-
-
Notifications
You must be signed in to change notification settings - Fork 7
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
Disable Wayland access #14
Conversation
Testing on debian 11 with system python 3.9, I noticed that Do you want to add a statement to upgrade |
That's odd - the version of pip should be installed at the end of the build. I thought ensurepip always resulted in the most-recent pip... but if it doesn't, then yeah - I guess we should update pip before performing package installs. |
(I'll push an update shortly) |
For posterity....(I added
|
I spent way too long looking at this... When you start the flatpak with Qt debugging enabled, you find this error:
From what I've gathered about Anyway, as for avoiding Wayland and forcing X11....what if instead of letting Qt try to use Wayland, fail, and fallback to X11, we just tell Qt to use X11 from the beginning. If I add Thoughts? |
Also, in looking for prior art for putting PySide6 in to a Flatpak, I only found the information...which may offer inspiration in general for our goals. Coolero: https://github.com/flathub/org.coolero.coolero |
I got nothing :-) My original hope was that Flatpak was going to be a more reliable cross-platform single binary solution than AppImage, but it's clearly... not. Or, if it is, it requires a lot more platform expertise that they're going to great efforts to not document. This is essentially why I've been working on the So - my inclination is that we either merge this, as is, knowing that it's very imperfect, but marginally better than before; or we close it, document the limitations of the flatpak backend, and focus on getting deb/rpm/pacman backends working so we can look at adopting "native platform packaging" as the default for Linux. Thoughts? :-) |
That's fair; I don't think either of these solutions are "great" by any means; so, I'm fine with merging this and disallowing access to Wayland by default given these issues with PySide. Ive also confirmed that Wayland can easily be re-enabled by end-users with My primary takeaway from this exercise is packaging PySide in to a Flatpak is a largely unexplored area. |
In my testing, PySide6 doesn't display window chrome if the wayland socket is open. Ultimately, we're going to want to make these permissions user-configurable; but for now, a set of defaults that work for the demo apps seems preferable.
This also re-orders the build so that bootstrap and resources (which are less likely to change) occur earlier, so they (in theory) should result in faster builds due to not spoiling the build cache.
PR Checklist: