https://bugzilla.redhat.com/show_bug.cgi?id=1931183
--- Comment #25 from Andy Mender <andymenderunix(a)gmail.com> ---
(In reply to Ankur Sinha (FranciscoD) from comment #24)
OK, so more deps:
- exdir
- hdf5storage (surprised this isn't in Fedora yet)
- MEArec
- shybrid
We can disable their tests and proceed in the meantime, and then keep
packaging them up and enabling the tests. I'll start with exdir next.
About the buggy tests, we should report them upstream first and see what
they say. If they're not trivial and upstream can't fix them (or suggest a
workaround/fix) soon, these can also be disabled with a comment that
includes the issue URL. You could take a quick look to see if it's something
obvious (like a change in py3.10 perhaps) and see if it's possible to submit
a patch?
Cheers,
A quick update:
- I forgot to mention last time, but because the test cases run as
unittest-style methods from a test class, it's not possible to silence only
specific tests. This was supposedly fixed upstream in pytest (compatibility
with unittest), but I couldn't get it to work with the suggested workaround.
- spikeextractors was now merged into spikeinterface:
https://github.com/SpikeInterface/spikeinterface
Not sure how this impacts your use case, but the old project will be used for
bug fixes only and submitting new issues is already blocked:
https://github.com/SpikeInterface/spikeextractors
- The exdir unit test still fails on the same assert call as before, because it
somehow fails to recognize the exdir module.
- The value issues in the tests look like something related to h5py or overall
compatibility on the line Python - numpy - h5py.
I think it might be worth opening a new review request for spikeinterface
instead, since that's the umbrella project and perhaps it's also packaged a
little differently. At least I see the test layout is different:
https://github.com/SpikeInterface/spikeinterface/tree/master/spikeinterfa...
I would hold off with opening extra review requests for spikeextractors
dependencies as these may no longer be needed :).
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
https://bugzilla.redhat.com/show_bug.cgi?id=1931183