• Joined on 2021-03-22
LPTK commented on issue M-Labs/nac3#51 2022-04-19 17:25:55 +08:00
escape analysis

That seems like a pretty nice tradeoff!

Here are some remarks off the top of my head:

Abstract Interpretation

I see that you are using the term "unification" a lot. Here I assume…

LPTK opened issue M-Labs/nac3#144 2021-12-21 17:17:06 +08:00
Add optional types
LPTK commented on issue M-Labs/nac3#52 2021-12-21 16:47:26 +08:00
Objects with non-primitive fields

See related issue: M-Labs/nac3#51

LPTK opened issue M-Labs/nac3#143 2021-12-21 16:44:27 +08:00
Sanity check for primitive operations type signatures
LPTK commented on issue M-Labs/nac3#52 2021-11-08 18:50:11 +08:00
Objects with non-primitive fields

I agree inlining is not a satisfactory solution. We'd still have the problem for all other functions, and there's no particular reason to have special treatment for constructors.…