = static_cast<SizeT>(__builtin_strlen(msg))},
+    };
+    e.params[0] = param0;
+    e.params[1] = param1;
+    e.params[2] = param2;
+    __nac3_raise(reinterpret_cast<void*>(&e));
+    __builtin_unreachable();
+}
+}  // namespace
+
+/**
+ * @brief Raise an exception with location details (location in the IRRT source files).
+ * @param SizeT The runtime `size_t` type.
+ * @param id The ID of the exception to raise.
+ * @param msg A global constant C-string of the error message.
+ *
+ * `param0` to `param2` are optional format arguments of `msg`. They should be set to
+ * `NO_PARAM` to indicate they are unused.
+ */
+#define raise_exception(SizeT, id, msg, param0, param1, param2) \
+    _raise_exception_helper<SizeT>(id, __FILE__, __LINE__, __FUNCTION__, msg, param0, param1, param2)

nac3core/irrt/irrt/int_types.hpp

@@ -0,0 +1,27 @@
+#pragma once
+
+#if __STDC_VERSION__ >= 202000
+using int8_t = _BitInt(8);
+using uint8_t = unsigned _BitInt(8);
+using int32_t = _BitInt(32);
+using uint32_t = unsigned _BitInt(32);
+using int64_t = _BitInt(64);
+using uint64_t = unsigned _BitInt(64);
+#else
+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Wdeprecated-type"
+using int8_t = _ExtInt(8);
+using uint8_t = unsigned _ExtInt(8);
+using int32_t = _ExtInt(32);
+using uint32_t = unsigned _ExtInt(32);
+using int64_t = _ExtInt(64);
+using uint64_t = unsigned _ExtInt(64);
+#pragma clang diagnostic pop
+
+#endif
+
+// NDArray indices are always `uint32_t`.
+using NDIndex = uint32_t;
+// The type of an index or a value describing the length of a range/slice is always `int32_t`.
+using SliceIndex = int32_t;

nac3core/irrt/irrt/list.hpp

@@ -0,0 +1,81 @@
+#pragma once
+
+#include "irrt/int_types.hpp"
+#include "irrt/math_util.hpp"
+
+extern "C" {
+// Handle list assignment and dropping part of the list when
+// both dest_step and src_step are +1.
+// - All the index must *not* be out-of-bound or negative,
+// - The end index is *inclusive*,
+// - The length of src and dest slice size should already
+// be checked: if dest.step == 1 then len(src) <= len(dest) else len(src) == len(dest)
+SliceIndex __nac3_list_slice_assign_var_size(SliceIndex dest_start,
+                                             SliceIndex dest_end,
+                                             SliceIndex dest_step,
+                                             void* dest_arr,
+                                             SliceIndex dest_arr_len,
+                                             SliceIndex src_start,
+                                             SliceIndex src_end,
+                                             SliceIndex src_step,
+                                             void* src_arr,
+                                             SliceIndex src_arr_len,
+                                             const SliceIndex size) {
+    /* if dest_arr_len == 0, do nothing since we do not support extending list */
+    if (dest_arr_len == 0)
+        return dest_arr_len;
+    /* if both step is 1, memmove directly, handle the dropping of the list, and shrink size */
+    if (src_step == dest_step && dest_step == 1) {
+        const SliceIndex src_len = (src_end >= src_start) ? (src_end - src_start + 1) : 0;
+        const SliceIndex dest_len = (dest_end >= dest_start) ? (dest_end - dest_start + 1) : 0;
+        if (src_len > 0) {
+            __builtin_memmove(static_cast<uint8_t*>(dest_arr) + dest_start * size,
+                              static_cast<uint8_t*>(src_arr) + src_start * size, src_len * size);
+        }
+        if (dest_len > 0) {
+            /* dropping */
+            __builtin_memmove(static_cast<uint8_t*>(dest_arr) + (dest_start + src_len) * size,
+                              static_cast<uint8_t*>(dest_arr) + (dest_end + 1) * size,
+                              (dest_arr_len - dest_end - 1) * size);
+        }
+        /* shrink size */
+        return dest_arr_len - (dest_len - src_len);
+    }
+    /* if two range overlaps, need alloca */
+    uint8_t need_alloca = (dest_arr == src_arr)
+                          && !(max(dest_start, dest_end) < min(src_start, src_end)
+                               || max(src_start, src_end) < min(dest_start, dest_end));
+    if (need_alloca) {
+        void* tmp = __builtin_alloca(src_arr_len * size);
+        __builtin_memcpy(tmp, src_arr, src_arr_len * size);
+        src_arr = tmp;
+    }
+    SliceIndex src_ind = src_start;
+    SliceIndex dest_ind = dest_start;
+    for (; (src_step > 0) ? (src_ind <= src_end) : (src_ind >= src_end); src_ind += src_step, dest_ind += dest_step) {
+        /* for constant optimization */
+        if (size == 1) {
+            __builtin_memcpy(static_cast<uint8_t*>(dest_arr) + dest_ind, static_cast<uint8_t*>(src_arr) + src_ind, 1);
+        } else if (size == 4) {
+            __builtin_memcpy(static_cast<uint8_t*>(dest_arr) + dest_ind * 4,
+                             static_cast<uint8_t*>(src_arr) + src_ind * 4, 4);
+        } else if (size == 8) {
+            __builtin_memcpy(static_cast<uint8_t*>(dest_arr) + dest_ind * 8,
+                             static_cast<uint8_t*>(src_arr) + src_ind * 8, 8);
+        } else {
+            /* memcpy for var size, cannot overlap after previous alloca */
+            __builtin_memcpy(static_cast<uint8_t*>(dest_arr) + dest_ind * size,
+                             static_cast<uint8_t*>(src_arr) + src_ind * size, size);
+        }
+    }
+    /* only dest_step == 1 can we shrink the dest list. */
+    /* size should be ensured prior to calling this function */
+    if (dest_step == 1 && dest_end >= dest_start) {
+        __builtin_memmove(static_cast<uint8_t*>(dest_arr) + dest_ind * size,
+                          static_cast<uint8_t*>(dest_arr) + (dest_end + 1) * size,
+                          (dest_arr_len - dest_end - 1) * size);
+        return dest_arr_len - (dest_end - dest_ind) - 1;
+    }
+    return dest_arr_len;
+}
+}  // extern "C"

nac3core/irrt/irrt/math.hpp

@@ -0,0 +1,93 @@
+#pragma once
+
+namespace {
+// adapted from GNU Scientific Library: https://git.savannah.gnu.org/cgit/gsl.git/tree/sys/pow_int.c
+// need to make sure `exp >= 0` before calling this function
+template<typename T>
+T __nac3_int_exp_impl(T base, T exp) {
+    T res = 1;
+    /* repeated squaring method */
+    do {
+        if (exp & 1) {
+            res *= base; /* for n odd */
+        }
+        exp >>= 1;
+        base *= base;
+    } while (exp);
+    return res;
+}
+}  // namespace
+
+#define DEF_nac3_int_exp_(T)                   \
+    T __nac3_int_exp_##T(T base, T exp) {      \
+        return __nac3_int_exp_impl(base, exp); \
+    }
+
+extern "C" {
+
+// Putting semicolons here to make clang-format not reformat this into
+// a stair shape.
+DEF_nac3_int_exp_(int32_t);
+DEF_nac3_int_exp_(int64_t);
+DEF_nac3_int_exp_(uint32_t);
+DEF_nac3_int_exp_(uint64_t);
+
+int32_t __nac3_isinf(double x) {
+    return __builtin_isinf(x);
+}
+
+int32_t __nac3_isnan(double x) {
+    return __builtin_isnan(x);
+}
+
+double tgamma(double arg);
+
+double __nac3_gamma(double z) {
+    // Handling for denormals
+    //     | x                 | Python gamma(x) | C tgamma(x) |
+    // --- | ----------------- | --------------- | ----------- |
+    // (1) | nan               | nan             | nan         |
+    // (2) | -inf              | -inf            | inf         |
+    // (3) | inf               | inf             | inf         |
+    // (4) | 0.0               | inf             | inf         |
+    // (5) | {-1.0, -2.0, ...} | inf             | nan         |
+
+    // (1)-(3)
+    if (__builtin_isinf(z) || __builtin_isnan(z)) {
+        return z;
+    }
+
+    double v = tgamma(z);
+
+    // (4)-(5)
+    return __builtin_isinf(v) || __builtin_isnan(v) ? __builtin_inf() : v;
+}
+
+double lgamma(double arg);
+
+double __nac3_gammaln(double x) {
+    // libm's handling of value overflows differs from scipy:
+    // - scipy: gammaln(-inf) -> -inf
+    // - libm : lgamma(-inf) -> inf
+
+    if (__builtin_isinf(x)) {
+        return x;
+    }
+
+    return lgamma(x);
+}
+
+double j0(double x);
+
+double __nac3_j0(double x) {
+    // libm's handling of value overflows differs from scipy:
+    // - scipy: j0(inf) -> nan
+    // - libm : j0(inf) -> 0.0
+
+    if (__builtin_isinf(x)) {
+        return __builtin_nan("");
+    }
+
+    return j0(x);
+}
+}  // namespace

nac3core/irrt/irrt/math_util.hpp

@@ -0,0 +1,13 @@
+#pragma once
+
+namespace {
+template<typename T>
+const T& max(const T& a, const T& b) {
+    return a > b ? a : b;
+}
+
+template<typename T>
+const T& min(const T& a, const T& b) {
+    return a > b ? b : a;
+}
+}  // namespace

nac3core/irrt/irrt/ndarray.hpp

@@ -0,0 +1,144 @@
+#pragma once
+
+#include "irrt/int_types.hpp"
+
+namespace {
+template<typename SizeT>
+SizeT __nac3_ndarray_calc_size_impl(const SizeT* list_data, SizeT list_len, SizeT begin_idx, SizeT end_idx) {
+    __builtin_assume(end_idx <= list_len);
+
+    SizeT num_elems = 1;
+    for (SizeT i = begin_idx; i < end_idx; ++i) {
+        SizeT val = list_data[i];
+        __builtin_assume(val > 0);
+        num_elems *= val;
+    }
+    return num_elems;
+}
+
+template<typename SizeT>
+void __nac3_ndarray_calc_nd_indices_impl(SizeT index, const SizeT* dims, SizeT num_dims, NDIndex* idxs) {
+    SizeT stride = 1;
+    for (SizeT dim = 0; dim < num_dims; dim++) {
+        SizeT i = num_dims - dim - 1;
+        __builtin_assume(dims[i] > 0);
+        idxs[i] = (index / stride) % dims[i];
+        stride *= dims[i];
+    }
+}
+
+template<typename SizeT>
+SizeT __nac3_ndarray_flatten_index_impl(const SizeT* dims, SizeT num_dims, const NDIndex* indices, SizeT num_indices) {
+    SizeT idx = 0;
+    SizeT stride = 1;
+    for (SizeT i = 0; i < num_dims; ++i) {
+        SizeT ri = num_dims - i - 1;
+        if (ri < num_indices) {
+            idx += stride * indices[ri];
+        }
+
+        __builtin_assume(dims[i] > 0);
+        stride *= dims[ri];
+    }
+    return idx;
+}
+
+template<typename SizeT>
+void __nac3_ndarray_calc_broadcast_impl(const SizeT* lhs_dims,
+                                        SizeT lhs_ndims,
+                                        const SizeT* rhs_dims,
+                                        SizeT rhs_ndims,
+                                        SizeT* out_dims) {
+    SizeT max_ndims = lhs_ndims > rhs_ndims ? lhs_ndims : rhs_ndims;
+
+    for (SizeT i = 0; i < max_ndims; ++i) {
+        const SizeT* lhs_dim_sz = i < lhs_ndims ? &lhs_dims[lhs_ndims - i - 1] : nullptr;
+        const SizeT* rhs_dim_sz = i < rhs_ndims ? &rhs_dims[rhs_ndims - i - 1] : nullptr;
+        SizeT* out_dim = &out_dims[max_ndims - i - 1];
+
+        if (lhs_dim_sz == nullptr) {
+            *out_dim = *rhs_dim_sz;
+        } else if (rhs_dim_sz == nullptr) {
+            *out_dim = *lhs_dim_sz;
+        } else if (*lhs_dim_sz == 1) {
+            *out_dim = *rhs_dim_sz;
+        } else if (*rhs_dim_sz == 1) {
+            *out_dim = *lhs_dim_sz;
+        } else if (*lhs_dim_sz == *rhs_dim_sz) {
+            *out_dim = *lhs_dim_sz;
+        } else {
+            __builtin_unreachable();
+        }
+    }
+}
+
+template<typename SizeT>
+void __nac3_ndarray_calc_broadcast_idx_impl(const SizeT* src_dims,
+                                            SizeT src_ndims,
+                                            const NDIndex* in_idx,
+                                            NDIndex* out_idx) {
+    for (SizeT i = 0; i < src_ndims; ++i) {
+        SizeT src_i = src_ndims - i - 1;
+        out_idx[src_i] = src_dims[src_i] == 1 ? 0 : in_idx[src_i];
+    }
+}
+}  // namespace
+
+extern "C" {
+uint32_t __nac3_ndarray_calc_size(const uint32_t* list_data, uint32_t list_len, uint32_t begin_idx, uint32_t end_idx) {
+    return __nac3_ndarray_calc_size_impl(list_data, list_len, begin_idx, end_idx);
+}
+
+uint64_t
+__nac3_ndarray_calc_size64(const uint64_t* list_data, uint64_t list_len, uint64_t begin_idx, uint64_t end_idx) {
+    return __nac3_ndarray_calc_size_impl(list_data, list_len, begin_idx, end_idx);
+}
+
+void __nac3_ndarray_calc_nd_indices(uint32_t index, const uint32_t* dims, uint32_t num_dims, NDIndex* idxs) {
+    __nac3_ndarray_calc_nd_indices_impl(index, dims, num_dims, idxs);
+}
+
+void __nac3_ndarray_calc_nd_indices64(uint64_t index, const uint64_t* dims, uint64_t num_dims, NDIndex* idxs) {
+    __nac3_ndarray_calc_nd_indices_impl(index, dims, num_dims, idxs);
+}
+
+uint32_t
+__nac3_ndarray_flatten_index(const uint32_t* dims, uint32_t num_dims, const NDIndex* indices, uint32_t num_indices) {
+    return __nac3_ndarray_flatten_index_impl(dims, num_dims, indices, num_indices);
+}
+
+uint64_t
+__nac3_ndarray_flatten_index64(const uint64_t* dims, uint64_t num_dims, const NDIndex* indices, uint64_t num_indices) {
+    return __nac3_ndarray_flatten_index_impl(dims, num_dims, indices, num_indices);
+}
+
+void __nac3_ndarray_calc_broadcast(const uint32_t* lhs_dims,
+                                   uint32_t lhs_ndims,
+                                   const uint32_t* rhs_dims,
+                                   uint32_t rhs_ndims,
+                                   uint32_t* out_dims) {
+    return __nac3_ndarray_calc_broadcast_impl(lhs_dims, lhs_ndims, rhs_dims, rhs_ndims, out_dims);
+}
+
+void __nac3_ndarray_calc_broadcast64(const uint64_t* lhs_dims,
+                                     uint64_t lhs_ndims,
+                                     const uint64_t* rhs_dims,
+                                     uint64_t rhs_ndims,
+                                     uint64_t* out_dims) {
+    return __nac3_ndarray_calc_broadcast_impl(lhs_dims, lhs_ndims, rhs_dims, rhs_ndims, out_dims);
+}
+
+void __nac3_ndarray_calc_broadcast_idx(const uint32_t* src_dims,
+                                       uint32_t src_ndims,
+                                       const NDIndex* in_idx,
+                                       NDIndex* out_idx) {
+    __nac3_ndarray_calc_broadcast_idx_impl(src_dims, src_ndims, in_idx, out_idx);
+}
+
+void __nac3_ndarray_calc_broadcast_idx64(const uint64_t* src_dims,
+                                         uint64_t src_ndims,
+                                         const NDIndex* in_idx,
+                                         NDIndex* out_idx) {
+    __nac3_ndarray_calc_broadcast_idx_impl(src_dims, src_ndims, in_idx, out_idx);
+}
+}  // namespace

nac3core/irrt/irrt/slice.hpp

@@ -0,0 +1,28 @@
+#pragma once
+
+#include "irrt/int_types.hpp"
+
+extern "C" {
+SliceIndex __nac3_slice_index_bound(SliceIndex i, const SliceIndex len) {
+    if (i < 0) {
+        i = len + i;
+    }
+    if (i < 0) {
+        return 0;
+    } else if (i > len) {
+        return len;
+    }
+    return i;
+}
+
+SliceIndex __nac3_range_slice_len(const SliceIndex start, const SliceIndex end, const SliceIndex step) {
+    SliceIndex diff = end - start;
+    if (diff > 0 && step > 0) {
+        return ((diff - 1) / step) + 1;
+    } else if (diff < 0 && step < 0) {
+        return ((diff + 1) / step) + 1;
+    } else {
+        return 0;
+    }
+}
+}  // namespace

nac3core/nac3core_derive/Cargo.toml

@@ -0,0 +1,21 @@
+[package]
+name = "nac3core_derive"
+version = "0.1.0"
+edition = "2021"
+
+[lib]
+proc-macro = true
+
+[[test]]
+name = "structfields_tests"
+path = "tests/structfields_test.rs"
+
+[dev-dependencies]
+nac3core = { path = ".." }
+trybuild = { version = "1.0", features = ["diff"] }
+
+[dependencies]
+proc-macro2 = "1.0"
+proc-macro-error = "1.0"
+syn = "2.0"
+quote = "1.0"

nac3core/nac3core_derive/src/lib.rs

@@ -0,0 +1,320 @@
+use proc_macro::TokenStream;
+use proc_macro_error::{abort, proc_macro_error};
+use quote::quote;
+use syn::{
+    parse_macro_input, spanned::Spanned, Data, DataStruct, Expr, ExprField, ExprMethodCall,
+    ExprPath, GenericArgument, Ident, LitStr, Path, PathArguments, Type, TypePath,
+};
+
+/// Extracts all generic arguments of a [`Type`] into a [`Vec`].
+///
+/// Returns [`Some`] of a possibly-empty [`Vec`] if the path of `ty` matches with
+/// `expected_ty_name`, otherwise returns [`None`].
+fn extract_generic_args(expected_ty_name: &'static str, ty: &Type) -> Option<Vec<GenericArgument>> {
+    let Type::Path(TypePath { qself: None, path, .. }) = ty else {
+        return None;
+    };
+
+    let segments = &path.segments;
+    if segments.len() != 1 {
+        return None;
+    };
+
+    let segment = segments.iter().next().unwrap();
+    if segment.ident != expected_ty_name {
+        return None;
+    }
+
+    let PathArguments::AngleBracketed(path_args) = &segment.arguments else {
+        return Some(Vec::new());
+    };
+    let args = &path_args.args;
+
+    Some(args.iter().cloned().collect::<Vec<_>>())
+}
+
+/// Maps a `path` matching one of the `target_idents` into the `replacement` [`Ident`].
+fn map_path_to_ident(path: &Path, target_idents: &[&str], replacement: &str) -> Option<Ident> {
+    path.require_ident()
+        .ok()
+        .filter(|ident| target_idents.iter().any(|target| ident == target))
+        .map(|ident| Ident::new(replacement, ident.span()))
+}
+
+/// Extracts the left-hand side of a dot-expression.
+fn extract_dot_operand(expr: &Expr) -> Option<&Expr> {
+    match expr {
+        Expr::MethodCall(ExprMethodCall { receiver: operand, .. })
+        | Expr::Field(ExprField { base: operand, .. }) => Some(operand),
+        _ => None,
+    }
+}
+
+/// Replaces the top-level receiver of a dot-expression with an [`Ident`], returning `Some(&mut expr)` if the
+/// replacement is performed.
+///
+/// The top-level receiver is the left-most receiver expression, e.g. the top-level receiver of `a.b.c.foo()` is `a`.
+fn replace_top_level_receiver(expr: &mut Expr, ident: Ident) -> Option<&mut Expr> {
+    if let Expr::MethodCall(ExprMethodCall { receiver: operand, .. })
+    | Expr::Field(ExprField { base: operand, .. }) = expr
+    {
+        return if extract_dot_operand(operand).is_some() {
+            if replace_top_level_receiver(operand, ident).is_some() {
+                Some(expr)
+            } else {
+                None
+            }
+        } else {
+            *operand = Box::new(Expr::Path(ExprPath {
+                attrs: Vec::default(),
+                qself: None,
+                path: ident.into(),
+            }));
+
+            Some(expr)
+        };
+    }
+
+    None
+}
+
+/// Iterates all operands to the left-hand side of the `.` of an [expression][`Expr`], i.e. the container operand of all
+/// [`Expr::Field`] and the receiver operand of all [`Expr::MethodCall`].
+///
+/// The iterator will return the operand expressions in reverse order of appearance. For example, `a.b.c.func()` will
+/// return `vec![c, b, a]`.
+fn iter_dot_operands(expr: &Expr) -> impl Iterator<Item = &Expr> {
+    let mut o = extract_dot_operand(expr);
+
+    std::iter::from_fn(move || {
+        let this = o;
+        o = o.as_ref().and_then(|o| extract_dot_operand(o));
+
+        this
+    })
+}
+
+/// Normalizes a value expression for use when creating an instance of this structure, returning a
+/// [`proc_macro2::TokenStream`] of tokens representing the normalized expression.
+fn normalize_value_expr(expr: &Expr) -> proc_macro2::TokenStream {
+    match &expr {
+        Expr::Path(ExprPath { qself: None, path, .. }) => {
+            if let Some(ident) = map_path_to_ident(path, &["usize", "size_t"], "llvm_usize") {
+                quote! { #ident }
+            } else {
+                abort!(
+                    path,
+                    format!(
+                        "Expected one of `size_t`, `usize`, or an implicit call expression in #[value_type(...)], found {}", 
+                        quote!(#expr).to_string(),
+                    )
+                )
+            }
+        }
+
+        Expr::Call(_) => {
+            quote! { ctx.#expr }
+        }
+
+        Expr::MethodCall(_) => {
+            let base_receiver = iter_dot_operands(expr).last();
+
+            match base_receiver {
+                // `usize.{...}`, `size_t.{...}` -> Rewrite the identifiers to `llvm_usize`
+                Some(Expr::Path(ExprPath { qself: None, path, .. }))
+                    if map_path_to_ident(path, &["usize", "size_t"], "llvm_usize").is_some() =>
+                {
+                    let ident =
+                        map_path_to_ident(path, &["usize", "size_t"], "llvm_usize").unwrap();
+
+                    let mut expr = expr.clone();
+                    let expr = replace_top_level_receiver(&mut expr, ident).unwrap();
+
+                    quote!(#expr)
+                }
+
+                // `ctx.{...}`, `context.{...}` -> Rewrite the identifiers to `ctx`
+                Some(Expr::Path(ExprPath { qself: None, path, .. }))
+                    if map_path_to_ident(path, &["ctx", "context"], "ctx").is_some() =>
+                {
+                    let ident = map_path_to_ident(path, &["ctx", "context"], "ctx").unwrap();
+
+                    let mut expr = expr.clone();
+                    let expr = replace_top_level_receiver(&mut expr, ident).unwrap();
+
+                    quote!(#expr)
+                }
+
+                // No reserved identifier prefix -> Prepend `ctx.` to the entire expression
+                _ => quote! { ctx.#expr },
+            }
+        }
+
+        _ => {
+            abort!(
+                expr,
+                format!(
+                    "Expected one of `size_t`, `usize`, or an implicit call expression in #[value_type(...)], found {}", 
+                    quote!(#expr).to_string(),
+                )
+            )
+        }
+    }
+}
+
+/// Derives an implementation of `codegen::types::structure::StructFields`.
+///
+/// The benefit of using `#[derive(StructFields)]` is that all index- or order-dependent logic required by
+/// `impl StructFields` is automatically generated by this implementation, including the field index as required by
+/// `StructField::new` and the fields as returned by `StructFields::to_vec`.
+///
+/// # Prerequisites
+///
+/// In order to derive from [`StructFields`], you must implement (or derive) [`Eq`] and [`Copy`] as required by
+/// `StructFields`.
+///
+/// Moreover, `#[derive(StructFields)]` can only be used for `struct`s with named fields, and may only contain fields
+/// with either `StructField` or [`PhantomData`] types.
+///
+/// # Attributes for [`StructFields`]
+///
+/// Each `StructField` field must be declared with the `#[value_type(...)]` attribute. The argument of `value_type`
+/// accepts one of the following:
+///
+/// - An expression returning an instance of `inkwell::types::BasicType` (with or without the receiver `ctx`/`context`).
+/// For example, `context.i8_type()`, `ctx.i8_type()`, and `i8_type()` all refer to `i8`.
+/// - The reserved identifiers `usize` and `size_t` referring to an `inkwell::types::IntType` of the platform-dependent
+/// integer size. `usize` and `size_t` can also be used as the receiver to other method calls, e.g.
+/// `usize.array_type(3)`.
+///
+/// # Example
+///
+/// The following is an example of an LLVM slice implemented using `#[derive(StructFields)]`.
+///
+/// ```rust,ignore
+/// use nac3core::{
+///     codegen::types::structure::StructField,
+///     inkwell::{
+///         values::{IntValue, PointerValue},
+///         AddressSpace,
+///     },
+/// };
+/// use nac3core_derive::StructFields;
+///
+/// // All classes that implement StructFields must also implement Eq and Copy
+/// #[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+/// pub struct SliceValue<'ctx> {
+///     // Declares ptr have a value type of i8*
+///     //
+///     // Can also be written as `ctx.i8_type().ptr_type(...)` or `context.i8_type().ptr_type(...)`
+///     #[value_type(i8_type().ptr_type(AddressSpace::default()))]
+///     ptr: StructField<'ctx, PointerValue<'ctx>>,
+///
+///     // Declares len have a value type of usize, depending on the target compilation platform
+///     #[value_type(usize)]
+///     len: StructField<'ctx, IntValue<'ctx>>,
+/// }
+/// ```
+#[proc_macro_derive(StructFields, attributes(value_type))]
+#[proc_macro_error]
+pub fn derive(input: TokenStream) -> TokenStream {
+    let input = parse_macro_input!(input as syn::DeriveInput);
+    let ident = &input.ident;
+
+    let Data::Struct(DataStruct { fields, .. }) = &input.data else {
+        abort!(input, "Only structs with named fields are supported");
+    };
+    if let Err(err_span) =
+        fields
+            .iter()
+            .try_for_each(|field| if field.ident.is_some() { Ok(()) } else { Err(field.span()) })
+    {
+        abort!(err_span, "Only structs with named fields are supported");
+    };
+
+    // Check if struct<'ctx>
+    if input.generics.params.len() != 1 {
+        abort!(input.generics, "Expected exactly 1 generic parameter")
+    }
+
+    let phantom_info = fields
+        .iter()
+        .filter(|field| extract_generic_args("PhantomData", &field.ty).is_some())
+        .map(|field| field.ident.as_ref().unwrap())
+        .cloned()
+        .collect::<Vec<_>>();
+
+    let field_info = fields
+        .iter()
+        .filter(|field| extract_generic_args("PhantomData", &field.ty).is_none())
+        .map(|field| {
+            let ident = field.ident.as_ref().unwrap();
+            let ty = &field.ty;
+
+            let Some(_) = extract_generic_args("StructField", ty) else {
+                abort!(field, "Only StructField and PhantomData are allowed")
+            };
+
+            let attrs = &field.attrs;
+            let Some(value_type_attr) =
+                attrs.iter().find(|attr| attr.path().is_ident("value_type"))
+            else {
+                abort!(field, "Expected #[value_type(...)] attribute for field");
+            };
+
+            let Ok(value_type_expr) = value_type_attr.parse_args::<Expr>() else {
+                abort!(value_type_attr, "Expected expression in #[value_type(...)]");
+            };
+
+            let value_expr_toks = normalize_value_expr(&value_type_expr);
+
+            (ident.clone(), value_expr_toks)
+        })
+        .collect::<Vec<_>>();
+
+    // `<*>::new` impl of `StructField` and `PhantomData` for `StructFields::new`
+    let phantoms_create = phantom_info
+        .iter()
+        .map(|id| quote! { #id: ::std::marker::PhantomData })
+        .collect::<Vec<_>>();
+    let fields_create = field_info
+        .iter()
+        .map(|(id, ty)| {
+            let id_lit = LitStr::new(&id.to_string(), id.span());
+            quote! {
+                #id: ::nac3core::codegen::types::structure::StructField::create(
+                    &mut counter,
+                    #id_lit,
+                    #ty,
+                )
+            }
+        })
+        .collect::<Vec<_>>();
+
+    // `.into()` impl of `StructField` for `StructFields::to_vec`
+    let fields_into =
+        field_info.iter().map(|(id, _)| quote! { self.#id.into() }).collect::<Vec<_>>();
+
+    let impl_block = quote! {
+        impl<'ctx> ::nac3core::codegen::types::structure::StructFields<'ctx> for #ident<'ctx> {
+            fn new(ctx: impl ::nac3core::inkwell::context::AsContextRef<'ctx>, llvm_usize: ::nac3core::inkwell::types::IntType<'ctx>) -> Self {
+                let ctx = unsafe { ::nac3core::inkwell::context::ContextRef::new(ctx.as_ctx_ref()) };
+
+                let mut counter = ::nac3core::codegen::types::structure::FieldIndexCounter::default();
+
+                #ident {
+                    #(#fields_create),*
+                    #(#phantoms_create),*
+                }
+            }
+
+            fn to_vec(&self) -> ::std::vec::Vec<(&'static str, ::nac3core::inkwell::types::BasicTypeEnum<'ctx>)> {
+                vec![
+                    #(#fields_into),*
+                ]
+            }
+        }
+    };
+
+    impl_block.into()
+}

nac3core/nac3core_derive/tests/structfields_empty.rs

@@ -0,0 +1,9 @@
+use nac3core_derive::StructFields;
+use std::marker::PhantomData;
+
+#[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+pub struct EmptyValue<'ctx> {
+    _phantom: PhantomData<&'ctx ()>,
+}
+
+fn main() {}

nac3core/nac3core_derive/tests/structfields_ndarray.rs

@@ -0,0 +1,20 @@
+use nac3core::{
+    codegen::types::structure::StructField,
+    inkwell::{
+        values::{IntValue, PointerValue},
+        AddressSpace,
+    },
+};
+use nac3core_derive::StructFields;
+
+#[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+pub struct NDArrayValue<'ctx> {
+    #[value_type(usize)]
+    ndims: StructField<'ctx, IntValue<'ctx>>,
+    #[value_type(usize.ptr_type(AddressSpace::default()))]
+    shape: StructField<'ctx, PointerValue<'ctx>>,
+    #[value_type(i8_type().ptr_type(AddressSpace::default()))]
+    data: StructField<'ctx, PointerValue<'ctx>>,
+}
+
+fn main() {}

nac3core/nac3core_derive/tests/structfields_slice.rs

@@ -0,0 +1,18 @@
+use nac3core::{
+    codegen::types::structure::StructField,
+    inkwell::{
+        values::{IntValue, PointerValue},
+        AddressSpace,
+    },
+};
+use nac3core_derive::StructFields;
+
+#[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+pub struct SliceValue<'ctx> {
+    #[value_type(i8_type().ptr_type(AddressSpace::default()))]
+    ptr: StructField<'ctx, PointerValue<'ctx>>,
+    #[value_type(usize)]
+    len: StructField<'ctx, IntValue<'ctx>>,
+}
+
+fn main() {}

nac3core/nac3core_derive/tests/structfields_slice_context.rs

@@ -0,0 +1,18 @@
+use nac3core::{
+    codegen::types::structure::StructField,
+    inkwell::{
+        values::{IntValue, PointerValue},
+        AddressSpace,
+    },
+};
+use nac3core_derive::StructFields;
+
+#[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+pub struct SliceValue<'ctx> {
+    #[value_type(context.i8_type().ptr_type(AddressSpace::default()))]
+    ptr: StructField<'ctx, PointerValue<'ctx>>,
+    #[value_type(usize)]
+    len: StructField<'ctx, IntValue<'ctx>>,
+}
+
+fn main() {}

nac3core/nac3core_derive/tests/structfields_slice_ctx.rs

@@ -0,0 +1,18 @@
+use nac3core::{
+    codegen::types::structure::StructField,
+    inkwell::{
+        values::{IntValue, PointerValue},
+        AddressSpace,
+    },
+};
+use nac3core_derive::StructFields;
+
+#[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+pub struct SliceValue<'ctx> {
+    #[value_type(ctx.i8_type().ptr_type(AddressSpace::default()))]
+    ptr: StructField<'ctx, PointerValue<'ctx>>,
+    #[value_type(usize)]
+    len: StructField<'ctx, IntValue<'ctx>>,
+}
+
+fn main() {}

nac3core/nac3core_derive/tests/structfields_slice_sizet.rs

@@ -0,0 +1,18 @@
+use nac3core::{
+    codegen::types::structure::StructField,
+    inkwell::{
+        values::{IntValue, PointerValue},
+        AddressSpace,
+    },
+};
+use nac3core_derive::StructFields;
+
+#[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+pub struct SliceValue<'ctx> {
+    #[value_type(i8_type().ptr_type(AddressSpace::default()))]
+    ptr: StructField<'ctx, PointerValue<'ctx>>,
+    #[value_type(size_t)]
+    len: StructField<'ctx, IntValue<'ctx>>,
+}
+
+fn main() {}

nac3core/nac3core_derive/tests/structfields_test.rs

@@ -0,0 +1,10 @@
+#[test]
+fn test_parse_empty() {
+    let t = trybuild::TestCases::new();
+    t.pass("tests/structfields_empty.rs");
+    t.pass("tests/structfields_slice.rs");
+    t.pass("tests/structfields_slice_ctx.rs");
+    t.pass("tests/structfields_slice_context.rs");
+    t.pass("tests/structfields_slice_sizet.rs");
+    t.pass("tests/structfields_ndarray.rs");
+}

nac3core/src/codegen/concrete_type.rs

@@ -1,3 +1,9 @@
+use std::collections::HashMap;
+
+use indexmap::IndexMap;
+
+use nac3parser::ast::StrRef;
+
 use crate::{
     symbol_resolver::SymbolValue,
     toplevel::DefinitionId,
@@ -9,10 +15,6 @@ use crate::{
     },
 };
 
-use indexmap::IndexMap;
-use nac3parser::ast::StrRef;
-use std::collections::HashMap;
-
 pub struct ConcreteTypeStore {
     store: Vec<ConcreteTypeEnum>,
 }
@@ -25,6 +27,7 @@ pub struct ConcreteFuncArg {
     pub name: StrRef,
     pub ty: ConcreteType,
     pub default_value: Option<SymbolValue>,
+    pub is_vararg: bool,
 }
 
 #[derive(Clone, Debug)]
@@ -46,9 +49,7 @@ pub enum ConcreteTypeEnum {
     TPrimitive(Primitive),
     TTuple {
         ty: Vec<ConcreteType>,
-    },
-    TList {
-        ty: ConcreteType,
+        is_vararg_ctx: bool,
     },
     TObj {
         obj_id: DefinitionId,
@@ -105,8 +106,16 @@ impl ConcreteTypeStore {
                 .iter()
                 .map(|arg| ConcreteFuncArg {
                     name: arg.name,
-                    ty: self.from_unifier_type(unifier, primitives, arg.ty, cache),
+                    ty: if arg.is_vararg {
+                        let tuple_ty = unifier
+                            .add_ty(TypeEnum::TTuple { ty: vec![arg.ty], is_vararg_ctx: true });
+
+                        self.from_unifier_type(unifier, primitives, tuple_ty, cache)
+                    } else {
+                        self.from_unifier_type(unifier, primitives, arg.ty, cache)
+                    },
                     default_value: arg.default_value.clone(),
+                    is_vararg: arg.is_vararg,
                 })
                 .collect(),
             ret: self.from_unifier_type(unifier, primitives, signature.ret, cache),
@@ -161,14 +170,12 @@ impl ConcreteTypeStore {
             cache.insert(ty, None);
             let ty_enum = unifier.get_ty(ty);
             let result = match &*ty_enum {
-                TypeEnum::TTuple { ty } => ConcreteTypeEnum::TTuple {
+                TypeEnum::TTuple { ty, is_vararg_ctx } => ConcreteTypeEnum::TTuple {
                     ty: ty
                         .iter()
                         .map(|t| self.from_unifier_type(unifier, primitives, *t, cache))
                         .collect(),
-                },
-                TypeEnum::TList { ty } => ConcreteTypeEnum::TList {
-                    ty: self.from_unifier_type(unifier, primitives, *ty, cache),
+                    is_vararg_ctx: *is_vararg_ctx,
                 },
                 TypeEnum::TObj { obj_id, fields, params } => ConcreteTypeEnum::TObj {
                     obj_id: *obj_id,
@@ -254,15 +261,13 @@ impl ConcreteTypeStore {
                 *cache.get_mut(&cty).unwrap() = Some(ty);
                 return ty;
             }
-            ConcreteTypeEnum::TTuple { ty } => TypeEnum::TTuple {
+            ConcreteTypeEnum::TTuple { ty, is_vararg_ctx } => TypeEnum::TTuple {
                 ty: ty
                     .iter()
                     .map(|cty| self.to_unifier_type(unifier, primitives, *cty, cache))
                     .collect(),
+                is_vararg_ctx: *is_vararg_ctx,
             },
-            ConcreteTypeEnum::TList { ty } => {
-                TypeEnum::TList { ty: self.to_unifier_type(unifier, primitives, *ty, cache) }
-            }
             ConcreteTypeEnum::TVirtual { ty } => {
                 TypeEnum::TVirtual { ty: self.to_unifier_type(unifier, primitives, *ty, cache) }
             }
@@ -286,6 +291,7 @@ impl ConcreteTypeStore {
                         name: arg.name,
                         ty: self.to_unifier_type(unifier, primitives, arg.ty, cache),
                         default_value: arg.default_value.clone(),
+                        is_vararg: false,
                     })
                     .collect(),
                 ret: self.to_unifier_type(unifier, primitives, *ret, cache),

nac3core/src/codegen/extern_fns.rs

@@ -1,517 +1,102 @@
-use inkwell::attributes::{Attribute, AttributeLoc};
-use inkwell::values::{BasicValueEnum, CallSiteValue, FloatValue, IntValue};
+use inkwell::{
+    attributes::{Attribute, AttributeLoc},
+    values::{BasicValueEnum, CallSiteValue, FloatValue, IntValue},
+};
 use itertools::Either;
 
-use crate::codegen::CodeGenContext;
+use super::CodeGenContext;
 
-/// Invokes the [`tan`](https://en.cppreference.com/w/c/numeric/math/tan) function.
-pub fn call_tan<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
+/// Macro to generate extern function
+/// Both function return type and function parameter type are `FloatValue`
+///
+/// Arguments:
+/// * `unary/binary`: Whether the extern function requires one (unary) or two (binary) operands
+/// * `$fn_name:ident`: The identifier of the rust function to be generated
+/// * `$extern_fn:literal`: Name of underlying extern function
+///
+/// Optional Arguments:
+/// * `$(,$attributes:literal)*)`: Attributes linked with the extern function.
+///   The default attributes are "mustprogress", "nofree", "nounwind", "willreturn", and "writeonly".
+///   These will be used unless other attributes are specified
+/// * `$(,$args:ident)*`: Operands of the extern function
+///   The data type of these operands will be set to `FloatValue`
+///  
+macro_rules! generate_extern_fn {
+    ("unary", $fn_name:ident, $extern_fn:literal) => {
+        generate_extern_fn!($fn_name, $extern_fn, arg, "mustprogress", "nofree", "nounwind", "willreturn", "writeonly");
+    };
+    ("unary", $fn_name:ident, $extern_fn:literal $(,$attributes:literal)*) => {
+        generate_extern_fn!($fn_name, $extern_fn, arg $(,$attributes)*);
+    };
+    ("binary", $fn_name:ident, $extern_fn:literal) => {
+        generate_extern_fn!($fn_name, $extern_fn, arg1, arg2, "mustprogress", "nofree", "nounwind", "willreturn", "writeonly");
+    };
+    ("binary", $fn_name:ident, $extern_fn:literal $(,$attributes:literal)*) => {
+        generate_extern_fn!($fn_name, $extern_fn, arg1, arg2 $(,$attributes)*);
+    };
+    ($fn_name:ident, $extern_fn:literal $(,$args:ident)* $(,$attributes:literal)*) => {
+        #[doc = concat!("Invokes the [`", stringify!($extern_fn), "`](https://en.cppreference.com/w/c/numeric/math/", stringify!($llvm_name), ") function." )]
+        pub fn $fn_name<'ctx>(
+            ctx: &CodeGenContext<'ctx, '_>
+            $(,$args: FloatValue<'ctx>)*,
             name: Option<&str>,
         ) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "tan";
+            const FN_NAME: &str = $extern_fn;
 
             let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
+            $(debug_assert_eq!($args.get_type(), llvm_f64);)*
 
             let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
+                let fn_type = llvm_f64.fn_type(&[$($args.get_type().into()),*], false);
                 let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
+                for attr in [$($attributes),*] {
                     func.add_attribute(
                         AttributeLoc::Function,
                         ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
                     );
                 }
-
                 func
             });
 
             ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
+                .build_call(extern_fn, &[$($args.into()),*], name.unwrap_or_default())
                 .map(CallSiteValue::try_as_basic_value)
                 .map(|v| v.map_left(BasicValueEnum::into_float_value))
                 .map(Either::unwrap_left)
                 .unwrap()
         }
+    };
+}
 
-/// Invokes the [`asin`](https://en.cppreference.com/w/c/numeric/math/asin) function.
-pub fn call_asin<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "asin";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
+generate_extern_fn!("unary", call_tan, "tan");
+generate_extern_fn!("unary", call_asin, "asin");
+generate_extern_fn!("unary", call_acos, "acos");
+generate_extern_fn!("unary", call_atan, "atan");
+generate_extern_fn!("unary", call_sinh, "sinh");
+generate_extern_fn!("unary", call_cosh, "cosh");
+generate_extern_fn!("unary", call_tanh, "tanh");
+generate_extern_fn!("unary", call_asinh, "asinh");
+generate_extern_fn!("unary", call_acosh, "acosh");
+generate_extern_fn!("unary", call_atanh, "atanh");
+generate_extern_fn!("unary", call_expm1, "expm1");
+generate_extern_fn!(
+    "unary",
+    call_cbrt,
+    "cbrt",
+    "mustprogress",
+    "nofree",
+    "nosync",
+    "nounwind",
+    "readonly",
+    "willreturn"
 );
-        }
+generate_extern_fn!("unary", call_erf, "erf", "nounwind");
+generate_extern_fn!("unary", call_erfc, "erfc", "nounwind");
+generate_extern_fn!("unary", call_j1, "j1", "nounwind");
 
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`acos`](https://en.cppreference.com/w/c/numeric/math/acos) function.
-pub fn call_acos<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "acos";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`atan`](https://en.cppreference.com/w/c/numeric/math/atan) function.
-pub fn call_atan<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "atan";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`sinh`](https://en.cppreference.com/w/c/numeric/math/sinh) function.
-pub fn call_sinh<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "sinh";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`cosh`](https://en.cppreference.com/w/c/numeric/math/cosh) function.
-pub fn call_cosh<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "cosh";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`tanh`](https://en.cppreference.com/w/c/numeric/math/tanh) function.
-pub fn call_tanh<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "tanh";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`asinh`](https://en.cppreference.com/w/c/numeric/math/asinh) function.
-pub fn call_asinh<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "asinh";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`acosh`](https://en.cppreference.com/w/c/numeric/math/acosh) function.
-pub fn call_acosh<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "acosh";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`atanh`](https://en.cppreference.com/w/c/numeric/math/atanh) function.
-pub fn call_atanh<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "atanh";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`expm1`](https://en.cppreference.com/w/c/numeric/math/expm1) function.
-pub fn call_expm1<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "expm1";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`cbrt`](https://en.cppreference.com/w/c/numeric/math/cbrt) function.
-pub fn call_cbrt<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "cbrt";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nosync", "nounwind", "readonly", "willreturn"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`erf`](https://en.cppreference.com/w/c/numeric/math/erf) function.
-pub fn call_erf<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "erf";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        func.add_attribute(
-            AttributeLoc::Function,
-            ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id("nounwind"), 0),
-        );
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`erfc`](https://en.cppreference.com/w/c/numeric/math/erfc) function.
-pub fn call_erfc<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "erfc";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        func.add_attribute(
-            AttributeLoc::Function,
-            ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id("nounwind"), 0),
-        );
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`j1`](https://www.gnu.org/software/libc/manual/html_node/Special-Functions.html#index-j1)
-/// function.
-pub fn call_j1<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    arg: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "j1";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(arg.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        func.add_attribute(
-            AttributeLoc::Function,
-            ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id("nounwind"), 0),
-        );
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[arg.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`atan2`](https://en.cppreference.com/w/c/numeric/math/atan2) function.
-pub fn call_atan2<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    y: FloatValue<'ctx>,
-    x: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "atan2";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(y.get_type(), llvm_f64);
-    debug_assert_eq!(x.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into(), llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
-            func.add_attribute(
-                AttributeLoc::Function,
-                ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
-            );
-        }
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[y.into(), x.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
+generate_extern_fn!("binary", call_atan2, "atan2");
+generate_extern_fn!("binary", call_hypot, "hypot", "nounwind");
+generate_extern_fn!("binary", call_nextafter, "nextafter", "nounwind");
 
 /// Invokes the [`ldexp`](https://en.cppreference.com/w/c/numeric/math/ldexp) function.
 pub fn call_ldexp<'ctx>(
@@ -548,66 +133,61 @@ pub fn call_ldexp<'ctx>(
         .unwrap()
 }
 
-/// Invokes the [`hypot`](https://en.cppreference.com/w/c/numeric/math/hypot) function.
-pub fn call_hypot<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    x: FloatValue<'ctx>,
-    y: FloatValue<'ctx>,
+/// Macro to generate `np_linalg` and `sp_linalg` functions
+/// The function takes as input `NDArray` and returns ()
+///
+/// Arguments:
+/// * `$fn_name:ident`: The identifier of the rust function to be generated
+/// * `$extern_fn:literal`: Name of underlying extern function
+/// * (2/3/4): Number of `NDArray` that function takes as input
+///
+/// Note:
+/// The operands and resulting `NDArray` are both passed as input to the funcion
+/// It is the responsibility of caller to ensure that output `NDArray` is properly allocated on stack
+/// The function changes the content of the output `NDArray` in-place
+macro_rules! generate_linalg_extern_fn {
+    ($fn_name:ident, $extern_fn:literal, 2) => {
+        generate_linalg_extern_fn!($fn_name, $extern_fn, mat1, mat2);
+    };
+    ($fn_name:ident, $extern_fn:literal, 3) => {
+        generate_linalg_extern_fn!($fn_name, $extern_fn, mat1, mat2, mat3);
+    };
+    ($fn_name:ident, $extern_fn:literal, 4) => {
+        generate_linalg_extern_fn!($fn_name, $extern_fn, mat1, mat2, mat3, mat4);
+    };
+    ($fn_name:ident, $extern_fn:literal $(,$input_matrix:ident)*) => {
+        #[doc = concat!("Invokes the linalg `", stringify!($extern_fn), " function." )]
+        pub fn $fn_name<'ctx>(
+            ctx: &mut CodeGenContext<'ctx, '_>
+            $(,$input_matrix: BasicValueEnum<'ctx>)*,
             name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "hypot";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(x.get_type(), llvm_f64);
-    debug_assert_eq!(y.get_type(), llvm_f64);
-
+        ){
+            const FN_NAME: &str = $extern_fn;
             let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into(), llvm_f64.into()], false);
+                let fn_type = ctx.ctx.void_type().fn_type(&[$($input_matrix.get_type().into()),*], false);
+
                 let func = ctx.module.add_function(FN_NAME, fn_type, None);
+                for attr in ["mustprogress", "nofree", "nounwind", "willreturn", "writeonly"] {
                     func.add_attribute(
                         AttributeLoc::Function,
-            ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id("nounwind"), 0),
+                        ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
                     );
-
+                }
                 func
             });
 
-    ctx.builder
-        .build_call(extern_fn, &[x.into(), y.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
+            ctx.builder.build_call(extern_fn, &[$($input_matrix.into(),)*], name.unwrap_or_default()).unwrap();
+        }
+    };
 }
 
-/// Invokes the [`nextafter`](https://en.cppreference.com/w/c/numeric/math/nextafter) function.
-pub fn call_nextafter<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    from: FloatValue<'ctx>,
-    to: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "nextafter";
-
-    let llvm_f64 = ctx.ctx.f64_type();
-    debug_assert_eq!(from.get_type(), llvm_f64);
-    debug_assert_eq!(to.get_type(), llvm_f64);
-
-    let extern_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into(), llvm_f64.into()], false);
-        let func = ctx.module.add_function(FN_NAME, fn_type, None);
-        func.add_attribute(
-            AttributeLoc::Function,
-            ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id("nounwind"), 0),
-        );
-
-        func
-    });
-
-    ctx.builder
-        .build_call(extern_fn, &[from.into(), to.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
+generate_linalg_extern_fn!(call_np_linalg_cholesky, "np_linalg_cholesky", 2);
+generate_linalg_extern_fn!(call_np_linalg_qr, "np_linalg_qr", 3);
+generate_linalg_extern_fn!(call_np_linalg_svd, "np_linalg_svd", 4);
+generate_linalg_extern_fn!(call_np_linalg_inv, "np_linalg_inv", 2);
+generate_linalg_extern_fn!(call_np_linalg_pinv, "np_linalg_pinv", 2);
+generate_linalg_extern_fn!(call_np_linalg_matrix_power, "np_linalg_matrix_power", 3);
+generate_linalg_extern_fn!(call_np_linalg_det, "np_linalg_det", 2);
+generate_linalg_extern_fn!(call_sp_linalg_lu, "sp_linalg_lu", 3);
+generate_linalg_extern_fn!(call_sp_linalg_schur, "sp_linalg_schur", 3);
+generate_linalg_extern_fn!(call_sp_linalg_hessenberg, "sp_linalg_hessenberg", 3);

nac3core/src/codegen/generator.rs

@@ -1,16 +1,18 @@
-use crate::{
-    codegen::{bool_to_i1, bool_to_i8, classes::ArraySliceValue, expr::*, stmt::*, CodeGenContext},
-    symbol_resolver::ValueEnum,
-    toplevel::{DefinitionId, TopLevelDef},
-    typecheck::typedef::{FunSignature, Type},
-};
 use inkwell::{
     context::Context,
     types::{BasicTypeEnum, IntType},
     values::{BasicValueEnum, IntValue, PointerValue},
 };
+
 use nac3parser::ast::{Expr, Stmt, StrRef};
 
+use super::{bool_to_i1, bool_to_i8, expr::*, stmt::*, values::ArraySliceValue, CodeGenContext};
+use crate::{
+    symbol_resolver::ValueEnum,
+    toplevel::{DefinitionId, TopLevelDef},
+    typecheck::typedef::{FunSignature, Type},
+};
+
 pub trait CodeGenerator {
     /// Return the module name for the code generator.
     fn get_name(&self) -> &str;
@@ -57,6 +59,7 @@ pub trait CodeGenerator {
     /// - fun: Function signature, definition ID and the substitution key.
     /// - params: Function parameters. Note that this does not include the object even if the
     ///   function is a class method.
+    ///
     /// Note that this function should check if the function is generated in another thread (due to
     /// possible race condition), see the default implementation for an example.
     fn gen_func_instance<'ctx>(
@@ -123,11 +126,45 @@ pub trait CodeGenerator {
         ctx: &mut CodeGenContext<'ctx, '_>,
         target: &Expr<Option<Type>>,
         value: ValueEnum<'ctx>,
+        value_ty: Type,
     ) -> Result<(), String>
     where
         Self: Sized,
     {
-        gen_assign(self, ctx, target, value)
+        gen_assign(self, ctx, target, value, value_ty)
+    }
+
+    /// Generate code for an assignment expression where LHS is a `"target_list"`.
+    ///
+    /// See <https://docs.python.org/3/reference/simple_stmts.html#assignment-statements>.
+    fn gen_assign_target_list<'ctx>(
+        &mut self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        targets: &Vec<Expr<Option<Type>>>,
+        value: ValueEnum<'ctx>,
+        value_ty: Type,
+    ) -> Result<(), String>
+    where
+        Self: Sized,
+    {
+        gen_assign_target_list(self, ctx, targets, value, value_ty)
+    }
+
+    /// Generate code for an item assignment.
+    ///
+    /// i.e., `target[key] = value`
+    fn gen_setitem<'ctx>(
+        &mut self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        target: &Expr<Option<Type>>,
+        key: &Expr<Option<Type>>,
+        value: ValueEnum<'ctx>,
+        value_ty: Type,
+    ) -> Result<(), String>
+    where
+        Self: Sized,
+    {
+        gen_setitem(self, ctx, target, key, value, value_ty)
     }
 
     /// Generate code for a while expression.

nac3core/src/codegen/irrt/irrt.c

@@ -1,389 +0,0 @@
-typedef _BitInt(8) int8_t;
-typedef unsigned _BitInt(8) uint8_t;
-typedef _BitInt(32) int32_t;
-typedef unsigned _BitInt(32) uint32_t;
-typedef _BitInt(64) int64_t;
-typedef unsigned _BitInt(64) uint64_t;
-
-# define MAX(a, b) (a > b ? a : b)
-# define MIN(a, b) (a > b ? b : a)
-
-# define NULL ((void *) 0)
-
-// adapted from GNU Scientific Library: https://git.savannah.gnu.org/cgit/gsl.git/tree/sys/pow_int.c
-// need to make sure `exp >= 0` before calling this function
-#define DEF_INT_EXP(T) T __nac3_int_exp_##T( \
-    T base, \
-    T exp \
-) { \
-    T res = (T)1; \
-    /* repeated squaring method */ \
-    do { \
-       if (exp & 1) res *= base;  /* for n odd */ \
-       exp >>= 1; \
-       base *= base; \
-    } while (exp); \
-    return res; \
-} \
-
-DEF_INT_EXP(int32_t)
-DEF_INT_EXP(int64_t)
-DEF_INT_EXP(uint32_t)
-DEF_INT_EXP(uint64_t)
-
-
-int32_t __nac3_slice_index_bound(int32_t i, const int32_t len) {
-    if (i < 0) {
-        i = len + i;
-    }
-    if (i < 0) {
-        return 0;
-    } else if (i > len) {
-        return len;
-    }
-    return i;
-}
-
-int32_t __nac3_range_slice_len(const int32_t start, const int32_t end, const int32_t step) {
-    int32_t diff = end - start;
-    if (diff > 0 && step > 0) {
-        return ((diff - 1) / step) + 1;
-    } else if (diff < 0 && step < 0) {
-        return ((diff + 1) / step) + 1;
-    } else {
-        return 0;
-    }
-}
-
-// Handle list assignment and dropping part of the list when
-// both dest_step and src_step are +1.
-// - All the index must *not* be out-of-bound or negative,
-// - The end index is *inclusive*,
-// - The length of src and dest slice size should already
-// be checked: if dest.step == 1 then len(src) <= len(dest) else len(src) == len(dest)
-int32_t __nac3_list_slice_assign_var_size(
-    int32_t dest_start,
-    int32_t dest_end,
-    int32_t dest_step,
-    uint8_t *dest_arr,
-    int32_t dest_arr_len,
-    int32_t src_start,
-    int32_t src_end,
-    int32_t src_step,
-    uint8_t *src_arr,
-    int32_t src_arr_len,
-    const int32_t size
-) {
-    /* if dest_arr_len == 0, do nothing since we do not support extending list */
-    if (dest_arr_len == 0) return dest_arr_len;
-    /* if both step is 1, memmove directly, handle the dropping of the list, and shrink size */
-    if (src_step == dest_step && dest_step == 1) {
-        const int32_t src_len = (src_end >= src_start) ? (src_end - src_start + 1) : 0;
-        const int32_t dest_len = (dest_end >= dest_start) ? (dest_end - dest_start + 1) : 0;
-        if (src_len > 0) {
-            __builtin_memmove(
-                dest_arr + dest_start * size,
-                src_arr + src_start * size,
-                src_len * size
-            );
-        }
-        if (dest_len > 0) {
-            /* dropping */
-            __builtin_memmove(
-                dest_arr + (dest_start + src_len) * size,
-                dest_arr + (dest_end + 1) * size,
-                (dest_arr_len - dest_end - 1) * size
-            );
-        }
-        /* shrink size */
-        return dest_arr_len - (dest_len - src_len);
-    }
-    /* if two range overlaps, need alloca */
-    uint8_t need_alloca =
-        (dest_arr == src_arr)
-        && !(
-            MAX(dest_start, dest_end) < MIN(src_start, src_end)
-            || MAX(src_start, src_end) < MIN(dest_start, dest_end)
-        );
-    if (need_alloca) {
-        uint8_t *tmp = __builtin_alloca(src_arr_len * size);
-        __builtin_memcpy(tmp, src_arr, src_arr_len * size);
-        src_arr = tmp;
-    }
-    int32_t src_ind = src_start;
-    int32_t dest_ind = dest_start;
-    for (;
-        (src_step > 0) ? (src_ind <= src_end) : (src_ind >= src_end);
-        src_ind += src_step, dest_ind += dest_step
-    ) {
-        /* for constant optimization */
-        if (size == 1) {
-            __builtin_memcpy(dest_arr + dest_ind, src_arr + src_ind, 1);
-        } else if (size == 4) {
-            __builtin_memcpy(dest_arr + dest_ind * 4, src_arr + src_ind * 4, 4);
-        } else if (size == 8) {
-            __builtin_memcpy(dest_arr + dest_ind * 8, src_arr + src_ind * 8, 8);
-        } else {
-            /* memcpy for var size, cannot overlap after previous alloca */
-            __builtin_memcpy(dest_arr + dest_ind * size, src_arr + src_ind * size, size);
-        }
-    }
-    /* only dest_step == 1 can we shrink the dest list. */
-    /* size should be ensured prior to calling this function */
-    if (dest_step == 1 && dest_end >= dest_start) {
-        __builtin_memmove(
-            dest_arr + dest_ind * size,
-            dest_arr + (dest_end + 1) * size,
-            (dest_arr_len - dest_end - 1) * size
-        );
-        return dest_arr_len - (dest_end - dest_ind) - 1;
-    }
-    return dest_arr_len;
-}
-
-int32_t __nac3_isinf(double x) {
-    return __builtin_isinf(x);
-}
-
-int32_t __nac3_isnan(double x) {
-    return __builtin_isnan(x);
-}
-
-double tgamma(double arg);
-
-double __nac3_gamma(double z) {
-    // Handling for denormals
-    //     | x                 | Python gamma(x) | C tgamma(x) |
-    // --- | ----------------- | --------------- | ----------- |
-    // (1) | nan               | nan             | nan         |
-    // (2) | -inf              | -inf            | inf         |
-    // (3) | inf               | inf             | inf         |
-    // (4) | 0.0               | inf             | inf         |
-    // (5) | {-1.0, -2.0, ...} | inf             | nan         |
-
-    // (1)-(3)
-    if (__builtin_isinf(z) || __builtin_isnan(z)) {
-        return z;
-    }
-
-    double v = tgamma(z);
-
-    // (4)-(5)
-    return __builtin_isinf(v) || __builtin_isnan(v) ? __builtin_inf() : v;
-}
-
-double lgamma(double arg);
-
-double __nac3_gammaln(double x) {
-    // libm's handling of value overflows differs from scipy:
-    // - scipy: gammaln(-inf) -> -inf
-    // - libm : lgamma(-inf) -> inf
-
-    if (__builtin_isinf(x)) {
-        return x;
-    }
-
-    return lgamma(x);
-}
-
-double j0(double x);
-
-double __nac3_j0(double x) {
-    // libm's handling of value overflows differs from scipy:
-    // - scipy: j0(inf) -> nan
-    // - libm : j0(inf) -> 0.0
-
-    if (__builtin_isinf(x)) {
-        return __builtin_nan("");
-    }
-
-    return j0(x);
-}
-
-uint32_t __nac3_ndarray_calc_size(
-    const uint64_t *list_data,
-    uint32_t list_len,
-    uint32_t begin_idx,
-    uint32_t end_idx
-) {
-    __builtin_assume(end_idx <= list_len);
-
-    uint32_t num_elems = 1;
-    for (uint32_t i = begin_idx; i < end_idx; ++i) {
-        uint64_t val = list_data[i];
-        __builtin_assume(val > 0);
-        num_elems *= val;
-    }
-    return num_elems;
-}
-
-uint64_t __nac3_ndarray_calc_size64(
-    const uint64_t *list_data,
-    uint64_t list_len,
-    uint64_t begin_idx,
-    uint64_t end_idx
-) {
-    __builtin_assume(end_idx <= list_len);
-
-    uint64_t num_elems = 1;
-    for (uint64_t i = begin_idx; i < end_idx; ++i) {
-        uint64_t val = list_data[i];
-        __builtin_assume(val > 0);
-       num_elems *= val;
-    }
-    return num_elems;
-}
-
-void __nac3_ndarray_calc_nd_indices(
-    uint32_t index,
-    const uint32_t* dims,
-    uint32_t num_dims,
-    uint32_t* idxs
-) {
-    uint32_t stride = 1;
-    for (uint32_t dim = 0; dim < num_dims; dim++) {
-        uint32_t i = num_dims - dim - 1;
-        __builtin_assume(dims[i] > 0);
-        idxs[i] = (index / stride) % dims[i];
-        stride *= dims[i];
-    }
-}
-
-void __nac3_ndarray_calc_nd_indices64(
-    uint64_t index,
-    const uint64_t* dims,
-    uint64_t num_dims,
-    uint32_t* idxs
-) {
-    uint64_t stride = 1;
-    for (uint64_t dim = 0; dim < num_dims; dim++) {
-        uint64_t i = num_dims - dim - 1;
-        __builtin_assume(dims[i] > 0);
-        idxs[i] = (uint32_t) ((index / stride) % dims[i]);
-        stride *= dims[i];
-    }
-}
-
-uint32_t __nac3_ndarray_flatten_index(
-    const uint32_t* dims,
-    uint32_t num_dims,
-    const uint32_t* indices,
-    uint32_t num_indices
-) {
-    uint32_t idx = 0;
-    uint32_t stride = 1;
-    for (uint32_t i = 0; i < num_dims; ++i) {
-        uint32_t ri = num_dims - i - 1;
-        if (ri < num_indices) {
-            idx += (stride * indices[ri]);
-        }
-
-        __builtin_assume(dims[i] > 0);
-        stride *= dims[ri];
-    }
-    return idx;
-}
-
-uint64_t __nac3_ndarray_flatten_index64(
-    const uint64_t* dims,
-    uint64_t num_dims,
-    const uint32_t* indices,
-    uint64_t num_indices
-) {
-    uint64_t idx = 0;
-    uint64_t stride = 1;
-    for (uint64_t i = 0; i < num_dims; ++i) {
-        uint64_t ri = num_dims - i - 1;
-        if (ri < num_indices) {
-            idx += (stride * indices[ri]);
-        }
-
-        __builtin_assume(dims[i] > 0);
-        stride *= dims[ri];
-    }
-    return idx;
-}
-
-void __nac3_ndarray_calc_broadcast(
-    const uint32_t *lhs_dims,
-    uint32_t lhs_ndims,
-    const uint32_t *rhs_dims,
-    uint32_t rhs_ndims,
-    uint32_t *out_dims
-) {
-    uint32_t max_ndims = lhs_ndims > rhs_ndims ? lhs_ndims : rhs_ndims;
-
-    for (uint32_t i = 0; i < max_ndims; ++i) {
-        uint32_t *lhs_dim_sz = i < lhs_ndims ? &lhs_dims[lhs_ndims - i - 1] : NULL;
-        uint32_t *rhs_dim_sz = i < rhs_ndims ? &rhs_dims[rhs_ndims - i - 1] : NULL;
-        uint32_t *out_dim = &out_dims[max_ndims - i - 1];
-
-        if (lhs_dim_sz == NULL) {
-            *out_dim = *rhs_dim_sz;
-        } else if (rhs_dim_sz == NULL) {
-            *out_dim = *lhs_dim_sz;
-        } else if (*lhs_dim_sz == 1) {
-            *out_dim = *rhs_dim_sz;
-        } else if (*rhs_dim_sz == 1) {
-            *out_dim = *lhs_dim_sz;
-        } else if (*lhs_dim_sz == *rhs_dim_sz) {
-            *out_dim = *lhs_dim_sz;
-        } else {
-            __builtin_unreachable();
-        }
-    }
-}
-
-void __nac3_ndarray_calc_broadcast64(
-    const uint64_t *lhs_dims,
-    uint64_t lhs_ndims,
-    const uint64_t *rhs_dims,
-    uint64_t rhs_ndims,
-    uint64_t *out_dims
-) {
-    uint64_t max_ndims = lhs_ndims > rhs_ndims ? lhs_ndims : rhs_ndims;
-
-    for (uint64_t i = 0; i < max_ndims; ++i) {
-        uint64_t *lhs_dim_sz = i < lhs_ndims ? &lhs_dims[lhs_ndims - i - 1] : NULL;
-        uint64_t *rhs_dim_sz = i < rhs_ndims ? &rhs_dims[rhs_ndims - i - 1] : NULL;
-        uint64_t *out_dim = &out_dims[max_ndims - i - 1];
-
-        if (lhs_dim_sz == NULL) {
-            *out_dim = *rhs_dim_sz;
-        } else if (rhs_dim_sz == NULL) {
-            *out_dim = *lhs_dim_sz;
-        } else if (*lhs_dim_sz == 1) {
-            *out_dim = *rhs_dim_sz;
-        } else if (*rhs_dim_sz == 1) {
-            *out_dim = *lhs_dim_sz;
-        } else if (*lhs_dim_sz == *rhs_dim_sz) {
-            *out_dim = *lhs_dim_sz;
-        } else {
-            __builtin_unreachable();
-        }
-    }
-}
-
-void __nac3_ndarray_calc_broadcast_idx(
-    const uint32_t *src_dims,
-    uint32_t src_ndims,
-    const uint32_t *in_idx,
-    uint32_t *out_idx
-) {
-    for (uint32_t i = 0; i < src_ndims; ++i) {
-        uint32_t src_i = src_ndims - i - 1;
-        out_idx[src_i] = src_dims[src_i] == 1 ? 0 : in_idx[src_i];
-    }
-}
-
-void __nac3_ndarray_calc_broadcast_idx64(
-    const uint64_t *src_dims,
-    uint64_t src_ndims,
-    const uint32_t *in_idx,
-    uint32_t *out_idx
-) {
-    for (uint64_t i = 0; i < src_ndims; ++i) {
-        uint64_t src_i = src_ndims - i - 1;
-        out_idx[src_i] = src_dims[src_i] == 1 ? 0 : (uint32_t) in_idx[src_i];
-    }
-}

nac3core/src/codegen/irrt/list.rs

@@ -0,0 +1,162 @@
+use inkwell::{
+    types::BasicTypeEnum,
+    values::{BasicValueEnum, CallSiteValue, IntValue},
+    AddressSpace, IntPredicate,
+};
+use itertools::Either;
+
+use super::calculate_len_for_slice_range;
+use crate::codegen::{
+    macros::codegen_unreachable,
+    values::{ArrayLikeValue, ListValue},
+    CodeGenContext, CodeGenerator,
+};
+
+/// This function handles 'end' **inclusively**.
+/// Order of tuples `assign_idx` and `value_idx` is ('start', 'end', 'step').
+/// Negative index should be handled before entering this function
+pub fn list_slice_assignment<'ctx, G: CodeGenerator + ?Sized>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    ty: BasicTypeEnum<'ctx>,
+    dest_arr: ListValue<'ctx>,
+    dest_idx: (IntValue<'ctx>, IntValue<'ctx>, IntValue<'ctx>),
+    src_arr: ListValue<'ctx>,
+    src_idx: (IntValue<'ctx>, IntValue<'ctx>, IntValue<'ctx>),
+) {
+    let size_ty = generator.get_size_type(ctx.ctx);
+    let int8_ptr = ctx.ctx.i8_type().ptr_type(AddressSpace::default());
+    let int32 = ctx.ctx.i32_type();
+    let (fun_symbol, elem_ptr_type) = ("__nac3_list_slice_assign_var_size", int8_ptr);
+    let slice_assign_fun = {
+        let ty_vec = vec![
+            int32.into(),         // dest start idx
+            int32.into(),         // dest end idx
+            int32.into(),         // dest step
+            elem_ptr_type.into(), // dest arr ptr
+            int32.into(),         // dest arr len
+            int32.into(),         // src start idx
+            int32.into(),         // src end idx
+            int32.into(),         // src step
+            elem_ptr_type.into(), // src arr ptr
+            int32.into(),         // src arr len
+            int32.into(),         // size
+        ];
+        ctx.module.get_function(fun_symbol).unwrap_or_else(|| {
+            let fn_t = int32.fn_type(ty_vec.as_slice(), false);
+            ctx.module.add_function(fun_symbol, fn_t, None)
+        })
+    };
+
+    let zero = int32.const_zero();
+    let one = int32.const_int(1, false);
+    let dest_arr_ptr = dest_arr.data().base_ptr(ctx, generator);
+    let dest_arr_ptr =
+        ctx.builder.build_pointer_cast(dest_arr_ptr, elem_ptr_type, "dest_arr_ptr_cast").unwrap();
+    let dest_len = dest_arr.load_size(ctx, Some("dest.len"));
+    let dest_len = ctx.builder.build_int_truncate_or_bit_cast(dest_len, int32, "srclen32").unwrap();
+    let src_arr_ptr = src_arr.data().base_ptr(ctx, generator);
+    let src_arr_ptr =
+        ctx.builder.build_pointer_cast(src_arr_ptr, elem_ptr_type, "src_arr_ptr_cast").unwrap();
+    let src_len = src_arr.load_size(ctx, Some("src.len"));
+    let src_len = ctx.builder.build_int_truncate_or_bit_cast(src_len, int32, "srclen32").unwrap();
+
+    // index in bound and positive should be done
+    // assert if dest.step == 1 then len(src) <= len(dest) else len(src) == len(dest), and
+    // throw exception if not satisfied
+    let src_end = ctx
+        .builder
+        .build_select(
+            ctx.builder.build_int_compare(IntPredicate::SLT, src_idx.2, zero, "is_neg").unwrap(),
+            ctx.builder.build_int_sub(src_idx.1, one, "e_min_one").unwrap(),
+            ctx.builder.build_int_add(src_idx.1, one, "e_add_one").unwrap(),
+            "final_e",
+        )
+        .map(BasicValueEnum::into_int_value)
+        .unwrap();
+    let dest_end = ctx
+        .builder
+        .build_select(
+            ctx.builder.build_int_compare(IntPredicate::SLT, dest_idx.2, zero, "is_neg").unwrap(),
+            ctx.builder.build_int_sub(dest_idx.1, one, "e_min_one").unwrap(),
+            ctx.builder.build_int_add(dest_idx.1, one, "e_add_one").unwrap(),
+            "final_e",
+        )
+        .map(BasicValueEnum::into_int_value)
+        .unwrap();
+    let src_slice_len =
+        calculate_len_for_slice_range(generator, ctx, src_idx.0, src_end, src_idx.2);
+    let dest_slice_len =
+        calculate_len_for_slice_range(generator, ctx, dest_idx.0, dest_end, dest_idx.2);
+    let src_eq_dest = ctx
+        .builder
+        .build_int_compare(IntPredicate::EQ, src_slice_len, dest_slice_len, "slice_src_eq_dest")
+        .unwrap();
+    let src_slt_dest = ctx
+        .builder
+        .build_int_compare(IntPredicate::SLT, src_slice_len, dest_slice_len, "slice_src_slt_dest")
+        .unwrap();
+    let dest_step_eq_one = ctx
+        .builder
+        .build_int_compare(
+            IntPredicate::EQ,
+            dest_idx.2,
+            dest_idx.2.get_type().const_int(1, false),
+            "slice_dest_step_eq_one",
+        )
+        .unwrap();
+    let cond_1 = ctx.builder.build_and(dest_step_eq_one, src_slt_dest, "slice_cond_1").unwrap();
+    let cond = ctx.builder.build_or(src_eq_dest, cond_1, "slice_cond").unwrap();
+    ctx.make_assert(
+        generator,
+        cond,
+        "0:ValueError",
+        "attempt to assign sequence of size {0} to slice of size {1} with step size {2}",
+        [Some(src_slice_len), Some(dest_slice_len), Some(dest_idx.2)],
+        ctx.current_loc,
+    );
+
+    let new_len = {
+        let args = vec![
+            dest_idx.0.into(),   // dest start idx
+            dest_idx.1.into(),   // dest end idx
+            dest_idx.2.into(),   // dest step
+            dest_arr_ptr.into(), // dest arr ptr
+            dest_len.into(),     // dest arr len
+            src_idx.0.into(),    // src start idx
+            src_idx.1.into(),    // src end idx
+            src_idx.2.into(),    // src step
+            src_arr_ptr.into(),  // src arr ptr
+            src_len.into(),      // src arr len
+            {
+                let s = match ty {
+                    BasicTypeEnum::FloatType(t) => t.size_of(),
+                    BasicTypeEnum::IntType(t) => t.size_of(),
+                    BasicTypeEnum::PointerType(t) => t.size_of(),
+                    BasicTypeEnum::StructType(t) => t.size_of().unwrap(),
+                    _ => codegen_unreachable!(ctx),
+                };
+                ctx.builder.build_int_truncate_or_bit_cast(s, int32, "size").unwrap()
+            }
+            .into(),
+        ];
+        ctx.builder
+            .build_call(slice_assign_fun, args.as_slice(), "slice_assign")
+            .map(CallSiteValue::try_as_basic_value)
+            .map(|v| v.map_left(BasicValueEnum::into_int_value))
+            .map(Either::unwrap_left)
+            .unwrap()
+    };
+    // update length
+    let need_update =
+        ctx.builder.build_int_compare(IntPredicate::NE, new_len, dest_len, "need_update").unwrap();
+    let current = ctx.builder.get_insert_block().unwrap().get_parent().unwrap();
+    let update_bb = ctx.ctx.append_basic_block(current, "update");
+    let cont_bb = ctx.ctx.append_basic_block(current, "cont");
+    ctx.builder.build_conditional_branch(need_update, update_bb, cont_bb).unwrap();
+    ctx.builder.position_at_end(update_bb);
+    let new_len = ctx.builder.build_int_z_extend_or_bit_cast(new_len, size_ty, "new_len").unwrap();
+    dest_arr.store_size(ctx, generator, new_len);
+    ctx.builder.build_unconditional_branch(cont_bb).unwrap();
+    ctx.builder.position_at_end(cont_bb);
+}

nac3core/src/codegen/irrt/math.rs

@@ -0,0 +1,152 @@
+use inkwell::{
+    values::{BasicValueEnum, CallSiteValue, FloatValue, IntValue},
+    IntPredicate,
+};
+use itertools::Either;
+
+use crate::codegen::{
+    macros::codegen_unreachable,
+    {CodeGenContext, CodeGenerator},
+};
+
+// repeated squaring method adapted from GNU Scientific Library:
+// https://git.savannah.gnu.org/cgit/gsl.git/tree/sys/pow_int.c
+pub fn integer_power<'ctx, G: CodeGenerator + ?Sized>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    base: IntValue<'ctx>,
+    exp: IntValue<'ctx>,
+    signed: bool,
+) -> IntValue<'ctx> {
+    let symbol = match (base.get_type().get_bit_width(), exp.get_type().get_bit_width(), signed) {
+        (32, 32, true) => "__nac3_int_exp_int32_t",
+        (64, 64, true) => "__nac3_int_exp_int64_t",
+        (32, 32, false) => "__nac3_int_exp_uint32_t",
+        (64, 64, false) => "__nac3_int_exp_uint64_t",
+        _ => codegen_unreachable!(ctx),
+    };
+    let base_type = base.get_type();
+    let pow_fun = ctx.module.get_function(symbol).unwrap_or_else(|| {
+        let fn_type = base_type.fn_type(&[base_type.into(), base_type.into()], false);
+        ctx.module.add_function(symbol, fn_type, None)
+    });
+    // throw exception when exp < 0
+    let ge_zero = ctx
+        .builder
+        .build_int_compare(
+            IntPredicate::SGE,
+            exp,
+            exp.get_type().const_zero(),
+            "assert_int_pow_ge_0",
+        )
+        .unwrap();
+    ctx.make_assert(
+        generator,
+        ge_zero,
+        "0:ValueError",
+        "integer power must be positive or zero",
+        [None, None, None],
+        ctx.current_loc,
+    );
+    ctx.builder
+        .build_call(pow_fun, &[base.into(), exp.into()], "call_int_pow")
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_int_value))
+        .map(Either::unwrap_left)
+        .unwrap()
+}
+
+/// Generates a call to `isinf` in IR. Returns an `i1` representing the result.
+pub fn call_isinf<'ctx, G: CodeGenerator + ?Sized>(
+    generator: &mut G,
+    ctx: &CodeGenContext<'ctx, '_>,
+    v: FloatValue<'ctx>,
+) -> IntValue<'ctx> {
+    let intrinsic_fn = ctx.module.get_function("__nac3_isinf").unwrap_or_else(|| {
+        let fn_type = ctx.ctx.i32_type().fn_type(&[ctx.ctx.f64_type().into()], false);
+        ctx.module.add_function("__nac3_isinf", fn_type, None)
+    });
+
+    let ret = ctx
+        .builder
+        .build_call(intrinsic_fn, &[v.into()], "isinf")
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_int_value))
+        .map(Either::unwrap_left)
+        .unwrap();
+
+    generator.bool_to_i1(ctx, ret)
+}
+
+/// Generates a call to `isnan` in IR. Returns an `i1` representing the result.
+pub fn call_isnan<'ctx, G: CodeGenerator + ?Sized>(
+    generator: &mut G,
+    ctx: &CodeGenContext<'ctx, '_>,
+    v: FloatValue<'ctx>,
+) -> IntValue<'ctx> {
+    let intrinsic_fn = ctx.module.get_function("__nac3_isnan").unwrap_or_else(|| {
+        let fn_type = ctx.ctx.i32_type().fn_type(&[ctx.ctx.f64_type().into()], false);
+        ctx.module.add_function("__nac3_isnan", fn_type, None)
+    });
+
+    let ret = ctx
+        .builder
+        .build_call(intrinsic_fn, &[v.into()], "isnan")
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_int_value))
+        .map(Either::unwrap_left)
+        .unwrap();
+
+    generator.bool_to_i1(ctx, ret)
+}
+
+/// Generates a call to `gamma` in IR. Returns an `f64` representing the result.
+pub fn call_gamma<'ctx>(ctx: &CodeGenContext<'ctx, '_>, v: FloatValue<'ctx>) -> FloatValue<'ctx> {
+    let llvm_f64 = ctx.ctx.f64_type();
+
+    let intrinsic_fn = ctx.module.get_function("__nac3_gamma").unwrap_or_else(|| {
+        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
+        ctx.module.add_function("__nac3_gamma", fn_type, None)
+    });
+
+    ctx.builder
+        .build_call(intrinsic_fn, &[v.into()], "gamma")
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_float_value))
+        .map(Either::unwrap_left)
+        .unwrap()
+}
+
+/// Generates a call to `gammaln` in IR. Returns an `f64` representing the result.
+pub fn call_gammaln<'ctx>(ctx: &CodeGenContext<'ctx, '_>, v: FloatValue<'ctx>) -> FloatValue<'ctx> {
+    let llvm_f64 = ctx.ctx.f64_type();
+
+    let intrinsic_fn = ctx.module.get_function("__nac3_gammaln").unwrap_or_else(|| {
+        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
+        ctx.module.add_function("__nac3_gammaln", fn_type, None)
+    });
+
+    ctx.builder
+        .build_call(intrinsic_fn, &[v.into()], "gammaln")
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_float_value))
+        .map(Either::unwrap_left)
+        .unwrap()
+}
+
+/// Generates a call to `j0` in IR. Returns an `f64` representing the result.
+pub fn call_j0<'ctx>(ctx: &CodeGenContext<'ctx, '_>, v: FloatValue<'ctx>) -> FloatValue<'ctx> {
+    let llvm_f64 = ctx.ctx.f64_type();
+
+    let intrinsic_fn = ctx.module.get_function("__nac3_j0").unwrap_or_else(|| {
+        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
+        ctx.module.add_function("__nac3_j0", fn_type, None)
+    });
+
+    ctx.builder
+        .build_call(intrinsic_fn, &[v.into()], "j0")
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_float_value))
+        .map(Either::unwrap_left)
+        .unwrap()
+}

nac3core/src/codegen/irrt/mod.rs

@@ -1,28 +1,28 @@
-use crate::typecheck::typedef::Type;
-
-use super::{
-    classes::{
-        ArrayLikeIndexer, ArrayLikeValue, ArraySliceValue, ListValue, NDArrayValue,
-        TypedArrayLikeAdapter, UntypedArrayLikeAccessor,
-    },
-    llvm_intrinsics, CodeGenContext, CodeGenerator,
-};
-use crate::codegen::classes::TypedArrayLikeAccessor;
-use crate::codegen::stmt::gen_for_callback_incrementing;
 use inkwell::{
     attributes::{Attribute, AttributeLoc},
     context::Context,
     memory_buffer::MemoryBuffer,
     module::Module,
-    types::{BasicTypeEnum, IntType},
-    values::{BasicValueEnum, CallSiteValue, FloatValue, IntValue},
-    AddressSpace, IntPredicate,
+    values::{BasicValue, BasicValueEnum, IntValue},
+    IntPredicate,
 };
-use itertools::Either;
+
 use nac3parser::ast::Expr;
 
+use super::{CodeGenContext, CodeGenerator};
+use crate::{symbol_resolver::SymbolResolver, typecheck::typedef::Type};
+pub use list::*;
+pub use math::*;
+pub use ndarray::*;
+pub use slice::*;
+
+mod list;
+mod math;
+mod ndarray;
+mod slice;
+
 #[must_use]
-pub fn load_irrt(ctx: &Context) -> Module {
+pub fn load_irrt<'ctx>(ctx: &'ctx Context, symbol_resolver: &dyn SymbolResolver) -> Module<'ctx> {
     let bitcode_buf = MemoryBuffer::create_from_memory_range(
         include_bytes!(concat!(env!("OUT_DIR"), "/irrt.bc")),
         "irrt_bitcode_buffer",
@@ -38,91 +38,28 @@ pub fn load_irrt(ctx: &Context) -> Module {
         let function = irrt_mod.get_function(symbol).unwrap();
         function.add_attribute(AttributeLoc::Function, ctx.create_enum_attribute(inline_attr, 0));
     }
+
+    // Initialize all global `EXN_*` exception IDs in IRRT with the [`SymbolResolver`].
+    let exn_id_type = ctx.i32_type();
+    let errors = &[
+        ("EXN_INDEX_ERROR", "0:IndexError"),
+        ("EXN_VALUE_ERROR", "0:ValueError"),
+        ("EXN_ASSERTION_ERROR", "0:AssertionError"),
+        ("EXN_TYPE_ERROR", "0:TypeError"),
+    ];
+    for (irrt_name, symbol_name) in errors {
+        let exn_id = symbol_resolver.get_string_id(symbol_name);
+        let exn_id = exn_id_type.const_int(exn_id as u64, false).as_basic_value_enum();
+
+        let global = irrt_mod.get_global(irrt_name).unwrap_or_else(|| {
+            panic!("Exception symbol name '{irrt_name}' should exist in the IRRT LLVM module")
+        });
+        global.set_initializer(&exn_id);
+    }
+
     irrt_mod
 }
 
-// repeated squaring method adapted from GNU Scientific Library:
-// https://git.savannah.gnu.org/cgit/gsl.git/tree/sys/pow_int.c
-pub fn integer_power<'ctx, G: CodeGenerator + ?Sized>(
-    generator: &mut G,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    base: IntValue<'ctx>,
-    exp: IntValue<'ctx>,
-    signed: bool,
-) -> IntValue<'ctx> {
-    let symbol = match (base.get_type().get_bit_width(), exp.get_type().get_bit_width(), signed) {
-        (32, 32, true) => "__nac3_int_exp_int32_t",
-        (64, 64, true) => "__nac3_int_exp_int64_t",
-        (32, 32, false) => "__nac3_int_exp_uint32_t",
-        (64, 64, false) => "__nac3_int_exp_uint64_t",
-        _ => unreachable!(),
-    };
-    let base_type = base.get_type();
-    let pow_fun = ctx.module.get_function(symbol).unwrap_or_else(|| {
-        let fn_type = base_type.fn_type(&[base_type.into(), base_type.into()], false);
-        ctx.module.add_function(symbol, fn_type, None)
-    });
-    // throw exception when exp < 0
-    let ge_zero = ctx
-        .builder
-        .build_int_compare(
-            IntPredicate::SGE,
-            exp,
-            exp.get_type().const_zero(),
-            "assert_int_pow_ge_0",
-        )
-        .unwrap();
-    ctx.make_assert(
-        generator,
-        ge_zero,
-        "0:ValueError",
-        "integer power must be positive or zero",
-        [None, None, None],
-        ctx.current_loc,
-    );
-    ctx.builder
-        .build_call(pow_fun, &[base.into(), exp.into()], "call_int_pow")
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-pub fn calculate_len_for_slice_range<'ctx, G: CodeGenerator + ?Sized>(
-    generator: &mut G,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    start: IntValue<'ctx>,
-    end: IntValue<'ctx>,
-    step: IntValue<'ctx>,
-) -> IntValue<'ctx> {
-    const SYMBOL: &str = "__nac3_range_slice_len";
-    let len_func = ctx.module.get_function(SYMBOL).unwrap_or_else(|| {
-        let i32_t = ctx.ctx.i32_type();
-        let fn_t = i32_t.fn_type(&[i32_t.into(), i32_t.into(), i32_t.into()], false);
-        ctx.module.add_function(SYMBOL, fn_t, None)
-    });
-
-    // assert step != 0, throw exception if not
-    let not_zero = ctx
-        .builder
-        .build_int_compare(IntPredicate::NE, step, step.get_type().const_zero(), "range_step_ne")
-        .unwrap();
-    ctx.make_assert(
-        generator,
-        not_zero,
-        "0:ValueError",
-        "step must not be zero",
-        [None, None, None],
-        ctx.current_loc,
-    );
-    ctx.builder
-        .build_call(len_func, &[start.into(), end.into(), step.into()], "calc_len")
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
 /// NOTE: the output value of the end index of this function should be compared ***inclusively***,
 /// because python allows `a[2::-1]`, whose semantic is `[a[2], a[1], a[0]]`, which is equivalent to
 /// NO numeric slice in python.
@@ -288,644 +225,3 @@ pub fn handle_slice_indices<'ctx, G: CodeGenerator>(
         }
     }))
 }
-
-/// this function allows index out of range, since python
-/// allows index out of range in slice (`a = [1,2,3]; a[1:10] == [2,3]`).
-pub fn handle_slice_index_bound<'ctx, G: CodeGenerator>(
-    i: &Expr<Option<Type>>,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    generator: &mut G,
-    length: IntValue<'ctx>,
-) -> Result<Option<IntValue<'ctx>>, String> {
-    const SYMBOL: &str = "__nac3_slice_index_bound";
-    let func = ctx.module.get_function(SYMBOL).unwrap_or_else(|| {
-        let i32_t = ctx.ctx.i32_type();
-        let fn_t = i32_t.fn_type(&[i32_t.into(), i32_t.into()], false);
-        ctx.module.add_function(SYMBOL, fn_t, None)
-    });
-
-    let i = if let Some(v) = generator.gen_expr(ctx, i)? {
-        v.to_basic_value_enum(ctx, generator, i.custom.unwrap())?
-    } else {
-        return Ok(None);
-    };
-    Ok(Some(
-        ctx.builder
-            .build_call(func, &[i.into(), length.into()], "bounded_ind")
-            .map(CallSiteValue::try_as_basic_value)
-            .map(|v| v.map_left(BasicValueEnum::into_int_value))
-            .map(Either::unwrap_left)
-            .unwrap(),
-    ))
-}
-
-/// This function handles 'end' **inclusively**.
-/// Order of tuples `assign_idx` and `value_idx` is ('start', 'end', 'step').
-/// Negative index should be handled before entering this function
-pub fn list_slice_assignment<'ctx, G: CodeGenerator + ?Sized>(
-    generator: &mut G,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    ty: BasicTypeEnum<'ctx>,
-    dest_arr: ListValue<'ctx>,
-    dest_idx: (IntValue<'ctx>, IntValue<'ctx>, IntValue<'ctx>),
-    src_arr: ListValue<'ctx>,
-    src_idx: (IntValue<'ctx>, IntValue<'ctx>, IntValue<'ctx>),
-) {
-    let size_ty = generator.get_size_type(ctx.ctx);
-    let int8_ptr = ctx.ctx.i8_type().ptr_type(AddressSpace::default());
-    let int32 = ctx.ctx.i32_type();
-    let (fun_symbol, elem_ptr_type) = ("__nac3_list_slice_assign_var_size", int8_ptr);
-    let slice_assign_fun = {
-        let ty_vec = vec![
-            int32.into(),         // dest start idx
-            int32.into(),         // dest end idx
-            int32.into(),         // dest step
-            elem_ptr_type.into(), // dest arr ptr
-            int32.into(),         // dest arr len
-            int32.into(),         // src start idx
-            int32.into(),         // src end idx
-            int32.into(),         // src step
-            elem_ptr_type.into(), // src arr ptr
-            int32.into(),         // src arr len
-            int32.into(),         // size
-        ];
-        ctx.module.get_function(fun_symbol).unwrap_or_else(|| {
-            let fn_t = int32.fn_type(ty_vec.as_slice(), false);
-            ctx.module.add_function(fun_symbol, fn_t, None)
-        })
-    };
-
-    let zero = int32.const_zero();
-    let one = int32.const_int(1, false);
-    let dest_arr_ptr = dest_arr.data().base_ptr(ctx, generator);
-    let dest_arr_ptr =
-        ctx.builder.build_pointer_cast(dest_arr_ptr, elem_ptr_type, "dest_arr_ptr_cast").unwrap();
-    let dest_len = dest_arr.load_size(ctx, Some("dest.len"));
-    let dest_len = ctx.builder.build_int_truncate_or_bit_cast(dest_len, int32, "srclen32").unwrap();
-    let src_arr_ptr = src_arr.data().base_ptr(ctx, generator);
-    let src_arr_ptr =
-        ctx.builder.build_pointer_cast(src_arr_ptr, elem_ptr_type, "src_arr_ptr_cast").unwrap();
-    let src_len = src_arr.load_size(ctx, Some("src.len"));
-    let src_len = ctx.builder.build_int_truncate_or_bit_cast(src_len, int32, "srclen32").unwrap();
-
-    // index in bound and positive should be done
-    // assert if dest.step == 1 then len(src) <= len(dest) else len(src) == len(dest), and
-    // throw exception if not satisfied
-    let src_end = ctx
-        .builder
-        .build_select(
-            ctx.builder.build_int_compare(IntPredicate::SLT, src_idx.2, zero, "is_neg").unwrap(),
-            ctx.builder.build_int_sub(src_idx.1, one, "e_min_one").unwrap(),
-            ctx.builder.build_int_add(src_idx.1, one, "e_add_one").unwrap(),
-            "final_e",
-        )
-        .map(BasicValueEnum::into_int_value)
-        .unwrap();
-    let dest_end = ctx
-        .builder
-        .build_select(
-            ctx.builder.build_int_compare(IntPredicate::SLT, dest_idx.2, zero, "is_neg").unwrap(),
-            ctx.builder.build_int_sub(dest_idx.1, one, "e_min_one").unwrap(),
-            ctx.builder.build_int_add(dest_idx.1, one, "e_add_one").unwrap(),
-            "final_e",
-        )
-        .map(BasicValueEnum::into_int_value)
-        .unwrap();
-    let src_slice_len =
-        calculate_len_for_slice_range(generator, ctx, src_idx.0, src_end, src_idx.2);
-    let dest_slice_len =
-        calculate_len_for_slice_range(generator, ctx, dest_idx.0, dest_end, dest_idx.2);
-    let src_eq_dest = ctx
-        .builder
-        .build_int_compare(IntPredicate::EQ, src_slice_len, dest_slice_len, "slice_src_eq_dest")
-        .unwrap();
-    let src_slt_dest = ctx
-        .builder
-        .build_int_compare(IntPredicate::SLT, src_slice_len, dest_slice_len, "slice_src_slt_dest")
-        .unwrap();
-    let dest_step_eq_one = ctx
-        .builder
-        .build_int_compare(
-            IntPredicate::EQ,
-            dest_idx.2,
-            dest_idx.2.get_type().const_int(1, false),
-            "slice_dest_step_eq_one",
-        )
-        .unwrap();
-    let cond_1 = ctx.builder.build_and(dest_step_eq_one, src_slt_dest, "slice_cond_1").unwrap();
-    let cond = ctx.builder.build_or(src_eq_dest, cond_1, "slice_cond").unwrap();
-    ctx.make_assert(
-        generator,
-        cond,
-        "0:ValueError",
-        "attempt to assign sequence of size {0} to slice of size {1} with step size {2}",
-        [Some(src_slice_len), Some(dest_slice_len), Some(dest_idx.2)],
-        ctx.current_loc,
-    );
-
-    let new_len = {
-        let args = vec![
-            dest_idx.0.into(),   // dest start idx
-            dest_idx.1.into(),   // dest end idx
-            dest_idx.2.into(),   // dest step
-            dest_arr_ptr.into(), // dest arr ptr
-            dest_len.into(),     // dest arr len
-            src_idx.0.into(),    // src start idx
-            src_idx.1.into(),    // src end idx
-            src_idx.2.into(),    // src step
-            src_arr_ptr.into(),  // src arr ptr
-            src_len.into(),      // src arr len
-            {
-                let s = match ty {
-                    BasicTypeEnum::FloatType(t) => t.size_of(),
-                    BasicTypeEnum::IntType(t) => t.size_of(),
-                    BasicTypeEnum::PointerType(t) => t.size_of(),
-                    BasicTypeEnum::StructType(t) => t.size_of().unwrap(),
-                    _ => unreachable!(),
-                };
-                ctx.builder.build_int_truncate_or_bit_cast(s, int32, "size").unwrap()
-            }
-            .into(),
-        ];
-        ctx.builder
-            .build_call(slice_assign_fun, args.as_slice(), "slice_assign")
-            .map(CallSiteValue::try_as_basic_value)
-            .map(|v| v.map_left(BasicValueEnum::into_int_value))
-            .map(Either::unwrap_left)
-            .unwrap()
-    };
-    // update length
-    let need_update =
-        ctx.builder.build_int_compare(IntPredicate::NE, new_len, dest_len, "need_update").unwrap();
-    let current = ctx.builder.get_insert_block().unwrap().get_parent().unwrap();
-    let update_bb = ctx.ctx.append_basic_block(current, "update");
-    let cont_bb = ctx.ctx.append_basic_block(current, "cont");
-    ctx.builder.build_conditional_branch(need_update, update_bb, cont_bb).unwrap();
-    ctx.builder.position_at_end(update_bb);
-    let new_len = ctx.builder.build_int_z_extend_or_bit_cast(new_len, size_ty, "new_len").unwrap();
-    dest_arr.store_size(ctx, generator, new_len);
-    ctx.builder.build_unconditional_branch(cont_bb).unwrap();
-    ctx.builder.position_at_end(cont_bb);
-}
-
-/// Generates a call to `isinf` in IR. Returns an `i1` representing the result.
-pub fn call_isinf<'ctx, G: CodeGenerator + ?Sized>(
-    generator: &mut G,
-    ctx: &CodeGenContext<'ctx, '_>,
-    v: FloatValue<'ctx>,
-) -> IntValue<'ctx> {
-    let intrinsic_fn = ctx.module.get_function("__nac3_isinf").unwrap_or_else(|| {
-        let fn_type = ctx.ctx.i32_type().fn_type(&[ctx.ctx.f64_type().into()], false);
-        ctx.module.add_function("__nac3_isinf", fn_type, None)
-    });
-
-    let ret = ctx
-        .builder
-        .build_call(intrinsic_fn, &[v.into()], "isinf")
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap();
-
-    generator.bool_to_i1(ctx, ret)
-}
-
-/// Generates a call to `isnan` in IR. Returns an `i1` representing the result.
-pub fn call_isnan<'ctx, G: CodeGenerator + ?Sized>(
-    generator: &mut G,
-    ctx: &CodeGenContext<'ctx, '_>,
-    v: FloatValue<'ctx>,
-) -> IntValue<'ctx> {
-    let intrinsic_fn = ctx.module.get_function("__nac3_isnan").unwrap_or_else(|| {
-        let fn_type = ctx.ctx.i32_type().fn_type(&[ctx.ctx.f64_type().into()], false);
-        ctx.module.add_function("__nac3_isnan", fn_type, None)
-    });
-
-    let ret = ctx
-        .builder
-        .build_call(intrinsic_fn, &[v.into()], "isnan")
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap();
-
-    generator.bool_to_i1(ctx, ret)
-}
-
-/// Generates a call to `gamma` in IR. Returns an `f64` representing the result.
-pub fn call_gamma<'ctx>(ctx: &CodeGenContext<'ctx, '_>, v: FloatValue<'ctx>) -> FloatValue<'ctx> {
-    let llvm_f64 = ctx.ctx.f64_type();
-
-    let intrinsic_fn = ctx.module.get_function("__nac3_gamma").unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        ctx.module.add_function("__nac3_gamma", fn_type, None)
-    });
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[v.into()], "gamma")
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Generates a call to `gammaln` in IR. Returns an `f64` representing the result.
-pub fn call_gammaln<'ctx>(ctx: &CodeGenContext<'ctx, '_>, v: FloatValue<'ctx>) -> FloatValue<'ctx> {
-    let llvm_f64 = ctx.ctx.f64_type();
-
-    let intrinsic_fn = ctx.module.get_function("__nac3_gammaln").unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        ctx.module.add_function("__nac3_gammaln", fn_type, None)
-    });
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[v.into()], "gammaln")
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Generates a call to `j0` in IR. Returns an `f64` representing the result.
-pub fn call_j0<'ctx>(ctx: &CodeGenContext<'ctx, '_>, v: FloatValue<'ctx>) -> FloatValue<'ctx> {
-    let llvm_f64 = ctx.ctx.f64_type();
-
-    let intrinsic_fn = ctx.module.get_function("__nac3_j0").unwrap_or_else(|| {
-        let fn_type = llvm_f64.fn_type(&[llvm_f64.into()], false);
-        ctx.module.add_function("__nac3_j0", fn_type, None)
-    });
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[v.into()], "j0")
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Generates a call to `__nac3_ndarray_calc_size`. Returns an [`IntValue`] representing the
-/// calculated total size.
-///
-/// * `dims` - An [`ArrayLikeIndexer`] containing the size of each dimension.
-/// * `range` - The dimension index to begin and end (exclusively) calculating the dimensions for,
-/// or [`None`] if starting from the first dimension and ending at the last dimension respectively.
-pub fn call_ndarray_calc_size<'ctx, G, Dims>(
-    generator: &G,
-    ctx: &CodeGenContext<'ctx, '_>,
-    dims: &Dims,
-    (begin, end): (Option<IntValue<'ctx>>, Option<IntValue<'ctx>>),
-) -> IntValue<'ctx>
-where
-    G: CodeGenerator + ?Sized,
-    Dims: ArrayLikeIndexer<'ctx>,
-{
-    let llvm_i64 = ctx.ctx.i64_type();
-    let llvm_usize = generator.get_size_type(ctx.ctx);
-
-    let llvm_pi64 = llvm_i64.ptr_type(AddressSpace::default());
-
-    let ndarray_calc_size_fn_name = match llvm_usize.get_bit_width() {
-        32 => "__nac3_ndarray_calc_size",
-        64 => "__nac3_ndarray_calc_size64",
-        bw => unreachable!("Unsupported size type bit width: {}", bw),
-    };
-    let ndarray_calc_size_fn_t = llvm_usize.fn_type(
-        &[llvm_pi64.into(), llvm_usize.into(), llvm_usize.into(), llvm_usize.into()],
-        false,
-    );
-    let ndarray_calc_size_fn =
-        ctx.module.get_function(ndarray_calc_size_fn_name).unwrap_or_else(|| {
-            ctx.module.add_function(ndarray_calc_size_fn_name, ndarray_calc_size_fn_t, None)
-        });
-
-    let begin = begin.unwrap_or_else(|| llvm_usize.const_zero());
-    let end = end.unwrap_or_else(|| dims.size(ctx, generator));
-    ctx.builder
-        .build_call(
-            ndarray_calc_size_fn,
-            &[
-                dims.base_ptr(ctx, generator).into(),
-                dims.size(ctx, generator).into(),
-                begin.into(),
-                end.into(),
-            ],
-            "",
-        )
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Generates a call to `__nac3_ndarray_calc_nd_indices`. Returns a [`TypeArrayLikeAdpater`]
-/// containing `i32` indices of the flattened index.
-///
-/// * `index` - The index to compute the multidimensional index for.
-/// * `ndarray` - LLVM pointer to the `NDArray`. This value must be the LLVM representation of an
-/// `NDArray`.
-pub fn call_ndarray_calc_nd_indices<'ctx, G: CodeGenerator + ?Sized>(
-    generator: &G,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    index: IntValue<'ctx>,
-    ndarray: NDArrayValue<'ctx>,
-) -> TypedArrayLikeAdapter<'ctx, IntValue<'ctx>> {
-    let llvm_void = ctx.ctx.void_type();
-    let llvm_i32 = ctx.ctx.i32_type();
-    let llvm_usize = generator.get_size_type(ctx.ctx);
-    let llvm_pi32 = llvm_i32.ptr_type(AddressSpace::default());
-    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
-
-    let ndarray_calc_nd_indices_fn_name = match llvm_usize.get_bit_width() {
-        32 => "__nac3_ndarray_calc_nd_indices",
-        64 => "__nac3_ndarray_calc_nd_indices64",
-        bw => unreachable!("Unsupported size type bit width: {}", bw),
-    };
-    let ndarray_calc_nd_indices_fn =
-        ctx.module.get_function(ndarray_calc_nd_indices_fn_name).unwrap_or_else(|| {
-            let fn_type = llvm_void.fn_type(
-                &[llvm_usize.into(), llvm_pusize.into(), llvm_usize.into(), llvm_pi32.into()],
-                false,
-            );
-
-            ctx.module.add_function(ndarray_calc_nd_indices_fn_name, fn_type, None)
-        });
-
-    let ndarray_num_dims = ndarray.load_ndims(ctx);
-    let ndarray_dims = ndarray.dim_sizes();
-
-    let indices = ctx.builder.build_array_alloca(llvm_i32, ndarray_num_dims, "").unwrap();
-
-    ctx.builder
-        .build_call(
-            ndarray_calc_nd_indices_fn,
-            &[
-                index.into(),
-                ndarray_dims.base_ptr(ctx, generator).into(),
-                ndarray_num_dims.into(),
-                indices.into(),
-            ],
-            "",
-        )
-        .unwrap();
-
-    TypedArrayLikeAdapter::from(
-        ArraySliceValue::from_ptr_val(indices, ndarray_num_dims, None),
-        Box::new(|_, v| v.into_int_value()),
-        Box::new(|_, v| v.into()),
-    )
-}
-
-fn call_ndarray_flatten_index_impl<'ctx, G, Indices>(
-    generator: &G,
-    ctx: &CodeGenContext<'ctx, '_>,
-    ndarray: NDArrayValue<'ctx>,
-    indices: &Indices,
-) -> IntValue<'ctx>
-where
-    G: CodeGenerator + ?Sized,
-    Indices: ArrayLikeIndexer<'ctx>,
-{
-    let llvm_i32 = ctx.ctx.i32_type();
-    let llvm_usize = generator.get_size_type(ctx.ctx);
-
-    let llvm_pi32 = llvm_i32.ptr_type(AddressSpace::default());
-    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
-
-    debug_assert_eq!(
-        IntType::try_from(indices.element_type(ctx, generator))
-            .map(IntType::get_bit_width)
-            .unwrap_or_default(),
-        llvm_i32.get_bit_width(),
-        "Expected i32 value for argument `indices` to `call_ndarray_flatten_index_impl`"
-    );
-    debug_assert_eq!(
-        indices.size(ctx, generator).get_type().get_bit_width(),
-        llvm_usize.get_bit_width(),
-        "Expected usize integer value for argument `indices_size` to `call_ndarray_flatten_index_impl`"
-    );
-
-    let ndarray_flatten_index_fn_name = match llvm_usize.get_bit_width() {
-        32 => "__nac3_ndarray_flatten_index",
-        64 => "__nac3_ndarray_flatten_index64",
-        bw => unreachable!("Unsupported size type bit width: {}", bw),
-    };
-    let ndarray_flatten_index_fn =
-        ctx.module.get_function(ndarray_flatten_index_fn_name).unwrap_or_else(|| {
-            let fn_type = llvm_usize.fn_type(
-                &[llvm_pusize.into(), llvm_usize.into(), llvm_pi32.into(), llvm_usize.into()],
-                false,
-            );
-
-            ctx.module.add_function(ndarray_flatten_index_fn_name, fn_type, None)
-        });
-
-    let ndarray_num_dims = ndarray.load_ndims(ctx);
-    let ndarray_dims = ndarray.dim_sizes();
-
-    let index = ctx
-        .builder
-        .build_call(
-            ndarray_flatten_index_fn,
-            &[
-                ndarray_dims.base_ptr(ctx, generator).into(),
-                ndarray_num_dims.into(),
-                indices.base_ptr(ctx, generator).into(),
-                indices.size(ctx, generator).into(),
-            ],
-            "",
-        )
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap();
-
-    index
-}
-
-/// Generates a call to `__nac3_ndarray_flatten_index`. Returns the flattened index for the
-/// multidimensional index.
-///
-/// * `ndarray` - LLVM pointer to the `NDArray`. This value must be the LLVM representation of an
-/// `NDArray`.
-/// * `indices` - The multidimensional index to compute the flattened index for.
-pub fn call_ndarray_flatten_index<'ctx, G, Index>(
-    generator: &mut G,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    ndarray: NDArrayValue<'ctx>,
-    indices: &Index,
-) -> IntValue<'ctx>
-where
-    G: CodeGenerator + ?Sized,
-    Index: ArrayLikeIndexer<'ctx>,
-{
-    call_ndarray_flatten_index_impl(generator, ctx, ndarray, indices)
-}
-
-/// Generates a call to `__nac3_ndarray_calc_broadcast`. Returns a tuple containing the number of
-/// dimension and size of each dimension of the resultant `ndarray`.
-pub fn call_ndarray_calc_broadcast<'ctx, G: CodeGenerator + ?Sized>(
-    generator: &mut G,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    lhs: NDArrayValue<'ctx>,
-    rhs: NDArrayValue<'ctx>,
-) -> TypedArrayLikeAdapter<'ctx, IntValue<'ctx>> {
-    let llvm_usize = generator.get_size_type(ctx.ctx);
-    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
-
-    let ndarray_calc_broadcast_fn_name = match llvm_usize.get_bit_width() {
-        32 => "__nac3_ndarray_calc_broadcast",
-        64 => "__nac3_ndarray_calc_broadcast64",
-        bw => unreachable!("Unsupported size type bit width: {}", bw),
-    };
-    let ndarray_calc_broadcast_fn =
-        ctx.module.get_function(ndarray_calc_broadcast_fn_name).unwrap_or_else(|| {
-            let fn_type = llvm_usize.fn_type(
-                &[
-                    llvm_pusize.into(),
-                    llvm_usize.into(),
-                    llvm_pusize.into(),
-                    llvm_usize.into(),
-                    llvm_pusize.into(),
-                ],
-                false,
-            );
-
-            ctx.module.add_function(ndarray_calc_broadcast_fn_name, fn_type, None)
-        });
-
-    let lhs_ndims = lhs.load_ndims(ctx);
-    let rhs_ndims = rhs.load_ndims(ctx);
-    let min_ndims = llvm_intrinsics::call_int_umin(ctx, lhs_ndims, rhs_ndims, None);
-
-    gen_for_callback_incrementing(
-        generator,
-        ctx,
-        llvm_usize.const_zero(),
-        (min_ndims, false),
-        |generator, ctx, idx| {
-            let idx = ctx.builder.build_int_sub(min_ndims, idx, "").unwrap();
-            let (lhs_dim_sz, rhs_dim_sz) = unsafe {
-                (
-                    lhs.dim_sizes().get_typed_unchecked(ctx, generator, &idx, None),
-                    rhs.dim_sizes().get_typed_unchecked(ctx, generator, &idx, None),
-                )
-            };
-
-            let llvm_usize_const_one = llvm_usize.const_int(1, false);
-            let lhs_eqz = ctx
-                .builder
-                .build_int_compare(IntPredicate::EQ, lhs_dim_sz, llvm_usize_const_one, "")
-                .unwrap();
-            let rhs_eqz = ctx
-                .builder
-                .build_int_compare(IntPredicate::EQ, rhs_dim_sz, llvm_usize_const_one, "")
-                .unwrap();
-            let lhs_or_rhs_eqz = ctx.builder.build_or(lhs_eqz, rhs_eqz, "").unwrap();
-
-            let lhs_eq_rhs = ctx
-                .builder
-                .build_int_compare(IntPredicate::EQ, lhs_dim_sz, rhs_dim_sz, "")
-                .unwrap();
-
-            let is_compatible = ctx.builder.build_or(lhs_or_rhs_eqz, lhs_eq_rhs, "").unwrap();
-
-            ctx.make_assert(
-                generator,
-                is_compatible,
-                "0:ValueError",
-                "operands could not be broadcast together",
-                [None, None, None],
-                ctx.current_loc,
-            );
-
-            Ok(())
-        },
-        llvm_usize.const_int(1, false),
-    )
-    .unwrap();
-
-    let max_ndims = llvm_intrinsics::call_int_umax(ctx, lhs_ndims, rhs_ndims, None);
-    let lhs_dims = lhs.dim_sizes().base_ptr(ctx, generator);
-    let lhs_ndims = lhs.load_ndims(ctx);
-    let rhs_dims = rhs.dim_sizes().base_ptr(ctx, generator);
-    let rhs_ndims = rhs.load_ndims(ctx);
-    let out_dims = ctx.builder.build_array_alloca(llvm_usize, max_ndims, "").unwrap();
-    let out_dims = ArraySliceValue::from_ptr_val(out_dims, max_ndims, None);
-
-    ctx.builder
-        .build_call(
-            ndarray_calc_broadcast_fn,
-            &[
-                lhs_dims.into(),
-                lhs_ndims.into(),
-                rhs_dims.into(),
-                rhs_ndims.into(),
-                out_dims.base_ptr(ctx, generator).into(),
-            ],
-            "",
-        )
-        .unwrap();
-
-    TypedArrayLikeAdapter::from(
-        out_dims,
-        Box::new(|_, v| v.into_int_value()),
-        Box::new(|_, v| v.into()),
-    )
-}
-
-/// Generates a call to `__nac3_ndarray_calc_broadcast_idx`. Returns an [`ArrayAllocaValue`]
-/// containing the indices used for accessing `array` corresponding to the index of the broadcasted
-/// array `broadcast_idx`.
-pub fn call_ndarray_calc_broadcast_index<
-    'ctx,
-    G: CodeGenerator + ?Sized,
-    BroadcastIdx: UntypedArrayLikeAccessor<'ctx>,
->(
-    generator: &mut G,
-    ctx: &mut CodeGenContext<'ctx, '_>,
-    array: NDArrayValue<'ctx>,
-    broadcast_idx: &BroadcastIdx,
-) -> TypedArrayLikeAdapter<'ctx, IntValue<'ctx>> {
-    let llvm_i32 = ctx.ctx.i32_type();
-    let llvm_usize = generator.get_size_type(ctx.ctx);
-    let llvm_pi32 = llvm_i32.ptr_type(AddressSpace::default());
-    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
-
-    let ndarray_calc_broadcast_fn_name = match llvm_usize.get_bit_width() {
-        32 => "__nac3_ndarray_calc_broadcast_idx",
-        64 => "__nac3_ndarray_calc_broadcast_idx64",
-        bw => unreachable!("Unsupported size type bit width: {}", bw),
-    };
-    let ndarray_calc_broadcast_fn =
-        ctx.module.get_function(ndarray_calc_broadcast_fn_name).unwrap_or_else(|| {
-            let fn_type = llvm_usize.fn_type(
-                &[llvm_pusize.into(), llvm_usize.into(), llvm_pi32.into(), llvm_pi32.into()],
-                false,
-            );
-
-            ctx.module.add_function(ndarray_calc_broadcast_fn_name, fn_type, None)
-        });
-
-    let broadcast_size = broadcast_idx.size(ctx, generator);
-    let out_idx = ctx.builder.build_array_alloca(llvm_i32, broadcast_size, "").unwrap();
-
-    let array_dims = array.dim_sizes().base_ptr(ctx, generator);
-    let array_ndims = array.load_ndims(ctx);
-    let broadcast_idx_ptr = unsafe {
-        broadcast_idx.ptr_offset_unchecked(ctx, generator, &llvm_usize.const_zero(), None)
-    };
-
-    ctx.builder
-        .build_call(
-            ndarray_calc_broadcast_fn,
-            &[array_dims.into(), array_ndims.into(), broadcast_idx_ptr.into(), out_idx.into()],
-            "",
-        )
-        .unwrap();
-
-    TypedArrayLikeAdapter::from(
-        ArraySliceValue::from_ptr_val(out_idx, broadcast_size, None),
-        Box::new(|_, v| v.into_int_value()),
-        Box::new(|_, v| v.into()),
-    )
-}

nac3core/src/codegen/irrt/ndarray.rs

@@ -0,0 +1,384 @@
+use inkwell::{
+    types::IntType,
+    values::{BasicValueEnum, CallSiteValue, IntValue},
+    AddressSpace, IntPredicate,
+};
+use itertools::Either;
+
+use crate::codegen::{
+    llvm_intrinsics,
+    macros::codegen_unreachable,
+    stmt::gen_for_callback_incrementing,
+    values::{
+        ArrayLikeIndexer, ArrayLikeValue, ArraySliceValue, NDArrayValue, TypedArrayLikeAccessor,
+        TypedArrayLikeAdapter, UntypedArrayLikeAccessor,
+    },
+    CodeGenContext, CodeGenerator,
+};
+
+/// Generates a call to `__nac3_ndarray_calc_size`. Returns an [`IntValue`] representing the
+/// calculated total size.
+///
+/// * `dims` - An [`ArrayLikeIndexer`] containing the size of each dimension.
+/// * `range` - The dimension index to begin and end (exclusively) calculating the dimensions for,
+///   or [`None`] if starting from the first dimension and ending at the last dimension
+///   respectively.
+pub fn call_ndarray_calc_size<'ctx, G, Dims>(
+    generator: &G,
+    ctx: &CodeGenContext<'ctx, '_>,
+    dims: &Dims,
+    (begin, end): (Option<IntValue<'ctx>>, Option<IntValue<'ctx>>),
+) -> IntValue<'ctx>
+where
+    G: CodeGenerator + ?Sized,
+    Dims: ArrayLikeIndexer<'ctx>,
+{
+    let llvm_usize = generator.get_size_type(ctx.ctx);
+    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
+
+    let ndarray_calc_size_fn_name = match llvm_usize.get_bit_width() {
+        32 => "__nac3_ndarray_calc_size",
+        64 => "__nac3_ndarray_calc_size64",
+        bw => codegen_unreachable!(ctx, "Unsupported size type bit width: {}", bw),
+    };
+    let ndarray_calc_size_fn_t = llvm_usize.fn_type(
+        &[llvm_pusize.into(), llvm_usize.into(), llvm_usize.into(), llvm_usize.into()],
+        false,
+    );
+    let ndarray_calc_size_fn =
+        ctx.module.get_function(ndarray_calc_size_fn_name).unwrap_or_else(|| {
+            ctx.module.add_function(ndarray_calc_size_fn_name, ndarray_calc_size_fn_t, None)
+        });
+
+    let begin = begin.unwrap_or_else(|| llvm_usize.const_zero());
+    let end = end.unwrap_or_else(|| dims.size(ctx, generator));
+    ctx.builder
+        .build_call(
+            ndarray_calc_size_fn,
+            &[
+                dims.base_ptr(ctx, generator).into(),
+                dims.size(ctx, generator).into(),
+                begin.into(),
+                end.into(),
+            ],
+            "",
+        )
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_int_value))
+        .map(Either::unwrap_left)
+        .unwrap()
+}
+
+/// Generates a call to `__nac3_ndarray_calc_nd_indices`. Returns a [`TypeArrayLikeAdpater`]
+/// containing `i32` indices of the flattened index.
+///
+/// * `index` - The index to compute the multidimensional index for.
+/// * `ndarray` - LLVM pointer to the `NDArray`. This value must be the LLVM representation of an
+///   `NDArray`.
+pub fn call_ndarray_calc_nd_indices<'ctx, G: CodeGenerator + ?Sized>(
+    generator: &G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    index: IntValue<'ctx>,
+    ndarray: NDArrayValue<'ctx>,
+) -> TypedArrayLikeAdapter<'ctx, IntValue<'ctx>> {
+    let llvm_void = ctx.ctx.void_type();
+    let llvm_i32 = ctx.ctx.i32_type();
+    let llvm_usize = generator.get_size_type(ctx.ctx);
+    let llvm_pi32 = llvm_i32.ptr_type(AddressSpace::default());
+    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
+
+    let ndarray_calc_nd_indices_fn_name = match llvm_usize.get_bit_width() {
+        32 => "__nac3_ndarray_calc_nd_indices",
+        64 => "__nac3_ndarray_calc_nd_indices64",
+        bw => codegen_unreachable!(ctx, "Unsupported size type bit width: {}", bw),
+    };
+    let ndarray_calc_nd_indices_fn =
+        ctx.module.get_function(ndarray_calc_nd_indices_fn_name).unwrap_or_else(|| {
+            let fn_type = llvm_void.fn_type(
+                &[llvm_usize.into(), llvm_pusize.into(), llvm_usize.into(), llvm_pi32.into()],
+                false,
+            );
+
+            ctx.module.add_function(ndarray_calc_nd_indices_fn_name, fn_type, None)
+        });
+
+    let ndarray_num_dims = ndarray.load_ndims(ctx);
+    let ndarray_dims = ndarray.shape();
+
+    let indices = ctx.builder.build_array_alloca(llvm_i32, ndarray_num_dims, "").unwrap();
+
+    ctx.builder
+        .build_call(
+            ndarray_calc_nd_indices_fn,
+            &[
+                index.into(),
+                ndarray_dims.base_ptr(ctx, generator).into(),
+                ndarray_num_dims.into(),
+                indices.into(),
+            ],
+            "",
+        )
+        .unwrap();
+
+    TypedArrayLikeAdapter::from(
+        ArraySliceValue::from_ptr_val(indices, ndarray_num_dims, None),
+        Box::new(|_, v| v.into_int_value()),
+        Box::new(|_, v| v.into()),
+    )
+}
+
+fn call_ndarray_flatten_index_impl<'ctx, G, Indices>(
+    generator: &G,
+    ctx: &CodeGenContext<'ctx, '_>,
+    ndarray: NDArrayValue<'ctx>,
+    indices: &Indices,
+) -> IntValue<'ctx>
+where
+    G: CodeGenerator + ?Sized,
+    Indices: ArrayLikeIndexer<'ctx>,
+{
+    let llvm_i32 = ctx.ctx.i32_type();
+    let llvm_usize = generator.get_size_type(ctx.ctx);
+
+    let llvm_pi32 = llvm_i32.ptr_type(AddressSpace::default());
+    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
+
+    debug_assert_eq!(
+        IntType::try_from(indices.element_type(ctx, generator))
+            .map(IntType::get_bit_width)
+            .unwrap_or_default(),
+        llvm_i32.get_bit_width(),
+        "Expected i32 value for argument `indices` to `call_ndarray_flatten_index_impl`"
+    );
+    debug_assert_eq!(
+        indices.size(ctx, generator).get_type().get_bit_width(),
+        llvm_usize.get_bit_width(),
+        "Expected usize integer value for argument `indices_size` to `call_ndarray_flatten_index_impl`"
+    );
+
+    let ndarray_flatten_index_fn_name = match llvm_usize.get_bit_width() {
+        32 => "__nac3_ndarray_flatten_index",
+        64 => "__nac3_ndarray_flatten_index64",
+        bw => codegen_unreachable!(ctx, "Unsupported size type bit width: {}", bw),
+    };
+    let ndarray_flatten_index_fn =
+        ctx.module.get_function(ndarray_flatten_index_fn_name).unwrap_or_else(|| {
+            let fn_type = llvm_usize.fn_type(
+                &[llvm_pusize.into(), llvm_usize.into(), llvm_pi32.into(), llvm_usize.into()],
+                false,
+            );
+
+            ctx.module.add_function(ndarray_flatten_index_fn_name, fn_type, None)
+        });
+
+    let ndarray_num_dims = ndarray.load_ndims(ctx);
+    let ndarray_dims = ndarray.shape();
+
+    let index = ctx
+        .builder
+        .build_call(
+            ndarray_flatten_index_fn,
+            &[
+                ndarray_dims.base_ptr(ctx, generator).into(),
+                ndarray_num_dims.into(),
+                indices.base_ptr(ctx, generator).into(),
+                indices.size(ctx, generator).into(),
+            ],
+            "",
+        )
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_int_value))
+        .map(Either::unwrap_left)
+        .unwrap();
+
+    index
+}
+
+/// Generates a call to `__nac3_ndarray_flatten_index`. Returns the flattened index for the
+/// multidimensional index.
+///
+/// * `ndarray` - LLVM pointer to the `NDArray`. This value must be the LLVM representation of an
+///   `NDArray`.
+/// * `indices` - The multidimensional index to compute the flattened index for.
+pub fn call_ndarray_flatten_index<'ctx, G, Index>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    ndarray: NDArrayValue<'ctx>,
+    indices: &Index,
+) -> IntValue<'ctx>
+where
+    G: CodeGenerator + ?Sized,
+    Index: ArrayLikeIndexer<'ctx>,
+{
+    call_ndarray_flatten_index_impl(generator, ctx, ndarray, indices)
+}
+
+/// Generates a call to `__nac3_ndarray_calc_broadcast`. Returns a tuple containing the number of
+/// dimension and size of each dimension of the resultant `ndarray`.
+pub fn call_ndarray_calc_broadcast<'ctx, G: CodeGenerator + ?Sized>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    lhs: NDArrayValue<'ctx>,
+    rhs: NDArrayValue<'ctx>,
+) -> TypedArrayLikeAdapter<'ctx, IntValue<'ctx>> {
+    let llvm_usize = generator.get_size_type(ctx.ctx);
+    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
+
+    let ndarray_calc_broadcast_fn_name = match llvm_usize.get_bit_width() {
+        32 => "__nac3_ndarray_calc_broadcast",
+        64 => "__nac3_ndarray_calc_broadcast64",
+        bw => codegen_unreachable!(ctx, "Unsupported size type bit width: {}", bw),
+    };
+    let ndarray_calc_broadcast_fn =
+        ctx.module.get_function(ndarray_calc_broadcast_fn_name).unwrap_or_else(|| {
+            let fn_type = llvm_usize.fn_type(
+                &[
+                    llvm_pusize.into(),
+                    llvm_usize.into(),
+                    llvm_pusize.into(),
+                    llvm_usize.into(),
+                    llvm_pusize.into(),
+                ],
+                false,
+            );
+
+            ctx.module.add_function(ndarray_calc_broadcast_fn_name, fn_type, None)
+        });
+
+    let lhs_ndims = lhs.load_ndims(ctx);
+    let rhs_ndims = rhs.load_ndims(ctx);
+    let min_ndims = llvm_intrinsics::call_int_umin(ctx, lhs_ndims, rhs_ndims, None);
+
+    gen_for_callback_incrementing(
+        generator,
+        ctx,
+        None,
+        llvm_usize.const_zero(),
+        (min_ndims, false),
+        |generator, ctx, _, idx| {
+            let idx = ctx.builder.build_int_sub(min_ndims, idx, "").unwrap();
+            let (lhs_dim_sz, rhs_dim_sz) = unsafe {
+                (
+                    lhs.shape().get_typed_unchecked(ctx, generator, &idx, None),
+                    rhs.shape().get_typed_unchecked(ctx, generator, &idx, None),
+                )
+            };
+
+            let llvm_usize_const_one = llvm_usize.const_int(1, false);
+            let lhs_eqz = ctx
+                .builder
+                .build_int_compare(IntPredicate::EQ, lhs_dim_sz, llvm_usize_const_one, "")
+                .unwrap();
+            let rhs_eqz = ctx
+                .builder
+                .build_int_compare(IntPredicate::EQ, rhs_dim_sz, llvm_usize_const_one, "")
+                .unwrap();
+            let lhs_or_rhs_eqz = ctx.builder.build_or(lhs_eqz, rhs_eqz, "").unwrap();
+
+            let lhs_eq_rhs = ctx
+                .builder
+                .build_int_compare(IntPredicate::EQ, lhs_dim_sz, rhs_dim_sz, "")
+                .unwrap();
+
+            let is_compatible = ctx.builder.build_or(lhs_or_rhs_eqz, lhs_eq_rhs, "").unwrap();
+
+            ctx.make_assert(
+                generator,
+                is_compatible,
+                "0:ValueError",
+                "operands could not be broadcast together",
+                [None, None, None],
+                ctx.current_loc,
+            );
+
+            Ok(())
+        },
+        llvm_usize.const_int(1, false),
+    )
+    .unwrap();
+
+    let max_ndims = llvm_intrinsics::call_int_umax(ctx, lhs_ndims, rhs_ndims, None);
+    let lhs_dims = lhs.shape().base_ptr(ctx, generator);
+    let lhs_ndims = lhs.load_ndims(ctx);
+    let rhs_dims = rhs.shape().base_ptr(ctx, generator);
+    let rhs_ndims = rhs.load_ndims(ctx);
+    let out_dims = ctx.builder.build_array_alloca(llvm_usize, max_ndims, "").unwrap();
+    let out_dims = ArraySliceValue::from_ptr_val(out_dims, max_ndims, None);
+
+    ctx.builder
+        .build_call(
+            ndarray_calc_broadcast_fn,
+            &[
+                lhs_dims.into(),
+                lhs_ndims.into(),
+                rhs_dims.into(),
+                rhs_ndims.into(),
+                out_dims.base_ptr(ctx, generator).into(),
+            ],
+            "",
+        )
+        .unwrap();
+
+    TypedArrayLikeAdapter::from(
+        out_dims,
+        Box::new(|_, v| v.into_int_value()),
+        Box::new(|_, v| v.into()),
+    )
+}
+
+/// Generates a call to `__nac3_ndarray_calc_broadcast_idx`. Returns an [`ArrayAllocaValue`]
+/// containing the indices used for accessing `array` corresponding to the index of the broadcasted
+/// array `broadcast_idx`.
+pub fn call_ndarray_calc_broadcast_index<
+    'ctx,
+    G: CodeGenerator + ?Sized,
+    BroadcastIdx: UntypedArrayLikeAccessor<'ctx>,
+>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    array: NDArrayValue<'ctx>,
+    broadcast_idx: &BroadcastIdx,
+) -> TypedArrayLikeAdapter<'ctx, IntValue<'ctx>> {
+    let llvm_i32 = ctx.ctx.i32_type();
+    let llvm_usize = generator.get_size_type(ctx.ctx);
+    let llvm_pi32 = llvm_i32.ptr_type(AddressSpace::default());
+    let llvm_pusize = llvm_usize.ptr_type(AddressSpace::default());
+
+    let ndarray_calc_broadcast_fn_name = match llvm_usize.get_bit_width() {
+        32 => "__nac3_ndarray_calc_broadcast_idx",
+        64 => "__nac3_ndarray_calc_broadcast_idx64",
+        bw => codegen_unreachable!(ctx, "Unsupported size type bit width: {}", bw),
+    };
+    let ndarray_calc_broadcast_fn =
+        ctx.module.get_function(ndarray_calc_broadcast_fn_name).unwrap_or_else(|| {
+            let fn_type = llvm_usize.fn_type(
+                &[llvm_pusize.into(), llvm_usize.into(), llvm_pi32.into(), llvm_pi32.into()],
+                false,
+            );
+
+            ctx.module.add_function(ndarray_calc_broadcast_fn_name, fn_type, None)
+        });
+
+    let broadcast_size = broadcast_idx.size(ctx, generator);
+    let out_idx = ctx.builder.build_array_alloca(llvm_i32, broadcast_size, "").unwrap();
+
+    let array_dims = array.shape().base_ptr(ctx, generator);
+    let array_ndims = array.load_ndims(ctx);
+    let broadcast_idx_ptr = unsafe {
+        broadcast_idx.ptr_offset_unchecked(ctx, generator, &llvm_usize.const_zero(), None)
+    };
+
+    ctx.builder
+        .build_call(
+            ndarray_calc_broadcast_fn,
+            &[array_dims.into(), array_ndims.into(), broadcast_idx_ptr.into(), out_idx.into()],
+            "",
+        )
+        .unwrap();
+
+    TypedArrayLikeAdapter::from(
+        ArraySliceValue::from_ptr_val(out_idx, broadcast_size, None),
+        Box::new(|_, v| v.into_int_value()),
+        Box::new(|_, v| v.into()),
+    )
+}

nac3core/src/codegen/irrt/slice.rs

@@ -0,0 +1,76 @@
+use inkwell::{
+    values::{BasicValueEnum, CallSiteValue, IntValue},
+    IntPredicate,
+};
+use itertools::Either;
+use nac3parser::ast::Expr;
+
+use crate::{
+    codegen::{CodeGenContext, CodeGenerator},
+    typecheck::typedef::Type,
+};
+
+/// this function allows index out of range, since python
+/// allows index out of range in slice (`a = [1,2,3]; a[1:10] == [2,3]`).
+pub fn handle_slice_index_bound<'ctx, G: CodeGenerator>(
+    i: &Expr<Option<Type>>,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    generator: &mut G,
+    length: IntValue<'ctx>,
+) -> Result<Option<IntValue<'ctx>>, String> {
+    const SYMBOL: &str = "__nac3_slice_index_bound";
+    let func = ctx.module.get_function(SYMBOL).unwrap_or_else(|| {
+        let i32_t = ctx.ctx.i32_type();
+        let fn_t = i32_t.fn_type(&[i32_t.into(), i32_t.into()], false);
+        ctx.module.add_function(SYMBOL, fn_t, None)
+    });
+
+    let i = if let Some(v) = generator.gen_expr(ctx, i)? {
+        v.to_basic_value_enum(ctx, generator, i.custom.unwrap())?
+    } else {
+        return Ok(None);
+    };
+    Ok(Some(
+        ctx.builder
+            .build_call(func, &[i.into(), length.into()], "bounded_ind")
+            .map(CallSiteValue::try_as_basic_value)
+            .map(|v| v.map_left(BasicValueEnum::into_int_value))
+            .map(Either::unwrap_left)
+            .unwrap(),
+    ))
+}
+
+pub fn calculate_len_for_slice_range<'ctx, G: CodeGenerator + ?Sized>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    start: IntValue<'ctx>,
+    end: IntValue<'ctx>,
+    step: IntValue<'ctx>,
+) -> IntValue<'ctx> {
+    const SYMBOL: &str = "__nac3_range_slice_len";
+    let len_func = ctx.module.get_function(SYMBOL).unwrap_or_else(|| {
+        let i32_t = ctx.ctx.i32_type();
+        let fn_t = i32_t.fn_type(&[i32_t.into(), i32_t.into(), i32_t.into()], false);
+        ctx.module.add_function(SYMBOL, fn_t, None)
+    });
+
+    // assert step != 0, throw exception if not
+    let not_zero = ctx
+        .builder
+        .build_int_compare(IntPredicate::NE, step, step.get_type().const_zero(), "range_step_ne")
+        .unwrap();
+    ctx.make_assert(
+        generator,
+        not_zero,
+        "0:ValueError",
+        "step must not be zero",
+        [None, None, None],
+        ctx.current_loc,
+    );
+    ctx.builder
+        .build_call(len_func, &[start.into(), end.into(), step.into()], "calc_len")
+        .map(CallSiteValue::try_as_basic_value)
+        .map(|v| v.map_left(BasicValueEnum::into_int_value))
+        .map(Either::unwrap_left)
+        .unwrap()
+}

nac3core/src/codegen/llvm_intrinsics.rs

@@ -1,12 +1,14 @@
-use crate::codegen::CodeGenContext;
-use inkwell::context::Context;
-use inkwell::intrinsics::Intrinsic;
-use inkwell::types::AnyTypeEnum::IntType;
-use inkwell::types::FloatType;
-use inkwell::values::{BasicValueEnum, CallSiteValue, FloatValue, IntValue, PointerValue};
-use inkwell::AddressSpace;
+use inkwell::{
+    context::Context,
+    intrinsics::Intrinsic,
+    types::{AnyTypeEnum::IntType, FloatType},
+    values::{BasicValueEnum, CallSiteValue, FloatValue, IntValue, PointerValue},
+    AddressSpace,
+};
 use itertools::Either;
 
+use super::CodeGenContext;
+
 /// Returns the string representation for the floating-point type `ft` when used in intrinsic
 /// functions.
 fn get_float_intrinsic_repr(ctx: &Context, ft: FloatType) -> &'static str {
@@ -35,6 +37,40 @@ fn get_float_intrinsic_repr(ctx: &Context, ft: FloatType) -> &'static str {
     unreachable!()
 }
 
+/// Invokes the [`llvm.va_start`](https://llvm.org/docs/LangRef.html#llvm-va-start-intrinsic)
+/// intrinsic.
+pub fn call_va_start<'ctx>(ctx: &CodeGenContext<'ctx, '_>, arglist: PointerValue<'ctx>) {
+    const FN_NAME: &str = "llvm.va_start";
+
+    let intrinsic_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
+        let llvm_void = ctx.ctx.void_type();
+        let llvm_i8 = ctx.ctx.i8_type();
+        let llvm_p0i8 = llvm_i8.ptr_type(AddressSpace::default());
+        let fn_type = llvm_void.fn_type(&[llvm_p0i8.into()], false);
+
+        ctx.module.add_function(FN_NAME, fn_type, None)
+    });
+
+    ctx.builder.build_call(intrinsic_fn, &[arglist.into()], "").unwrap();
+}
+
+/// Invokes the [`llvm.va_start`](https://llvm.org/docs/LangRef.html#llvm-va-start-intrinsic)
+/// intrinsic.
+pub fn call_va_end<'ctx>(ctx: &CodeGenContext<'ctx, '_>, arglist: PointerValue<'ctx>) {
+    const FN_NAME: &str = "llvm.va_end";
+
+    let intrinsic_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
+        let llvm_void = ctx.ctx.void_type();
+        let llvm_i8 = ctx.ctx.i8_type();
+        let llvm_p0i8 = llvm_i8.ptr_type(AddressSpace::default());
+        let fn_type = llvm_void.fn_type(&[llvm_p0i8.into()], false);
+
+        ctx.module.add_function(FN_NAME, fn_type, None)
+    });
+
+    ctx.builder.build_call(intrinsic_fn, &[arglist.into()], "").unwrap();
+}
+
 /// Invokes the [`llvm.stacksave`](https://llvm.org/docs/LangRef.html#llvm-stacksave-intrinsic)
 /// intrinsic.
 pub fn call_stacksave<'ctx>(
@@ -62,145 +98,30 @@ pub fn call_stacksave<'ctx>(
 pub fn call_stackrestore<'ctx>(ctx: &CodeGenContext<'ctx, '_>, ptr: PointerValue<'ctx>) {
     const FN_NAME: &str = "llvm.stackrestore";
 
-    let llvm_i8 = ctx.ctx.i8_type();
-    let llvm_p0i8 = llvm_i8.ptr_type(AddressSpace::default());
+    /*
+        SEE https://github.com/TheDan64/inkwell/issues/496
 
+        We want `llvm.stackrestore`, but the following would generate `llvm.stackrestore.p0i8`.
+        ```ignore
         let intrinsic_fn = Intrinsic::find(FN_NAME)
             .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_p0i8.into()]))
             .unwrap();
+        ```
+
+        Temp workaround by manually declaring the intrinsic with the correct function name instead.
+    */
+    let intrinsic_fn = ctx.module.get_function(FN_NAME).unwrap_or_else(|| {
+        let llvm_void = ctx.ctx.void_type();
+        let llvm_i8 = ctx.ctx.i8_type();
+        let llvm_p0i8 = llvm_i8.ptr_type(AddressSpace::default());
+        let fn_type = llvm_void.fn_type(&[llvm_p0i8.into()], false);
+
+        ctx.module.add_function(FN_NAME, fn_type, None)
+    });
 
     ctx.builder.build_call(intrinsic_fn, &[ptr.into()], "").unwrap();
 }
 
-/// Invokes the [`llvm.abs`](https://llvm.org/docs/LangRef.html#llvm-abs-intrinsic) intrinsic.
-///
-/// * `src` - The value for which the absolute value is to be returned.
-/// * `is_int_min_poison` - Whether `poison` is to be returned if `src` is `INT_MIN`.
-pub fn call_int_abs<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    src: IntValue<'ctx>,
-    is_int_min_poison: IntValue<'ctx>,
-    name: Option<&str>,
-) -> IntValue<'ctx> {
-    const FN_NAME: &str = "llvm.abs";
-
-    debug_assert_eq!(is_int_min_poison.get_type().get_bit_width(), 1);
-    debug_assert!(is_int_min_poison.is_const());
-
-    let llvm_src_t = src.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_src_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[src.into(), is_int_min_poison.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.smax`](https://llvm.org/docs/LangRef.html#llvm-smax-intrinsic) intrinsic.
-pub fn call_int_smax<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    a: IntValue<'ctx>,
-    b: IntValue<'ctx>,
-    name: Option<&str>,
-) -> IntValue<'ctx> {
-    const FN_NAME: &str = "llvm.smax";
-
-    debug_assert_eq!(a.get_type().get_bit_width(), b.get_type().get_bit_width());
-
-    let llvm_int_t = a.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_int_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[a.into(), b.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.smin`](https://llvm.org/docs/LangRef.html#llvm-smin-intrinsic) intrinsic.
-pub fn call_int_smin<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    a: IntValue<'ctx>,
-    b: IntValue<'ctx>,
-    name: Option<&str>,
-) -> IntValue<'ctx> {
-    const FN_NAME: &str = "llvm.smin";
-
-    debug_assert_eq!(a.get_type().get_bit_width(), b.get_type().get_bit_width());
-
-    let llvm_int_t = a.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_int_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[a.into(), b.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.umax`](https://llvm.org/docs/LangRef.html#llvm-umax-intrinsic) intrinsic.
-pub fn call_int_umax<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    a: IntValue<'ctx>,
-    b: IntValue<'ctx>,
-    name: Option<&str>,
-) -> IntValue<'ctx> {
-    const FN_NAME: &str = "llvm.umax";
-
-    debug_assert_eq!(a.get_type().get_bit_width(), b.get_type().get_bit_width());
-
-    let llvm_int_t = a.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_int_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[a.into(), b.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.umin`](https://llvm.org/docs/LangRef.html#llvm-umin-intrinsic) intrinsic.
-pub fn call_int_umin<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    a: IntValue<'ctx>,
-    b: IntValue<'ctx>,
-    name: Option<&str>,
-) -> IntValue<'ctx> {
-    const FN_NAME: &str = "llvm.umin";
-
-    debug_assert_eq!(a.get_type().get_bit_width(), b.get_type().get_bit_width());
-
-    let llvm_int_t = a.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_int_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[a.into(), b.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
 /// Invokes the [`llvm.memcpy`](https://llvm.org/docs/LangRef.html#llvm-memcpy-intrinsic) intrinsic.
 ///
 /// * `dest` - The pointer to the destination. Must be a pointer to an integer type.
@@ -264,7 +185,7 @@ pub fn call_memcpy_generic<'ctx>(
         dest
     } else {
         ctx.builder
-            .build_bitcast(dest, llvm_p0i8, "")
+            .build_bit_cast(dest, llvm_p0i8, "")
             .map(BasicValueEnum::into_pointer_value)
             .unwrap()
     };
@@ -272,7 +193,7 @@ pub fn call_memcpy_generic<'ctx>(
         src
     } else {
         ctx.builder
-            .build_bitcast(src, llvm_p0i8, "")
+            .build_bit_cast(src, llvm_p0i8, "")
             .map(BasicValueEnum::into_pointer_value)
             .unwrap()
     };
@@ -280,27 +201,122 @@ pub fn call_memcpy_generic<'ctx>(
     call_memcpy(ctx, dest, src, len, is_volatile);
 }
 
-/// Invokes the [`llvm.sqrt`](https://llvm.org/docs/LangRef.html#llvm-sqrt-intrinsic) intrinsic.
-pub fn call_float_sqrt<'ctx>(
+/// Macro to find and generate build call for llvm intrinsic (body of llvm intrinsic function)
+///
+/// Arguments:
+/// * `$ctx:ident`: Reference to the current Code Generation Context
+/// * `$name:ident`: Optional name to be assigned to the llvm build call (Option<&str>)
+/// * `$llvm_name:literal`: Name of underlying llvm intrinsic function
+/// * `$map_fn:ident`: Mapping function to be applied on `BasicValue` (`BasicValue` -> Function Return Type).
+///   Use `BasicValueEnum::into_int_value` for Integer return type and
+///   `BasicValueEnum::into_float_value` for Float return type
+/// * `$llvm_ty:ident`: Type of first operand
+/// * `,($val:ident)*`: Comma separated list of operands
+macro_rules! generate_llvm_intrinsic_fn_body {
+    ($ctx:ident, $name:ident, $llvm_name:literal, $map_fn:expr, $llvm_ty:ident $(,$val:ident)*) => {{
+        const FN_NAME: &str = concat!("llvm.", $llvm_name);
+        let intrinsic_fn = Intrinsic::find(FN_NAME).and_then(|intrinsic| intrinsic.get_declaration(&$ctx.module, &[$llvm_ty.into()])).unwrap();
+        $ctx.builder.build_call(intrinsic_fn, &[$($val.into()),*],  $name.unwrap_or_default()).map(CallSiteValue::try_as_basic_value).map(|v| v.map_left($map_fn)).map(Either::unwrap_left).unwrap()
+    }};
+}
+
+/// Macro to generate the llvm intrinsic function using [`generate_llvm_intrinsic_fn_body`].
+///
+/// Arguments:
+/// * `float/int`: Indicates the return and argument type of the function
+/// * `$fn_name:ident`: The identifier of the rust function to be generated
+/// * `$llvm_name:literal`: Name of underlying llvm intrinsic function.
+///   Omit "llvm." prefix from the function name i.e. use "ceil" instead of "llvm.ceil"
+/// * `$val:ident`: The operand for unary operations
+/// * `$val1:ident`, `$val2:ident`: The operands for binary operations
+macro_rules! generate_llvm_intrinsic_fn {
+    ("float", $fn_name:ident, $llvm_name:literal, $val:ident) => {
+        #[doc = concat!("Invokes the [`", stringify!($llvm_name), "`](https://llvm.org/docs/LangRef.html#llvm-", stringify!($llvm_name), "-intrinsic) intrinsic." )]
+        pub fn $fn_name<'ctx> (
             ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
+            $val: FloatValue<'ctx>,
             name: Option<&str>,
         ) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.sqrt";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
+            let llvm_ty = $val.get_type();
+            generate_llvm_intrinsic_fn_body!(ctx, name, $llvm_name, BasicValueEnum::into_float_value, llvm_ty, $val)
         }
+    };
+    ("float", $fn_name:ident, $llvm_name:literal, $val1:ident, $val2:ident) => {
+        #[doc = concat!("Invokes the [`", stringify!($llvm_name), "`](https://llvm.org/docs/LangRef.html#llvm-", stringify!($llvm_name), "-intrinsic) intrinsic." )]
+        pub fn $fn_name<'ctx> (
+            ctx: &CodeGenContext<'ctx, '_>,
+            $val1: FloatValue<'ctx>,
+            $val2: FloatValue<'ctx>,
+            name: Option<&str>,
+        ) -> FloatValue<'ctx> {
+            debug_assert_eq!($val1.get_type(), $val2.get_type());
+            let llvm_ty = $val1.get_type();
+            generate_llvm_intrinsic_fn_body!(ctx, name, $llvm_name, BasicValueEnum::into_float_value, llvm_ty, $val1, $val2)
+        }
+    };
+    ("int", $fn_name:ident, $llvm_name:literal, $val1:ident, $val2:ident) => {
+        #[doc = concat!("Invokes the [`", stringify!($llvm_name), "`](https://llvm.org/docs/LangRef.html#llvm-", stringify!($llvm_name), "-intrinsic) intrinsic." )]
+        pub fn $fn_name<'ctx> (
+            ctx: &CodeGenContext<'ctx, '_>,
+            $val1: IntValue<'ctx>,
+            $val2: IntValue<'ctx>,
+            name: Option<&str>,
+        ) -> IntValue<'ctx> {
+            debug_assert_eq!($val1.get_type().get_bit_width(), $val2.get_type().get_bit_width());
+            let llvm_ty = $val1.get_type();
+            generate_llvm_intrinsic_fn_body!(ctx, name, $llvm_name, BasicValueEnum::into_int_value, llvm_ty, $val1, $val2)
+        }
+    };
+}
+
+/// Invokes the [`llvm.abs`](https://llvm.org/docs/LangRef.html#llvm-abs-intrinsic) intrinsic.
+///
+/// * `src` - The value for which the absolute value is to be returned.
+/// * `is_int_min_poison` - Whether `poison` is to be returned if `src` is `INT_MIN`.
+pub fn call_int_abs<'ctx>(
+    ctx: &CodeGenContext<'ctx, '_>,
+    src: IntValue<'ctx>,
+    is_int_min_poison: IntValue<'ctx>,
+    name: Option<&str>,
+) -> IntValue<'ctx> {
+    debug_assert_eq!(is_int_min_poison.get_type().get_bit_width(), 1);
+    debug_assert!(is_int_min_poison.is_const());
+
+    let src_type = src.get_type();
+    generate_llvm_intrinsic_fn_body!(
+        ctx,
+        name,
+        "abs",
+        BasicValueEnum::into_int_value,
+        src_type,
+        src,
+        is_int_min_poison
+    )
+}
+
+generate_llvm_intrinsic_fn!("int", call_int_smax, "smax", a, b);
+generate_llvm_intrinsic_fn!("int", call_int_smin, "smin", a, b);
+generate_llvm_intrinsic_fn!("int", call_int_umax, "umax", a, b);
+generate_llvm_intrinsic_fn!("int", call_int_umin, "umin", a, b);
+generate_llvm_intrinsic_fn!("int", call_expect, "expect", val, expected_val);
+
+generate_llvm_intrinsic_fn!("float", call_float_sqrt, "sqrt", val);
+generate_llvm_intrinsic_fn!("float", call_float_sin, "sin", val);
+generate_llvm_intrinsic_fn!("float", call_float_cos, "cos", val);
+generate_llvm_intrinsic_fn!("float", call_float_pow, "pow", val, power);
+generate_llvm_intrinsic_fn!("float", call_float_exp, "exp", val);
+generate_llvm_intrinsic_fn!("float", call_float_exp2, "exp2", val);
+generate_llvm_intrinsic_fn!("float", call_float_log, "log", val);
+generate_llvm_intrinsic_fn!("float", call_float_log10, "log10", val);
+generate_llvm_intrinsic_fn!("float", call_float_log2, "log2", val);
+generate_llvm_intrinsic_fn!("float", call_float_fabs, "fabs", src);
+generate_llvm_intrinsic_fn!("float", call_float_minnum, "minnum", val, power);
+generate_llvm_intrinsic_fn!("float", call_float_maxnum, "maxnum", val, power);
+generate_llvm_intrinsic_fn!("float", call_float_copysign, "copysign", mag, sgn);
+generate_llvm_intrinsic_fn!("float", call_float_floor, "floor", val);
+generate_llvm_intrinsic_fn!("float", call_float_ceil, "ceil", val);
+generate_llvm_intrinsic_fn!("float", call_float_round, "round", val);
+generate_llvm_intrinsic_fn!("float", call_float_rint, "rint", val);
 
 /// Invokes the [`llvm.powi`](https://llvm.org/docs/LangRef.html#llvm-powi-intrinsic) intrinsic.
 pub fn call_float_powi<'ctx>(
@@ -327,393 +343,3 @@ pub fn call_float_powi<'ctx>(
         .map(Either::unwrap_left)
         .unwrap()
 }
-
-/// Invokes the [`llvm.sin`](https://llvm.org/docs/LangRef.html#llvm-sin-intrinsic) intrinsic.
-pub fn call_float_sin<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.sin";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.cos`](https://llvm.org/docs/LangRef.html#llvm-cos-intrinsic) intrinsic.
-pub fn call_float_cos<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.cos";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.pow`](https://llvm.org/docs/LangRef.html#llvm-pow-intrinsic) intrinsic.
-pub fn call_float_pow<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    power: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.pow";
-
-    debug_assert_eq!(val.get_type(), power.get_type());
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into(), power.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.exp`](https://llvm.org/docs/LangRef.html#llvm-exp-intrinsic) intrinsic.
-pub fn call_float_exp<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.exp";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.exp2`](https://llvm.org/docs/LangRef.html#llvm-exp2-intrinsic) intrinsic.
-pub fn call_float_exp2<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.exp2";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.log`](https://llvm.org/docs/LangRef.html#llvm-log-intrinsic) intrinsic.
-pub fn call_float_log<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.log";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.log10`](https://llvm.org/docs/LangRef.html#llvm-log10-intrinsic) intrinsic.
-pub fn call_float_log10<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.log10";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.log2`](https://llvm.org/docs/LangRef.html#llvm-log2-intrinsic) intrinsic.
-pub fn call_float_log2<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.log2";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.fabs`](https://llvm.org/docs/LangRef.html#llvm-fabs-intrinsic) intrinsic.
-pub fn call_float_fabs<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    src: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.fabs";
-
-    let llvm_src_t = src.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_src_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[src.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.minnum`](https://llvm.org/docs/LangRef.html#llvm-minnum-intrinsic) intrinsic.
-pub fn call_float_minnum<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val1: FloatValue<'ctx>,
-    val2: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.minnum";
-
-    debug_assert_eq!(val1.get_type(), val2.get_type());
-
-    let llvm_float_t = val1.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val1.into(), val2.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.maxnum`](https://llvm.org/docs/LangRef.html#llvm-maxnum-intrinsic) intrinsic.
-pub fn call_float_maxnum<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val1: FloatValue<'ctx>,
-    val2: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.maxnum";
-
-    debug_assert_eq!(val1.get_type(), val2.get_type());
-
-    let llvm_float_t = val1.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val1.into(), val2.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.copysign`](https://llvm.org/docs/LangRef.html#llvm-copysign-intrinsic) intrinsic.
-pub fn call_float_copysign<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    mag: FloatValue<'ctx>,
-    sgn: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.copysign";
-
-    debug_assert_eq!(mag.get_type(), sgn.get_type());
-
-    let llvm_float_t = mag.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[mag.into(), sgn.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.floor`](https://llvm.org/docs/LangRef.html#llvm-floor-intrinsic) intrinsic.
-pub fn call_float_floor<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.floor";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.ceil`](https://llvm.org/docs/LangRef.html#llvm-ceil-intrinsic) intrinsic.
-pub fn call_float_ceil<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.ceil";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.round`](https://llvm.org/docs/LangRef.html#llvm-round-intrinsic) intrinsic.
-pub fn call_float_round<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.round";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the
-/// [`llvm.roundeven`](https://llvm.org/docs/LangRef.html#llvm-roundeven-intrinsic) intrinsic.
-pub fn call_float_roundeven<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: FloatValue<'ctx>,
-    name: Option<&str>,
-) -> FloatValue<'ctx> {
-    const FN_NAME: &str = "llvm.roundeven";
-
-    let llvm_float_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_float_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_float_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}
-
-/// Invokes the [`llvm.expect`](https://llvm.org/docs/LangRef.html#llvm-expect-intrinsic) intrinsic.
-pub fn call_expect<'ctx>(
-    ctx: &CodeGenContext<'ctx, '_>,
-    val: IntValue<'ctx>,
-    expected_val: IntValue<'ctx>,
-    name: Option<&str>,
-) -> IntValue<'ctx> {
-    const FN_NAME: &str = "llvm.expect";
-
-    debug_assert_eq!(val.get_type().get_bit_width(), expected_val.get_type().get_bit_width());
-
-    let llvm_int_t = val.get_type();
-
-    let intrinsic_fn = Intrinsic::find(FN_NAME)
-        .and_then(|intrinsic| intrinsic.get_declaration(&ctx.module, &[llvm_int_t.into()]))
-        .unwrap();
-
-    ctx.builder
-        .build_call(intrinsic_fn, &[val.into(), expected_val.into()], name.unwrap_or_default())
-        .map(CallSiteValue::try_as_basic_value)
-        .map(|v| v.map_left(BasicValueEnum::into_int_value))
-        .map(Either::unwrap_left)
-        .unwrap()
-}

nac3core/src/codegen/mod.rs

@@ -1,12 +1,12 @@
-use crate::{
-    codegen::classes::{ListType, NDArrayType, ProxyType, RangeType},
-    symbol_resolver::{StaticValue, SymbolResolver},
-    toplevel::{helper::PrimDef, numpy::unpack_ndarray_var_tys, TopLevelContext, TopLevelDef},
-    typecheck::{
-        type_inferencer::{CodeLocation, PrimitiveStore},
-        typedef::{CallId, FuncArg, Type, TypeEnum, Unifier},
+use std::{
+    collections::{HashMap, HashSet},
+    sync::{
+        atomic::{AtomicBool, Ordering},
+        Arc,
     },
+    thread,
 };
+
 use crossbeam::channel::{unbounded, Receiver, Sender};
 use inkwell::{
     attributes::{Attribute, AttributeLoc},
@@ -24,17 +24,23 @@ use inkwell::{
     AddressSpace, IntPredicate, OptimizationLevel,
 };
 use itertools::Itertools;
-use nac3parser::ast::{Location, Stmt, StrRef};
 use parking_lot::{Condvar, Mutex};
-use std::collections::{HashMap, HashSet};
-use std::sync::{
-    atomic::{AtomicBool, Ordering},
-    Arc,
+
+use nac3parser::ast::{Location, Stmt, StrRef};
+
+use crate::{
+    symbol_resolver::{StaticValue, SymbolResolver},
+    toplevel::{helper::PrimDef, numpy::unpack_ndarray_var_tys, TopLevelContext, TopLevelDef},
+    typecheck::{
+        type_inferencer::{CodeLocation, PrimitiveStore},
+        typedef::{CallId, FuncArg, Type, TypeEnum, Unifier},
+    },
 };
-use std::thread;
+use concrete_type::{ConcreteType, ConcreteTypeEnum, ConcreteTypeStore};
+pub use generator::{CodeGenerator, DefaultCodeGenerator};
+use types::{ListType, NDArrayType, ProxyType, RangeType};
 
 pub mod builtin_fns;
-pub mod classes;
 pub mod concrete_type;
 pub mod expr;
 pub mod extern_fns;
@@ -43,12 +49,27 @@ pub mod irrt;
 pub mod llvm_intrinsics;
 pub mod numpy;
 pub mod stmt;
+pub mod types;
+pub mod values;
 
 #[cfg(test)]
 mod test;
 
-use concrete_type::{ConcreteType, ConcreteTypeEnum, ConcreteTypeStore};
-pub use generator::{CodeGenerator, DefaultCodeGenerator};
+mod macros {
+    /// Codegen-variant of [`std::unreachable`] which accepts an instance of [`CodeGenContext`] as
+    /// its first argument to provide Python source information to indicate the codegen location
+    /// causing the assertion.
+    macro_rules! codegen_unreachable {
+        ($ctx:expr $(,)?) => {
+            std::unreachable!("unreachable code while processing {}", &$ctx.current_loc)
+        };
+        ($ctx:expr, $($arg:tt)*) => {
+            std::unreachable!("unreachable code while processing {}: {}", &$ctx.current_loc, std::format!("{}", std::format_args!($($arg)+)))
+        };
+    }
+
+    pub(crate) use codegen_unreachable;
+}
 
 #[derive(Default)]
 pub struct StaticValueStore {
@@ -68,6 +89,16 @@ pub struct CodeGenLLVMOptions {
     pub target: CodeGenTargetMachineOptions,
 }
 
+impl CodeGenLLVMOptions {
+    /// Creates a [`TargetMachine`] using the target options specified by this struct.
+    ///
+    /// See [`Target::create_target_machine`].
+    #[must_use]
+    pub fn create_target_machine(&self) -> Option<TargetMachine> {
+        self.target.create_target_machine(self.opt_level)
+    }
+}
+
 /// Additional options for code generation for the target machine.
 #[derive(Clone, Debug, Eq, PartialEq)]
 pub struct CodeGenTargetMachineOptions {
@@ -338,6 +369,10 @@ impl WorkerRegistry {
         let mut builder = context.create_builder();
         let mut module = context.create_module(generator.get_name());
 
+        let target_machine = self.llvm_options.create_target_machine().unwrap();
+        module.set_data_layout(&target_machine.get_target_data().get_data_layout());
+        module.set_triple(&target_machine.get_triple());
+
         module.add_basic_value_flag(
             "Debug Info Version",
             inkwell::module::FlagBehavior::Warning,
@@ -361,6 +396,10 @@ impl WorkerRegistry {
                     errors.insert(e);
                     // create a new empty module just to continue codegen and collect errors
                     module = context.create_module(&format!("{}_recover", generator.get_name()));
+
+                    let target_machine = self.llvm_options.create_target_machine().unwrap();
+                    module.set_data_layout(&target_machine.get_target_data().get_data_layout());
+                    module.set_triple(&target_machine.get_triple());
                 }
             }
             *self.task_count.lock() -= 1;
@@ -426,7 +465,7 @@ pub struct CodeGenTask {
 fn get_llvm_type<'ctx, G: CodeGenerator + ?Sized>(
     ctx: &'ctx Context,
     module: &Module<'ctx>,
-    generator: &mut G,
+    generator: &G,
     unifier: &mut Unifier,
     top_level: &TopLevelContext,
     type_cache: &mut HashMap<Type, BasicTypeEnum<'ctx>>,
@@ -456,6 +495,20 @@ fn get_llvm_type<'ctx, G: CodeGenerator + ?Sized>(
                             .into()
                         }
 
+                        TObj { obj_id, params, .. } if *obj_id == PrimDef::List.id() => {
+                            let element_type = get_llvm_type(
+                                ctx,
+                                module,
+                                generator,
+                                unifier,
+                                top_level,
+                                type_cache,
+                                *params.iter().next().unwrap().1,
+                            );
+
+                            ListType::new(generator, ctx, element_type).as_base_type().into()
+                        }
+
                         TObj { obj_id, .. } if *obj_id == PrimDef::NDArray.id() => {
                             let (dtype, _) = unpack_ndarray_var_tys(unifier, ty);
                             let element_type = get_llvm_type(
@@ -506,8 +559,10 @@ fn get_llvm_type<'ctx, G: CodeGenerator + ?Sized>(
                 };
                 return ty;
             }
-            TTuple { ty } => {
+            TTuple { ty, is_vararg_ctx } => {
                 // a struct with fields in the order present in the tuple
+                assert!(!is_vararg_ctx, "Tuples in vararg context must be instantiated with the correct number of arguments before calling get_llvm_type");
+
                 let fields = ty
                     .iter()
                     .map(|ty| {
@@ -516,12 +571,6 @@ fn get_llvm_type<'ctx, G: CodeGenerator + ?Sized>(
                     .collect_vec();
                 ctx.struct_type(&fields, false).into()
             }
-            TList { ty } => {
-                let element_type =
-                    get_llvm_type(ctx, module, generator, unifier, top_level, type_cache, *ty);
-
-                ListType::new(generator, ctx, element_type).as_base_type().into()
-            }
             TVirtual { .. } => unimplemented!(),
             _ => unreachable!("{}", ty_enum.get_type_name()),
         };
@@ -543,7 +592,7 @@ fn get_llvm_type<'ctx, G: CodeGenerator + ?Sized>(
 fn get_llvm_abi_type<'ctx, G: CodeGenerator + ?Sized>(
     ctx: &'ctx Context,
     module: &Module<'ctx>,
-    generator: &mut G,
+    generator: &G,
     unifier: &mut Unifier,
     top_level: &TopLevelContext,
     type_cache: &mut HashMap<Type, BasicTypeEnum<'ctx>>,
@@ -552,11 +601,11 @@ fn get_llvm_abi_type<'ctx, G: CodeGenerator + ?Sized>(
 ) -> BasicTypeEnum<'ctx> {
     // If the type is used in the definition of a function, return `i1` instead of `i8` for ABI
     // consistency.
-    return if unifier.unioned(ty, primitives.bool) {
+    if unifier.unioned(ty, primitives.bool) {
         ctx.bool_type().into()
     } else {
         get_llvm_type(ctx, module, generator, unifier, top_level, type_cache, ty)
-    };
+    }
 }
 
 /// Whether `sret` is needed for a return value with type `ty`.
@@ -581,6 +630,40 @@ fn need_sret(ty: BasicTypeEnum) -> bool {
     need_sret_impl(ty, true)
 }
 
+/// Returns the [`BasicTypeEnum`] representing a `va_list` struct for variadic arguments.
+fn get_llvm_valist_type<'ctx>(ctx: &'ctx Context, triple: &TargetTriple) -> BasicTypeEnum<'ctx> {
+    let triple = TargetMachine::normalize_triple(triple);
+    let triple = triple.as_str().to_str().unwrap();
+    let arch = triple.split('-').next().unwrap();
+
+    let llvm_pi8 = ctx.i8_type().ptr_type(AddressSpace::default());
+
+    // Referenced from parseArch() in llvm/lib/Support/Triple.cpp
+    match arch {
+        "i386" | "i486" | "i586" | "i686" | "riscv32" => {
+            ctx.i8_type().ptr_type(AddressSpace::default()).into()
+        }
+        "amd64" | "x86_64" | "x86_64h" => {
+            let llvm_i32 = ctx.i32_type();
+
+            let va_list_tag = ctx.opaque_struct_type("struct.__va_list_tag");
+            va_list_tag.set_body(
+                &[llvm_i32.into(), llvm_i32.into(), llvm_pi8.into(), llvm_pi8.into()],
+                false,
+            );
+            va_list_tag.into()
+        }
+        "armv7" => {
+            let va_list = ctx.opaque_struct_type("struct.__va_list");
+            va_list.set_body(&[llvm_pi8.into()], false);
+            va_list.into()
+        }
+        triple => {
+            todo!("Unsupported platform for varargs: {triple}")
+        }
+    }
+}
+
 /// Implementation for generating LLVM IR for a function.
 pub fn gen_func_impl<
     'ctx,
@@ -692,6 +775,7 @@ pub fn gen_func_impl<
                 name: arg.name,
                 ty: task.store.to_unifier_type(&mut unifier, &primitives, arg.ty, &mut cache),
                 default_value: arg.default_value.clone(),
+                is_vararg: arg.is_vararg,
             })
             .collect_vec(),
         task.store.to_unifier_type(&mut unifier, &primitives, *ret, &mut cache),
@@ -714,7 +798,10 @@ pub fn gen_func_impl<
     let has_sret = ret_type.map_or(false, |ty| need_sret(ty));
     let mut params = args
         .iter()
+        .filter(|arg| !arg.is_vararg)
         .map(|arg| {
+            debug_assert!(!arg.is_vararg);
+
             get_llvm_abi_type(
                 context,
                 &module,
@@ -733,9 +820,12 @@ pub fn gen_func_impl<
         params.insert(0, ret_type.unwrap().ptr_type(AddressSpace::default()).into());
     }
 
+    debug_assert!(matches!(args.iter().filter(|arg| arg.is_vararg).count(), 0..=1));
+    let vararg_arg = args.iter().find(|arg| arg.is_vararg);
+
     let fn_type = match ret_type {
-        Some(ret_type) if !has_sret => ret_type.fn_type(&params, false),
-        _ => context.void_type().fn_type(&params, false),
+        Some(ret_type) if !has_sret => ret_type.fn_type(&params, vararg_arg.is_some()),
+        _ => context.void_type().fn_type(&params, vararg_arg.is_some()),
     };
 
     let symbol = &task.symbol_name;
@@ -763,9 +853,10 @@ pub fn gen_func_impl<
     builder.position_at_end(init_bb);
     let body_bb = context.append_basic_block(fn_val, "body");
 
+    // Store non-vararg argument values into local variables
     let mut var_assignment = HashMap::new();
     let offset = u32::from(has_sret);
-    for (n, arg) in args.iter().enumerate() {
+    for (n, arg) in args.iter().enumerate().filter(|(_, arg)| !arg.is_vararg) {
         let param = fn_val.get_nth_param((n as u32) + offset).unwrap();
         let local_type = get_llvm_type(
             context,
@@ -798,6 +889,8 @@ pub fn gen_func_impl<
         var_assignment.insert(arg.name, (alloca, None, 0));
     }
 
+    // TODO: Save vararg parameters as list
+
     let return_buffer = if has_sret {
         Some(fn_val.get_nth_param(0).unwrap().into_pointer_value())
     } else {
@@ -1020,3 +1113,9 @@ fn gen_in_range_check<'ctx>(
 
     ctx.builder.build_int_compare(IntPredicate::SLT, lo, hi, "cmp").unwrap()
 }
+
+/// Returns the internal name for the `va_count` argument, used to indicate the number of arguments
+/// passed to the variadic function.
+fn get_va_count_arg_name(arg_name: StrRef) -> StrRef {
+    format!("__{}_va_count", &arg_name).into()
+}

nac3core/src/codegen/stmt.rs

@@ -1,18 +1,3 @@
-use super::{
-    super::symbol_resolver::ValueEnum,
-    expr::destructure_range,
-    irrt::{handle_slice_indices, list_slice_assignment},
-    CodeGenContext, CodeGenerator,
-};
-use crate::{
-    codegen::{
-        classes::{ArrayLikeIndexer, ArraySliceValue, ListValue, RangeValue},
-        expr::gen_binop_expr,
-        gen_in_range_check,
-    },
-    toplevel::{helper::PrimDef, numpy::unpack_ndarray_var_tys, DefinitionId, TopLevelDef},
-    typecheck::typedef::{FunSignature, Type, TypeEnum},
-};
 use inkwell::{
     attributes::{Attribute, AttributeLoc},
     basic_block::BasicBlock,
@@ -20,10 +5,28 @@ use inkwell::{
     values::{BasicValue, BasicValueEnum, FunctionValue, IntValue, PointerValue},
     IntPredicate,
 };
+use itertools::{izip, Itertools};
+
 use nac3parser::ast::{
     Constant, ExcepthandlerKind, Expr, ExprKind, Location, Stmt, StmtKind, StrRef,
 };
-use std::convert::TryFrom;
+
+use super::{
+    expr::{destructure_range, gen_binop_expr},
+    gen_in_range_check,
+    irrt::{handle_slice_indices, list_slice_assignment},
+    macros::codegen_unreachable,
+    values::{ArrayLikeIndexer, ArraySliceValue, ListValue, RangeValue},
+    CodeGenContext, CodeGenerator,
+};
+use crate::{
+    symbol_resolver::ValueEnum,
+    toplevel::{DefinitionId, TopLevelDef},
+    typecheck::{
+        magic_methods::Binop,
+        typedef::{iter_type_vars, FunSignature, Type, TypeEnum},
+    },
+};
 
 /// See [`CodeGenerator::gen_var_alloc`].
 pub fn gen_var<'ctx>(
@@ -94,8 +97,6 @@ pub fn gen_store_target<'ctx, G: CodeGenerator>(
     pattern: &Expr<Option<Type>>,
     name: Option<&str>,
 ) -> Result<Option<PointerValue<'ctx>>, String> {
-    let llvm_usize = generator.get_size_type(ctx.ctx);
-
     // very similar to gen_expr, but we don't do an extra load at the end
     // and we flatten nested tuples
     Ok(Some(match &pattern.node {
@@ -120,7 +121,7 @@ pub fn gen_store_target<'ctx, G: CodeGenerator>(
                 return Ok(None);
             };
             let BasicValueEnum::PointerValue(ptr) = val else {
-                unreachable!();
+                codegen_unreachable!(ctx);
             };
             unsafe {
                 ctx.builder.build_in_bounds_gep(
@@ -134,66 +135,7 @@ pub fn gen_store_target<'ctx, G: CodeGenerator>(
             }
             .unwrap()
         }
-        ExprKind::Subscript { value, slice, .. } => {
-            match ctx.unifier.get_ty_immutable(value.custom.unwrap()).as_ref() {
-                TypeEnum::TList { .. } => {
-                    let v = generator
-                        .gen_expr(ctx, value)?
-                        .unwrap()
-                        .to_basic_value_enum(ctx, generator, value.custom.unwrap())?
-                        .into_pointer_value();
-                    let v = ListValue::from_ptr_val(v, llvm_usize, None);
-                    let len = v.load_size(ctx, Some("len"));
-                    let raw_index = generator
-                        .gen_expr(ctx, slice)?
-                        .unwrap()
-                        .to_basic_value_enum(ctx, generator, slice.custom.unwrap())?
-                        .into_int_value();
-                    let raw_index = ctx
-                        .builder
-                        .build_int_s_extend(raw_index, generator.get_size_type(ctx.ctx), "sext")
-                        .unwrap();
-                    // handle negative index
-                    let is_negative = ctx
-                        .builder
-                        .build_int_compare(
-                            IntPredicate::SLT,
-                            raw_index,
-                            generator.get_size_type(ctx.ctx).const_zero(),
-                            "is_neg",
-                        )
-                        .unwrap();
-                    let adjusted = ctx.builder.build_int_add(raw_index, len, "adjusted").unwrap();
-                    let index = ctx
-                        .builder
-                        .build_select(is_negative, adjusted, raw_index, "index")
-                        .map(BasicValueEnum::into_int_value)
-                        .unwrap();
-                    // unsigned less than is enough, because negative index after adjustment is
-                    // bigger than the length (for unsigned cmp)
-                    let bound_check = ctx
-                        .builder
-                        .build_int_compare(IntPredicate::ULT, index, len, "inbound")
-                        .unwrap();
-                    ctx.make_assert(
-                        generator,
-                        bound_check,
-                        "0:IndexError",
-                        "index {0} out of bounds 0:{1}",
-                        [Some(raw_index), Some(len), None],
-                        slice.location,
-                    );
-                    v.data().ptr_offset(ctx, generator, &index, name)
-                }
-
-                TypeEnum::TObj { obj_id, .. } if *obj_id == PrimDef::NDArray.id() => {
-                    todo!()
-                }
-
-                _ => unreachable!(),
-            }
-        }
-        _ => unreachable!(),
+        _ => codegen_unreachable!(ctx),
     }))
 }
 
@@ -203,68 +145,20 @@ pub fn gen_assign<'ctx, G: CodeGenerator>(
     ctx: &mut CodeGenContext<'ctx, '_>,
     target: &Expr<Option<Type>>,
     value: ValueEnum<'ctx>,
+    value_ty: Type,
 ) -> Result<(), String> {
-    let llvm_usize = generator.get_size_type(ctx.ctx);
-
+    // See https://docs.python.org/3/reference/simple_stmts.html#assignment-statements.
     match &target.node {
-        ExprKind::Tuple { elts, .. } => {
-            let BasicValueEnum::StructValue(v) =
-                value.to_basic_value_enum(ctx, generator, target.custom.unwrap())?
-            else {
-                unreachable!()
-            };
-
-            for (i, elt) in elts.iter().enumerate() {
-                let v = ctx
-                    .builder
-                    .build_extract_value(v, u32::try_from(i).unwrap(), "struct_elem")
-                    .unwrap();
-                generator.gen_assign(ctx, elt, v.into())?;
+        ExprKind::Subscript { value: target, slice: key, .. } => {
+            // Handle "slicing" or "subscription"
+            generator.gen_setitem(ctx, target, key, value, value_ty)?;
         }
-        }
-        ExprKind::Subscript { value: ls, slice, .. }
-            if matches!(&slice.node, ExprKind::Slice { .. }) =>
-        {
-            let ExprKind::Slice { lower, upper, step } = &slice.node else { unreachable!() };
-
-            let ls = generator
-                .gen_expr(ctx, ls)?
-                .unwrap()
-                .to_basic_value_enum(ctx, generator, ls.custom.unwrap())?
-                .into_pointer_value();
-            let ls = ListValue::from_ptr_val(ls, llvm_usize, None);
-            let Some((start, end, step)) =
-                handle_slice_indices(lower, upper, step, ctx, generator, ls.load_size(ctx, None))?
-            else {
-                return Ok(());
-            };
-            let value = value
-                .to_basic_value_enum(ctx, generator, target.custom.unwrap())?
-                .into_pointer_value();
-            let value = ListValue::from_ptr_val(value, llvm_usize, None);
-            let ty = match &*ctx.unifier.get_ty_immutable(target.custom.unwrap()) {
-                TypeEnum::TList { ty } => *ty,
-                TypeEnum::TObj { obj_id, .. } if *obj_id == PrimDef::NDArray.id() => {
-                    unpack_ndarray_var_tys(&mut ctx.unifier, target.custom.unwrap()).0
-                }
-                _ => unreachable!(),
-            };
-
-            let ty = ctx.get_llvm_type(generator, ty);
-            let Some(src_ind) = handle_slice_indices(
-                &None,
-                &None,
-                &None,
-                ctx,
-                generator,
-                value.load_size(ctx, None),
-            )?
-            else {
-                return Ok(());
-            };
-            list_slice_assignment(generator, ctx, ty, ls, (start, end, step), value, src_ind);
+        ExprKind::Tuple { elts, .. } | ExprKind::List { elts, .. } => {
+            // Fold on `"[" [target_list] "]"` and `"(" [target_list] ")"`
+            generator.gen_assign_target_list(ctx, elts, value, value_ty)?;
         }
         _ => {
+            // Handle attribute and direct variable assignments.
             let name = if let ExprKind::Name { id, .. } = &target.node {
                 format!("{id}.addr")
             } else {
@@ -282,19 +176,259 @@ pub fn gen_assign<'ctx, G: CodeGenerator>(
                 }
             }
             let val = value.to_basic_value_enum(ctx, generator, target.custom.unwrap())?;
+
+            // Perform i1 <-> i8 conversion as needed
+            let val = if ctx.unifier.unioned(target.custom.unwrap(), ctx.primitives.bool) {
+                generator.bool_to_i8(ctx, val.into_int_value()).into()
+            } else {
+                val
+            };
+
             ctx.builder.build_store(ptr, val).unwrap();
         }
     };
     Ok(())
 }
 
+/// See [`CodeGenerator::gen_assign_target_list`].
+pub fn gen_assign_target_list<'ctx, G: CodeGenerator>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    targets: &Vec<Expr<Option<Type>>>,
+    value: ValueEnum<'ctx>,
+    value_ty: Type,
+) -> Result<(), String> {
+    // Deconstruct the tuple `value`
+    let BasicValueEnum::StructValue(tuple) = value.to_basic_value_enum(ctx, generator, value_ty)?
+    else {
+        codegen_unreachable!(ctx)
+    };
+
+    // NOTE: Currently, RHS's type is forced to be a Tuple by the type inferencer.
+    let TypeEnum::TTuple { ty: tuple_tys, .. } = &*ctx.unifier.get_ty(value_ty) else {
+        codegen_unreachable!(ctx);
+    };
+
+    assert_eq!(tuple.get_type().count_fields() as usize, tuple_tys.len());
+
+    let tuple = (0..tuple.get_type().count_fields())
+        .map(|i| ctx.builder.build_extract_value(tuple, i, "item").unwrap())
+        .collect_vec();
+
+    // Find the starred target if it exists.
+    let mut starred_target_index: Option<usize> = None; // Index of the "starred" target. If it exists, there may only be one.
+    for (i, target) in targets.iter().enumerate() {
+        if matches!(target.node, ExprKind::Starred { .. }) {
+            assert!(starred_target_index.is_none()); // The typechecker ensures this
+            starred_target_index = Some(i);
+        }
+    }
+
+    if let Some(starred_target_index) = starred_target_index {
+        assert!(tuple_tys.len() >= targets.len() - 1); // The typechecker ensures this
+
+        let a = starred_target_index; // Number of RHS values before the starred target
+        let b = tuple_tys.len() - (targets.len() - 1 - starred_target_index); // Number of RHS values after the starred target
+                                                                              // Thus `tuple[a..b]` is assigned to the starred target.
+
+        // Handle assignment before the starred target
+        for (target, val, val_ty) in
+            izip!(&targets[..starred_target_index], &tuple[..a], &tuple_tys[..a])
+        {
+            generator.gen_assign(ctx, target, ValueEnum::Dynamic(*val), *val_ty)?;
+        }
+
+        // Handle assignment to the starred target
+        if let ExprKind::Starred { value: target, .. } = &targets[starred_target_index].node {
+            let vals = &tuple[a..b];
+            let val_tys = &tuple_tys[a..b];
+
+            // Create a sub-tuple from `value` for the starred target.
+            let sub_tuple_ty = ctx
+                .ctx
+                .struct_type(&vals.iter().map(BasicValueEnum::get_type).collect_vec(), false);
+            let psub_tuple_val =
+                ctx.builder.build_alloca(sub_tuple_ty, "starred_target_value_ptr").unwrap();
+            for (i, val) in vals.iter().enumerate() {
+                let pitem = ctx
+                    .builder
+                    .build_struct_gep(psub_tuple_val, i as u32, "starred_target_value_item")
+                    .unwrap();
+                ctx.builder.build_store(pitem, *val).unwrap();
+            }
+            let sub_tuple_val =
+                ctx.builder.build_load(psub_tuple_val, "starred_target_value").unwrap();
+
+            // Create the typechecker type of the sub-tuple
+            let sub_tuple_ty =
+                ctx.unifier.add_ty(TypeEnum::TTuple { ty: val_tys.to_vec(), is_vararg_ctx: false });
+
+            // Now assign with that sub-tuple to the starred target.
+            generator.gen_assign(ctx, target, ValueEnum::Dynamic(sub_tuple_val), sub_tuple_ty)?;
+        } else {
+            codegen_unreachable!(ctx) // The typechecker ensures this
+        }
+
+        // Handle assignment after the starred target
+        for (target, val, val_ty) in
+            izip!(&targets[starred_target_index + 1..], &tuple[b..], &tuple_tys[b..])
+        {
+            generator.gen_assign(ctx, target, ValueEnum::Dynamic(*val), *val_ty)?;
+        }
+    } else {
+        assert_eq!(tuple_tys.len(), targets.len()); // The typechecker ensures this
+
+        for (target, val, val_ty) in izip!(targets, tuple, tuple_tys) {
+            generator.gen_assign(ctx, target, ValueEnum::Dynamic(val), *val_ty)?;
+        }
+    }
+    Ok(())
+}
+
+/// See [`CodeGenerator::gen_setitem`].
+pub fn gen_setitem<'ctx, G: CodeGenerator>(
+    generator: &mut G,
+    ctx: &mut CodeGenContext<'ctx, '_>,
+    target: &Expr<Option<Type>>,
+    key: &Expr<Option<Type>>,
+    value: ValueEnum<'ctx>,
+    value_ty: Type,
+) -> Result<(), String> {
+    let target_ty = target.custom.unwrap();
+    let key_ty = key.custom.unwrap();
+
+    match &*ctx.unifier.get_ty(target_ty) {
+        TypeEnum::TObj { obj_id, params: list_params, .. }
+            if *obj_id == ctx.primitives.list.obj_id(&ctx.unifier).unwrap() =>
+        {
+            // Handle list item assignment
+            let llvm_usize = generator.get_size_type(ctx.ctx);
+            let target_item_ty = iter_type_vars(list_params).next().unwrap().ty;
+
+            let target = generator
+                .gen_expr(ctx, target)?
+                .unwrap()
+                .to_basic_value_enum(ctx, generator, target_ty)?
+                .into_pointer_value();
+            let target = ListValue::from_pointer_value(target, llvm_usize, None);
+
+            if let ExprKind::Slice { .. } = &key.node {
+                // Handle assigning to a slice
+                let ExprKind::Slice { lower, upper, step } = &key.node else {
+                    codegen_unreachable!(ctx)
+                };
+                let Some((start, end, step)) = handle_slice_indices(
+                    lower,
+                    upper,
+                    step,
+                    ctx,
+                    generator,
+                    target.load_size(ctx, None),
+                )?
+                else {
+                    return Ok(());
+                };
+
+                let value =
+                    value.to_basic_value_enum(ctx, generator, value_ty)?.into_pointer_value();
+                let value = ListValue::from_pointer_value(value, llvm_usize, None);
+
+                let target_item_ty = ctx.get_llvm_type(generator, target_item_ty);
+                let Some(src_ind) = handle_slice_indices(
+                    &None,
+                    &None,
+                    &None,
+                    ctx,
+                    generator,
+                    value.load_size(ctx, None),
+                )?
+                else {
+                    return Ok(());
+                };
+                list_slice_assignment(
+                    generator,
+                    ctx,
+                    target_item_ty,
+                    target,
+                    (start, end, step),
+                    value,
+                    src_ind,
+                );
+            } else {
+                // Handle assigning to an index
+                let len = target.load_size(ctx, Some("len"));
+
+                let index = generator
+                    .gen_expr(ctx, key)?
+                    .unwrap()
+                    .to_basic_value_enum(ctx, generator, key_ty)?
+                    .into_int_value();
+                let index = ctx
+                    .builder
+                    .build_int_s_extend(index, generator.get_size_type(ctx.ctx), "sext")
+                    .unwrap();
+
+                // handle negative index
+                let is_negative = ctx
+                    .builder
+                    .build_int_compare(
+                        IntPredicate::SLT,
+                        index,
+                        generator.get_size_type(ctx.ctx).const_zero(),
+                        "is_neg",
+                    )
+                    .unwrap();
+                let adjusted = ctx.builder.build_int_add(index, len, "adjusted").unwrap();
+                let index = ctx
+                    .builder
+                    .build_select(is_negative, adjusted, index, "index")
+                    .map(BasicValueEnum::into_int_value)
+                    .unwrap();
+
+                // unsigned less than is enough, because negative index after adjustment is
+                // bigger than the length (for unsigned cmp)
+                let bound_check = ctx
+                    .builder
+                    .build_int_compare(IntPredicate::ULT, index, len, "inbound")
+                    .unwrap();
+                ctx.make_assert(
+                    generator,
+                    bound_check,
+                    "0:IndexError",
+                    "index {0} out of bounds 0:{1}",
+                    [Some(index), Some(len), None],
+                    key.location,
+                );
+
+                // Write value to index on list
+                let item_ptr =
+                    target.data().ptr_offset(ctx, generator, &index, Some("list_item_ptr"));
+                let value = value.to_basic_value_enum(ctx, generator, value_ty)?;
+                ctx.builder.build_store(item_ptr, value).unwrap();
+            }
+        }
+        TypeEnum::TObj { obj_id, .. }
+            if *obj_id == ctx.primitives.ndarray.obj_id(&ctx.unifier).unwrap() =>
+        {
+            // Handle NDArray item assignment
+            todo!("ndarray subscript assignment is not yet implemented");
+        }
+        _ => {
+            panic!("encountered unknown target type: {}", ctx.unifier.stringify(target_ty));
+        }
+    }
+    Ok(())
+}
+
 /// See [`CodeGenerator::gen_for`].
 pub fn gen_for<G: CodeGenerator>(
     generator: &mut G,
     ctx: &mut CodeGenContext<'_, '_>,
     stmt: &Stmt<Option<Type>>,
 ) -> Result<(), String> {
-    let StmtKind::For { iter, target, body, orelse, .. } = &stmt.node else { unreachable!() };
+    let StmtKind::For { iter, target, body, orelse, .. } = &stmt.node else {
+        codegen_unreachable!(ctx)
+    };
 
     // var_assignment static values may be changed in another branch
     // if so, remove the static value as it may not be correct in this branch
@@ -310,9 +444,6 @@ pub fn gen_for<G: CodeGenerator>(
     let orelse_bb =
         if orelse.is_empty() { cont_bb } else { ctx.ctx.append_basic_block(current, "for.orelse") };
 
-    // Whether the iterable is a range() expression
-    let is_iterable_range_expr = ctx.unifier.unioned(iter.custom.unwrap(), ctx.primitives.range);
-
     // The BB containing the increment expression
     let incr_bb = ctx.ctx.append_basic_block(current, "for.incr");
     // The BB containing the loop condition check
@@ -321,27 +452,36 @@ pub fn gen_for<G: CodeGenerator>(
     // store loop bb information and restore it later
     let loop_bb = ctx.loop_target.replace((incr_bb, cont_bb));
 
+    let iter_ty = iter.custom.unwrap();
     let iter_val = if let Some(v) = generator.gen_expr(ctx, iter)? {
-        v.to_basic_value_enum(ctx, generator, iter.custom.unwrap())?
+        v.to_basic_value_enum(ctx, generator, iter_ty)?
     } else {
         return Ok(());
     };
-    if is_iterable_range_expr {
-        let iter_val = RangeValue::from_ptr_val(iter_val.into_pointer_value(), Some("range"));
+
+    match &*ctx.unifier.get_ty(iter_ty) {
+        TypeEnum::TObj { obj_id, .. }
+            if *obj_id == ctx.primitives.range.obj_id(&ctx.unifier).unwrap() =>
+        {
+            let iter_val =
+                RangeValue::from_pointer_value(iter_val.into_pointer_value(), Some("range"));
             // Internal variable for loop; Cannot be assigned
             let i = generator.gen_var_alloc(ctx, int32.into(), Some("for.i.addr"))?;
             // Variable declared in "target" expression of the loop; Can be reassigned *or* shadowed
-        let Some(target_i) = generator.gen_store_target(ctx, target, Some("for.target.addr"))?
+            let Some(target_i) =
+                generator.gen_store_target(ctx, target, Some("for.target.addr"))?
             else {
-            unreachable!()
+                codegen_unreachable!(ctx)
             };
             let (start, stop, step) = destructure_range(ctx, iter_val);
 
             ctx.builder.build_store(i, start).unwrap();
 
             // Check "If step is zero, ValueError is raised."
-        let rangenez =
-            ctx.builder.build_int_compare(IntPredicate::NE, step, int32.const_zero(), "").unwrap();
+            let rangenez = ctx
+                .builder
+                .build_int_compare(IntPredicate::NE, step, int32.const_zero(), "")
+                .unwrap();
             ctx.make_assert(
                 generator,
                 rangenez,
@@ -358,7 +498,10 @@ pub fn gen_for<G: CodeGenerator>(
                     .build_conditional_branch(
                         gen_in_range_check(
                             ctx,
-                        ctx.builder.build_load(i, "").map(BasicValueEnum::into_int_value).unwrap(),
+                            ctx.builder
+                                .build_load(i, "")
+                                .map(BasicValueEnum::into_int_value)
+                                .unwrap(),
                             stop,
                             step,
                         ),
@@ -388,7 +531,10 @@ pub fn gen_for<G: CodeGenerator>(
                 )
                 .unwrap();
             generator.gen_block(ctx, body.iter())?;
-    } else {
+        }
+        TypeEnum::TObj { obj_id, params: list_params, .. }
+            if *obj_id == ctx.primitives.list.obj_id(&ctx.unifier).unwrap() =>
+        {
             let index_addr = generator.gen_var_alloc(ctx, size_t.into(), Some("for.index.addr"))?;
             ctx.builder.build_store(index_addr, size_t.const_zero()).unwrap();
             let len = ctx
@@ -426,9 +572,14 @@ pub fn gen_for<G: CodeGenerator>(
                 .map(BasicValueEnum::into_int_value)
                 .unwrap();
             let val = ctx.build_gep_and_load(arr_ptr, &[index], Some("val"));
-        generator.gen_assign(ctx, target, val.into())?;
+            let val_ty = iter_type_vars(list_params).next().unwrap().ty;
+            generator.gen_assign(ctx, target, val.into(), val_ty)?;
             generator.gen_block(ctx, body.iter())?;
         }
+        _ => {
+            panic!("unsupported for loop iterator type: {}", ctx.unifier.stringify(iter_ty));
+        }
+    }
 
     for (k, (_, _, counter)) in &var_assignment {
         let (_, static_val, counter2) = ctx.var_assignment.get_mut(k).unwrap();
@@ -462,6 +613,16 @@ pub fn gen_for<G: CodeGenerator>(
     Ok(())
 }
 
+#[derive(PartialEq, Eq, Debug, Clone, Copy, Hash)]
+pub struct BreakContinueHooks<'ctx> {
+    /// The [exit block][`BasicBlock`] to branch to when `break`-ing out of a loop.
+    pub exit_bb: BasicBlock<'ctx>,
+
+    /// The [latch basic block][`BasicBlock`] to branch to for `continue`-ing to the next iteration
+    /// of the loop.
+    pub latch_bb: BasicBlock<'ctx>,
+}
+
 /// Generates a C-style `for` construct using lambdas, similar to the following C code:
 ///
 /// ```c
@@ -479,6 +640,7 @@ pub fn gen_for<G: CodeGenerator>(
 pub fn gen_for_callback<'ctx, 'a, G, I, InitFn, CondFn, BodyFn, UpdateFn>(
     generator: &mut G,
     ctx: &mut CodeGenContext<'ctx, 'a>,
+    label: Option<&str>,
     init: InitFn,
     cond: CondFn,
     body: BodyFn,
@@ -489,17 +651,24 @@ where
     I: Clone,
     InitFn: FnOnce(&mut G, &mut CodeGenContext<'ctx, 'a>) -> Result<I, String>,
     CondFn: FnOnce(&mut G, &mut CodeGenContext<'ctx, 'a>, I) -> Result<IntValue<'ctx>, String>,
-    BodyFn: FnOnce(&mut G, &mut CodeGenContext<'ctx, 'a>, I) -> Result<(), String>,
+    BodyFn: FnOnce(
+        &mut G,
+        &mut CodeGenContext<'ctx, 'a>,
+        BreakContinueHooks<'ctx>,
+        I,
+    ) -> Result<(), String>,
     UpdateFn: FnOnce(&mut G, &mut CodeGenContext<'ctx, 'a>, I) -> Result<(), String>,
 {
+    let label = label.unwrap_or("for");
+
     let current_bb = ctx.builder.get_insert_block().unwrap();
-    let init_bb = ctx.ctx.insert_basic_block_after(current_bb, "for.init");
+    let init_bb = ctx.ctx.insert_basic_block_after(current_bb, &format!("{label}.init"));
     // The BB containing the loop condition check
-    let cond_bb = ctx.ctx.insert_basic_block_after(init_bb, "for.cond");
-    let body_bb = ctx.ctx.insert_basic_block_after(cond_bb, "for.body");
+    let cond_bb = ctx.ctx.insert_basic_block_after(init_bb, &format!("{label}.cond"));
+    let body_bb = ctx.ctx.insert_basic_block_after(cond_bb, &format!("{label}.body"));
     // The BB containing the increment expression
-    let update_bb = ctx.ctx.insert_basic_block_after(body_bb, "for.update");
-    let cont_bb = ctx.ctx.insert_basic_block_after(update_bb, "for.end");
+    let update_bb = ctx.ctx.insert_basic_block_after(body_bb, &format!("{label}.update"));
+    let cont_bb = ctx.ctx.insert_basic_block_after(update_bb, &format!("{label}.end"));
 
     // store loop bb information and restore it later
     let loop_bb = ctx.loop_target.replace((update_bb, cont_bb));
@@ -520,7 +689,8 @@ where
     }
 
     ctx.builder.position_at_end(body_bb);
-    body(generator, ctx, loop_var.clone())?;
+    let hooks = BreakContinueHooks { exit_bb: cont_bb, latch_bb: update_bb };
+    body(generator, ctx, hooks, loop_var.clone())?;
     if !ctx.is_terminated() {
         ctx.builder.build_unconditional_branch(update_bb).unwrap();
     }
@@ -555,6 +725,7 @@ where
 pub fn gen_for_callback_incrementing<'ctx, 'a, G, BodyFn>(
     generator: &mut G,
     ctx: &mut CodeGenContext<'ctx, 'a>,
+    label: Option<&str>,
     init_val: IntValue<'ctx>,
     max_val: (IntValue<'ctx>, bool),
     body: BodyFn,
@@ -562,13 +733,19 @@ pub fn gen_for_callback_incrementing<'ctx, 'a, G, BodyFn>(
 ) -> Result<(), String>
 where
     G: CodeGenerator + ?Sized,
-    BodyFn: FnOnce(&mut G, &mut CodeGenContext<'ctx, 'a>, IntValue<'ctx>) -> Result<(), String>,
+    BodyFn: FnOnce(
+        &mut G,
+        &mut CodeGenContext<'ctx, 'a>,
+        BreakContinueHooks<'ctx>,
+        IntValue<'ctx>,
+    ) -> Result<(), String>,
 {
     let init_val_t = init_val.get_type();
 
     gen_for_callback(
         generator,
         ctx,
+        label,
         |generator, ctx| {
             let i_addr = generator.gen_var_alloc(ctx, init_val_t.into(), None)?;
             ctx.builder.build_store(i_addr, init_val).unwrap();
@@ -584,10 +761,10 @@ where
 
             Ok(ctx.builder.build_int_compare(cmp_op, i, max_val, "").unwrap())
         },
-        |generator, ctx, i_addr| {
+        |generator, ctx, hooks, i_addr| {
             let i = ctx.builder.build_load(i_addr, "").map(BasicValueEnum::into_int_value).unwrap();
 
-            body(generator, ctx, i)
+            body(generator, ctx, hooks, i)
         },
         |_, ctx, i_addr| {
             let i = ctx.builder.build_load(i_addr, "").map(BasicValueEnum::into_int_value).unwrap();
@@ -620,9 +797,11 @@ where
 /// - `step_fn`: A lambda of IR statements that retrieves the `step` value of the  `range`-like
 ///   iterable. This value will be extended to the size of `start`.
 /// - `body_fn`: A lambda of IR statements within the loop body.
+#[allow(clippy::too_many_arguments)]
 pub fn gen_for_range_callback<'ctx, 'a, G, StartFn, StopFn, StepFn, BodyFn>(
     generator: &mut G,
     ctx: &mut CodeGenContext<'ctx, 'a>,
+    label: Option<&str>,
     is_unsigned: bool,
     start_fn: StartFn,
     (stop_fn, stop_inclusive): (StopFn, bool),
@@ -634,13 +813,19 @@ where
     StartFn: Fn(&mut G, &mut CodeGenContext<'ctx, 'a>) -> Result<IntValue<'ctx>, String>,
     StopFn: Fn(&mut G, &mut CodeGenContext<'ctx, 'a>) -> Result<IntValue<'ctx>, String>,
     StepFn: Fn(&mut G, &mut CodeGenContext<'ctx, 'a>) -> Result<IntValue<'ctx>, String>,
-    BodyFn: FnOnce(&mut G, &mut CodeGenContext<'ctx, 'a>, IntValue<'ctx>) -> Result<(), String>,
+    BodyFn: FnOnce(
+        &mut G,
+        &mut CodeGenContext<'ctx, 'a>,
+        BreakContinueHooks<'ctx>,
+        IntValue<'ctx>,
+    ) -> Result<(), String>,
 {
     let init_val_t = start_fn(generator, ctx).map(IntValue::get_type).unwrap();
 
     gen_for_callback(
         generator,
         ctx,
+        label,
         |generator, ctx| {
             let i_addr = generator.gen_var_alloc(ctx, init_val_t.into(), None)?;
 
@@ -698,10 +883,10 @@ where
 
             Ok(cond)
         },
-        |generator, ctx, (i_addr, _)| {
+        |generator, ctx, hooks, (i_addr, _)| {
             let i = ctx.builder.build_load(i_addr, "").map(BasicValueEnum::into_int_value).unwrap();
 
-            body_fn(generator, ctx, i)
+            body_fn(generator, ctx, hooks, i)
         },
         |generator, ctx, (i_addr, _)| {
             let i = ctx.builder.build_load(i_addr, "").map(BasicValueEnum::into_int_value).unwrap();
@@ -729,7 +914,7 @@ pub fn gen_while<G: CodeGenerator>(
     ctx: &mut CodeGenContext<'_, '_>,
     stmt: &Stmt<Option<Type>>,
 ) -> Result<(), String> {
-    let StmtKind::While { test, body, orelse, .. } = &stmt.node else { unreachable!() };
+    let StmtKind::While { test, body, orelse, .. } = &stmt.node else { codegen_unreachable!(ctx) };
 
     // var_assignment static values may be changed in another branch
     // if so, remove the static value as it may not be correct in this branch
@@ -759,7 +944,7 @@ pub fn gen_while<G: CodeGenerator>(
 
         return Ok(());
     };
-    let BasicValueEnum::IntValue(test) = test else { unreachable!() };
+    let BasicValueEnum::IntValue(test) = test else { codegen_unreachable!(ctx) };
 
     ctx.builder
         .build_conditional_branch(generator.bool_to_i1(ctx, test), body_bb, orelse_bb)
@@ -907,7 +1092,7 @@ pub fn gen_if<G: CodeGenerator>(
     ctx: &mut CodeGenContext<'_, '_>,
     stmt: &Stmt<Option<Type>>,
 ) -> Result<(), String> {
-    let StmtKind::If { test, body, orelse, .. } = &stmt.node else { unreachable!() };
+    let StmtKind::If { test, body, orelse, .. } = &stmt.node else { codegen_unreachable!(ctx) };
 
     // var_assignment static values may be changed in another branch
     // if so, remove the static value as it may not be correct in this branch
@@ -1030,11 +1215,11 @@ pub fn exn_constructor<'ctx>(
     let zelf_id = if let TypeEnum::TObj { obj_id, .. } = &*ctx.unifier.get_ty(zelf_ty) {
         obj_id.0
     } else {
-        unreachable!()
+        codegen_unreachable!(ctx)
     };
     let defs = ctx.top_level.definitions.read();
     let def = defs[zelf_id].read();
-    let TopLevelDef::Class { name: zelf_name, .. } = &*def else { unreachable!() };
+    let TopLevelDef::Class { name: zelf_name, .. } = &*def else { codegen_unreachable!(ctx) };
     let exception_name = format!("{}:{}", ctx.resolver.get_exception_id(zelf_id), zelf_name);
     unsafe {
         let id_ptr = ctx.builder.build_in_bounds_gep(zelf, &[zero, zero], "exn.id").unwrap();
@@ -1142,7 +1327,7 @@ pub fn gen_try<'ctx, 'a, G: CodeGenerator>(
     target: &Stmt<Option<Type>>,
 ) -> Result<(), String> {
     let StmtKind::Try { body, handlers, orelse, finalbody, .. } = &target.node else {
-        unreachable!()
+        codegen_unreachable!(ctx)
     };
 
     // if we need to generate anything related to exception, we must have personality defined
@@ -1219,7 +1404,7 @@ pub fn gen_try<'ctx, 'a, G: CodeGenerator>(
             if let TypeEnum::TObj { obj_id, .. } = &*ctx.unifier.get_ty(type_.custom.unwrap()) {
                 *obj_id
             } else {
-                unreachable!()
+                codegen_unreachable!(ctx)
             };
         let exception_name = format!("{}:{}", ctx.resolver.get_exception_id(obj_id.0), exn_name);
         let exn_id = ctx.resolver.get_string_id(&exception_name);
@@ -1491,6 +1676,23 @@ pub fn gen_return<G: CodeGenerator>(
     } else {
         None
     };
+
+    // Remap boolean return type into i1
+    let value = value.map(|ret_val| {
+        // The "return type" of a sret function is in the first parameter
+        let expected_ty = if ctx.need_sret {
+            func.get_type().get_param_types()[0]
+        } else {
+            func.get_type().get_return_type().unwrap()
+        };
+
+        if matches!(expected_ty, BasicTypeEnum::IntType(ty) if ty.get_bit_width() == 1) {
+            generator.bool_to_i1(ctx, ret_val.into_int_value()).into()
+        } else {
+            ret_val
+        }
+    });
+
     if let Some(return_target) = ctx.return_target {
         if let Some(value) = value {
             ctx.builder.build_store(ctx.return_buffer.unwrap(), value).unwrap();
@@ -1501,25 +1703,6 @@ pub fn gen_return<G: CodeGenerator>(
         ctx.builder.build_store(ctx.return_buffer.unwrap(), value.unwrap()).unwrap();
         ctx.builder.build_return(None).unwrap();
     } else {
-        // Remap boolean return type into i1
-        let value = value.map(|v| {
-            let expected_ty = func.get_type().get_return_type().unwrap();
-            let ret_val = v.as_basic_value_enum();
-
-            if expected_ty.is_int_type() && ret_val.is_int_value() {
-                let ret_type = expected_ty.into_int_type();
-                let ret_val = ret_val.into_int_value();
-
-                if ret_type.get_bit_width() == 1 && ret_val.get_type().get_bit_width() != 1 {
-                    generator.bool_to_i1(ctx, ret_val)
-                } else {
-                    ret_val
-                }
-                .into()
-            } else {
-                ret_val
-            }
-        });
         let value = value.as_ref().map(|v| v as &dyn BasicValue);
         ctx.builder.build_return(value).unwrap();
     }
@@ -1553,14 +1736,14 @@ pub fn gen_stmt<G: CodeGenerator>(
         }
         StmtKind::AnnAssign { target, value, .. } => {
             if let Some(value) = value {
-                let Some(value) = generator.gen_expr(ctx, value)? else { return Ok(()) };
-                generator.gen_assign(ctx, target, value)?;
+                let Some(value_enum) = generator.gen_expr(ctx, value)? else { return Ok(()) };
+                generator.gen_assign(ctx, target, value_enum, value.custom.unwrap())?;
             }
         }
         StmtKind::Assign { targets, value, .. } => {
-            let Some(value) = generator.gen_expr(ctx, value)? else { return Ok(()) };
+            let Some(value_enum) = generator.gen_expr(ctx, value)? else { return Ok(()) };
             for target in targets {
-                generator.gen_assign(ctx, target, value.clone())?;
+                generator.gen_assign(ctx, target, value_enum.clone(), value.custom.unwrap())?;
             }
         }
         StmtKind::Continue { .. } => {
@@ -1574,13 +1757,44 @@ pub fn gen_stmt<G: CodeGenerator>(
         StmtKind::For { .. } => generator.gen_for(ctx, stmt)?,
         StmtKind::With { .. } => generator.gen_with(ctx, stmt)?,
         StmtKind::AugAssign { target, op, value, .. } => {
-            let value = gen_binop_expr(generator, ctx, target, *op, value, stmt.location, true)?;
-            generator.gen_assign(ctx, target, value.unwrap())?;
+            let value_enum = gen_binop_expr(
+                generator,
+                ctx,
+                target,
+                Binop::aug_assign(*op),
+                value,
+                stmt.location,
+            )?
+            .unwrap();
+            generator.gen_assign(ctx, target, value_enum, value.custom.unwrap())?;
         }
         StmtKind::Try { .. } => gen_try(generator, ctx, stmt)?,
         StmtKind::Raise { exc, .. } => {
             if let Some(exc) = exc {
-                let exc = if let Some(v) = generator.gen_expr(ctx, exc)? {
+                let exn = if let ExprKind::Name { id, .. } = &exc.node {
+                    // Handle "raise Exception" short form
+                    let def_id = ctx.resolver.get_identifier_def(*id).map_err(|e| {
+                        format!("{} (at {})", e.iter().next().unwrap(), exc.location)
+                    })?;
+                    let def = ctx.top_level.definitions.read();
+                    let TopLevelDef::Class { constructor, .. } = *def[def_id.0].read() else {
+                        return Err(format!("Failed to resolve symbol {id} (at {})", exc.location));
+                    };
+
+                    let TypeEnum::TFunc(signature) =
+                        ctx.unifier.get_ty(constructor.unwrap()).as_ref().clone()
+                    else {
+                        return Err(format!("Failed to resolve symbol {id} (at {})", exc.location));
+                    };
+
+                    generator
+                        .gen_call(ctx, None, (&signature, def_id), Vec::default())?
+                        .map(Into::into)
+                } else {
+                    generator.gen_expr(ctx, exc)?
+                };
+
+                let exc = if let Some(v) = exn {
                     v.to_basic_value_enum(ctx, generator, exc.custom.unwrap())?
                 } else {
                     return Ok(());
@@ -1604,17 +1818,48 @@ pub fn gen_stmt<G: CodeGenerator>(
                         return Ok(());
                     }
                 }
-                None => ctx.gen_string(generator, ""),
+                None => ctx.gen_string(generator, "").into(),
             };
             ctx.make_assert_impl(
                 generator,
-                test.into_int_value(),
+                generator.bool_to_i1(ctx, test.into_int_value()),
                 "0:AssertionError",
                 err_msg,
                 [None, None, None],
                 stmt.location,
             );
         }
+        StmtKind::Global { names, .. } => {
+            let registered_globals = ctx
+                .top_level
+                .definitions
+                .read()
+                .iter()
+                .filter_map(|def| {
+                    if let TopLevelDef::Variable { simple_name, ty, .. } = &*def.read() {
+                        Some((*simple_name, *ty))
+                    } else {
+                        None
+                    }
+                })
+                .collect_vec();
+
+            for id in names {
+                let Some((_, ty)) = registered_globals.iter().find(|(name, _)| name == id) else {
+                    return Err(format!("{id} is not a global at {}", stmt.location));
+                };
+
+                let resolver = ctx.resolver.clone();
+                let ptr = resolver
+                    .get_symbol_value(*id, ctx, generator)
+                    .map(|val| val.to_basic_value_enum(ctx, generator, *ty))
+                    .transpose()?
+                    .map(BasicValueEnum::into_pointer_value)
+                    .unwrap();
+
+                ctx.var_assignment.insert(*id, (ptr, None, 0));
+            }
+        }
         _ => unimplemented!(),
     };
     Ok(())

nac3core/src/codegen/test.rs

@@ -1,34 +1,37 @@
-use crate::{
-    codegen::{
-        classes::{ListType, NDArrayType, ProxyType, RangeType},
-        concrete_type::ConcreteTypeStore,
-        CodeGenContext, CodeGenLLVMOptions, CodeGenTargetMachineOptions, CodeGenTask,
-        CodeGenerator, DefaultCodeGenerator, WithCall, WorkerRegistry,
-    },
-    symbol_resolver::{SymbolResolver, ValueEnum},
-    toplevel::{
-        composer::{ComposerConfig, TopLevelComposer},
-        DefinitionId, FunInstance, TopLevelContext, TopLevelDef,
-    },
-    typecheck::{
-        type_inferencer::{FunctionData, Inferencer, PrimitiveStore},
-        typedef::{FunSignature, FuncArg, Type, TypeEnum, Unifier, VarMap},
-    },
+use std::{
+    collections::{HashMap, HashSet},
+    sync::Arc,
 };
+
 use indexmap::IndexMap;
 use indoc::indoc;
 use inkwell::{
     targets::{InitializationConfig, Target},
     OptimizationLevel,
 };
-use nac3parser::ast::FileName;
 use nac3parser::{
-    ast::{fold::Fold, StrRef},
+    ast::{fold::Fold, FileName, StrRef},
     parser::parse_program,
 };
 use parking_lot::RwLock;
-use std::collections::{HashMap, HashSet};
-use std::sync::Arc;
+
+use super::{
+    concrete_type::ConcreteTypeStore,
+    types::{ListType, NDArrayType, ProxyType, RangeType},
+    CodeGenContext, CodeGenLLVMOptions, CodeGenTargetMachineOptions, CodeGenTask, CodeGenerator,
+    DefaultCodeGenerator, WithCall, WorkerRegistry,
+};
+use crate::{
+    symbol_resolver::{SymbolResolver, ValueEnum},
+    toplevel::{
+        composer::{ComposerConfig, TopLevelComposer},
+        DefinitionId, FunInstance, TopLevelContext, TopLevelDef,
+    },
+    typecheck::{
+        type_inferencer::{FunctionData, IdentifierInfo, Inferencer, PrimitiveStore},
+        typedef::{FunSignature, FuncArg, Type, TypeEnum, Unifier, VarMap},
+    },
+};
 
 struct Resolver {
     id_to_type: HashMap<StrRef, Type>,
@@ -64,6 +67,7 @@ impl SymbolResolver for Resolver {
         &self,
         _: StrRef,
         _: &mut CodeGenContext<'ctx, '_>,
+        _: &mut dyn CodeGenerator,
     ) -> Option<ValueEnum<'ctx>> {
         unimplemented!()
     }
@@ -94,7 +98,7 @@ fn test_primitives() {
         "};
     let statements = parse_program(source, FileName::default()).unwrap();
 
-    let composer = TopLevelComposer::new(Vec::new(), ComposerConfig::default(), 32).0;
+    let composer = TopLevelComposer::new(Vec::new(), Vec::new(), ComposerConfig::default(), 32).0;
     let mut unifier = composer.unifier.clone();
     let primitives = composer.primitives_ty;
     let top_level = Arc::new(composer.make_top_level_context());
@@ -109,8 +113,18 @@ fn test_primitives() {
     let threads = vec![DefaultCodeGenerator::new("test".into(), 32).into()];
     let signature = FunSignature {
         args: vec![
-            FuncArg { name: "a".into(), ty: primitives.int32, default_value: None },
-            FuncArg { name: "b".into(), ty: primitives.int32, default_value: None },
+            FuncArg {
+                name: "a".into(),
+                ty: primitives.int32,
+                default_value: None,
+                is_vararg: false,
+            },
+            FuncArg {
+                name: "b".into(),
+                ty: primitives.int32,
+                default_value: None,
+                is_vararg: false,
+            },
         ],
         ret: primitives.int32,
         vars: VarMap::new(),
@@ -128,7 +142,8 @@ fn test_primitives() {
     };
     let mut virtual_checks = Vec::new();
     let mut calls = HashMap::new();
-    let mut identifiers: HashSet<_> = ["a".into(), "b".into()].into();
+    let mut identifiers: HashMap<_, _> =
+        ["a".into(), "b".into()].map(|id| (id, IdentifierInfo::default())).into();
     let mut inferencer = Inferencer {
         top_level: &top_level,
         function_data: &mut function_data,
@@ -189,6 +204,8 @@ fn test_primitives() {
         let expected = indoc! {"
             ; ModuleID = 'test'
             source_filename = \"test\"
+            target datalayout = \"e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128\"
+            target triple = \"x86_64-unknown-linux-gnu\"
 
             ; Function Attrs: mustprogress nofree norecurse nosync nounwind readnone willreturn
             define i32 @testing(i32 %0, i32 %1) local_unnamed_addr #0 !dbg !4 {
@@ -246,14 +263,19 @@ fn test_simple_call() {
         "};
     let statements_2 = parse_program(source_2, FileName::default()).unwrap();
 
-    let composer = TopLevelComposer::new(Vec::new(), ComposerConfig::default(), 32).0;
+    let composer = TopLevelComposer::new(Vec::new(), Vec::new(), ComposerConfig::default(), 32).0;
     let mut unifier = composer.unifier.clone();
     let primitives = composer.primitives_ty;
     let top_level = Arc::new(composer.make_top_level_context());
     unifier.top_level = Some(top_level.clone());
 
     let signature = FunSignature {
-        args: vec![FuncArg { name: "a".into(), ty: primitives.int32, default_value: None }],
+        args: vec![FuncArg {
+            name: "a".into(),
+            ty: primitives.int32,
+            default_value: None,
+            is_vararg: false,
+        }],
         ret: primitives.int32,
         vars: VarMap::new(),
     };
@@ -300,7 +322,8 @@ fn test_simple_call() {
     };
     let mut virtual_checks = Vec::new();
     let mut calls = HashMap::new();
-    let mut identifiers: HashSet<_> = ["a".into(), "foo".into()].into();
+    let mut identifiers: HashMap<_, _> =
+        ["a".into(), "foo".into()].map(|id| (id, IdentifierInfo::default())).into();
     let mut inferencer = Inferencer {
         top_level: &top_level,
         function_data: &mut function_data,
@@ -368,6 +391,8 @@ fn test_simple_call() {
         let expected = indoc! {"
             ; ModuleID = 'test'
             source_filename = \"test\"
+            target datalayout = \"e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128\"
+            target triple = \"x86_64-unknown-linux-gnu\"
 
             ; Function Attrs: mustprogress nofree norecurse nosync nounwind readnone willreturn
             define i32 @testing(i32 %0) local_unnamed_addr #0 !dbg !5 {
@@ -427,7 +452,7 @@ fn test_classes_list_type_new() {
     let llvm_usize = generator.get_size_type(&ctx);
 
     let llvm_list = ListType::new(&generator, &ctx, llvm_i32.into());
-    assert!(ListType::is_type(llvm_list.as_base_type(), llvm_usize).is_ok());
+    assert!(ListType::is_representable(llvm_list.as_base_type(), llvm_usize).is_ok());
 }
 
 #[test]
@@ -435,7 +460,7 @@ fn test_classes_range_type_new() {
     let ctx = inkwell::context::Context::create();
 
     let llvm_range = RangeType::new(&ctx);
-    assert!(RangeType::is_type(llvm_range.as_base_type()).is_ok());
+    assert!(RangeType::is_representable(llvm_range.as_base_type()).is_ok());
 }
 
 #[test]
@@ -447,5 +472,5 @@ fn test_classes_ndarray_type_new() {
     let llvm_usize = generator.get_size_type(&ctx);
 
     let llvm_ndarray = NDArrayType::new(&generator, &ctx, llvm_i32.into());
-    assert!(NDArrayType::is_type(llvm_ndarray.as_base_type(), llvm_usize).is_ok());
+    assert!(NDArrayType::is_representable(llvm_ndarray.as_base_type(), llvm_usize).is_ok());
 }

nac3core/src/codegen/types/list.rs

@@ -0,0 +1,192 @@
+use inkwell::{
+    context::Context,
+    types::{AnyTypeEnum, BasicType, BasicTypeEnum, IntType, PointerType},
+    values::IntValue,
+    AddressSpace,
+};
+
+use super::ProxyType;
+use crate::codegen::{
+    values::{ArraySliceValue, ListValue, ProxyValue},
+    CodeGenContext, CodeGenerator,
+};
+
+/// Proxy type for a `list` type in LLVM.
+#[derive(Debug, PartialEq, Eq, Clone, Copy)]
+pub struct ListType<'ctx> {
+    ty: PointerType<'ctx>,
+    llvm_usize: IntType<'ctx>,
+}
+
+impl<'ctx> ListType<'ctx> {
+    /// Checks whether `llvm_ty` represents a `list` type, returning [Err] if it does not.
+    pub fn is_representable(
+        llvm_ty: PointerType<'ctx>,
+        llvm_usize: IntType<'ctx>,
+    ) -> Result<(), String> {
+        let llvm_list_ty = llvm_ty.get_element_type();
+        let AnyTypeEnum::StructType(llvm_list_ty) = llvm_list_ty else {
+            return Err(format!("Expected struct type for `list` type, got {llvm_list_ty}"));
+        };
+        if llvm_list_ty.count_fields() != 2 {
+            return Err(format!(
+                "Expected 2 fields in `list`, got {}",
+                llvm_list_ty.count_fields()
+            ));
+        }
+
+        let list_size_ty = llvm_list_ty.get_field_type_at_index(0).unwrap();
+        let Ok(_) = PointerType::try_from(list_size_ty) else {
+            return Err(format!("Expected pointer type for `list.0`, got {list_size_ty}"));
+        };
+
+        let list_data_ty = llvm_list_ty.get_field_type_at_index(1).unwrap();
+        let Ok(list_data_ty) = IntType::try_from(list_data_ty) else {
+            return Err(format!("Expected int type for `list.1`, got {list_data_ty}"));
+        };
+        if list_data_ty.get_bit_width() != llvm_usize.get_bit_width() {
+            return Err(format!(
+                "Expected {}-bit int type for `list.1`, got {}-bit int",
+                llvm_usize.get_bit_width(),
+                list_data_ty.get_bit_width()
+            ));
+        }
+
+        Ok(())
+    }
+
+    /// Creates an LLVM type corresponding to the expected structure of a `List`.
+    #[must_use]
+    fn llvm_type(
+        ctx: &'ctx Context,
+        element_type: BasicTypeEnum<'ctx>,
+        llvm_usize: IntType<'ctx>,
+    ) -> PointerType<'ctx> {
+        // struct List { data: T*, size: size_t }
+        let field_tys = [element_type.ptr_type(AddressSpace::default()).into(), llvm_usize.into()];
+
+        ctx.struct_type(&field_tys, false).ptr_type(AddressSpace::default())
+    }
+
+    /// Creates an instance of [`ListType`].
+    #[must_use]
+    pub fn new<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        element_type: BasicTypeEnum<'ctx>,
+    ) -> Self {
+        let llvm_usize = generator.get_size_type(ctx);
+        let llvm_list = Self::llvm_type(ctx, element_type, llvm_usize);
+
+        ListType::from_type(llvm_list, llvm_usize)
+    }
+
+    /// Creates an [`ListType`] from a [`PointerType`].
+    #[must_use]
+    pub fn from_type(ptr_ty: PointerType<'ctx>, llvm_usize: IntType<'ctx>) -> Self {
+        debug_assert!(Self::is_representable(ptr_ty, llvm_usize).is_ok());
+
+        ListType { ty: ptr_ty, llvm_usize }
+    }
+
+    /// Returns the type of the `size` field of this `list` type.
+    #[must_use]
+    pub fn size_type(&self) -> IntType<'ctx> {
+        self.as_base_type()
+            .get_element_type()
+            .into_struct_type()
+            .get_field_type_at_index(1)
+            .map(BasicTypeEnum::into_int_type)
+            .unwrap()
+    }
+
+    /// Returns the element type of this `list` type.
+    #[must_use]
+    pub fn element_type(&self) -> AnyTypeEnum<'ctx> {
+        self.as_base_type()
+            .get_element_type()
+            .into_struct_type()
+            .get_field_type_at_index(0)
+            .map(BasicTypeEnum::into_pointer_type)
+            .map(PointerType::get_element_type)
+            .unwrap()
+    }
+}
+
+impl<'ctx> ProxyType<'ctx> for ListType<'ctx> {
+    type Base = PointerType<'ctx>;
+    type Value = ListValue<'ctx>;
+
+    fn is_type<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        llvm_ty: impl BasicType<'ctx>,
+    ) -> Result<(), String> {
+        if let BasicTypeEnum::PointerType(ty) = llvm_ty.as_basic_type_enum() {
+            <Self as ProxyType<'ctx>>::is_representable(generator, ctx, ty)
+        } else {
+            Err(format!("Expected pointer type, got {llvm_ty:?}"))
+        }
+    }
+
+    fn is_representable<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        llvm_ty: Self::Base,
+    ) -> Result<(), String> {
+        Self::is_representable(llvm_ty, generator.get_size_type(ctx))
+    }
+
+    fn new_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        name: Option<&'ctx str>,
+    ) -> Self::Value {
+        self.map_value(
+            generator
+                .gen_var_alloc(
+                    ctx,
+                    self.as_base_type().get_element_type().into_struct_type().into(),
+                    name,
+                )
+                .unwrap(),
+            name,
+        )
+    }
+
+    fn new_array_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        size: IntValue<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> ArraySliceValue<'ctx> {
+        generator
+            .gen_array_var_alloc(
+                ctx,
+                self.as_base_type().get_element_type().into_struct_type().into(),
+                size,
+                name,
+            )
+            .unwrap()
+    }
+
+    fn map_value(
+        &self,
+        value: <Self::Value as ProxyValue<'ctx>>::Base,
+        name: Option<&'ctx str>,
+    ) -> Self::Value {
+        Self::Value::from_pointer_value(value, self.llvm_usize, name)
+    }
+
+    fn as_base_type(&self) -> Self::Base {
+        self.ty
+    }
+}
+
+impl<'ctx> From<ListType<'ctx>> for PointerType<'ctx> {
+    fn from(value: ListType<'ctx>) -> Self {
+        value.as_base_type()
+    }
+}

nac3core/src/codegen/types/mod.rs

@@ -0,0 +1,64 @@
+use inkwell::{context::Context, types::BasicType, values::IntValue};
+
+use super::{
+    values::{ArraySliceValue, ProxyValue},
+    {CodeGenContext, CodeGenerator},
+};
+pub use list::*;
+pub use ndarray::*;
+pub use range::*;
+
+mod list;
+mod ndarray;
+mod range;
+pub mod structure;
+
+/// A LLVM type that is used to represent a corresponding type in NAC3.
+pub trait ProxyType<'ctx>: Into<Self::Base> {
+    /// The LLVM type of which values of this type possess. This is usually a
+    /// [LLVM pointer type][PointerType] for any non-primitive types.
+    type Base: BasicType<'ctx>;
+
+    /// The type of values represented by this type.
+    type Value: ProxyValue<'ctx, Type = Self>;
+
+    fn is_type<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        llvm_ty: impl BasicType<'ctx>,
+    ) -> Result<(), String>;
+
+    /// Checks whether `llvm_ty` can be represented by this [`ProxyType`].
+    fn is_representable<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        llvm_ty: Self::Base,
+    ) -> Result<(), String>;
+
+    /// Creates a new value of this type.
+    fn new_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        name: Option<&'ctx str>,
+    ) -> Self::Value;
+
+    /// Creates a new array value of this type.
+    fn new_array_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        size: IntValue<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> ArraySliceValue<'ctx>;
+
+    /// Converts an existing value into a [`ProxyValue`] of this type.
+    fn map_value(
+        &self,
+        value: <Self::Value as ProxyValue<'ctx>>::Base,
+        name: Option<&'ctx str>,
+    ) -> Self::Value;
+
+    /// Returns the [base type][Self::Base] of this proxy.
+    fn as_base_type(&self) -> Self::Base;
+}

nac3core/src/codegen/types/ndarray.rs

@@ -0,0 +1,258 @@
+use inkwell::{
+    context::Context,
+    types::{AnyTypeEnum, BasicType, BasicTypeEnum, IntType, PointerType},
+    values::{IntValue, PointerValue},
+    AddressSpace,
+};
+use itertools::Itertools;
+
+use nac3core_derive::StructFields;
+
+use super::{
+    structure::{StructField, StructFields},
+    ProxyType,
+};
+use crate::codegen::{
+    values::{ArraySliceValue, NDArrayValue, ProxyValue},
+    {CodeGenContext, CodeGenerator},
+};
+
+/// Proxy type for a `ndarray` type in LLVM.
+#[derive(Debug, PartialEq, Eq, Clone, Copy)]
+pub struct NDArrayType<'ctx> {
+    ty: PointerType<'ctx>,
+    dtype: BasicTypeEnum<'ctx>,
+    llvm_usize: IntType<'ctx>,
+}
+
+#[derive(PartialEq, Eq, Clone, Copy, StructFields)]
+pub struct NDArrayStructFields<'ctx> {
+    #[value_type(usize)]
+    pub ndims: StructField<'ctx, IntValue<'ctx>>,
+    #[value_type(usize.ptr_type(AddressSpace::default()))]
+    pub shape: StructField<'ctx, PointerValue<'ctx>>,
+    #[value_type(i8_type().ptr_type(AddressSpace::default()))]
+    pub data: StructField<'ctx, PointerValue<'ctx>>,
+}
+
+impl<'ctx> NDArrayType<'ctx> {
+    /// Checks whether `llvm_ty` represents a `ndarray` type, returning [Err] if it does not.
+    pub fn is_representable(
+        llvm_ty: PointerType<'ctx>,
+        llvm_usize: IntType<'ctx>,
+    ) -> Result<(), String> {
+        let llvm_ndarray_ty = llvm_ty.get_element_type();
+        let AnyTypeEnum::StructType(llvm_ndarray_ty) = llvm_ndarray_ty else {
+            return Err(format!("Expected struct type for `NDArray` type, got {llvm_ndarray_ty}"));
+        };
+        if llvm_ndarray_ty.count_fields() != 3 {
+            return Err(format!(
+                "Expected 3 fields in `NDArray`, got {}",
+                llvm_ndarray_ty.count_fields()
+            ));
+        }
+
+        let ndarray_ndims_ty = llvm_ndarray_ty.get_field_type_at_index(0).unwrap();
+        let Ok(ndarray_ndims_ty) = IntType::try_from(ndarray_ndims_ty) else {
+            return Err(format!("Expected int type for `ndarray.0`, got {ndarray_ndims_ty}"));
+        };
+        if ndarray_ndims_ty.get_bit_width() != llvm_usize.get_bit_width() {
+            return Err(format!(
+                "Expected {}-bit int type for `ndarray.0`, got {}-bit int",
+                llvm_usize.get_bit_width(),
+                ndarray_ndims_ty.get_bit_width()
+            ));
+        }
+
+        let ndarray_dims_ty = llvm_ndarray_ty.get_field_type_at_index(1).unwrap();
+        let Ok(ndarray_pdims) = PointerType::try_from(ndarray_dims_ty) else {
+            return Err(format!("Expected pointer type for `ndarray.1`, got {ndarray_dims_ty}"));
+        };
+        let ndarray_dims = ndarray_pdims.get_element_type();
+        let Ok(ndarray_dims) = IntType::try_from(ndarray_dims) else {
+            return Err(format!(
+                "Expected pointer-to-int type for `ndarray.1`, got pointer-to-{ndarray_dims}"
+            ));
+        };
+        if ndarray_dims.get_bit_width() != llvm_usize.get_bit_width() {
+            return Err(format!(
+                "Expected pointer-to-{}-bit int type for `ndarray.1`, got pointer-to-{}-bit int",
+                llvm_usize.get_bit_width(),
+                ndarray_dims.get_bit_width()
+            ));
+        }
+
+        let ndarray_data_ty = llvm_ndarray_ty.get_field_type_at_index(2).unwrap();
+        let Ok(ndarray_pdata) = PointerType::try_from(ndarray_data_ty) else {
+            return Err(format!("Expected pointer type for `ndarray.2`, got {ndarray_data_ty}"));
+        };
+        let ndarray_data = ndarray_pdata.get_element_type();
+        let Ok(ndarray_data) = IntType::try_from(ndarray_data) else {
+            return Err(format!(
+                "Expected pointer-to-int type for `ndarray.2`, got pointer-to-{ndarray_data}"
+            ));
+        };
+        if ndarray_data.get_bit_width() != 8 {
+            return Err(format!(
+                "Expected pointer-to-8-bit int type for `ndarray.1`, got pointer-to-{}-bit int",
+                ndarray_data.get_bit_width()
+            ));
+        }
+
+        Ok(())
+    }
+
+    // TODO: Move this into e.g. StructProxyType
+    #[must_use]
+    fn fields(ctx: &'ctx Context, llvm_usize: IntType<'ctx>) -> NDArrayStructFields<'ctx> {
+        NDArrayStructFields::new(ctx, llvm_usize)
+    }
+
+    // TODO: Move this into e.g. StructProxyType
+    #[must_use]
+    pub fn get_fields(
+        &self,
+        ctx: &'ctx Context,
+        llvm_usize: IntType<'ctx>,
+    ) -> NDArrayStructFields<'ctx> {
+        Self::fields(ctx, llvm_usize)
+    }
+
+    /// Creates an LLVM type corresponding to the expected structure of an `NDArray`.
+    #[must_use]
+    fn llvm_type(ctx: &'ctx Context, llvm_usize: IntType<'ctx>) -> PointerType<'ctx> {
+        // struct NDArray { num_dims: size_t, dims: size_t*, data: i8* }
+        //
+        // * data    : Pointer to an array containing the array data
+        // * itemsize: The size of each NDArray elements in bytes
+        // * ndims   : Number of dimensions in the array
+        // * shape   : Pointer to an array containing the shape of the NDArray
+        // * strides : Pointer to an array indicating the number of bytes between each element at a dimension
+        let field_tys =
+            Self::fields(ctx, llvm_usize).into_iter().map(|field| field.1).collect_vec();
+
+        ctx.struct_type(&field_tys, false).ptr_type(AddressSpace::default())
+    }
+
+    /// Creates an instance of [`NDArrayType`].
+    #[must_use]
+    pub fn new<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        dtype: BasicTypeEnum<'ctx>,
+    ) -> Self {
+        let llvm_usize = generator.get_size_type(ctx);
+        let llvm_ndarray = Self::llvm_type(ctx, llvm_usize);
+
+        NDArrayType { ty: llvm_ndarray, dtype, llvm_usize }
+    }
+
+    /// Creates an [`NDArrayType`] from a [`PointerType`] representing an `NDArray`.
+    #[must_use]
+    pub fn from_type(
+        ptr_ty: PointerType<'ctx>,
+        dtype: BasicTypeEnum<'ctx>,
+        llvm_usize: IntType<'ctx>,
+    ) -> Self {
+        debug_assert!(Self::is_representable(ptr_ty, llvm_usize).is_ok());
+
+        NDArrayType { ty: ptr_ty, dtype, llvm_usize }
+    }
+
+    /// Returns the type of the `size` field of this `ndarray` type.
+    #[must_use]
+    pub fn size_type(&self) -> IntType<'ctx> {
+        self.as_base_type()
+            .get_element_type()
+            .into_struct_type()
+            .get_field_type_at_index(0)
+            .map(BasicTypeEnum::into_int_type)
+            .unwrap()
+    }
+
+    /// Returns the element type of this `ndarray` type.
+    #[must_use]
+    pub fn element_type(&self) -> BasicTypeEnum<'ctx> {
+        self.dtype
+    }
+}
+
+impl<'ctx> ProxyType<'ctx> for NDArrayType<'ctx> {
+    type Base = PointerType<'ctx>;
+    type Value = NDArrayValue<'ctx>;
+
+    fn is_type<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        llvm_ty: impl BasicType<'ctx>,
+    ) -> Result<(), String> {
+        if let BasicTypeEnum::PointerType(ty) = llvm_ty.as_basic_type_enum() {
+            <Self as ProxyType<'ctx>>::is_representable(generator, ctx, ty)
+        } else {
+            Err(format!("Expected pointer type, got {llvm_ty:?}"))
+        }
+    }
+
+    fn is_representable<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        llvm_ty: Self::Base,
+    ) -> Result<(), String> {
+        Self::is_representable(llvm_ty, generator.get_size_type(ctx))
+    }
+
+    fn new_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        name: Option<&'ctx str>,
+    ) -> Self::Value {
+        self.map_value(
+            generator
+                .gen_var_alloc(
+                    ctx,
+                    self.as_base_type().get_element_type().into_struct_type().into(),
+                    name,
+                )
+                .unwrap(),
+            name,
+        )
+    }
+
+    fn new_array_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        size: IntValue<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> ArraySliceValue<'ctx> {
+        generator
+            .gen_array_var_alloc(
+                ctx,
+                self.as_base_type().get_element_type().into_struct_type().into(),
+                size,
+                name,
+            )
+            .unwrap()
+    }
+
+    fn map_value(
+        &self,
+        value: <Self::Value as ProxyValue<'ctx>>::Base,
+        name: Option<&'ctx str>,
+    ) -> Self::Value {
+        debug_assert_eq!(value.get_type(), self.as_base_type());
+
+        NDArrayValue::from_pointer_value(value, self.dtype, self.llvm_usize, name)
+    }
+
+    fn as_base_type(&self) -> Self::Base {
+        self.ty
+    }
+}
+
+impl<'ctx> From<NDArrayType<'ctx>> for PointerType<'ctx> {
+    fn from(value: NDArrayType<'ctx>) -> Self {
+        value.as_base_type()
+    }
+}

nac3core/src/codegen/types/range.rs

@@ -0,0 +1,159 @@
+use inkwell::{
+    context::Context,
+    types::{AnyTypeEnum, BasicType, BasicTypeEnum, IntType, PointerType},
+    values::IntValue,
+    AddressSpace,
+};
+
+use super::ProxyType;
+use crate::codegen::{
+    values::{ArraySliceValue, ProxyValue, RangeValue},
+    {CodeGenContext, CodeGenerator},
+};
+
+/// Proxy type for a `range` type in LLVM.
+#[derive(Debug, PartialEq, Eq, Clone, Copy)]
+pub struct RangeType<'ctx> {
+    ty: PointerType<'ctx>,
+}
+
+impl<'ctx> RangeType<'ctx> {
+    /// Checks whether `llvm_ty` represents a `range` type, returning [Err] if it does not.
+    pub fn is_representable(llvm_ty: PointerType<'ctx>) -> Result<(), String> {
+        let llvm_range_ty = llvm_ty.get_element_type();
+        let AnyTypeEnum::ArrayType(llvm_range_ty) = llvm_range_ty else {
+            return Err(format!("Expected array type for `range` type, got {llvm_range_ty}"));
+        };
+        if llvm_range_ty.len() != 3 {
+            return Err(format!(
+                "Expected 3 elements for `range` type, got {}",
+                llvm_range_ty.len()
+            ));
+        }
+
+        let llvm_range_elem_ty = llvm_range_ty.get_element_type();
+        let Ok(llvm_range_elem_ty) = IntType::try_from(llvm_range_elem_ty) else {
+            return Err(format!(
+                "Expected int type for `range` element type, got {llvm_range_elem_ty}"
+            ));
+        };
+        if llvm_range_elem_ty.get_bit_width() != 32 {
+            return Err(format!(
+                "Expected 32-bit int type for `range` element type, got {}",
+                llvm_range_elem_ty.get_bit_width()
+            ));
+        }
+
+        Ok(())
+    }
+
+    /// Creates an LLVM type corresponding to the expected structure of a `Range`.
+    #[must_use]
+    fn llvm_type(ctx: &'ctx Context) -> PointerType<'ctx> {
+        // typedef int32_t Range[3];
+        let llvm_i32 = ctx.i32_type();
+        llvm_i32.array_type(3).ptr_type(AddressSpace::default())
+    }
+
+    /// Creates an instance of [`RangeType`].
+    #[must_use]
+    pub fn new(ctx: &'ctx Context) -> Self {
+        let llvm_range = Self::llvm_type(ctx);
+
+        RangeType::from_type(llvm_range)
+    }
+
+    /// Creates an [`RangeType`] from a [`PointerType`].
+    #[must_use]
+    pub fn from_type(ptr_ty: PointerType<'ctx>) -> Self {
+        debug_assert!(Self::is_representable(ptr_ty).is_ok());
+
+        RangeType { ty: ptr_ty }
+    }
+
+    /// Returns the type of all fields of this `range` type.
+    #[must_use]
+    pub fn value_type(&self) -> IntType<'ctx> {
+        self.as_base_type().get_element_type().into_array_type().get_element_type().into_int_type()
+    }
+}
+
+impl<'ctx> ProxyType<'ctx> for RangeType<'ctx> {
+    type Base = PointerType<'ctx>;
+    type Value = RangeValue<'ctx>;
+
+    fn is_type<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        llvm_ty: impl BasicType<'ctx>,
+    ) -> Result<(), String> {
+        if let BasicTypeEnum::PointerType(ty) = llvm_ty.as_basic_type_enum() {
+            <Self as ProxyType<'ctx>>::is_representable(generator, ctx, ty)
+        } else {
+            Err(format!("Expected pointer type, got {llvm_ty:?}"))
+        }
+    }
+
+    fn is_representable<G: CodeGenerator + ?Sized>(
+        _: &G,
+        _: &'ctx Context,
+        llvm_ty: Self::Base,
+    ) -> Result<(), String> {
+        Self::is_representable(llvm_ty)
+    }
+
+    fn new_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        name: Option<&'ctx str>,
+    ) -> Self::Value {
+        self.map_value(
+            generator
+                .gen_var_alloc(
+                    ctx,
+                    self.as_base_type().get_element_type().into_struct_type().into(),
+                    name,
+                )
+                .unwrap(),
+            name,
+        )
+    }
+
+    fn new_array_value<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        size: IntValue<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> ArraySliceValue<'ctx> {
+        generator
+            .gen_array_var_alloc(
+                ctx,
+                self.as_base_type().get_element_type().into_struct_type().into(),
+                size,
+                name,
+            )
+            .unwrap()
+    }
+
+    fn map_value(
+        &self,
+        value: <Self::Value as ProxyValue<'ctx>>::Base,
+        name: Option<&'ctx str>,
+    ) -> Self::Value {
+        debug_assert_eq!(value.get_type(), self.as_base_type());
+
+        RangeValue::from_pointer_value(value, name)
+    }
+
+    fn as_base_type(&self) -> Self::Base {
+        self.ty
+    }
+}
+
+impl<'ctx> From<RangeType<'ctx>> for PointerType<'ctx> {
+    fn from(value: RangeType<'ctx>) -> Self {
+        value.as_base_type()
+    }
+}

nac3core/src/codegen/types/structure.rs

@@ -0,0 +1,203 @@
+use std::marker::PhantomData;
+
+use inkwell::{
+    context::AsContextRef,
+    types::{BasicTypeEnum, IntType},
+    values::{BasicValue, BasicValueEnum, IntValue, PointerValue, StructValue},
+};
+
+use crate::codegen::CodeGenContext;
+
+/// Trait indicating that the structure is a field-wise representation of an LLVM structure.
+///
+/// # Usage
+///
+/// For example, for a simple C-slice LLVM structure:
+///
+/// ```ignore
+/// struct CSliceFields<'ctx> {
+///     ptr: StructField<'ctx, PointerValue<'ctx>>,
+///     len: StructField<'ctx, IntValue<'ctx>>
+/// }
+/// ```
+pub trait StructFields<'ctx>: Eq + Copy {
+    /// Creates an instance of [`StructFields`] using the given `ctx` and `size_t` types.
+    fn new(ctx: impl AsContextRef<'ctx>, llvm_usize: IntType<'ctx>) -> Self;
+
+    /// Returns a [`Vec`] that contains the fields of the structure in the order as they appear in
+    /// the type definition.
+    #[must_use]
+    fn to_vec(&self) -> Vec<(&'static str, BasicTypeEnum<'ctx>)>;
+
+    /// Returns a [`Iterator`] that contains the fields of the structure in the order as they appear
+    /// in the type definition.
+    #[must_use]
+    fn iter(&self) -> impl Iterator<Item = (&'static str, BasicTypeEnum<'ctx>)> {
+        self.to_vec().into_iter()
+    }
+
+    /// Returns a [`Vec`] that contains the fields of the structure in the order as they appear in
+    /// the type definition.
+    #[must_use]
+    fn into_vec(self) -> Vec<(&'static str, BasicTypeEnum<'ctx>)>
+    where
+        Self: Sized,
+    {
+        self.to_vec()
+    }
+
+    /// Returns a [`Iterator`] that contains the fields of the structure in the order as they appear
+    /// in the type definition.
+    #[must_use]
+    fn into_iter(self) -> impl Iterator<Item = (&'static str, BasicTypeEnum<'ctx>)>
+    where
+        Self: Sized,
+    {
+        self.into_vec().into_iter()
+    }
+}
+
+/// A single field of an LLVM structure.
+#[derive(Debug, PartialEq, Eq, Clone, Copy)]
+pub struct StructField<'ctx, Value>
+where
+    Value: BasicValue<'ctx> + TryFrom<BasicValueEnum<'ctx>, Error = ()>,
+{
+    /// The index of this field within the structure.
+    index: u32,
+
+    /// The name of this field.
+    name: &'static str,
+
+    /// The type of this field.
+    ty: BasicTypeEnum<'ctx>,
+
+    /// Instance of [`PhantomData`] containing [`Value`], used to implement automatic downcasts.
+    _value_ty: PhantomData<Value>,
+}
+
+impl<'ctx, Value> StructField<'ctx, Value>
+where
+    Value: BasicValue<'ctx> + TryFrom<BasicValueEnum<'ctx>, Error = ()>,
+{
+    /// Creates an instance of [`StructField`].
+    ///
+    /// * `idx_counter` - The instance of [`FieldIndexCounter`] used to track the current field
+    ///   index.
+    /// * `name` - Name of the field.
+    /// * `ty` - The type of this field.
+    pub fn create(
+        idx_counter: &mut FieldIndexCounter,
+        name: &'static str,
+        ty: impl Into<BasicTypeEnum<'ctx>>,
+    ) -> Self {
+        StructField { index: idx_counter.increment(), name, ty: ty.into(), _value_ty: PhantomData }
+    }
+
+    /// Creates an instance of [`StructField`] with a given index.
+    ///
+    /// * `index` - The index of this field within its enclosing structure.
+    /// * `name` - Name of the field.
+    /// * `ty` - The type of this field.
+    pub fn create_at(index: u32, name: &'static str, ty: impl Into<BasicTypeEnum<'ctx>>) -> Self {
+        StructField { index, name, ty: ty.into(), _value_ty: PhantomData }
+    }
+
+    /// Creates a pointer to this field in an arbitrary structure by performing a `getelementptr i32
+    /// {idx...}, i32 {self.index}`.
+    pub fn ptr_by_array_gep(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        pobj: PointerValue<'ctx>,
+        idx: &[IntValue<'ctx>],
+    ) -> PointerValue<'ctx> {
+        unsafe {
+            ctx.builder.build_in_bounds_gep(
+                pobj,
+                &[idx, &[ctx.ctx.i32_type().const_int(u64::from(self.index), false)]].concat(),
+                "",
+            )
+        }
+        .unwrap()
+    }
+
+    /// Creates a pointer to this field in an arbitrary structure by performing the equivalent of
+    /// `getelementptr i32 0, i32 {self.index}`.
+    pub fn ptr_by_gep(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        pobj: PointerValue<'ctx>,
+        obj_name: Option<&'ctx str>,
+    ) -> PointerValue<'ctx> {
+        ctx.builder
+            .build_struct_gep(
+                pobj,
+                self.index,
+                &obj_name.map(|name| format!("{name}.{}.addr", self.name)).unwrap_or_default(),
+            )
+            .unwrap()
+    }
+
+    /// Gets the value of this field for a given `obj`.
+    #[must_use]
+    pub fn get_from_value(&self, obj: StructValue<'ctx>) -> Value {
+        obj.get_field_at_index(self.index).and_then(|value| Value::try_from(value).ok()).unwrap()
+    }
+
+    /// Sets the value of this field for a given `obj`.
+    pub fn set_from_value(&self, obj: StructValue<'ctx>, value: Value) {
+        obj.set_field_at_index(self.index, value);
+    }
+
+    /// Gets the value of this field for a pointer-to-structure.
+    pub fn get(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        pobj: PointerValue<'ctx>,
+        obj_name: Option<&'ctx str>,
+    ) -> Value {
+        ctx.builder
+            .build_load(
+                self.ptr_by_gep(ctx, pobj, obj_name),
+                &obj_name.map(|name| format!("{name}.{}", self.name)).unwrap_or_default(),
+            )
+            .map_err(|_| ())
+            .and_then(|value| Value::try_from(value))
+            .unwrap()
+    }
+
+    /// Sets the value of this field for a pointer-to-structure.
+    pub fn set(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        pobj: PointerValue<'ctx>,
+        value: Value,
+        obj_name: Option<&'ctx str>,
+    ) {
+        ctx.builder.build_store(self.ptr_by_gep(ctx, pobj, obj_name), value).unwrap();
+    }
+}
+
+impl<'ctx, Value> From<StructField<'ctx, Value>> for (&'static str, BasicTypeEnum<'ctx>)
+where
+    Value: BasicValue<'ctx> + TryFrom<BasicValueEnum<'ctx>, Error = ()>,
+{
+    fn from(value: StructField<'ctx, Value>) -> Self {
+        (value.name, value.ty)
+    }
+}
+
+/// A counter that tracks the next index of a field using a monotonically increasing counter.
+#[derive(Default, Debug, PartialEq, Eq, Clone, Copy)]
+pub struct FieldIndexCounter(u32);
+
+impl FieldIndexCounter {
+    /// Increments the number stored by this counter, returning the previous value.
+    ///
+    /// Functionally equivalent to `i++` in C-based languages.
+    pub fn increment(&mut self) -> u32 {
+        let v = self.0;
+        self.0 += 1;
+        v
+    }
+}

nac3core/src/codegen/values/array.rs

@@ -0,0 +1,426 @@
+use inkwell::{
+    types::AnyTypeEnum,
+    values::{BasicValueEnum, IntValue, PointerValue},
+    IntPredicate,
+};
+
+use crate::codegen::{CodeGenContext, CodeGenerator};
+
+/// An LLVM value that is array-like, i.e. it contains a contiguous, sequenced collection of
+/// elements.
+pub trait ArrayLikeValue<'ctx> {
+    /// Returns the element type of this array-like value.
+    fn element_type<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> AnyTypeEnum<'ctx>;
+
+    /// Returns the base pointer to the array.
+    fn base_ptr<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> PointerValue<'ctx>;
+
+    /// Returns the size of this array-like value.
+    fn size<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> IntValue<'ctx>;
+
+    /// Returns a [`ArraySliceValue`] representing this value.
+    fn as_slice_value<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> ArraySliceValue<'ctx> {
+        ArraySliceValue::from_ptr_val(
+            self.base_ptr(ctx, generator),
+            self.size(ctx, generator),
+            None,
+        )
+    }
+}
+
+/// An array-like value that can be indexed by memory offset.
+pub trait ArrayLikeIndexer<'ctx, Index = IntValue<'ctx>>: ArrayLikeValue<'ctx> {
+    /// # Safety
+    ///
+    /// This function should be called with a valid index.
+    unsafe fn ptr_offset_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx>;
+
+    /// Returns the pointer to the data at the `idx`-th index.
+    fn ptr_offset<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx>;
+}
+
+/// An array-like value that can have its array elements accessed as a [`BasicValueEnum`].
+pub trait UntypedArrayLikeAccessor<'ctx, Index = IntValue<'ctx>>:
+    ArrayLikeIndexer<'ctx, Index>
+{
+    /// # Safety
+    ///
+    /// This function should be called with a valid index.
+    unsafe fn get_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> BasicValueEnum<'ctx> {
+        let ptr = unsafe { self.ptr_offset_unchecked(ctx, generator, idx, name) };
+        ctx.builder.build_load(ptr, name.unwrap_or_default()).unwrap()
+    }
+
+    /// Returns the data at the `idx`-th index.
+    fn get<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> BasicValueEnum<'ctx> {
+        let ptr = self.ptr_offset(ctx, generator, idx, name);
+        ctx.builder.build_load(ptr, name.unwrap_or_default()).unwrap()
+    }
+}
+
+/// An array-like value that can have its array elements mutated as a [`BasicValueEnum`].
+pub trait UntypedArrayLikeMutator<'ctx, Index = IntValue<'ctx>>:
+    ArrayLikeIndexer<'ctx, Index>
+{
+    /// # Safety
+    ///
+    /// This function should be called with a valid index.
+    unsafe fn set_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        value: BasicValueEnum<'ctx>,
+    ) {
+        let ptr = unsafe { self.ptr_offset_unchecked(ctx, generator, idx, None) };
+        ctx.builder.build_store(ptr, value).unwrap();
+    }
+
+    /// Sets the data at the `idx`-th index.
+    fn set<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        value: BasicValueEnum<'ctx>,
+    ) {
+        let ptr = self.ptr_offset(ctx, generator, idx, None);
+        ctx.builder.build_store(ptr, value).unwrap();
+    }
+}
+
+/// An array-like value that can have its array elements accessed as an arbitrary type `T`.
+pub trait TypedArrayLikeAccessor<'ctx, T, Index = IntValue<'ctx>>:
+    UntypedArrayLikeAccessor<'ctx, Index>
+{
+    /// Casts an element from [`BasicValueEnum`] into `T`.
+    fn downcast_to_type(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        value: BasicValueEnum<'ctx>,
+    ) -> T;
+
+    /// # Safety
+    ///
+    /// This function should be called with a valid index.
+    unsafe fn get_typed_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> T {
+        let value = unsafe { self.get_unchecked(ctx, generator, idx, name) };
+        self.downcast_to_type(ctx, value)
+    }
+
+    /// Returns the data at the `idx`-th index.
+    fn get_typed<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> T {
+        let value = self.get(ctx, generator, idx, name);
+        self.downcast_to_type(ctx, value)
+    }
+}
+
+/// An array-like value that can have its array elements mutated as an arbitrary type `T`.
+pub trait TypedArrayLikeMutator<'ctx, T, Index = IntValue<'ctx>>:
+    UntypedArrayLikeMutator<'ctx, Index>
+{
+    /// Casts an element from T into [`BasicValueEnum`].
+    fn upcast_from_type(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        value: T,
+    ) -> BasicValueEnum<'ctx>;
+
+    /// # Safety
+    ///
+    /// This function should be called with a valid index.
+    unsafe fn set_typed_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        value: T,
+    ) {
+        let value = self.upcast_from_type(ctx, value);
+        unsafe { self.set_unchecked(ctx, generator, idx, value) }
+    }
+
+    /// Sets the data at the `idx`-th index.
+    fn set_typed<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        value: T,
+    ) {
+        let value = self.upcast_from_type(ctx, value);
+        self.set(ctx, generator, idx, value);
+    }
+}
+
+/// Type alias for a function that casts a [`BasicValueEnum`] into a `T`.
+type ValueDowncastFn<'ctx, T> =
+    Box<dyn Fn(&mut CodeGenContext<'ctx, '_>, BasicValueEnum<'ctx>) -> T>;
+/// Type alias for a function that casts a `T` into a [`BasicValueEnum`].
+type ValueUpcastFn<'ctx, T> = Box<dyn Fn(&mut CodeGenContext<'ctx, '_>, T) -> BasicValueEnum<'ctx>>;
+
+/// An adapter for constraining untyped array values as typed values.
+pub struct TypedArrayLikeAdapter<'ctx, T, Adapted: ArrayLikeValue<'ctx> = ArraySliceValue<'ctx>> {
+    adapted: Adapted,
+    downcast_fn: ValueDowncastFn<'ctx, T>,
+    upcast_fn: ValueUpcastFn<'ctx, T>,
+}
+
+impl<'ctx, T, Adapted> TypedArrayLikeAdapter<'ctx, T, Adapted>
+where
+    Adapted: ArrayLikeValue<'ctx>,
+{
+    /// Creates a [`TypedArrayLikeAdapter`].
+    ///
+    /// * `adapted` - The value to be adapted.
+    /// * `downcast_fn` - The function converting a [`BasicValueEnum`] into a `T`.
+    /// * `upcast_fn` - The function converting a T into a [`BasicValueEnum`].
+    pub fn from(
+        adapted: Adapted,
+        downcast_fn: ValueDowncastFn<'ctx, T>,
+        upcast_fn: ValueUpcastFn<'ctx, T>,
+    ) -> Self {
+        TypedArrayLikeAdapter { adapted, downcast_fn, upcast_fn }
+    }
+}
+
+impl<'ctx, T, Adapted> ArrayLikeValue<'ctx> for TypedArrayLikeAdapter<'ctx, T, Adapted>
+where
+    Adapted: ArrayLikeValue<'ctx>,
+{
+    fn element_type<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> AnyTypeEnum<'ctx> {
+        self.adapted.element_type(ctx, generator)
+    }
+
+    fn base_ptr<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> PointerValue<'ctx> {
+        self.adapted.base_ptr(ctx, generator)
+    }
+
+    fn size<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> IntValue<'ctx> {
+        self.adapted.size(ctx, generator)
+    }
+}
+
+impl<'ctx, T, Index, Adapted> ArrayLikeIndexer<'ctx, Index>
+    for TypedArrayLikeAdapter<'ctx, T, Adapted>
+where
+    Adapted: ArrayLikeIndexer<'ctx, Index>,
+{
+    unsafe fn ptr_offset_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        unsafe { self.adapted.ptr_offset_unchecked(ctx, generator, idx, name) }
+    }
+
+    fn ptr_offset<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &Index,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        self.adapted.ptr_offset(ctx, generator, idx, name)
+    }
+}
+
+impl<'ctx, T, Index, Adapted> UntypedArrayLikeAccessor<'ctx, Index>
+    for TypedArrayLikeAdapter<'ctx, T, Adapted>
+where
+    Adapted: UntypedArrayLikeAccessor<'ctx, Index>,
+{
+}
+impl<'ctx, T, Index, Adapted> UntypedArrayLikeMutator<'ctx, Index>
+    for TypedArrayLikeAdapter<'ctx, T, Adapted>
+where
+    Adapted: UntypedArrayLikeMutator<'ctx, Index>,
+{
+}
+
+impl<'ctx, T, Index, Adapted> TypedArrayLikeAccessor<'ctx, T, Index>
+    for TypedArrayLikeAdapter<'ctx, T, Adapted>
+where
+    Adapted: UntypedArrayLikeAccessor<'ctx, Index>,
+{
+    fn downcast_to_type(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        value: BasicValueEnum<'ctx>,
+    ) -> T {
+        (self.downcast_fn)(ctx, value)
+    }
+}
+
+impl<'ctx, T, Index, Adapted> TypedArrayLikeMutator<'ctx, T, Index>
+    for TypedArrayLikeAdapter<'ctx, T, Adapted>
+where
+    Adapted: UntypedArrayLikeMutator<'ctx, Index>,
+{
+    fn upcast_from_type(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        value: T,
+    ) -> BasicValueEnum<'ctx> {
+        (self.upcast_fn)(ctx, value)
+    }
+}
+
+/// An LLVM value representing an array slice, consisting of a pointer to the data and the size of
+/// the slice.
+#[derive(Copy, Clone)]
+pub struct ArraySliceValue<'ctx>(PointerValue<'ctx>, IntValue<'ctx>, Option<&'ctx str>);
+
+impl<'ctx> ArraySliceValue<'ctx> {
+    /// Creates an [`ArraySliceValue`] from a [`PointerValue`] and its size.
+    #[must_use]
+    pub fn from_ptr_val(
+        ptr: PointerValue<'ctx>,
+        size: IntValue<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> Self {
+        ArraySliceValue(ptr, size, name)
+    }
+}
+
+impl<'ctx> From<ArraySliceValue<'ctx>> for PointerValue<'ctx> {
+    fn from(value: ArraySliceValue<'ctx>) -> Self {
+        value.0
+    }
+}
+
+impl<'ctx> ArrayLikeValue<'ctx> for ArraySliceValue<'ctx> {
+    fn element_type<G: CodeGenerator + ?Sized>(
+        &self,
+        _: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> AnyTypeEnum<'ctx> {
+        self.0.get_type().get_element_type()
+    }
+
+    fn base_ptr<G: CodeGenerator + ?Sized>(
+        &self,
+        _: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> PointerValue<'ctx> {
+        self.0
+    }
+
+    fn size<G: CodeGenerator + ?Sized>(
+        &self,
+        _: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> IntValue<'ctx> {
+        self.1
+    }
+}
+
+impl<'ctx> ArrayLikeIndexer<'ctx> for ArraySliceValue<'ctx> {
+    unsafe fn ptr_offset_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let var_name = name.map(|v| format!("{v}.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(self.base_ptr(ctx, generator), &[*idx], var_name.as_str())
+                .unwrap()
+        }
+    }
+
+    fn ptr_offset<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        debug_assert_eq!(idx.get_type(), generator.get_size_type(ctx.ctx));
+
+        let size = self.size(ctx, generator);
+        let in_range = ctx.builder.build_int_compare(IntPredicate::ULT, *idx, size, "").unwrap();
+        ctx.make_assert(
+            generator,
+            in_range,
+            "0:IndexError",
+            "list index out of range",
+            [None, None, None],
+            ctx.current_loc,
+        );
+
+        unsafe { self.ptr_offset_unchecked(ctx, generator, idx, name) }
+    }
+}
+
+impl<'ctx> UntypedArrayLikeAccessor<'ctx> for ArraySliceValue<'ctx> {}
+impl<'ctx> UntypedArrayLikeMutator<'ctx> for ArraySliceValue<'ctx> {}

nac3core/src/codegen/values/list.rs

@@ -0,0 +1,241 @@
+use inkwell::{
+    types::{AnyTypeEnum, BasicType, BasicTypeEnum, IntType},
+    values::{BasicValueEnum, IntValue, PointerValue},
+    AddressSpace, IntPredicate,
+};
+
+use super::{
+    ArrayLikeIndexer, ArrayLikeValue, ProxyValue, UntypedArrayLikeAccessor, UntypedArrayLikeMutator,
+};
+use crate::codegen::{
+    types::ListType,
+    {CodeGenContext, CodeGenerator},
+};
+
+/// Proxy type for accessing a `list` value in LLVM.
+#[derive(Copy, Clone)]
+pub struct ListValue<'ctx> {
+    value: PointerValue<'ctx>,
+    llvm_usize: IntType<'ctx>,
+    name: Option<&'ctx str>,
+}
+
+impl<'ctx> ListValue<'ctx> {
+    /// Checks whether `value` is an instance of `list`, returning [Err] if `value` is not an
+    /// instance.
+    pub fn is_representable(
+        value: PointerValue<'ctx>,
+        llvm_usize: IntType<'ctx>,
+    ) -> Result<(), String> {
+        ListType::is_representable(value.get_type(), llvm_usize)
+    }
+
+    /// Creates an [`ListValue`] from a [`PointerValue`].
+    #[must_use]
+    pub fn from_pointer_value(
+        ptr: PointerValue<'ctx>,
+        llvm_usize: IntType<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> Self {
+        debug_assert!(Self::is_representable(ptr, llvm_usize).is_ok());
+
+        ListValue { value: ptr, llvm_usize, name }
+    }
+
+    /// Returns the double-indirection pointer to the `data` array, as if by calling `getelementptr`
+    /// on the field.
+    fn pptr_to_data(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        let llvm_i32 = ctx.ctx.i32_type();
+        let var_name = self.name.map(|v| format!("{v}.data.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.as_base_value(),
+                    &[llvm_i32.const_zero(), llvm_i32.const_zero()],
+                    var_name.as_str(),
+                )
+                .unwrap()
+        }
+    }
+
+    /// Returns the pointer to the field storing the size of this `list`.
+    fn ptr_to_size(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        let llvm_i32 = ctx.ctx.i32_type();
+        let var_name = self.name.map(|v| format!("{v}.size.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.as_base_value(),
+                    &[llvm_i32.const_zero(), llvm_i32.const_int(1, true)],
+                    var_name.as_str(),
+                )
+                .unwrap()
+        }
+    }
+
+    /// Stores the array of data elements `data` into this instance.
+    fn store_data(&self, ctx: &CodeGenContext<'ctx, '_>, data: PointerValue<'ctx>) {
+        ctx.builder.build_store(self.pptr_to_data(ctx), data).unwrap();
+    }
+
+    /// Convenience method for creating a new array storing data elements with the given element
+    /// type `elem_ty` and `size`.
+    ///
+    /// If `size` is [None], the size stored in the field of this instance is used instead.
+    pub fn create_data(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        elem_ty: BasicTypeEnum<'ctx>,
+        size: Option<IntValue<'ctx>>,
+    ) {
+        let size = size.unwrap_or_else(|| self.load_size(ctx, None));
+
+        let data = ctx
+            .builder
+            .build_select(
+                ctx.builder
+                    .build_int_compare(IntPredicate::NE, size, self.llvm_usize.const_zero(), "")
+                    .unwrap(),
+                ctx.builder.build_array_alloca(elem_ty, size, "").unwrap(),
+                elem_ty.ptr_type(AddressSpace::default()).const_zero(),
+                "",
+            )
+            .map(BasicValueEnum::into_pointer_value)
+            .unwrap();
+        self.store_data(ctx, data);
+    }
+
+    /// Returns the double-indirection pointer to the `data` array, as if by calling `getelementptr`
+    /// on the field.
+    #[must_use]
+    pub fn data(&self) -> ListDataProxy<'ctx, '_> {
+        ListDataProxy(self)
+    }
+
+    /// Stores the `size` of this `list` into this instance.
+    pub fn store_size<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+        size: IntValue<'ctx>,
+    ) {
+        debug_assert_eq!(size.get_type(), generator.get_size_type(ctx.ctx));
+
+        let psize = self.ptr_to_size(ctx);
+        ctx.builder.build_store(psize, size).unwrap();
+    }
+
+    /// Returns the size of this `list` as a value.
+    pub fn load_size(&self, ctx: &CodeGenContext<'ctx, '_>, name: Option<&str>) -> IntValue<'ctx> {
+        let psize = self.ptr_to_size(ctx);
+        let var_name = name
+            .map(ToString::to_string)
+            .or_else(|| self.name.map(|v| format!("{v}.size")))
+            .unwrap_or_default();
+
+        ctx.builder
+            .build_load(psize, var_name.as_str())
+            .map(BasicValueEnum::into_int_value)
+            .unwrap()
+    }
+}
+
+impl<'ctx> ProxyValue<'ctx> for ListValue<'ctx> {
+    type Base = PointerValue<'ctx>;
+    type Type = ListType<'ctx>;
+
+    fn get_type(&self) -> Self::Type {
+        ListType::from_type(self.as_base_value().get_type(), self.llvm_usize)
+    }
+
+    fn as_base_value(&self) -> Self::Base {
+        self.value
+    }
+}
+
+impl<'ctx> From<ListValue<'ctx>> for PointerValue<'ctx> {
+    fn from(value: ListValue<'ctx>) -> Self {
+        value.as_base_value()
+    }
+}
+
+/// Proxy type for accessing the `data` array of an `list` instance in LLVM.
+#[derive(Copy, Clone)]
+pub struct ListDataProxy<'ctx, 'a>(&'a ListValue<'ctx>);
+
+impl<'ctx> ArrayLikeValue<'ctx> for ListDataProxy<'ctx, '_> {
+    fn element_type<G: CodeGenerator + ?Sized>(
+        &self,
+        _: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> AnyTypeEnum<'ctx> {
+        self.0.value.get_type().get_element_type()
+    }
+
+    fn base_ptr<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> PointerValue<'ctx> {
+        let var_name = self.0.name.map(|v| format!("{v}.data")).unwrap_or_default();
+
+        ctx.builder
+            .build_load(self.0.pptr_to_data(ctx), var_name.as_str())
+            .map(BasicValueEnum::into_pointer_value)
+            .unwrap()
+    }
+
+    fn size<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> IntValue<'ctx> {
+        self.0.load_size(ctx, None)
+    }
+}
+
+impl<'ctx> ArrayLikeIndexer<'ctx> for ListDataProxy<'ctx, '_> {
+    unsafe fn ptr_offset_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let var_name = name.map(|v| format!("{v}.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(self.base_ptr(ctx, generator), &[*idx], var_name.as_str())
+                .unwrap()
+        }
+    }
+
+    fn ptr_offset<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        debug_assert_eq!(idx.get_type(), generator.get_size_type(ctx.ctx));
+
+        let size = self.size(ctx, generator);
+        let in_range = ctx.builder.build_int_compare(IntPredicate::ULT, *idx, size, "").unwrap();
+        ctx.make_assert(
+            generator,
+            in_range,
+            "0:IndexError",
+            "list index out of range",
+            [None, None, None],
+            ctx.current_loc,
+        );
+
+        unsafe { self.ptr_offset_unchecked(ctx, generator, idx, name) }
+    }
+}
+
+impl<'ctx> UntypedArrayLikeAccessor<'ctx> for ListDataProxy<'ctx, '_> {}
+impl<'ctx> UntypedArrayLikeMutator<'ctx> for ListDataProxy<'ctx, '_> {}

nac3core/src/codegen/values/mod.rs

@@ -0,0 +1,47 @@
+use inkwell::{context::Context, values::BasicValue};
+
+use super::types::ProxyType;
+use crate::codegen::CodeGenerator;
+pub use array::*;
+pub use list::*;
+pub use ndarray::*;
+pub use range::*;
+
+mod array;
+mod list;
+mod ndarray;
+mod range;
+
+/// A LLVM type that is used to represent a non-primitive value in NAC3.
+pub trait ProxyValue<'ctx>: Into<Self::Base> {
+    /// The type of LLVM values represented by this instance. This is usually the
+    /// [LLVM pointer type][PointerValue].
+    type Base: BasicValue<'ctx>;
+
+    /// The type of this value.
+    type Type: ProxyType<'ctx, Value = Self>;
+
+    /// Checks whether `value` can be represented by this [`ProxyValue`].
+    fn is_instance<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        value: impl BasicValue<'ctx>,
+    ) -> Result<(), String> {
+        Self::Type::is_type(generator, ctx, value.as_basic_value_enum().get_type())
+    }
+
+    /// Checks whether `value` can be represented by this [`ProxyValue`].
+    fn is_representable<G: CodeGenerator + ?Sized>(
+        generator: &G,
+        ctx: &'ctx Context,
+        value: Self::Base,
+    ) -> Result<(), String> {
+        Self::is_instance(generator, ctx, value.as_basic_value_enum())
+    }
+
+    /// Returns the [type][ProxyType] of this value.
+    fn get_type(&self) -> Self::Type;
+
+    /// Returns the [base value][Self::Base] of this proxy.
+    fn as_base_value(&self) -> Self::Base;
+}

nac3core/src/codegen/values/ndarray.rs

@@ -0,0 +1,523 @@
+use inkwell::{
+    types::{AnyType, AnyTypeEnum, BasicType, BasicTypeEnum, IntType},
+    values::{BasicValueEnum, IntValue, PointerValue},
+    AddressSpace, IntPredicate,
+};
+
+use super::{
+    ArrayLikeIndexer, ArrayLikeValue, ProxyValue, TypedArrayLikeAccessor, TypedArrayLikeMutator,
+    UntypedArrayLikeAccessor, UntypedArrayLikeMutator,
+};
+use crate::codegen::{
+    irrt::{call_ndarray_calc_size, call_ndarray_flatten_index},
+    llvm_intrinsics::call_int_umin,
+    stmt::gen_for_callback_incrementing,
+    types::NDArrayType,
+    CodeGenContext, CodeGenerator,
+};
+
+/// Proxy type for accessing an `NDArray` value in LLVM.
+#[derive(Copy, Clone)]
+pub struct NDArrayValue<'ctx> {
+    value: PointerValue<'ctx>,
+    dtype: BasicTypeEnum<'ctx>,
+    llvm_usize: IntType<'ctx>,
+    name: Option<&'ctx str>,
+}
+
+impl<'ctx> NDArrayValue<'ctx> {
+    /// Checks whether `value` is an instance of `NDArray`, returning [Err] if `value` is not an
+    /// instance.
+    pub fn is_representable(
+        value: PointerValue<'ctx>,
+        llvm_usize: IntType<'ctx>,
+    ) -> Result<(), String> {
+        NDArrayType::is_representable(value.get_type(), llvm_usize)
+    }
+
+    /// Creates an [`NDArrayValue`] from a [`PointerValue`].
+    #[must_use]
+    pub fn from_pointer_value(
+        ptr: PointerValue<'ctx>,
+        dtype: BasicTypeEnum<'ctx>,
+        llvm_usize: IntType<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> Self {
+        debug_assert!(Self::is_representable(ptr, llvm_usize).is_ok());
+
+        NDArrayValue { value: ptr, dtype, llvm_usize, name }
+    }
+
+    /// Returns the pointer to the field storing the number of dimensions of this `NDArray`.
+    fn ptr_to_ndims(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        self.get_type()
+            .get_fields(ctx.ctx, self.llvm_usize)
+            .ndims
+            .ptr_by_gep(ctx, self.value, self.name)
+    }
+
+    /// Stores the number of dimensions `ndims` into this instance.
+    pub fn store_ndims<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+        ndims: IntValue<'ctx>,
+    ) {
+        debug_assert_eq!(ndims.get_type(), generator.get_size_type(ctx.ctx));
+
+        let pndims = self.ptr_to_ndims(ctx);
+        ctx.builder.build_store(pndims, ndims).unwrap();
+    }
+
+    /// Returns the number of dimensions of this `NDArray` as a value.
+    pub fn load_ndims(&self, ctx: &CodeGenContext<'ctx, '_>) -> IntValue<'ctx> {
+        let pndims = self.ptr_to_ndims(ctx);
+        ctx.builder.build_load(pndims, "").map(BasicValueEnum::into_int_value).unwrap()
+    }
+
+    /// Returns the double-indirection pointer to the `shape` array, as if by calling
+    /// `getelementptr` on the field.
+    fn ptr_to_shape(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        self.get_type()
+            .get_fields(ctx.ctx, self.llvm_usize)
+            .shape
+            .ptr_by_gep(ctx, self.value, self.name)
+    }
+
+    /// Stores the array of dimension sizes `dims` into this instance.
+    fn store_shape(&self, ctx: &CodeGenContext<'ctx, '_>, dims: PointerValue<'ctx>) {
+        ctx.builder.build_store(self.ptr_to_shape(ctx), dims).unwrap();
+    }
+
+    /// Convenience method for creating a new array storing dimension sizes with the given `size`.
+    pub fn create_shape(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        llvm_usize: IntType<'ctx>,
+        size: IntValue<'ctx>,
+    ) {
+        self.store_shape(ctx, ctx.builder.build_array_alloca(llvm_usize, size, "").unwrap());
+    }
+
+    /// Returns a proxy object to the field storing the size of each dimension of this `NDArray`.
+    #[must_use]
+    pub fn shape(&self) -> NDArrayShapeProxy<'ctx, '_> {
+        NDArrayShapeProxy(self)
+    }
+
+    /// Returns the double-indirection pointer to the `data` array, as if by calling `getelementptr`
+    /// on the field.
+    pub fn ptr_to_data(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        self.get_type()
+            .get_fields(ctx.ctx, self.llvm_usize)
+            .data
+            .ptr_by_gep(ctx, self.value, self.name)
+    }
+
+    /// Stores the array of data elements `data` into this instance.
+    fn store_data(&self, ctx: &CodeGenContext<'ctx, '_>, data: PointerValue<'ctx>) {
+        let data = ctx
+            .builder
+            .build_bit_cast(data, ctx.ctx.i8_type().ptr_type(AddressSpace::default()), "")
+            .unwrap();
+        ctx.builder.build_store(self.ptr_to_data(ctx), data).unwrap();
+    }
+
+    /// Convenience method for creating a new array storing data elements with the given element
+    /// type `elem_ty` and `size`.
+    pub fn create_data(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        elem_ty: BasicTypeEnum<'ctx>,
+        size: IntValue<'ctx>,
+    ) {
+        let itemsize =
+            ctx.builder.build_int_cast(elem_ty.size_of().unwrap(), size.get_type(), "").unwrap();
+        let nbytes = ctx.builder.build_int_mul(size, itemsize, "").unwrap();
+
+        // TODO: What about alignment?
+        self.store_data(
+            ctx,
+            ctx.builder.build_array_alloca(ctx.ctx.i8_type(), nbytes, "").unwrap(),
+        );
+    }
+
+    /// Returns a proxy object to the field storing the data of this `NDArray`.
+    #[must_use]
+    pub fn data(&self) -> NDArrayDataProxy<'ctx, '_> {
+        NDArrayDataProxy(self)
+    }
+}
+
+impl<'ctx> ProxyValue<'ctx> for NDArrayValue<'ctx> {
+    type Base = PointerValue<'ctx>;
+    type Type = NDArrayType<'ctx>;
+
+    fn get_type(&self) -> Self::Type {
+        NDArrayType::from_type(self.as_base_value().get_type(), self.dtype, self.llvm_usize)
+    }
+
+    fn as_base_value(&self) -> Self::Base {
+        self.value
+    }
+}
+
+impl<'ctx> From<NDArrayValue<'ctx>> for PointerValue<'ctx> {
+    fn from(value: NDArrayValue<'ctx>) -> Self {
+        value.as_base_value()
+    }
+}
+
+/// Proxy type for accessing the `dims` array of an `NDArray` instance in LLVM.
+#[derive(Copy, Clone)]
+pub struct NDArrayShapeProxy<'ctx, 'a>(&'a NDArrayValue<'ctx>);
+
+impl<'ctx> ArrayLikeValue<'ctx> for NDArrayShapeProxy<'ctx, '_> {
+    fn element_type<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> AnyTypeEnum<'ctx> {
+        self.0.shape().base_ptr(ctx, generator).get_type().get_element_type()
+    }
+
+    fn base_ptr<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> PointerValue<'ctx> {
+        let var_name = self.0.name.map(|v| format!("{v}.data")).unwrap_or_default();
+
+        ctx.builder
+            .build_load(self.0.ptr_to_shape(ctx), var_name.as_str())
+            .map(BasicValueEnum::into_pointer_value)
+            .unwrap()
+    }
+
+    fn size<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> IntValue<'ctx> {
+        self.0.load_ndims(ctx)
+    }
+}
+
+impl<'ctx> ArrayLikeIndexer<'ctx, IntValue<'ctx>> for NDArrayShapeProxy<'ctx, '_> {
+    unsafe fn ptr_offset_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let var_name = name.map(|v| format!("{v}.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(self.base_ptr(ctx, generator), &[*idx], var_name.as_str())
+                .unwrap()
+        }
+    }
+
+    fn ptr_offset<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let size = self.size(ctx, generator);
+        let in_range = ctx.builder.build_int_compare(IntPredicate::ULT, *idx, size, "").unwrap();
+        ctx.make_assert(
+            generator,
+            in_range,
+            "0:IndexError",
+            "index {0} is out of bounds for axis 0 with size {1}",
+            [Some(*idx), Some(self.0.load_ndims(ctx)), None],
+            ctx.current_loc,
+        );
+
+        unsafe { self.ptr_offset_unchecked(ctx, generator, idx, name) }
+    }
+}
+
+impl<'ctx> UntypedArrayLikeAccessor<'ctx, IntValue<'ctx>> for NDArrayShapeProxy<'ctx, '_> {}
+impl<'ctx> UntypedArrayLikeMutator<'ctx, IntValue<'ctx>> for NDArrayShapeProxy<'ctx, '_> {}
+
+impl<'ctx> TypedArrayLikeAccessor<'ctx, IntValue<'ctx>> for NDArrayShapeProxy<'ctx, '_> {
+    fn downcast_to_type(
+        &self,
+        _: &mut CodeGenContext<'ctx, '_>,
+        value: BasicValueEnum<'ctx>,
+    ) -> IntValue<'ctx> {
+        value.into_int_value()
+    }
+}
+
+impl<'ctx> TypedArrayLikeMutator<'ctx, IntValue<'ctx>> for NDArrayShapeProxy<'ctx, '_> {
+    fn upcast_from_type(
+        &self,
+        _: &mut CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> BasicValueEnum<'ctx> {
+        value.into()
+    }
+}
+
+/// Proxy type for accessing the `data` array of an `NDArray` instance in LLVM.
+#[derive(Copy, Clone)]
+pub struct NDArrayDataProxy<'ctx, 'a>(&'a NDArrayValue<'ctx>);
+
+impl<'ctx> ArrayLikeValue<'ctx> for NDArrayDataProxy<'ctx, '_> {
+    fn element_type<G: CodeGenerator + ?Sized>(
+        &self,
+        _: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> AnyTypeEnum<'ctx> {
+        self.0.dtype.as_any_type_enum()
+    }
+
+    fn base_ptr<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        _: &G,
+    ) -> PointerValue<'ctx> {
+        let var_name = self.0.name.map(|v| format!("{v}.data")).unwrap_or_default();
+
+        ctx.builder
+            .build_load(self.0.ptr_to_data(ctx), var_name.as_str())
+            .map(BasicValueEnum::into_pointer_value)
+            .unwrap()
+    }
+
+    fn size<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        generator: &G,
+    ) -> IntValue<'ctx> {
+        call_ndarray_calc_size(generator, ctx, &self.as_slice_value(ctx, generator), (None, None))
+    }
+}
+
+impl<'ctx> ArrayLikeIndexer<'ctx> for NDArrayDataProxy<'ctx, '_> {
+    unsafe fn ptr_offset_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let sizeof_elem = ctx
+            .builder
+            .build_int_truncate_or_bit_cast(
+                self.element_type(ctx, generator).size_of().unwrap(),
+                idx.get_type(),
+                "",
+            )
+            .unwrap();
+        let idx = ctx.builder.build_int_mul(*idx, sizeof_elem, "").unwrap();
+        let ptr = unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.base_ptr(ctx, generator),
+                    &[idx],
+                    name.unwrap_or_default(),
+                )
+                .unwrap()
+        };
+
+        // Current implementation is transparent - The returned pointer type is
+        // already cast into the expected type, allowing for immediately
+        // load/store.
+        ctx.builder
+            .build_pointer_cast(
+                ptr,
+                BasicTypeEnum::try_from(self.element_type(ctx, generator))
+                    .unwrap()
+                    .ptr_type(AddressSpace::default()),
+                "",
+            )
+            .unwrap()
+    }
+
+    fn ptr_offset<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        idx: &IntValue<'ctx>,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let data_sz = self.size(ctx, generator);
+        let in_range = ctx.builder.build_int_compare(IntPredicate::ULT, *idx, data_sz, "").unwrap();
+        ctx.make_assert(
+            generator,
+            in_range,
+            "0:IndexError",
+            "index {0} is out of bounds with size {1}",
+            [Some(*idx), Some(self.0.load_ndims(ctx)), None],
+            ctx.current_loc,
+        );
+
+        let ptr = unsafe { self.ptr_offset_unchecked(ctx, generator, idx, name) };
+
+        // Current implementation is transparent - The returned pointer type is
+        // already cast into the expected type, allowing for immediately
+        // load/store.
+        ctx.builder
+            .build_pointer_cast(
+                ptr,
+                BasicTypeEnum::try_from(self.element_type(ctx, generator))
+                    .unwrap()
+                    .ptr_type(AddressSpace::default()),
+                "",
+            )
+            .unwrap()
+    }
+}
+
+impl<'ctx> UntypedArrayLikeAccessor<'ctx, IntValue<'ctx>> for NDArrayDataProxy<'ctx, '_> {}
+impl<'ctx> UntypedArrayLikeMutator<'ctx, IntValue<'ctx>> for NDArrayDataProxy<'ctx, '_> {}
+
+impl<'ctx, Index: UntypedArrayLikeAccessor<'ctx>> ArrayLikeIndexer<'ctx, Index>
+    for NDArrayDataProxy<'ctx, '_>
+{
+    unsafe fn ptr_offset_unchecked<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        indices: &Index,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let llvm_usize = generator.get_size_type(ctx.ctx);
+
+        let indices_elem_ty = indices
+            .ptr_offset(ctx, generator, &llvm_usize.const_zero(), None)
+            .get_type()
+            .get_element_type();
+        let Ok(indices_elem_ty) = IntType::try_from(indices_elem_ty) else {
+            panic!("Expected list[int32] but got {indices_elem_ty}")
+        };
+        assert_eq!(
+            indices_elem_ty.get_bit_width(),
+            32,
+            "Expected list[int32] but got list[int{}]",
+            indices_elem_ty.get_bit_width()
+        );
+
+        let index = call_ndarray_flatten_index(generator, ctx, *self.0, indices);
+        let sizeof_elem = ctx
+            .builder
+            .build_int_truncate_or_bit_cast(
+                self.element_type(ctx, generator).size_of().unwrap(),
+                index.get_type(),
+                "",
+            )
+            .unwrap();
+        let index = ctx.builder.build_int_mul(index, sizeof_elem, "").unwrap();
+
+        let ptr = unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.base_ptr(ctx, generator),
+                    &[index],
+                    name.unwrap_or_default(),
+                )
+                .unwrap()
+        };
+        // TODO: Current implementation is transparent
+        ctx.builder
+            .build_pointer_cast(
+                ptr,
+                BasicTypeEnum::try_from(self.element_type(ctx, generator))
+                    .unwrap()
+                    .ptr_type(AddressSpace::default()),
+                "",
+            )
+            .unwrap()
+    }
+
+    fn ptr_offset<G: CodeGenerator + ?Sized>(
+        &self,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut G,
+        indices: &Index,
+        name: Option<&str>,
+    ) -> PointerValue<'ctx> {
+        let llvm_usize = generator.get_size_type(ctx.ctx);
+
+        let indices_size = indices.size(ctx, generator);
+        let nidx_leq_ndims = ctx
+            .builder
+            .build_int_compare(IntPredicate::SLE, indices_size, self.0.load_ndims(ctx), "")
+            .unwrap();
+        ctx.make_assert(
+            generator,
+            nidx_leq_ndims,
+            "0:IndexError",
+            "invalid index to scalar variable",
+            [None, None, None],
+            ctx.current_loc,
+        );
+
+        let indices_len = indices.size(ctx, generator);
+        let ndarray_len = self.0.load_ndims(ctx);
+        let len = call_int_umin(ctx, indices_len, ndarray_len, None);
+        gen_for_callback_incrementing(
+            generator,
+            ctx,
+            None,
+            llvm_usize.const_zero(),
+            (len, false),
+            |generator, ctx, _, i| {
+                let (dim_idx, dim_sz) = unsafe {
+                    (
+                        indices.get_unchecked(ctx, generator, &i, None).into_int_value(),
+                        self.0.shape().get_typed_unchecked(ctx, generator, &i, None),
+                    )
+                };
+                let dim_idx = ctx
+                    .builder
+                    .build_int_z_extend_or_bit_cast(dim_idx, dim_sz.get_type(), "")
+                    .unwrap();
+
+                let dim_lt =
+                    ctx.builder.build_int_compare(IntPredicate::SLT, dim_idx, dim_sz, "").unwrap();
+
+                ctx.make_assert(
+                    generator,
+                    dim_lt,
+                    "0:IndexError",
+                    "index {0} is out of bounds for axis 0 with size {1}",
+                    [Some(dim_idx), Some(dim_sz), None],
+                    ctx.current_loc,
+                );
+
+                Ok(())
+            },
+            llvm_usize.const_int(1, false),
+        )
+        .unwrap();
+
+        let ptr = unsafe { self.ptr_offset_unchecked(ctx, generator, indices, name) };
+        // TODO: Current implementation is transparent
+        ctx.builder
+            .build_pointer_cast(
+                ptr,
+                BasicTypeEnum::try_from(self.element_type(ctx, generator))
+                    .unwrap()
+                    .ptr_type(AddressSpace::default()),
+                "",
+            )
+            .unwrap()
+    }
+}
+
+impl<'ctx, Index: UntypedArrayLikeAccessor<'ctx>> UntypedArrayLikeAccessor<'ctx, Index>
+    for NDArrayDataProxy<'ctx, '_>
+{
+}
+impl<'ctx, Index: UntypedArrayLikeAccessor<'ctx>> UntypedArrayLikeMutator<'ctx, Index>
+    for NDArrayDataProxy<'ctx, '_>
+{
+}

nac3core/src/codegen/values/range.rs

@@ -0,0 +1,153 @@
+use inkwell::values::{BasicValueEnum, IntValue, PointerValue};
+
+use super::ProxyValue;
+use crate::codegen::{types::RangeType, CodeGenContext};
+
+/// Proxy type for accessing a `range` value in LLVM.
+#[derive(Copy, Clone)]
+pub struct RangeValue<'ctx> {
+    value: PointerValue<'ctx>,
+    name: Option<&'ctx str>,
+}
+
+impl<'ctx> RangeValue<'ctx> {
+    /// Checks whether `value` is an instance of `range`, returning [Err] if `value` is not an instance.
+    pub fn is_representable(value: PointerValue<'ctx>) -> Result<(), String> {
+        RangeType::is_representable(value.get_type())
+    }
+
+    /// Creates an [`RangeValue`] from a [`PointerValue`].
+    #[must_use]
+    pub fn from_pointer_value(ptr: PointerValue<'ctx>, name: Option<&'ctx str>) -> Self {
+        debug_assert!(Self::is_representable(ptr).is_ok());
+
+        RangeValue { value: ptr, name }
+    }
+
+    fn ptr_to_start(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        let llvm_i32 = ctx.ctx.i32_type();
+        let var_name = self.name.map(|v| format!("{v}.start.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.as_base_value(),
+                    &[llvm_i32.const_zero(), llvm_i32.const_int(0, false)],
+                    var_name.as_str(),
+                )
+                .unwrap()
+        }
+    }
+
+    fn ptr_to_end(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        let llvm_i32 = ctx.ctx.i32_type();
+        let var_name = self.name.map(|v| format!("{v}.end.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.as_base_value(),
+                    &[llvm_i32.const_zero(), llvm_i32.const_int(1, false)],
+                    var_name.as_str(),
+                )
+                .unwrap()
+        }
+    }
+
+    fn ptr_to_step(&self, ctx: &CodeGenContext<'ctx, '_>) -> PointerValue<'ctx> {
+        let llvm_i32 = ctx.ctx.i32_type();
+        let var_name = self.name.map(|v| format!("{v}.step.addr")).unwrap_or_default();
+
+        unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.as_base_value(),
+                    &[llvm_i32.const_zero(), llvm_i32.const_int(2, false)],
+                    var_name.as_str(),
+                )
+                .unwrap()
+        }
+    }
+
+    /// Stores the `start` value into this instance.
+    pub fn store_start(&self, ctx: &CodeGenContext<'ctx, '_>, start: IntValue<'ctx>) {
+        debug_assert_eq!(start.get_type().get_bit_width(), 32);
+
+        let pstart = self.ptr_to_start(ctx);
+        ctx.builder.build_store(pstart, start).unwrap();
+    }
+
+    /// Returns the `start` value of this `range`.
+    pub fn load_start(&self, ctx: &CodeGenContext<'ctx, '_>, name: Option<&str>) -> IntValue<'ctx> {
+        let pstart = self.ptr_to_start(ctx);
+        let var_name = name
+            .map(ToString::to_string)
+            .or_else(|| self.name.map(|v| format!("{v}.start")))
+            .unwrap_or_default();
+
+        ctx.builder
+            .build_load(pstart, var_name.as_str())
+            .map(BasicValueEnum::into_int_value)
+            .unwrap()
+    }
+
+    /// Stores the `end` value into this instance.
+    pub fn store_end(&self, ctx: &CodeGenContext<'ctx, '_>, end: IntValue<'ctx>) {
+        debug_assert_eq!(end.get_type().get_bit_width(), 32);
+
+        let pend = self.ptr_to_end(ctx);
+        ctx.builder.build_store(pend, end).unwrap();
+    }
+
+    /// Returns the `end` value of this `range`.
+    pub fn load_end(&self, ctx: &CodeGenContext<'ctx, '_>, name: Option<&str>) -> IntValue<'ctx> {
+        let pend = self.ptr_to_end(ctx);
+        let var_name = name
+            .map(ToString::to_string)
+            .or_else(|| self.name.map(|v| format!("{v}.end")))
+            .unwrap_or_default();
+
+        ctx.builder.build_load(pend, var_name.as_str()).map(BasicValueEnum::into_int_value).unwrap()
+    }
+
+    /// Stores the `step` value into this instance.
+    pub fn store_step(&self, ctx: &CodeGenContext<'ctx, '_>, step: IntValue<'ctx>) {
+        debug_assert_eq!(step.get_type().get_bit_width(), 32);
+
+        let pstep = self.ptr_to_step(ctx);
+        ctx.builder.build_store(pstep, step).unwrap();
+    }
+
+    /// Returns the `step` value of this `range`.
+    pub fn load_step(&self, ctx: &CodeGenContext<'ctx, '_>, name: Option<&str>) -> IntValue<'ctx> {
+        let pstep = self.ptr_to_step(ctx);
+        let var_name = name
+            .map(ToString::to_string)
+            .or_else(|| self.name.map(|v| format!("{v}.step")))
+            .unwrap_or_default();
+
+        ctx.builder
+            .build_load(pstep, var_name.as_str())
+            .map(BasicValueEnum::into_int_value)
+            .unwrap()
+    }
+}
+
+impl<'ctx> ProxyValue<'ctx> for RangeValue<'ctx> {
+    type Base = PointerValue<'ctx>;
+    type Type = RangeType<'ctx>;
+
+    fn get_type(&self) -> Self::Type {
+        RangeType::from_type(self.value.get_type())
+    }
+
+    fn as_base_value(&self) -> Self::Base {
+        self.value
+    }
+}
+
+impl<'ctx> From<RangeValue<'ctx>> for PointerValue<'ctx> {
+    fn from(value: RangeValue<'ctx>) -> Self {
+        value.as_base_value()
+    }
+}

nac3core/src/lib.rs

@@ -1,10 +1,4 @@
-#![deny(
-    future_incompatible,
-    let_underscore,
-    nonstandard_style,
-    rust_2024_compatibility,
-    clippy::all
-)]
+#![deny(future_incompatible, let_underscore, nonstandard_style, clippy::all)]
 #![warn(clippy::pedantic)]
 #![allow(
     dead_code,
@@ -19,7 +13,13 @@
     clippy::wildcard_imports
 )]
 
+// users of nac3core need to use the same version of these dependencies, so expose them as nac3core::*
+pub use inkwell;
+pub use nac3parser;
+
 pub mod codegen;
 pub mod symbol_resolver;
 pub mod toplevel;
 pub mod typecheck;
+
+extern crate self as nac3core;

nac3core/src/symbol_resolver.rs

@@ -1,7 +1,15 @@
-use std::fmt::Debug;
-use std::rc::Rc;
-use std::sync::Arc;
-use std::{collections::HashMap, collections::HashSet, fmt::Display};
+use std::{
+    collections::{HashMap, HashSet},
+    fmt::{Debug, Display},
+    rc::Rc,
+    sync::Arc,
+};
+
+use inkwell::values::{BasicValueEnum, FloatValue, IntValue, PointerValue, StructValue};
+use itertools::{chain, izip, Itertools};
+use parking_lot::RwLock;
+
+use nac3parser::ast::{Constant, Expr, Location, StrRef};
 
 use crate::{
     codegen::{CodeGenContext, CodeGenerator},
@@ -11,10 +19,6 @@ use crate::{
         typedef::{Type, TypeEnum, Unifier, VarMap},
     },
 };
-use inkwell::values::{BasicValueEnum, FloatValue, IntValue, PointerValue, StructValue};
-use itertools::{chain, izip, Itertools};
-use nac3parser::ast::{Constant, Expr, Location, StrRef};
-use parking_lot::RwLock;
 
 #[derive(Clone, PartialEq, Debug)]
 pub enum SymbolValue {
@@ -78,14 +82,14 @@ impl SymbolValue {
             }
             Constant::Tuple(t) => {
                 let expected_ty = unifier.get_ty(expected_ty);
-                let TypeEnum::TTuple { ty } = expected_ty.as_ref() else {
+                let TypeEnum::TTuple { ty, is_vararg_ctx } = expected_ty.as_ref() else {
                     return Err(format!(
                         "Expected {:?}, but got Tuple",
                         expected_ty.get_type_name()
                     ));
                 };
 
-                assert_eq!(ty.len(), t.len());
+                assert!(*is_vararg_ctx || ty.len() == t.len());
 
                 let elems = t
                     .iter()
@@ -155,7 +159,7 @@ impl SymbolValue {
             SymbolValue::Bool(_) => primitives.bool,
             SymbolValue::Tuple(vs) => {
                 let vs_tys = vs.iter().map(|v| v.get_type(primitives, unifier)).collect::<Vec<_>>();
-                unifier.add_ty(TypeEnum::TTuple { ty: vs_tys })
+                unifier.add_ty(TypeEnum::TTuple { ty: vs_tys, is_vararg_ctx: false })
             }
             SymbolValue::OptionSome(_) | SymbolValue::OptionNone => primitives.option,
         }
@@ -365,6 +369,7 @@ pub trait SymbolResolver {
         &self,
         str: StrRef,
         ctx: &mut CodeGenContext<'ctx, '_>,
+        generator: &mut dyn CodeGenerator,
     ) -> Option<ValueEnum<'ctx>>;
 
     fn get_default_param_value(&self, expr: &Expr) -> Option<SymbolValue>;
@@ -382,13 +387,12 @@ pub trait SymbolResolver {
 }
 
 thread_local! {
-    static IDENTIFIER_ID: [StrRef; 12] = [
+    static IDENTIFIER_ID: [StrRef; 11] = [
         "int32".into(),
         "int64".into(),
         "float".into(),
         "bool".into(),
         "virtual".into(),
-        "list".into(),
         "tuple".into(),
         "str".into(),
         "Exception".into(),
@@ -413,13 +417,12 @@ pub fn parse_type_annotation<T>(
     let float_id = ids[2];
     let bool_id = ids[3];
     let virtual_id = ids[4];
-    let list_id = ids[5];
-    let tuple_id = ids[6];
-    let str_id = ids[7];
-    let exn_id = ids[8];
-    let uint32_id = ids[9];
-    let uint64_id = ids[10];
-    let literal_id = ids[11];
+    let tuple_id = ids[5];
+    let str_id = ids[6];
+    let exn_id = ids[7];
+    let uint32_id = ids[8];
+    let uint64_id = ids[9];
+    let literal_id = ids[10];
 
     let name_handling = |id: &StrRef, loc: Location, unifier: &mut Unifier| {
         if *id == int32_id {
@@ -476,9 +479,6 @@ pub fn parse_type_annotation<T>(
         if *id == virtual_id {
             let ty = parse_type_annotation(resolver, top_level_defs, unifier, primitives, slice)?;
             Ok(unifier.add_ty(TypeEnum::TVirtual { ty }))
-        } else if *id == list_id {
-            let ty = parse_type_annotation(resolver, top_level_defs, unifier, primitives, slice)?;
-            Ok(unifier.add_ty(TypeEnum::TList { ty }))
         } else if *id == tuple_id {
             if let Tuple { elts, .. } = &slice.node {
                 let ty = elts
@@ -487,7 +487,7 @@ pub fn parse_type_annotation<T>(
                         parse_type_annotation(resolver, top_level_defs, unifier, primitives, elt)
                     })
                     .collect::<Result<Vec<_>, _>>()?;
-                Ok(unifier.add_ty(TypeEnum::TTuple { ty }))
+                Ok(unifier.add_ty(TypeEnum::TTuple { ty, is_vararg_ctx: false }))
             } else {
                 Err(HashSet::from(["Expected multiple elements for tuple".into()]))
             }

nac3core/src/toplevel/composer.rs

@@ -1,17 +1,17 @@
-use nac3parser::ast::fold::Fold;
 use std::rc::Rc;
 
+use nac3parser::ast::{fold::Fold, ExprKind, Ident};
+
+use super::*;
 use crate::{
     codegen::{expr::get_subst_key, stmt::exn_constructor},
     symbol_resolver::SymbolValue,
     typecheck::{
-        type_inferencer::{FunctionData, Inferencer},
+        type_inferencer::{FunctionData, IdentifierInfo, Inferencer},
         typedef::{TypeVar, VarMap},
     },
 };
 
-use super::*;
-
 pub struct ComposerConfig {
     pub kernel_ann: Option<&'static str>,
     pub kernel_invariant_ann: &'static str,
@@ -23,7 +23,7 @@ impl Default for ComposerConfig {
     }
 }
 
-type DefAst = (Arc<RwLock<TopLevelDef>>, Option<Stmt<()>>);
+pub type DefAst = (Arc<RwLock<TopLevelDef>>, Option<Stmt<()>>);
 pub struct TopLevelComposer {
     // list of top level definitions, same as top level context
     pub definition_ast_list: Vec<DefAst>,
@@ -44,12 +44,27 @@ pub struct TopLevelComposer {
     pub size_t: u32,
 }
 
+/// The specification for a builtin function, consisting of the function name, the function
+/// signature, and a [code generation callback][`GenCall`].
+pub type BuiltinFuncSpec = (StrRef, FunSignature, Arc<GenCall>);
+
+/// A function that creates a [`BuiltinFuncSpec`] using the provided [`PrimitiveStore`] and
+/// [`Unifier`].
+pub type BuiltinFuncCreator = dyn Fn(&PrimitiveStore, &mut Unifier) -> BuiltinFuncSpec;
+
 impl TopLevelComposer {
     /// return a composer and things to make a "primitive" symbol resolver, so that the symbol
-    /// resolver can later figure out primitive type definitions when passed a primitive type name
+    /// resolver can later figure out primitive tye definitions when passed a primitive type name
+    ///
+    /// `lateinit_builtins` are specifically for the ARTIQ module. Since the [`Unifier`] instance
+    /// used to create builtin functions do not persist until method compilation, any types
+    /// created (e.g. [`TypeEnum::TVar`]) also do not persist. Those functions should be instead put
+    /// in `lateinit_builtins`, where they will be instantiated with the [`Unifier`] instance used
+    /// for method compilation.
     #[must_use]
     pub fn new(
-        builtins: Vec<(StrRef, FunSignature, Arc<GenCall>)>,
+        builtins: Vec<BuiltinFuncSpec>,
+        lateinit_builtins: Vec<Box<BuiltinFuncCreator>>,
         core_config: ComposerConfig,
         size_t: u32,
     ) -> (Self, HashMap<StrRef, DefinitionId>, HashMap<StrRef, Type>) {
@@ -86,7 +101,8 @@ impl TopLevelComposer {
             .iter()
             .map(|def_ast| match *def_ast.0.read() {
                 TopLevelDef::Class { name, .. } => name.to_string(),
-                TopLevelDef::Function { simple_name, .. } => simple_name.to_string(),
+                TopLevelDef::Function { simple_name, .. }
+                | TopLevelDef::Variable { simple_name, .. } => simple_name.to_string(),
             })
             .collect_vec();
 
@@ -119,7 +135,13 @@ impl TopLevelComposer {
             }
         }
 
-        for (name, sig, codegen_callback) in builtins {
+        // Materialize lateinit_builtins, now that the unifier is ready
+        let lateinit_builtins = lateinit_builtins
+            .into_iter()
+            .map(|builtin| builtin(&primitives_ty, &mut unifier))
+            .collect_vec();
+
+        for (name, sig, codegen_callback) in builtins.into_iter().chain(lateinit_builtins) {
             let fun_sig = unifier.add_ty(TypeEnum::TFunc(sig));
             builtin_ty.insert(name, fun_sig);
             builtin_id.insert(name, DefinitionId(definition_ast_list.len()));
@@ -360,13 +382,87 @@ impl TopLevelComposer {
                 ))
             }
 
+            ast::StmtKind::Assign { .. } => {
+                // Assignment statements can assign to (and therefore create) more than one 
+                // variable, but this function only allows returning one set of symbol information.
+                // We want to avoid changing this to return a `Vec` of symbol info, as this would 
+                // require `iter().next().unwrap()` on every variable created from a non-Assign 
+                // statement.
+                //
+                // Make callers use `register_top_level_var` instead, as it provides more
+                // fine-grained control over which symbols to register, while also simplifying the
+                // usage of this function.
+                panic!("Registration of top-level Assign statements must use TopLevelComposer::register_top_level_var (at {})", ast.location);
+            }
+
+            ast::StmtKind::AnnAssign { target, annotation, .. } => {
+                let ExprKind::Name { id: name, .. } = target.node else {
+                    return Err(format!(
+                        "global variable declaration must be an identifier (at {})",
+                        target.location
+                    ));
+                };
+
+                self.register_top_level_var(
+                    name,
+                    Some(annotation.as_ref().clone()),
+                    resolver,
+                    mod_path,
+                    target.location,
+                )
+            }
+
             _ => Err(format!(
-                "registrations of constructs other than top level classes/functions are not supported (at {})",
+                "registrations of constructs other than top level classes/functions/variables are not supported (at {})",
                 ast.location
             )),
         }
     }
 
+    /// Registers a top-level variable with the given `name` into the composer.
+    ///
+    /// `annotation` - The type annotation of the top-level variable, or [`None`] if no type
+    /// annotation is provided.
+    /// `location` - The location of the top-level variable.
+    pub fn register_top_level_var(
+        &mut self,
+        name: Ident,
+        annotation: Option<Expr>,
+        resolver: Option<Arc<dyn SymbolResolver + Send + Sync>>,
+        mod_path: &str,
+        location: Location,
+    ) -> Result<(StrRef, DefinitionId, Option<Type>), String> {
+        if self.keyword_list.contains(&name) {
+            return Err(format!("cannot use keyword `{name}` as a class name (at {location})"));
+        }
+
+        let global_var_name =
+            if mod_path.is_empty() { name.to_string() } else { format!("{mod_path}.{name}") };
+
+        if !self.defined_names.insert(global_var_name.clone()) {
+            return Err(format!(
+                "global variable `{global_var_name}` defined twice (at {location})"
+            ));
+        }
+
+        let ty_to_be_unified = self.unifier.get_dummy_var().ty;
+        self.definition_ast_list.push((
+            RwLock::new(Self::make_top_level_variable_def(
+                global_var_name,
+                name,
+                // dummy here, unify with correct type later,
+                ty_to_be_unified,
+                annotation,
+                resolver,
+                Some(location),
+            ))
+            .into(),
+            None,
+        ));
+
+        Ok((name, DefinitionId(self.definition_ast_list.len() - 1), Some(ty_to_be_unified)))
+    }
+
     pub fn start_analysis(&mut self, inference: bool) -> Result<(), HashSet<String>> {
         self.analyze_top_level_class_type_var()?;
         self.analyze_top_level_class_bases()?;
@@ -375,6 +471,7 @@ impl TopLevelComposer {
         if inference {
             self.analyze_function_instance()?;
         }
+        self.analyze_top_level_variables()?;
         Ok(())
     }
 
@@ -412,7 +509,7 @@ impl TopLevelComposer {
                     // things like `class A(Generic[T, V, ImportedModule.T])` is not supported
                     // i.e. only simple names are allowed in the subscript
                     // should update the TopLevelDef::Class.typevars and the TypeEnum::TObj.params
-                    ast::ExprKind::Subscript { value, slice, .. }
+                    ExprKind::Subscript { value, slice, .. }
                         if {
                             matches!(
                                 &value.node,
@@ -428,9 +525,9 @@ impl TopLevelComposer {
                         }
                         is_generic = true;
 
-                        let type_var_list: Vec<&ast::Expr<()>>;
+                        let type_var_list: Vec<&Expr<()>>;
                         // if `class A(Generic[T, V, G])`
-                        if let ast::ExprKind::Tuple { elts, .. } = &slice.node {
+                        if let ExprKind::Tuple { elts, .. } = &slice.node {
                             type_var_list = elts.iter().collect_vec();
                         // `class A(Generic[T])`
                         } else {
@@ -479,6 +576,7 @@ impl TopLevelComposer {
             }
             Ok(())
         };
+
         let mut errors = HashSet::new();
         for (class_def, class_ast) in def_list.iter().skip(self.builtin_num) {
             if class_ast.is_none() {
@@ -766,6 +864,7 @@ impl TopLevelComposer {
             let target_ty = get_type_from_type_annotation_kinds(
                 &temp_def_list,
                 unifier,
+                primitives,
                 &def,
                 &mut subst_list,
             )?;
@@ -831,7 +930,6 @@ impl TopLevelComposer {
         let unifier = self.unifier.borrow_mut();
         let primitives_store = &self.primitives_ty;
 
-        let mut errors = HashSet::new();
         let mut analyze = |function_def: &Arc<RwLock<TopLevelDef>>, function_ast: &Option<Stmt>| {
             let mut function_def = function_def.write();
             let function_def = &mut *function_def;
@@ -859,7 +957,73 @@ impl TopLevelComposer {
             let resolver = &**resolver;
 
             let mut function_var_map = VarMap::new();
-            let arg_types = {
+
+            let vararg = args
+                .vararg
+                .as_ref()
+                .map(|vararg| -> Result<_, HashSet<String>> {
+                    let vararg = vararg.as_ref();
+
+                    let annotation = vararg
+                        .node
+                        .annotation
+                        .as_ref()
+                        .ok_or_else(|| {
+                            HashSet::from([format!(
+                                "function parameter `{}` needs type annotation at {}",
+                                vararg.node.arg, vararg.location
+                            )])
+                        })?
+                        .as_ref();
+
+                    let type_annotation = parse_ast_to_type_annotation_kinds(
+                        resolver,
+                        temp_def_list.as_slice(),
+                        unifier,
+                        primitives_store,
+                        annotation,
+                        // NOTE: since only class need this, for function
+                        // it should be fine to be empty map
+                        HashMap::new(),
+                    )?;
+
+                    let type_vars_within =
+                        get_type_var_contained_in_type_annotation(&type_annotation)
+                            .into_iter()
+                            .map(|x| -> Result<TypeVar, HashSet<String>> {
+                                let TypeAnnotation::TypeVar(ty) = x else {
+                                    unreachable!("must be type var annotation kind")
+                                };
+
+                                let id = Self::get_var_id(ty, unifier)?;
+                                Ok(TypeVar { id, ty })
+                            })
+                            .collect::<Result<Vec<_>, _>>()?;
+                    for var in type_vars_within {
+                        if let Some(prev_ty) = function_var_map.insert(var.id, var.ty) {
+                            // if already have the type inserted, make sure they are the same thing
+                            assert_eq!(prev_ty, var.ty);
+                        }
+                    }
+
+                    let ty = get_type_from_type_annotation_kinds(
+                        temp_def_list.as_ref(),
+                        unifier,
+                        primitives_store,
+                        &type_annotation,
+                        &mut None,
+                    )?;
+
+                    Ok(FuncArg {
+                        name: vararg.node.arg,
+                        ty,
+                        default_value: Some(SymbolValue::Tuple(Vec::default())),
+                        is_vararg: true,
+                    })
+                })
+                .transpose()?;
+
+            let mut arg_types = {
                 // make sure no duplicate parameter
                 let mut defined_parameter_name: HashSet<_> = HashSet::new();
                 for x in &args.args {
@@ -874,15 +1038,15 @@ impl TopLevelComposer {
                     }
                 }
 
-                let arg_with_default: Vec<(&ast::Located<ast::ArgData<()>>, Option<&ast::Expr>)> =
-                    args.args
+                let arg_with_default: Vec<(&ast::Located<ast::ArgData<()>>, Option<&Expr>)> = args
+                    .args
                     .iter()
                     .rev()
                     .zip(
                         args.defaults
                             .iter()
                             .rev()
-                                .map(|x| -> Option<&ast::Expr> { Some(x) })
+                            .map(|x| -> Option<&Expr> { Some(x) })
                             .chain(std::iter::repeat(None)),
                     )
                     .collect_vec();
@@ -936,6 +1100,7 @@ impl TopLevelComposer {
                         let ty = get_type_from_type_annotation_kinds(
                             temp_def_list.as_ref(),
                             unifier,
+                            primitives_store,
                             &type_annotation,
                             &mut None,
                         )?;
@@ -959,11 +1124,18 @@ impl TopLevelComposer {
                                     v
                                 }),
                             },
+                            is_vararg: false,
                         })
                     })
                     .collect::<Result<Vec<_>, _>>()?
             };
 
+            if let Some(vararg) = vararg {
+                arg_types.push(vararg);
+            };
+
+            let arg_types = arg_types;
+
             let return_ty = {
                 if let Some(returns) = returns {
                     let return_ty_annotation = {
@@ -1002,6 +1174,7 @@ impl TopLevelComposer {
                     get_type_from_type_annotation_kinds(
                         &temp_def_list,
                         unifier,
+                        primitives_store,
                         &return_ty_annotation,
                         &mut None,
                     )?
@@ -1031,6 +1204,8 @@ impl TopLevelComposer {
             })?;
             Ok(())
         };
+
+        let mut errors = HashSet::new();
         for (function_def, function_ast) in def_list.iter().skip(self.builtin_num) {
             if function_ast.is_none() {
                 continue;
@@ -1132,7 +1307,7 @@ impl TopLevelComposer {
 
                         let arg_with_default: Vec<(
                             &ast::Located<ast::ArgData<()>>,
-                            Option<&ast::Expr>,
+                            Option<&Expr>,
                         )> = args
                             .args
                             .iter()
@@ -1141,7 +1316,7 @@ impl TopLevelComposer {
                                 args.defaults
                                     .iter()
                                     .rev()
-                                    .map(|x| -> Option<&ast::Expr> { Some(x) })
+                                    .map(|x| -> Option<&Expr> { Some(x) })
                                     .chain(std::iter::repeat(None)),
                             )
                             .collect_vec();
@@ -1214,6 +1389,7 @@ impl TopLevelComposer {
                                             })
                                         }
                                     },
+                                    is_vararg: false,
                                 };
                                 // push the dummy type and the type annotation
                                 // into the list for later unification
@@ -1297,7 +1473,7 @@ impl TopLevelComposer {
                         .map_err(|e| HashSet::from([e.to_display(unifier).to_string()]))?;
                 }
                 ast::StmtKind::AnnAssign { target, annotation, value, .. } => {
-                    if let ast::ExprKind::Name { id: attr, .. } = &target.node {
+                    if let ExprKind::Name { id: attr, .. } = &target.node {
                         if defined_fields.insert(attr.to_string()) {
                             let dummy_field_type = unifier.get_dummy_var().ty;
 
@@ -1305,7 +1481,7 @@ impl TopLevelComposer {
                                 None => {
                                     // handle Kernel[T], KernelInvariant[T]
                                     let (annotation, mutable) = match &annotation.node {
-                                        ast::ExprKind::Subscript { value, slice, .. }
+                                        ExprKind::Subscript { value, slice, .. }
                                             if matches!(
                                                 &value.node,
                                                 ast::ExprKind::Name { id, .. } if id == &core_config.kernel_invariant_ann.into()
@@ -1313,7 +1489,7 @@ impl TopLevelComposer {
                                         {
                                             (slice, false)
                                         }
-                                        ast::ExprKind::Subscript { value, slice, .. }
+                                        ExprKind::Subscript { value, slice, .. }
                                             if matches!(
                                                 &value.node,
                                                 ast::ExprKind::Name { id, .. } if core_config.kernel_ann.map_or(false, |c| id == &c.into())
@@ -1331,13 +1507,13 @@ impl TopLevelComposer {
                                 Some(boxed_expr) => {
                                     // Class attributes are set as immutable regardless
                                     let (annotation, _) = match &annotation.node {
-                                        ast::ExprKind::Subscript { slice, .. } => (slice, false),
+                                        ExprKind::Subscript { slice, .. } => (slice, false),
                                         _ if core_config.kernel_ann.is_none() => (annotation, false),
                                         _ => continue,
                                     };
 
                                     match &**boxed_expr {
-                                        ast::Located {location: _, custom: (), node: ast::ExprKind::Constant { value: v, kind: _ }} => {
+                                        ast::Located {location: _, custom: (), node: ExprKind::Constant { value: v, kind: _ }} => {
                                             // Restricting the types allowed to be defined as class attributes
                                             match v {
                                                 ast::Constant::Bool(_) | ast::Constant::Str(_) | ast::Constant::Int(_) | ast::Constant::Float(_) => {}
@@ -1604,7 +1780,6 @@ impl TopLevelComposer {
             }
         }
 
-        let mut errors = HashSet::new();
         let mut analyze = |i, def: &Arc<RwLock<TopLevelDef>>, ast: &Option<Stmt>| {
             let class_def = def.read();
             if let TopLevelDef::Class {
@@ -1622,6 +1797,7 @@ impl TopLevelComposer {
                 let self_type = get_type_from_type_annotation_kinds(
                     &def_list,
                     unifier,
+                    primitives_ty,
                     &make_self_type_annotation(type_vars, *object_id),
                     &mut None,
                 )?;
@@ -1638,21 +1814,25 @@ impl TopLevelComposer {
                                 name: "msg".into(),
                                 ty: string,
                                 default_value: Some(SymbolValue::Str(String::new())),
+                                is_vararg: false,
                             },
                             FuncArg {
                                 name: "param0".into(),
                                 ty: int64,
                                 default_value: Some(SymbolValue::I64(0)),
+                                is_vararg: false,
                             },
                             FuncArg {
                                 name: "param1".into(),
                                 ty: int64,
                                 default_value: Some(SymbolValue::I64(0)),
+                                is_vararg: false,
                             },
                             FuncArg {
                                 name: "param2".into(),
                                 ty: int64,
                                 default_value: Some(SymbolValue::I64(0)),
+                                is_vararg: false,
                             },
                         ],
                         ret: self_type,
@@ -1719,7 +1899,12 @@ impl TopLevelComposer {
                         if *name != init_str_id {
                             unreachable!("must be init function here")
                         }
-                        let all_inited = Self::get_all_assigned_field(body.as_slice())?;
+
+                        let all_inited = Self::get_all_assigned_field(
+                            object_id.0,
+                            definition_ast_list,
+                            body.as_slice(),
+                        )?;
                         for (f, _, _) in fields {
                             if !all_inited.contains(f) {
                                 return Err(HashSet::from([
@@ -1737,6 +1922,8 @@ impl TopLevelComposer {
             }
             Ok(())
         };
+
+        let mut errors = HashSet::new();
         for (i, (def, ast)) in definition_ast_list.iter().enumerate().skip(self.builtin_num) {
             if ast.is_none() {
                 continue;
@@ -1774,19 +1961,20 @@ impl TopLevelComposer {
             if ast.is_none() {
                 return Ok(());
             }
-            let mut function_def = def.write();
-            if let TopLevelDef::Function {
-                instance_to_stmt,
-                instance_to_symbol,
-                name,
-                simple_name,
-                signature,
-                resolver,
-                ..
-            } = &mut *function_def
-            {
-                let signature_ty_enum = unifier.get_ty(*signature);
-                let TypeEnum::TFunc(FunSignature { args, ret, vars }) = signature_ty_enum.as_ref()
+
+            let (name, simple_name, signature, resolver) = {
+                let function_def = def.read();
+                let TopLevelDef::Function { name, simple_name, signature, resolver, .. } =
+                    &*function_def
+                else {
+                    return Ok(());
+                };
+
+                (name.clone(), *simple_name, *signature, resolver.clone())
+            };
+
+            let signature_ty_enum = unifier.get_ty(signature);
+            let TypeEnum::TFunc(FunSignature { args, ret, vars, .. }) = signature_ty_enum.as_ref()
             else {
                 unreachable!("must be typeenum::tfunc")
             };
@@ -1803,7 +1991,11 @@ impl TopLevelComposer {
 
                     let ty_ann = make_self_type_annotation(type_vars, *class_id);
                     let self_ty = get_type_from_type_annotation_kinds(
-                            &def_list, unifier, &ty_ann, &mut None,
+                        &def_list,
+                        unifier,
+                        primitives_ty,
+                        &ty_ann,
+                        &mut None,
                     )?;
                     vars.extend(type_vars.iter().map(|ty| {
                         let TypeEnum::TVar { id, .. } = &*unifier.get_ty(*ty) else {
@@ -1858,6 +2050,7 @@ impl TopLevelComposer {
                             name: a.name,
                             ty: unifier.subst(a.ty, &subst).unwrap_or(a.ty),
                             default_value: a.default_value.clone(),
+                            is_vararg: false,
                         })
                         .collect_vec()
                 };
@@ -1889,11 +2082,11 @@ impl TopLevelComposer {
                     })
                 };
                 let mut identifiers = {
-                        let mut result: HashSet<_> = HashSet::new();
+                    let mut result = HashMap::new();
                     if self_type.is_some() {
-                            result.insert("self".into());
+                        result.insert("self".into(), IdentifierInfo::default());
                     }
-                        result.extend(inst_args.iter().map(|x| x.name));
+                    result.extend(inst_args.iter().map(|x| (x.name, IdentifierInfo::default())));
                     result
                 };
                 let mut calls: HashMap<CodeLocation, CallId> = HashMap::new();
@@ -1930,23 +2123,43 @@ impl TopLevelComposer {
                 else {
                     unreachable!("must be function def ast")
                 };
-                    if !decorator_list.is_empty()
-                        && matches!(&decorator_list[0].node,
-                            ast::ExprKind::Name{ id, .. } if id == &"extern".into())
-                    {
-                        instance_to_symbol.insert(String::new(), simple_name.to_string());
-                        continue;
-                    }
-                    if !decorator_list.is_empty()
-                        && matches!(&decorator_list[0].node,
-                            ast::ExprKind::Name{ id, .. } if id == &"rpc".into())
+
+                if !decorator_list.is_empty() {
+                    if matches!(&decorator_list[0].node, ExprKind::Name { id, .. } if id == &"extern".into())
                     {
+                        let TopLevelDef::Function { instance_to_symbol, .. } = &mut *def.write()
+                        else {
+                            unreachable!()
+                        };
                         instance_to_symbol.insert(String::new(), simple_name.to_string());
                         continue;
                     }
 
-                    let fun_body = body
-                        .into_iter()
+                    if matches!(&decorator_list[0].node, ExprKind::Name { id, .. } if id == &"rpc".into())
+                    {
+                        let TopLevelDef::Function { instance_to_symbol, .. } = &mut *def.write()
+                        else {
+                            unreachable!()
+                        };
+                        instance_to_symbol.insert(String::new(), simple_name.to_string());
+                        continue;
+                    }
+
+                    if let ExprKind::Call { func, .. } = &decorator_list[0].node {
+                        if matches!(&func.node, ExprKind::Name { id, .. } if id == &"rpc".into()) {
+                            let TopLevelDef::Function { instance_to_symbol, .. } =
+                                &mut *def.write()
+                            else {
+                                unreachable!()
+                            };
+                            instance_to_symbol.insert(String::new(), simple_name.to_string());
+                            continue;
+                        }
+                    }
+                }
+
+                let fun_body =
+                    body.into_iter()
                         .map(|b| inferencer.fold_stmt(b))
                         .collect::<Result<Vec<_>, _>>()?;
 
@@ -2016,6 +2229,9 @@ impl TopLevelComposer {
                     )]));
                 }
 
+                let TopLevelDef::Function { instance_to_stmt, .. } = &mut *def.write() else {
+                    unreachable!()
+                };
                 instance_to_stmt.insert(
                     get_subst_key(
                         unifier,
@@ -2031,10 +2247,10 @@ impl TopLevelComposer {
                     },
                 );
             }
-            }
 
             Ok(())
         };
+
         for (id, (def, ast)) in self.definition_ast_list.iter().enumerate().skip(self.builtin_num) {
             if ast.is_none() {
                 continue;
@@ -2048,4 +2264,59 @@ impl TopLevelComposer {
         }
         Ok(())
     }
+
+    /// Step 6. Analyze and populate the types of global variables.
+    fn analyze_top_level_variables(&mut self) -> Result<(), HashSet<String>> {
+        let def_list = &self.definition_ast_list;
+        let temp_def_list = self.extract_def_list();
+        let unifier = &mut self.unifier;
+        let primitives_store = &self.primitives_ty;
+
+        let mut analyze = |variable_def: &Arc<RwLock<TopLevelDef>>| -> Result<_, HashSet<String>> {
+            let TopLevelDef::Variable { ty: dummy_ty, ty_decl, resolver, loc, .. } =
+                &*variable_def.read()
+            else {
+                // not top level variable def, skip
+                return Ok(());
+            };
+
+            let resolver = &**resolver.as_ref().unwrap();
+
+            if let Some(ty_decl) = ty_decl {
+                let ty_annotation = parse_ast_to_type_annotation_kinds(
+                    resolver,
+                    &temp_def_list,
+                    unifier,
+                    primitives_store,
+                    ty_decl,
+                    HashMap::new(),
+                )?;
+                let ty_from_ty_annotation = get_type_from_type_annotation_kinds(
+                    &temp_def_list,
+                    unifier,
+                    primitives_store,
+                    &ty_annotation,
+                    &mut None,
+                )?;
+
+                unifier.unify(*dummy_ty, ty_from_ty_annotation).map_err(|e| {
+                    HashSet::from([e.at(Some(loc.unwrap())).to_display(unifier).to_string()])
+                })?;
+            }
+
+            Ok(())
+        };
+
+        let mut errors = HashSet::new();
+        for (variable_def, _) in def_list.iter().skip(self.builtin_num) {
+            if let Err(e) = analyze(variable_def) {
+                errors.extend(e);
+            }
+        }
+        if !errors.is_empty() {
+            return Err(errors);
+        }
+
+        Ok(())
+    }
 }

nac3core/src/toplevel/helper.rs

@@ -1,17 +1,20 @@
 use std::convert::TryInto;
 
-use crate::symbol_resolver::SymbolValue;
-use crate::toplevel::numpy::unpack_ndarray_var_tys;
-use crate::typecheck::typedef::{into_var_map, Mapping, TypeVarId, VarMap};
-use nac3parser::ast::{Constant, Location};
 use strum::IntoEnumIterator;
 use strum_macros::EnumIter;
 
-use super::*;
+use nac3parser::ast::{Constant, ExprKind, Location};
+
+use super::{numpy::unpack_ndarray_var_tys, *};
+use crate::{
+    symbol_resolver::SymbolValue,
+    typecheck::typedef::{into_var_map, iter_type_vars, Mapping, TypeVarId, VarMap},
+};
 
 /// All primitive types and functions in nac3core.
 #[derive(Clone, Copy, Debug, EnumIter, PartialEq, Eq)]
 pub enum PrimDef {
+    // Classes
     Int32,
     Int64,
     Float,
@@ -23,17 +26,25 @@ pub enum PrimDef {
     UInt32,
     UInt64,
     Option,
-    OptionIsSome,
-    OptionIsNone,
-    OptionUnwrap,
+    List,
     NDArray,
-    NDArrayCopy,
-    NDArrayFill,
-    FunInt32,
-    FunInt64,
-    FunUInt32,
-    FunUInt64,
-    FunFloat,
+
+    // Option methods
+    FunOptionIsSome,
+    FunOptionIsNone,
+    FunOptionUnwrap,
+
+    // Option-related functions
+    FunSome,
+
+    // NDArray methods
+    FunNDArrayCopy,
+    FunNDArrayFill,
+
+    // Range methods
+    FunRangeInit,
+
+    // NumPy factory functions
     FunNpNDArray,
     FunNpEmpty,
     FunNpZeros,
@@ -42,26 +53,17 @@ pub enum PrimDef {
     FunNpArray,
     FunNpEye,
     FunNpIdentity,
-    FunRound,
-    FunRound64,
+
+    // Miscellaneous NumPy & SciPy functions
     FunNpRound,
-    FunRange,
-    FunStr,
-    FunBool,
-    FunFloor,
-    FunFloor64,
     FunNpFloor,
-    FunCeil,
-    FunCeil64,
     FunNpCeil,
-    FunLen,
-    FunMin,
     FunNpMin,
     FunNpMinimum,
-    FunMax,
+    FunNpArgmin,
     FunNpMax,
     FunNpMaximum,
-    FunAbs,
+    FunNpArgmax,
     FunNpIsNan,
     FunNpIsInf,
     FunNpSin,
@@ -99,13 +101,46 @@ pub enum PrimDef {
     FunNpLdExp,
     FunNpHypot,
     FunNpNextAfter,
-    FunSome,
+    FunNpTranspose,
+    FunNpReshape,
+
+    // Linalg functions
+    FunNpDot,
+    FunNpLinalgCholesky,
+    FunNpLinalgQr,
+    FunNpLinalgSvd,
+    FunNpLinalgInv,
+    FunNpLinalgPinv,
+    FunNpLinalgMatrixPower,
+    FunNpLinalgDet,
+    FunSpLinalgLu,
+    FunSpLinalgSchur,
+    FunSpLinalgHessenberg,
+
+    // Miscellaneous Python & NAC3 functions
+    FunInt32,
+    FunInt64,
+    FunUInt32,
+    FunUInt64,
+    FunFloat,
+    FunRound,
+    FunRound64,
+    FunStr,
+    FunBool,
+    FunFloor,
+    FunFloor64,
+    FunCeil,
+    FunCeil64,
+    FunLen,
+    FunMin,
+    FunMax,
+    FunAbs,
 }
 
 /// Associated details of a [`PrimDef`]
 pub enum PrimDefDetails {
     PrimFunction { name: &'static str, simple_name: &'static str },
-    PrimClass { name: &'static str },
+    PrimClass { name: &'static str, get_ty_fn: fn(&PrimitiveStore) -> Type },
 }
 
 impl PrimDef {
@@ -147,15 +182,17 @@ impl PrimDef {
     #[must_use]
     pub fn name(&self) -> &'static str {
         match self.details() {
-            PrimDefDetails::PrimFunction { name, .. } | PrimDefDetails::PrimClass { name } => name,
+            PrimDefDetails::PrimFunction { name, .. } | PrimDefDetails::PrimClass { name, .. } => {
+                name
+            }
         }
     }
 
     /// Get the associated details of this [`PrimDef`]
     #[must_use]
     pub fn details(self) -> PrimDefDetails {
-        fn class(name: &'static str) -> PrimDefDetails {
-            PrimDefDetails::PrimClass { name }
+        fn class(name: &'static str, get_ty_fn: fn(&PrimitiveStore) -> Type) -> PrimDefDetails {
+            PrimDefDetails::PrimClass { name, get_ty_fn }
         }
 
         fn fun(name: &'static str, simple_name: Option<&'static str>) -> PrimDefDetails {
@@ -163,28 +200,37 @@ impl PrimDef {
         }
 
         match self {
-            PrimDef::Int32 => class("int32"),
-            PrimDef::Int64 => class("int64"),
-            PrimDef::Float => class("float"),
-            PrimDef::Bool => class("bool"),
-            PrimDef::None => class("none"),
-            PrimDef::Range => class("range"),
-            PrimDef::Str => class("str"),
-            PrimDef::Exception => class("Exception"),
-            PrimDef::UInt32 => class("uint32"),
-            PrimDef::UInt64 => class("uint64"),
-            PrimDef::Option => class("Option"),
-            PrimDef::OptionIsSome => fun("Option.is_some", Some("is_some")),
-            PrimDef::OptionIsNone => fun("Option.is_none", Some("is_none")),
-            PrimDef::OptionUnwrap => fun("Option.unwrap", Some("unwrap")),
-            PrimDef::NDArray => class("ndarray"),
-            PrimDef::NDArrayCopy => fun("ndarray.copy", Some("copy")),
-            PrimDef::NDArrayFill => fun("ndarray.fill", Some("fill")),
-            PrimDef::FunInt32 => fun("int32", None),
-            PrimDef::FunInt64 => fun("int64", None),
-            PrimDef::FunUInt32 => fun("uint32", None),
-            PrimDef::FunUInt64 => fun("uint64", None),
-            PrimDef::FunFloat => fun("float", None),
+            // Classes
+            PrimDef::Int32 => class("int32", |primitives| primitives.int32),
+            PrimDef::Int64 => class("int64", |primitives| primitives.int64),
+            PrimDef::Float => class("float", |primitives| primitives.float),
+            PrimDef::Bool => class("bool", |primitives| primitives.bool),
+            PrimDef::None => class("none", |primitives| primitives.none),
+            PrimDef::Range => class("range", |primitives| primitives.range),
+            PrimDef::Str => class("str", |primitives| primitives.str),
+            PrimDef::Exception => class("Exception", |primitives| primitives.exception),
+            PrimDef::UInt32 => class("uint32", |primitives| primitives.uint32),
+            PrimDef::UInt64 => class("uint64", |primitives| primitives.uint64),
+            PrimDef::Option => class("Option", |primitives| primitives.option),
+            PrimDef::List => class("list", |primitives| primitives.list),
+            PrimDef::NDArray => class("ndarray", |primitives| primitives.ndarray),
+
+            // Option methods
+            PrimDef::FunOptionIsSome => fun("Option.is_some", Some("is_some")),
+            PrimDef::FunOptionIsNone => fun("Option.is_none", Some("is_none")),
+            PrimDef::FunOptionUnwrap => fun("Option.unwrap", Some("unwrap")),
+
+            // Option-related functions
+            PrimDef::FunSome => fun("Some", None),
+
+            // NDArray methods
+            PrimDef::FunNDArrayCopy => fun("ndarray.copy", Some("copy")),
+            PrimDef::FunNDArrayFill => fun("ndarray.fill", Some("fill")),
+
+            // Range methods
+            PrimDef::FunRangeInit => fun("range.__init__", Some("__init__")),
+
+            // NumPy factory functions
             PrimDef::FunNpNDArray => fun("np_ndarray", None),
             PrimDef::FunNpEmpty => fun("np_empty", None),
             PrimDef::FunNpZeros => fun("np_zeros", None),
@@ -193,26 +239,17 @@ impl PrimDef {
             PrimDef::FunNpArray => fun("np_array", None),
             PrimDef::FunNpEye => fun("np_eye", None),
             PrimDef::FunNpIdentity => fun("np_identity", None),
-            PrimDef::FunRound => fun("round", None),
-            PrimDef::FunRound64 => fun("round64", None),
+
+            // Miscellaneous NumPy & SciPy functions
             PrimDef::FunNpRound => fun("np_round", None),
-            PrimDef::FunRange => fun("range", None),
-            PrimDef::FunStr => fun("str", None),
-            PrimDef::FunBool => fun("bool", None),
-            PrimDef::FunFloor => fun("floor", None),
-            PrimDef::FunFloor64 => fun("floor64", None),
             PrimDef::FunNpFloor => fun("np_floor", None),
-            PrimDef::FunCeil => fun("ceil", None),
-            PrimDef::FunCeil64 => fun("ceil64", None),
             PrimDef::FunNpCeil => fun("np_ceil", None),
-            PrimDef::FunLen => fun("len", None),
-            PrimDef::FunMin => fun("min", None),
             PrimDef::FunNpMin => fun("np_min", None),
             PrimDef::FunNpMinimum => fun("np_minimum", None),
-            PrimDef::FunMax => fun("max", None),
+            PrimDef::FunNpArgmin => fun("np_argmin", None),
             PrimDef::FunNpMax => fun("np_max", None),
             PrimDef::FunNpMaximum => fun("np_maximum", None),
-            PrimDef::FunAbs => fun("abs", None),
+            PrimDef::FunNpArgmax => fun("np_argmax", None),
             PrimDef::FunNpIsNan => fun("np_isnan", None),
             PrimDef::FunNpIsInf => fun("np_isinf", None),
             PrimDef::FunNpSin => fun("np_sin", None),
@@ -250,7 +287,40 @@ impl PrimDef {
             PrimDef::FunNpLdExp => fun("np_ldexp", None),
             PrimDef::FunNpHypot => fun("np_hypot", None),
             PrimDef::FunNpNextAfter => fun("np_nextafter", None),
-            PrimDef::FunSome => fun("Some", None),
+            PrimDef::FunNpTranspose => fun("np_transpose", None),
+            PrimDef::FunNpReshape => fun("np_reshape", None),
+
+            // Linalg functions
+            PrimDef::FunNpDot => fun("np_dot", None),
+            PrimDef::FunNpLinalgCholesky => fun("np_linalg_cholesky", None),
+            PrimDef::FunNpLinalgQr => fun("np_linalg_qr", None),
+            PrimDef::FunNpLinalgSvd => fun("np_linalg_svd", None),
+            PrimDef::FunNpLinalgInv => fun("np_linalg_inv", None),
+            PrimDef::FunNpLinalgPinv => fun("np_linalg_pinv", None),
+            PrimDef::FunNpLinalgMatrixPower => fun("np_linalg_matrix_power", None),
+            PrimDef::FunNpLinalgDet => fun("np_linalg_det", None),
+            PrimDef::FunSpLinalgLu => fun("sp_linalg_lu", None),
+            PrimDef::FunSpLinalgSchur => fun("sp_linalg_schur", None),
+            PrimDef::FunSpLinalgHessenberg => fun("sp_linalg_hessenberg", None),
+
+            // Miscellaneous Python & NAC3 functions
+            PrimDef::FunInt32 => fun("int32", None),
+            PrimDef::FunInt64 => fun("int64", None),
+            PrimDef::FunUInt32 => fun("uint32", None),
+            PrimDef::FunUInt64 => fun("uint64", None),
+            PrimDef::FunFloat => fun("float", None),
+            PrimDef::FunRound => fun("round", None),
+            PrimDef::FunRound64 => fun("round64", None),
+            PrimDef::FunStr => fun("str", None),
+            PrimDef::FunBool => fun("bool", None),
+            PrimDef::FunFloor => fun("floor", None),
+            PrimDef::FunFloor64 => fun("floor64", None),
+            PrimDef::FunCeil => fun("ceil", None),
+            PrimDef::FunCeil64 => fun("ceil64", None),
+            PrimDef::FunLen => fun("len", None),
+            PrimDef::FunMin => fun("min", None),
+            PrimDef::FunMax => fun("max", None),
+            PrimDef::FunAbs => fun("abs", None),
         }
     }
 }
@@ -319,6 +389,9 @@ impl TopLevelDef {
                     r
                 }
             ),
+            TopLevelDef::Variable { name, ty, .. } => {
+                format!("Variable {{ name: {name:?}, ty: {:?} }}", unifier.stringify(*ty),)
+            }
         }
     }
 }
@@ -354,7 +427,13 @@ impl TopLevelComposer {
         });
         let range = unifier.add_ty(TypeEnum::TObj {
             obj_id: PrimDef::Range.id(),
-            fields: HashMap::new(),
+            fields: [
+                ("start".into(), (int32, true)),
+                ("stop".into(), (int32, true)),
+                ("step".into(), (int32, true)),
+            ]
+            .into_iter()
+            .collect(),
             params: VarMap::new(),
         });
         let str = unifier.add_ty(TypeEnum::TObj {
@@ -395,9 +474,9 @@ impl TopLevelComposer {
         let option = unifier.add_ty(TypeEnum::TObj {
             obj_id: PrimDef::Option.id(),
             fields: vec![
-                (PrimDef::OptionIsSome.simple_name().into(), (is_some_type_fun_ty, true)),
-                (PrimDef::OptionIsNone.simple_name().into(), (is_some_type_fun_ty, true)),
-                (PrimDef::OptionUnwrap.simple_name().into(), (unwrap_fun_ty, true)),
+                (PrimDef::FunOptionIsSome.simple_name().into(), (is_some_type_fun_ty, true)),
+                (PrimDef::FunOptionIsNone.simple_name().into(), (is_some_type_fun_ty, true)),
+                (PrimDef::FunOptionUnwrap.simple_name().into(), (unwrap_fun_ty, true)),
             ]
             .into_iter()
             .collect::<HashMap<_, _>>(),
@@ -410,6 +489,13 @@ impl TopLevelComposer {
             _ => unreachable!(),
         };
 
+        let list_elem_tvar = unifier.get_fresh_var(Some("list_elem".into()), None);
+        let list = unifier.add_ty(TypeEnum::TObj {
+            obj_id: PrimDef::List.id(),
+            fields: Mapping::new(),
+            params: into_var_map([list_elem_tvar]),
+        });
+
         let ndarray_dtype_tvar = unifier.get_fresh_var(Some("ndarray_dtype".into()), None);
         let ndarray_ndims_tvar =
             unifier.get_fresh_const_generic_var(size_t_ty, Some("ndarray_ndims".into()), None);
@@ -424,6 +510,7 @@ impl TopLevelComposer {
                 name: "value".into(),
                 ty: ndarray_dtype_tvar.ty,
                 default_value: None,
+                is_vararg: false,
             }],
             ret: none,
             vars: into_var_map([ndarray_dtype_tvar, ndarray_ndims_tvar]),
@@ -431,8 +518,8 @@ impl TopLevelComposer {
         let ndarray = unifier.add_ty(TypeEnum::TObj {
             obj_id: PrimDef::NDArray.id(),
             fields: Mapping::from([
-                (PrimDef::NDArrayCopy.simple_name().into(), (ndarray_copy_fun_ty, true)),
-                (PrimDef::NDArrayFill.simple_name().into(), (ndarray_fill_fun_ty, true)),
+                (PrimDef::FunNDArrayCopy.simple_name().into(), (ndarray_copy_fun_ty, true)),
+                (PrimDef::FunNDArrayFill.simple_name().into(), (ndarray_fill_fun_ty, true)),
             ]),
             params: into_var_map([ndarray_dtype_tvar, ndarray_ndims_tvar]),
         });
@@ -451,6 +538,7 @@ impl TopLevelComposer {
             str,
             exception,
             option,
+            list,
             ndarray,
             size_t,
         };
@@ -505,6 +593,18 @@ impl TopLevelComposer {
         }
     }
 
+    #[must_use]
+    pub fn make_top_level_variable_def(
+        name: String,
+        simple_name: StrRef,
+        ty: Type,
+        ty_decl: Option<Expr>,
+        resolver: Option<Arc<dyn SymbolResolver + Send + Sync>>,
+        loc: Option<Location>,
+    ) -> TopLevelDef {
+        TopLevelDef::Variable { name, simple_name, ty, ty_decl, resolver, loc }
+    }
+
     #[must_use]
     pub fn make_class_method_name(mut class_name: String, method_name: &str) -> String {
         class_name.push('.');
@@ -650,7 +750,16 @@ impl TopLevelComposer {
         )
     }
 
-    pub fn get_all_assigned_field(stmts: &[Stmt<()>]) -> Result<HashSet<StrRef>, HashSet<String>> {
+    /// This function returns the fields that have been initialized in the `__init__` function of a class
+    /// The function takes as input:
+    ///  * `class_id`: The `object_id` of the class whose function is being evaluated (check `TopLevelDef::Class`)
+    ///  * `definition_ast_list`: A list of ast definitions and statements defined in `TopLevelComposer`
+    ///  * `stmts`: The body of function being parsed. Each statment is analyzed to check varaible initialization statements
+    pub fn get_all_assigned_field(
+        class_id: usize,
+        definition_ast_list: &Vec<DefAst>,
+        stmts: &[Stmt<()>],
+    ) -> Result<HashSet<StrRef>, HashSet<String>> {
         let mut result = HashSet::new();
         for s in stmts {
             match &s.node {
@@ -686,30 +795,138 @@ impl TopLevelComposer {
                 // TODO: do not check for For and While?
                 ast::StmtKind::For { body, orelse, .. }
                 | ast::StmtKind::While { body, orelse, .. } => {
-                    result.extend(Self::get_all_assigned_field(body.as_slice())?);
-                    result.extend(Self::get_all_assigned_field(orelse.as_slice())?);
+                    result.extend(Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        body.as_slice(),
+                    )?);
+                    result.extend(Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        orelse.as_slice(),
+                    )?);
                 }
                 ast::StmtKind::If { body, orelse, .. } => {
-                    let inited_for_sure = Self::get_all_assigned_field(body.as_slice())?
-                        .intersection(&Self::get_all_assigned_field(orelse.as_slice())?)
+                    let inited_for_sure = Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        body.as_slice(),
+                    )?
+                    .intersection(&Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        orelse.as_slice(),
+                    )?)
                     .copied()
                     .collect::<HashSet<_>>();
                     result.extend(inited_for_sure);
                 }
                 ast::StmtKind::Try { body, orelse, finalbody, .. } => {
-                    let inited_for_sure = Self::get_all_assigned_field(body.as_slice())?
-                        .intersection(&Self::get_all_assigned_field(orelse.as_slice())?)
+                    let inited_for_sure = Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        body.as_slice(),
+                    )?
+                    .intersection(&Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        orelse.as_slice(),
+                    )?)
                     .copied()
                     .collect::<HashSet<_>>();
                     result.extend(inited_for_sure);
-                    result.extend(Self::get_all_assigned_field(finalbody.as_slice())?);
+                    result.extend(Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        finalbody.as_slice(),
+                    )?);
                 }
                 ast::StmtKind::With { body, .. } => {
-                    result.extend(Self::get_all_assigned_field(body.as_slice())?);
+                    result.extend(Self::get_all_assigned_field(
+                        class_id,
+                        definition_ast_list,
+                        body.as_slice(),
+                    )?);
+                }
+                // Variables Initialized in function calls
+                ast::StmtKind::Expr { value, .. } => {
+                    let ExprKind::Call { func, .. } = &value.node else {
+                        continue;
+                    };
+                    let ExprKind::Attribute { value, attr, .. } = &func.node else {
+                        continue;
+                    };
+                    let ExprKind::Name { id, .. } = &value.node else {
+                        continue;
+                    };
+                    // Need to consider the two cases:
+                    //  Case 1) Call to class function i.e. id = `self`
+                    //  Case 2) Call to class ancestor function i.e. id = ancestor_name
+                    // We leave checking whether function in case 2 belonged to class ancestor or not to type checker
+                    //
+                    // According to current handling of `self`, function definition are fixed and do not change regardless
+                    // of which object is passed as `self` i.e. virtual polymorphism is not supported
+                    // Therefore, we change class id for case 2 to reflect behavior of our compiler
+
+                    let class_name = if *id == "self".into() {
+                        let ast::StmtKind::ClassDef { name, .. } =
+                            &definition_ast_list[class_id].1.as_ref().unwrap().node
+                        else {
+                            unreachable!()
+                        };
+                        name
+                    } else {
+                        id
+                    };
+
+                    let parent_method = definition_ast_list.iter().find_map(|def| {
+                        let (
+                            class_def,
+                            Some(ast::Located {
+                                node: ast::StmtKind::ClassDef { name, body, .. },
+                                ..
+                            }),
+                        ) = &def
+                        else {
+                            return None;
+                        };
+                        let TopLevelDef::Class { object_id: class_id, .. } = &*class_def.read()
+                        else {
+                            unreachable!()
+                        };
+
+                        if name == class_name {
+                            body.iter().find_map(|m| {
+                                let ast::StmtKind::FunctionDef { name, body, .. } = &m.node else {
+                                    return None;
+                                };
+                                if *name == *attr {
+                                    return Some((body.clone(), class_id.0));
+                                }
+                                None
+                            })
+                        } else {
+                            None
+                        }
+                    });
+
+                    // If method body is none then method does not exist
+                    if let Some((method_body, class_id)) = parent_method {
+                        result.extend(Self::get_all_assigned_field(
+                            class_id,
+                            definition_ast_list,
+                            method_body.as_slice(),
+                        )?);
+                    } else {
+                        return Err(HashSet::from([format!(
+                            "{}.{} not found in class {class_name} at {}",
+                            *id, *attr, value.location
+                        )]));
+                    }
                 }
                 ast::StmtKind::Pass { .. }
                 | ast::StmtKind::Assert { .. }
-                | ast::StmtKind::Expr { .. } => {}
+                | ast::StmtKind::AnnAssign { .. } => {}
 
                 _ => {
                     unimplemented!()
@@ -888,7 +1105,9 @@ pub fn arraylike_flatten_element_type(unifier: &mut Unifier, ty: Type) -> Type {
             unpack_ndarray_var_tys(unifier, ty).0
         }
 
-        TypeEnum::TList { ty } => arraylike_flatten_element_type(unifier, *ty),
+        TypeEnum::TObj { obj_id, params, .. } if *obj_id == PrimDef::List.id() => {
+            arraylike_flatten_element_type(unifier, iter_type_vars(params).next().unwrap().ty)
+        }
         _ => ty,
     }
 }
@@ -909,7 +1128,29 @@ pub fn arraylike_get_ndims(unifier: &mut Unifier, ty: Type) -> u64 {
             u64::try_from(values[0].clone()).unwrap()
         }
 
-        TypeEnum::TList { ty } => arraylike_get_ndims(unifier, *ty) + 1,
+        TypeEnum::TObj { obj_id, params, .. } if *obj_id == PrimDef::List.id() => {
+            arraylike_get_ndims(unifier, iter_type_vars(params).next().unwrap().ty) + 1
+        }
         _ => 0,
     }
 }
+
+/// Extract an ndarray's `ndims` [type][`Type`] in `u64`. Panic if not possible.
+/// The `ndims` must only contain 1 value.
+#[must_use]
+pub fn extract_ndims(unifier: &Unifier, ndims_ty: Type) -> u64 {
+    let ndims_ty_enum = unifier.get_ty_immutable(ndims_ty);
+    let TypeEnum::TLiteral { values, .. } = &*ndims_ty_enum else {
+        panic!("ndims_ty should be a TLiteral");
+    };
+
+    assert_eq!(values.len(), 1, "ndims_ty TLiteral should only contain 1 value");
+
+    let ndims = values[0].clone();
+    u64::try_from(ndims).unwrap()
+}
+
+/// Return an ndarray's `ndims` as a typechecker [`Type`] from its `u64` value.
+pub fn create_ndims(unifier: &mut Unifier, ndims: u64) -> Type {
+    unifier.get_fresh_literal(vec![SymbolValue::U64(ndims)], None)
+}

nac3core/src/toplevel/mod.rs

@@ -6,36 +6,36 @@ use std::{
     sync::Arc,
 };
 
-use super::codegen::CodeGenContext;
-use super::typecheck::type_inferencer::PrimitiveStore;
-use super::typecheck::typedef::{
-    FunSignature, FuncArg, SharedUnifier, Type, TypeEnum, Unifier, VarMap,
-};
-use crate::{
-    codegen::CodeGenerator,
-    symbol_resolver::{SymbolResolver, ValueEnum},
-    typecheck::{
-        type_inferencer::CodeLocation,
-        typedef::{CallId, TypeVarId},
-    },
-};
 use inkwell::values::BasicValueEnum;
 use itertools::Itertools;
-use nac3parser::ast::{self, Location, Stmt, StrRef};
 use parking_lot::RwLock;
 
-#[derive(PartialEq, Eq, PartialOrd, Ord, Clone, Copy, Hash, Debug)]
-pub struct DefinitionId(pub usize);
+use nac3parser::ast::{self, Expr, Location, Stmt, StrRef};
+
+use crate::{
+    codegen::{CodeGenContext, CodeGenerator},
+    symbol_resolver::{SymbolResolver, ValueEnum},
+    typecheck::{
+        type_inferencer::{CodeLocation, PrimitiveStore},
+        typedef::{
+            CallId, FunSignature, FuncArg, SharedUnifier, Type, TypeEnum, TypeVarId, Unifier,
+            VarMap,
+        },
+    },
+};
+use composer::*;
+use type_annotation::*;
 
 pub mod builtins;
 pub mod composer;
 pub mod helper;
 pub mod numpy;
-pub mod type_annotation;
-use composer::*;
-use type_annotation::*;
 #[cfg(test)]
 mod test;
+pub mod type_annotation;
+
+#[derive(PartialEq, Eq, PartialOrd, Ord, Clone, Copy, Hash, Debug)]
+pub struct DefinitionId(pub usize);
 
 type GenCallCallback = dyn for<'ctx, 'a> Fn(
         &mut CodeGenContext<'ctx, 'a>,
@@ -148,6 +148,25 @@ pub enum TopLevelDef {
         /// Definition location.
         loc: Option<Location>,
     },
+    Variable {
+        /// Qualified name of the global variable, should be unique globally.
+        name: String,
+
+        /// Simple name, the same as in method/function definition.
+        simple_name: StrRef,
+
+        /// Type of the global variable.
+        ty: Type,
+
+        /// The declared type of the global variable, or [`None`] if no type annotation is provided.
+        ty_decl: Option<Expr>,
+
+        /// Symbol resolver of the module defined the class.
+        resolver: Option<Arc<dyn SymbolResolver + Send + Sync>>,
+
+        /// Definition location.
+        loc: Option<Location>,
+    },
 }
 
 pub struct TopLevelContext {

nac3core/src/toplevel/numpy.rs

@@ -1,11 +1,10 @@
-use crate::{
-    toplevel::helper::PrimDef,
-    typecheck::{
+use itertools::Itertools;
+
+use super::helper::PrimDef;
+use crate::typecheck::{
     type_inferencer::PrimitiveStore,
     typedef::{Type, TypeEnum, TypeVarId, Unifier, VarMap},
-    },
 };
-use itertools::Itertools;
 
 /// Creates a `ndarray` [`Type`] with the given type arguments.
 ///

nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__generic_class.snap

@@ -5,7 +5,7 @@ expression: res_vec
 [
     "Class {\nname: \"Generic_A\",\nancestors: [\"Generic_A[V]\", \"B\"],\nfields: [\"aa\", \"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"foo\", \"fn[[b:T], none]\"), (\"fun\", \"fn[[a:int32], V]\")],\ntype_vars: [\"V\"]\n}\n",
     "Function {\nname: \"Generic_A.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
-    "Function {\nname: \"Generic_A.fun\",\nsig: \"fn[[a:int32], V]\",\nvar_id: [TypeVarId(240)]\n}\n",
+    "Function {\nname: \"Generic_A.fun\",\nsig: \"fn[[a:int32], V]\",\nvar_id: [TypeVarId(241)]\n}\n",
     "Class {\nname: \"B\",\nancestors: [\"B\"],\nfields: [\"aa\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"foo\", \"fn[[b:T], none]\")],\ntype_vars: []\n}\n",
     "Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
     "Function {\nname: \"B.foo\",\nsig: \"fn[[b:T], none]\",\nvar_id: []\n}\n",

nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__inheritance_override.snap

@@ -7,7 +7,7 @@ expression: res_vec
     "Function {\nname: \"A.__init__\",\nsig: \"fn[[t:T], none]\",\nvar_id: []\n}\n",
     "Function {\nname: \"A.fun\",\nsig: \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\",\nvar_id: []\n}\n",
     "Function {\nname: \"A.foo\",\nsig: \"fn[[c:C], none]\",\nvar_id: []\n}\n",
-    "Class {\nname: \"B\",\nancestors: [\"B[typevar229]\", \"A[float]\"],\nfields: [\"a\", \"b\", \"c\", \"d\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\"), (\"foo\", \"fn[[c:C], none]\")],\ntype_vars: [\"typevar229\"]\n}\n",
+    "Class {\nname: \"B\",\nancestors: [\"B[typevar230]\", \"A[float]\"],\nfields: [\"a\", \"b\", \"c\", \"d\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\"), (\"foo\", \"fn[[c:C], none]\")],\ntype_vars: [\"typevar230\"]\n}\n",
     "Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
     "Function {\nname: \"B.fun\",\nsig: \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\",\nvar_id: []\n}\n",
     "Class {\nname: \"C\",\nancestors: [\"C\", \"B[bool]\", \"A[float]\"],\nfields: [\"a\", \"b\", \"c\", \"d\", \"e\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\"), (\"foo\", \"fn[[c:C], none]\")],\ntype_vars: []\n}\n",

nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__list_tuple_generic.snap

@@ -5,8 +5,8 @@ expression: res_vec
 [
     "Function {\nname: \"foo\",\nsig: \"fn[[a:list[int32], b:tuple[T, float]], A[B, bool]]\",\nvar_id: []\n}\n",
     "Class {\nname: \"A\",\nancestors: [\"A[T, V]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[v:V], none]\"), (\"fun\", \"fn[[a:T], V]\")],\ntype_vars: [\"T\", \"V\"]\n}\n",
-    "Function {\nname: \"A.__init__\",\nsig: \"fn[[v:V], none]\",\nvar_id: [TypeVarId(242)]\n}\n",
-    "Function {\nname: \"A.fun\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(247)]\n}\n",
+    "Function {\nname: \"A.__init__\",\nsig: \"fn[[v:V], none]\",\nvar_id: [TypeVarId(243)]\n}\n",
+    "Function {\nname: \"A.fun\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(248)]\n}\n",
     "Function {\nname: \"gfun\",\nsig: \"fn[[a:A[list[float], int32]], none]\",\nvar_id: []\n}\n",
     "Class {\nname: \"B\",\nancestors: [\"B\"],\nfields: [],\nmethods: [(\"__init__\", \"fn[[], none]\")],\ntype_vars: []\n}\n",
     "Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",

nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__self1.snap

@@ -3,7 +3,7 @@ source: nac3core/src/toplevel/test.rs
 expression: res_vec
 ---
 [
-    "Class {\nname: \"A\",\nancestors: [\"A[typevar228, typevar229]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[a:A[float, bool], b:B], none]\"), (\"fun\", \"fn[[a:A[float, bool]], A[bool, int32]]\")],\ntype_vars: [\"typevar228\", \"typevar229\"]\n}\n",
+    "Class {\nname: \"A\",\nancestors: [\"A[typevar229, typevar230]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[a:A[float, bool], b:B], none]\"), (\"fun\", \"fn[[a:A[float, bool]], A[bool, int32]]\")],\ntype_vars: [\"typevar229\", \"typevar230\"]\n}\n",
     "Function {\nname: \"A.__init__\",\nsig: \"fn[[a:A[float, bool], b:B], none]\",\nvar_id: []\n}\n",
     "Function {\nname: \"A.fun\",\nsig: \"fn[[a:A[float, bool]], A[bool, int32]]\",\nvar_id: []\n}\n",
     "Class {\nname: \"B\",\nancestors: [\"B\", \"A[int64, bool]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:A[float, bool]], A[bool, int32]]\"), (\"foo\", \"fn[[b:B], B]\"), (\"bar\", \"fn[[a:A[list[B], int32]], tuple[A[virtual[A[B, int32]], bool], B]]\")],\ntype_vars: []\n}\n",

nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__simple_class_compose.snap

@@ -6,12 +6,12 @@ expression: res_vec
     "Class {\nname: \"A\",\nancestors: [\"A\"],\nfields: [\"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[b:B], none]\"), (\"foo\", \"fn[[a:T, b:V], none]\")],\ntype_vars: []\n}\n",
     "Function {\nname: \"A.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
     "Function {\nname: \"A.fun\",\nsig: \"fn[[b:B], none]\",\nvar_id: []\n}\n",
-    "Function {\nname: \"A.foo\",\nsig: \"fn[[a:T, b:V], none]\",\nvar_id: [TypeVarId(248)]\n}\n",
+    "Function {\nname: \"A.foo\",\nsig: \"fn[[a:T, b:V], none]\",\nvar_id: [TypeVarId(249)]\n}\n",
     "Class {\nname: \"B\",\nancestors: [\"B\", \"C\", \"A\"],\nfields: [\"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[b:B], none]\"), (\"foo\", \"fn[[a:T, b:V], none]\")],\ntype_vars: []\n}\n",
     "Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
     "Class {\nname: \"C\",\nancestors: [\"C\", \"A\"],\nfields: [\"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[b:B], none]\"), (\"foo\", \"fn[[a:T, b:V], none]\")],\ntype_vars: []\n}\n",
     "Function {\nname: \"C.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
     "Function {\nname: \"C.fun\",\nsig: \"fn[[b:B], none]\",\nvar_id: []\n}\n",
     "Function {\nname: \"foo\",\nsig: \"fn[[a:A], none]\",\nvar_id: []\n}\n",
-    "Function {\nname: \"ff\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(256)]\n}\n",
+    "Function {\nname: \"ff\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(257)]\n}\n",
 ]

nac3core/src/toplevel/test.rs

@@ -1,20 +1,23 @@
+use std::{collections::HashMap, sync::Arc};
+
+use indoc::indoc;
+use parking_lot::Mutex;
+use test_case::test_case;
+
+use nac3parser::{
+    ast::{fold::Fold, FileName},
+    parser::parse_program,
+};
+
+use super::{helper::PrimDef, DefinitionId, *};
 use crate::{
     codegen::CodeGenContext,
     symbol_resolver::{SymbolResolver, ValueEnum},
-    toplevel::DefinitionId,
     typecheck::{
         type_inferencer::PrimitiveStore,
-        typedef::{Type, Unifier},
+        typedef::{into_var_map, Type, Unifier},
     },
 };
-use indoc::indoc;
-use nac3parser::ast::FileName;
-use nac3parser::{ast::fold::Fold, parser::parse_program};
-use parking_lot::Mutex;
-use std::{collections::HashMap, sync::Arc};
-use test_case::test_case;
-
-use super::*;
 
 struct ResolverInternal {
     id_to_type: Mutex<HashMap<StrRef, Type>>,
@@ -61,6 +64,7 @@ impl SymbolResolver for Resolver {
         &self,
         _: StrRef,
         _: &mut CodeGenContext<'ctx, '_>,
+        _: &mut dyn CodeGenerator,
     ) -> Option<ValueEnum<'ctx>> {
         unimplemented!()
     }
@@ -116,7 +120,8 @@ impl SymbolResolver for Resolver {
     "register"
 )]
 fn test_simple_register(source: Vec<&str>) {
-    let mut composer = TopLevelComposer::new(Vec::new(), ComposerConfig::default(), 64).0;
+    let mut composer =
+        TopLevelComposer::new(Vec::new(), Vec::new(), ComposerConfig::default(), 64).0;
 
     for s in source {
         let ast = parse_program(s, FileName::default()).unwrap();
@@ -136,7 +141,8 @@ fn test_simple_register(source: Vec<&str>) {
     "register"
 )]
 fn test_simple_register_without_constructor(source: &str) {
-    let mut composer = TopLevelComposer::new(Vec::new(), ComposerConfig::default(), 64).0;
+    let mut composer =
+        TopLevelComposer::new(Vec::new(), Vec::new(), ComposerConfig::default(), 64).0;
     let ast = parse_program(source, FileName::default()).unwrap();
     let ast = ast[0].clone();
     composer.register_top_level(ast, None, "", true).unwrap();
@@ -170,7 +176,8 @@ fn test_simple_register_without_constructor(source: &str) {
     "function compose"
 )]
 fn test_simple_function_analyze(source: &[&str], tys: &[&str], names: &[&str]) {
-    let mut composer = TopLevelComposer::new(Vec::new(), ComposerConfig::default(), 64).0;
+    let mut composer =
+        TopLevelComposer::new(Vec::new(), Vec::new(), ComposerConfig::default(), 64).0;
 
     let internal_resolver = Arc::new(ResolverInternal {
         id_to_def: Mutex::default(),
@@ -518,7 +525,8 @@ fn test_simple_function_analyze(source: &[&str], tys: &[&str], names: &[&str]) {
 )]
 fn test_analyze(source: &[&str], res: &[&str]) {
     let print = false;
-    let mut composer = TopLevelComposer::new(Vec::new(), ComposerConfig::default(), 64).0;
+    let mut composer =
+        TopLevelComposer::new(Vec::new(), Vec::new(), ComposerConfig::default(), 64).0;
 
     let internal_resolver = make_internal_resolver_with_tvar(
         vec![
@@ -695,7 +703,8 @@ fn test_analyze(source: &[&str], res: &[&str]) {
 )]
 fn test_inference(source: Vec<&str>, res: &[&str]) {
     let print = true;
-    let mut composer = TopLevelComposer::new(Vec::new(), ComposerConfig::default(), 64).0;
+    let mut composer =
+        TopLevelComposer::new(Vec::new(), Vec::new(), ComposerConfig::default(), 64).0;
 
     let internal_resolver = make_internal_resolver_with_tvar(
         vec![
@@ -775,8 +784,15 @@ fn make_internal_resolver_with_tvar(
     unifier: &mut Unifier,
     print: bool,
 ) -> Arc<ResolverInternal> {
+    let list_elem_tvar = unifier.get_fresh_var(Some("list_elem".into()), None);
+    let list = unifier.add_ty(TypeEnum::TObj {
+        obj_id: PrimDef::List.id(),
+        fields: HashMap::new(),
+        params: into_var_map([list_elem_tvar]),
+    });
+
     let res: Arc<ResolverInternal> = ResolverInternal {
-        id_to_def: Mutex::default(),
+        id_to_def: Mutex::new(HashMap::from([("list".into(), PrimDef::List.id())])),
         id_to_type: tvars
             .into_iter()
             .map(|(name, range)| {
@@ -790,7 +806,7 @@ fn make_internal_resolver_with_tvar(
             })
             .collect::<HashMap<_, _>>()
             .into(),
-        class_names: Mutex::default(),
+        class_names: Mutex::new(HashMap::from([("list".into(), list)])),
     }
     .into();
     if print {

nac3core/src/toplevel/type_annotation.rs

@@ -1,9 +1,13 @@
-use super::*;
-use crate::symbol_resolver::SymbolValue;
-use crate::toplevel::helper::PrimDef;
-use crate::typecheck::typedef::VarMap;
+use strum::IntoEnumIterator;
+
 use nac3parser::ast::Constant;
 
+use super::{
+    helper::{PrimDef, PrimDefDetails},
+    *,
+};
+use crate::{symbol_resolver::SymbolValue, typecheck::typedef::VarMap};
+
 #[derive(Clone, Debug)]
 pub enum TypeAnnotation {
     Primitive(Type),
@@ -18,7 +22,6 @@ pub enum TypeAnnotation {
     TypeVar(Type),
     /// A `Literal` allowing a subset of literals.
     Literal(Vec<Constant>),
-    List(Box<TypeAnnotation>),
     Tuple(Vec<TypeAnnotation>),
 }
 
@@ -51,7 +54,6 @@ impl TypeAnnotation {
                 format!("Literal({})", values.iter().map(|v| format!("{v:?}")).join(", "))
             }
             Virtual(ty) => format!("virtual[{}]", ty.stringify(unifier)),
-            List(ty) => format!("list[{}]", ty.stringify(unifier)),
             Tuple(types) => {
                 format!(
                     "tuple[{}]",
@@ -145,9 +147,7 @@ pub fn parse_ast_to_type_annotation_kinds<T, S: std::hash::BuildHasher + Clone>(
          slice: &ast::Expr<T>,
          unifier: &mut Unifier,
          mut locked: HashMap<DefinitionId, Vec<Type>, S>| {
-            if ["virtual".into(), "Generic".into(), "list".into(), "tuple".into(), "Option".into()]
-                .contains(id)
-            {
+            if ["virtual".into(), "Generic".into(), "tuple".into(), "Option".into()].contains(id) {
                 return Err(HashSet::from([format!(
                     "keywords cannot be class name (at {})",
                     expr.location
@@ -236,23 +236,6 @@ pub fn parse_ast_to_type_annotation_kinds<T, S: std::hash::BuildHasher + Clone>(
             Ok(TypeAnnotation::Virtual(def.into()))
         }
 
-        // list
-        ast::ExprKind::Subscript { value, slice, .. }
-            if {
-                matches!(&value.node, ast::ExprKind::Name { id, .. } if id == &"list".into())
-            } =>
-        {
-            let def_ann = parse_ast_to_type_annotation_kinds(
-                resolver,
-                top_level_defs,
-                unifier,
-                primitives,
-                slice.as_ref(),
-                locked,
-            )?;
-            Ok(TypeAnnotation::List(def_ann.into()))
-        }
-
         // option
         ast::ExprKind::Subscript { value, slice, .. }
             if {
@@ -378,6 +361,7 @@ pub fn parse_ast_to_type_annotation_kinds<T, S: std::hash::BuildHasher + Clone>(
 pub fn get_type_from_type_annotation_kinds(
     top_level_defs: &[Arc<RwLock<TopLevelDef>>],
     unifier: &mut Unifier,
+    primitives: &PrimitiveStore,
     ann: &TypeAnnotation,
     subst_list: &mut Option<Vec<Type>>,
 ) -> Result<Type, HashSet<String>> {
@@ -400,10 +384,43 @@ pub fn get_type_from_type_annotation_kinds(
             let param_ty = params
                 .iter()
                 .map(|x| {
-                    get_type_from_type_annotation_kinds(top_level_defs, unifier, x, subst_list)
+                    get_type_from_type_annotation_kinds(
+                        top_level_defs,
+                        unifier,
+                        primitives,
+                        x,
+                        subst_list,
+                    )
                 })
                 .collect::<Result<Vec<_>, _>>()?;
 
+            let ty = if let Some(prim_def) = PrimDef::iter().find(|prim| prim.id() == *obj_id) {
+                // Primitive TopLevelDefs do not contain all fields that are present in their Type
+                // counterparts, so directly perform subst on the Type instead.
+
+                let PrimDefDetails::PrimClass { get_ty_fn, .. } = prim_def.details() else {
+                    unreachable!()
+                };
+
+                let base_ty = get_ty_fn(primitives);
+                let params =
+                    if let TypeEnum::TObj { params, .. } = &*unifier.get_ty_immutable(base_ty) {
+                        params.clone()
+                    } else {
+                        unreachable!()
+                    };
+
+                unifier
+                    .subst(
+                        get_ty_fn(primitives),
+                        &params
+                            .iter()
+                            .zip(param_ty)
+                            .map(|(obj_tv, param)| (*obj_tv.0, param))
+                            .collect(),
+                    )
+                    .unwrap_or(base_ty)
+            } else {
                 let subst = {
                     // check for compatible range
                     // TODO: if allow type var to be applied(now this disallowed in the parse_to_type_annotation), need more check
@@ -445,12 +462,15 @@ pub fn get_type_from_type_annotation_kinds(
                                 }
                             }
 
-                        TypeEnum::TVar { id, range, name, loc, is_const_generic: true, .. } => {
+                            TypeEnum::TVar {
+                                id, range, name, loc, is_const_generic: true, ..
+                            } => {
                                 let ty = range[0];
                                 let ok: bool = {
                                     // create a temp type var and unify to check compatibility
                                     p == *tvar || {
-                                    let temp = unifier.get_fresh_const_generic_var(ty, *name, *loc);
+                                        let temp =
+                                            unifier.get_fresh_const_generic_var(ty, *name, *loc);
                                         unifier.unify(temp.ty, p).is_ok()
                                     }
                                 };
@@ -489,11 +509,16 @@ pub fn get_type_from_type_annotation_kinds(
                     fields: tobj_fields,
                     params: subst,
                 });
+
                 if need_subst {
                     if let Some(wl) = subst_list.as_mut() {
                         wl.push(ty);
                     }
                 }
+
+                ty
+            };
+
             Ok(ty)
         }
         TypeAnnotation::Primitive(ty) | TypeAnnotation::TypeVar(ty) => Ok(*ty),
@@ -511,28 +536,26 @@ pub fn get_type_from_type_annotation_kinds(
             let ty = get_type_from_type_annotation_kinds(
                 top_level_defs,
                 unifier,
+                primitives,
                 ty.as_ref(),
                 subst_list,
             )?;
             Ok(unifier.add_ty(TypeEnum::TVirtual { ty }))
         }
-        TypeAnnotation::List(ty) => {
-            let ty = get_type_from_type_annotation_kinds(
-                top_level_defs,
-                unifier,
-                ty.as_ref(),
-                subst_list,
-            )?;
-            Ok(unifier.add_ty(TypeEnum::TList { ty }))
-        }
         TypeAnnotation::Tuple(tys) => {
             let tys = tys
                 .iter()
                 .map(|x| {
-                    get_type_from_type_annotation_kinds(top_level_defs, unifier, x, subst_list)
+                    get_type_from_type_annotation_kinds(
+                        top_level_defs,
+                        unifier,
+                        primitives,
+                        x,
+                        subst_list,
+                    )
                 })
                 .collect::<Result<Vec<_>, _>>()?;
-            Ok(unifier.add_ty(TypeEnum::TTuple { ty: tys }))
+            Ok(unifier.add_ty(TypeEnum::TTuple { ty: tys, is_vararg_ctx: false }))
         }
     }
 }
@@ -565,7 +588,7 @@ pub fn get_type_var_contained_in_type_annotation(ann: &TypeAnnotation) -> Vec<Ty
     let mut result: Vec<TypeAnnotation> = Vec::new();
     match ann {
         TypeAnnotation::TypeVar(..) => result.push(ann.clone()),
-        TypeAnnotation::Virtual(ann) | TypeAnnotation::List(ann) => {
+        TypeAnnotation::Virtual(ann) => {
             result.extend(get_type_var_contained_in_type_annotation(ann.as_ref()));
         }
         TypeAnnotation::CustomClass { params, .. } => {
@@ -606,8 +629,7 @@ pub fn check_overload_type_annotation_compatible(
 
             a == b
         }
-        (TypeAnnotation::Virtual(a), TypeAnnotation::Virtual(b))
-        | (TypeAnnotation::List(a), TypeAnnotation::List(b)) => {
+        (TypeAnnotation::Virtual(a), TypeAnnotation::Virtual(b)) => {
             check_overload_type_annotation_compatible(a.as_ref(), b.as_ref(), unifier)
         }
 

nac3core/src/typecheck/function_check.rs

@@ -1,13 +1,19 @@
-use crate::typecheck::typedef::TypeEnum;
+use std::{
+    collections::{HashMap, HashSet},
+    iter::once,
+};
 
-use super::type_inferencer::Inferencer;
-use super::typedef::Type;
 use nac3parser::ast::{
     self, Constant, Expr, ExprKind,
     Operator::{LShift, RShift},
     Stmt, StmtKind, StrRef,
 };
-use std::{collections::HashSet, iter::once};
+
+use super::{
+    type_inferencer::{DeclarationSource, IdentifierInfo, Inferencer},
+    typedef::{Type, TypeEnum},
+};
+use crate::toplevel::helper::PrimDef;
 
 impl<'a> Inferencer<'a> {
     fn should_have_value(&mut self, expr: &Expr<Option<Type>>) -> Result<(), HashSet<String>> {
@@ -21,26 +27,45 @@ impl<'a> Inferencer<'a> {
     fn check_pattern(
         &mut self,
         pattern: &Expr<Option<Type>>,
-        defined_identifiers: &mut HashSet<StrRef>,
+        defined_identifiers: &mut HashMap<StrRef, IdentifierInfo>,
     ) -> Result<(), HashSet<String>> {
         match &pattern.node {
             ExprKind::Name { id, .. } if id == &"none".into() => {
                 Err(HashSet::from([format!("cannot assign to a `none` (at {})", pattern.location)]))
             }
             ExprKind::Name { id, .. } => {
-                if !defined_identifiers.contains(id) {
-                    defined_identifiers.insert(*id);
+                // If `id` refers to a declared symbol, reject this assignment if it is used in the
+                // context of an (implicit) global variable
+                if let Some(id_info) = defined_identifiers.get(id) {
+                    if matches!(
+                        id_info.source,
+                        DeclarationSource::Global { is_explicit: Some(false) }
+                    ) {
+                        return Err(HashSet::from([format!(
+                            "cannot access local variable '{id}' before it is declared (at {})",
+                            pattern.location
+                        )]));
+                    }
+                }
+
+                if !defined_identifiers.contains_key(id) {
+                    defined_identifiers.insert(*id, IdentifierInfo::default());
                 }
                 self.should_have_value(pattern)?;
                 Ok(())
             }
-            ExprKind::Tuple { elts, .. } => {
+            ExprKind::List { elts, .. } | ExprKind::Tuple { elts, .. } => {
                 for elt in elts {
                     self.check_pattern(elt, defined_identifiers)?;
                     self.should_have_value(elt)?;
                 }
                 Ok(())
             }
+            ExprKind::Starred { value, .. } => {
+                self.check_pattern(value, defined_identifiers)?;
+                self.should_have_value(value)?;
+                Ok(())
+            }
             ExprKind::Subscript { value, slice, .. } => {
                 self.check_expr(value, defined_identifiers)?;
                 self.should_have_value(value)?;
@@ -64,11 +89,12 @@ impl<'a> Inferencer<'a> {
     fn check_expr(
         &mut self,
         expr: &Expr<Option<Type>>,
-        defined_identifiers: &mut HashSet<StrRef>,
+        defined_identifiers: &mut HashMap<StrRef, IdentifierInfo>,
     ) -> Result<(), HashSet<String>> {
         // there are some cases where the custom field is None
         if let Some(ty) = &expr.custom {
             if !matches!(&expr.node, ExprKind::Constant { value: Constant::Ellipsis, .. })
+                && !ty.obj_id(self.unifier).is_some_and(|id| id == PrimDef::List.id())
                 && !self.unifier.is_concrete(*ty, &self.function_data.bound_variables)
             {
                 return Err(HashSet::from([format!(
@@ -84,7 +110,7 @@ impl<'a> Inferencer<'a> {
                     return Ok(());
                 }
                 self.should_have_value(expr)?;
-                if !defined_identifiers.contains(id) {
+                if !defined_identifiers.contains_key(id) {
                     match self.function_data.resolver.get_symbol_type(
                         self.unifier,
                         &self.top_level.definitions.read(),
@@ -92,7 +118,22 @@ impl<'a> Inferencer<'a> {
                         *id,
                     ) {
                         Ok(_) => {
-                            self.defined_identifiers.insert(*id);
+                            let is_global = self.is_id_global(*id);
+
+                            defined_identifiers.insert(
+                                *id,
+                                IdentifierInfo {
+                                    source: match is_global {
+                                        Some(true) => {
+                                            DeclarationSource::Global { is_explicit: Some(false) }
+                                        }
+                                        Some(false) => {
+                                            DeclarationSource::Global { is_explicit: None }
+                                        }
+                                        None => DeclarationSource::Local,
+                                    },
+                                },
+                            );
                         }
                         Err(e) => {
                             return Err(HashSet::from([format!(
@@ -165,9 +206,7 @@ impl<'a> Inferencer<'a> {
                 let mut defined_identifiers = defined_identifiers.clone();
                 for arg in &args.args {
                     // TODO: should we check the types here?
-                    if !defined_identifiers.contains(&arg.node.arg) {
-                        defined_identifiers.insert(arg.node.arg);
-                    }
+                    defined_identifiers.entry(arg.node.arg).or_default();
                 }
                 self.check_expr(body, &mut defined_identifiers)?;
             }
@@ -206,6 +245,9 @@ impl<'a> Inferencer<'a> {
     /// This is a workaround preventing the caller from using a variable `alloca`-ed in the body, which
     /// is freed when the function returns.
     fn check_return_value_ty(&mut self, ret_ty: Type) -> bool {
+        if cfg!(feature = "no-escape-analysis") {
+            true
+        } else {
             match &*self.unifier.get_ty_immutable(ret_ty) {
                 TypeEnum::TObj { .. } => [
                     self.primitives.int32,
@@ -217,16 +259,17 @@ impl<'a> Inferencer<'a> {
                 ]
                 .iter()
                 .any(|allowed_ty| self.unifier.unioned(ret_ty, *allowed_ty)),
-            TypeEnum::TTuple { ty } => ty.iter().all(|t| self.check_return_value_ty(*t)),
+                TypeEnum::TTuple { ty, .. } => ty.iter().all(|t| self.check_return_value_ty(*t)),
                 _ => false,
             }
         }
+    }
 
     // check statements for proper identifier def-use and return on all paths
     fn check_stmt(
         &mut self,
         stmt: &Stmt<Option<Type>>,
-        defined_identifiers: &mut HashSet<StrRef>,
+        defined_identifiers: &mut HashMap<StrRef, IdentifierInfo>,
     ) -> Result<bool, HashSet<String>> {
         match &stmt.node {
             StmtKind::For { target, iter, body, orelse, .. } => {
@@ -252,9 +295,11 @@ impl<'a> Inferencer<'a> {
                 let body_returned = self.check_block(body, &mut body_identifiers)?;
                 let orelse_returned = self.check_block(orelse, &mut orelse_identifiers)?;
 
-                for ident in &body_identifiers {
-                    if !defined_identifiers.contains(ident) && orelse_identifiers.contains(ident) {
-                        defined_identifiers.insert(*ident);
+                for ident in body_identifiers.keys() {
+                    if !defined_identifiers.contains_key(ident)
+                        && orelse_identifiers.contains_key(ident)
+                    {
+                        defined_identifiers.insert(*ident, IdentifierInfo::default());
                     }
                 }
                 Ok(body_returned && orelse_returned)
@@ -285,7 +330,7 @@ impl<'a> Inferencer<'a> {
                     let mut defined_identifiers = defined_identifiers.clone();
                     let ast::ExcepthandlerKind::ExceptHandler { name, body, .. } = &handler.node;
                     if let Some(name) = name {
-                        defined_identifiers.insert(*name);
+                        defined_identifiers.insert(*name, IdentifierInfo::default());
                     }
                     self.check_block(body, &mut defined_identifiers)?;
                 }
@@ -349,6 +394,44 @@ impl<'a> Inferencer<'a> {
                 }
                 Ok(true)
             }
+            StmtKind::Global { names, .. } => {
+                for id in names {
+                    if let Some(id_info) = defined_identifiers.get(id) {
+                        if id_info.source == DeclarationSource::Local {
+                            return Err(HashSet::from([format!(
+                                "name '{id}' is referenced prior to global declaration at {}",
+                                stmt.location,
+                            )]));
+                        }
+
+                        continue;
+                    }
+
+                    match self.function_data.resolver.get_symbol_type(
+                        self.unifier,
+                        &self.top_level.definitions.read(),
+                        self.primitives,
+                        *id,
+                    ) {
+                        Ok(_) => {
+                            defined_identifiers.insert(
+                                *id,
+                                IdentifierInfo {
+                                    source: DeclarationSource::Global { is_explicit: Some(true) },
+                                },
+                            );
+                        }
+                        Err(e) => {
+                            return Err(HashSet::from([format!(
+                                "type error at identifier `{}` ({}) at {}",
+                                id, e, stmt.location
+                            )]))
+                        }
+                    }
+                }
+
+                Ok(false)
+            }
             // break, raise, etc.
             _ => Ok(false),
         }
@@ -357,7 +440,7 @@ impl<'a> Inferencer<'a> {
     pub fn check_block(
         &mut self,
         block: &[Stmt<Option<Type>>],
-        defined_identifiers: &mut HashSet<StrRef>,
+        defined_identifiers: &mut HashMap<StrRef, IdentifierInfo>,
     ) -> Result<bool, HashSet<String>> {
         let mut ret = false;
         for stmt in block {

nac3core/src/typecheck/magic_methods.rs

@@ -1,79 +1,154 @@
-use crate::symbol_resolver::SymbolValue;
-use crate::toplevel::helper::PrimDef;
-use crate::toplevel::numpy::{make_ndarray_ty, unpack_ndarray_var_tys};
-use crate::typecheck::{
+use std::{cmp::max, collections::HashMap, rc::Rc};
+
+use itertools::{iproduct, Itertools};
+use strum::IntoEnumIterator;
+
+use nac3parser::ast::{Cmpop, Operator, StrRef, Unaryop};
+
+use super::{
     type_inferencer::*,
     typedef::{FunSignature, FuncArg, Type, TypeEnum, Unifier, VarMap},
 };
-use itertools::Itertools;
-use nac3parser::ast::StrRef;
-use nac3parser::ast::{Cmpop, Operator, Unaryop};
-use std::cmp::max;
-use std::collections::HashMap;
-use std::rc::Rc;
-use strum::IntoEnumIterator;
+use crate::{
+    symbol_resolver::SymbolValue,
+    toplevel::{
+        helper::PrimDef,
+        numpy::{make_ndarray_ty, unpack_ndarray_var_tys},
+    },
+};
 
+/// The variant of a binary operator.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum BinopVariant {
+    /// The normal variant.
+    /// For addition, it would be `+`.
+    Normal,
+    /// The "Augmented Assigning Operator" variant.
+    /// For addition, it would be `+=`.
+    AugAssign,
+}
+
+/// A binary operator with its variant.
+#[derive(Debug, Clone, Copy)]
+pub struct Binop {
+    /// The base [`Operator`] of this binary operator.
+    pub base: Operator,
+    /// The variant of this binary operator.
+    pub variant: BinopVariant,
+}
+
+impl Binop {
+    /// Make a [`Binop`] of the normal variant from an [`Operator`].
     #[must_use]
-pub fn binop_name(op: Operator) -> &'static str {
-    match op {
-        Operator::Add => "__add__",
-        Operator::Sub => "__sub__",
-        Operator::Div => "__truediv__",
-        Operator::Mod => "__mod__",
-        Operator::Mult => "__mul__",
-        Operator::Pow => "__pow__",
-        Operator::BitOr => "__or__",
-        Operator::BitXor => "__xor__",
-        Operator::BitAnd => "__and__",
-        Operator::LShift => "__lshift__",
-        Operator::RShift => "__rshift__",
-        Operator::FloorDiv => "__floordiv__",
-        Operator::MatMult => "__matmul__",
+    pub fn normal(base: Operator) -> Self {
+        Binop { base, variant: BinopVariant::Normal }
+    }
+
+    /// Make a [`Binop`] of the aug assign variant from an [`Operator`].
+    #[must_use]
+    pub fn aug_assign(base: Operator) -> Self {
+        Binop { base, variant: BinopVariant::AugAssign }
     }
 }
 
-#[must_use]
-pub fn binop_assign_name(op: Operator) -> &'static str {
-    match op {
-        Operator::Add => "__iadd__",
-        Operator::Sub => "__isub__",
-        Operator::Div => "__itruediv__",
-        Operator::Mod => "__imod__",
-        Operator::Mult => "__imul__",
-        Operator::Pow => "__ipow__",
-        Operator::BitOr => "__ior__",
-        Operator::BitXor => "__ixor__",
-        Operator::BitAnd => "__iand__",
-        Operator::LShift => "__ilshift__",
-        Operator::RShift => "__irshift__",
-        Operator::FloorDiv => "__ifloordiv__",
-        Operator::MatMult => "__imatmul__",
-    }
+/// Details about an operator (unary, binary, etc...) in Python
+#[derive(Debug, Clone, Copy)]
+pub struct OpInfo {
+    /// The method name of the binary operator.
+    /// For addition, this would be `__add__`, and `__iadd__` if
+    /// it is the augmented assigning variant.
+    pub method_name: &'static str,
+    /// The symbol of the binary operator.
+    /// For addition, this would be `+`, and `+=` if
+    /// it is the augmented assigning variant.
+    pub symbol: &'static str,
 }
 
-#[must_use]
-pub fn unaryop_name(op: Unaryop) -> &'static str {
-    match op {
-        Unaryop::UAdd => "__pos__",
-        Unaryop::USub => "__neg__",
-        Unaryop::Not => "__not__",
-        Unaryop::Invert => "__inv__",
-    }
+/// Helper macro to conveniently build an [`OpInfo`].
+///
+/// Example usage: `make_info("add", "+")` generates `OpInfo { name: "__add__", symbol: "+" }`
+macro_rules! make_op_info {
+    ($name:expr, $symbol:expr) => {
+        OpInfo { method_name: concat!("__", $name, "__"), symbol: $symbol }
+    };
 }
 
-#[must_use]
-pub fn comparison_name(op: Cmpop) -> Option<&'static str> {
+pub trait HasOpInfo {
+    fn op_info(&self) -> OpInfo;
+}
+
+fn try_get_cmpop_info(op: Cmpop) -> Option<OpInfo> {
     match op {
-        Cmpop::Lt => Some("__lt__"),
-        Cmpop::LtE => Some("__le__"),
-        Cmpop::Gt => Some("__gt__"),
-        Cmpop::GtE => Some("__ge__"),
-        Cmpop::Eq => Some("__eq__"),
-        Cmpop::NotEq => Some("__ne__"),
+        Cmpop::Lt => Some(make_op_info!("lt", "<")),
+        Cmpop::LtE => Some(make_op_info!("le", "<=")),
+        Cmpop::Gt => Some(make_op_info!("gt", ">")),
+        Cmpop::GtE => Some(make_op_info!("ge", ">=")),
+        Cmpop::Eq => Some(make_op_info!("eq", "==")),
+        Cmpop::NotEq => Some(make_op_info!("ne", "!=")),
         _ => None,
     }
 }
 
+impl OpInfo {
+    #[must_use]
+    pub fn supports_cmpop(op: Cmpop) -> bool {
+        try_get_cmpop_info(op).is_some()
+    }
+}
+
+impl HasOpInfo for Cmpop {
+    fn op_info(&self) -> OpInfo {
+        try_get_cmpop_info(*self).expect("{self:?} is not supported")
+    }
+}
+
+impl HasOpInfo for Binop {
+    fn op_info(&self) -> OpInfo {
+        // Helper macro to generate both the normal variant [`OpInfo`] and the
+        // augmented assigning variant [`OpInfo`] for a binary operator conveniently.
+        macro_rules! info {
+            ($name:literal, $symbol:literal) => {
+                (
+                    make_op_info!($name, $symbol),
+                    make_op_info!(concat!("i", $name), concat!($symbol, "=")),
+                )
+            };
+        }
+
+        let (normal_variant, aug_assign_variant) = match self.base {
+            Operator::Add => info!("add", "+"),
+            Operator::Sub => info!("sub", "-"),
+            Operator::Div => info!("truediv", "/"),
+            Operator::Mod => info!("mod", "%"),
+            Operator::Mult => info!("mul", "*"),
+            Operator::Pow => info!("pow", "**"),
+            Operator::BitOr => info!("or", "|"),
+            Operator::BitXor => info!("xor", "^"),
+            Operator::BitAnd => info!("and", "&"),
+            Operator::LShift => info!("lshift", "<<"),
+            Operator::RShift => info!("rshift", ">>"),
+            Operator::FloorDiv => info!("floordiv", "//"),
+            Operator::MatMult => info!("matmul", "@"),
+        };
+
+        match self.variant {
+            BinopVariant::Normal => normal_variant,
+            BinopVariant::AugAssign => aug_assign_variant,
+        }
+    }
+}
+
+impl HasOpInfo for Unaryop {
+    fn op_info(&self) -> OpInfo {
+        match self {
+            Unaryop::UAdd => make_op_info!("pos", "+"),
+            Unaryop::USub => make_op_info!("neg", "-"),
+            Unaryop::Not => make_op_info!("not", "not"), // i.e., `not False`, so the symbol is just `not`.
+            Unaryop::Invert => make_op_info!("inv", "~"),
+        }
+    }
+}
+
 pub(super) fn with_fields<F>(unifier: &mut Unifier, ty: Type, f: F)
 where
     F: FnOnce(&mut Unifier, &mut HashMap<StrRef, (Type, bool)>),
@@ -115,23 +190,9 @@ pub fn impl_binop(
 
         let ret_ty = ret_ty.unwrap_or_else(|| unifier.get_fresh_var(None, None).ty);
 
-        for op in ops {
-            fields.insert(binop_name(*op).into(), {
-                (
-                    unifier.add_ty(TypeEnum::TFunc(FunSignature {
-                        ret: ret_ty,
-                        vars: function_vars.clone(),
-                        args: vec![FuncArg {
-                            ty: other_ty,
-                            default_value: None,
-                            name: "other".into(),
-                        }],
-                    })),
-                    false,
-                )
-            });
-
-            fields.insert(binop_assign_name(*op).into(), {
+        for (base_op, variant) in iproduct!(ops, [BinopVariant::Normal, BinopVariant::AugAssign]) {
+            let op = Binop { base: *base_op, variant };
+            fields.insert(op.op_info().method_name.into(), {
                 (
                     unifier.add_ty(TypeEnum::TFunc(FunSignature {
                         ret: ret_ty,
@@ -140,6 +201,7 @@ pub fn impl_binop(
                             ty: other_ty,
                             default_value: None,
                             name: "other".into(),
+                            is_vararg: false,
                         }],
                     })),
                     false,
@@ -155,7 +217,7 @@ pub fn impl_unaryop(unifier: &mut Unifier, ty: Type, ret_ty: Option<Type>, ops:
 
         for op in ops {
             fields.insert(
-                unaryop_name(*op).into(),
+                op.op_info().method_name.into(),
                 (
                     unifier.add_ty(TypeEnum::TFunc(FunSignature {
                         ret: ret_ty,
@@ -195,7 +257,7 @@ pub fn impl_cmpop(
 
         for op in ops {
             fields.insert(
-                comparison_name(*op).unwrap().into(),
+                op.op_info().method_name.into(),
                 (
                     unifier.add_ty(TypeEnum::TFunc(FunSignature {
                         ret: ret_ty,
@@ -204,6 +266,7 @@ pub fn impl_cmpop(
                             ty: other_ty,
                             default_value: None,
                             name: "other".into(),
+                            is_vararg: false,
                         }],
                     })),
                     false,
@@ -429,12 +492,29 @@ pub fn typeof_binop(
     lhs: Type,
     rhs: Type,
 ) -> Result<Option<Type>, String> {
+    let op = Binop { base: op, variant: BinopVariant::Normal };
+
+    let is_left_list = lhs.obj_id(unifier).is_some_and(|id| id == PrimDef::List.id());
+    let is_right_list = rhs.obj_id(unifier).is_some_and(|id| id == PrimDef::List.id());
     let is_left_ndarray = lhs.obj_id(unifier).is_some_and(|id| id == PrimDef::NDArray.id());
     let is_right_ndarray = rhs.obj_id(unifier).is_some_and(|id| id == PrimDef::NDArray.id());
 
-    Ok(Some(match op {
+    Ok(Some(match op.base {
         Operator::Add | Operator::Sub | Operator::Mult | Operator::Mod | Operator::FloorDiv => {
-            if is_left_ndarray || is_right_ndarray {
+            if is_left_list || is_right_list {
+                if ![Operator::Add, Operator::Mult].contains(&op.base) {
+                    return Err(format!(
+                        "Binary operator {} not supported for list",
+                        op.op_info().symbol
+                    ));
+                }
+
+                if is_left_list {
+                    lhs
+                } else {
+                    rhs
+                }
+            } else if is_left_ndarray || is_right_ndarray {
                 typeof_ndarray_broadcast(unifier, primitives, lhs, rhs)?
             } else if unifier.unioned(lhs, rhs) {
                 lhs
@@ -444,6 +524,23 @@ pub fn typeof_binop(
         }
 
         Operator::MatMult => {
+            // NOTE: NumPy matmul's LHS and RHS must both be ndarrays. Scalars are not allowed.
+            match (&*unifier.get_ty(lhs), &*unifier.get_ty(rhs)) {
+                (
+                    TypeEnum::TObj { obj_id: lhs_obj_id, .. },
+                    TypeEnum::TObj { obj_id: rhs_obj_id, .. },
+                ) if *lhs_obj_id == primitives.ndarray.obj_id(unifier).unwrap()
+                    && *rhs_obj_id == primitives.ndarray.obj_id(unifier).unwrap() =>
+                {
+                    // LHS and RHS have valid types
+                }
+                _ => {
+                    let lhs_str = unifier.stringify(lhs);
+                    let rhs_str = unifier.stringify(rhs);
+                    return Err(format!("ndarray.__matmul__ only accepts ndarray operands, but left operand has type {lhs_str}, and right operand has type {rhs_str}"));
+                }
+            }
+
             let (_, lhs_ndims) = unpack_ndarray_var_tys(unifier, lhs);
             let lhs_ndims = match &*unifier.get_ty_immutable(lhs_ndims) {
                 TypeEnum::TLiteral { values, .. } => {
@@ -604,6 +701,8 @@ pub fn set_primitives_magic_methods(store: &PrimitiveStore, unifier: &mut Unifie
         bool: bool_t,
         uint32: uint32_t,
         uint64: uint64_t,
+        str: str_t,
+        list: list_t,
         ndarray: ndarray_t,
         ..
     } = *store;
@@ -648,6 +747,14 @@ pub fn set_primitives_magic_methods(store: &PrimitiveStore, unifier: &mut Unifie
     impl_sign(unifier, store, bool_t, Some(int32_t));
     impl_eq(unifier, store, bool_t, &[bool_t, ndarray_bool_t], None);
 
+    /* str ========= */
+    impl_cmpop(unifier, store, str_t, &[str_t], &[Cmpop::Eq, Cmpop::NotEq], Some(bool_t));
+
+    /* list ======== */
+    impl_binop(unifier, store, list_t, &[list_t], Some(list_t), &[Operator::Add]);
+    impl_binop(unifier, store, list_t, &[int32_t, int64_t], Some(list_t), &[Operator::Mult]);
+    impl_cmpop(unifier, store, list_t, &[list_t], &[Cmpop::Eq, Cmpop::NotEq], Some(bool_t));
+
     /* ndarray ===== */
     let ndarray_usized_ndims_tvar =
         unifier.get_fresh_const_generic_var(size_t, Some("ndarray_ndims".into()), None);

nac3core/src/typecheck/type_error.rs

@@ -1,11 +1,13 @@
-use std::collections::HashMap;
-use std::fmt::Display;
+use std::{collections::HashMap, fmt::Display};
 
-use crate::typecheck::typedef::TypeEnum;
-
-use super::typedef::{RecordKey, Type, Unifier};
 use itertools::Itertools;
-use nac3parser::ast::{Location, StrRef};
+
+use nac3parser::ast::{Cmpop, Location, StrRef};
+
+use super::{
+    magic_methods::{Binop, HasOpInfo},
+    typedef::{RecordKey, Type, TypeEnum, Unifier},
+};
 
 #[derive(Debug, Clone)]
 pub enum TypeErrorKind {
@@ -26,6 +28,18 @@ pub enum TypeErrorKind {
         expected: Type,
         got: Type,
     },
+    UnsupportedBinaryOpTypes {
+        operator: Binop,
+        lhs_type: Type,
+        rhs_type: Type,
+        expected_rhs_type: Type,
+    },
+    UnsupportedComparsionOpTypes {
+        operator: Cmpop,
+        lhs_type: Type,
+        rhs_type: Type,
+        expected_rhs_type: Type,
+    },
     FieldUnificationError {
         field: RecordKey,
         types: (Type, Type),
@@ -101,6 +115,26 @@ impl<'a> Display for DisplayTypeError<'a> {
                 let args = missing_arg_names.iter().join(", ");
                 write!(f, "Missing arguments: {args}")
             }
+            UnsupportedBinaryOpTypes { operator, lhs_type, rhs_type, expected_rhs_type } => {
+                let op_symbol = operator.op_info().symbol;
+
+                let lhs_type_str = self.unifier.stringify_with_notes(*lhs_type, &mut notes);
+                let rhs_type_str = self.unifier.stringify_with_notes(*rhs_type, &mut notes);
+                let expected_rhs_type_str =
+                    self.unifier.stringify_with_notes(*expected_rhs_type, &mut notes);
+
+                write!(f, "Unsupported operand type(s) for {op_symbol}: '{lhs_type_str}' and '{rhs_type_str}' (right operand should have type {expected_rhs_type_str})")
+            }
+            UnsupportedComparsionOpTypes { operator, lhs_type, rhs_type, expected_rhs_type } => {
+                let op_symbol = operator.op_info().symbol;
+
+                let lhs_type_str = self.unifier.stringify_with_notes(*lhs_type, &mut notes);
+                let rhs_type_str = self.unifier.stringify_with_notes(*rhs_type, &mut notes);
+                let expected_rhs_type_str =
+                    self.unifier.stringify_with_notes(*expected_rhs_type, &mut notes);
+
+                write!(f, "'{op_symbol}' not supported between instances of '{lhs_type_str}' and '{rhs_type_str}' (right operand should have type {expected_rhs_type_str})")
+            }
             UnknownArgName(name) => {
                 write!(f, "Unknown argument name: {name}")
             }
@@ -148,9 +182,10 @@ impl<'a> Display for DisplayTypeError<'a> {
                         }
                         result
                     }
-                    (TypeEnum::TTuple { ty: ty1 }, TypeEnum::TTuple { ty: ty2 })
-                        if ty1.len() != ty2.len() =>
-                    {
+                    (
+                        TypeEnum::TTuple { ty: ty1, is_vararg_ctx: is_vararg1 },
+                        TypeEnum::TTuple { ty: ty2, is_vararg_ctx: is_vararg2 },
+                    ) if !is_vararg1 && !is_vararg2 && ty1.len() != ty2.len() => {
                         let t1 = self.unifier.stringify_with_notes(*t1, &mut notes);
                         let t2 = self.unifier.stringify_with_notes(*t2, &mut notes);
                         write!(f, "Tuple length mismatch: got {t1} and {t2}")

nac3core/src/typecheck/type_inferencer/test.rs

@@ -1,18 +1,20 @@
-use super::super::{magic_methods::with_fields, typedef::*};
-use super::*;
-use crate::{
-    codegen::CodeGenContext,
-    symbol_resolver::ValueEnum,
-    toplevel::{helper::PrimDef, DefinitionId, TopLevelDef},
-};
+use std::iter::zip;
+
 use indexmap::IndexMap;
 use indoc::indoc;
-use nac3parser::ast::FileName;
-use nac3parser::parser::parse_program;
 use parking_lot::RwLock;
-use std::iter::zip;
 use test_case::test_case;
 
+use nac3parser::{ast::FileName, parser::parse_program};
+
+use super::*;
+use crate::{
+    codegen::{CodeGenContext, CodeGenerator},
+    symbol_resolver::ValueEnum,
+    toplevel::{helper::PrimDef, DefinitionId, TopLevelDef},
+    typecheck::{magic_methods::with_fields, typedef::*},
+};
+
 struct Resolver {
     id_to_type: HashMap<StrRef, Type>,
     id_to_def: HashMap<StrRef, DefinitionId>,
@@ -41,6 +43,7 @@ impl SymbolResolver for Resolver {
         &self,
         _: StrRef,
         _: &mut CodeGenContext<'ctx, '_>,
+        _: &mut dyn CodeGenerator,
     ) -> Option<ValueEnum<'ctx>> {
         unimplemented!()
     }
@@ -83,7 +86,12 @@ impl TestEnvironment {
         });
         with_fields(&mut unifier, int32, |unifier, fields| {
             let add_ty = unifier.add_ty(TypeEnum::TFunc(FunSignature {
-                args: vec![FuncArg { name: "other".into(), ty: int32, default_value: None }],
+                args: vec![FuncArg {
+                    name: "other".into(),
+                    ty: int32,
+                    default_value: None,
+                    is_vararg: false,
+                }],
                 ret: int32,
                 vars: VarMap::new(),
             }));
@@ -139,6 +147,12 @@ impl TestEnvironment {
             fields: HashMap::new(),
             params: VarMap::new(),
         });
+        let list_elem_tvar = unifier.get_fresh_var(Some("list_elem".into()), None);
+        let list = unifier.add_ty(TypeEnum::TObj {
+            obj_id: PrimDef::List.id(),
+            fields: HashMap::new(),
+            params: into_var_map([list_elem_tvar]),
+        });
         let ndarray_dtype_tvar = unifier.get_fresh_var(Some("ndarray_dtype".into()), None);
         let ndarray_ndims_tvar =
             unifier.get_fresh_const_generic_var(uint64, Some("ndarray_ndims".into()), None);
@@ -159,6 +173,7 @@ impl TestEnvironment {
             uint32,
             uint64,
             option,
+            list,
             ndarray,
             size_t: 64,
         };
@@ -217,7 +232,12 @@ impl TestEnvironment {
         });
         with_fields(&mut unifier, int32, |unifier, fields| {
             let add_ty = unifier.add_ty(TypeEnum::TFunc(FunSignature {
-                args: vec![FuncArg { name: "other".into(), ty: int32, default_value: None }],
+                args: vec![FuncArg {
+                    name: "other".into(),
+                    ty: int32,
+                    default_value: None,
+                    is_vararg: false,
+                }],
                 ret: int32,
                 vars: VarMap::new(),
             }));
@@ -273,13 +293,33 @@ impl TestEnvironment {
             fields: HashMap::new(),
             params: VarMap::new(),
         });
+        let list_elem_tvar = unifier.get_fresh_var(Some("list_elem".into()), None);
+        let list = unifier.add_ty(TypeEnum::TObj {
+            obj_id: PrimDef::List.id(),
+            fields: HashMap::new(),
+            params: into_var_map([list_elem_tvar]),
+        });
         let ndarray = unifier.add_ty(TypeEnum::TObj {
             obj_id: PrimDef::NDArray.id(),
             fields: HashMap::new(),
             params: VarMap::new(),
         });
         identifier_mapping.insert("None".into(), none);
-        for (i, name) in ["int32", "int64", "float", "bool", "none", "range", "str", "Exception"]
+        for (i, name) in [
+            "int32",
+            "int64",
+            "float",
+            "bool",
+            "none",
+            "range",
+            "str",
+            "Exception",
+            "uint32",
+            "uint64",
+            "Option",
+            "list",
+            "ndarray",
+        ]
         .iter()
         .enumerate()
         {
@@ -299,7 +339,7 @@ impl TestEnvironment {
                 .into(),
             );
         }
-        let defs = 7;
+        let defs = 12;
 
         let primitives = PrimitiveStore {
             int32,
@@ -313,6 +353,7 @@ impl TestEnvironment {
             uint32,
             uint64,
             option,
+            list,
             ndarray,
             size_t: 64,
         };
@@ -424,6 +465,11 @@ impl TestEnvironment {
             "range".into(),
             "str".into(),
             "exception".into(),
+            "uint32".into(),
+            "uint64".into(),
+            "option".into(),
+            "list".into(),
+            "ndarray".into(),
             "Foo".into(),
             "Bar".into(),
             "Bar2".into(),
@@ -474,7 +520,7 @@ impl TestEnvironment {
             primitives: &mut self.primitives,
             virtual_checks: &mut self.virtual_checks,
             calls: &mut self.calls,
-            defined_identifiers: HashSet::default(),
+            defined_identifiers: HashMap::default(),
             in_handler: false,
         }
     }
@@ -550,8 +596,9 @@ fn test_basic(source: &str, mapping: &HashMap<&str, &str>, virtuals: &[(&str, &s
     println!("source:\n{source}");
     let mut env = TestEnvironment::new();
     let id_to_name = std::mem::take(&mut env.id_to_name);
-    let mut defined_identifiers: HashSet<_> = env.identifier_mapping.keys().copied().collect();
-    defined_identifiers.insert("virtual".into());
+    let mut defined_identifiers: HashMap<_, _> =
+        env.identifier_mapping.keys().copied().map(|id| (id, IdentifierInfo::default())).collect();
+    defined_identifiers.insert("virtual".into(), IdentifierInfo::default());
     let mut inferencer = env.get_inferencer();
     inferencer.defined_identifiers.clone_from(&defined_identifiers);
     let statements = parse_program(source, FileName::default()).unwrap();
@@ -696,8 +743,9 @@ fn test_primitive_magic_methods(source: &str, mapping: &HashMap<&str, &str>) {
     println!("source:\n{source}");
     let mut env = TestEnvironment::basic_test_env();
     let id_to_name = std::mem::take(&mut env.id_to_name);
-    let mut defined_identifiers: HashSet<_> = env.identifier_mapping.keys().copied().collect();
-    defined_identifiers.insert("virtual".into());
+    let mut defined_identifiers: HashMap<_, _> =
+        env.identifier_mapping.keys().copied().map(|id| (id, IdentifierInfo::default())).collect();
+    defined_identifiers.insert("virtual".into(), IdentifierInfo::default());
     let mut inferencer = env.get_inferencer();
     inferencer.defined_identifiers.clone_from(&defined_identifiers);
     let statements = parse_program(source, FileName::default()).unwrap();

nac3core/src/typecheck/typedef/mod.rs

@@ -1,20 +1,28 @@
+use std::{
+    borrow::Cow,
+    cell::RefCell,
+    collections::{HashMap, HashSet},
+    fmt::{self, Display},
+    iter::{repeat, zip},
+    rc::Rc,
+    sync::{Arc, Mutex},
+};
+
 use indexmap::IndexMap;
-use itertools::Itertools;
-use std::cell::RefCell;
-use std::collections::HashMap;
-use std::fmt::{self, Display};
-use std::iter::zip;
-use std::rc::Rc;
-use std::sync::{Arc, Mutex};
-use std::{borrow::Cow, collections::HashSet};
+use itertools::{repeat_n, Itertools};
 
-use nac3parser::ast::{Location, StrRef};
+use nac3parser::ast::{Cmpop, Location, StrRef, Unaryop};
 
-use super::type_error::{TypeError, TypeErrorKind};
-use super::unification_table::{UnificationKey, UnificationTable};
-use crate::symbol_resolver::SymbolValue;
-use crate::toplevel::{DefinitionId, TopLevelContext, TopLevelDef};
-use crate::typecheck::type_inferencer::PrimitiveStore;
+use super::{
+    magic_methods::{Binop, HasOpInfo, OpInfo},
+    type_error::{TypeError, TypeErrorKind},
+    type_inferencer::PrimitiveStore,
+    unification_table::{UnificationKey, UnificationTable},
+};
+use crate::{
+    symbol_resolver::SymbolValue,
+    toplevel::{helper::PrimDef, DefinitionId, TopLevelContext, TopLevelDef},
+};
 
 #[cfg(test)]
 mod test;
@@ -73,6 +81,28 @@ pub fn iter_type_vars(var_map: &VarMap) -> impl Iterator<Item = TypeVar> + '_ {
     var_map.iter().map(|(&id, &ty)| TypeVar { id, ty })
 }
 
+#[derive(Debug, Clone)]
+pub enum OperatorInfo {
+    /// The call was written as an unary operation, e.g., `~a` or `not a`.
+    IsUnaryOp {
+        /// The [`Type`] of the `self` object
+        self_type: Type,
+        operator: Unaryop,
+    },
+    /// The call was written as a binary operation, e.g., `a + b` or `a += b`.
+    IsBinaryOp {
+        /// The [`Type`] of the `self` object
+        self_type: Type,
+        operator: Binop,
+    },
+    /// The call was written as a binary comparison operation, e.g., `a < b`.
+    IsComparisonOp {
+        /// The [`Type`] of the `self` object
+        self_type: Type,
+        operator: Cmpop,
+    },
+}
+
 #[derive(Clone)]
 pub struct Call {
     pub posargs: Vec<Type>,
@@ -80,6 +110,9 @@ pub struct Call {
     pub ret: Type,
     pub fun: RefCell<Option<Type>>,
     pub loc: Option<Location>,
+
+    /// Details about the associated Python user operator expression of this call, if any.
+    pub operator_info: Option<OperatorInfo>,
 }
 
 #[derive(Debug, Clone)]
@@ -87,6 +120,7 @@ pub struct FuncArg {
     pub name: StrRef,
     pub ty: Type,
     pub default_value: Option<SymbolValue>,
+    pub is_vararg: bool,
 }
 
 impl FuncArg {
@@ -205,12 +239,12 @@ pub enum TypeEnum {
     TTuple {
         /// The types of elements present in this tuple.
         ty: Vec<Type>,
-    },
 
-    /// A list type.
-    TList {
-        /// The type of elements present in this list.
-        ty: Type,
+        /// Whether this tuple is used in a vararg context.
+        ///
+        /// If `true`, `ty` must only contain one type, and the tuple is assumed to contain any
+        /// number of `ty`-typed values.
+        is_vararg_ctx: bool,
     },
 
     /// An object type.
@@ -246,7 +280,6 @@ impl TypeEnum {
             TypeEnum::TVar { .. } => "TVar",
             TypeEnum::TLiteral { .. } => "TConstant",
             TypeEnum::TTuple { .. } => "TTuple",
-            TypeEnum::TList { .. } => "TList",
             TypeEnum::TObj { .. } => "TObj",
             TypeEnum::TVirtual { .. } => "TVirtual",
             TypeEnum::TCall { .. } => "TCall",
@@ -482,13 +515,31 @@ impl Unifier {
                     )
                 }
             }
-            TypeEnum::TList { ty } => self
-                .get_instantiations(*ty)
-                .map(|ty| ty.iter().map(|&ty| self.add_ty(TypeEnum::TList { ty })).collect_vec()),
+            TypeEnum::TObj { obj_id, params, .. } if *obj_id == PrimDef::List.id() => {
+                let tv = iter_type_vars(params).nth(0).unwrap();
+
+                let tv_id = if let TypeEnum::TVar { id, .. } =
+                    self.unification_table.probe_value(tv.ty).as_ref()
+                {
+                    *id
+                } else {
+                    tv.id
+                };
+
+                self.get_instantiations(tv.ty).map(|ty_insts| {
+                    ty_insts
+                        .iter()
+                        .map(|&ty_inst| {
+                            self.subst(ty, &into_var_map([TypeVar { id: tv_id, ty: ty_inst }]))
+                                .unwrap_or(ty)
+                        })
+                        .collect()
+                })
+            }
             TypeEnum::TVirtual { ty } => self.get_instantiations(*ty).map(|ty| {
                 ty.iter().map(|&ty| self.add_ty(TypeEnum::TVirtual { ty })).collect_vec()
             }),
-            TypeEnum::TTuple { ty } => {
+            TypeEnum::TTuple { ty, is_vararg_ctx } => {
                 let tuples = ty
                     .iter()
                     .map(|ty| self.get_instantiations(*ty).unwrap_or_else(|| vec![*ty]))
@@ -498,7 +549,12 @@ impl Unifier {
                     None
                 } else {
                     Some(
-                        tuples.into_iter().map(|ty| self.add_ty(TypeEnum::TTuple { ty })).collect(),
+                        tuples
+                            .into_iter()
+                            .map(|ty| {
+                                self.add_ty(TypeEnum::TTuple { ty, is_vararg_ctx: *is_vararg_ctx })
+                            })
+                            .collect(),
                     )
                 }
             }
@@ -541,10 +597,8 @@ impl Unifier {
 
             TVar { .. } => allowed_typevars.iter().any(|b| self.unification_table.unioned(a, *b)),
             TCall { .. } => false,
-            TList { ty }
-            | TVirtual { ty } => self.is_concrete(*ty, allowed_typevars),
-
-            TTuple { ty } => ty.iter().all(|ty| self.is_concrete(*ty, allowed_typevars)),
+            TVirtual { ty } => self.is_concrete(*ty, allowed_typevars),
+            TTuple { ty, .. } => ty.iter().all(|ty| self.is_concrete(*ty, allowed_typevars)),
             TObj { params: vars, .. } => {
                 vars.values().all(|ty| self.is_concrete(*ty, allowed_typevars))
             }
@@ -612,16 +666,74 @@ impl Unifier {
 
         // Get details about the function signature/parameters.
         let num_params = signature.args.len();
+        let is_vararg = signature.args.iter().any(|arg| arg.is_vararg);
 
         // Force the type vars in `b` and `signature' to be up-to-date.
         let b = self.instantiate_fun(b, signature);
         let TypeEnum::TFunc(signature) = &*self.get_ty(b) else { unreachable!() };
 
         // Get details about the input arguments
-        let Call { posargs, kwargs, ret, fun, loc } = call;
+        let Call { posargs, kwargs, ret, fun, loc, operator_info } = call;
         let num_args = posargs.len() + kwargs.len();
 
-        // Now we check the arguments against the parameters
+        // Now we check the arguments against the parameters,
+        // and depending on what `call_info` is, we might change how `unify_call()` behaves
+        // to improve user error messages when type checking fails.
+        match operator_info {
+            Some(OperatorInfo::IsBinaryOp { self_type, operator }) => {
+                // The call is written in the form of (say) `a + b`.
+                // Technically, it is `a.__add__(b)`, and they have the following constraints:
+                assert_eq!(posargs.len(), 1);
+                assert_eq!(kwargs.len(), 0);
+                assert_eq!(num_params, 1);
+
+                let other_type = posargs[0]; // the second operand
+                let expected_other_type = signature.args[0].ty;
+
+                let ok = self.unify_impl(expected_other_type, other_type, false).is_ok();
+                if !ok {
+                    self.restore_snapshot();
+                    return Err(TypeError::new(
+                        TypeErrorKind::UnsupportedBinaryOpTypes {
+                            operator: *operator,
+                            lhs_type: *self_type,
+                            rhs_type: other_type,
+                            expected_rhs_type: expected_other_type,
+                        },
+                        *loc,
+                    ));
+                }
+            }
+            Some(OperatorInfo::IsComparisonOp { self_type, operator })
+                if OpInfo::supports_cmpop(*operator) // Otherwise that comparison operator is not supported.
+                =>
+            {
+                // The call is written in the form of (say) `a <= b`.
+                // Technically, it is `a.__le__(b)`, and they have the following constraints:
+                assert_eq!(posargs.len(), 1);
+                assert_eq!(kwargs.len(), 0);
+                assert_eq!(num_params, 1);
+
+                let other_type = posargs[0]; // the second operand
+                let expected_other_type = signature.args[0].ty;
+
+                let ok = self.unify_impl(expected_other_type, other_type, false).is_ok();
+                if !ok {
+                    self.restore_snapshot();
+                    return Err(TypeError::new(
+                        TypeErrorKind::UnsupportedComparsionOpTypes {
+                            operator: *operator,
+                            lhs_type: *self_type,
+                            rhs_type: other_type,
+                            expected_rhs_type: expected_other_type,
+                        },
+                        *loc,
+                    ));
+                }
+            }
+            _ => {
+                // Handle [`CallInfo::IsNormalFunctionCall`] and other uninteresting variants
+                // of [`CallInfo`] (e.g, `CallInfo::IsUnaryOp` and unsupported comparison operators)
 
                 // Helper lambdas
                 let mut type_check_arg = |param_name, expected_arg_ty, arg_ty| {
@@ -643,7 +755,7 @@ impl Unifier {
                 };
 
                 // Check for "too many arguments"
-        if num_params < posargs.len() {
+                if !is_vararg && num_params < posargs.len() {
                     let expected_min_count =
                         signature.args.iter().filter(|param| param.is_required()).count();
                     let expected_max_count = num_params;
@@ -676,13 +788,29 @@ impl Unifier {
                     type_check_arg(param.name, param.ty, arg_ty)?;
                 }
 
+                if is_vararg {
+                    debug_assert!(!signature.args.is_empty());
+
+                    let vararg_args = posargs.iter().skip(signature.args.len());
+                    let vararg_param = signature.args.last().unwrap();
+
+                    for (&arg_ty, param) in zip(vararg_args, repeat(vararg_param)) {
+                        // `param_info` for this argument would've already been marked as supplied
+                        // during non-vararg posarg typecheck
+                        type_check_arg(param.name, param.ty, arg_ty)?;
+                    }
+                }
+
                 // Now consume all keyword arguments and typecheck them.
                 for (&param_name, &arg_ty) in kwargs {
                     // We will also use this opportunity to check if this keyword argument is "legal".
 
                     let Some(param_info) = param_info_by_name.get_mut(&param_name) else {
                         self.restore_snapshot();
-                return Err(TypeError::new(TypeErrorKind::UnknownArgName(param_name), *loc));
+                        return Err(TypeError::new(
+                            TypeErrorKind::UnknownArgName(param_name),
+                            *loc,
+                        ));
                     };
 
                     if param_info.has_been_supplied {
@@ -707,12 +835,17 @@ impl Unifier {
                 // unsupplied required parameters, and throw an error if they exist.
                 let missing_arg_names = param_info_by_name
                     .values()
-            .filter(|param_info| param_info.param.is_required() && !param_info.has_been_supplied)
+                    .filter(|param_info| {
+                        param_info.param.is_required() && !param_info.has_been_supplied
+                    })
                     .map(|param_info| param_info.param.name)
                     .collect_vec();
                 if !missing_arg_names.is_empty() {
                     self.restore_snapshot();
-            return Err(TypeError::new(TypeErrorKind::MissingArgs { missing_arg_names }, *loc));
+                    return Err(TypeError::new(
+                        TypeErrorKind::MissingArgs { missing_arg_names },
+                        *loc,
+                    ));
                 }
 
                 // Finally, check the Call's return type
@@ -723,6 +856,8 @@ impl Unifier {
                     }
                     err
                 })?;
+            }
+        }
 
         *fun.borrow_mut() = Some(b);
 
@@ -855,7 +990,10 @@ impl Unifier {
                 self.unify_impl(x, b, false)?;
                 self.set_a_to_b(a, x);
             }
-            (TVar { fields: Some(fields), range, is_const_generic: false, .. }, TTuple { ty }) => {
+            (
+                TVar { fields: Some(fields), range, is_const_generic: false, .. },
+                TTuple { ty, .. },
+            ) => {
                 let len = i32::try_from(ty.len()).unwrap();
                 for (k, v) in fields {
                     match *k {
@@ -876,31 +1014,25 @@ impl Unifier {
                             self.unify_impl(v.ty, ty[ind as usize], false)
                                 .map_err(|e| e.at(v.loc))?;
                         }
-                        RecordKey::Str(_) => {
-                            return Err(TypeError::new(TypeErrorKind::NoSuchField(*k, b), v.loc))
-                        }
-                    }
-                }
-                let x = self.check_var_compatibility(b, range)?.unwrap_or(b);
-                self.unify_impl(x, b, false)?;
-                self.set_a_to_b(a, x);
-            }
-            (TVar { fields: Some(fields), range, is_const_generic: false, .. }, TList { ty }) => {
-                for (k, v) in fields {
-                    match *k {
-                        RecordKey::Int(_) => {
-                            self.unify_impl(v.ty, *ty, false).map_err(|e| e.at(v.loc))?;
-                        }
-                        RecordKey::Str(_) => {
-                            return Err(TypeError::new(TypeErrorKind::NoSuchField(*k, b), v.loc))
-                        }
-                    }
-                }
-                let x = self.check_var_compatibility(b, range)?.unwrap_or(b);
-                self.unify_impl(x, b, false)?;
-                self.set_a_to_b(a, x);
-            }
+                        RecordKey::Str(s) => {
+                            let tuple_fns = [
+                                Cmpop::Eq.op_info().method_name,
+                                Cmpop::NotEq.op_info().method_name,
+                            ];
 
+                            if !tuple_fns.into_iter().any(|op| s.to_string() == op) {
+                                return Err(TypeError::new(
+                                    TypeErrorKind::NoSuchField(*k, b),
+                                    v.loc,
+                                ));
+                            }
+                        }
+                    }
+                }
+                let x = self.check_var_compatibility(b, range)?.unwrap_or(b);
+                self.unify_impl(x, b, false)?;
+                self.set_a_to_b(a, x);
+            }
             (
                 TVar { id: id1, range: ty1, is_const_generic: true, .. },
                 TVar { id: id2, range: ty2, .. },
@@ -968,24 +1100,50 @@ impl Unifier {
                 self.set_a_to_b(a, b);
             }
 
-            (TTuple { ty: ty1 }, TTuple { ty: ty2 }) => {
-                if ty1.len() != ty2.len() {
-                    return Err(TypeError::new(TypeErrorKind::IncompatibleTypes(a, b), None));
+            (
+                TTuple { ty: ty1, is_vararg_ctx: is_vararg1 },
+                TTuple { ty: ty2, is_vararg_ctx: is_vararg2 },
+            ) => {
+                // Rules for Tuples:
+                // - ty1: is_vararg && ty2: is_vararg -> ty1[0] == ty2[0]
+                // - ty1: is_vararg && ty2: !is_vararg -> type error (not enough info to infer the correct number of arguments)
+                // - ty1: !is_vararg && ty2: is_vararg -> ty1[..] == ty2[0]
+                // - ty1: !is_vararg && ty2: !is_vararg -> ty1.len() == ty2.len() && ty1[i] == ty2[i]
+
+                debug_assert!(!is_vararg1 || ty1.len() == 1);
+                debug_assert!(!is_vararg2 || ty2.len() == 1);
+
+                match (*is_vararg1, *is_vararg2) {
+                    (true, true) => {
+                        if self.unify_impl(ty1[0], ty2[0], false).is_err() {
+                            return Self::incompatible_types(a, b);
                         }
+                    }
+                    (true, false) => return Self::incompatible_types(a, b),
+
+                    (false, true) => {
+                        for y in ty2 {
+                            if self.unify_impl(ty1[0], *y, false).is_err() {
+                                return Self::incompatible_types(a, b);
+                            }
+                        }
+                    }
+                    (false, false) => {
+                        if ty1.len() != ty2.len() {
+                            return Self::incompatible_types(a, b);
+                        }
+
                         for (x, y) in ty1.iter().zip(ty2.iter()) {
                             if self.unify_impl(*x, *y, false).is_err() {
-                        return Err(TypeError::new(TypeErrorKind::IncompatibleTypes(a, b), None));
+                                return Self::incompatible_types(a, b);
                             }
                         }
+                    }
+                }
+
                 self.set_a_to_b(a, b);
             }
-            (TList { ty: ty1 }, TList { ty: ty2 }) => {
-                if self.unify_impl(*ty1, *ty2, false).is_err() {
-                    return Err(TypeError::new(TypeErrorKind::IncompatibleTypes(a, b), None));
-                }
-                self.set_a_to_b(a, b);
-            }
-            (TVar { fields: Some(map), range, .. }, TObj { fields, .. }) => {
+            (TVar { fields: Some(map), range, .. }, TObj { obj_id, fields, params }) => {
                 for (k, field) in map {
                     match *k {
                         RecordKey::Str(s) => {
@@ -1004,10 +1162,18 @@ impl Unifier {
                             self.unify_impl(field.ty, ty, false).map_err(|v| v.at(field.loc))?;
                         }
                         RecordKey::Int(_) => {
+                            // Allow expressions such as list[0]
+                            if *obj_id == PrimDef::List.id() {
+                                let ty = iter_type_vars(params).nth(0).unwrap().ty;
+
+                                self.unify_impl(field.ty, ty, false)
+                                    .map_err(|e| e.at(field.loc))?;
+                            } else {
                                 return Err(TypeError::new(
                                     TypeErrorKind::NoSuchField(*k, b),
                                     field.loc,
-                            ))
+                                ));
+                            }
                         }
                     }
                 }
@@ -1217,13 +1383,22 @@ impl Unifier {
             TypeEnum::TLiteral { values, .. } => {
                 format!("const({})", values.iter().map(|v| format!("{v:?}")).join(", "))
             }
-            TypeEnum::TTuple { ty } => {
-                let mut fields =
-                    ty.iter().map(|v| self.internal_stringify(*v, obj_to_name, var_to_name, notes));
+            TypeEnum::TTuple { ty, is_vararg_ctx } => {
+                if *is_vararg_ctx {
+                    debug_assert_eq!(ty.len(), 1);
+                    let field = self.internal_stringify(
+                        *ty.iter().next().unwrap(),
+                        obj_to_name,
+                        var_to_name,
+                        notes,
+                    );
+                    format!("tuple[*{field}]")
+                } else {
+                    let mut fields = ty
+                        .iter()
+                        .map(|v| self.internal_stringify(*v, obj_to_name, var_to_name, notes));
                     format!("tuple[{}]", fields.join(", "))
                 }
-            TypeEnum::TList { ty } => {
-                format!("list[{}]", self.internal_stringify(*ty, obj_to_name, var_to_name, notes))
             }
             TypeEnum::TVirtual { ty } => {
                 format!(
@@ -1248,17 +1423,21 @@ impl Unifier {
                     .args
                     .iter()
                     .map(|arg| {
+                        let vararg_prefix = if arg.is_vararg { "*" } else { "" };
+
                         if let Some(dv) = &arg.default_value {
                             format!(
-                                "{}:{}={}",
+                                "{}:{}{}={}",
                                 arg.name,
+                                vararg_prefix,
                                 self.internal_stringify(arg.ty, obj_to_name, var_to_name, notes),
                                 dv
                             )
                         } else {
                             format!(
-                                "{}:{}",
+                                "{}:{}{}",
                                 arg.name,
+                                vararg_prefix,
                                 self.internal_stringify(arg.ty, obj_to_name, var_to_name, notes)
                             )
                         }
@@ -1344,7 +1523,7 @@ impl Unifier {
         match &*ty {
             TypeEnum::TRigidVar { .. } | TypeEnum::TLiteral { .. } => None,
             TypeEnum::TVar { id, .. } => mapping.get(id).copied(),
-            TypeEnum::TTuple { ty } => {
+            TypeEnum::TTuple { ty, is_vararg_ctx } => {
                 let mut new_ty = Cow::from(ty);
                 for (i, t) in ty.iter().enumerate() {
                     if let Some(t1) = self.subst_impl(*t, mapping, cache) {
@@ -1352,14 +1531,14 @@ impl Unifier {
                     }
                 }
                 if matches!(new_ty, Cow::Owned(_)) {
-                    Some(self.add_ty(TypeEnum::TTuple { ty: new_ty.into_owned() }))
+                    Some(self.add_ty(TypeEnum::TTuple {
+                        ty: new_ty.into_owned(),
+                        is_vararg_ctx: *is_vararg_ctx,
+                    }))
                 } else {
                     None
                 }
             }
-            TypeEnum::TList { ty } => {
-                self.subst_impl(*ty, mapping, cache).map(|t| self.add_ty(TypeEnum::TList { ty: t }))
-            }
             TypeEnum::TVirtual { ty } => self
                 .subst_impl(*ty, mapping, cache)
                 .map(|t| self.add_ty(TypeEnum::TVirtual { ty: t })),
@@ -1370,6 +1549,7 @@ impl Unifier {
                 // This is also used to prevent infinite substitution...
                 let need_subst = params.values().any(|v| {
                     let ty = self.unification_table.probe_value(*v);
+                    // TODO(Derppening): #444
                     if let TypeEnum::TVar { id, .. } = ty.as_ref() {
                         mapping.contains_key(id)
                     } else {
@@ -1514,20 +1694,55 @@ impl Unifier {
                 }
             }
             (TVar { range, .. }, _) => self.check_var_compatibility(b, range).or(Err(())),
-            (TTuple { ty: ty1 }, TTuple { ty: ty2 }) if ty1.len() == ty2.len() => {
-                let ty: Vec<_> = zip(ty1.iter(), ty2.iter())
-                    .map(|(a, b)| self.get_intersection(*a, *b))
-                    .try_collect()?;
-                if ty.iter().any(Option::is_some) {
-                    Ok(Some(self.add_ty(TTuple {
-                        ty: zip(ty, ty1.iter()).map(|(a, b)| a.unwrap_or(*b)).collect(),
-                    })))
+            (
+                TTuple { ty: ty1, is_vararg_ctx: is_vararg1 },
+                TTuple { ty: ty2, is_vararg_ctx: is_vararg2 },
+            ) => {
+                if *is_vararg1 && *is_vararg2 {
+                    let isect_ty = self.get_intersection(ty1[0], ty2[0])?;
+                    Ok(isect_ty.map(|ty| self.add_ty(TTuple { ty: vec![ty], is_vararg_ctx: true })))
                 } else {
-                    Ok(None)
+                    let zip_iter: Box<dyn Iterator<Item = (&Type, &Type)>> =
+                        match (*is_vararg1, *is_vararg2) {
+                            (true, _) => Box::new(repeat_n(&ty1[0], ty2.len()).zip(ty2.iter())),
+                            (_, false) => Box::new(ty1.iter().zip(repeat_n(&ty2[0], ty1.len()))),
+                            _ => {
+                                if ty1.len() != ty2.len() {
+                                    return Err(());
+                                }
+
+                                Box::new(ty1.iter().zip(ty2.iter()))
+                            }
+                        };
+
+                    let ty: Vec<_> =
+                        zip_iter.map(|(a, b)| self.get_intersection(*a, *b)).try_collect()?;
+                    Ok(if ty.iter().any(Option::is_some) {
+                        Some(self.add_ty(TTuple {
+                            ty: zip(ty, ty1.iter()).map(|(a, b)| a.unwrap_or(*b)).collect(),
+                            is_vararg_ctx: false,
+                        }))
+                    } else {
+                        None
+                    })
                 }
             }
-            (TList { ty: ty1 }, TList { ty: ty2 }) => {
-                Ok(self.get_intersection(*ty1, *ty2)?.map(|ty| self.add_ty(TList { ty })))
+            // TODO(Derppening): #444
+            (
+                TObj { obj_id: id1, fields, params: params1 },
+                TObj { obj_id: id2, params: params2, .. },
+            ) if *id1 == PrimDef::List.id() && *id2 == PrimDef::List.id() => {
+                let tv_id = iter_type_vars(params1).nth(0).unwrap().id;
+                let ty1 = iter_type_vars(params1).nth(0).unwrap().ty;
+                let ty2 = iter_type_vars(params2).nth(0).unwrap().ty;
+
+                Ok(self.get_intersection(ty1, ty2)?.map(|ty| {
+                    self.add_ty(TObj {
+                        obj_id: *id1,
+                        fields: fields.clone(),
+                        params: into_var_map([TypeVar { id: tv_id, ty }]),
+                    })
+                }))
             }
             (TVirtual { ty: ty1 }, TVirtual { ty: ty2 }) => {
                 Ok(self.get_intersection(*ty1, *ty2)?.map(|ty| self.add_ty(TVirtual { ty })))

nac3core/src/typecheck/typedef/test.rs

@@ -1,10 +1,12 @@
-use super::super::magic_methods::with_fields;
-use super::*;
+use std::collections::HashMap;
+
 use indoc::indoc;
 use itertools::Itertools;
-use std::collections::HashMap;
 use test_case::test_case;
 
+use super::*;
+use crate::typecheck::magic_methods::with_fields;
+
 impl Unifier {
     /// Check whether two types are equal.
     fn eq(&mut self, a: Type, b: Type) -> bool {
@@ -28,14 +30,14 @@ impl Unifier {
                 TypeEnum::TVar { fields: Some(map1), .. },
                 TypeEnum::TVar { fields: Some(map2), .. },
             ) => self.map_eq2(map1, map2),
-            (TypeEnum::TTuple { ty: ty1 }, TypeEnum::TTuple { ty: ty2 }) => {
+            (
+                TypeEnum::TTuple { ty: ty1, is_vararg_ctx: false },
+                TypeEnum::TTuple { ty: ty2, is_vararg_ctx: false },
+            ) => {
                 ty1.len() == ty2.len()
                     && ty1.iter().zip(ty2.iter()).all(|(t1, t2)| self.eq(*t1, *t2))
             }
-            (TypeEnum::TList { ty: ty1 }, TypeEnum::TList { ty: ty2 })
-            | (TypeEnum::TVirtual { ty: ty1 }, TypeEnum::TVirtual { ty: ty2 }) => {
-                self.eq(*ty1, *ty2)
-            }
+            (TypeEnum::TVirtual { ty: ty1 }, TypeEnum::TVirtual { ty: ty2 }) => self.eq(*ty1, *ty2),
             (
                 TypeEnum::TObj { obj_id: id1, params: params1, .. },
                 TypeEnum::TObj { obj_id: id2, params: params2, .. },
@@ -119,6 +121,15 @@ impl TestEnvironment {
                 params: into_var_map([tvar]),
             }),
         );
+        let tvar = unifier.get_dummy_var();
+        type_mapping.insert(
+            "list".into(),
+            unifier.add_ty(TypeEnum::TObj {
+                obj_id: PrimDef::List.id(),
+                fields: HashMap::new(),
+                params: into_var_map([tvar]),
+            }),
+        );
 
         TestEnvironment { unifier, type_mapping }
     }
@@ -133,6 +144,36 @@ impl TestEnvironment {
         // for testing only, so we can just panic when the input is malformed
         let end = typ.find(|c| ['[', ',', ']', '='].contains(&c)).unwrap_or(typ.len());
         match &typ[..end] {
+            "list" => {
+                let mut s = &typ[end..];
+                assert_eq!(&s[0..1], "[");
+                let mut ty = Vec::new();
+                while &s[0..1] != "]" {
+                    let result = self.internal_parse(&s[1..], mapping);
+                    ty.push(result.0);
+                    s = result.1;
+                }
+
+                assert_eq!(ty.len(), 1);
+
+                let list_elem_tvar = if let TypeEnum::TObj { params, .. } =
+                    &*self.unifier.get_ty_immutable(self.type_mapping["list"])
+                {
+                    iter_type_vars(params).next().unwrap()
+                } else {
+                    unreachable!()
+                };
+
+                (
+                    self.unifier
+                        .subst(
+                            self.type_mapping["list"],
+                            &into_var_map([TypeVar { id: list_elem_tvar.id, ty: ty[0] }]),
+                        )
+                        .unwrap(),
+                    &s[1..],
+                )
+            }
             "tuple" => {
                 let mut s = &typ[end..];
                 assert_eq!(&s[0..1], "[");
@@ -142,13 +183,7 @@ impl TestEnvironment {
                     ty.push(result.0);
                     s = result.1;
                 }
-                (self.unifier.add_ty(TypeEnum::TTuple { ty }), &s[1..])
-            }
-            "list" => {
-                assert_eq!(&typ[end..=end], "[");
-                let (ty, s) = self.internal_parse(&typ[end + 1..], mapping);
-                assert_eq!(&s[0..1], "]");
-                (self.unifier.add_ty(TypeEnum::TList { ty }), &s[1..])
+                (self.unifier.add_ty(TypeEnum::TTuple { ty, is_vararg_ctx: false }), &s[1..])
             }
             "Record" => {
                 let mut s = &typ[end..];
@@ -274,7 +309,7 @@ fn test_unify(
         ("v1", "tuple[int]"),
         ("v2", "list[int]"),
     ],
-    (("v1", "v2"), "Incompatible types: list[0] and tuple[0]")
+    (("v1", "v2"), "Incompatible types: 11[0] and tuple[0]")
     ; "type mismatch"
 )]
 #[test_case(2,
@@ -298,7 +333,7 @@ fn test_unify(
         ("v1", "Record[a=float,b=int]"),
         ("v2", "Foo[v3]"),
     ],
-    (("v1", "v2"), "`3[typevar4]::b` field/method does not exist")
+    (("v1", "v2"), "`3[typevar5]::b` field/method does not exist")
     ; "record obj merge"
 )]
 /// Test cases for invalid unifications.
@@ -388,6 +423,14 @@ fn test_typevar_range() {
     let int_list = env.parse("list[int]", &HashMap::new());
     let float_list = env.parse("list[float]", &HashMap::new());
 
+    let list_elem_tvar = if let TypeEnum::TObj { params, .. } =
+        &*env.unifier.get_ty_immutable(env.type_mapping["list"])
+    {
+        iter_type_vars(params).next().unwrap()
+    } else {
+        unreachable!()
+    };
+
     // unification between v and int
     // where v in (int, bool)
     let v = env.unifier.get_fresh_var_with_range(&[int, boolean], None, None).ty;
@@ -398,7 +441,7 @@ fn test_typevar_range() {
     let v = env.unifier.get_fresh_var_with_range(&[int, boolean], None, None).ty;
     assert_eq!(
         env.unify(int_list, v),
-        Err("Expected any one of these types: 0, 2, but got list[0]".to_string())
+        Err("Expected any one of these types: 0, 2, but got 11[0]".to_string())
     );
 
     // unification between v and float
@@ -410,7 +453,11 @@ fn test_typevar_range() {
     );
 
     let v1 = env.unifier.get_fresh_var_with_range(&[int, boolean], None, None).ty;
-    let v1_list = env.unifier.add_ty(TypeEnum::TList { ty: v1 });
+    let v1_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: v1 }]),
+    });
     let v = env.unifier.get_fresh_var_with_range(&[int, v1_list], None, None).ty;
     // unification between v and int
     // where v in (int, list[v1]), v1 in (int, bool)
@@ -424,9 +471,10 @@ fn test_typevar_range() {
     let v = env.unifier.get_fresh_var_with_range(&[int, v1_list], None, None).ty;
     // unification between v and list[float]
     // where v in (int, list[v1]), v1 in (int, bool)
+    println!("float_list: {}, v: {}", env.unifier.stringify(float_list), env.unifier.stringify(v));
     assert_eq!(
         env.unify(float_list, v),
-        Err("Expected any one of these types: 0, list[typevar5], but got list[1]\n\nNotes:\n    typevar5 ∈ {0, 2}".to_string())
+        Err("Expected any one of these types: 0, 11[typevar6], but got 11[1]\n\nNotes:\n    typevar6 ∈ {0, 2}".to_string())
     );
 
     let a = env.unifier.get_fresh_var_with_range(&[int, float], None, None).ty;
@@ -441,34 +489,66 @@ fn test_typevar_range() {
 
     let a = env.unifier.get_fresh_var_with_range(&[int, float], None, None).ty;
     let b = env.unifier.get_fresh_var_with_range(&[boolean, float], None, None).ty;
-    let a_list = env.unifier.add_ty(TypeEnum::TList { ty: a });
+    let a_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: a }]),
+    });
     let a_list = env.unifier.get_fresh_var_with_range(&[a_list], None, None).ty;
-    let b_list = env.unifier.add_ty(TypeEnum::TList { ty: b });
+    let b_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: b }]),
+    });
     let b_list = env.unifier.get_fresh_var_with_range(&[b_list], None, None).ty;
     env.unifier.unify(a_list, b_list).unwrap();
-    let float_list = env.unifier.add_ty(TypeEnum::TList { ty: float });
+    let float_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: float }]),
+    });
     env.unifier.unify(a_list, float_list).unwrap();
     // previous unifications should not affect a and b
     env.unifier.unify(a, int).unwrap();
 
     let a = env.unifier.get_fresh_var_with_range(&[int, float], None, None).ty;
     let b = env.unifier.get_fresh_var_with_range(&[boolean, float], None, None).ty;
-    let a_list = env.unifier.add_ty(TypeEnum::TList { ty: a });
-    let b_list = env.unifier.add_ty(TypeEnum::TList { ty: b });
+    let a_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: a }]),
+    });
+    let b_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: b }]),
+    });
     env.unifier.unify(a_list, b_list).unwrap();
-    let int_list = env.unifier.add_ty(TypeEnum::TList { ty: int });
+    let int_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: int }]),
+    });
     assert_eq!(
         env.unify(a_list, int_list),
-        Err("Incompatible types: list[typevar22] and list[0]\
-            \n\nNotes:\n    typevar22 ∈ {1}"
+        Err("Incompatible types: 11[typevar23] and 11[0]\
+            \n\nNotes:\n    typevar23 ∈ {1}"
             .into())
     );
 
     let a = env.unifier.get_fresh_var_with_range(&[int, float], None, None).ty;
     let b = env.unifier.get_dummy_var().ty;
-    let a_list = env.unifier.add_ty(TypeEnum::TList { ty: a });
+    let a_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: a }]),
+    });
     let a_list = env.unifier.get_fresh_var_with_range(&[a_list], None, None).ty;
-    let b_list = env.unifier.add_ty(TypeEnum::TList { ty: b });
+    let b_list = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: b }]),
+    });
     env.unifier.unify(a_list, b_list).unwrap();
     assert_eq!(
         env.unify(b, boolean),
@@ -482,16 +562,25 @@ fn test_rigid_var() {
     let a = env.unifier.get_fresh_rigid_var(None, None).ty;
     let b = env.unifier.get_fresh_rigid_var(None, None).ty;
     let x = env.unifier.get_dummy_var().ty;
-    let list_a = env.unifier.add_ty(TypeEnum::TList { ty: a });
-    let list_x = env.unifier.add_ty(TypeEnum::TList { ty: x });
+    let list_elem_tvar = env.unifier.get_fresh_var(Some("list_elem".into()), None);
+    let list_a = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: a }]),
+    });
+    let list_x = env.unifier.add_ty(TypeEnum::TObj {
+        obj_id: env.type_mapping["list"].obj_id(&env.unifier).unwrap(),
+        fields: Mapping::default(),
+        params: into_var_map([TypeVar { id: list_elem_tvar.id, ty: x }]),
+    });
     let int = env.parse("int", &HashMap::new());
     let list_int = env.parse("list[int]", &HashMap::new());
 
-    assert_eq!(env.unify(a, b), Err("Incompatible types: typevar3 and typevar2".to_string()));
+    assert_eq!(env.unify(a, b), Err("Incompatible types: typevar4 and typevar3".to_string()));
     env.unifier.unify(list_a, list_x).unwrap();
     assert_eq!(
         env.unify(list_x, list_int),
-        Err("Incompatible types: list[typevar2] and list[0]".to_string())
+        Err("Incompatible types: 11[typevar3] and 11[0]".to_string())
     );
 
     env.unifier.replace_rigid_var(a, int);
@@ -506,14 +595,25 @@ fn test_instantiation() {
     let float = env.parse("float", &HashMap::new());
     let list_int = env.parse("list[int]", &HashMap::new());
 
-    let obj_map: HashMap<_, _> = [(0usize, "int"), (1, "float"), (2, "bool")].into();
+    let list_elem_tvar = if let TypeEnum::TObj { params, .. } =
+        &*env.unifier.get_ty_immutable(env.type_mapping["list"])
+    {
+        iter_type_vars(params).next().unwrap()
+    } else {
+        unreachable!()
+    };
+
+    let obj_map: HashMap<_, _> = [(0usize, "int"), (1, "float"), (2, "bool"), (11, "list")].into();
 
     let v = env.unifier.get_fresh_var_with_range(&[int, boolean], None, None).ty;
-    let list_v = env.unifier.add_ty(TypeEnum::TList { ty: v });
+    let list_v = env
+        .unifier
+        .subst(env.type_mapping["list"], &into_var_map([TypeVar { id: list_elem_tvar.id, ty: v }]))
+        .unwrap();
     let v1 = env.unifier.get_fresh_var_with_range(&[list_v, int], None, None).ty;
     let v2 = env.unifier.get_fresh_var_with_range(&[list_int, float], None, None).ty;
     let t = env.unifier.get_dummy_var().ty;
-    let tuple = env.unifier.add_ty(TypeEnum::TTuple { ty: vec![v, v1, v2] });
+    let tuple = env.unifier.add_ty(TypeEnum::TTuple { ty: vec![v, v1, v2], is_vararg_ctx: false });
     let v3 = env.unifier.get_fresh_var_with_range(&[tuple, t], None, None).ty;
     // t = TypeVar('t')
     // v = TypeVar('v', int, bool)
@@ -536,7 +636,7 @@ fn test_instantiation() {
         tuple[int, list[bool], list[int]]
         tuple[int, list[int], float]
         tuple[int, list[int], list[int]]
-        v5"
+        v6"
     }
     .split('\n')
     .collect_vec();

nac3ld/src/dwarf.rs

@@ -238,7 +238,7 @@ impl<'a> EH_Frame<'a> {
 /// From the [specification](https://refspecs.linuxfoundation.org/LSB_5.0.0/LSB-Core-generic/LSB-Core-generic/ehframechpt.html):
 ///
 /// > Each CFI record contains a Common Information Entry (CIE) record followed by 1 or more Frame
-/// Description Entry (FDE) records.
+/// > Description Entry (FDE) records.
 pub struct CFI_Record<'a> {
     // It refers to the augmentation data that corresponds to 'R' in the augmentation string
     fde_pointer_encoding: u8,

nac3ld/src/lib.rs

@@ -1,10 +1,4 @@
-#![deny(
-    future_incompatible,
-    let_underscore,
-    nonstandard_style,
-    rust_2024_compatibility,
-    clippy::all
-)]
+#![deny(future_incompatible, let_underscore, nonstandard_style, clippy::all)]
 #![warn(clippy::pedantic)]
 #![allow(
     clippy::cast_possible_truncation,
@@ -21,13 +15,12 @@
     clippy::wildcard_imports
 )]
 
+use std::{collections::HashMap, mem, ptr, slice, str};
+
+use byteorder::{ByteOrder, LittleEndian};
+
 use dwarf::*;
 use elf::*;
-use std::collections::HashMap;
-use std::{mem, ptr, slice, str};
-
-extern crate byteorder;
-use byteorder::{ByteOrder, LittleEndian};
 
 mod dwarf;
 mod elf;

nac3parser/Cargo.toml

@@ -8,15 +8,15 @@ license = "MIT"
 edition = "2021"
 
 [build-dependencies]
-lalrpop = "0.20"
+lalrpop = "0.22"
 
 [dependencies]
 nac3ast = { path = "../nac3ast" }
-lalrpop-util = "0.20"
+lalrpop-util = "0.22"
 log = "0.4"
 unic-emoji-char = "0.9"
 unic-ucd-ident  = "0.9"
-unicode_names2 = "1.2"
+unicode_names2 = "1.3"
 phf = { version = "0.11", features = ["macros"] }
 ahash = "0.8"
 

nac3parser/src/config_comment_helper.rs

@@ -1,8 +1,10 @@
-use crate::ast::Ident;
-use crate::ast::Location;
-use crate::error::*;
-use crate::token::Tok;
+use crate::{
+    ast::{Ident, Location},
+    error::*,
+    token::Tok,
+};
 use lalrpop_util::ParseError;
+
 use nac3ast::*;
 
 pub fn make_config_comment(

nac3parser/src/error.rs

@@ -1,13 +1,12 @@
 //! Define internal parse error types
 //! The goal is to provide a matching and a safe error API, maksing errors from LALR
-use lalrpop_util::ParseError as LalrpopError;
-
-use crate::ast::Location;
-use crate::token::Tok;
-
 use std::error::Error;
 use std::fmt;
 
+use lalrpop_util::ParseError as LalrpopError;
+
+use crate::{ast::Location, token::Tok};
+
 /// Represents an error during lexical scanning.
 #[derive(Debug, PartialEq)]
 pub struct LexicalError {

nac3parser/src/fstring.rs

@@ -1,12 +1,11 @@
-use std::iter;
-use std::mem;
-use std::str;
-
-use crate::ast::{Constant, ConversionFlag, Expr, ExprKind, Location};
-use crate::error::{FStringError, FStringErrorType, ParseError};
-use crate::parser::parse_expression;
+use std::{iter, mem, str};
 
 use self::FStringErrorType::*;
+use crate::{
+    ast::{Constant, ConversionFlag, Expr, ExprKind, Location},
+    error::{FStringError, FStringErrorType, ParseError},
+    parser::parse_expression,
+};
 
 struct FStringParser<'a> {
     chars: iter::Peekable<str::Chars<'a>>,

nac3parser/src/function.rs

@@ -1,8 +1,11 @@
-use ahash::RandomState;
 use std::collections::HashSet;
 
-use crate::ast;
-use crate::error::{LexicalError, LexicalErrorType};
+use ahash::RandomState;
+
+use crate::{
+    ast,
+    error::{LexicalError, LexicalErrorType},
+};
 
 pub struct ArgumentList {
     pub args: Vec<ast::Expr>,

nac3parser/src/lexer.rs

@@ -1,17 +1,17 @@
 //! This module takes care of lexing python source text.
 //!
 //! This means source code is translated into separate tokens.
+use std::{char, cmp::Ordering, num::IntErrorKind, str::FromStr};
 
-pub use super::token::Tok;
-use crate::ast::{FileName, Location};
-use crate::error::{LexicalError, LexicalErrorType};
-use std::char;
-use std::cmp::Ordering;
-use std::num::IntErrorKind;
-use std::str::FromStr;
 use unic_emoji_char::is_emoji_presentation;
 use unic_ucd_ident::{is_xid_continue, is_xid_start};
 
+pub use super::token::Tok;
+use crate::{
+    ast::{FileName, Location},
+    error::{LexicalError, LexicalErrorType},
+};
+
 #[derive(Clone, Copy, PartialEq, Debug, Default)]
 struct IndentationLevel {
     tabs: usize,

nac3parser/src/lib.rs

@@ -15,13 +15,7 @@
 //!
 //! ```
 
-#![deny(
-    future_incompatible,
-    let_underscore,
-    nonstandard_style,
-    rust_2024_compatibility,
-    clippy::all
-)]
+#![deny(future_incompatible, let_underscore, nonstandard_style, clippy::all)]
 #![warn(clippy::pedantic)]
 #![allow(
     clippy::enum_glob_use,
@@ -49,11 +43,11 @@ lalrpop_mod!(
         future_incompatible,
         let_underscore,
         nonstandard_style,
-        rust_2024_compatibility,
         unused,
         clippy::all,
         clippy::pedantic
     )]
+    #[warn(rust_2024_compatibility)]
     python
 );
 pub mod config_comment_helper;

nac3parser/src/parser.rs

@@ -5,14 +5,16 @@
 //! parse a whole program, a single statement, or a single
 //! expression.
 
-use nac3ast::Location;
 use std::iter;
 
-use crate::ast::{self, FileName};
-use crate::error::ParseError;
-use crate::lexer;
+use nac3ast::Location;
+
 pub use crate::mode::Mode;
-use crate::python;
+use crate::{
+    ast::{self, FileName},
+    error::ParseError,
+    lexer, python,
+};
 
 /*
  * Parse python code.

nac3parser/src/token.rs

@@ -1,8 +1,9 @@
 //! Different token definitions.
 //! Loosely based on token.h from CPython source:
-use crate::ast;
 use std::fmt::{self, Write};
 
+use crate::ast;
+
 /// Python source code can be tokenized in a sequence of these tokens.
 #[derive(Clone, Debug, PartialEq)]
 pub enum Tok {

nac3standalone/Cargo.toml

@@ -4,16 +4,13 @@ version = "0.1.0"
 authors = ["M-Labs"]
 edition = "2021"
 
+[features]
+no-escape-analysis = ["nac3core/no-escape-analysis"]
+
 [dependencies]
 parking_lot = "0.12"
-nac3parser = { path = "../nac3parser" }
 nac3core = { path = "../nac3core" }
 
 [dependencies.clap]
 version = "4.5"
 features = ["derive"]
-
-[dependencies.inkwell]
-version = "0.4"
-default-features = false
-features = ["llvm14-0", "target-x86", "target-arm", "target-riscv", "no-libffi-linking"]

nac3standalone/demo/check_demo.sh

@@ -3,23 +3,66 @@
 set -e
 
 if [ -z "$1" ]; then
-  echo "Requires at least one argument"
+    echo "No argument supplied"
     exit 1
 fi
 
 declare -a nac3args
+while [ $# -gt 1 ]; do
+  case "$1" in
+    --help)
+      echo "Usage: check_demo.sh [--debug] [-i686] -- [NAC3ARGS...] demo"
+      exit
+      ;;
+    --debug)
+      debug=1
+      ;;
+    -i686)
+      i686=1
+      ;;
+    --)
+      shift
+      break
+      ;;
+    *)
+      echo "Unrecognized argument \"$1\""
+      exit 1
+      ;;
+  esac
+  shift
+done
+
 while [ $# -gt 1 ]; do
   nac3args+=("$1")
   shift
 done
 demo="$1"
 
-echo -n "Checking $demo... "
-./interpret_demo.py "$demo" > interpreted.log
-./run_demo.sh --out run.log "${nac3args[@]}" "$demo"
-./run_demo.sh --lli --out run_lli.log "${nac3args[@]}" "$demo"
-diff -Nau interpreted.log run.log
-diff -Nau interpreted.log run_lli.log
-echo "ok"
 
-rm -f interpreted.log run.log run_lli.log
+echo "### Checking $demo..."
+
+echo ">>>>>> Running $demo with the Python interpreter"
+./interpret_demo.py "$demo" > interpreted.log
+
+if [ -n "$i686" ]; then
+  echo "...... Trying NAC3's 32-bit code generator output"
+  if [ -n "$debug" ]; then
+    ./run_demo.sh --debug -i686 --out run_32.log -- "${nac3args[@]}" "$demo"
+  else
+    ./run_demo.sh -i686 --out run_32.log -- "${nac3args[@]}" "$demo"
+  fi
+  diff -Nau interpreted.log run_32.log
+fi
+
+echo "...... Trying NAC3's 64-bit code generator output"
+if [ -n "$debug" ]; then
+  ./run_demo.sh --debug --out run_64.log -- "${nac3args[@]}" "$demo"
+else
+  ./run_demo.sh --out run_64.log -- "${nac3args[@]}" "$demo"
+fi
+diff -Nau interpreted.log run_64.log
+
+echo "...... OK"
+
+rm -f interpreted.log \
+  run_32.log run_64.log