07:34 <phlebas> Hodgestar: i wonder if something like Coverage is really worthwhile - it seems so tied to internals, I wonder if HPy should really aspire to replace the C API for that kind of extension at the current stage

08:30 <Hodgestar> phlebas: It is a good question. I kept it in the list because it's a whole new area of the API we haven't looked at. Presumably HPy would have to cover it eventually.

08:30 <Hodgestar> Perhaps I should add the document to the HPy wiki and then we can all edit it and leave observations more easily.

12:39 <Hodgestar> https://github.com/hpyproject/hpy/pull/237 -- documentation cleanup PR (thank you mattip for catching the failing build ages ago -- sorry it took so long to start sorting it out)

17:09 <ronan> I'm wondering what test_cant_use_closed_handle() is actually supposed to do on pypy. In the use-after-close part, there is no way to reach a valid object from a closed handle, so HPy_Repr() can only crash

19:21 <antocuni> ronan: yes, I think you are right

19:22 <antocuni> note that by default, if you use a closed handle the process aborts with a fatal error, so it "works" also on pypy

19:23 <antocuni> the on_invalid_handle callback is there only to be able to write this test, basically. But it's more a test of the debug mode itself than a "generic" HPy test

19:23 <antocuni> I think we can just skip it on pypy