Stream: git-wasmtime

Topic: wasmtime / PR #11206 Allow marking data objects as used f...

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 10:26):

bjorn3 opened PR #11206 from bjorn3:used_data_object to bytecodealliance:main:

Fixes https://github.com/bytecodealliance/wasmtime/issues/10758

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 10:26):

bjorn3 requested abrown for a review on PR #11206.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 10:26):

bjorn3 requested wasmtime-compiler-reviewers for a review on PR #11206.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 10:29):

bjorn3 submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 10:29):

bjorn3 created PR review comment:

I couldn't find what is necessary for Windows.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 10:29):

bjorn3 edited PR review comment.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 10:30):

bjorn3 updated PR #11206.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 09 2025 at 11:40):

bjorn3 commented on PR #11206:

Also could this get a backport to the branch for the next Cranelift release. https://github.com/rust-lang/rust/pull/140525 has removed a workaround for #[used] previously defaulting to #[used(compiler)] rather than #[used(linker)], so without the Cranelift changes, .init_array support in cg_clif is broken with lld.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 10 2025 at 17:01):

abrown submitted PR review:

I'm no expert here but this seems reasonable.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 10 2025 at 17:24):

abrown merged PR #11206.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 10 2025 at 18:56):

bjorn3 submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 10 2025 at 18:56):

bjorn3 created PR review comment:

This should have also included elf::SHF_ALLOC | elf::SHF_WRITE (or elf::SHF_ALLOC only when the data object is neither writable nor contains relocations).

view this post on Zulip Wasmtime GitHub notifications bot (Jul 10 2025 at 19:01):

bjorn3 created PR review comment:

And I think this is missing N_WEAK_REF/N_WEAK_DEF for weak symbols.

view this post on Zulip Wasmtime GitHub notifications bot (Jul 10 2025 at 19:01):

bjorn3 submitted PR review.

Last updated: Dec 06 2025 at 07:03 UTC