[07:40] “This message contains an event that has already been processed.” Dear Thunderbird, I have no idea what that means. [07:41] :> [07:44] I lost the overlay scrollbars on wily after an update this morning, how can I get it back? [07:46] jibel, you can't, those have been deprecated [07:46] in favor of the gtk equivalent [07:46] but those are not o-s, they are in the win [07:48] seb128, ah, currently I see a large light grey scrollbar, is it the equivalent? or there is a way to have a thin one like the overlay was? [07:49] jibel, that's the "equivalent", but they are unthemed atm [07:49] they are going to be changed to be thinner and more integrated [07:49] unsure why Laney and larsu decided to upload GTK before doing the minimal theming tweaks though :-/ [07:51] seb128, ok, thanks. I'll wait but this large light border is visually annoying and distractive [07:51] yeah, agreed [08:02] sigh [08:02] please don't beat on us [08:03] morning Laney :) [08:03] hi larsu [08:03] (working on totem now, didn't manage to last night) [08:03] np [08:05] hey Laney [08:06] Laney, larsu, is there anything that prevent us to just delete the custom scrollbar theming so they look less weird/buggy? [08:08] seb128: they don't look buggy. They look exactly like they looked before when you disabled overlay-scrollbars [08:08] I seriously don't know what the fuzz is about [08:08] (fuss?) [08:08] well, for somebody used to o-s they look like quite a regression [08:09] "oh, all my windows have the same scrollbar as firefox now" [08:09] see jibel's comment about them being "visually annoying and distractive" [08:09] I hope he doesn't look at firefox too often, then [08:10] except that in firefox they are there and don't change [08:10] in gedit they keep fading in and out [08:10] I find it quite distractive as well [08:11] fair enough [08:12] oh, that makes me realize a difference of behaviour with o-s [08:12] I guess I naively hoped that people on the development release could put up with things being in flux for a while. [08:12] with o-s the colored thumb doesn't fade out when you move the mouse out of the win [08:12] Instead of blocking the update and everything that depends on gtk 3.16 until we have the final experience [08:13] Laney, I might naively think that using the upstream styling would be deleting a blob of 30 css lines in our theme where it's likely more complicated? [08:14] I'm not speaking about having an updated look&feel [08:14] Yes [08:14] just about deleting our buggy rules [08:14] Unless I'm wrong this is GTK's default style [08:14] you are wrong :p [08:14] cf backlog from yesterday afternoon on this channel [08:14] default upstream is fine, their bars are thinner [08:15] OK then I'll review your merge proposal [08:15] Laney, https://blogs.gnome.org/mclasen/files/2013/08/zoomscrollbar.png [08:15] is how upstream looks like [08:16] k :-) [08:16] * seb128 bzr fetch the theme [08:17] What I resent is implying that we are somehow being lazy or deliberately making it crap [08:17] Maybe that's not what you mean but it feels like it to me [08:17] ;9 [08:17] :9 [08:17] FFS [08:17] :( [08:17] sorry :-/ [08:17] no, I meant that deleting those few lines in the theme seemed like an easy change [08:17] and I didn't get why that was not done before uploading [08:19] seems I wrongly thought that this was the default style & adwaita had themed it differently themselves [08:21] seb128: don't worry, I can do it today if people feel so strongly about it [08:21] * larsu has a 1:1 first though and needs fooooooooood [08:21] larsu, shall we begin? [08:21] oh [08:21] hrm [08:22] When your day starts with an email from "Have I been pwned".... [08:22] spam / phish? [08:22] willcooke: morning :) [08:23] Laney, larsu, sorry I didn't mean to make a fuzz about it either, they are just unpleasant and we started having users asking about them [08:23] no, some l33t haxors got the passwords for some forum I must have signed up to a long time ago [08:23] uh oh [08:23] seems I signed up with a different email address to usual [08:27] Laney, do you have any idea about https://launchpadlibrarian.net/208330071/buildlog_ubuntu_wily_i386_ubuntu-desktop-next_BUILDING.txt.gz ? [08:27] mv: cannot stat 'device.tar.gz': No such file or directory [08:29] nope, this is where you get out of my area of knowledge I'm afraid [08:30] the code is in livecd-rootfs [08:30] but I have no idea what device.tar.gz is meant to be [08:31] o_gra is probably the person to ask [08:31] yeah, I'm moving to -devel [08:31] cjwatson might also know [08:31] wait, is this just a bug? [08:31] pull-lp-source livecd-rootfs [08:31] look at live-build/auto/config [08:31] line 343 and line 356 [08:32] seems to do the same thing twice no? [08:32] http://bazaar.launchpad.net/~ubuntu-core-dev/livecd-rootfs/trunk/view/head:/live-build/auto/config#L343 ? [08:33] oops /auto/build [08:34] https://bazaar.launchpad.net/~ubuntu-core-dev/livecd-rootfs/trunk/view/head:/live-build/auto/build#L343 [08:35] hum [08:36] that was part of didrocks changes in https://code.launchpad.net/~ubuntu-desktop/livecd-rootfs/desktop-next/+merge/257056 [08:37] yeah, feels like a bug [08:37] if you look on that mp, on the first version he duplicated the section [08:38] and Colin in the review said to "Rather than duplicating this between ubuntu-core and ubuntu-desktop-next, maybe share this code? " [08:38] I guess he should have deleted the extra section then [08:38] * seb128 does that [08:39] nod [10:17] Laney: how do you want the backported totem patch? [10:18] larsu: on the bug is fine [10:18] will it fix the other issues? [10:19] which are the other issues? [10:19] it won't add the fullscreen menu button back [10:19] I don't know that we can SRU a patch which removes that [10:20] hm, I was planning to do this work on gtk master :/ [10:20] backporting all of that would be quite a lot of work [10:20] ya [10:20] I guess we could leave the bug in vivid? [10:20] or just try it and see what the SRU guys say [10:21] how about removing those two sections in the fullscreen menu? [10:21] Languages and Subtitles, I mean [10:21] not sure where people use them more [10:22] are they going to be broken there? [10:23] if we remove the prefers_app_menu check for the fullscreen case [10:23] yes [10:24] this is the whole issue: we can't have dynamic menus in two different sections of the ui [10:24] oh right [10:24] gtk somehow is deciding which one to use [10:28] yes, it simply uses the first [10:28] it can find [10:28] (it's by id) [10:30] what do you think we should do? [10:30] SRU and delete the fullscreen app menu, SRU and keep it without those dynamic bits, don't SRU? [10:30] I was about to ask you the same :) [10:30] what's the main fix for the backport? [10:31] willcooke: right now, you can only select subtitles from the fullscreen menu in totem [10:31] not from the menu bar [10:32] of course it should be in both, but we've never had the case where an application needed to insert dynamic menus in two places at the same time [10:32] and we're fixing that, but that's too big a change for an SRU [10:32] I imagine that once it's in full screen mode then people generally wont want to change menu items [10:32] like, [10:32] yes, I agree [10:32] you get it set up and then once your done you go fullscreen [10:32] better in the "real" menu than in the fullscreen one, if we have to chose [10:32] *choose [10:32] +1 [10:33] Laney: d'accord? [10:33] fair enough [10:33] The SRU team gets the final decision here anyway [10:33] did you fix the empty menu case? [10:33] ok, thats good [10:34] no, I simply left it removed [10:34] doing that now [10:35] I don't know if removing it, making it insensitive (if possible) or showing "No languages" is preferable [10:35] probably avoid a new string [10:35] No languages would be wrong, no? [10:35] No selectable languages [10:35] this is clunky :) [10:36] well, you could still select them in the menu bar [10:36] I'm talking about the other issue [10:36] empty submenu if there are no languages to choose from [10:36] ah. should be insensitive in that case, not hidden [10:36] cool [10:36] oh wait, weird [10:36] I guess you're right [10:37] I'd even remove it [10:37] but then that wouldn't signify that there are no languages to choose from [10:38] * larsu changes opinion to "should be hidden" [10:38] that's quite a ui change for an sru thoguh [10:38] ya maybe insensitive is best there [10:38] * Laney doesn't know [10:54] argh! [10:55] so, this needs much more work [10:55] all of the plugins use dynamic menus as well [10:55] so the fullscreen gear menu won't have screenshot, rotation, etc [10:55] * larsu was wondering why this menu was so sparse [10:57] flexiondotorg: you here? We've got a test failure in nemo: https://jenkins.qa.ubuntu.com/job/wily-adt-nemo/lastBuild/ARCH=amd64,label=adt/console [10:57] Should these files be added to POTFILES.in? [10:57] Laney: this is not really fixable unless we touch every plugin [10:58] because they don't know about the possibility of multiple menus [10:58] haha [10:58] there just like: where should I insert my items? Ok, insert(), insert(), insert() [10:58] can totem handle it itself? [10:58] no [10:58] connect to a changed signal and populate the other menu ... [10:58] like the rb link menus thing [10:58] (or however that works) [10:58] urgh [10:59] it could... but really? [10:59] :) [10:59] and it would need to be hardcoded for every plugin [10:59] because we need to know the section id [11:00] they way plugins insert menus in totem is kind of bad [11:00] but that's not something to fix in an SRU === MacSlow is now known as MacSlow|lunch [11:02] agreed [11:02] Laney: so, I think we should backport with the fullscreen menu removed [11:02] let's see what SRU people say [11:03] but I'd be uncomfortable with bigger patches in there [11:03] this is already quite bug [11:03] *big [11:16] sure [11:16] so just the insensitive/remove/... empty menu thing I think? [11:18] yes [11:18] and I think we should go with insensitive [11:18] (again) [11:19] to signify "there's something here sometimes" [11:19] I wonder what upstream does [11:19] ah, empty sub sections [11:26] lemme know when it's ready! [11:26] * larsu is cursing desrt [11:26] things that should be simple: making a submenu item insenstive [11:27] now let me tell you about something that is impossible… [11:30] larsu: why? [11:30] desrt: morning :) [11:30] hi :) [11:30] how do I make a menu item that has a submenu insensitive? [11:30] you don't, apparently :) [11:31] ya... [11:31] only thing i can guess about is via submenu-action [11:31] doesn't work [11:31] well, it was just a guess :) [11:31] my first guess as well [11:35] Laney, can you retry the desktop-next image build? [11:35] yus [11:35] livecd-rootfs is in wily [11:35] thanks [11:36] should be running [11:38] thx === alan_g is now known as alan_g|lunch [11:56] Laney: new string is impossible to SRU? [11:56] hiding and making insensitive is impossible without patching gtkmenutracker [11:56] actually hiding might work if I move the creation into the .c file [11:57] remove and add an insensitive item? [11:57] it wouldn't look like a submenu item then [11:57] i.e., missing the triangl === benonsoftware is now known as MerryChristmas [11:58] indeed [11:58] removing altogether is probably better than a new string [11:58] unless you can find one which is already translated === MerryChristmas is now known as benonsoftware [11:59] I was thinking about "Empty", which would be a very small patch [11:59] if I remove I might as well make insensitive [11:59] thought you couldn't do that [11:59] do what? [11:59] or you can in code? [11:59] yes, I can move the whole thing into code [11:59] really ugly [12:00] but better than what we have now === MacSlow|lunch is now known as MacSlow [12:00] maybe Empty is already translated? [12:01] it is, but only in a couple of languages [12:01] probably an old string [12:01] * larsu checks in the only language that matters, fr [12:02] what are you talking about? [12:03] that was much quicker than I thought :) [12:03] seb128: totem shows empty sub menus when no language/subtitles exist for a movie [12:03] sound the fr alarm [12:03] we're looking for the least-intrusive way to SRU a fix [12:03] what should it display? [12:04] either hide the menu item or make it insensitive [12:04] honestly, I think we could leave it as is [12:04] and find a proper way to do it in W [12:04] (but don't tell Laney that) [12:05] har de har [12:07] Laney: I've attached the backport for now. Gotta grab some lunch [12:08] yeah, if the menu is supposed to be empty it's just cosmetic [12:08] not the end of the world [12:09] never said it was [12:09] I am surprised that this is unfixable though [12:24] Laney: me neither. I'm disliking GMenu more and more :/ [12:24] * larsu actually goes to lunch now [12:24] oh well, let's take what you have done already [12:25] thanks for the patch! [12:26] Laney: sorry :( [12:26] no sweat [12:27] larsu, Laney, did ctrl-f/g still works for you in gedit to cycle through search matches? [12:33] ctrl-g does, f doesn't do that but raises the search window again [12:33] seems the same on a utopic vm [12:34] oh there is a difference: ctrl-g doesn't work if the search field is focused [12:39] Laney, right, that's what I meant [12:39] usually I do ctrl-f then press ctrl-g to cycle [12:39] cftrl-f used to work as well in the focus box [12:39] iirc [12:45] seems to be ctrl-g and shift-ctrl-g but that doesn't work inside the box [12:45] up/down does, that's what I always used [12:45] but probably is a bug I guess [12:46] I think it's due to the gtk update [12:46] need to try with old gtk [12:47] other keys work [12:47] eg ctrl-shift-k to clear the highlight [12:49] Laney, I guess we need https://git.gnome.org/browse/gedit/commit/?h=gnome-3-16&id=18cd0a02e0c654da0ce9817385989acc61a0263c [12:50] https://git.gnome.org/browse/gedit/commit/?h=gnome-3-16&id=611d788b15c68722b7d5aeaf42634ada2e6ccd06 [12:50] esc doesn't work either [12:51] ya just found that [12:52] I will look this afternoon [12:52] thanks [12:52] still trying to fix autopkgtest fun [12:54] flexiondotorg: I uploaded that nemo thing, could you take a look/forward/fix differently if you want to? [12:54] or let me know what you want [12:55] (assuming it works) === alan_g|lunch is now known as alan_g [13:31] seb128: weird, indeed [13:31] I wonder if this is because of hergetme's accel changes [13:31] Laney: let me know if that commit seb128 linked to doesn't work out [14:35] Laney, when you are back can you retry the desktop-next build? [14:45] seb128: another one or is that obsolete? [14:46] Laney, another one [14:46] ok! [14:46] thanks :-) [14:46] we should set up a release team for desktop-next then you can do it yourself from the iso tracker [14:47] hopefully once they start working I/we don't need manual work anymore though [14:47] ha [14:48] seeing the amount of stuff they have in there for core I bet a few rounds of tweaks will be wanted [14:48] running [14:51] thanks [14:54] https://launchpadlibrarian.net/208366019/buildlog_ubuntu_wily_i386_ubuntu-desktop-next_BUILDING.txt.gz [14:54] nearly there ;-) [14:55] (not a clue why we have this) [14:56] bah! [15:04] Laney, do you understand what's the group mismatch from the log? is that just the md5 from the log to copy to the config? [15:05] I guess you have to update the file in the source with some piece of output [15:05] can't think what it is for though [15:05] maybe we don't need it for desktop-next [15:06] yeah, I don't understand much those bits [15:10] bah, I don't understand [15:11] fairly sure it is og rah's work if you want an explanation [15:12] lol [15:15] Laney, looks like 02-add_user_to_groups.chroot isn't used [15:15] dunno why [15:15] and dunno why ubuntu-core doesn't have the same issue [15:16] echo "I: add $USER to ($DEFGROUPS) group(s)" [15:16] that's not showing in https://launchpadlibrarian.net/208331223/buildlog_ubuntu_wily_arm64_ubuntu-core_BUILDING.txt.gz [15:16] nor https://launchpadlibrarian.net/208366019/buildlog_ubuntu_wily_i386_ubuntu-desktop-next_BUILDING.txt.gz [15:18] it's the 00-uid-... one [15:19] Laney, you mean? that fails? [15:20] yeah, but what I don't get is that the config there includes the "docker" user [15:20] but docker is not installed [15:21] unsure why ubuntu-core doesn't get the same issue [15:21] it seems to overwrite the files [15:21] if they match some known hashes at the start of the run [15:22] yeah, I'm unsure why the md5 are different between -core and -desktop [15:25] I guess we have some different stuff installed [15:25] I'm pondering just updating the md5 from the config with the ones from the log :p [15:26] trying to follow https://lists.ubuntu.com/archives/ubuntu-devel/2011-June/033458.html to see if I can debug that locally instead first though [15:26] would be good to know what it's for [15:26] ogra_: can you educate us or share a link about the passwd/groups hash stuff in livecd-rootfs? [15:29] Laney, cant link, no ... i dont think it was ever documented ... effectively livecd-rootfs has checksums for all readonly password files and is supposed to fail if the passwd db differs in any way from that so you can manually adjust the values ... (UIDs and GIDs can not change since we are readonly) [15:29] ogra_, why does it need a copy? [15:30] ogra_, is there an easy way to see the differences? [15:30] seb128, it doesnt have a copy ... only a checksum [15:30] the build log shows a diff if the files differ [15:31] ogra_, e.g https://launchpadlibrarian.net/208366019/buildlog_ubuntu_wily_i386_ubuntu-desktop-next_BUILDING.txt.gz is a failing build [15:31] unsure what is wrong though [15:31] seb128, right, so you scroll to the end and there are the diffs ... then you compare to an installed system [15:32] ogra_, "installed system"? [15:32] we don't have one, we are trying to get that iso going :p [15:32] seb128, if you do your very first build i guess you first need to do a local debootstrap and create the right checksums [15:32] ogra_, any idea why they are different from ubuntu-core? we should have a similar base [15:32] creatin the starting point is all manual [15:33] you likely have other packages (recommends) installed [15:35] ogra_, I should compared the content from http://bazaar.launchpad.net/~ubuntu-core-dev/livecd-rootfs/trunk/view/head:/live-build/ubuntu-desktop-next/hooks/00-uid-gid-fix.chroot_early to the bootstrap one then? [15:36] ogra_, diff gives me http://paste.ubuntu.com/11590387/ [15:36] seb128, well, adjust your check to match your file and you are done ... [15:36] ogra_, but I don't understand why ubuntu-core has docker it is config (http://bazaar.launchpad.net/~ubuntu-core-dev/livecd-rootfs/trunk/view/head:/live-build/ubuntu-core/hooks/00-uid-gid-fix.chroot_early) where docker is not installed [15:37] probably a hack that mvo added [15:38] (ask him on monday) ... [15:39] yeah [15:39] ogra_, thanks [15:40] oh, and that passwd db thing will definitely bite you more [15:40] you mean? [15:40] the writable passwd db is handled by libnss-extrausers ... [15:40] which adduser can not properly work with [15:40] (there is a bug open since ages) [15:40] it is an overlay ... [15:41] well, desktop-next is a copy of ubuntu-core, if core works desktop should as well? [15:41] core doesnt add users :) [15:41] and touch doesnt either [15:41] so both dont have that problem [15:42] (i mean. both add users, but hardcoded ones during build time) [15:42] desktop-next shouldn't add users either [15:42] ah, k [15:43] i thought you were aiming for an actual desktop install [15:43] (with adding a personalized user etc) [15:43] installer might need to exist first :) [15:43] heh, ok [15:43] ogra_, no, aiming at doing an image with a different seed [15:43] so more installed from there [15:43] well, once we need to add users that issue will bubble up again [15:44] he's just trying to get a first image to work from atm [15:44] yeah [15:44] basically core-system-image with the desktop-next seed installed [15:45] yup [15:45] btw, if you have any idea how we can solve the UID/GID thing any better .... [15:45] this obviously scales very very badly :) [15:45] shrug, I don't understand [15:45] "/etc/passwd post-debootstrap hash doesn't match record" [15:45] would need to understand the problem it solves first [15:45] but passwd_bootstrap="1050cdb2fcde4d10c86399616f64dfdb" [15:46] $ md5sum cache/bootstrap/etc/passwd [15:46] 1050cdb2fcde4d10c86399616f64dfdb cache/bootstrap/etc/passwd [15:47] Laney, direcotries on disk are owned by a certain UID/GID ... during build UID and GID for users are assigned dynamically ... so each build can differ ... but the readonly rootfs wont adjust the dir ownership so we need to make sure UID and GID always stay the same [15:48] so if a UID or GID changes during build, we dont want that image to finish building but tell us there was a change so we can force the ID back to what the rootfs expects [15:48] the wonderful world of readonly rootfses :) [15:50] Alright I'll try to think about it at some point [15:50] btw,m there are two ID checks ... one after bootstrap and one after the chroot was filled [15:51] seb128: I think just update the hash and the contents of the files that get outputted [15:51] (the latter indeed adds more system users) [15:51] right [15:51] to establish the baseline [15:51] for your first image just update to what you get manually [15:51] * Laney switches topics *swoosh* [15:52] Noskcaj: darkxst: can you update the gnome shell extensions? [15:52] Noskcaj: darkxst: holding gtk and shell back in proposed because they dep on specific versions of shell [15:52] I did -extensions itself but the other seem more irritating :) [16:12] cyphermox: yo [16:12] cyphermox: how can I start ubiquity to the slideshow step? [16:17] Laney, http://paste.ubuntu.com/11591093/ seems fine to you? [16:17] doubt we need the docker hack [16:18] can't you just copy and paste passwd/group/etc into the file? [16:18] Laney, what docker hack? [16:18] you kept the docker user [16:18] well [16:20] you want me to replace the ones in that file with the system ones? [16:21] I think it makes sense to start from those [16:21] well, I don't understand why the docker hack got added [16:21] but I think that's something useful for snappy [16:21] probably to demo a snap of docker [16:23] but that is just my advice/feeling, keep it if you think it is better [16:23] well, I don't understand snappy enough to know if it's needed for some snappy feature to work [16:23]