afonso360 opened PR #4914 from disable-nan-verifier
to main
:
We are currently running the verifier twice, once after the nan canonicalization pass, and again when JIT compiling the code.
The verifier first runs in the NaN Canonicalization pass. If it fails it prevents us from getting a nice
cargo fuzz fmt
test case.So disable the verifier there, but ensure its enabled when JIT compiling.
@jameysharp I think this solves the issue in this comment: https://github.com/bytecodealliance/wasmtime/pull/4896#pullrequestreview-1104415283
I think it makes sense to keep NaN canonicalization in fuzzgen, otherwise the testcase reported in
fmt
is different than the one that is actually executed and that can be really confusing.
afonso360 submitted PR review.
afonso360 created PR review comment:
Instead of doing this we can also explicitly run the verifier here, I don't know which one is preferable.
afonso360 edited PR #4914 from disable-nan-verifier
to main
:
We are currently running the verifier twice, once after the nan canonicalization pass, and again when JIT compiling the code.
The verifier first runs in the NaN Canonicalization pass. If it fails it prevents us from getting a nice
cargo fuzz fmt
test case.So disable the verifier there, but ensure its enabled when JIT compiling.
@jameysharp I think this solves the issue in this comment: https://github.com/bytecodealliance/wasmtime/pull/4896#pullrequestreview-1104415283
I also think it makes sense to keep NaN canonicalization in fuzzgen, otherwise the testcase reported in
fmt
is different than the one that is actually executed and that can be really confusing.
jameysharp submitted PR review.
afonso360 updated PR #4914 from disable-nan-verifier
to main
.
jameysharp has enabled auto merge for PR #4914.
jameysharp merged PR #4914.
Last updated: Dec 23 2024 at 13:07 UTC