Package Details: j4-dmenu-desktop 2.17-1

Git Clone URL: https://aur.archlinux.org/j4-dmenu-desktop.git (read-only, click to copy)
Package Base: j4-dmenu-desktop
Description: A much faster replacement for i3-dmenu-desktop. Its purpose is to find .desktop files and offer you a menu to start an application using dmenu.
Upstream URL: https://github.com/enkore/j4-dmenu-desktop
Licenses: GPL3
Conflicts: j4-dmenu-desktop-git
Submitter: justin8
Maintainer: austinpray (physkets)
Last Packager: physkets
Votes: 118
Popularity: 1.85
First Submitted: 2013-12-30 00:14
Last Updated: 2019-03-16 07:02

Pinned Comments

physkets commented on 2019-03-16 07:35

The pkgbuild now performs a signature check, and you will have to import and (locally) sign the key, once.

gpg --search-keys A1774C1B37DC1DCEDB65EE469B8450B91D1362C1
gpg --lsign-key A1774C1B37DC1DCEDB65EE469B8450B91D1362C1

Latest Comments

« First ‹ Previous 1 2 3 4 Next › Last »

justin8 commented on 2015-07-02 23:10

Checksum seems fine to me. I just redownloaded it, and no change.

LeonardK commented on 2015-07-02 17:42

checksum out of date

justin8 commented on 2015-06-23 00:10

Updated to 2.13

bebehei commented on 2015-06-22 15:29

v2.13 is out.

laser_b commented on 2015-06-21 08:17

It seems the md5sum changed, as validation failed. It expects the md5sum to be f657ca76bf27642f4a35db4abb0868ed.

justin8 commented on 2014-05-26 23:18

Thanks for that, I saw your issue opened upstream, Once that's in I can revert it to a basic make once more, for now I'll add your patch to this package.

I would like to leave tests in place where possible since they usually exist for a reason; probably not so bad when we're compiling a tagged release version I suppose.

By the way, when I tested it, doing a single threaded build was only ~0.5 seconds slower than multi-threaded (on a quad-core xeon).

carstene1ns commented on 2014-05-26 16:00

This should fix the problem: https://gist.github.com/carstene1ns/01b7cbfa58d71415a0ac Also, because we compile the test sources, this also runs it in the check() function.
Another fix would be (as already mentioned) to define NO_TESTS, this would skip the git checkout of 'catch' and not compile the test sources.

justin8 commented on 2014-05-25 22:45

That might be something to raise an issue on the upstream bug tracker if that is the case.

As for building in a dirty directory, I'm not sure, that's very hard to make a directory with the same leftover files as someone else after possibly multiple versions, I also suggested it since I tested it and it did work for me, and still does. Catch doesn't appear to compile as it reports no patch/configure/build/install steps, then continues on with the rest of the build process.

@IBeArjen: If you're certain that setting it to use a single threaded build will fix this issue, then I'll gladly add it to the pkgbuild.

carstene1ns commented on 2014-05-25 18:56

'catch' is only required to run some tests that could also be disabled with -DNO_TESTS as we do not run them for building this package anyways (no 'check' function).

But a proper fix would be to fix the cmakelists.txt by adding a hard dependency.

IBeArjen commented on 2014-05-25 16:03

Actually, recompiling it in a _dirty_ directory might work. That, or compiling using "-j1" (i.e. single threaded build). The build process pulls catch from git and starts further steps before getting catch has completed, and fails because of this.