Hi Étienne On Fri, 23 Dec 2022 at 20:03, Étienne Mollier <emoll...@emlwks999.eu> wrote: > Hmn, I'm dry on the python-pysam failure to build from source > with python3.11, and upstream remained silent for now[1]. If I > were dealing with C code, I would have suspected something like > an uncaught segmentation fault, or uninitialized memory usage, > but cython seems to be a slightly different beast, however > similar it might sound. > > [1]: https://github.com/pysam-developers/pysam/issues/1151 > > As far as I can tell, the purpose of the test is to check the > behavior of an iterator once it ran out of values, to make sure > ValueErrors can be sent to callers. The test code seems to rely > on the iterator to return NULL once out, or a memory address in > a normal situation. But according to the debugger, the address > obtained is 0x100000000 (or 2^32), and for some reason, I don't > even manage to get an ugly workaround checking for that value > running as expected. > > so I formally request help on this one,
What about skipping this test for now? I tried this, and it allowed at least the three reverse-dependencies I tried (kinetic stools, python-pybedtools, umis) to build with Python 3.11. Regards Graham
