[02:04] RAOF, are you in WA or were just working late? [02:04] duflu: I was just on my phone. [02:05] Ooh, right. IRC/matrix/whatever [05:35] good morning [06:10] good morning desktoppers [06:13] Morning didrocks and oSoMoN [06:15] hey duflu, salut didrocks [06:22] how goes duflu, oSoMoN? [06:22] Going well, didrocks. You? [06:27] doing fine, it's gonna be another beautiful and warm spring day, and hopefully productive, too [06:28] Saturday will be the first day of winter, and 26 degrees :) [06:49] duflu: I'm good, thanks ;) (Sorry, was busy trello-github-carding) === pstolowski|afk is now known as pstolowski [06:59] morning all [06:59] or just didrocks oSoMoN and duflu :) [07:04] Hi marcustomlinson [07:05] hey marcustomlinson, we are "all" ;) [07:06] didrocks: ah, there's a public holiday somewhere today hey? [07:06] marcustomlinson: yes, France & Germany at least [07:07] didrocks: oh yes, and you opted for tomorrow instead. Nice one [07:07] yep ;) [07:17] good morning desktopers [07:17] Morning seb128 [07:18] hey duflu [07:18] hey seb128 [07:18] hey marcustomlinson [07:18] you're allowed to take a break hey ;p [07:25] yeah, I'm going to swap it for tomorrow [07:25] I didn't finish things I wanted to get done yesterday [07:25] so better to not cut context and have a long w.e [08:01] morning! [08:02] good morning Laney [08:02] Morning Laney [08:02] moin didrocks duflu [08:10] hey Laney, how are you? [08:11] yo seb128 [08:15] sorry, got distracted *cough* [08:15] i'm good, we saw a nice band at the folk club last night, so i'm happy about that [08:15] you? [08:30] I'm good! [08:31] we were supposed to go out for dinner but the baby sitter was stucked at another place where the parents had to come back some hours later for some reasons [08:31] so we ended up to not do that and I went to tennis instead, which was nice :) [08:34] brb, changing location [08:49] k, back! [08:50] o/ [08:51] \o [08:51] \o/ [08:52] :-) [08:52] good morning Wimpress [09:20] Laney, Trevinho, duflu, do you have any idea about how https://gitlab.gnome.org/GNOME/gjs/issues/248 could be done technically? [09:20] GNOME issue 248 in gjs "gnome-shell JS errors don't trigger report tools" [1. Enhancement, 2. Needs Information, Closed] [09:21] seb128, I think you need to dump core. So abort() etc [09:22] Or if we are after just a JS stack then maybe exit(1) ?? [09:22] duflu, see the comments, upstream rejected that idea because it would bypass proper JS engine shutdown [09:22] seb128, yes I know. It's meant to bypass shutdown :S [09:22] Sounds like the job for a distro patch [09:23] well, the way we do it for python we get the backtrace without having to bail out/exit [09:23] I don't know how it's done and if the same thing would be possible in JS though [09:23] there's some kind of way to hook into the exception handler [09:23] for python [09:23] do you know if a similar mechanism exists for js? [09:24] nope, but that is presumably something that "could" be done by gjs [09:24] That sounds desirable. gnome-shell logs stack traces all the time without exiting [09:24] nope I don't know (sorry) [09:24] no worry, I asked in case :) [09:24] oSoMoN: jamesh: do you know how snap dependencies work, and if this is the right thing to do https://code.launchpad.net/~ken-vandine/ubuntu-seeds/+git/ubuntu/+merge/368102 ? cause like desktop isos should be offline snap seedable. or something like that. [09:24] (not nope there isn't) [09:25] I don't yet buy Phillip's reasoning though. We are only talking about replacing exit() with abort(). If you see leaks in either case you should ignore them [09:25] Or Philip's [09:25] duflu, ah good point, feel free to comment about that on the gitlab issue [09:25] I mean in the latter case at least [09:26] Valgrind can only detect leaks properly using exit() [09:26] xnox: as I understand it, snapd's seeding process is offline: it expects the set of seeded snaps to fulfil all dependencies [09:26] And to worry about leaking on exceptions only matters for caught exceptions. Uncaught exceptions don't leak because the whole process is gone [09:27] jamesh: yeah, but i don't know how to check which deps are needed. i guess i should download an iso and check? [09:27] I wouldn't get into that kind of discussion in #ubuntu-desktop [09:28] and actually personally I wouldn't push back on ptomato either unless there's someone who is actually going to work on it [09:28] xnox: I guess you'd need to download all the snaps in the seed, and check that (a) anything referenced as a base snap is included, and (b) anything referenced by a content interface plug is included [09:28] it's not likely to result in the problem being addressed [09:28] Sounds simple enough. But I'm not starting on another tangent for now [09:28] k, well let's keep it this way for now then [09:29] we might get back to it if someone finds cycles/interest to poke at the problem [09:29] duflu, Laney, thx for the input [09:29] the core dump you'd get from abort() might not even be very useful btw [09:30] i.e. you might actually want to do something from the js side [09:32] Laney, yeah it's not to use the core dump, but to actually get a problem report [09:32] If there's another way to that then great [09:32] Different topic (almost) - I wonder if we could get problem reports for ALL JavaScript stack traces that don't terminate gnome-shell [09:33] That might be very useful for statistical analysis [09:34] something like what we have with https://docs.python.org/3/library/sys.html#sys.excepthook [09:34] I don't think we want to pop an apport dialog every time we have one of those, at least not today [09:34] that's what apport uses for python [09:34] Fair point. We may drown out crashes and not prioritise them [09:34] and it's what Philip was hinting towards with "Is there some API that could be used to trigger the reporting tool that doesn't shut down the program without cleanup?" [09:34] xnox, what james_h said, base snaps and content interfaces, not sure if snapd exposes this information somehow, without requesting an actual install [09:35] I'll comment with a link to the Python thing [09:35] +1 [09:35] Laney, thx [09:37] jamesh: oSoMoN: so i used the snap download tool from livecd-rootfs, and mounted the snap by hand, and I see inside the snap, meta/snap.yaml has things like: [09:37] gnome-3-28-1804: [09:37] default-provider: gnome-3-28-1804 [09:37] interface: content [09:37] target: $SNAP/gnome-platform [09:37] so i guess i will check all of these manually [09:38] xnox: the default-provider: line is the important part, yeah. also base: at the top level (but that should only reference "core" or "core18") [09:39] xnox, right, unfortunately it’s still very much a manual process, it'd be nice if snapd offered a way to query for "dependencies" [09:39] yeah [09:40] but it's interesting how i can use the plug mechanism to probably provide prebuilt initrds for the kernel snaps [09:40] hmmmmmm [09:49] kenvandine: jamesh: oSoMoN: it looks like gtk-common-themes is the only snap that doesn't declare "base: core18" [09:49] if it did, we would be able to build images with just: snapd core18 [09:50] but because of gtk-common-themes it looks like we will end up with: core core18 [09:50] as the required bases. [09:51] xnox: it could probably be switched to "base: bare". While that is one extra base snap, it is 4K [09:52] oooooh [09:52] i like "base: bare" [09:52] I guess it would also depend on whether snapd is ready to run on classic systems without "core" [09:52] never heard of it. [09:52] i believe it is. [09:53] jamesh: i guess we should experiment with like stable/ubuntu-19.10 channel first (i.e. switch gtk-common-themes to base:bare and check if we can drop dual-core) [09:54] or just test on edge. [09:55] eoan dailies are built with stable/ubuntu-19.10 but i think we can request livefs build with an alternative channel. [09:55] (to do a one off edge build) [10:18] GunnarHj, I sponsored {,ibus-}libpinyin for bionic and they're in the unapproved queue, gonna handle cosmic now [10:32] and sponsored to cosmic too [12:01] xnox, jamesh: that is intentional. We need something to ensure core gets installed. At least at the time 19.04 shipped we still needed that. [12:06] hey kenvandine, how are you? [12:09] Hey seb128 [12:10] I'll check with mvo to see if we can drop that yet [12:38] kenvandine: snap-tool seeds core/core18 as needed in livecd-rootfs..... ditto snapd. can you explain why it is intentional to pull in core? [12:39] The snapd snap wasn't quite ready to work without core installed [12:39] They thought it was, but there was a bug [12:39] Might be fixed now [12:40] When we switched all the snaps to core18 it broke the images [12:41] So we switched the themes to core just as a work around to get 19.04 out the door [12:41] The themes snap doesn't actually need any base [12:41] So that was a workaround [12:53] ah [12:53] thanks! [13:28] Thanks oSoMoN! [13:55] seb128: i'm sorry, can't reply to your email i'm stuck in a loop-one singing and dancing spice girls songs, because your email had trigger words [13:55] xnox, no worry, it's almost friday then you will be able to get some rest and recover! [14:00] seb128: seriously my facebook is impossible. their spice world tour is on now, and i'm just flooded with girl power [14:02] click on the "X" button in the corner and be happy! [14:42] JUJUUUUUUUUUUUUUUUUUUUUUU [14:43] ERROR profile "default": juju does not support IPv6. Disable IPv6 in LXD via: [14:43] lxc network set lxdbr0 ipv6.address none [14:43] /o\ [14:49] :-( [14:49] k, so in case that's interesting to someone, I hacked a sort of bionic/desktop version report mostly for GNOME components [14:50] it's not a proper service and there are weird things like Debian version forced to 0, but it can help to see where we potentially have GNOME components that have a stable update that could be SRUed [14:50] so mentioning it in case it's useful to others, https://people.canonical.com/~platform/desktop/bionic/versions.html [14:50] or https://people.canonical.com/~platform/desktop/bionic/desktop.html for our set [15:54] morning! [15:54] good morning Trevinho [15:59] oSoMoN: hey === pstolowski is now known as pstolowski|afk [17:09] moin and bye Trevinho [17:09] seb128: sorry forgot to say earlier, nice one [17:09] * Laney has been trying to understand the crowd [17:10] > [17:10] byeeee [18:21] Laney, :-) [18:21] k, calling it a day [18:22] which means a week since I'm off tomorrow, have a nice evening/w.e desktopers! see you on monday [18:27] bye seb128, enjoy :) [19:35] enjoy the long week-end seb128 [19:35] good night all