Arch Linux User Repository

If the patch file fails the checksum validity check, try clearing your cache and rebuilding the package before flagging the package.

The libseat dependency needs to be changed to libseat.so: community/seatd 0.5.0-1 provided libseat, but community/seatd 0.6.1-2 has changed to providing libseat.so=1-64 instead.

(There’s still aur/seatd 0.6.1-1 which does still provide libseat, but there’s a pending request on that package which I presume will be a deletion request, because of the conflict with community/seatd which does the same thing but a smidgeon better in its compilation options and PKGBUILD.)

The patch file is failing the checksum when clean (so there’s no cache issue). I suspect that every time wlroots’s master branch gets new commits, it’s going to break, though I haven’t confirmed that hypothesis.

I’d say you’re going about it the wrong way. You can’t properly do a checksum on a dynamically-fetched patch.

I see three solutions:

1. SKIP that checksum too. These patches will now potentially fall out of date and out of sync, so clean rebuilds (or removing the diff file locally) will be needed to be sure of being up to date.

2. Track MisterDA:xwayland_hidpi instead of swaywm:master. The meaning of *-git is obscure when there are two things you could be tracking! Arguably, the current behaviour is more like wlroots-git-hidpi-git…

3. Add MisterDA:xwayland_hidpi as a second Git source, and replace the patch application in prepare with cherry-picking from this second repository. Oh, and probably making pkgver pull from both repositories (.r[swaywm].g[swaywm].r[MisterDA].g[MisterDA]). This is the most correct way to do things with the fewest negative side-effects (the main one is downloading everything twice, but that’s not harmful, whereas losing the ability to track updates is bad). If you’re interested, I can figure this out and provide a patch.

If the patch file fails the checksum validity check, try clearing your cache and rebuilding the package before flagging the package.

@LawnGnome plot twist: MisterDA, Rucikir and me are the same person using different nicknames. I've rebased, thanks!

This just broke with https://github.com/swaywm/wlroots/commit/1eb38e001503636a8e5058051e9256d70f1bffcc — the patch no longer applies cleanly since the WLR_HAS_XCB_ERRORS constant changed to HAS_XCB_ERRORS.

As a very hacky stopgap, this will build again by replacing the constant name in the patch, but I presume the right fix here would be to ask MisterDA to rebase his branch on top of the current wlroots master?

It seems that after latest update Android Studio and Android Emulator don't present any windows on screen. They don't crash, they just don't draw any windows.

@Rucikir Yes you are right, sorry for the incoveniences. I always forget that yay keeps caches for git packages.

@Yegorius The package is not outdated, it correctly builds wlroots master 526ae594. The checksum for the diff is correct. Try building it from a clean state, clear your cache.

@zrl yes, the branch doesn’t merge on master since https://github.com/swaywm/wlroots/commit/27609ba0d9d140660f0cc077ad7af8670fa0ed21. My rebase could be enabled by editing the PKGBUILD, but for now I’ve made it the default. I’ll change back to the PR if something new happens.

Hi, I'm having trouble building this:

==> Starting build()...
patching file include/wlr/xwayland.h
Hunk #1 succeeded at 27 (offset -3 lines).
Hunk #2 succeeded at 240 (offset 33 lines).
patching file include/xwayland/xwm.h
patching file xwayland/xwayland.c
Hunk #1 FAILED at 409.
Hunk #2 FAILED at 436.
2 out of 2 hunks FAILED -- saving rejects to file xwayland/xwayland.c.rej
patching file xwayland/xwm.c
Hunk #5 succeeded at 1452 (offset 2 lines).
Hunk #6 succeeded at 1524 (offset 2 lines).
Hunk #7 succeeded at 1556 (offset 2 lines).
Hunk #8 succeeded at 1884 (offset -2 lines).
==> ERROR: A failure occurred in build().
    Aborting...
Error making: wlroots-hidpi-git