=== guiverc2 is now known as guiverc
[06:01] <oSoMoN> good morning desktoppers
[06:02] <duflu> Hi oSoMoN 
[06:13] <oSoMoN> hey duflu 
[06:13] <ricotz> hello desktoppers
[06:14] <oSoMoN> hey ricotz 
[06:27] <duflu> Hi ricotz 
[06:47] <seb128> goood morning desktopers!
[07:19] <duflu> Hi seb128 
[07:19] <duflu> Hi diddledani 
[07:19] <duflu> Argh
[07:19] <duflu> Hi didrocks 
[07:20] <seb128> hey duflu, how are you?
[07:20] <duflu> seb128, not bad. I made an effort to actually stop working (mostly) at 6pm yesterday and it really helped. How are you?
[07:21] <seb128> I'm alright, went to play tennis yesterday, manage to warm up and do a few games when it starting raining hard, the courts went so wet that we couldn't resume ... I will do another try tonight!
[07:22] <didrocks> hey duflu, seb128 
[07:23] <seb128> lut didrocks, en forme ?
[08:17] <lissyx> seb128, do we know when https://github.com/snapcore/snapd/pull/12081 will ship?
[08:17] -ubottu:#ubuntu-desktop- Pull 12081 in snapcore/snapd "interfaces: grant access to speech-dispatcher socket (bug 1787245)" [Merged]
[08:17] <lissyx> looking at releases of the project there has been multiple releases since it was merged but it seems none includes it?
[08:17] <oSoMoN> salut seb128, didrocks & lissyx 
[08:18] <lissyx> and it's marked as poart of 2.57 milestone https://github.com/snapcore/snapd/milestone/44 but 2.57.4 does not include it?
[08:23] <seb128> lissyx, right, I asked for a cherrypick and they did it, if you snap refresh --edge snapd it's in there
[08:23] <seb128> lissyx, it should be in stable by next week
[08:23] <lissyx> nice
[08:23] <seb128> lissyx, I pinged jamesh with a socket dir question yesterday but he didn't reply
[08:23] <seb128> I was asking about at-spi but it's an issue for speech-dispatcher as well
[08:24] <seb128> with the new snapd the snap can access the socket, but the socket is in the system dir and no the snap private one XDG_RUNTIME_DIR points to
[08:24] <jamesh> seb128: sorry, I missed that. It shouldn't be a problem in the at-spi case because they get the socket path from a D-Bus call.
[08:24] <seb128> jamesh, ah, that explains
[08:24] <seb128> it is for speech-dispatcher
[08:25] <seb128> so I guess we need another hack in desktop-launch?
[08:25] <lissyx> I though tthe XDG_RUNTIME_DIR was good?
[08:25] <lissyx> if it's in edge I can verify later today
[08:25] <jamesh> if you've got d-feet installed, find org.a11y.Bus and execute the org.a11y.Bus.GetAddress() method
[08:26] <jamesh> yeah. We'd need to update desktop-launch to set an environment variable to find the speech-dispatcher socket
[08:26] <seb128> lissyx, I was trying in gnome-calculator and I get
[08:26] <seb128> $ env | grep XDG_RUN
[08:26] <seb128> XDG_RUNTIME_DIR=/run/user/1000/snap.gnome-calculator
[08:26] <seb128> $ LD_LIBRARY_PATH=./ ./spd-say -L
[08:26] <seb128> Failed to connect to Speech Dispatcher:
[08:26] <seb128> Error: Can't connect to unix socket /run/user/1000/snap.gnome-calculator/speech-dispatcher/speechd.sock: No such file or directory
[08:26] <seb128> so we probably need to export SPEECHD_ADDRESS=unix_socket:/run/user/1000/speech-dispatcher/speechd.sock
[08:26] <seb128> or something similar
[08:27] <lissyx> seb128, dont we also depend on speechd's socket activation support?
[08:27] <lissyx> it recently landed, I dont think it has made it so ubuntu's yet?
[08:28] <seb128> lissyx, we cherrypicked that in kinetic yes
[08:28] <lissyx> (also I was wondering whether I should keep my test env on 22.04 or move to 22.10)
[08:28] <seb128> https://launchpad.net/ubuntu/+source/speech-dispatcher/0.11.3-1ubuntu1
[08:28] <seb128> I'm testing on kinetic
[08:28] <seb128> once it's working I will handle the SRUing
[08:28] <lissyx> ok
[08:29] <seb128> but you can probably rebuild that speech-dispatcher version on 22.04 if you don't want to upgrade
[08:29] <seb128> but for some reason even with SPEECHD_ADDRESS it was not working for me
[08:39] <seb128> no, in fact it works with 
[08:39] <seb128> SPEECHD_ADDRESS=unix_socket:/var/run/user/1000/speech-dispatcher/speechd.sock
[08:39] <seb128> so I guess we need to make desktop-launch set that one
[08:40] <seb128> except I get spd-day to segfault randomly now, but it's the same outside the snap and I don't have the issue with the previous revision
[08:40] <seb128> so I wonder if there is a bug in the socket activation patch :/
[08:40] <seb128> $ spd-say -L
[08:40] <seb128> Invalid language fr-FR!
[08:40] <seb128> Erreur de segmentation (core dumped)
[08:41] <seb128> but it's random, sometime it displays no output and sometime it works
[08:42] <seb128> shrug, and now that I tried outside the snap I got in a state where it doesn't work in the snap anymore
[08:42] <seb128> Error: Can't connect to unix socket /var/run/user/1000/speech-dispatcher/speechd.sock: Connection refused.
[08:42] <seb128> so I think there is bugfixing still needed :/
[08:43] <lissyx> seb128, pretty sure I saw some mention of intermittent bug on CI on this PR
[08:44] <seb128> yes, but they ended up having the same error in CI on master without the change so they acked the merge and decided it was not a regression
[08:44] <seb128> maybe it changes timing in a way that makes it more likely to hit the error though...
[08:46] <oSoMoN> ricotz, FYI, I pushed a patch to the firefox-beta.bionic branch to fix the i386 build failures (https://bazaar.launchpad.net/~mozillateam/firefox/firefox-beta.bionic/revision/1468), and uploaded a .2 package to the firefox-next PPA
[10:37] <ricotz> oSoMoN, thank you!
[10:48] <KGB-2> geary Iain Lane 340728 * commented merge request !2 * https://deb.li/S72n
[14:19] <diddledani> Hi duflu 🤪