New comment by tornaria on void-packages repository

https://github.com/void-linux/void-packages/pull/34030#issuecomment-1002212369

Comment:
@dkwo: are you using `m4ri` from the repo with musl 1.2.2? If that is the case, can you recompile it and try again with the one built for 1.2.2?

This is what should happen:
 - The test is makes `m4ri` to allocate more memory than available.
 - On failure to allocate `m4ri` calls `abort()` which signals `SIGABRT`.
 - cysignals traps the `SIGABRT` signal and then it uses `longjmp(3)` to roll back to the `sig_str()` function which then raises `RuntimeError` with the given error message.

I don't know why cysignals is raising a different exception than `RuntimeError`; I wonder if there's any reason for `abort()` to behave different on musl-1.2.2 (perhaps when the program was compiled on musl-1.1).

As for the timeout in matrix2.py, it's just a test that is intended to be labeled `# long` so it shouldn't run, but it runs nevertheless (reported it upstream: https://trac.sagemath.org/ticket/33091). It's very easy to workaround just for that test, which I'll do in the next update.