This happens in both master and core-updates, although I don't know if the former fails for the same reason. It also happens in both real machine and virtual machine. ## Steps to reproduce Assuming you are using the Guix System with GNOME: 1. Go to Activities → Show Applications (nine dots button) → All 2. Click on Maps ## Expected result GNOME Maps starts. ## Unexpected result GNOME Maps does not starts. You see the GNOME Maps application menu loading next to the "Activities" button in the GNOME desktop for a few seconds, but the application never starts. ## Additional information In both, real and virtual machines, GNOME Maps starts when run from a terminal. It prints the following message though: ``` $ gnome-maps (org.gnome.Maps:22887): Gjs-WARNING **: 13:35:50.971: JS ERROR: Gio.IOErrorEnum: Ya existe un objeto exportado para la interfaz org.gnome.Sysprof3.Profiler en /org/gtk/Profiler vfunc_dbus_unregister@resource:///org/gnome/Maps/js/application.js:269:9 main@resource:///org/gnome/Maps/js/main.js:57:12 run@resource:///org/gnome/gjs/modules/package.js:222:12 start@resource:///org/gnome/gjs/modules/package.js:206:5 @/gnu/store/agrz0vryg6ipcz4qh3q1addwdiizbkrw-gnome-maps-3.34.2/share/gnome-maps/org.gnome.Maps:2:1 ``` In both, real and virtual machines, GNOME Maps also starts when run from the minishell (Alt+F2 and type gnome-maps). ## System information ``` GNOME Maps 3.34.1 $ LANG=C guix describe Generation 78 May 02 2020 18:00:20 (current) guix 7b07852 repository URL: https://git.savannah.gnu.org/git/guix.git branch: core-updates commit: 7b07852ddb334c92bcef69666f21c599f1f0fa79 sirgazil-x ef6ac93 repository URL: https://gitlab.com/sirgazil/guix-channel-x.git branch: master commit: ef6ac9331a7847fdec6f0eb199524b1b755ba0df ``` --- https://sirgazil.bitbucket.io/
[-- Attachment #1: Type: text/plain, Size: 336 bytes --] sirgazil via Bug reports for GNU Guix <bug-guix@gnu.org> writes: > This happens in both master and core-updates, although I don't know if the former fails for the same reason. It also happens in both real machine and virtual machine. Fixed on core-updates with commit 7d7fedf6f022e6114959684658f927e191ab40b1. Thanks for the report! [-- Attachment #2: signature.asc --] [-- Type: application/pgp-signature, Size: 487 bytes --]
I still can reproduce the problem though, in both real and virtual machines. ★★★★★★★★★★★★★★★★★★★★ $ LANG=C guix describe Generation 80 May 08 2020 15:18:18 (current) sirgazil-x ef6ac93 repository URL: https://gitlab.com/sirgazil/guix-channel-x.git branch: master commit: ef6ac9331a7847fdec6f0eb199524b1b755ba0df guix 35c43fc repository URL: https://git.savannah.gnu.org/git/guix.git branch: master commit: 35c43fcdbb408a5755efebc13241dd9082360aa1 ★★★★★★★★★★★★★★★★★★★★
reopen 41058
retitle GNOME Maps fails to start in GNOME
thanks
sirgazil via Bug reports for GNU Guix <bug-guix@gnu.org> writes:
> I still can reproduce the problem though, in both real and virtual machines.
Oh, I had missed that you mentioned it worked from a terminal (it did
not for me on i3).
I hope one of the GNOME users around here can have a look.
[-- Attachment #1: Type: text/plain, Size: 753 bytes --] On Sun, 10 May 2020, Marius Bakke wrote: > sirgazil via Bug reports for GNU Guix <bug-guix@gnu.org> writes: > >> I still can reproduce the problem though, in both real and virtual machines. > > Oh, I had missed that you mentioned it worked from a terminal (it did > not for me on i3). > > I hope one of the GNOME users around here can have a look. I think that this is probably the same issue as: https://issues.guix.gnu.org/issue/39324 In that case the problem application was GNOME-Weather. Pierre worked around it there by calling the executable directly in the .desktop files rather than using gapplication launch… It would be nice to figure out what's going wrong with gapplication launch though. Should the issues be merged? Best, Jack
[-- Attachment #1: Type: text/plain, Size: 891 bytes --] I think, I figured out the root cause of this. GApplication launches apps through dbus, looking up the respective service in share/dbus-1. The service for GNOME Maps does not start the actual program, but much rather a gjs wrapper, which we don't patch -- and which we can't simply patch, because that leads to other errors, that rival our own when it comes to how cryptic they are. I've attached a patch, in which I rewrite the DBUS service to use the actual program. I have not fully tested this patch, as `guix environment --ad-hoc gnome-maps` appears not to be strong enough to influence the gapplication lookup. However, if you launch `gnome-maps --gapplication-service` in one terminal, then quickly switch to some other and launch `gnome-maps` normally, the one that spawned the service is the one that gets used, so claiming the bus in this fashion *should* work. Regards, Leo [-- Attachment #2: 0001-gnu-gnome-maps-Patch-DBUS-service.patch --] [-- Type: text/x-patch, Size: 1157 bytes --] From d709692ba02e7f3a4c2f9bd8ca43374131a81a7a Mon Sep 17 00:00:00 2001 From: Leo Prikler <leo.prikler@student.tugraz.at> Date: Mon, 11 May 2020 13:30:44 +0200 Subject: [PATCH] gnu: gnome-maps: Patch DBUS service. * gnu/packages/geo.scm (gnome-maps): [arguments]<#:phases>: Add patch-dbus-service. --- gnu/packages/geo.scm | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/gnu/packages/geo.scm b/gnu/packages/geo.scm index 6c68d40..b0d33d9 100644 --- a/gnu/packages/geo.scm +++ b/gnu/packages/geo.scm @@ -160,6 +160,12 @@ topology functions.") (substitute* "meson_post_install.py" (("gtk-update-icon-cache") "true")) #t)) + (add-after 'unpack 'patch-dbus-service + (lambda* (#:key outputs #:allow-other-keys) + (substitute* "data/org.gnome.Maps.service.in" + (("@pkgdatadir@/org.gnome.Maps") + (string-append (assoc-ref outputs "out") "/bin/gnome-maps"))) + #t)) (add-after 'install 'wrap (lambda* (#:key inputs outputs #:allow-other-keys) (let ((out (assoc-ref outputs "out")) -- libgit2 1.0.0
Bumping this, as it has been months since I've posted my fix. For the record, it might also be worth looking into #39324, which uses a different workaround, and seeing whether the dbus service can be patched in a similar manner. Regards, Leo
Hello Guix, I've just pushed a variant of the patch I proposed as a2d5f9f60d0f440e5d5ce1a1cc81743b27d6cede. I tested it yesterday in a VM; it should work. Regards, Leo