10:13 <fangerer> mattip: antocuni: next (rather small) PR is up for review: https://github.com/hpyproject/hpy/pull/385 (Enable slots HPy_tp_str and HPy_tp_hash)

10:54 <mattip> lgtm. Waiting for CI to pass

10:59 <fangerer> thank you

12:01 <mattip> wow, the valgrind test is exceptionally slow

12:05 <mattip> ubuntu is ~8 minutes, macos is ~15 min, windows ~32 min, valgrind 1hr 50 min and counting

12:51 <fangerer> yes, I know 😞

12:53 <fangerer> I'm thinking of splitting it up. We could clone the job and run each _mode_ (i.e. universal, cpython, ...) in a separate instance.

12:53 <fangerer> but that would increase the overall costs (in terms of CI minutes) a bit

15:12 <Hodgestar> We could also run valgrind nightly and on request.

15:13 <Hodgestar> Although it has sometimes picked up bugs for us in individual PRs, which is nice.

15:47 <antocuni> catching bug early >> faster tests

15:48 <antocuni> if we completely delete all our tests, the speed of our CI jobs would skyrocket 😂

15:48 <fangerer> 😃

15:48 <Hodgestar> Yes, but two hours is a pretty long time to wait.

15:48 <fangerer> I would also keep it but as I said, I think we can split the job and run the tests in parallel

15:49 <Hodgestar> We should definitely keep it, but it would be great to also avoid the two hours.

15:49 <fangerer> agreed. I'll try to improve the situation (tomorrow)

15:51 <antocuni> maybe we can configure it in such a way that the build becomes green even if valgrind is still running

15:51 <fangerer> since I have you here: please remind that we will have our dev call on Thursday (Dec 1st, 10a CET / 9a UTC)

16:14 <antocuni> ouch, I forgot about it. I won't be able to attend because of a doctor visit at the exact same time