Package Details: sway-git r6506.1be66c98-2

Git Clone URL: (read-only, click to copy)
Package Base: sway-git
Description: Tiling Wayland compositor and replacement for the i3 window manager
Upstream URL:
Keywords: wayland
Licenses: MIT
Conflicts: sway
Provides: sway
Submitter: SirCmpwn
Maintainer: Rucikir
Last Packager: Rucikir
Votes: 133
Popularity: 2.35
First Submitted: 2015-08-16 14:05
Last Updated: 2020-10-24 00:34

Dependencies (25)

Required by (23)

Sources (2)

Latest Comments

1 2 3 4 5 6 ... Next › Last »

Rucikir commented on 2020-10-24 00:34

Your reasoning feels right; if deleting the build cache breaks the "debug, not stripped" package, then we shouldn't force that option by default. Having a second debug package is safer. Thanks for raising that issue.

Also, @SirCmpwn wrote that contrib scripts should not be included, neither here, nor in the stable package. I’m not here to start a meaningless war, my personal opinion is that the scripts come in handy, and I’m going to follow the stable package on that point.

EDIT: oh, I see there are AUR packages for these scripts. Well… in that case, dropping contrib seems fair.

Ferdi265 commented on 2020-10-23 21:50

Hello @Rucikir,

thanks for taking over maintainership :)

Since the previous maintainer never addressed my question I asked 10 months ago, I'm gonna add it here again. I'd really like to know your take on this:

-- begin old question from 2020-01-09 17:33 --

Is there any reason why the PKGBUILD overrides the strip option to off?

Having both debug and strip activated produces a separate debug package that will be installed to /usr/src/debug and /usr/lib/debug.

Having debug and !strip activated produces a single packages with binaries containing debug info, which reference the build folder however.

Many AUR helpers clean the build directory after packaging, which removes the sources, making the debug infos often less useful in case of a crash.

If the strip option is not overridden to off, users are able to choose on their own, whether they want a separate debug package or not.

-- end old question from 2020-01-09 17:33 --

I ended up creating a second AUR package, sway-no-strip-override-git for my needs, but in the long run, it would be very helpful if the official sway-git package would not break my workflow.

If you do decide to keep the override, I will most likely continue maintaining sway-no-strip-override-git, but change it to be only a small patch on top of this one.

Rucikir commented on 2020-10-23 21:06

Hello everyone,

I’ve picked up the maintainership of this package. My intend is to follow what is done for the community/sway package, plus the few quirks we need for developing and debugging sway. I also maintain sway-hidpi-git.

@VannTenn: done.

@j.r: I think that's a bad idea for sway, for instance the 1.5 tag is not present in the master branch. As of now, the latest tag in master is 1.5-rc2. That isn't useful. The version could be extracted from the meson file, but I think that the commit id is sufficient.

VannTen commented on 2020-10-19 10:17


Would you consider adding in that package the script used in the official sway package (this one for importing env variable into the systemd environment (particularly the user instance) ?

This would allow dbus-activated services (in my example, that was mako) to start without trouble.

Thank you very much.

j.r commented on 2020-08-06 06:30

As the upstream repo has tags you should consider to use printf "%s" "$(git describe --long --tags | sed 's/\([^-]*-\)g/r\1/;s/-/./g')" for the pkgver function.

Fandekasp commented on 2020-07-16 00:34

After reading this article and the PKGBUILD, I understand that the pkgver() function returns the proper hash (line 4 "pkgver=r5942.3f77591b" is confusing and could be removed).

Reading up on yay documentation, I understand that it wasn't updating the hash cache db (yay -Y --gendb).

Sorry for the disturb.

SirCmpwn commented on 2020-07-16 00:06

That's not how -git packages work.

Fandekasp commented on 2020-07-16 00:05

@SirCmpwn hi, thanks for the quick reply. I realized that yay never prompts for updates on sway-git because you're not pinning the git commit hash. In other aur repos, a hook exists to auto-update the PKGBUILD file with the latest commit hash on master, which helps understanding that an update is available. Could you consider adding this feature? Or am I missing some other way that people use to update this kind of PKGBUILD tracking master branches?

SirCmpwn commented on 2020-07-15 23:59

I don't know what you expect. This package tracks master.

rpodgorny commented on 2020-05-18 17:10

@SirCmpwn: i'd also like to hear the reasoning behind the "strip" and "debug" override. ;-) (i also think it should go) thanks!