[05:02] Good morning [05:56] good morning [07:26] happyaron, weekly report please [07:26] also [07:27] good morning! [07:31] hey willcooke [07:31] morning willcooke [07:31] morning! [07:34] ah, life! good morning everyone! [07:34] hey pitti [07:34] pitti, wie gehts? [07:36] cute http://www.omgubuntu.co.uk/wp-content/uploads/2015/05/release-schedule.jpg [07:36] morning pitti :) [07:37] seb128: gut, danke! ein bisschen muede, vom Sport [07:37] larsu: oooh! [07:37] pitti, same here, I played tennis for over 2 hours yesterday evening ;-) [07:37] seb128: wow, great! so it stopped raining for you too? [07:38] yes, back to blue sky & sun since yesterday afternoon [07:43] seb128: still gray here, but at least dry again [07:43] * larsu wonders what this "rain" is that people keep talking about [07:43] larsu, did you move in south of France with didrocks?! [07:44] haha no they speak a weird dialect there [07:45] Hey willcooke, hey folks. [07:45] hey TheMuso, all [07:46] hey TheMuso [07:47] pitti, still looking at this locale filtering issue ... is there a way to say that the "sr" locales dir is for the "sr_RS" locale? [07:47] pitti, out of reading /usr/share/language-tools/main-countries [07:48] seb128: ah, interesting file! that's from accountsservice [07:48] probably something that GunnarHj wrote? [07:48] pitti, yeah, I think one from Gunnar [07:48] likely yes [07:48] there is too much hackery in those scripts to my taste [07:48] seb128: that's not built into locales by themselves, as the concept of a local doesn't have a "main" country (it's a language and region where you are) [07:49] but at the same time it works and nobody else looked at those issues [07:50] pitti, so my new heuristic is to get the list is to keep the locales that are in that main-countries list or have a matching dir in /usr/share/locale-langpack ... how does it sound to you? [07:51] pitti, so sr_RS would stay because it's a "main" language, en_GB because it's a disk dir, en_NZ would match neither and be filtered out [07:52] same de_DE is a main language so it would stay [07:52] seb128: what's wrong with showing all locales that we have translations for? [07:52] pitti, define "have translation for" [07:52] or how do we check that? [07:53] e. g. for "de", I thought launchpad's policy was to only ship "de/" translations, not "de_*" [07:53] (that should be true for the majority of languages, except zh, en, es, and pt) [07:53] seb128: has a dir in /usr/share/locale-langpack/ ? [07:53] pitti, well, we have a "de_DE" locale, we can't check for /usr/share/locale-langpack/de_DE [07:53] since that doesn't exist [07:53] we do have en_NZ translations, so why would we hide them? [07:54] thx happyaron [07:54] or rather, if we don't want them, we should already disable them in launchpad and the langpacks [07:54] pitti, we wouldn't, my heuristic is keeping those which have a dir on disk [07:54] no point in offering them for translations? [07:54] pitti, "keep the locales that are in that main-countries list or have a matching dir in /usr/share/locale-langpack ." [07:54] seb128: why do you need de_DE? [07:55] because we want to propose "German" in the languages? [07:55] seb128: is whatever you are doing purely for translations, or does it have a regional aspect too (time, currency, paper format, etc.)? [07:55] just display language translations [07:55] willcooke: :) [07:55] seb128: right, but why would that not be de, but de_DE? [07:55] pitti, we use the output from "locale -a" to build the list [07:56] maybe that bit is wrong and to revisit [07:56] that's a good thing to start from [07:56] but that includes "de_DE" not "de" [07:56] but might be easier to start from /usr/share/locale-langpack/* [07:57] and then substract the entries from that which we dont' have locales for [07:57] e. g. de, en, en_AU would all stay [07:57] but en@quot would drop [07:57] how do you determine what's the locale for e.g "de" [07:57] or for "sr" [07:58] just prefix matching [07:58] we need the locale [07:58] to have a nice display name [07:58] hey ho [07:58] this topic again! [07:58] seb128: hm, you just told me you don't need a region? [07:58] hey Laney! [07:59] pitti, well, we .displayName() in the UI to show nice names in the list [07:59] /usr/share/language-tools/language2locale de [07:59] de_DE.UTF-8 [07:59] hm, this perhaps? [08:00] hey Laney [08:00] language-selector is able to figure this all out, what does that use? [08:00] right [08:00] that uses /usr/share/language-tools/main-countries [08:00] which is basically my current heuristic [08:00] look to that list and to existing langpack dirs [08:02] Laney, do you like that patch better? http://paste.ubuntu.com/11259919/ [08:02] Laney, ignore the code comment I didn't rewrite it yet [08:02] haha [08:03] I was confused [08:03] can you explain the indexOf in english? [08:03] Laney, basically I keep locales that are "main language ones" (= likely) or that have a matching dir [08:03] it's hard to figure it out from the code [08:03] ok [08:03] I think that's what I was asking for, so yes [08:03] "is the locale in the /usr/share/locale-langpack/*" [08:03] the table is the ls from that dir [08:04] and the indexOf is "is that value in the list" [08:04] good :-) [08:04] I might still look at simplifying that code/reworking it later [08:04] but that diff works and should do the job with minimal changes [08:04] Laney, pitti, thanks for the input! [08:05] an improvement that I suggested last night is to have that script take multiple inputs and process them all [08:05] it is a bit grim to call it once for each locale [08:12] Laney, yeah, going to look at that next (or at replacing the script, I like better to build a table from the text file and dealing with that than having to call some script calling sed & co) [08:12] Laney, http://paste.ubuntu.com/11260039/ with updated comment, if that looks ok to you I'm going to put it up for review [08:14] any reason you use indexOf instead of contains? [08:18] Laney, I think contains() does substring matching [08:18] ["de_DE", "en_GB"] contains "en" but wouldn't have an indexOf "en" I think [08:19] that would surprise me [08:19] no, contains is == [08:19] weird [08:19] http://doc.qt.io/qt-4.8/qstringlist.html#contains [08:19] is this in a loop? [08:20] oh, I misread the example [08:20] nevermind ;-) [08:20] I think you should list the directory before it [08:20] the api doc was not responding yesterday and I was unsure [08:20] I guess you could do .filter("de")[0] [08:21] but you'd need to check that [0] exists [08:23] * Laney filters larsu [08:24] * larsu disappears [08:25] ah wait, I guess I should be the only one left if you filter on "larsu" [08:25] hello? [08:25] k [08:25] installedTranslations.contains(languageLocale.localeName.split(".utf8")[0])) [08:25] that's probably easier to read [08:25] all this talk of lists makes me want to write Haskell again [08:25] larsu, can't see you! [08:27] seb128: did you see my comment about the dir list? [08:28] Laney, that's a very good point, just did, thanks ;-) [08:28] ok, now SHIP IT [08:36] Laney, https://code.launchpad.net/~seb128/ubuntu-system-settings/language-filter-translations/+merge/259730 [08:38] wait, what's that script that GunnarHj talks about? [08:39] dunno, something that doesn't do what we need by default :p [08:39] it lists e.g en_NZ or de_CH for me [08:40] that's because it uses /usr/share/locale as well [08:41] ok [08:41] I feel like this isn't the end of the story on listing languages [08:42] no, but it seems that script is supposed to do what we want [08:42] right [08:42] what does it do on your phone? [08:43] Laney, http://paste.ubuntu.com/11260387/ [08:44] too many? [08:44] unsure [08:44] too many on my desktop for sure [08:45] ok let's think about this later [08:45] I think we probably will end up wanting a mechanism for adding any language [08:46] which will somehow install the langpack and then it will be listed anyway [08:47] I need to go back to mpt [08:47] the issue there/the flackiness is because e.g [08:47] $ ls /usr/share/locale/de_CH/LC_MESSAGES/ [08:47] encfs.mo virt-viewer.mo [08:47] [08:47] so you get 1 thing on the system including a translation for a weird locale and it makes it be listed [08:48] unsure if we maybe we would like to list only languages with a >%system_strings translated [08:48] like only if it's > 50% translated [08:48] you'd probably exclude all variants then [08:49] I bet en_GB doesn't have many differences from en_US [08:49] hum [08:50] so apply that filter only for non variants ;-) [08:50] that topic is more complicated that I thought it was :-/ [08:50] maybe, but then they are presumably only installed when a user has done it on purpose [08:50] we would only ship major languages by default [08:51] ahhhhhh [08:51] oh, hello mpt [08:51] hey mpt [08:51] seb128, I am so sorry, I thought I had commented again about the locale stuff [08:52] oh god, stop reading now, he's going to make you change it [08:52] mpt, trying to get that list of language simplified, we have a solution which is not perfect but should be a good first step [08:52] lol [08:52] seb128, I was all geared up to separate out locale stuff into its own screen, but Andreea (reasonably) says that so much is changing in the first-run setup, that we should leave that for later [08:53] mpt, yeah, well our first step solution should be good enough for now [08:53] it basically keep languages which have langpack translations installed [08:54] so no more e.g de_CH [08:54] or en_NZ (sorry ;-) [08:55] should we remove the (Country) for the likely locale? [08:56] to not make New Zealanders have to click on English (United States) ... [08:58] The currency setting in locales is pretty much useless, right? I.e. I won’t have £ appearing anywhere merely because in the absence of en_NZ I choose en_GB [08:59] good point, dunno where that's used [08:59] or rather I don't know of any place where it's used [08:59] but it doesn't mean it's not used [09:00] anything using strfmon() [09:01] also, start of weekday, summer time, paper formats etc. are region specific [09:01] pitti, yeah, I just don't know if we have lot of code doing that [09:01] but why are we talking monetary/time/ etc. now? [09:01] I thought we are *just* talking about a language? [09:02] if this is for picking a *locale*, don't filter out anything [09:02] pitti, because the current list shows language+locale [09:02] list for what? [09:02] seb128’s change drops the locales that don’t have a corresponding language [09:02] selecting a desktop translation doesn't need that; selecting a system locale very much does [09:02] System Settings > Language & Text > Display language [09:02] ah; that'll just set LC_MESSAGES and LANGUAGE, right? [09:02] This is because I was in a mad rush in 2013 and thought, “eh, let’s think about locale later” [09:03] And now is later [09:03] and we presumably have another setting for the region then [09:03] It currently sets the language and locale [09:03] We don’t [09:04] oh! [09:04] seb128: so previously you said that this is just about a language [09:04] And when we do, it should we changed in both System Settings and the first-run setup, otherwise we’re inconsistent [09:04] what is true now? [09:04] And Andreea doesn’t want to change it in the first-run setup for now. So. [09:05] well, first-time setup does include the countries [09:05] Only because System Settings does too [09:05] (at least for German, i haven't checked the others) [09:05] It is true, the change will mean you can only have a locale if there are translations [09:05] And the lack of locales will start mattering when bug 1447504 is fixed [09:05] bug 1447504 in Ubuntu UX "[system settings] Dialer-app: Wrong number is shown in call history" [Medium,Triaged] https://launchpad.net/bugs/1447504 [09:05] ideally these become two separate settings, like in ubiquity and language selector [09:05] yes [09:05] I can be in the US and still speak Spanish [09:06] Should we wait until there is a design for locale then? [09:06] I think so, sorry seb128 [09:06] and one language and a completely independent country/region setting are both smaller individually and also more logical [09:06] I thought I had commented to that effect, but I guess I did not [09:06] * pitti knows some fr_NL dude here :-) [09:07] * pitti hugs seb128 ^ [09:07] The code is still good, it just needs to wait for a few weeks-ish [09:08] shrug [09:08] it feels like I just wasted a day :-/ [09:08] this discussion wouldn't have happeed otherwise [09:09] seb128: well, as mpt said I think the filtering *is* good for a language list [09:09] * seb128 hugs pitti back [09:09] seb128: the thing that's missing is a separate region list [09:09] pitti, well, it's just that mpt made it sounds like that having locales was an important usability problem for that list [09:09] "what do I speak" and "where am I" are two independent things [09:09] which motivated to work on it [09:10] right [09:10] but that was meant to be a languages list [09:10] the whole thing is just frustrating at this point [09:10] we probably get to discuss these issues again in a few weeks [09:10] right, the current first-time setup is bad, it mangles both, has a long list, and doesn't have most combinations [09:10] yay! [09:10] "oh, we really need to fix that now" [09:10] "there is a fix" [09:10] seb128, to be fair, this bug is #107 on the list of ubuntu-system-settings bugs and wasn’t Triaged or linked to a spec yet, I didn’t know you were going to work on it yesterday [09:10] "sorry but in fact we keep it the current way" [09:11] mpt, well you IRC pinged about it and commented on the bug recently, which I took as an hint that it needed to be looked at [09:11] mpt, and where do you get your ranking from? ;-) [09:12] seb128, https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings [09:13] seb128, in my comment I said “we can separate out the choice of locale into its own step. If you have allowed location detection, possibly the step will be automatic, or at least the default can be based on your location” [09:14] None of that is designed yet, I was just documenting thoughts in progress (a) in case I forgot them later and (b) so you and others could tell me if I was making a mistake [09:14] (c) in case I get hit by a bus [09:14] k, sorry for assuming that was something to look at now [09:14] I guess it's was not all waster work [09:15] No, it’s not wasted, it just has a dependency [09:15] so I should put the change on hold until we get something else for the locale selection? [09:16] yes please [09:16] done [09:16] mpt, thanks [09:16] * mpt hugs seb128 [09:16] * seb128 hugs mpt back [09:17] mpt: please look left and right when crossing the street [09:18] err, right and left [09:27] it's London, they tell you which way to look [09:27] that is why nobody gets hurt there! [09:29] so... the other day someone showed me their laptop running 12.04 because they were having a problem with it [09:30] turns out that update manager was offering them an update to 12.10 and when they clicked it it failed [09:30] which was worrying them [09:30] anyone know if this is fixed in current u-m? [09:33] is 12.10 still available? [09:33] no [09:33] I thought we just supported LTS to LTS upgrades in those cases [09:33] oh [09:33] I see what you mean [09:33] it should have suggested an upgrade to 14.04 not 12.10 [09:34] that's a mvo question :-) [09:34] mvo, ^ you might know? [09:34] I wonder if it cached it [09:34] or if it is still a current issue [09:34] boot a 12.04 iso in a vm and try :-) [09:35] ya [09:35] Laney: uh, did he/she changed /etc/update-manager/release-upgrades at some point? the default for a lts is lts there [09:36] mvo: It's possible, I don't know - the problem is offering an upgrade to an EOL release [09:36] mvo, well, even if that was changed to "next version" if the next version is out of support we should suggest the lts no? [09:38] I did change it to LTS and then it updated okay [09:39] bug 1199060 [09:39] bug 1199060 in update-manager (Ubuntu) "installing/updating tools need to detect EOL versions" [Wishlist,Triaged] https://launchpad.net/bugs/1199060 [09:40] I guess it's still current then [09:41] mpt: The problem here is that the *next* release is EOL [09:41] understood [09:41] should we fall-back to trying the next LTS? [09:41] I guess only if the current release is LTS [09:41] otherwise, the upgrade path isn't supported [09:41] yes, otherwise you die [09:42] and so it should tell "please reinstall" [09:42] "Sorry pal, better reinstall" [09:42] oh man, that makes me sad [09:42] don't be sad :p [09:43] We talked at UDS about just supporting this but it never made it through to being policy [09:44] Is it not possible to upgrade to an EOL LTS just as a step towards the current LTS? [09:45] I suppose so, using the old-releases archive [09:46] This applies even (maybe moreso) on non-LTS releases too [09:46] Right. I guess what I really mean is, how much fiddling would a user need to do to do that. (E.g. would they have to edit text files.) [09:46] because the only supported upgrade path there is via the next normal release [09:47] but then, what's support anyway if you're upgrading to a release which itself is EOL? [09:47] I think you'd have to use apt yourself and point sources.list at old-releases.ubuntu.com [09:50] Not sure we make any guarantees about the availability of that archive either. And I don't think it's mirrored so IS would probably cry if any amount of traffic were to hit it. [09:51] I'll probably skirt around this hell hole and simply implement "if we're on an LTS and the next regular release is EOL, then try the next LTS instead". [09:52] Something like this? [09:53] Basically this, yes [09:53] +1 [09:53] Although the installer is supposed to let you keep your files [10:18] * Laney commits a sin [10:23] git commit -a -m "I have sinned" [10:27] it's for the greater good [10:27] * Laney 's fingers slip [10:48] hi all from Cambridge! [10:50] hi Sweet5hark! [10:56] good choice Sweet5hark [10:56] get yourself down the devonshire arms [10:57] Laney: will try! [10:58] you could also get the guided bus and go visit my mum ^o) [10:58] insert your mum joke here === rickspencer3_ is now known as rickspencer3 [12:03] chrisccoulson, hihi, Ubuntu Browser on Desktop Next crashes on startup. http://paste.ubuntu.com/11262972/ any ideas? [12:03] chrisccoulson, oh, and the log file mentions "legacy browser". Am I running a really old version? [12:05] willcooke, is EGL_PLATFORM set in the environment for the process? [12:06] how do I tell? [12:06] willcooke, env | grep EGL_PLATFORM from a terminal [12:07] chrisccoulson, thx - so, no it's not. I expect I need to set that in the /etc/environment? [12:07] what do I set it to? [12:09] Mir? [12:09] I'll try that [12:10] why is that required? [12:11] dunno. But it fixed it [12:11] "We don't know how it works.... BUT IT DOES" [12:12] Would be good if it could detect the right value and only use env vars to override it [12:12] but... don't know about this stuff :) [12:17] willcooke, I reported bug 1457458 [12:17] bug 1457458 in Oxide ""No suitable EGL configs found" on desktop-next" [Medium,Triaged] https://launchpad.net/bugs/1457458 [12:17] note, I can't actually run desktop-next (I just get a blank screen when I log in), so I can't test this [12:18] chrisccoulson, hey, did you try on real hwd or in a vm? [12:18] what Ubuntu release are you on? [12:18] seb128, real hardware (intel). And this is on vivid [12:20] thx chrisccoulson [12:20] FWIW, I installed from an old ISO yesterday and it works: [12:20] https://chinstrap.canonical.com/~will/vivid-desktop-amd64-FINAL.iso [12:21] it's not really final [12:21] is that the same as http://cdimage.ubuntu.com/ubuntu-desktop-next/backup-20150422/ ? [12:21] yes! [12:21] that one is on a non secret url ;-) [12:22] good call, thx Laney [12:26] chrisccoulson, can you share your ~/.cache/upstart/unity8.log ? [12:29] oh FFS, now what is up with gvfs? [12:32] seb128, one second, I need to install desktop-next again [12:34] I was just approaching having working autopkgtests for everything again [12:34] and now this starts failing [12:34] * Laney weeps [12:39] seb128, there's only 2 lines in it [12:39] () [12:39] QXcbConnection: Could not connect to display [12:40] chrisccoulson, how do you start the session? using unity8-desktop-session from lightdm? [12:40] seb128, yeah [12:41] weird, it looks like it's using the wrong backend or something [12:41] is there anything useful in the lightmdm greeter log? [12:41] or the system compositor log? [12:41] maybe try to wipe .cache/upstart, try to log in and see what's in there [12:42] ChrisTownsend, ^ does that sort of issues ring any bell to you? [12:43] seb128: Yeah, it's using the wrong backend. bschaefer was hitting that yesterday, but we thought it was due to him mucking around with Mir and getting his libs screwed up. It works fine for me. [12:45] Is there any way I can tell network manager to stop using the ofono settings plugin? (I have to keep uninstalling unity 8 because of bug 1391842) [12:45] bug 1391842 in network-manager (Ubuntu) "Mobile broadband doesn't work in utopic (worked in trusty)" [High,Confirmed] https://launchpad.net/bugs/1391842 [12:45] chrisccoulson, do you have some touch init scripts installed? [12:45] cyphermox, ^ [12:45] seb128, no, this is a fresh profile [12:45] chrisccoulson, "profile" means system install? [12:46] seb128, no, I just created a new user account to test unity 8 [12:46] chrisccoulson: remove "ofono" from /etc/NetworkManager/NetworkManager.conf [12:47] cyphermox, I have. It's still messed up [12:47] well, you also need to restart NM [12:47] that's all there is to it [12:48] otherwise, you're looking at a bug in MM rather than an issue with ofono [12:49] cyphermox, I restarted NM and I still have a bunch of non-functioning entries in the menu for me 3G adapter [12:49] The only way I've ever been able to make it work is to uninstall ofono [12:50] chrisccoulson: I don't know, you shouldn't have ofono entries if the plugin is disabled [12:50] try to see if they are in /etc/NetworkManager/system-connections? [12:50] but they shouldn't be [12:50] chrisccoulson, right, but you don't have e.g ubuntu-touch-session installed right? [12:51] (unless you tried to edit them) [12:51] chrisccoulson, bug #1238151 [12:51] bug 1238151 in ubuntu-touch-session (Ubuntu) "ubuntu-touch-session is installable on amd64 but prevents graphical login" [Undecided,New] https://launchpad.net/bugs/1238151 [12:51] seb128, I don't have that installed [12:51] hum, k :-/ [12:51] chrisccoulson, anything useful in the greeter log? [12:52] ChrisTownsend, did you debug that a bit with Brandon? any clue what to look at? [12:53] seb128: Unfortunately, I went offline before he said reported any of his findings. Maybe bregma heard from him... [12:53] bregma, ^ [12:53] I have heard nothing more [12:55] seb128, it doesn't look like it. But unity-system-compositor.log has this - http://paste.ubuntu.com/11263708/ [12:56] Looks like Mir doesn't like your monitor. Should probably check with the Mir guys about why. [12:58] chrisccoulson, how many screens do you have? [12:59] if you are docked maybe try undocked [12:59] seb128, I'm docked. But I had a blank screen the last time I tried it undocked too [12:59] chrisccoulson, try reporting a bug on mir, and/or ask in #ubuntu-mir [13:06] oh dear, I just re-discovered http://people.canonical.com/~pitti/photos/pitti-virtual-desktop.png (from 2009) [13:06] waouh [13:06] things changed a bit since ;-) [13:07] * pitti is currently cleaning up his public_html, but I'll keep that :) [13:07] seb128: well, the terminal/weechat layout hasn't for me :) [13:07] and I still only have one panel at the top and nothign else [13:07] pitti, what was the dog icon? ;-) [13:08] seb128: I *think* it was a custom launcher to run offlineimap [13:08] seb128: it's the mutt icon [13:08] ah ok, I saw that one before but I couldn't remember where/what it was [13:09] beautiful theme :) [13:09] and pidgin! [13:28] larsu: and ekiga! [13:28] larsu: I actually still use pidgin [13:28] there was some talk about teaching telepathy about OTR, but it seems it never went anywhere? :-( [13:28] seems nothing is going anywhere with telepathy [13:36] pitti, there are patches, and even package in my ppa [13:37] but then nobody really cared to push further to get it merged, and I lost interest [13:37]