Force pillow version to 9.5 - #106
Conversation
There was a problem hiding this comment.
Thanks for opening this pull request! Be sure to follow the pull request template!
|
I am a bot, here are the test results for this PR:
|
|
This pull request has been automatically marked as stale because it has not had recent activity. This might be due to missing feedback from OP. It will be closed if no further activity occurs. Thank you for your contributions. |
|
putting a comment so maybe it won't be marked as stale anymore? I've taken to publishing my own image (but have to manually update it, would be good if upstream worked). but like I said in the original PR text, pinning pillow or installing imagemagick both fix it. imagemagick is actually the preferred image resizing solution in the beets code, so that's what i'm using in my image, but it does increase the image size significantly (and i assume it's available for arm, but don't know). |
|
Imagemagick is about 100Mb extra, but https://github.com/linuxserver/docker-mods/tree/universal-package-install is an option if it's preferred. I've finally had time to look at our broken 3.18 rebase PRs, fix them, and include your changes from this PR. I'm going to try and get them merged today. |
|
Going to see what I can do with Imagemagick instead of Pillow and the nightly branch, it'll add about 70Mb to the image so but that's perfectly workable if it provides significant benefits. |
|
Any chance you could give |
|
Not the OP and by no means an exhaustive test, but I changed my docker compose from an older image that was from before issues with PIL to yours: ...and I just processed an album without the artwork errors I was getting from all the recent 'latest' images. |
|
Stable branch is now pinned to Pillow 9.5.0, nightly is using Imagemagick. If/when there's a new upstream stable release we'll roll Imagemagick into that too. |
Description:
There are various crashes in the embedart plugin, as it's not compatible with Pillow 10. The worst of these blocks every import, since it's using parameters to resize images that don't exist in the code anymore.
I feel like the images should probably install ImageMagick instead of falling back to pillow, but I don't have a way to test the non-x86 image.
Benefits of this PR and context:
How Has This Been Tested?
I manually rolled back to pillow 9.5 on my image, which allows imports to continue without crashing
Source / References: