-
-
Notifications
You must be signed in to change notification settings - Fork 2.2k
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
Move non-x86 tests to GHA #5088
Conversation
Pillow uses the images from Docker that have been pushed by docker-images stable jobs. So when the next Pillow release occurs, docker-images stable jobs will have an updated version of Pillow that specifies 0.7.1 in 'depends/install_raqm.sh' and then it will be updated to 0.7.1 here. |
Oh right, I was looking at the latest jobs, not stable. |
Thank you! Similar to my comment at python-pillow/docker-images#98 (comment), these timings aren't too bad (especially if Travis is essentially dead!), and it's worth trying it out. Let's also put these slower ones first in the matrix here as well. |
I reordered the images and opened #5091 with just the xfail part of this. |
For #5028 (comment).
Companion to python-pillow/docker-images#98.
Includes #5091.
Move tests of arm64v8, ppc64le, s390x from Travis CI to GHA Docker tests using aptman/qus.
Other changes:
test_libimagequant_quantize
failing on ppc64le (not sure why).The new tests will fail until the new images get pushed, see run in my branch for now: https://github.com/nulano/Pillow/actions/runs/416245841
Speed comparison (approximate)
*: The ppc64le and s390x image builds currently fail a bit early due to new test failures, the install time will be slightly longer.
The total is a bit faster than I saw in #5028 (comment) mostly because it installs WebP from
apt-get
instead of compiling locally. More importantly, only the test times should matter, which are in line with cache-miss Windows tests and slightly slower than macOS tests.Dependency versions:
Travis CI (removed):
Ubuntu Focal in Docker (python-pillow/docker-images#98):