Hi, >I've checked MACS2/IO/PeakIO.pyx and I think the typing of the function >call should be fine - so probably the problem is somewhere else - but >where?
I don't have the answer, but upstream seems to be aware of a new bug report https://github.com/taoliu/MACS/issues/140 what I'm really not sure is: why cython Developers and Maintainers are opening RC bugs against packages using it, without a proper transition, and with a cython RC buggy package? cython is not built on s390x because of testsuite failures File "/«PKGBUILDDIR»/build/work-dir/run/c/int_float_builtins_as_casts_T400/int_float_builtins_as_casts_T400.so", line ?, in int_float_builtins_as_casts_T400.__test__.long_double_to_float_int (line 195) Failed example: long_double_to_float_int(4) Expected: 4.0 Got: 0.0 some errors in casting double/float types sounds maybe related to your issue. So, at the end, I would open an RC against cython (FTBFS on s390x), wait for it to be fixed, and then start looking at reverse-dependencies, with a particular eye on cython regressions, specially when the failures looks strange like this one Hope this helps, G.
