Update the WASI libc to 30094b6ed05f19cee102115215863d185f2db4f0

[jedisct1]

Update our copy of wasi-libc up to commit 30094b6ed05f19cee102115215863d185f2db4f0 of the upstream repository.

[andrewrk]

Would you mind mentioning in the commit message the version number, git tag name, or sha1 of the wasi libc that this updates to?

[kubkon]

LGTM! One question though as I've not been keeping tabs on the libc development: have any new emulations been added that we should be aware of and potentially expose them as configurable build flags for wasi-libc? Thinking here of something like emulated-process-clocks, etc.

[jedisct1]

I think the most recent emulation remains emulated-pid that we already had.

[kubkon]

I think the most recent emulation remains emulated-pid that we already had.

Thanks @jedisct1!

[jedisct1]

2022-06-15T00:38:17.8225490Z compiler [1074/1143] adding_numbers_at_runtime_and_comptime (stage2, x86_64-linux) [2/3] CBE: control flow failed: ChildProcessExecution
2022-06-15T00:38:17.8226360Z 1 tests failed

That doesn't smell good. But that doesn't seem to be related either.

[kubkon]

2022-06-15T00:38:17.8225490Z compiler [1074/1143] adding_numbers_at_runtime_and_comptime (stage2, x86_64-linux) [2/3] CBE: control flow failed: ChildProcessExecution
2022-06-15T00:38:17.8226360Z 1 tests failed

That doesn't smell good. But that doesn't seem to be related either.

Yeah, I don't think this is related in any way. If you look at the CI state for current master we fail in linux CI also in incremental tests for no reason. I wonder if this is related to moving to running tests in parallel introduced in #11862

[andrewrk]

That doesn't smell good. But that doesn't seem to be related either.

Fixed in 8caa206. Unfortunately there are two more bugs floating around that do not yet have reproducers:

• warning(module): unexpected EOF reading cached ZIR when running test-cases multi-threaded #11871
• compiler crash with generic function instantiation, invalid decl_index when adding new instantiation to the set #11848

I think the key move here is going to be investing in expanding on our debug infrastructure in order to diagnose these issues rather than trying to do some kind of heroic one-off debugging effort.