2014/05/14/#ubuntu-desktop.txt

[05:23] <pitti> Good morning
[05:23] <TheMuso> pitti: Good morning.
[05:23] <pitti> hey TheMuso, long time no see! how are you?
[05:23] <TheMuso> pitti: Very well thanks, yourself
[05:23] <TheMuso> ?
[05:24] <pitti> TheMuso: quite fine, thanks; shepherding distro tests these days, and refining the tools for that :)
[05:27] <TheMuso> pitti: Will you be in Malta?
[05:27] <pitti> TheMuso: yes I will, in the second week
[05:28] <TheMuso> Thats the last week of May right? I'll be there in the last week of May.
[05:28] <pitti> right, me too; 23rd to 30st
[05:29] <TheMuso> Yeah same as me, I'll probably see you around at some point.
[07:46] <darkxst> pitti, are you familiar with the sbuild launchpad stuff?
[07:46] <darkxst> is it possible to add aliases to that will pull in PPA's?
[07:46] <pitti> darkxst: I'm fairly familiar with sbuild, and I know that LP uses it; what do you mean?
[07:46] <pitti> oh
[07:46] <pitti> I saw a G+ post not too long ago about someone who wrote clever hooks
[07:49] <pitti> darkxst: https://plus.google.com/u/0/+ChristopherHalseRogers/posts/24T9xSgTMEF
[07:49] <pitti> darkxst: https://plus.google.com/u/0/+St%C3%A9phaneGraber/posts/KPTfpLKKF8J
[07:49] <pitti> those might be useful
[07:51] <darkxst> pitti, yeh I am using the sbuild-launchpad-chroot
[07:52] <pitti> but then again, sbuild hooks are rather simple to write
[07:54] <darkxst> pitti, right, the alias' are defined in a horrible shell script
[07:54] <darkxst> I somehow assumed they would be pulled from lp
[07:59] <pitti> darkxst: yes, that's supposed to download LP's schroot tarballs
[08:06] <seb128> good morning desktopers!
[08:07] <didrocks> hey seb128!
[08:07] <seb128> didrocks, lut ;-)
[08:07] <Laney> hallo
[08:08] <seb128> hey Laney, how are you?
[08:08] <darkxst> pitti, yes it downloads a tarball, but seems the actual apt sources are just generated by lots of awk-ish
[08:14] <mvo> 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] <didrocks> hey Laney!
[08:15] <Laney> hey seb128 & didrocks
[08:15] <mvo> 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] <Laney> I'm alright thanks, you?
[08:15] <pitti> hey seb128
[08:15]  * pitti waves to didrocks and Laney too
[08:15] <didrocks> I'm good, thanks! still a little bit cold here, but at least sunny
[08:15] <didrocks> guten morgen pitti :)
[08:15] <Laney> howdy pitti
[08:16] <Laney> we're supposed to have 17° and sun today
[08:16] <Laney> luxury
[08:17] <mvo> hey Laney and didrocks
[08:17] <didrocks> hey mvo!
[08:17] <didrocks> Laney: waow ;)
[08:17] <seb128> mvo, hey, no, try asking mardy?
[08:17] <seb128> hey pitti
[08:18] <seb128> Laney, what happened to the rain?!
[08:18] <Laney> I told its mum that it was being mean to me and it got sent home
[08:18] <seb128> hehe
[08:19] <seb128> I hope it doesn't leave around here
[08:19] <seb128> so far we have sun here as well at least ;-)
[10:28] <mlankhorst> rickspencer3: your boot failure, does it fail to boot or does it not boot into xorg?
[10:29] <rickspencer3> mlankhorst, I think it fails to boot into xorg
[10:30] <mlankhorst> can you try utopic? it has an updated mesa
[10:32] <mlankhorst> from what i can tell your pc should just work, does ssh work?
[10:35] <mlankhorst> if so, try apport-collect -p xorg 1319149
[10:42] <mlankhorst> rickspencer3: for the touchscreen you need qtbase-opensource-src from the sru queue :-)
[10:42] <rickspencer3> mlankhorst, I could not try utopic, as I had a lot of trouble getting the installer o run
[10:43] <rickspencer3> mlankhorst, unfortunately, I won't be able to debug much today, I have calls all day
[10:43] <mlankhorst> ok you can work on utopic, that's fine
[10:44] <rickspencer3> mlankhorst, you are certain it will work utopic?
[10:44] <rickspencer3> I'm not so sure
[10:44] <mlankhorst> neither
[10:44] <mlankhorst> what exactly is the issue?
[10:44] <rickspencer3> mlankhorst, first, uniquity does not load 9 times out of 10
[10:45] <rickspencer3> it stops at the screen with Ubuntu and the dots
[10:45] <rickspencer3> then, when I did get Trusty to install ...
[10:45] <rickspencer3> it booted once or twice, but also several times booted to a black screen
[10:45] <rickspencer3> though I could boot into recovery mode
[10:45] <mlankhorst> ok
=== MacSlow is now known as MacSlow|lunch
[10:55] <rickspencer3> good morning desktoppers!
[10:55] <rickspencer3> seb128, didrocks, and others in Europe who are up o/
[10:55] <Sweet5hark> moin!
[10:55] <rickspencer3> Sweet5hark,
[10:55] <rickspencer3> !
[10:55] <rickspencer3> moin you say?
[10:55] <rickspencer3> ===coffee activity===
[10:56] <rickspencer3>  * Action: rickspencer3 to sip coffee
[10:56] <Sweet5hark> rickspencer3: Its a universal greeting in northern germany. ;)
[10:57] <rickspencer3> Sweet5hark, nice to know!
[10:57]  * rickspencer3 presumes Sweet5hark refers to "moin" and not * to sip coffee
[10:57] <Sweet5hark> indeed.
[10:58] <didrocks> hey rickspencer3!
[10:58]  * Sweet5hark sips his sugar-and-caffeine based coffee surrogate ...
[10:59] <rickspencer3> bonjour didrocks
[11:01]  * rickspencer3 wonders why seb128 is avoiding him
[11:02] <rickspencer3> 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] <Laney> must self drive my car
[11:04] <Laney> actually we have some prelimenary images there
[11:04] <Laney> 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] <Laney> then see about getting it built in the normal way with dailies
[11:11] <seb128> rickspencer3, salut
[11:11] <rickspencer3> moin, seb128
[11:11] <rickspencer3> ;)
[11:12] <seb128> rickspencer3, sorry but there was food before you in the line ;-)
[11:12] <rickspencer3> 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] <seb128> lol
[11:13] <seb128> rickspencer3, I'm still amazed you manage to get up that early btw!
[11:15] <rickspencer3> seb128, it's 7:15am
[11:15] <rickspencer3> I live on teh East Coast now :)
[11:15] <seb128> rickspencer3, right, but I saw joining like an hour ago
[11:16] <rickspencer3> hehe
[11:16] <seb128> saw you*
[11:16] <rickspencer3> right
[11:16]  * didrocks notes that seb128 put food before manager. Bad for you review dude!
[11:16] <seb128> didrocks, next reviews are in a year right?
[11:16] <rickspencer3> when I get up I check my email for irc for fires first thing ;)
[11:16] <seb128> but then we are going to have a new manager anyway :p
[11:16] <rickspencer3> lol
[11:16] <didrocks> yeah, you can "too late!" :)
[11:16] <rickspencer3> seb128, knows he can do no wrong in my eyes
[11:17]  * seb128 hugs rickspencer3
[11:17]  * rickspencer3 hugs seb128
[11:18] <seb128> Laney, testing the iso on real hardware and tweaking the seed is still on my todolist
[11:18] <seb128> I didn't manage to get to it after the meeting yesterday, I had accumulated some pings and backlog
[11:18] <seb128> I hope to do that this afternoon ;-)
[11:18] <rickspencer3> seb128, when you do you think we can install that?
[11:18] <rickspencer3> couplle of weeks?
[11:18] <seb128> rickspencer3, boot and test, by Malta
[11:18] <rickspencer3> seb128, sweet
[11:19] <rickspencer3> I'll have my new 11.6 inch dell with me
[11:19] <rickspencer3> we can convert it
[11:19] <seb128> install ... let's see, with some luck that's working without needing to tweak too much
[11:19] <seb128> good
[11:19] <Laney> install should work if live does ...
[11:19] <seb128> which one is that? inspiron 11?
[11:19]  * Laney makes a bold statement
[11:19]  * Laney regrets it
[11:19] <seb128> lol
[11:20] <seb128> Laney, I was thinking the same but I went for the "with some luck" statement ;-)
[11:25] <Sweet5hark> Laney: to boldly go where no unittest has gone before!
[11:25] <rickspencer3> seb128, yeah, I got the inspiron 11
[11:25] <rickspencer3> I was told that it "just works"
[11:25] <seb128> rickspencer3, it does indeed ;-)
[11:26] <rickspencer3> my plan is to be all Unity 8/Mir asap
[11:26] <seb128> rickspencer3, that's the machine I'm testing unity8/Mir on btw
[11:26] <rickspencer3> :0
[11:26] <rickspencer3> so, I will need chromium
[11:26] <rickspencer3> which supposedly already has a Mir backend
[11:26] <seb128> qengho probably knows about that
[11:26] <rickspencer3> and I will need Qt, which supposedly already works :)
[11:26] <seb128> right
[11:26] <rickspencer3> seb128, I'm setting up a call so that a few of us can discuss Gtk
[11:27] <seb128> ok
[11:27] <rickspencer3> I'm getting slightly different stories from different people, so I thought we could just work it out on a call quickly
[11:27] <seb128> would it make sense to wait for Malta to discuss it more?
[11:27] <seb128> k
[11:27] <rickspencer3> seb128, no, I would like to have at least a rough plan by Malta
[11:27] <rickspencer3> so we can work the plan at the sprint :)
[11:27] <seb128> k
[11:30]  * darkxst is also wondering what is happening with gtk+
[11:31] <seb128> you mean?
[11:31] <seb128> upstream?
[11:32] <darkxst> seb128, no, I know what is happening upstream! in utopic!
[11:35] <seb128> what do you want to know about GTK in utopic?
[11:40] <darkxst> seb128, what is blocking 3.12 would be a good start!
[11:43] <seb128> it's being tested, we discussed it at the weekly meeting here (logs are available if you are interested)
[11:43] <seb128> the main problem at the moment in GtkHeaderBar use in dialogs
[11:43] <seb128> larsu is working on that to see what we can do to make them work on e.g Unity
[11:44] <seb128> one of the changes required is to update our themes
[11:44] <seb128> Lars said he's going to work on that
[11:44] <seb128> we might need some Unity work as well
[11:44] <seb128> then we need to decide if that's going enough or we need to patch GTK as well
[11:45] <larsu> we definitely need support from unity for border-less windows
[11:45] <larsu> it can't resize them right now
[11:45] <larsu> and we'll need a patch if we want traditional title bars on dialogs
[11:46] <darkxst> seb128, right, I see, can't really help with unity/theming issues
[11:46] <darkxst> but gtk patches are more probable
[11:47] <darkxst> I would actually attend your meetings... if they weren't at a god awful time for me!
[11:47] <larsu> darkxst: what do you mean by "more probable"
[11:48] <darkxst> larsu, that I could probably help there
[11:48] <larsu> ah, right
[11:48] <darkxst> (as long as it doesn't involve theming!)
[11:48] <larsu> you're really afraid of theming ;)
[11:49] <larsu> tbh, most of this work involves theming at least a bit
[11:50] <darkxst> larsu, nope, just find C much more readable than CSS!
[11:51] <larsu> ha, it's almost the same syntax man ;)
[11:54] <darkxst> larsu, syntax is not the problem, it just feels like I end up wasting hours anytime I start messing CSS
[11:54] <darkxst> (same for the couple of websites I have written)
[11:55] <larsu> right, I know the feeling
[11:55] <larsu> it's possible to write good css, but almost noone does it
[11:56] <darkxst> larsu, its possible to write good code in any language, but there are a bunch where no-one does!
[11:57] <larsu> C isn't necessarily a good example of that, though
[11:59] <ogra_> bah
[11:59] <darkxst> I think GNOME C is pretty good in general
[11:59] <ogra_> so i just got a firefox update that told me to restart the browser with a popup ...
[11:59] <ogra_> when i clicked it didnt restart
[12:00] <ogra_> and now trying to start FF i only get a message about FF already running
=== MacSlow|lunch is now known as MacSlow
[12:00] <darkxst> 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] <larsu> ogra_: firefox should keep those even if it crashes (it shows you a "restore last session?" tab)
[12:00] <larsu> darkxst: fair enoguh
[12:01] <larsu> *enough
[12:01] <ogra_> larsu, well, it shows me "another instance is already running" ...
[12:01] <larsu> ogra_: ya, kill -9 it first
[12:01] <ogra_> i'm looking for a safe way out of that :)
[12:01] <ogra_> ok
[12:01] <larsu> wait.
[12:01] <larsu> don't make me responsible
[12:01] <larsu> if you loose all those tabs
[12:01] <ogra_> hah
[12:02] <ogra_> to late ... now i will blame you :)
[12:02] <darkxst> ogra_, why do you need 300 tabs?
[12:02] <ogra_> oh, wow
[12:02] <ogra_>  3188 ?        Sl   1799:52 /usr/lib/firefox/firefox
[12:02] <larsu> having 300 tabs open is some kind of insanity anyway
[12:02] <ogra_> that did run since quite some time
[12:02] <ogra_> well, i do actual work on my machine :P
[12:03] <larsu> 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] <larsu> seb128: ?
[12:03] <ogra_> phew
[12:03] <ogra_> all fine ...
[12:03] <larsu> ogra_: awesome :)
[12:03] <ogra_> it always restores all windows on one workspace though ... but thats something to blame unity for i guess ... not FF
[12:04] <seb128> larsu, the less reverts the better, maybe check with design/mpt, what they don't flag "no no no" can stay?
[12:05] <larsu> seb128: good idea. I'll note those down and schedule a call / session in Malta with mpt
[12:06] <seb128> larsu, looks like mpt should be in Malta the same week as we are, good ;-)
[12:06] <larsu> yep
[12:06] <larsu> \o/
[12:14]  * Laney tries g-t with the transparency patch
[12:15] <larsu> wheeeee
[12:15] <Laney> it's a bit buggy
[12:15] <Laney> it doesn't live update
[12:16] <Laney> you have to toggle "use colours from system theme" to make changes take effect
[12:16] <larsu> ugh
[12:16] <larsu> that's annoying
[12:16] <Laney> yeah probably just some missing signal bindings though
[12:18] <larsu> I can have a look at that later
[12:31] <Sweet5hark> seb128: would there be a chance in Malta for me to present a widely a crazy lunatic vision for Ubuntu?
[12:32] <mlankhorst> libreoffice replacing unity as shell?
[12:34] <seb128> Sweet5hark, to who? the team? Mark? ;-)
[12:35] <ogra_> will you show us the LibO click packages finally ?
[12:37] <Sweet5hark> seb128: the team, not (yet) Mark. It would be about packaging, so foundations would likely be the ones to consider it ultimately ...
[12:37] <seb128> Sweet5hark, yeah, I'm sure we can organize that
[12:38] <seb128> humpf
[12:38] <seb128> Laney, would you take your rain back, please?
[12:38] <ogra_> no ! ... keep it in france ... we dont want it up here north
[12:39] <Laney> 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] <mlankhorst> more red tape!
[12:40] <seb128> :-(
[12:41] <Sweet5hark> 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] <ogra_> Sweet5hark, well, the phone image will eventually take over the desktop
[12:42] <ogra_> i woudl actually expect us to move away from debs over the next few years
[12:42] <ogra_> (for endusers that is)
[12:43] <ogra_> using image based updates on the desktop (which means readonly rootfs) etc etc
[12:43] <Sweet5hark> 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] <seb128> ogra_, that's going to be challenging
[12:43] <seb128> ogra_, did you read https://lists.ubuntu.com/archives/technical-board/2014-March/001856.html ?
[12:44] <ogra_> seb128, why ? just drop debs
[12:44] <ogra_> yes
[12:44] <seb128> ogra_, if we "just drop debs" we don't have an usable desktop anymore
[12:44] <seb128> we need to find a way to convert those to clicks first
[12:44] <ogra_> 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] <seb128> ogra_, except if you don't care for libreoffice, firefox, chromium, gimp, inkscape, etc
[12:45] <ogra_> for the rest just make the desktop identical to what we have on the phone today
[12:45] <ogra_> seb128, right, then you switch your image to writable mode and use it like you did before
[12:46] <seb128> I though that was not supported?
[12:46] <ogra_> for the typical desktop user the image based approach works better ...
[12:46] <seb128> "typical"
[12:46] <ogra_> it is not supported in the specific image design we use on the phones
[12:46] <seb128> I guess you mean "if we include libreoffice, firefox, etc in the base image"?
[12:47] <ogra_> 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] <ogra_> seb128, no, these should be clicks
[12:47] <ogra_> built from plain upstream ... without distro hacks we have to care for etc
[12:47] <ogra_> no more maintainer scripts ...
[12:48] <ogra_> system upgrades done in minutes instead of hours with predictable installations etc
[12:48] <seb128> right
[12:48] <seb128> well, we "just" have to convert our archive to clicks
[12:49] <asac> 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] <ogra_> nah
[12:49] <Sweet5hark> seb128, orga_:  are you both in Malta? from what I see you are both wrong ;)
[12:49] <ogra_> we only have to convert the UI apps to clicks
[12:49] <seb128> Sweet5hark, lol
[12:49] <seb128> ogra_, it's still a large stack of apps
[12:49] <ogra_> seb128, sure
[12:50] <ogra_> thats why i said "within the next years" :)
[12:50] <seb128> right, fair enough
[12:50] <seb128> asac, and sure, engaging with upstream to ramp that up is going to be useful/needed
[12:50] <asac> 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] <ogra_> why would you use binaries ?
[12:51] <seb128> 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] <ogra_> just use the source ... we have it already ... you just need to make the archive build a click package alongside
[12:51] <asac> seb128: yes, just upload :)
[12:51] <Sweet5hark> asac: using upstream binaries is insane for anything nontrivial
[12:51] <asac> seb128: i know mozilla has been waiting for this ability for ages
[12:51] <ogra_> seb128, yeah, just push your locally built binary
[12:52] <asac> mozilla etc. have infrastructure to validate that their build works on ubuntu
[12:52] <asac> etc.
[12:52] <asac> they dont have many users, but if they would have thea bility to really distribute their stuff alongside our distro
[12:53] <asac> they surely would even invest more to make their experience shine by default on ubuntu
[12:53] <asac> 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] <ogra_> we dont :P
[12:55] <ogra_> we need to make sure that if you switch away from system image you can use debs
[12:55] <asac> ogra_: we dont?
[12:56] <didrocks> one source tree to rule them all :)
[12:56] <ogra_> imho we dont need any deb support in system-image setups
[12:56] <didrocks> but I have already made the point multiple times :p
[12:56] <seb128> didrocks, well, the source tree is orthogonal to the apps distribution problem
[12:56] <asac> ogra_: well, so guess we could say that and pair it with the ability to convert things into your own system image
[12:56] <asac> to re-lock down
[12:56] <asac> :)
[12:56] <asac> lie what was on canoni-tech
[12:57] <ogra_> fi you have system-image you go with clicks ... if you want debs you switch to readwrite ... and cant go back
[12:57] <seb128> didrocks, having one unified source tree doesn't give you libreoffice packages in the click store
[12:57] <ogra_> but then debs need to be properly supported indeed
[12:57] <asac> disable system image, use apt to fiddle, snapshot a system image, easy distribute :)
[12:57] <didrocks> seb128: you should the subdirectory built binary and project that to the installed system
[12:57] <asac> ogra_: the part of the story where you cant go bad just doesnt feel good enough for our standards :)
[12:57] <ogra_> asac, and how would that work with the server side generated deltas ?
[12:57] <asac> sounds like a technical detail
[12:57] <asac> :)
[12:57] <ogra_> lol
[12:58] <asac> i am not saying that story is fully developed
[12:58] <ogra_> well, the design wont allow that
[12:58] <asac> but its one vector to maybe look at this
[12:58] <ogra_> my vector is that 80% of the users should be fine with s-i
[12:58] <asac> yes, for thtat i agree :)
[12:58] <ogra_> and the other 20 will happily just go with apt-get dist-upgarde
[12:58] <asac> actually 95% at least
[12:58] <ogra_> and stay there
[12:59] <ogra_> we just need to make sure that both work fine and that the switch works properly
[12:59] <ogra_> i wouldnt provide a switch to go back ..
[12:59] <asac> maybe its the right way to look at it
[12:59] <asac> buit lets see; the more we think the better our solutions get
[13:00] <Laney> larsu: I uploaded it to ppa:laney/experimental if you want to take a look
[13:00]  * Laney biab
[13:00] <asac> and we have just started thinking :)
[13:00] <asac> hehe
[13:01] <seb128> ogra_, well, your 80% of users being happy assume that we have clicks for libreoffice, chromium, firefox, gimp, pidgin, etc?
[13:01] <asac> seb128: of course
[13:01] <asac> seb128: we want t store full of useful killer apps
[13:01] <asac> open source, proprietary, games etc.
[13:01] <ogra_> seb128, yes, which means our builders need to roll clicks alongside the debs
[13:02] <seb128> that seems like a pre-require
[13:02] <asac> how they get into the store and who is doing that is kind of secondary and technical detail i think.
[13:02] <seb128> before using system images for desktop
[13:03] <mdeslaur> I'm not sure anyone will be happy about rebooting their desktop for daily security updates
[13:03] <mdeslaur> I sure hate as hell when I need to reboot my phone for updates :)
[13:03] <mvo> 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] <ogra_> mvo, static builds ;)
[13:04] <mvo> ogra_: right, but even then /etc will be different, plugin loading location, support files (like images)
[13:04] <ogra_> they will all be relative to the click root
[13:04] <mvo>  ./configure with the right options can do that, but it would be a build for click and a build for deb
[13:06] <ogra_> build a static deb ... dpkg -x into the click root ... roll a click package from that root ... done
[13:06] <mvo> indeed, its totally doable, but "build alongside debs" sounds nicer than "repackage them"  :)
[13:06] <mvo> but why would we use deb in this case? we could as well skip this step
[13:06] <ogra_> but we still want to provide the deb
[13:07] <ogra_> and we build it already
[13:07] <ogra_> the click creation should just happen alongside
[13:07] <pitti> tjaalton, mlankhorst: is the provisional mesa MRE still a requirement from somewhere? we reviewed it during the last TB meeting
[13:07] <mdeslaur> ogra_: uhm, no, we're not going to start building everything statically
[13:07] <ogra_> indeed it could just build it differently or in aa separate way
[13:07] <mvo> so debs would be for older distro releases? or why would we keep building it in this future?
[13:08] <ogra_> mvo, for people wanting to use a traditional setup
[13:08] <pitti> 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] <ubot5> 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] <ogra_> 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] <seb128> pitti, aren't those updates required for new hardware enablement sometimes?
[13:09] <pitti> seb128: well, that's the tension -- breaking existing systems vs. enabling new ones
[13:09] <pitti> so I'm asking where that request comes from (OEM team, etc.?)
[13:09] <pitti> and whether these could rather live in e. g. backports
[13:10] <seb128> we can't really have an LTS that doesn't run on newer hardware though
[13:10] <pitti> well, we can't really have an LTS which breaks underneath you either :)
[13:10] <seb128> indeed, a tricky balance as always ;-)
[13:11] <pitti> 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] <pitti> 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] <tjaalton> pitti: yes, they're needed
[13:12] <pitti> so if it's e. g. enough to backport them for the 14.04.X releases, the -lts-backport approach seems better
[13:12] <pitti> (even though it still doesn't guarantee regressions)
[13:12] <pitti> so the TB is leaning towards revoking the MRE until we get a better approach here
[13:13] <pitti> at least a plan how to catch regressions (and at the *very* least, triage incoming regression reports)
[13:13] <pitti> tjaalton: by whom? for what targets?
[13:13] <pitti> ah, meeting time, will catch up later
[13:16] <pitti> (just an invalid noise from hangouts, still listening)
[13:28] <tjaalton> I tried to find the regression bug from the mesa bug list but couldn't
[13:28] <tjaalton> and the noisy list is only 177 bugs long
[13:31] <mdeslaur> 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] <mdeslaur> ie: how are we making sure we don't regress certified hardware from oems?
[13:32] <tjaalton> we use piglit
[13:33] <tjaalton> and used wider testing only when doing a major release update post-ff
[13:34] <tjaalton> haven't received a bug via oem projects due to an update
[13:35] <mdeslaur> hrm, I gather most oem projects use intel hardware
[13:35] <tjaalton> and blobs
[13:35] <tjaalton> on nvidia/amd
[13:35] <tjaalton> because they are sooo much easier
[13:35] <tjaalton> really :)
[13:35] <mdeslaur> hehe
[13:36] <tjaalton> still, upstream point releases get tested by others too, we're not in a void
[13:36] <tjaalton> and it's mostly intel updates in them anyway
[13:37] <tjaalton> hmm I lied
[13:41] <tjaalton> so we've done 11 point-release sru's, and got one regression that fell through the cracks?
[13:41] <tjaalton> which probably got fixed by the next point-release
[13:45] <tjaalton> 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] <mdeslaur> tjaalton: well, the regression that fell through the cracks was worrisome
=== mjohnson151 is now known as mjohnson15
[13:45] <tjaalton> which bug was it?
[13:45] <tjaalton> couldn't find it
[13:46] <pitti> bug 1134974 (see above)
[13:46] <ubot5> 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] <seb128> (thanks to whoever got the bot back here)
[13:47] <tjaalton> ah so it didn't even have severity set
[13:47] <pitti> yeah, as I said it wasn't triaged at all
[13:47] <tjaalton> well it's a hybrid so dunno
[13:51] <tjaalton> 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] <tjaalton> and I've been using IVB on my desktop for two years now, since it was beta
[13:52] <tjaalton> we'd know if it was something widespread
[14:02] <Sweet5hark> 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] <ubot5> Ubuntu bug 1253620 in LibreOffice Productivity Suite "Can't open a LibreOffice native file via CIFS share" [Critical,Confirmed]
[14:02] <Sweet5hark> seb128: ^^ volunteering?
[14:24] <Sweet5hark> seb128: *sigh* the joke just got serious.
[14:24] <ogra_> stop making jokes !
[14:25] <seb128> Sweet5hark, sorry, was in an hangout, reading backlog
[14:25] <seb128> urg
[14:25] <Sweet5hark> seb128: you have mail ;)
[14:25] <seb128> saw that!
[14:26] <Sweet5hark> seb128: can we agree on ogra_ SRUing all of gnome?
[14:26] <seb128> wfm
[14:27] <ogra_> thats scriptable ...
[14:28] <Sweet5hark> ogra_: the SRU team will likely script back though ...
[14:28] <ogra_> yeah, thats what i fear
[14:28] <Sweet5hark> ogra_: ... a drone that finds and kills you with a spoon.
[14:28] <ogra_> gnome is dead anyway :P
[14:28] <ogra_> (until the QML re-write)
[14:29] <seb128> ogra_, it's not friday yet!
[14:29] <Laney> I wish boring trolling was dead
[14:29] <ogra_> seb128, ah, damned ... just noticed
[14:29] <Sweet5hark> ogra_: well, "gnome" was a simplifcation. actually its all that depends on glib-2.0
[14:29] <Laney> all that uses those macros, at worst
[14:30] <seb128> I don't understand
[14:30] <seb128> http://ubuntu-codesearch.surgut.co.uk/search?q=GUINT32_SWAP_LE_BE
[14:30] <seb128> returns neither gvfs nor libreoffice
[14:31] <Sweet5hark> seb128: is that searching on precise?
[14:31] <seb128> it doesn't return glib either
[14:31] <seb128> Laney, codesearch is busted!
[14:32] <seb128> Sweet5hark, http://codesearch.debian.net/search?q=GUINT32_SWAP_LE_BE seems better
[14:32] <Laney> it's re-running on utopic now
[14:32] <seb128> "re-running"?
[14:32] <Laney> yes
[14:32] <seb128> means rebuilding an index/currently incomplete ?
[14:33] <Laney> I guess it could explain it, could also be buggy though :(
[14:33] <Sweet5hark> seb128: ... which shows on the first page that this macro is used in macros, so a simple search wont do ...
[14:34] <seb128> Sweet5hark, one solution might be to put a "fixed define" in gvfs to fix that specific issue
[14:34] <seb128> though I'm unsure how likely that's going to create interaction issues with other component using the wrong macro
[14:36] <Sweet5hark> seb128: yes, but that might break as client code of gvfs migh do the swap too: swap twice in a broken way => no error, swap correctly once (in fixed gvfs), incorrectly again (in lib linking against gvfs) => busted
[14:36] <seb128> right
[14:41] <seb128> Sweet5hark, https://lists.fedoraproject.org/pipermail/devel/2013-March/180325.html doesn't bring confidence
[14:43] <Sweet5hark> yep. When I read that back in the days, I felt for Stephan. Imagine debugging a networking issue vs. windows shares from LibreOffice down to glib ...
[14:43] <seb128> yeah
[14:46] <Sweet5hark> seb128: what we maybe could do, is putting these in officialish PPA and see if there are any issues for 6 months or so ... but even then. The audience (and thus test coverage) for that PPA would be very small (only people hitting this problem on a LTS, while there is a newer LTS already) ...
=== Zdra is now known as xclaesse
[14:47] <seb128> Sweet5hark, that might be a good enough solution for those users though
[14:47] <seb128> which would be the best way to make everyone happy
[14:52] <jagannath> hi can any body help me
[14:52] <jagannath> ??
[14:55] <ogra_> was that a trick question ?
=== rickspencer3_ is now known as rickspencer3
[15:03] <kenvandine> mpt, where do you stand on adding custom ringtones and notification sounds in system-settings?  Could we get something on the wiki about that?
[15:04] <kenvandine> mpt, like using the content picker to select sounds from your music library, or handle downloaded ringtones
=== JanC is now known as Guest12750
=== JanC_ is now known as JanC
=== gatox is now known as gatox_lunch
=== m_conley_away is now known as m_conley
=== ted is now known as tedg
[15:45] <tjaalton> I have two firefox profiles running, and when trying to open a link from thunderbird I get a prompt asking which profile to start, it doesn't let me open the link on an open profile
=== gatox_lunch is now known as gatox
[16:26] <didrocks> qengho: hey, in case you don't really remember our discussion, but now, I got multiple confirmation that the "lost tab" bugs is fixed in 36
[16:26] <didrocks> qengho: I hope it's even fixed in 35 so that your work for getting the fix in chromium will be easier…
[16:27] <qengho> didrocks: Yes, thanks.
[16:28] <didrocks> yw
=== beidl_ is now known as beidl
[16:31] <seb128> qengho, do you know when 35 is becoming stable?
[16:35] <qengho> seb128: Probably end of this month.
[16:36] <tedg> Laney, How about "ubuntu-unicorn"! :-)
[16:36] <ogra_> ubuntu-desktop-ng
[16:36] <ogra_> :)
[16:37] <tedg> omg-ubuntu
[16:37] <ogra_> ubuntu-da-fjutscha
[16:39] <Laney> ubuntu-me-harder
[16:46] <kenvandine> ubuntu-desktop-omg
[16:47] <seb128> ubuntu-finally-no-compiz
[16:47] <seb128> ;-)
[16:47] <ogra_> ++
[16:47] <seb128> no wm either though atm :/
[16:48] <seb128> unity8 on a desktop feels like a tablet UI
[16:50] <Laney> very early days
[16:50] <ogra_> just bribe the filemanager app guys into adding a desktop :)
[16:57] <stgraber> we could also call it ubutnu, that'd be in line with what we use for our pre-release services :)
[16:57] <stgraber> (just to confuse the hell out of everyone)
[17:03] <Laney> :D
[17:05] <seb128> Laney, thanks for the work on bootstraping and sending that email btw ;-)
[17:37] <seb128> Sweet5hark, what happened to your normal you?
[17:37] <Sweet5hark> seb128: normal me being "Sweetshark"?
[17:37] <seb128> Sweet5hark, yeah
[17:38] <seb128> Sweet5hark, you are also not on the Canonical IRC it seems
[17:38] <Sweet5hark> seb128: did a "do-release-update" to trusty on the pandaboard where that was running, and the machine didnt come up with a sshd :/
[17:39] <seb128> :-(
[17:39] <seb128> Sweet5hark, anyway, real question for you ... ;-)
[17:39] <Sweet5hark> seb128: yes?
[17:39] <seb128> Sweet5hark, what do you see as your focus on lo for u-cycle?
[17:40] <seb128> the packaging changes you mentioned, keeping up with upstream, bugfix I guess
[17:40] <seb128> GTK3/Mir work maybe?
[17:40] <seb128> Sweet5hark, I'm just trying to wrap my head around what the team is doing this cycle ;-)
[17:40] <Sweet5hark> seb128: yes, thats roughly the agenda. plus (but thats more upstream) getting more unittesting and CI in there ..
[17:41] <seb128> ok, makes sense
[17:41] <seb128> Sweet5hark, thanks ;-)
=== mjohnson151 is now known as mjohnson15
=== m_conley is now known as m_conley_away