[05:23] Good morning [05:23] pitti: Good morning. [05:23] hey TheMuso, long time no see! how are you? [05:23] pitti: Very well thanks, yourself [05:23] ? [05:24] TheMuso: quite fine, thanks; shepherding distro tests these days, and refining the tools for that :) [05:27] pitti: Will you be in Malta? [05:27] TheMuso: yes I will, in the second week [05:28] Thats the last week of May right? I'll be there in the last week of May. [05:28] right, me too; 23rd to 30st [05:29] Yeah same as me, I'll probably see you around at some point. [07:46] pitti, are you familiar with the sbuild launchpad stuff? [07:46] is it possible to add aliases to that will pull in PPA's? [07:46] darkxst: I'm fairly familiar with sbuild, and I know that LP uses it; what do you mean? [07:46] oh [07:46] I saw a G+ post not too long ago about someone who wrote clever hooks [07:49] darkxst: https://plus.google.com/u/0/+ChristopherHalseRogers/posts/24T9xSgTMEF [07:49] darkxst: https://plus.google.com/u/0/+St%C3%A9phaneGraber/posts/KPTfpLKKF8J [07:49] those might be useful [07:51] pitti, yeh I am using the sbuild-launchpad-chroot [07:52] but then again, sbuild hooks are rather simple to write [07:54] pitti, right, the alias' are defined in a horrible shell script [07:54] I somehow assumed they would be pulled from lp [07:59] darkxst: yes, that's supposed to download LP's schroot tarballs [08:06] good morning desktopers! [08:07] hey seb128! [08:07] didrocks, lut ;-) [08:07] hallo [08:08] hey Laney, how are you? [08:08] pitti, yes it downloads a tarball, but seems the actual apt sources are just generated by lots of awk-ish [08:14] seb128: hi, do you happen to know what "CRITICAL - findToken(): disabled account "ubuntuone" 1 " means in ubuntu-system-settings? I'm trying to add my u1 account on my n4 with utopic and its grayed out :/ [08:14] hey Laney! [08:15] hey seb128 & didrocks [08:15] seb128: it looks just like when I was playing with the u8/mir session on the desktop, not sure though the log message is the same. any hints how I can still add my u1 account on the phone? I need it to to the tests for a new click release [08:15] I'm alright thanks, you? [08:15] hey seb128 [08:15] * pitti waves to didrocks and Laney too [08:15] I'm good, thanks! still a little bit cold here, but at least sunny [08:15] guten morgen pitti :) [08:15] howdy pitti [08:16] we're supposed to have 17° and sun today [08:16] luxury [08:17] hey Laney and didrocks [08:17] hey mvo! [08:17] Laney: waow ;) [08:17] mvo, hey, no, try asking mardy? [08:17] hey pitti [08:18] Laney, what happened to the rain?! [08:18] I told its mum that it was being mean to me and it got sent home [08:18] hehe [08:19] I hope it doesn't leave around here [08:19] so far we have sun here as well at least ;-) [10:28] rickspencer3: your boot failure, does it fail to boot or does it not boot into xorg? [10:29] mlankhorst, I think it fails to boot into xorg [10:30] can you try utopic? it has an updated mesa [10:32] from what i can tell your pc should just work, does ssh work? [10:35] if so, try apport-collect -p xorg 1319149 [10:42] rickspencer3: for the touchscreen you need qtbase-opensource-src from the sru queue :-) [10:42] mlankhorst, I could not try utopic, as I had a lot of trouble getting the installer o run [10:43] mlankhorst, unfortunately, I won't be able to debug much today, I have calls all day [10:43] ok you can work on utopic, that's fine [10:44] mlankhorst, you are certain it will work utopic? [10:44] I'm not so sure [10:44] neither [10:44] what exactly is the issue? [10:44] mlankhorst, first, uniquity does not load 9 times out of 10 [10:45] it stops at the screen with Ubuntu and the dots [10:45] then, when I did get Trusty to install ... [10:45] it booted once or twice, but also several times booted to a black screen [10:45] though I could boot into recovery mode [10:45] ok === MacSlow is now known as MacSlow|lunch [10:55] good morning desktoppers! [10:55] seb128, didrocks, and others in Europe who are up o/ [10:55] moin! [10:55] Sweet5hark, [10:55] ! [10:55] moin you say? [10:55] ===coffee activity=== [10:56] * Action: rickspencer3 to sip coffee [10:56] rickspencer3: Its a universal greeting in northern germany. ;) [10:57] Sweet5hark, nice to know! [10:57] * rickspencer3 presumes Sweet5hark refers to "moin" and not * to sip coffee [10:57] indeed. [10:58] hey rickspencer3! [10:58] * Sweet5hark sips his sugar-and-caffeine based coffee surrogate ... [10:59] bonjour didrocks [11:01] * rickspencer3 wonders why seb128 is avoiding him [11:02] must be trying to figure out a way to get the Unity 8/Mir Desktop image off the plans [11:02] * rickspencer3 starts escalating the requirements [11:03] must self drive my car [11:04] actually we have some prelimenary images there [11:04] next step is that (IIRC) seb128 was going to boot one of them on his machine and see what more stuff we need to add to get it to work [11:04] then see about getting it built in the normal way with dailies [11:11] rickspencer3, salut [11:11] moin, seb128 [11:11] ;) [11:12] rickspencer3, sorry but there was food before you in the line ;-) [11:12] seb128, maybe I should install a web cam in my house so you can see when I am getting coffee and be ready for my jokes in the morning? [11:12] lol [11:13] rickspencer3, I'm still amazed you manage to get up that early btw! [11:15] seb128, it's 7:15am [11:15] I live on teh East Coast now :) [11:15] rickspencer3, right, but I saw joining like an hour ago [11:16] hehe [11:16] saw you* [11:16] right [11:16] * didrocks notes that seb128 put food before manager. Bad for you review dude! [11:16] didrocks, next reviews are in a year right? [11:16] when I get up I check my email for irc for fires first thing ;) [11:16] but then we are going to have a new manager anyway :p [11:16] lol [11:16] yeah, you can "too late!" :) [11:16] seb128, knows he can do no wrong in my eyes [11:17] * seb128 hugs rickspencer3 [11:17] * rickspencer3 hugs seb128 [11:18] Laney, testing the iso on real hardware and tweaking the seed is still on my todolist [11:18] I didn't manage to get to it after the meeting yesterday, I had accumulated some pings and backlog [11:18] I hope to do that this afternoon ;-) [11:18] seb128, when you do you think we can install that? [11:18] couplle of weeks? [11:18] rickspencer3, boot and test, by Malta [11:18] seb128, sweet [11:19] I'll have my new 11.6 inch dell with me [11:19] we can convert it [11:19] install ... let's see, with some luck that's working without needing to tweak too much [11:19] good [11:19] install should work if live does ... [11:19] which one is that? inspiron 11? [11:19] * Laney makes a bold statement [11:19] * Laney regrets it [11:19] lol [11:20] Laney, I was thinking the same but I went for the "with some luck" statement ;-) [11:25] Laney: to boldly go where no unittest has gone before! [11:25] seb128, yeah, I got the inspiron 11 [11:25] I was told that it "just works" [11:25] rickspencer3, it does indeed ;-) [11:26] my plan is to be all Unity 8/Mir asap [11:26] rickspencer3, that's the machine I'm testing unity8/Mir on btw [11:26] :0 [11:26] so, I will need chromium [11:26] which supposedly already has a Mir backend [11:26] qengho probably knows about that [11:26] and I will need Qt, which supposedly already works :) [11:26] right [11:26] seb128, I'm setting up a call so that a few of us can discuss Gtk [11:27] ok [11:27] I'm getting slightly different stories from different people, so I thought we could just work it out on a call quickly [11:27] would it make sense to wait for Malta to discuss it more? [11:27] k [11:27] seb128, no, I would like to have at least a rough plan by Malta [11:27] so we can work the plan at the sprint :) [11:27] k [11:30] * darkxst is also wondering what is happening with gtk+ [11:31] you mean? [11:31] upstream? [11:32] seb128, no, I know what is happening upstream! in utopic! [11:35] what do you want to know about GTK in utopic? [11:40] seb128, what is blocking 3.12 would be a good start! [11:43] it's being tested, we discussed it at the weekly meeting here (logs are available if you are interested) [11:43] the main problem at the moment in GtkHeaderBar use in dialogs [11:43] larsu is working on that to see what we can do to make them work on e.g Unity [11:44] one of the changes required is to update our themes [11:44] Lars said he's going to work on that [11:44] we might need some Unity work as well [11:44] then we need to decide if that's going enough or we need to patch GTK as well [11:45] we definitely need support from unity for border-less windows [11:45] it can't resize them right now [11:45] and we'll need a patch if we want traditional title bars on dialogs [11:46] seb128, right, I see, can't really help with unity/theming issues [11:46] but gtk patches are more probable [11:47] I would actually attend your meetings... if they weren't at a god awful time for me! [11:47] darkxst: what do you mean by "more probable" [11:48] larsu, that I could probably help there [11:48] ah, right [11:48] (as long as it doesn't involve theming!) [11:48] you're really afraid of theming ;) [11:49] tbh, most of this work involves theming at least a bit [11:50] larsu, nope, just find C much more readable than CSS! [11:51] ha, it's almost the same syntax man ;) [11:54] larsu, syntax is not the problem, it just feels like I end up wasting hours anytime I start messing CSS [11:54] (same for the couple of websites I have written) [11:55] right, I know the feeling [11:55] it's possible to write good css, but almost noone does it [11:56] larsu, its possible to write good code in any language, but there are a bunch where no-one does! [11:57] C isn't necessarily a good example of that, though [11:59] bah [11:59] I think GNOME C is pretty good in general [11:59] so i just got a firefox update that told me to restart the browser with a popup ... [11:59] when i clicked it didnt restart [12:00] and now trying to start FF i only get a message about FF already running === MacSlow|lunch is now known as MacSlow [12:00] but certainly I have seen horror code in many languages [12:00] * ogra_ wouldnt like to lose the ~300 tabs he has open across various windows [12:00] ogra_: firefox should keep those even if it crashes (it shows you a "restore last session?" tab) [12:00] darkxst: fair enoguh [12:01] *enough [12:01] larsu, well, it shows me "another instance is already running" ... [12:01] ogra_: ya, kill -9 it first [12:01] i'm looking for a safe way out of that :) [12:01] ok [12:01] wait. [12:01] don't make me responsible [12:01] if you loose all those tabs [12:01] hah [12:02] to late ... now i will blame you :) [12:02] ogra_, why do you need 300 tabs? [12:02] oh, wow [12:02] 3188 ? Sl 1799:52 /usr/lib/firefox/firefox [12:02] having 300 tabs open is some kind of insanity anyway [12:02] that did run since quite some time [12:02] well, i do actual work on my machine :P [12:03] hm, new gtk also loses images in messages dialogs and centers its labels... [12:03] * larsu wonders in how far we want to go to revert those [12:03] seb128: ? [12:03] phew [12:03] all fine ... [12:03] ogra_: awesome :) [12:03] it always restores all windows on one workspace though ... but thats something to blame unity for i guess ... not FF [12:04] larsu, the less reverts the better, maybe check with design/mpt, what they don't flag "no no no" can stay? [12:05] seb128: good idea. I'll note those down and schedule a call / session in Malta with mpt [12:06] larsu, looks like mpt should be in Malta the same week as we are, good ;-) [12:06] yep [12:06] \o/ [12:14] * Laney tries g-t with the transparency patch [12:15] wheeeee [12:15] it's a bit buggy [12:15] it doesn't live update [12:16] you have to toggle "use colours from system theme" to make changes take effect [12:16] ugh [12:16] that's annoying [12:16] yeah probably just some missing signal bindings though [12:18] I can have a look at that later [12:31] seb128: would there be a chance in Malta for me to present a widely a crazy lunatic vision for Ubuntu? [12:32] libreoffice replacing unity as shell? [12:34] Sweet5hark, to who? the team? Mark? ;-) [12:35] will you show us the LibO click packages finally ? [12:37] seb128: the team, not (yet) Mark. It would be about packaging, so foundations would likely be the ones to consider it ultimately ... [12:37] Sweet5hark, yeah, I'm sure we can organize that [12:38] humpf [12:38] Laney, would you take your rain back, please? [12:38] no ! ... keep it in france ... we dont want it up here north [12:39] please put your request in writing and it will be filed and processed in due course [12:39] * Laney sets up the special processing area [12:39] more red tape! [12:40] :-( [12:41] ogra_: hah, actually, it will most likely say click-packages are not the solution as they only solve a subset of the problem. ;) The idea goes much deeper, but might solve issues on a more fundamental level -- but it would be really something for the long run. [12:41] Sweet5hark, well, the phone image will eventually take over the desktop [12:42] i woudl actually expect us to move away from debs over the next few years [12:42] (for endusers that is) [12:43] using image based updates on the desktop (which means readonly rootfs) etc etc [12:43] ogra_: yeah, the phone image might be a starting point of what I had in mind. And also yeah: debs are not part of that way to deploy anymore. ;) [12:43] ogra_, that's going to be challenging [12:43] ogra_, did you read https://lists.ubuntu.com/archives/technical-board/2014-March/001856.html ? [12:44] seb128, why ? just drop debs [12:44] yes [12:44] ogra_, if we "just drop debs" we don't have an usable desktop anymore [12:44] we need to find a way to convert those to clicks first [12:44] we just need to leave the old cruft behind ... use debs for building images, give them to developers that run in rw mode and perhaps to server people [12:44] ogra_, except if you don't care for libreoffice, firefox, chromium, gimp, inkscape, etc [12:45] for the rest just make the desktop identical to what we have on the phone today [12:45] seb128, right, then you switch your image to writable mode and use it like you did before [12:46] I though that was not supported? [12:46] for the typical desktop user the image based approach works better ... [12:46] "typical" [12:46] it is not supported in the specific image design we use on the phones [12:46] I guess you mean "if we include libreoffice, firefox, etc in the base image"? [12:47] if you have an actual readonly partition instead of a bunch of loop mounted img files you can indeed do it in a supported way [12:47] seb128, no, these should be clicks [12:47] built from plain upstream ... without distro hacks we have to care for etc [12:47] no more maintainer scripts ... [12:48] system upgrades done in minutes instead of hours with predictable installations etc [12:48] right [12:48] well, we "just" have to convert our archive to clicks [12:49] if we feel we are facing this question: we could reach out to firefox and libreoffice and i am sure with some moderate convincing power they might be happy to feed into our clickstore if that gives them the ability to ship what they want [12:49] nah [12:49] seb128, orga_: are you both in Malta? from what I see you are both wrong ;) [12:49] we only have to convert the UI apps to clicks [12:49] Sweet5hark, lol [12:49] ogra_, it's still a large stack of apps [12:49] seb128, sure [12:50] thats why i said "within the next years" :) [12:50] right, fair enough [12:50] asac, and sure, engaging with upstream to ramp that up is going to be useful/needed [12:50] i think the ones that have upstreams that distribute binaries its not that hard. for those that only distribute sources and are not into the binary business like gnome, its indeed a bit more tricky [12:51] why would you use binaries ? [12:51] do we have a good story on "how to get your app published in the click store" for upstreams? === mjohnson151 is now known as mjohnson15 [12:51] * didrocks sees clouds now… wonder if it's because of juju or seb128 sending them my way now… [12:51] just use the source ... we have it already ... you just need to make the archive build a click package alongside [12:51] seb128: yes, just upload :) [12:51] asac: using upstream binaries is insane for anything nontrivial [12:51] seb128: i know mozilla has been waiting for this ability for ages [12:51] seb128, yeah, just push your locally built binary [12:52] mozilla etc. have infrastructure to validate that their build works on ubuntu [12:52] etc. [12:52] they dont have many users, but if they would have thea bility to really distribute their stuff alongside our distro [12:53] they surely would even invest more to make their experience shine by default on ubuntu [12:53] anyway, its clear we probably need a deb story in the click/syustem image world somehow. afaik mvo and foudnations are working on that [12:55] we dont :P [12:55] we need to make sure that if you switch away from system image you can use debs [12:55] ogra_: we dont? [12:56] one source tree to rule them all :) [12:56] imho we dont need any deb support in system-image setups [12:56] but I have already made the point multiple times :p [12:56] didrocks, well, the source tree is orthogonal to the apps distribution problem [12:56] ogra_: well, so guess we could say that and pair it with the ability to convert things into your own system image [12:56] to re-lock down [12:56] :) [12:56] lie what was on canoni-tech [12:57] fi you have system-image you go with clicks ... if you want debs you switch to readwrite ... and cant go back [12:57] didrocks, having one unified source tree doesn't give you libreoffice packages in the click store [12:57] but then debs need to be properly supported indeed [12:57] disable system image, use apt to fiddle, snapshot a system image, easy distribute :) [12:57] seb128: you should the subdirectory built binary and project that to the installed system [12:57] ogra_: the part of the story where you cant go bad just doesnt feel good enough for our standards :) [12:57] asac, and how would that work with the server side generated deltas ? [12:57] sounds like a technical detail [12:57] :) [12:57] lol [12:58] i am not saying that story is fully developed [12:58] well, the design wont allow that [12:58] but its one vector to maybe look at this [12:58] my vector is that 80% of the users should be fine with s-i [12:58] yes, for thtat i agree :) [12:58] and the other 20 will happily just go with apt-get dist-upgarde [12:58] actually 95% at least [12:58] and stay there [12:59] we just need to make sure that both work fine and that the switch works properly [12:59] i wouldnt provide a switch to go back .. [12:59] maybe its the right way to look at it [12:59] buit lets see; the more we think the better our solutions get [13:00] larsu: I uploaded it to ppa:laney/experimental if you want to take a look [13:00] * Laney biab [13:00] and we have just started thinking :) [13:00] hehe [13:01] ogra_, well, your 80% of users being happy assume that we have clicks for libreoffice, chromium, firefox, gimp, pidgin, etc? [13:01] seb128: of course [13:01] seb128: we want t store full of useful killer apps [13:01] open source, proprietary, games etc. [13:01] seb128, yes, which means our builders need to roll clicks alongside the debs [13:02] that seems like a pre-require [13:02] how they get into the store and who is doing that is kind of secondary and technical detail i think. [13:02] before using system images for desktop [13:03] I'm not sure anyone will be happy about rebooting their desktop for daily security updates [13:03] I sure hate as hell when I need to reboot my phone for updates :) [13:03] ogra_: building them alongside seems to be tricky given that e.g. the file layout will have to be different for clicks (different sysconfdir for example) [13:03] mvo, static builds ;) [13:04] ogra_: right, but even then /etc will be different, plugin loading location, support files (like images) [13:04] they will all be relative to the click root [13:04] ./configure with the right options can do that, but it would be a build for click and a build for deb [13:06] build a static deb ... dpkg -x into the click root ... roll a click package from that root ... done [13:06] indeed, its totally doable, but "build alongside debs" sounds nicer than "repackage them" :) [13:06] but why would we use deb in this case? we could as well skip this step [13:06] but we still want to provide the deb [13:07] and we build it already [13:07] the click creation should just happen alongside [13:07] tjaalton, mlankhorst: is the provisional mesa MRE still a requirement from somewhere? we reviewed it during the last TB meeting [13:07] ogra_: uhm, no, we're not going to start building everything statically [13:07] indeed it could just build it differently or in aa separate way [13:07] so debs would be for older distro releases? or why would we keep building it in this future? [13:08] mvo, for people wanting to use a traditional setup [13:08] tjaalton, mlankhorst: it provably caused at least one regression in a stable update (bug 1134974) which was never triaged, and in general it's by and large impossible to regression test that on all the world's hw [13:08] bug 1134974 in mesa (Ubuntu) "compiz and other display misbehavior on HD4000 after xatracker/mesa components upgraded to 9.0.2-0ubuntu0.1" [Undecided,New] https://launchpad.net/bugs/1134974 [13:09] mvo, the idea is: everyone gets a system-image install ... people that want to still use debs switch that over to an apt-get based design with a single command ... they cant go back but still use the system like debian or the former ubuntu ... so you please developers [13:09] pitti, aren't those updates required for new hardware enablement sometimes? [13:09] seb128: well, that's the tension -- breaking existing systems vs. enabling new ones [13:09] so I'm asking where that request comes from (OEM team, etc.?) [13:09] and whether these could rather live in e. g. backports [13:10] we can't really have an LTS that doesn't run on newer hardware though [13:10] well, we can't really have an LTS which breaks underneath you either :) [13:10] indeed, a tricky balance as always ;-) [13:11] so we either a way to do proper regression testing (but I think that's practically impossible) or confine the impact to new installations [13:11] of course if we put it into the lts-backports pacakges or into the -backports pocket you'd have the same problem one update later [13:11] pitti: yes, they're needed [13:12] so if it's e. g. enough to backport them for the 14.04.X releases, the -lts-backport approach seems better [13:12] (even though it still doesn't guarantee regressions) [13:12] so the TB is leaning towards revoking the MRE until we get a better approach here [13:13] at least a plan how to catch regressions (and at the *very* least, triage incoming regression reports) [13:13] tjaalton: by whom? for what targets? [13:13] ah, meeting time, will catch up later [13:16] (just an invalid noise from hangouts, still listening) [13:28] I tried to find the regression bug from the mesa bug list but couldn't [13:28] and the noisy list is only 177 bugs long [13:31] tjaalton: do mesa updates get testing on our whole lab of certified hardware before they are pushed out to a release? (just curious) [13:32] ie: how are we making sure we don't regress certified hardware from oems? [13:32] we use piglit [13:33] and used wider testing only when doing a major release update post-ff [13:34] haven't received a bug via oem projects due to an update [13:35] hrm, I gather most oem projects use intel hardware [13:35] and blobs [13:35] on nvidia/amd [13:35] because they are sooo much easier [13:35] really :) [13:35] hehe [13:36] still, upstream point releases get tested by others too, we're not in a void [13:36] and it's mostly intel updates in them anyway [13:37] hmm I lied [13:41] so we've done 11 point-release sru's, and got one regression that fell through the cracks? [13:41] which probably got fixed by the next point-release [13:45] pitti: by me for instance who's waiting for some of the broadwell diff in 10.1..10.2 to get in the next point-release [13:45] tjaalton: well, the regression that fell through the cracks was worrisome === mjohnson151 is now known as mjohnson15 [13:45] which bug was it? [13:45] couldn't find it [13:46] bug 1134974 (see above) [13:46] bug 1134974 in mesa (Ubuntu) "compiz and other display misbehavior on HD4000 after xatracker/mesa components upgraded to 9.0.2-0ubuntu0.1" [Undecided,New] https://launchpad.net/bugs/1134974 [13:46] (thanks to whoever got the bot back here) [13:47] ah so it didn't even have severity set [13:47] yeah, as I said it wasn't triaged at all [13:47] well it's a hybrid so dunno [13:51] still, 12.04.2 got the same update and most machines that used it were IVB based, so I'd say this was something that went wrong on his install only [13:51] and I've been using IVB on my desktop for two years now, since it was beta [13:52] we'd know if it was something widespread [14:02] Hey guys, its really simple: https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1253620/comments/77 so who of you will SRU all dependencies of glib on precise for once? [14:02] Ubuntu bug 1253620 in LibreOffice Productivity Suite "Can't open a LibreOffice native file via CIFS share" [Critical,Confirmed] [14:02] seb128: ^^ volunteering? [14:24] seb128: *sigh* the joke just got serious. [14:24] stop making jokes ! [14:25] Sweet5hark, sorry, was in an hangout, reading backlog [14:25] urg [14:25]