Speed up AppImage execution by running pre-extracted/mounted filesystems
Opened this issue · 3 comments
Pre-submit checks
- I checked for similar issues beforehand, but could not find any. I could not add my proposal to any existing issue.
- I am going to take the time to to fill in all the required details. I know that the bug report will be dismissed otherwise.
Feature description
some AppImages take a long take for me to start up. I'd like them to be faster
Proposed solution
Extract or mount Appimages at some point before running them.
Here are some tests I conducted with Joplin-2.14.17.AppImage (236MiB) (I stop the timing when the image in the default note loads)
Normal AppImage run: 12.9s 12.4s 12.5s
With appimage-mount done before: 8.5s 8.3s 8.5s
With appimage-extract done before: 8.3s 8.3 8.2s
Here are the ways I think we could improve startup time, and their problems:
- premount all AppImages on integration/system startup
- What implication does this have ? Maybe will consumes memory or CPU in background ?
- extract AppImages in a folder on integration
- This will take a lot of disk space. There should be an option to enable/disable that
- Scan all folders for an
AppRun
in the AppImage folder (so that users can manually extract AppImage there, and they will still be integrated)- I don't see any downside in this, but it could be combined with other options
Alternative solutions you considered
idk
Additional context
No response
You might wanna read this:
AppImage/AppImageKit#478 (comment)
This is mostly an issue with appimagetool setting to compression too high.
Thanks ! hope this get resolved. I can help if needed
Thanks ! hope this get resolved. I can help if needed
Well in that case it would be changing the default compression level and block size that appimagetool uses to zstd level 1 and whatever block size mksquashfs uses by default.