-
Notifications
You must be signed in to change notification settings - Fork 438
Build system and test suite update #759
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
I saw something similar occur when I rebased #754. The CI tests worked on my local branch but failed in python-control (even though the code was identical) with the DISPLAY error. I re-ran the tests in python-control and they completed, so figured it was a one-time glitch. I also noticed there was a warning about mambaforge, but I'm not able to find it. Something about how something was an experimental implementation and resolution of some dependencies may not be repeatable. |
I also saw it in my local branch several times. It fails, but which of the many jobs fail is random (see #756) |
d189ce1
to
a4766c1
Compare
Switching back from mamba to conda gets rid of the "Mamba support is still experimental and can result in differently solved environments!" warning you were mentioning @murrayrm, but even that didn't help: https://github.com/bnavigator/python-control/commits/update-CI |
I give up. I think the random "invalid DISPLAY" and segmentation faults are not caused by us but rather either by the conda environment or the Github Actions runners. This might even resolve itself through the next release cycle of either of these. I would say we just ignore the occassional random CI job fail and look at the remaining successful ones for determining whether a PR is good. We can always re-run a job and see if the failure is reproducible. |
e6f8e4d
to
6b07235
Compare
The X Display errors seem to be gone. New failures with the balred routines using conda Slycot producing possibly valid but differently ordered systems. Needs to be checked. |
Odd that Slycot from source is OK, but the other Slycots aren't... |
It's conda slycot using conda openblas vs. ubuntu netlib lapack/blas |
d08b0ae
to
9db91cf
Compare
Whack-a-mole! Now this cvxopt test is flaky: https://github.com/python-control/python-control/runs/7930805491?check_suite_focus=true#step:6:3450
|
35db505
to
7d6fdd8
Compare
7d6fdd8
to
7f13a07
Compare
7f13a07
to
69fb34a
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks good to go, but a few small questions as I looked through the code.
setup.py install
is deprecated. Use this opportunity to migrate to PEP517 style packaging: fixes Building wheel for dev version fails #758QtAgg
backend for the zoom toolbar in the extra full conda test run