Ptr instance offset_const can be negative + use i64_type()

To explicitly set the C++ variant and avoid inconsistencies.

.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

nac3artiq/src/lib.rs

@@ -557,6 +557,10 @@ impl Nac3 {
             .register_top_level(synthesized.pop().unwrap(), Some(resolver.clone()), "", false)
             .unwrap();
 
+        // Process IRRT
+        let context = inkwell::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();
@@ -727,7 +731,7 @@ impl Nac3 {
             membuffer.lock().push(buffer);
         });
 
-        let context = inkwell::context::Context::create();
+        // Link all modules into `main`.
         let buffers = membuffers.lock();
         let main = context
             .create_module_from_ir(MemoryBuffer::create_from_memory_range(&buffers[0], "main"))
@@ -756,8 +760,7 @@ impl Nac3 {
             )
             .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 {

nac3core/build.rs

@@ -8,37 +8,50 @@ use std::{
 };
 
 fn main() {
-    const FILE: &str = "src/codegen/irrt/irrt.cpp";
+    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",
         "-fno-exceptions",
         "-fno-rtti",
-        match env::var("PROFILE").as_deref() {
-            Ok("debug") => "-O0",
-            Ok("release") => "-O3",
-            flavor => panic!("Unknown or missing build flavor {flavor:?}"),
-        },
         "-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()
@@ -52,7 +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]);
@@ -63,18 +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 +1,6 @@
+#include <irrt/exception.hpp>
+#include <irrt/int_types.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 {
+    uint8_t* 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 +1,82 @@
+#pragma once
+
+#include <irrt/cslice.hpp>
+#include <irrt/int_types.hpp>
+
+/**
+ * @brief The int type of ARTIQ exception IDs.
+ */
+typedef int32_t ExceptionId;
+
+/*
+ * 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<const uint8_t*>(filename), .len = __builtin_strlen(filename)},
+        .line = line,
+        .column = 0,
+        .function = {.base = reinterpret_cast<const uint8_t*>(function), .len = __builtin_strlen(function)},
+        .msg = {.base = reinterpret_cast<const uint8_t*>(msg), .len = __builtin_strlen(msg)},
+    };
+    e.params[0] = param0;
+    e.params[1] = param1;
+    e.params[2] = param2;
+    __nac3_raise(reinterpret_cast<void*>(&e));
+    __builtin_unreachable();
+}
+
+/**
+ * @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)
+}  // namespace

nac3core/irrt/irrt/int_types.hpp

@@ -0,0 +1,13 @@
+#pragma once
+
+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);
+
+// 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 +1,75 @@
+#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,
+                                             uint8_t* dest_arr,
+                                             SliceIndex dest_arr_len,
+                                             SliceIndex src_start,
+                                             SliceIndex src_end,
+                                             SliceIndex src_step,
+                                             uint8_t* 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(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 = reinterpret_cast<uint8_t*>(__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(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;
+}
+}  // 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);
+}
+}

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);
+}
+}

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;
+    }
+}
+}

nac3core/src/codegen/irrt/irrt.cpp

@@ -1,414 +0,0 @@
-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);
-
-// 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;
-
-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;
-}
-
-// 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;
-}
-
-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" {
-#define DEF_nac3_int_exp_(T) \
-    T __nac3_int_exp_##T(T base, T exp) {\
-        return __nac3_int_exp_impl(base, exp);\
-    }
-
-DEF_nac3_int_exp_(int32_t)
-DEF_nac3_int_exp_(int64_t)
-DEF_nac3_int_exp_(uint32_t)
-DEF_nac3_int_exp_(uint64_t)
-
-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;
-    }
-}
-
-// 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,
-    uint8_t* dest_arr,
-    SliceIndex dest_arr_len,
-    SliceIndex src_start,
-    SliceIndex src_end,
-    SliceIndex src_step,
-    uint8_t* 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(
-                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 = reinterpret_cast<uint8_t *>(__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(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 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);
-}
-}  // extern "C"

nac3core/src/codegen/irrt/mod.rs

@@ -1,4 +1,4 @@
-use crate::typecheck::typedef::Type;
+use crate::{symbol_resolver::SymbolResolver, typecheck::typedef::Type};
 
 use super::{
     classes::{
@@ -15,14 +15,14 @@ use inkwell::{
     memory_buffer::MemoryBuffer,
     module::Module,
     types::{BasicTypeEnum, IntType},
-    values::{BasicValueEnum, CallSiteValue, FloatValue, IntValue},
+    values::{BasicValue, BasicValueEnum, CallSiteValue, FloatValue, IntValue},
     AddressSpace, IntPredicate,
 };
 use itertools::Either;
 use nac3parser::ast::Expr;
 
 #[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",
@@ -38,6 +38,25 @@ 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
 }
 

nac3core/src/codegen/mod.rs

@@ -41,6 +41,7 @@ pub mod extern_fns;
 mod generator;
 pub mod irrt;
 pub mod llvm_intrinsics;
+pub mod model;
 pub mod numpy;
 pub mod stmt;
 

nac3core/src/codegen/model/any.rs

@@ -0,0 +1,37 @@
+use inkwell::{
+    context::Context,
+    types::{BasicType, BasicTypeEnum, IntType},
+    values::IntValue,
+};
+
+use super::*;
+
+/// A [`Model`] of any [`BasicTypeEnum`].
+///
+/// Use this when it is infeasible to use model abstractions.
+#[derive(Debug, Clone, Copy)]
+pub struct Any<'ctx>(pub BasicTypeEnum<'ctx>);
+
+impl<'ctx> ModelBase<'ctx> for Any<'ctx> {
+    fn llvm_type_impl(&self, _size_t: IntType<'ctx>, _ctx: &'ctx Context) -> BasicTypeEnum<'ctx> {
+        self.0.as_basic_type_enum()
+    }
+
+    fn check_type_impl(
+        &self,
+        _size_t: IntType<'ctx>,
+        _ctx: &'ctx Context,
+        ty: BasicTypeEnum<'ctx>,
+    ) -> Result<(), ModelError> {
+        if ty == self.0 {
+            Ok(())
+        } else {
+            Err(ModelError(format!("Expecting {}, but got {}", self.0, ty)))
+        }
+    }
+}
+
+impl<'ctx> Model<'ctx> for Any<'ctx> {
+    type Type = IntType<'ctx>;
+    type Value = IntValue<'ctx>;
+}

nac3core/src/codegen/model/array.rs

@@ -0,0 +1,152 @@
+use std::{fmt, marker::PhantomData};
+
+use inkwell::{
+    context::Context,
+    types::{ArrayType, BasicType, BasicTypeEnum, IntType},
+    values::{ArrayValue, IntValue},
+};
+
+use crate::codegen::{CodeGenContext, CodeGenerator};
+
+use super::*;
+
+/// Trait for Rust structs identifying length values for [`Array`].
+pub trait ArrayLen: fmt::Debug + Clone + Copy {
+    fn length(&self) -> u32;
+}
+
+/// A statically known length.
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Len<const N: u32>;
+
+/// A dynamically known length.
+#[derive(Debug, Clone, Copy)]
+pub struct AnyLen(pub u32);
+
+impl<const N: u32> ArrayLen for Len<N> {
+    fn length(&self) -> u32 {
+        N
+    }
+}
+
+impl ArrayLen for AnyLen {
+    fn length(&self) -> u32 {
+        self.0
+    }
+}
+
+/// A Model for an [`ArrayType`].
+///
+/// `Len` should be of a [`LenKind`] and `Item` should be a of [`Model`].
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Array<'ctx, Len: ArrayLen, Item: ModelBase<'ctx>> {
+    /// Length of this array.
+    pub len: Len,
+    /// [`Model`] of the array items.
+    pub item: Item,
+    pub _phantom: PhantomData<&'ctx ()>,
+}
+
+impl<'ctx, Len: ArrayLen, Item: ModelBase<'ctx>> Array<'ctx, Len, Item> {
+    pub fn new(len: Len, item: Item) -> Self {
+        Array { len, item, _phantom: PhantomData }
+    }
+}
+
+impl<'ctx, Len: ArrayLen, Item: ModelBase<'ctx>> ModelBase<'ctx> for Array<'ctx, Len, Item> {
+    fn llvm_type_impl(&self, size_t: IntType<'ctx>, ctx: &'ctx Context) -> BasicTypeEnum<'ctx> {
+        let item = self.item.llvm_type_impl(size_t, ctx);
+        item.array_type(self.len.length()).into()
+    }
+
+    fn check_type_impl(
+        &self,
+        size_t: IntType<'ctx>,
+        ctx: &'ctx Context,
+        ty: BasicTypeEnum<'ctx>,
+    ) -> Result<(), ModelError> {
+        let BasicTypeEnum::ArrayType(ty) = ty else {
+            return Err(ModelError(format!("Expecting ArrayType, but got {ty:?}")));
+        };
+
+        if ty.len() != self.len.length() {
+            return Err(ModelError(format!(
+                "Expecting ArrayType with size {}, but got an ArrayType with size {}",
+                ty.len(),
+                self.len.length()
+            )));
+        }
+
+        self.item
+            .check_type_impl(size_t, ctx, ty.get_element_type())
+            .map_err(|err| err.under_context("an ArrayType"))?;
+
+        Ok(())
+    }
+}
+
+impl<'ctx, Len: ArrayLen, Item: Model<'ctx>> Model<'ctx> for Array<'ctx, Len, Item> {
+    type Type = ArrayType<'ctx>;
+    type Value = ArrayValue<'ctx>;
+}
+
+impl<'ctx, Len: ArrayLen, Item: Model<'ctx>> Instance<'ctx, Ptr<Array<'ctx, Len, Item>>> {
+    /// Get the pointer to the `i`-th (0-based) array element.
+    pub fn gep(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        i: IntValue<'ctx>,
+    ) -> Instance<'ctx, Ptr<Item>> {
+        let zero = ctx.ctx.i64_type().const_zero();
+        let ptr = unsafe { ctx.builder.build_in_bounds_gep(self.value, &[zero, i], "").unwrap() };
+
+        unsafe { Ptr::new(self.model.item.item).believe_value(ptr) }
+    }
+
+    /// Like `gep` but `i` is a constant.
+    pub fn gep_const(&self, ctx: &CodeGenContext<'ctx, '_>, i: u64) -> Instance<'ctx, Ptr<Item>> {
+        assert!(
+            i < u64::from(self.model.item.len.length()),
+            "Index {i} is out of bounds. Array length = {}",
+            self.model.item.len.length()
+        );
+
+        let i = ctx.ctx.i64_type().const_int(i, true);
+        self.gep(ctx, i)
+    }
+
+    /// Convenience function equivalent to `.gep(...).load(...)`.
+    pub fn get<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        i: IntValue<'ctx>,
+    ) -> Instance<'ctx, Item> {
+        self.gep(ctx, i).load(generator, ctx)
+    }
+
+    /// Like `get` but `i` is a constant.
+    pub fn get_const<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        i: u64,
+    ) -> Instance<'ctx, Item> {
+        self.gep_const(ctx, i).load(generator, ctx)
+    }
+
+    /// Convenience function equivalent to `.gep(...).store(...)`.
+    pub fn set(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        i: IntValue<'ctx>,
+        value: Instance<'ctx, Item>,
+    ) {
+        self.gep(ctx, i).store(ctx, value);
+    }
+
+    /// Like `set` but `i` is a constant.
+    pub fn set_const(&self, ctx: &CodeGenContext<'ctx, '_>, i: u64, value: Instance<'ctx, Item>) {
+        self.gep_const(ctx, i).store(ctx, value);
+    }
+}

nac3core/src/codegen/model/core.rs

@@ -0,0 +1,237 @@
+use std::fmt;
+
+use inkwell::{context::Context, types::*, values::*};
+use itertools::Itertools;
+
+use super::*;
+use crate::codegen::{CodeGenContext, CodeGenerator};
+
+/// A error type for reporting any [`Model`]-related error (e.g., a [`BasicType`] mismatch).
+#[derive(Debug, Clone)]
+pub struct ModelError(pub String);
+
+impl ModelError {
+    /// Append a context message to the error.
+    pub(super) fn under_context(mut self, context: &str) -> Self {
+        self.0.push_str(" ... in ");
+        self.0.push_str(context);
+        self
+    }
+}
+
+// NOTE: A watered down version of `Model` trait. Made to be object safe.
+pub trait ModelBase<'ctx> {
+    // NOTE: Taking `size_t` here instead of `CodeGenerator` to be object safe.
+    // In fact, all the entire model abstraction need from the `CodeGenerator` is its `get_size_type()`.
+
+    // NOTE: Model's llvm_type but object-safe and returns BasicTypeEnum, instead of a known BasicType variant.
+    fn llvm_type_impl(&self, size_t: IntType<'ctx>, ctx: &'ctx Context) -> BasicTypeEnum<'ctx>;
+
+    // NOTE: Model's check_type but object-safe.
+    fn check_type_impl(
+        &self,
+        size_t: IntType<'ctx>,
+        ctx: &'ctx Context,
+        scrutinee: BasicTypeEnum<'ctx>,
+    ) -> Result<(), ModelError>;
+}
+
+/// Trait for Rust structs identifying [`BasicType`]s in the context of a known [`CodeGenerator`] and [`CodeGenContext`].
+///
+/// For instance,
+/// - [`Int<Int32>`] identifies an [`IntType`] with 32-bits.
+/// - [`Int<SizeT>`] identifies an [`IntType`] with bit-width [`CodeGenerator::get_size_type`].
+/// - [`Ptr<Int<SizeT>>`] identifies a [`PointerType`] that points to an [`IntType`] with bit-width [`CodeGenerator::get_size_type`].
+/// - [`Int<AnyInt>`] identifies an [`IntType`] with bit-width of whatever is set in the [`AnyInt`] object.
+/// - [`Any`] identifies a [`BasicType`] set in the [`Any`] object itself.
+///
+/// You can get the [`BasicType`] out of a model with [`Model::llvm_type`].
+///
+/// Furthermore, [`Instance<'ctx, M>`] is a simple structure that carries a [`BasicValue`] with [`BasicType`] identified by model `M`.
+///
+/// The main purpose of this abstraction is to have a more Rust type-safe way to use Inkwell and give type-hints for programmers.
+///
+/// ### Notes on `Default` trait
+///
+/// For some models like [`Int<Int32>`] or [`Int<SizeT>`], they have a [`Default`] trait since just by looking at their types, it is possible
+/// to tell the [`BasicType`]s they are identifying.
+///
+/// This can be used to create strongly-typed interfaces accepting only values of a specific [`BasicType`] without having to worry about
+/// writing debug assertions to check, for example, if the programmer has passed in an [`IntValue`] with the wrong bit-width.
+/// ```ignore
+/// fn give_me_i32_and_get_a_size_t_back<'ctx>(i32: Instance<'ctx, Int<Int32>>) -> Instance<'ctx, Int<SizeT>> {
+///     // code...
+/// }
+/// ```
+///
+/// ### Notes on converting between Inkwell and model.
+///
+/// Suppose you have an [`IntValue`], and you want to pass it into a function that takes a [`Instance<'ctx, Int<Int32>>`]. You can do use
+/// [`Model::check_value`] or [`Model::believe_value`].
+/// ```ignore
+/// let my_value: IntValue<'ctx>;
+///
+/// let my_value = Int(Int32).check_value(my_value).unwrap(); // Panics if `my_value` is not 32-bit with a descriptive error message.
+///
+/// // or, if you are absolutely certain that `my_value` is 32-bit and doing extra checks is a waste of time:
+/// let my_value = Int(Int32).believe_value(my_value);
+/// ```
+pub trait Model<'ctx>: fmt::Debug + Clone + Copy + ModelBase<'ctx> {
+    /// The [`BasicType`] *variant* this model is identifying.
+    type Type: BasicType<'ctx> + TryFrom<BasicTypeEnum<'ctx>>;
+
+    /// The [`BasicValue`] type of the [`BasicType`] of this model.
+    type Value: BasicValue<'ctx> + TryFrom<BasicValueEnum<'ctx>>;
+
+    /// Return the [`BasicType`] of this model.
+    #[must_use]
+    fn llvm_type<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &G,
+        ctx: &'ctx Context,
+    ) -> Self::Type {
+        let size_t = generator.get_size_type(ctx);
+
+        let ty = self.llvm_type_impl(size_t, ctx);
+        match Self::Type::try_from(ty) {
+            Ok(ty) => ty,
+            _ => panic!("Model::Type is inconsistent with what is returned from ModelBase::get_type_impl()! Got {ty:?}."),
+        }
+    }
+
+    /// Get the number of bytes of the [`BasicType`] of this model.
+    fn size_of<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+    ) -> IntValue<'ctx> {
+        self.llvm_type(generator, ctx).size_of().unwrap()
+    }
+
+    /// Check if a [`BasicType`] matches the [`BasicType`] of this model.
+    fn check_type<T: BasicType<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+        ty: T,
+    ) -> Result<(), ModelError> {
+        let size_t = generator.get_size_type(ctx);
+        self.check_type_impl(size_t, ctx, ty.as_basic_type_enum())
+    }
+
+    /// Create an instance from a value.
+    ///
+    /// # Safety
+    ///
+    /// Caller must make sure the type of `value` and the type of this `model` are equivalent.
+    #[must_use]
+    unsafe fn believe_value(&self, value: Self::Value) -> Instance<'ctx, Self> {
+        Instance { model: *self, value }
+    }
+
+    /// Check if a [`BasicValue`]'s type is equivalent to the type of this model.
+    /// Wrap the [`BasicValue`] into an [`Instance`] if it is.
+    fn check_value<V: BasicValue<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+        value: V,
+    ) -> Result<Instance<'ctx, Self>, ModelError> {
+        let value = value.as_basic_value_enum();
+        self.check_type(generator, ctx, value.get_type())
+            .map_err(|err| err.under_context(format!("the value {value:?}").as_str()))?;
+
+        let Ok(value) = Self::Value::try_from(value) else {
+            unreachable!("check_type() has bad implementation")
+        };
+        Ok(unsafe { self.believe_value(value) })
+    }
+
+    // Allocate a value on the stack and return its pointer.
+    fn alloca<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+    ) -> Instance<'ctx, Ptr<Self>> {
+        let p = ctx.builder.build_alloca(self.llvm_type(generator, ctx.ctx), "").unwrap();
+        unsafe { Ptr::new(*self).believe_value(p) }
+    }
+
+    // Allocate an array on the stack and return its pointer.
+    fn array_alloca<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        len: IntValue<'ctx>,
+    ) -> Instance<'ctx, Ptr<Self>> {
+        let p =
+            ctx.builder.build_array_alloca(self.llvm_type(generator, ctx.ctx), len, "").unwrap();
+        unsafe { Ptr::new(*self).believe_value(p) }
+    }
+
+    fn var_alloca<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        name: Option<&str>,
+    ) -> Result<Instance<'ctx, Ptr<Self>>, String> {
+        let ty = self.llvm_type(generator, ctx.ctx).as_basic_type_enum();
+        let p = generator.gen_var_alloc(ctx, ty, name)?;
+        Ok(unsafe { Ptr::new(*self).believe_value(p) })
+    }
+
+    fn array_var_alloca<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &mut CodeGenContext<'ctx, '_>,
+        len: IntValue<'ctx>,
+        name: Option<&'ctx str>,
+    ) -> Result<Instance<'ctx, Ptr<Self>>, String> {
+        // TODO: Remove ArraySliceValue
+        let ty = self.llvm_type(generator, ctx.ctx).as_basic_type_enum();
+        let p = generator.gen_array_var_alloc(ctx, ty, len, name)?;
+        Ok(unsafe { Ptr::new(*self).believe_value(PointerValue::from(p)) })
+    }
+
+    /// Allocate a constant array.
+    fn const_array<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+        values: &[Instance<'ctx, Self>],
+    ) -> Instance<'ctx, Array<'ctx, AnyLen, Self>> {
+        macro_rules! make {
+            ($t:expr, $into_value:expr) => {
+                $t.const_array(
+                    &values
+                        .iter()
+                        .map(|x| $into_value(x.value.as_basic_value_enum()))
+                        .collect_vec(),
+                )
+            };
+        }
+
+        let value = match self.llvm_type(generator, ctx).as_basic_type_enum() {
+            BasicTypeEnum::ArrayType(t) => make!(t, BasicValueEnum::into_array_value),
+            BasicTypeEnum::IntType(t) => make!(t, BasicValueEnum::into_int_value),
+            BasicTypeEnum::FloatType(t) => make!(t, BasicValueEnum::into_float_value),
+            BasicTypeEnum::PointerType(t) => make!(t, BasicValueEnum::into_pointer_value),
+            BasicTypeEnum::StructType(t) => make!(t, BasicValueEnum::into_struct_value),
+            BasicTypeEnum::VectorType(t) => make!(t, BasicValueEnum::into_vector_value),
+        };
+
+        let model = Array::new(AnyLen(values.len() as u32), *self);
+        model.check_value(generator, ctx, value).unwrap()
+    }
+}
+
+#[derive(Debug, Clone, Copy)]
+pub struct Instance<'ctx, M: Model<'ctx>> {
+    /// The model of this instance.
+    pub model: M,
+
+    /// The value of this instance.
+    ///
+    /// It is guaranteed the [`BasicType`] of `value` is consistent with that of `model`.
+    pub value: M::Value,
+}

nac3core/src/codegen/model/float.rs

@@ -0,0 +1,80 @@
+use std::{fmt, marker::PhantomData};
+
+use inkwell::{
+    context::Context,
+    types::{BasicTypeEnum, FloatType, IntType},
+    values::FloatValue,
+};
+
+use super::*;
+
+pub trait FloatKind<'ctx>: fmt::Debug + Clone + Copy {
+    fn float_type(&self, ctx: &'ctx Context) -> FloatType<'ctx>;
+}
+
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Float32;
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Float64;
+
+impl<'ctx> FloatKind<'ctx> for Float32 {
+    fn float_type(&self, ctx: &'ctx Context) -> FloatType<'ctx> {
+        ctx.f32_type()
+    }
+}
+
+impl<'ctx> FloatKind<'ctx> for Float64 {
+    fn float_type(&self, ctx: &'ctx Context) -> FloatType<'ctx> {
+        ctx.f64_type()
+    }
+}
+
+#[derive(Debug, Clone, Copy)]
+pub struct AnyFloat<'ctx>(FloatType<'ctx>);
+
+impl<'ctx> FloatKind<'ctx> for AnyFloat<'ctx> {
+    fn float_type(&self, _ctx: &'ctx Context) -> FloatType<'ctx> {
+        self.0
+    }
+}
+
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Float<'ctx, N: FloatKind<'ctx>> {
+    kind: N,
+    _phantom: PhantomData<&'ctx ()>,
+}
+
+impl<'ctx, N: FloatKind<'ctx>> Float<'ctx, N> {
+    pub fn new(kind: N) -> Self {
+        Float { kind, _phantom: PhantomData }
+    }
+}
+
+impl<'ctx, N: FloatKind<'ctx>> ModelBase<'ctx> for Float<'ctx, N> {
+    fn llvm_type_impl(&self, _size_t: IntType<'ctx>, ctx: &'ctx Context) -> BasicTypeEnum<'ctx> {
+        self.kind.float_type(ctx).into()
+    }
+
+    fn check_type_impl(
+        &self,
+        _size_t: IntType<'ctx>,
+        ctx: &'ctx Context,
+        ty: BasicTypeEnum<'ctx>,
+    ) -> Result<(), ModelError> {
+        let Ok(ty) = FloatType::try_from(ty) else {
+            return Err(ModelError(format!("Expecting FloatType, but got {ty:?}")));
+        };
+
+        let expected_ty = self.kind.float_type(ctx);
+        if ty != expected_ty {
+            return Err(ModelError(format!("Expecting {expected_ty:?}, but got {ty:?}")));
+        }
+
+        Ok(())
+    }
+}
+
+impl<'ctx, N: FloatKind<'ctx>> Model<'ctx> for Float<'ctx, N> {
+    type Value = FloatValue<'ctx>;
+    type Type = FloatType<'ctx>;
+}

nac3core/src/codegen/model/function.rs

@@ -0,0 +1,122 @@
+use inkwell::{
+    attributes::{Attribute, AttributeLoc},
+    types::{BasicMetadataTypeEnum, BasicType, FunctionType},
+    values::{AnyValue, BasicMetadataValueEnum, BasicValue, BasicValueEnum, CallSiteValue},
+};
+use itertools::Itertools;
+
+use crate::codegen::{CodeGenContext, CodeGenerator};
+
+use super::*;
+
+#[derive(Debug, Clone, Copy)]
+struct Arg<'ctx> {
+    ty: BasicMetadataTypeEnum<'ctx>,
+    val: BasicMetadataValueEnum<'ctx>,
+}
+
+/// A convenience structure to construct & call an LLVM function.
+///
+/// ### Usage
+///
+/// The syntax is like this:
+/// ```ignore
+/// let result = CallFunction::begin("my_function_name")
+///   .attrs(...)
+///   .arg(arg1)
+///   .arg(arg2)
+///   .arg(arg3)
+///   .returning("my_function_result", Int32);
+/// ```
+///
+/// The function `my_function_name` is called when `.returning()` (or its variants) is called, returning
+/// the result as an `Instance<'ctx, Int<Int32>>`.
+///
+/// If `my_function_name` has not been declared in `ctx.module`, once `.returning()` is called, a function
+/// declaration of `my_function_name` is added to `ctx.module`, where the [`FunctionType`] is deduced from
+/// the argument types and returning type.
+pub struct FnCall<'ctx, 'a, 'b, G: CodeGenerator + ?Sized> {
+    generator: &'b mut G,
+    ctx: &'b CodeGenContext<'ctx, 'a>,
+    /// Function name
+    name: &'b str,
+    /// Call arguments
+    args: Vec<Arg<'ctx>>,
+    /// LLVM function Attributes
+    attrs: Vec<&'static str>,
+}
+
+impl<'ctx, 'a, 'b, G: CodeGenerator + ?Sized> FnCall<'ctx, 'a, 'b, G> {
+    pub fn begin(generator: &'b mut G, ctx: &'b CodeGenContext<'ctx, 'a>, name: &'b str) -> Self {
+        FnCall { generator, ctx, name, args: Vec::new(), attrs: Vec::new() }
+    }
+
+    /// Push a list of LLVM function attributes to the function declaration.
+    #[must_use]
+    pub fn attrs(mut self, attrs: Vec<&'static str>) -> Self {
+        self.attrs = attrs;
+        self
+    }
+
+    /// Push a call argument to the function call.
+    #[allow(clippy::needless_pass_by_value)]
+    #[must_use]
+    pub fn arg<M: Model<'ctx>>(mut self, arg: Instance<'ctx, M>) -> Self {
+        let arg = Arg {
+            ty: arg.model.llvm_type(self.generator, self.ctx.ctx).as_basic_type_enum().into(),
+            val: arg.value.as_basic_value_enum().into(),
+        };
+        self.args.push(arg);
+        self
+    }
+
+    /// Call the function and expect the function to return a value of type of `return_model`.
+    #[must_use]
+    pub fn returning<M: Model<'ctx>>(self, name: &str, return_model: M) -> Instance<'ctx, M> {
+        let ret_ty = return_model.llvm_type(self.generator, self.ctx.ctx);
+
+        let ret = self.call(|tys| ret_ty.fn_type(tys, false), name);
+        let ret = BasicValueEnum::try_from(ret.as_any_value_enum()).unwrap(); // Must work
+        let ret = return_model.check_value(self.generator, self.ctx.ctx, ret).unwrap(); // Must work
+        ret
+    }
+
+    /// Like [`CallFunction::returning_`] but `return_model` is automatically inferred.
+    #[must_use]
+    pub fn returning_auto<M: Model<'ctx> + Default>(self, name: &str) -> Instance<'ctx, M> {
+        self.returning(name, M::default())
+    }
+
+    /// Call the function and expect the function to return a void-type.
+    pub fn returning_void(self) {
+        let ret_ty = self.ctx.ctx.void_type();
+
+        let _ = self.call(|tys| ret_ty.fn_type(tys, false), "");
+    }
+
+    fn call<F>(&self, make_fn_type: F, return_value_name: &str) -> CallSiteValue<'ctx>
+    where
+        F: FnOnce(&[BasicMetadataTypeEnum<'ctx>]) -> FunctionType<'ctx>,
+    {
+        // Get the LLVM function.
+        let func = self.ctx.module.get_function(self.name).unwrap_or_else(|| {
+            // Declare the function if it doesn't exist.
+            let tys = self.args.iter().map(|arg| arg.ty).collect_vec();
+
+            let func_type = make_fn_type(&tys);
+            let func = self.ctx.module.add_function(self.name, func_type, None);
+
+            for attr in &self.attrs {
+                func.add_attribute(
+                    AttributeLoc::Function,
+                    self.ctx.ctx.create_enum_attribute(Attribute::get_named_enum_kind_id(attr), 0),
+                );
+            }
+
+            func
+        });
+
+        let vals = self.args.iter().map(|arg| arg.val).collect_vec();
+        self.ctx.builder.build_call(func, &vals, return_value_name).unwrap()
+    }
+}

nac3core/src/codegen/model/int.rs

@@ -0,0 +1,400 @@
+use std::{cmp::Ordering, fmt, marker::PhantomData};
+
+use inkwell::{
+    context::Context,
+    types::{BasicTypeEnum, IntType},
+    values::IntValue,
+    IntPredicate,
+};
+
+use crate::codegen::{CodeGenContext, CodeGenerator};
+
+use super::*;
+
+pub trait IntKind<'ctx>: fmt::Debug + Clone + Copy {
+    fn int_type(&self, size_t: IntType<'ctx>, ctx: &'ctx Context) -> IntType<'ctx>;
+}
+
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Bool;
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Byte;
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Int32;
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Int64;
+#[derive(Debug, Clone, Copy, Default)]
+pub struct SizeT;
+
+impl<'ctx> IntKind<'ctx> for Bool {
+    fn int_type(&self, _size_t: IntType<'ctx>, ctx: &'ctx Context) -> IntType<'ctx> {
+        ctx.bool_type()
+    }
+}
+
+impl<'ctx> IntKind<'ctx> for Byte {
+    fn int_type(&self, _size_t: IntType<'ctx>, ctx: &'ctx Context) -> IntType<'ctx> {
+        ctx.i8_type()
+    }
+}
+
+impl<'ctx> IntKind<'ctx> for Int32 {
+    fn int_type(&self, _size_t: IntType<'ctx>, ctx: &'ctx Context) -> IntType<'ctx> {
+        ctx.i32_type()
+    }
+}
+
+impl<'ctx> IntKind<'ctx> for Int64 {
+    fn int_type(&self, _size_t: IntType<'ctx>, ctx: &'ctx Context) -> IntType<'ctx> {
+        ctx.i64_type()
+    }
+}
+
+impl<'ctx> IntKind<'ctx> for SizeT {
+    fn int_type(&self, size_t: IntType<'ctx>, _ctx: &'ctx Context) -> IntType<'ctx> {
+        size_t
+    }
+}
+
+#[derive(Debug, Clone, Copy)]
+pub struct AnyInt<'ctx>(pub IntType<'ctx>);
+
+impl<'ctx> IntKind<'ctx> for AnyInt<'ctx> {
+    fn int_type(&self, _size_t: IntType<'ctx>, _ctx: &'ctx Context) -> IntType<'ctx> {
+        self.0
+    }
+}
+
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Int<'ctx, N: IntKind<'ctx>> {
+    kind: N,
+    _phantom: PhantomData<&'ctx ()>,
+}
+
+impl<'ctx, N: IntKind<'ctx>> ModelBase<'ctx> for Int<'ctx, N> {
+    fn llvm_type_impl(&self, size_t: IntType<'ctx>, ctx: &'ctx Context) -> BasicTypeEnum<'ctx> {
+        self.kind.int_type(size_t, ctx).into()
+    }
+
+    fn check_type_impl(
+        &self,
+        size_t: IntType<'ctx>,
+        ctx: &'ctx Context,
+        ty: BasicTypeEnum<'ctx>,
+    ) -> Result<(), ModelError> {
+        let Ok(ty) = IntType::try_from(ty) else {
+            return Err(ModelError(format!("Expecting IntType, but got {ty:?}")));
+        };
+
+        let exp_ty = self.kind.int_type(size_t, ctx);
+        if ty.get_bit_width() != exp_ty.get_bit_width() {
+            return Err(ModelError(format!(
+                "Expecting IntType to have {} bit(s), but got {} bit(s)",
+                exp_ty.get_bit_width(),
+                ty.get_bit_width()
+            )));
+        }
+
+        Ok(())
+    }
+}
+
+impl<'ctx, N: IntKind<'ctx>> Model<'ctx> for Int<'ctx, N> {
+    type Type = IntType<'ctx>;
+    type Value = IntValue<'ctx>;
+}
+
+impl<'ctx, N: IntKind<'ctx>> Int<'ctx, N> {
+    pub fn new(kind: N) -> Self {
+        Int { kind, _phantom: PhantomData }
+    }
+
+    pub fn const_int<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+        value: u64,
+        sign_extend: bool,
+    ) -> Instance<'ctx, Self> {
+        let value = self.llvm_type(generator, ctx).const_int(value, sign_extend);
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn const_0<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+    ) -> Instance<'ctx, Self> {
+        let value = self.llvm_type(generator, ctx).const_zero();
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn const_1<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+    ) -> Instance<'ctx, Self> {
+        self.const_int(generator, ctx, 1, false)
+    }
+
+    pub fn const_all_ones<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+    ) -> Instance<'ctx, Self> {
+        let value = self.llvm_type(generator, ctx).const_all_ones();
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn s_extend_or_bit_cast<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        assert!(
+            value.get_type().get_bit_width()
+                <= self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width()
+        );
+        let value = ctx
+            .builder
+            .build_int_s_extend_or_bit_cast(value, self.llvm_type(generator, ctx.ctx), "")
+            .unwrap();
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn s_extend<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        assert!(
+            value.get_type().get_bit_width()
+                < self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width()
+        );
+        let value =
+            ctx.builder.build_int_s_extend(value, self.llvm_type(generator, ctx.ctx), "").unwrap();
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn z_extend_or_bit_cast<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        assert!(
+            value.get_type().get_bit_width()
+                <= self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width()
+        );
+        let value = ctx
+            .builder
+            .build_int_z_extend_or_bit_cast(value, self.llvm_type(generator, ctx.ctx), "")
+            .unwrap();
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn z_extend<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        assert!(
+            value.get_type().get_bit_width()
+                < self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width()
+        );
+        let value =
+            ctx.builder.build_int_z_extend(value, self.llvm_type(generator, ctx.ctx), "").unwrap();
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn truncate_or_bit_cast<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        assert!(
+            value.get_type().get_bit_width()
+                >= self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width()
+        );
+        let value = ctx
+            .builder
+            .build_int_truncate_or_bit_cast(value, self.llvm_type(generator, ctx.ctx), "")
+            .unwrap();
+        unsafe { self.believe_value(value) }
+    }
+
+    pub fn truncate<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        assert!(
+            value.get_type().get_bit_width()
+                > self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width()
+        );
+        let value =
+            ctx.builder.build_int_truncate(value, self.llvm_type(generator, ctx.ctx), "").unwrap();
+        unsafe { self.believe_value(value) }
+    }
+
+    /// `sext` or `trunc` an int to this model's int type. Does nothing if equal bit-widths.
+    pub fn s_extend_or_truncate<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        let their_width = value.get_type().get_bit_width();
+        let our_width =
+            self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width();
+        match their_width.cmp(&our_width) {
+            Ordering::Less => self.s_extend(generator, ctx, value),
+            Ordering::Equal => unsafe { self.believe_value(value) },
+            Ordering::Greater => self.truncate(generator, ctx, value),
+        }
+    }
+
+    /// `zext` or `trunc` an int to this model's int type. Does nothing if equal bit-widths.
+    pub fn z_extend_or_truncate<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        value: IntValue<'ctx>,
+    ) -> Instance<'ctx, Self> {
+        let their_width = value.get_type().get_bit_width();
+        let our_width =
+            self.kind.int_type(generator.get_size_type(ctx.ctx), ctx.ctx).get_bit_width();
+        match their_width.cmp(&our_width) {
+            Ordering::Less => self.z_extend(generator, ctx, value),
+            Ordering::Equal => unsafe { self.believe_value(value) },
+            Ordering::Greater => self.truncate(generator, ctx, value),
+        }
+    }
+}
+
+impl<'ctx> Int<'ctx, Bool> {
+    #[must_use]
+    pub fn const_false<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+    ) -> Instance<'ctx, Self> {
+        self.const_int(generator, ctx, 0, false)
+    }
+
+    #[must_use]
+    pub fn const_true<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+    ) -> Instance<'ctx, Self> {
+        self.const_int(generator, ctx, 1, false)
+    }
+}
+
+impl<'ctx, N: IntKind<'ctx>> Instance<'ctx, Int<'ctx, N>> {
+    pub fn s_extend_or_bit_cast<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).s_extend_or_bit_cast(generator, ctx, self.value)
+    }
+
+    pub fn s_extend<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).s_extend(generator, ctx, self.value)
+    }
+
+    pub fn z_extend_or_bit_cast<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).z_extend_or_bit_cast(generator, ctx, self.value)
+    }
+
+    pub fn z_extend<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).z_extend(generator, ctx, self.value)
+    }
+
+    pub fn truncate_or_bit_cast<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).truncate_or_bit_cast(generator, ctx, self.value)
+    }
+
+    pub fn truncate<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).truncate(generator, ctx, self.value)
+    }
+
+    pub fn s_extend_or_truncate<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).s_extend_or_truncate(generator, ctx, self.value)
+    }
+
+    pub fn z_extend_or_truncate<NewN: IntKind<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        to_int_kind: NewN,
+    ) -> Instance<'ctx, Int<'ctx, NewN>> {
+        Int::new(to_int_kind).z_extend_or_truncate(generator, ctx, self.value)
+    }
+
+    #[must_use]
+    pub fn add(&self, ctx: &CodeGenContext<'ctx, '_>, other: Self) -> Self {
+        let value = ctx.builder.build_int_add(self.value, other.value, "").unwrap();
+        unsafe { self.model.believe_value(value) }
+    }
+
+    #[must_use]
+    pub fn sub(&self, ctx: &CodeGenContext<'ctx, '_>, other: Self) -> Self {
+        let value = ctx.builder.build_int_sub(self.value, other.value, "").unwrap();
+        unsafe { self.model.believe_value(value) }
+    }
+
+    #[must_use]
+    pub fn mul(&self, ctx: &CodeGenContext<'ctx, '_>, other: Self) -> Self {
+        let value = ctx.builder.build_int_mul(self.value, other.value, "").unwrap();
+        unsafe { self.model.believe_value(value) }
+    }
+
+    pub fn compare(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        op: IntPredicate,
+        other: Self,
+    ) -> Instance<'ctx, Int<'ctx, Bool>> {
+        let value = ctx.builder.build_int_compare(op, self.value, other.value, "").unwrap();
+        unsafe { Int::new(Bool).believe_value(value) }
+    }
+}

nac3core/src/codegen/model/mod.rs

@@ -0,0 +1,16 @@
+mod any;
+mod array;
+mod core;
+mod float;
+pub mod function;
+mod int;
+mod ptr;
+mod structure;
+
+pub use any::*;
+pub use array::*;
+pub use core::*;
+pub use float::*;
+pub use int::*;
+pub use ptr::*;
+pub use structure::*;

nac3core/src/codegen/model/ptr.rs

@@ -0,0 +1,223 @@
+use inkwell::{
+    context::Context,
+    types::{BasicType, BasicTypeEnum, IntType, PointerType},
+    values::{IntValue, PointerValue},
+    AddressSpace,
+};
+
+use crate::codegen::{llvm_intrinsics::call_memcpy_generic, CodeGenContext, CodeGenerator};
+
+use super::*;
+
+/// A model for [`PointerType`].
+///
+/// `Item` is the element type this pointer is pointing to, and should be of a [`Model`].
+///
+// TODO: LLVM 15: `Item` is a Rust type-hint for the LLVM type of value the `.store()/.load()` family
+// of functions return. If a truly opaque pointer is needed, tell the programmer to use `OpaquePtr`.
+//
+// NOTE: Do not put `Item: ModelBase<'ctx>`. See the LLVM 15 note above.
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Ptr<Item> {
+    pub item: Item,
+}
+
+/// An opaque pointer. Like [`Ptr`] but without any Rust type-hints about its element type.
+///
+/// `.load()/.store()` is not available for [`Instance`]s of opaque pointers.
+pub type OpaquePtr = Ptr<()>;
+
+impl<Item> Ptr<Item> {
+    pub fn new(item: Item) -> Self {
+        Ptr { item }
+    }
+}
+
+// TODO: LLVM 15: `Item: ModelBase<'ctx>` don't even need to be a model anymore. It will only be
+// a type-hint for the `.load()/.store()` functions for the `pointee_ty`.
+//
+// See https://thedan64.github.io/inkwell/inkwell/builder/struct.Builder.html#method.build_load.
+impl<'ctx, Item: ModelBase<'ctx>> ModelBase<'ctx> for Ptr<Item> {
+    fn llvm_type_impl(&self, size_t: IntType<'ctx>, ctx: &'ctx Context) -> BasicTypeEnum<'ctx> {
+        // TODO: LLVM 15: ctx.ptr_type(AddressSpace::default())
+        let item = self.item.llvm_type_impl(size_t, ctx);
+        item.ptr_type(AddressSpace::default()).into()
+    }
+
+    fn check_type_impl(
+        &self,
+        size_t: IntType<'ctx>,
+        ctx: &'ctx Context,
+        ty: BasicTypeEnum<'ctx>,
+    ) -> Result<(), ModelError> {
+        let Ok(ty) = PointerType::try_from(ty) else {
+            return Err(ModelError(format!("Expecting PointerType, but got {ty:?}")));
+        };
+
+        let elem_ty = ty.get_element_type();
+        let Ok(elem_ty) = BasicTypeEnum::try_from(elem_ty) else {
+            return Err(ModelError(format!(
+                "Expecting pointer element type to be a BasicTypeEnum, but got {elem_ty:?}"
+            )));
+        };
+
+        // TODO: inkwell `get_element_type()` will be deprecated.
+        // Remove the check for `get_element_type()` when the time comes.
+        self.item
+            .check_type_impl(size_t, ctx, elem_ty)
+            .map_err(|err| err.under_context("a PointerType"))?;
+
+        Ok(())
+    }
+}
+
+impl<'ctx, Item: Model<'ctx>> Model<'ctx> for Ptr<Item> {
+    type Type = PointerType<'ctx>;
+    type Value = PointerValue<'ctx>;
+}
+
+impl<'ctx, Item: Model<'ctx>> Ptr<Item> {
+    /// Return a ***constant*** nullptr.
+    pub fn nullptr<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+    ) -> Instance<'ctx, Ptr<Item>> {
+        // TODO: LLVM 15: Write in an impl where `Item` does not have to be `Model<'ctx>`.
+        let ptr = self.llvm_type(generator, ctx).const_null();
+        unsafe { self.believe_value(ptr) }
+    }
+
+    /// Cast a pointer into this model with [`inkwell::builder::Builder::build_pointer_cast`]
+    pub fn pointer_cast<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        ptr: PointerValue<'ctx>,
+    ) -> Instance<'ctx, Ptr<Item>> {
+        // TODO: LLVM 15: Write in an impl where `Item` does not have to be `Model<'ctx>`.
+        // TODO: LLVM 15: This function will only have to be:
+        // ```
+        // return self.believe_value(ptr);
+        // ```
+        let t = self.llvm_type(generator, ctx.ctx);
+        let ptr = ctx.builder.build_pointer_cast(ptr, t, "").unwrap();
+        unsafe { self.believe_value(ptr) }
+    }
+}
+
+impl<'ctx, Item: Model<'ctx>> Instance<'ctx, Ptr<Item>> {
+    /// Offset the pointer by [`inkwell::builder::Builder::build_in_bounds_gep`].
+    #[must_use]
+    pub fn offset(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        offset: IntValue<'ctx>,
+    ) -> Instance<'ctx, Ptr<Item>> {
+        let p = unsafe { ctx.builder.build_in_bounds_gep(self.value, &[offset], "").unwrap() };
+        unsafe { self.model.believe_value(p) }
+    }
+
+    /// Offset the pointer by [`inkwell::builder::Builder::build_in_bounds_gep`] by a constant offset.
+    #[must_use]
+    pub fn offset_const(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        offset: i64,
+    ) -> Instance<'ctx, Ptr<Item>> {
+        let offset = ctx.ctx.i64_type().const_int(offset as u64, true);
+        self.offset(ctx, offset)
+    }
+
+    pub fn set_index(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        index: IntValue<'ctx>,
+        value: Instance<'ctx, Item>,
+    ) {
+        self.offset(ctx, index).store(ctx, value);
+    }
+
+    pub fn set_index_const(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        index: i64,
+        value: Instance<'ctx, Item>,
+    ) {
+        self.offset_const(ctx, index).store(ctx, value);
+    }
+
+    pub fn get_index<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        index: IntValue<'ctx>,
+    ) -> Instance<'ctx, Item> {
+        self.offset(ctx, index).load(generator, ctx)
+    }
+
+    pub fn get_index_const<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        index: i64,
+    ) -> Instance<'ctx, Item> {
+        self.offset_const(ctx, index).load(generator, ctx)
+    }
+
+    /// Load the value with [`inkwell::builder::Builder::build_load`].
+    pub fn load<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+    ) -> Instance<'ctx, Item> {
+        let value = ctx.builder.build_load(self.value, "").unwrap();
+        self.model.item.check_value(generator, ctx.ctx, value).unwrap() // If unwrap() panics, there is a logic error.
+    }
+
+    /// Store a value with [`inkwell::builder::Builder::build_store`].
+    pub fn store(&self, ctx: &CodeGenContext<'ctx, '_>, value: Instance<'ctx, Item>) {
+        ctx.builder.build_store(self.value, value.value).unwrap();
+    }
+
+    /// Return a casted pointer of element type `NewElement` with [`inkwell::builder::Builder::build_pointer_cast`].
+    pub fn pointer_cast<NewItem: Model<'ctx>, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        new_item: NewItem,
+    ) -> Instance<'ctx, Ptr<NewItem>> {
+        // TODO: LLVM 15: Write in an impl where `Item` does not have to be `Model<'ctx>`.
+        Ptr::new(new_item).pointer_cast(generator, ctx, self.value)
+    }
+
+    /// Check if the pointer is null with [`inkwell::builder::Builder::build_is_null`].
+    pub fn is_null(&self, ctx: &CodeGenContext<'ctx, '_>) -> Instance<'ctx, Int<'ctx, Bool>> {
+        let value = ctx.builder.build_is_null(self.value, "").unwrap();
+        unsafe { Int::new(Bool).believe_value(value) }
+    }
+
+    /// Check if the pointer is not null with [`inkwell::builder::Builder::build_is_not_null`].
+    pub fn is_not_null(&self, ctx: &CodeGenContext<'ctx, '_>) -> Instance<'ctx, Int<'ctx, Bool>> {
+        let value = ctx.builder.build_is_not_null(self.value, "").unwrap();
+        unsafe { Int::new(Bool).believe_value(value) }
+    }
+
+    /// `memcpy` from another pointer.
+    pub fn copy_from<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        source: Self,
+        num_items: IntValue<'ctx>,
+    ) {
+        // Force extend `num_items` and `itemsize` to `i64` so their types would match.
+        let itemsize = self.model.size_of(generator, ctx.ctx);
+        let itemsize = Int::new(Int64).z_extend_or_truncate(generator, ctx, itemsize);
+        let num_items = Int::new(Int64).z_extend_or_truncate(generator, ctx, num_items);
+        let totalsize = itemsize.mul(ctx, num_items);
+
+        let is_volatile = ctx.ctx.bool_type().const_zero(); // is_volatile = false
+        call_memcpy_generic(ctx, self.value, source.value, totalsize.value, is_volatile);
+    }
+}

nac3core/src/codegen/model/structure.rs

@@ -0,0 +1,275 @@
+use std::{fmt, marker::PhantomData};
+
+use inkwell::{
+    context::Context,
+    types::{BasicType, BasicTypeEnum, IntType, StructType},
+    values::{BasicValueEnum, StructValue},
+};
+use itertools::{izip, Itertools};
+
+use crate::codegen::{CodeGenContext, CodeGenerator};
+
+use super::*;
+
+// pub trait StructKind2<'ctx>: fmt::Debug + Clone + Copy {
+//     type Fields<F: FieldTraversal2<'ctx>> = ;
+// }
+
+pub struct Field<M> {
+    gep_index: u32,
+    model: M,
+    name: &'static str,
+}
+
+// NOTE: Very similar to Field, but is forall on `M`, (and also uses ModelBase to get object safety for the `Box<dyn ____>`.
+pub struct Entry<'ctx> {
+    model: Box<dyn ModelBase<'ctx> + 'ctx>,
+    name: &'static str,
+}
+
+pub struct FieldMapper<'ctx> {
+    gep_index_counter: u32,
+    entries: Vec<Entry<'ctx>>,
+}
+
+impl<'ctx> FieldMapper<'ctx> {
+    fn add<M: 'ctx + Model<'ctx>>(&mut self, name: &'static str, model: M) -> Field<M> {
+        let entry = Entry { model: Box::new(model), name };
+        self.entries.push(entry);
+
+        let gep_index = self.gep_index_counter;
+        self.gep_index_counter += 1;
+        Field { gep_index, model, name }
+    }
+
+    fn add_auto<M: 'ctx + Model<'ctx> + Default>(&mut self, name: &'static str) -> Field<M> {
+        self.add(name, M::default())
+    }
+}
+
+pub trait StructKind<'ctx>: fmt::Debug + Clone + Copy {
+    type Fields;
+
+    fn iter_fields(&self, mapper: &mut FieldMapper<'ctx>) -> Self::Fields;
+
+    // Produce `Vec<Entry>` and `Self::Fields` simultaneously.
+    // The former is for doing field-wise type checks.
+    // The latter is for enabling the `.gep(|f| f.data)` syntax.
+    fn entries_and_fields(&self) -> (Vec<Entry<'ctx>>, Self::Fields) {
+        let mut mapper = FieldMapper { gep_index_counter: 0, entries: Vec::new() };
+        let fields = self.iter_fields(&mut mapper);
+        (mapper.entries, fields)
+    }
+
+    fn entries(&self) -> Vec<Entry<'ctx>> {
+        self.entries_and_fields().0
+    }
+
+    fn fields(&self) -> Self::Fields {
+        self.entries_and_fields().1
+    }
+
+    /// Get the LLVM [`StructType`] of this [`StructKind`].
+    fn get_struct_type(&self, size_t: IntType<'ctx>, ctx: &'ctx Context) -> StructType<'ctx> {
+        let entries = self.entries();
+        let entries =
+            entries.into_iter().map(|t| t.model.llvm_type_impl(size_t, ctx)).collect_vec();
+        ctx.struct_type(&entries, false)
+    }
+}
+
+#[derive(Debug, Clone, Copy, Default)]
+pub struct Struct<'ctx, S: StructKind<'ctx>> {
+    pub kind: S,
+    _phantom: PhantomData<&'ctx ()>,
+}
+
+impl<'ctx, S: StructKind<'ctx>> ModelBase<'ctx> for Struct<'ctx, S> {
+    fn llvm_type_impl(&self, size_t: IntType<'ctx>, ctx: &'ctx Context) -> BasicTypeEnum<'ctx> {
+        self.kind.get_struct_type(size_t, ctx).as_basic_type_enum()
+    }
+
+    fn check_type_impl(
+        &self,
+        size_t: IntType<'ctx>,
+        ctx: &'ctx Context,
+        ty: BasicTypeEnum<'ctx>,
+    ) -> Result<(), ModelError> {
+        let Ok(ty) = StructType::try_from(ty) else {
+            return Err(ModelError(format!("Expecting StructType, but got {ty:?}")));
+        };
+
+        let entries = self.kind.entries();
+        let field_types = ty.get_field_types();
+
+        // Check the number of fields.
+        if entries.len() != field_types.len() {
+            return Err(ModelError(format!(
+                "Expecting StructType with {} field(s), but got {}",
+                entries.len(),
+                field_types.len()
+            )));
+        }
+
+        // Check each field.
+        for (i, (entry, field_type)) in izip!(entries, field_types).enumerate() {
+            entry.model.check_type_impl(size_t, ctx, field_type).map_err(|err| {
+                let context = &format!("in field #{i} '{}'", entry.name);
+                err.under_context(context)
+            })?;
+        }
+
+        Ok(())
+    }
+}
+
+impl<'ctx, S: StructKind<'ctx>> Model<'ctx> for Struct<'ctx, S> {
+    type Type = StructType<'ctx>;
+    type Value = StructValue<'ctx>;
+}
+
+impl<'ctx, S: StructKind<'ctx>> Struct<'ctx, S> {
+    pub fn new(kind: S) -> Self {
+        Struct { kind, _phantom: PhantomData }
+    }
+
+    pub fn const_struct<G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+        fields: &[BasicValueEnum<'ctx>],
+    ) -> Instance<'ctx, Self> {
+        let val = ctx.const_struct(fields, false);
+        self.check_value(generator, ctx, val).unwrap()
+    }
+}
+
+impl<'ctx, S: StructKind<'ctx>> Instance<'ctx, Struct<'ctx, S>> {
+    /// Get a field with [`StructValue::get_field_at_index`].
+    pub fn get_field<G: CodeGenerator + ?Sized, M, GetField>(
+        &self,
+        generator: &mut G,
+        ctx: &'ctx Context,
+        get_field: GetField,
+    ) -> Instance<'ctx, M>
+    where
+        M: Model<'ctx>,
+        GetField: FnOnce(S::Fields) -> Field<M>,
+    {
+        let field = get_field(self.model.kind.fields());
+        let val = self.value.get_field_at_index(field.gep_index).unwrap();
+        field.model.check_value(generator, ctx, val).unwrap()
+    }
+}
+
+impl<'ctx, S: StructKind<'ctx>> Instance<'ctx, Ptr<Struct<'ctx, S>>> {
+    /// Get a pointer to a field with [`Builder::build_in_bounds_gep`].
+    pub fn gep<M, GetField>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        get_field: GetField,
+    ) -> Instance<'ctx, Ptr<M>>
+    where
+        M: Model<'ctx>,
+        GetField: FnOnce(S::Fields) -> Field<M>,
+    {
+        let field = get_field(self.model.item.kind.fields());
+        let llvm_i32 = ctx.ctx.i32_type();
+
+        let ptr = unsafe {
+            ctx.builder
+                .build_in_bounds_gep(
+                    self.value,
+                    &[llvm_i32.const_zero(), llvm_i32.const_int(u64::from(field.gep_index), false)],
+                    field.name,
+                )
+                .unwrap()
+        };
+
+        unsafe { Ptr::new(field.model).believe_value(ptr) }
+    }
+
+    /// Convenience function equivalent to `.gep(...).load(...)`.
+    pub fn get<M, GetField, G: CodeGenerator + ?Sized>(
+        &self,
+        generator: &mut G,
+        ctx: &CodeGenContext<'ctx, '_>,
+        get_field: GetField,
+    ) -> Instance<'ctx, M>
+    where
+        M: Model<'ctx>,
+        GetField: FnOnce(S::Fields) -> Field<M>,
+    {
+        self.gep(ctx, get_field).load(generator, ctx)
+    }
+
+    /// Convenience function equivalent to `.gep(...).store(...)`.
+    pub fn set<M, GetField>(
+        &self,
+        ctx: &CodeGenContext<'ctx, '_>,
+        get_field: GetField,
+        value: Instance<'ctx, M>,
+    ) where
+        M: Model<'ctx>,
+        GetField: FnOnce(S::Fields) -> Field<M>,
+    {
+        self.gep(ctx, get_field).store(ctx, value);
+    }
+}
+
+/////////////////////// Example; Delete later
+
+// Example: NDArray.
+//
+// Compared to List, it has no generic models.
+pub struct NDArrayFields<'ctx> {
+    data: Field<Ptr<Int<'ctx, Byte>>>,
+    itemsize: Field<Int<'ctx, SizeT>>,
+    ndims: Field<Int<'ctx, SizeT>>,
+    shape: Field<Ptr<Int<'ctx, SizeT>>>,
+    strides: Field<Ptr<Int<'ctx, SizeT>>>,
+}
+
+#[derive(Debug, Clone, Copy, Default)]
+struct NDArray;
+
+impl<'ctx> StructKind<'ctx> for NDArray {
+    type Fields = NDArrayFields<'ctx>;
+
+    fn iter_fields(&self, mapper: &mut FieldMapper<'ctx>) -> Self::Fields {
+        NDArrayFields {
+            data: mapper.add_auto("data"),
+            itemsize: mapper.add_auto("itemsize"),
+            ndims: mapper.add_auto("ndims"),
+            shape: mapper.add_auto("shape"),
+            strides: mapper.add_auto("strides"),
+        }
+    }
+}
+
+// Example: List.
+//
+// Compared to NDArray, it has generic models.
+pub struct ListFields<'ctx, Item: Model<'ctx>> {
+    items: Field<Ptr<Item>>,
+    len: Field<Int<'ctx, SizeT>>,
+    _phantom: PhantomData<&'ctx ()>,
+}
+
+#[derive(Debug, Clone, Copy, Default)]
+pub struct List<'ctx, Item: Model<'ctx>> {
+    item: Item,
+    _phantom: PhantomData<&'ctx ()>,
+}
+
+impl<'ctx, Item: Model<'ctx> + 'ctx> StructKind<'ctx> for List<'ctx, Item> {
+    type Fields = ListFields<'ctx, Item>;
+
+    fn iter_fields(&self, mapper: &mut FieldMapper<'ctx>) -> Self::Fields {
+        ListFields {
+            items: mapper.add("items", Ptr::new(self.item)),
+            len: mapper.add_auto("len"),
+            _phantom: PhantomData,
+        }
+    }
+}

nac3standalone/demo/demo.c

@@ -7,11 +7,11 @@
 #include <string.h>
 
 double dbl_nan(void) {
-  return NAN;
+    return NAN;
 }
 
 double dbl_inf(void) {
-  return INFINITY;
+    return INFINITY;
 }
 
 void output_bool(bool x) {
@@ -19,19 +19,19 @@ void output_bool(bool x) {
 }
 
 void output_int32(int32_t x) {
-    printf("%"PRId32"\n", x);
+    printf("%" PRId32 "\n", x);
 }
 
 void output_int64(int64_t x) {
-    printf("%"PRId64"\n", x);
+    printf("%" PRId64 "\n", x);
 }
 
 void output_uint32(uint32_t x) {
-    printf("%"PRIu32"\n", x);
+    printf("%" PRIu32 "\n", x);
 }
 
 void output_uint64(uint64_t x) {
-    printf("%"PRIu64"\n", x);
+    printf("%" PRIu64 "\n", x);
 }
 
 void output_float64(double x) {
@@ -52,7 +52,7 @@ void output_range(int32_t range[3]) {
 }
 
 void output_asciiart(int32_t x) {
-    static const char *chars = " .,-:;i+hHM$*#@    ";
+    static const char* chars = " .,-:;i+hHM$*#@    ";
     if (x < 0) {
         putchar('\n');
     } else {
@@ -61,12 +61,12 @@ void output_asciiart(int32_t x) {
 }
 
 struct cslice {
-    void *data;
+    void* data;
     size_t len;
 };
 
-void output_int32_list(struct cslice *slice) {
-    const int32_t *data = (int32_t *) slice->data;
+void output_int32_list(struct cslice* slice) {
+    const int32_t* data = reinterpret_cast<int32_t*>(slice->data);
 
     putchar('[');
     for (size_t i = 0; i < slice->len; ++i) {
@@ -80,23 +80,23 @@ void output_int32_list(struct cslice *slice) {
     putchar('\n');
 }
 
-void output_str(struct cslice *slice) {
-    const char *data = (const char *) slice->data;
+void output_str(struct cslice* slice) {
+    const char* data = reinterpret_cast<const char*>(slice->data);
 
     for (size_t i = 0; i < slice->len; ++i) {
         putchar(data[i]);
     }
 }
 
-void output_strln(struct cslice *slice) {
+void output_strln(struct cslice* slice) {
     output_str(slice);
     putchar('\n');
 }
 
-uint64_t dbg_stack_address(__attribute__((unused)) struct cslice *slice) {
+uint64_t dbg_stack_address(__attribute__((unused)) struct cslice* slice) {
     int i;
-    void *ptr = (void *) &i;
-    return (uintptr_t) ptr;
+    void* ptr = static_cast<void*>(&i);
+    return (uintptr_t)ptr;
 }
 
 uint32_t __nac3_personality(uint32_t state, uint32_t exception_object, uint32_t context) {
@@ -119,11 +119,12 @@ struct Exception {
 
 uint32_t __nac3_raise(struct Exception* e) {
     printf("__nac3_raise called. Exception details:\n");
-    printf("          ID: %"PRIu32"\n", e->id);
-    printf("    Location: %*s:%"PRIu32":%"PRIu32"\n" , (int) e->file.len, (const char*) e->file.data, e->line, e->column);
-    printf("    Function: %*s\n" , (int) e->function.len, (const char*) e->function.data);
-    printf("     Message: \"%*s\"\n" , (int) e->message.len, (const char*) e->message.data);
-    printf("      Params: {0}=%"PRId64", {1}=%"PRId64", {2}=%"PRId64"\n", e->param[0], e->param[1], e->param[2]);
+    printf("          ID: %" PRIu32 "\n", e->id);
+    printf("    Location: %*s:%" PRIu32 ":%" PRIu32 "\n", static_cast<int>(e->file.len),
+           reinterpret_cast<const char*>(e->file.data), e->line, e->column);
+    printf("    Function: %*s\n", static_cast<int>(e->function.len), reinterpret_cast<const char*>(e->function.data));
+    printf("     Message: \"%*s\"\n", static_cast<int>(e->message.len), reinterpret_cast<const char*>(e->message.data));
+    printf("      Params: {0}=%" PRId64 ", {1}=%" PRId64 ", {2}=%" PRId64 "\n", e->param[0], e->param[1], e->param[2]);
     exit(101);
     __builtin_unreachable();
 }

nac3standalone/src/main.rs

@@ -314,6 +314,15 @@ fn main() {
     let resolver =
         Arc::new(Resolver(internal_resolver.clone())) as Arc<dyn SymbolResolver + Send + Sync>;
 
+    let context = inkwell::context::Context::create();
+
+    // Process IRRT
+    let irrt = load_irrt(&context, resolver.as_ref());
+    if emit_llvm {
+        irrt.write_bitcode_to_path(Path::new("irrt.bc"));
+    }
+
+    // Process the Python script
     let parser_result = parser::parse_program(&program, file_name.into()).unwrap();
 
     for stmt in parser_result {
@@ -418,8 +427,8 @@ fn main() {
     registry.add_task(task);
     registry.wait_tasks_complete(handles);
 
+    // Link all modules together into `main`
     let buffers = membuffers.lock();
-    let context = inkwell::context::Context::create();
     let main = context
         .create_module_from_ir(MemoryBuffer::create_from_memory_range(&buffers[0], "main"))
         .unwrap();
@@ -439,12 +448,9 @@ fn main() {
         main.link_in_module(other).unwrap();
     }
 
-    let irrt = load_irrt(&context);
-    if emit_llvm {
-        irrt.write_bitcode_to_path(Path::new("irrt.bc"));
-    }
     main.link_in_module(irrt).unwrap();
 
+    // Private all functions except "run"
     let mut function_iter = main.get_first_function();
     while let Some(func) = function_iter {
         if func.count_basic_blocks() > 0 && func.get_name().to_str().unwrap() != "run" {
@@ -453,6 +459,7 @@ fn main() {
         function_iter = func.get_next_function();
     }
 
+    // Optimize `main`
     let target_machine = llvm_options
         .target
         .create_target_machine(llvm_options.opt_level)
@@ -466,6 +473,7 @@ fn main() {
         panic!("Failed to run optimization for module `main`: {}", err.to_string());
     }
 
+    // Write output
     target_machine
         .write_to_file(&main, FileType::Object, Path::new("module.o"))
         .expect("couldn't write module to file");