IRC channel logs


back to list of logs

<rekado>welp, time’s up for me. This will remain a mystery for a little while longer.
<civodul>uh, thanks for debugging this far already!
***rekado_ is now known as rekado
<PurpleSym>rekado: I saw you updated Jupyter in Guix proper, but disabled most test suites. guix-science has tests enabled on all packages, so I’ll try to merge that into Guix proper.
<PurpleSym>I’m using a python-jupyter-client-bootstrap to bootstrap python-ipykernel-bootstrap, which then goes into the real python-jupyter-client as native input.
<rekado>disabled most test suites…? Oh, because of the cycles.
<rekado>good, good. I’d be happy to see that fixed.
<efraim>sometimes I wonder if it'd be worth trying to create a build-as-bootstrap-input procedure to automagically generate some packages
<efraim>no native-inputs, disable all tests, I'll remove the rest of the inputs (or add them back in) as necessary
<zimoun>efraim: I just saw your report on GitHub about zstd. Are you building core-updates on PowerPC?
<PurpleSym>efraim: Yeah, I’d be nice if we could “classify” inputs (aka feature flags) for that, so we can easily separate build inputs from test inputs in native-inputs.
<zimoun>It seems the test-suite raises a warning then set to 19 instead of 22 for some reasons. And 19 is the buggy value.
<efraim>Yeah, it's how we did the other ones in the past and I've been sitting on this one for years
<efraim>This time I think it's about ready
<zimoun>efraim: about zstd «Warning : compression level higher than max, reduced to 19», could you try with the value 16 instead of 22? <> Otherwise, you can remove this particular test.
<efraim>Sure, I'll try it with 16
<efraim>Right after dinner :)