[PRC] How may I be of help?

[JJ]

I have been assigned this module for the September pull request challenge. Do you have something in mind I could help with?

[danaj]

Hi JJ, thanks for the help!

There are some generic things that apply to all projects. Quality checks in the documentation (spelling, missing, malformed, unclear, ambiguous, etc.). Build procedures. Test coverage, test performance, etc.

There is a big list of things in the TODO file, but most of those items are either not really high priority or not obvious. Many are only of interest if you really like doing the math and/or low level coding. I'd love it, but I don't expect it. They are in mostly chronilogical order, so things near the top have been sitting there for years, while things at the bottom are new.

If you're ambitious and know XS, I am stuck on how to use multicall in the forfactored call. I can find no examples that seem to be doing something similar. The XS documentation says it is possible but very tricky.

This month I'm working on changing some of the floating point routines to use quadmath if possible. I also have a bug report about the t/26-pillai.t test basically hanging. I cannot reproduce but will keep trying. I changed the test a bit so it should run faster, but 12+ hours is just ludicrously long so there is probably something else going on (the failing test machine was HP/UX 64-bit Itanium). Adding a timer to the test was the suggested mitigation, which certainly is the friendly thing to do though doesn't find the actual cause.

[JJ]

Tests work OK in 5.28, I'm not good at all with XS. Will see what I can do, but it really looks like an impressive, and flawless, work.
Test coverage might be something, but then part of the code is in XS, which I don't if it's examined or not. Will try that, writing tests should not be complicated.

[JJ]

Coverage can certainly be improved.

[JJ]

I'm rerunning it with the GMP modules installed. Most coverage problems seem to come, anyway, from errors that have not been tested.