Stream: git-wasmtime

Topic: wasmtime / PR #12457 fix: support `fmin` and `fmax` in in...

Wasmtime GitHub notifications bot (Jan 28 2026 at 00:27):

tommyscholly opened PR #12457 from tommyscholly:fmin_interp to bytecodealliance:main:

This addresses #11233, where neither fmin nor fmax were supported in the interpreter because is_nan and other method are not implemented on vector types.

Wasmtime GitHub notifications bot (Jan 28 2026 at 00:27):

tommyscholly requested cfallin for a review on PR #12457.

Wasmtime GitHub notifications bot (Jan 28 2026 at 00:27):

tommyscholly requested wasmtime-compiler-reviewers for a review on PR #12457.

Wasmtime GitHub notifications bot (Jan 28 2026 at 00:34):

cfallin commented on PR #12457:

Thanks -- can you update or add a runtest showing these opcodes working in the interpreter with vector inputs?

Wasmtime GitHub notifications bot (Jan 28 2026 at 01:01):

tommyscholly updated PR #12457.

Wasmtime GitHub notifications bot (Jan 28 2026 at 02:30):

github-actions[bot] added the label cranelift on PR #12457.

Wasmtime GitHub notifications bot (Jan 28 2026 at 04:23):

tommyscholly updated PR #12457.

Wasmtime GitHub notifications bot (Jan 28 2026 at 04:45):

tommyscholly updated PR #12457.

Wasmtime GitHub notifications bot (Jan 28 2026 at 22:12):

cfallin commented on PR #12457:

Thanks for the update. A few more requests:

• Since this isn't a fuzzbug/weird corner case regression but rather ordinary opcode implementation coverage, let's name the runtest something descriptive like fmin-vector.clif + fmax-vector.clif rather than issue-11233.clif.
• It's always best to look for existing runtests to add the target to first; I don't see any that test vector forms of these ops (e.g. fmin.clif is scalar-only). That said, since we're adding runtests for the opcodes, let's add all platforms that support them while we're at it. #11233 implies that x86-64 has lowerings too -- could you add it to the runtest if so?
• Could you make updates to your PR with additional commits rather than a squash and force-push? It will all be squashed together in the end when we merge.

Thanks!

Wasmtime GitHub notifications bot (Jan 29 2026 at 01:52):

tommyscholly updated PR #12457.

Wasmtime GitHub notifications bot (Jan 29 2026 at 01:54):

tommyscholly commented on PR #12457:

Did all of those, thanks for the feedback!

Wasmtime GitHub notifications bot (Jan 29 2026 at 01:57):

cfallin submitted PR review:

Thanks -- LGTM!

Wasmtime GitHub notifications bot (Jan 29 2026 at 01:57):

cfallin has enabled auto merge for PR #12457.

Wasmtime GitHub notifications bot (Jan 29 2026 at 02:14):

cfallin added PR #12457 fix: support fmin and fmax in interpreter on vector types to the merge queue.

Wasmtime GitHub notifications bot (Jan 29 2026 at 02:37):

cfallin merged PR #12457.

Wasmtime GitHub notifications bot (Jan 29 2026 at 02:37):

cfallin removed PR #12457 fix: support fmin and fmax in interpreter on vector types from the merge queue.

Last updated: Jan 29 2026 at 13:25 UTC