It is not a platform I have run any test on.

As far as I can see, several failures come from the shell or the signal handling in the tests not behaving as expected. Those tests run correctly on all the Linix/macOS platforms tested in the Github test grid, unless specifically disabled (but most likely they are only disabled on Windows and they are reasonably portable on *nix architectures).

The datetime failure is curious, especially because there are datetime tests much more extensive and comprehensive that passed no problem... or not? I don't see tests/types/test_datetime.py in your logs? I wonder why.

If you have MR to propose one-off, no problem accepting them. If you want us to consider the riscv64 platform supported you should add it to the CI test grid so we can monitor regressions.

Thank you for having a look. I'm not sure why tests/types/test_datetime.py did not run. I'll try to do some digging. I don't believe any tests are disabled in the RPM spec.

I agree the tests are not showing any issue on other platforms. Those tests passed here; https://koji.fedoraproject.org/koji/buildinfo?buildID=2518656 on others. Do you think there are any chances we're hitting a timeout? The riscv64 hardware at present is not particulary fast. We can skip them easily enough to build, but I thought I'd try to do a little digging in my spare time as well.

I apologize if my questions are a bit basic. I just got pypy bootstrapped for Fedora 41 on riscv64 and figured I'd build this and asv, which are the two packages currently dependent on pypy in Fedora and I have not done a lot of digging yet.