Stream: git-wasmtime

Topic: wasmtime / issue #13114 Should we define GC-related `VM` ...

Wasmtime GitHub notifications bot (Apr 15 2026 at 21:36):

fitzgen opened issue #13114:

We could probably even move the VM*Header and VMGcRef definitions into wasmtime-environ since VMGcRef is always 32 bits regardless of the target pointer size.

_Originally posted by @fitzgen in https://github.com/bytecodealliance/wasmtime/pull/13107#discussion_r3089116958_

Wasmtime GitHub notifications bot (Apr 15 2026 at 21:38):

fitzgen commented on issue #13114:

One thing to note is that the current VMDrcHeader and VMCopyingHeader types are defined in a module with visibility where they can only be used by the correct collector, and we have some unsafe trait implementations that rely on this visibility, and would instead become something like // SAFETY: nothing uses these when they shouldn't, which is much less satisfying.

But maybe is worth it to be able to just use offset_of! instead of defining constants manually and adding tests to assert that the constants match the actual struct's field's offsets?

Wasmtime GitHub notifications bot (Apr 23 2026 at 16:48):

alexcrichton added the wasm-proposal:gc label to Issue #13114.

Last updated: May 03 2026 at 22:13 UTC