Fedora Account System
Red Hat Associate
Red Hat Customer
python-obspy failed to build from source in Fedora rawhide/f45 https://koji.fedoraproject.org/koji/taskinfo?taskID=147901421 For details on the mass rebuild see: https://fedoraproject.org/wiki/Fedora_45_Mass_Rebuild Please fix python-obspy at your earliest convenience and set the bug's status to ASSIGNED when you start fixing it. If the bug remains in NEW state for 8 weeks, python-obspy will be orphaned. Before branching of Fedora 46, python-obspy will be retired, if it still fails to build. For more details on the FTBFS policy, please visit: https://docs.fedoraproject.org/en-US/fesco/Fails_to_build_from_source_Fails_to_install/
Created attachment 2151431 [details] build.log file build.log too big, will only attach last 32768 bytes
Created attachment 2151432 [details] root.log file root.log too big, will only attach last 32768 bytes
Created attachment 2151433 [details] state.log
In https://bodhi.fedoraproject.org/updates/FEDORA-2026-2551d0d66b we have shipped Python 3.15.0b4. All packages that link to libpython or have extension modules should be rebuilt, or they will segfault on import/usage. Packages using stable ABI might not need a rebuild. This package was not rebuilt because we were unable to do so. Please rebuild it with Python 3.15.0b4+ ASAP. Thanks
Fixed in python-obspy-1.5.0-4.fc45. For the record, the failure was not caused by Python 3.15 but by numpy 2.5: linalg.eig() and linalg.eigvals() now always return complex arrays, where previously the dtype depended on whether the eigenvalues happened to be real. ObsPy passed those eigenvectors on to ufuncs that only accept real input, which raised e.g. TypeError: ufunc 'degrees' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe'' That produced 30 test failures in obspy/io/nordic/tests/test_nordic.py and obspy/imaging/tests/test_beachball.py. See https://numpy.org/devdocs/release/2.5.0-notes.html#linalg-eig-and-linalg-eigvals-now-always-return-complex-arrays Fixed by backporting upstream PR obspy/obspy#3760, which adds a drop-in replacement for numpy.linalg.eig restoring the pre-2.5 behaviour. The fix is already on the upstream maintenance_1.5.x branch and should ship in 1.5.1, so the downstream patch can be dropped on the next rebase.