- Looked into a user-reported potential memory issue (#22947). It turns out to be a use-after-free problem in the user code. That being said, the existing guidelines need some reminder about this pitfall, it’s been added to this week’s ghc-wasm-meta update (ghc-wasm-meta!14).
- Looked into a user-reported build failure on aarch64-darwin (#22875). The exact failure couldn’t be reproduced, but I did find another build failure when cross-compiling base to wasm from aarch64-darwin host (#22981).
- Fixed a bug in IPE codegen logic that affects all registerised no-tables-next-to-code builds, including the wasm backend (!9967). The bug would make source locations be dropped in the generated code.
Previous update: https://discourse.haskell.org/t/ghc-webassembly-weekly-update-2023-02-08