Skip to content

Conversation

@cceckman-at-fastly
Copy link
Contributor

Roll all the way up to stable, see if that also reproduces the Windows crash.

Per Dustin's suggestion: update to version-aware resolver in 1.84, to
avoid triggering the incompatible SPDX upgrade. This should also let us
bisect between 1.84 and 1.83
@cceckman-at-fastly cceckman-at-fastly marked this pull request as draft August 11, 2025 20:22
@ulyssa
Copy link
Contributor

ulyssa commented Aug 12, 2025

It's not quite the same (we're seeing exit code: 0xc0000409, STATUS_STACK_BUFFER_OVERRUN, instead of 0xc0000028/STATUS_BAD_STACK), but I'm wondering if this is similar to bytecodealliance/wasmtime#10289, and if we can fix the issue by upgrading to windows-2025 like in bytecodealliance/wasmtime#10290.

@cceckman-at-fastly
Copy link
Contributor Author

Maybe - I do think it's from some of the unwinding stuff that came in 1.84 (#507).

I had wondered how upstream got around it, but hadn't found bytecodealliance/wasmtime#10290, nice!

@cceckman-at-fastly
Copy link
Contributor Author

Eh, panic in a function that cannot unwind is still turning up in trap-test.

@ulyssa
Copy link
Contributor

ulyssa commented Aug 12, 2025

And I see bytecodealliance/wasmtime#10923 about whether C-unwind should show up in more places in wasmtime...

I'll try building with -C force-unwind-tables like in bytecodealliance/wasmtime#11383 in case that helps here.

@ulyssa
Copy link
Contributor

ulyssa commented Aug 12, 2025

Updating to wasmtime 28 in 13ef41b seems to fix the windows issues: https://github.com/fastly/Viceroy/actions/runs/16918127336/job/47937127734

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants