[06:06] good morning desktoppers [06:11] Morning oSoMoN [06:14] hey duflu [06:38] bonjour duflu and oSoMoN [06:40] Hi jibel [06:44] salut jibel [09:03] ho de ho [09:05] hey Laney [09:06] hey oSoMoN [09:06] gggoood morning desktopers [09:12] * Laney nods seb128 [09:58] One more hackfest... [09:58] https://wiki.gnome.org/Hackfests/ShellExtensions2020 [10:08] Trevinho, hey Marco! someone was also mentioning a settings one yesterday I think, unsure if that's real though [10:08] nice to see a snap/flatpak on the list with Canonical suggesting it :-) [10:09] Yeah, nice indeed [10:09] seb128: good, nice to see Hackfests coming. [10:12] mobile user spotted! [10:23] default desktop snap packages are reinstalled during upgrade to focal. Is it really what we want to do? [10:24] removing them before the upgrade makes the upgrader crash. I'm looking for the right way to fix this. [10:24] like removed and installed? [10:24] yeah, removed manually from eoan then upgrade to focal [10:24] no wait [10:25] do you mean: all of the snaps are removed and installed again [10:25] or: if I remove a snap, it is put back onto my system [10:25] I removed them from eoan, upgraded to focal and they are reinstalled by the upgrader. [10:25] indeed, sounds buggy [10:34] jibel: looks like this function needs to check if the .deb was installed before installing the snap? https://git.launchpad.net/ubuntu-release-upgrader/tree/DistUpgrade/DistUpgradeQuirks.py#n515 [10:34] or in _prepare_snap_replacement_data [10:36] neat, it does have code to refresh to the stable/ubuntu-foo channel for the new series, was expecting that to be missing [10:36] Laney, yes, I'll fix it with the other crash [10:36] coolio [10:37] we need to make this support snap2deb too btw, not sure if that interests you at the same time :-) [10:37] just fixing automated tests ATM [10:42] okey [11:38] Laney, do you know offhand if SRU builds with proposed enabled? [11:39] yes they do [11:39] Trevinho, ^ so yeah, backport that vala fix [11:40] ricotz isnt around to complain about the fact that the vala bionic SRU creates build issue for other components [11:40] I will comment on the bug though [11:40] seb128: iirc I enabled it in my schroot as for this, probably Laney was also the one pushing me to do it [11:40] what are you on about [11:42] Laney, https://launchpad.net/ubuntu/+source/vala/0.40.17-0ubuntu1 makes gnome-calculator trigger https://gitlab.gnome.org/GNOME/gnome-calculator/merge_requests/17 [11:42] rather the error fixed by that mr [11:44] Trevinho notices while testing a fix for a regression in his currently gnome-calculator/bionic SRU [11:44] I see [13:19] Laney, seems like the glib update makes glib-networking autopkgtests fail (mentioning since I happened to look at the report and those don't trigger notifications, sorry for the noise if you were already aware) [13:22] yeah I saw thanks, going to look into it [15:12] I get this when I git push to LP [15:12] $ git push lp:~jibel/ubuntu-release-upgrader [15:12] fatal: remote error: Path translation timed out. [15:12] has anyone seen this or am I doing something wrong here? [15:39] jibel: think there's some kind of LP bug regarding that, perhaps try #launchpad [15:59] Laney, that was it, cjwatson fixed it. [16:01] cool [16:10] good morning folks! [16:14] hi hellsworth [16:15] hi jibel! [16:16] hey hellsworth [16:16] what's going on g [16:17] oh not much. getting into the thick of a cold i think [16:17] jsut gotta power through today and tomorrow :) [16:21] sux [16:21] eh it could be worse [16:22] looking forward to the break [17:11] kenvandine, do you think you could squeeze in backporting the commit mentioned on bug #1856927 to focal today if you have some free slot? it makes totem segfault on start on fiscal, I'm away from my main desktop until holidays and managed to screw my key on the laptop :/ [17:11] bug 1856927 in grilo-plugins (Ubuntu) "totem crashed with SIGSEGV in build_flavored_key() from insert_key_mapping()" [High,Fix committed] https://launchpad.net/bugs/1856927 [17:43] seb128: i can try [17:43] got several things scheduled for this afternoon [17:43] kenvandine, thx, if you don't have time don't worry that can wait for after holidays [17:43] i can probably get to it in the morning [17:43] thx === ijohnson is now known as ijohnson|lunch [20:00] does anyone have a good description of what cursors are for X11 apps (from xrestop for example) - I'm guessing it's an allocation to do transitions of some kind with the mouse cursor.. [20:02] to show up differently over different kinds of backgrounds, like over text vs other things.. [20:35] kenvandine, aday was asking about https://gitlab.gnome.org/GNOME/gnome-control-center/issues/681 - is that something you've been involved in? [20:35] I'm assuming willcooke is probably not going to follow that up at the moment :) [20:38] nope :) [20:39] i haven't been involved [21:01] hellsworth, what's the simplest way to build a GTK snap these days? I want to update https://github.com/robert-ancell/multipass-gui to be able to build. [21:01] does the snap not currently build? [21:02] oh it's missing the desktop-gtk3 section or you can use the gnome-3-28 extension [21:02] if you'd like, i can take a go at tryig to make it build.. [21:03] hellsworth, please do :) [21:03] okey dokey [21:03] I think it relied on the method of the day, which I'm guessing is no longer applicable. [21:07] robert_ancell: why is it classic confinement? [21:07] hellsworth, because it has to run the 'multipass' command [21:07] hmm ok so it needs to be classically confined [21:08] i wonder why the gnome-3-28 extension doesn't support classic confinement [21:08] This is not a project that is being worked on, but want it fixed because others are still using it (https://discourse.ubuntu.com/t/multipass-gui/8067/14) and just don't want to leave bad examples around [21:09] fair enough [21:14] hellsworth, I also have a not work related project that I want to snap (https://github.com/robert-ancell/microflash). I had a go back when I made it, but the need to access USB devices was tricky. I really wanted it to be strictly confined rather than classic. I ended up making a Snap like PPA, but it's not as nice. Would love some advice on if that's do-able these days. [21:15] Ah, now I remember. The big issue is registering a MIME type. [21:15] That's kind of essential to make it work. [21:16] i have not yet dealt with registering MIME types in snaps [21:16] kenvandine: what was the thought behind extensions only supporting strict and devmode confinement? [21:16] hellsworth, should you be looking for a simple example to try... :) :) :) [21:17] hint taken robert_ancell :) [21:17] Was I that subtle?? [21:19] robert_ancell: it looks like https://bugs.launchpad.net/snapd/+bug/1849094 is still open and I don't see a workaround [21:19] Ubuntu bug 1849094 in snapd "Feature request: allow snaps to register new mime types" [Undecided,Confirmed] [21:19] subscribed. [21:36] robert_ancell: https://github.com/robert-ancell/multipass-gui/pull/5 [21:36] hellsworth, thanks! [21:36] my pleasure [21:37] hellsworth: can't really work with classic [21:37] but why not [21:39] hellsworth, "Could not find a required package in 'build-packages': python3-distutils" [21:39] oh shoot [21:40] i have a hacked verison of snapcraft installed so let me try agian with the stable snapcraft :) [21:41] robert_ancell: did you see that when trying to build? [21:41] hellsworth, came out of the CI build [21:42] https://travis-ci.org/robert-ancell/multipass-gui/builds/627437203 if you can see that [21:42] yep thanks [21:44] i'm building it now as in the pr, but with the stable snapcraft [21:44] if it builds and runs fine here without adding python3-distutils, then i guess that's something that docker needs? [21:46] hellsworth, yeah, I wasn't sure where that error was coming from. Snapcraft is the only part of the source that has 'build-packages', so I would have thought it was required for that. [21:46] yeah [21:47] it built ok over here with teh stable snapcraft [21:47] hellsworth, actually, I'm using the ubuntudesktop/gnome-3-26-1604 docker image - kenvandine is that no longer appropriate? [21:47] I guess it should be an 18.04 image. [21:48] Not with core18 [21:50] well robert_ancell we could just add python3-distutils to the build section and see if that appeases the docker build? [21:51] it shouldn't affect the final snap anyways [21:52] hellsworth, I'm just thinking if it's not actually required in the snap then it should be done the level above, i.e. either use a newer image of apt install it. [21:53] I'll try switching to https://hub.docker.com/r/ubuntudesktop/gnome-3-28-1804 [21:54] The original issue was actually that it was built against a too old version of GTK for some desktop, so this is probably the right thing to do anyway. [21:56] Oh wait, snapcraft is going to build inside a VM isn't it? So the docker image is not so important. So many layers... [21:57] can you setup travis to use multipass? [21:59] I guess if you install multipass inside the docker image then snapcraft would use it? [22:00] It looks like it's being built locally, but working now. [22:00] Success! [22:02] woo hoo! so just switching your docker did the trick? [22:03] yes [22:03] hot dog! [22:04] Hmm, it's not showing up in snapcraft.io, but Travis said it uploaded it. [22:04] hmm. maybe check back in 5 min? [22:05] Previously it's been instantaneous, and haven't got an email saying it's in a queue. Yeah, hopefully will turn up soon. [22:05] indeed [22:06] hellsworth, thanks for the fixes! [22:07] no problem :) === heather is now known as hellsworth [22:19] ah, it was my shonky Travis config. [22:20] Looks like I need to update SNAP_TOKEN in Travis as well. [22:21] SNAP_TOKEN is just the macaroon, right? [22:22] https://docs.travis-ci.com/user/deployment/snaps/ is pretty vague [22:27] robert there is now an ubuntudesktop-3-28-1804 docker image [22:27] nevermind, you figured that out :) [22:27] yeah, and snapcraft using multipass is great :) [22:30] kenvandine: what would the env var be to add to the override-build section to add the built binary to the prime dir? [22:30] or maybe i just cp it [22:30] no env var necessary [22:30] ok nmv [22:30] yup [22:30] I think you want to use something like $SNAPCRAFT_PRIME [22:30] * kenvandine doesn't recall the env for that [22:31] but the most reliable way to do that [22:32] yeah my problem is that the binary built that lives in parts/cherrytree/build/cherrytree doesn't get primed: https://paste.ubuntu.com/p/2RzfhRk8qp/ [22:32] in fact, the prime directory doesn't have a usr component at all [22:33] so i guess i need to mkdir -p ../prime/usr/bin and then cp it to there, but that seems hacky [22:34] i figured it would be better to just snap the c++ version of cherrytree since the commits show that it's actively being developed, even if the developer is not the most communicative [22:34] rather than rewriting the app [22:34] before i had snapped the python2 version that is no longer seeing develoopment [22:35] are you around tomorrow ken? === ijohnson|lunch is now known as ijohnson === ijohnson is now known as ijohnson|EOY [23:32] hellsworth, The store was unable to accept this snap. [23:32] - confinement 'classic' not allowed with plugs/slots [23:33] I guess you can't make a classic snap that accesses the themes? [23:36] interesting [23:36] well lemme try to remove those plugs/slots and rebuilt [23:36] it was a hunch to add them [23:45] yeah it builds and runs just fine without the plugs [23:46] there are a lot of fontconfig warnings and errors though. i mean they were there even with the plugs. but fyi.. [23:48] oh it looks like you've already removed it [23:48] good on you robert_ancell :) [23:48] i was gonna make another mr but nvm :) [23:49] ok i gotta run. see some of yous tomorrow! [23:50] robert_ancell: if you still have issues with that snap, ping me on telegram so i'll see it [23:51] hellsworth, woohoo, it's uploaded to edge now.