GNOME Bugzilla – Bug 711359
Should have a way to silent "Couldn't connect" message
Last modified: 2021-07-05 10:48:37 UTC
Hello, A Debian user reported he was annoyed by the "Couldn't connect to accessibility bus" warning when running a gtk application through ssh. More details on http://bugs.debian.org/728385 AIUI, ideally we'd have accessibility working over ssh, but we don't yet. The warning makes sense anyway, but it is also annoying for users who do not need accessibility features. Perhaps we could have an environment variable that people could use to silent the warning if they really wish so? Samuel
I've eventually found on http://debbugs.gnu.org/cgi/bugreport.cgi?bug=15154 that NO_AT_BRIDGE=1 does this. Perhaps there could be a reference in the warning.
(In reply to comment #1) > I've eventually found on http://debbugs.gnu.org/cgi/bugreport.cgi?bug=15154 > that NO_AT_BRIDGE=1 does this. What NO_AT_BRIDGE=1 does is avoid to load the atk bridge. In short, atk_bridge_adaptor_init will just return early without doing anything. So yes, in the practice, that would avoid you getting that message. But that would mean that accessibility would be disabled. The main idea behind this envvar is debugging. > Perhaps there could be a reference in the > warning. I don't get this. Are you suggesting to add something like "Couldn't not connect to accessibility bus, use NO_AT_BRIDGE on your system if you don't want to use accessibility technologies"?
If I understand correctly, there is a warning because accessibility technologies do not work (under some conditions). So, if the user says that he doesn't want these accessibility technologies, he doesn't lose anything, i.e. the only difference is that he doesn't get the warning. Am I right? Of course, setting NO_AT_BRIDGE is just a workaround, but currently useful.
(In reply to comment #3) > If I understand correctly, there is a warning because accessibility > technologies do not work (under some conditions). So, if the user says that he > doesn't want these accessibility technologies, he doesn't lose anything, i.e. > the only difference is that he doesn't get the warning. Am I right? When a gtk application starts, in order to be accessible, it initializes the atk bridge. This bridge tries to find the accessibility bus, because accessibility technologies uses dbus in order to expose information and events to the accessible tools. In this case, as you are connecting through ssh, you are running an application that is not technically local. So if we want this to work, the remote application would need to connect to the local bus. And as far as I understand, if that is not possible, is because a limitation on DBUS itself, not because at-spi2 lacks to implement something already available: http://www.freedesktop.org/wiki/Software/DBusRemote/ > Of course, setting NO_AT_BRIDGE is just a workaround, but currently useful. Was just guessing about how that workaround would be used. NO_AT_BRIDGE is a envvar that the bridge checks. You opened this bug against at-spi2, so, how are you proposing to use that envvar here? PS: bridge is initialized on at-spi2-atk, moving to proper component.
> Are you suggesting to add something like "Couldn't not connect to accessibility bus, use NO_AT_BRIDGE on your system if you don't want to use accessibility technologies"? As I said in the original report: yes, the intent is to just have a way to avoid the warning when one does not need accessibility. The warning is emited by at-spi2-core, that's why the component was at-spi2-core, since that's where the warning would have to be modified. Now, it would be odd to make at-spi2-core mention a variable taken into account by at-spi2-atk, so a more coherent solution would be to have a variable to make *at-spi2-core* discard initialization without emitting a warning, so that the warning could mention it.
I opened the bug (on the Debian BTS, which was forwarded here by Samuel Thibault) against at-spi2-core just because the warning message came from it. The NO_AT_BRIDGE trick was just initially suggested on http://askubuntu.com/questions/227515/terminal-warning-when-opening-a-file-in-gedit (but then, I don't know the internals) which I found after a search concerning another "accessibility bus" warning, this time SSH wasn't involved at all. I don't know whether there is a bug somewhere triggering this warning or it should just be silenced like here. FYI: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=732618
GNOME is going to shut down bugzilla.gnome.org in favor of gitlab.gnome.org. As part of that, we are mass-closing older open tickets in bugzilla.gnome.org which have not seen updates for a longer time (resources are unfortunately quite limited so not every ticket can get handled). If you can still reproduce the situation described in this ticket in a recent and supported software version, then please follow https://wiki.gnome.org/GettingInTouch/BugReportingGuidelines and create a new ticket at https://gitlab.gnome.org/GNOME/at-spi2-atk/-/issues/ Thank you for your understanding and your help.