=== maclin1 is now known as maclin === timchen1` is now known as timchen119 === maclin1 is now known as maclin === Guest91097 is now known as fredp === fredp is now known as Guest327 [06:34] Hi === Guest82926 is now known as pitti === pitti is now known as Guest48131 === Guest48131 is now known as pitti [07:06] Good morning === pitti` is now known as pitti === pitti is now known as Guest54153 === Guest54153 is now known as pitti === faenil_ is now known as faenil [08:49] good morning desktopers [08:49] morning seb128 [08:49] bonjour seb128 ! [08:49] hey willcooke [08:50] g'morn pitti [08:52] hey willcooke pitti [09:06] yo [09:06] moin! [09:07] Happy Birthday, the Document Foundation! https://www.youtube.com/watch?v=IVkkZ64hJJA [09:07] hey Laney Sweet5hark1 [09:07] Sweet5hark1, ;-) [09:09] Sweet5hark1, so, I updated libreoffice in xenial and now my menus are not integrated in unity anymore ... known issue? [09:17] hey all [09:18] seb128: yes, because we install libreoffice-gtk3 in the default. That sneaked in as a dependency from Debian. Already fixed locally. [09:18] hoy all! [09:19] seb128, didrocks: is xorg-gtest specific to Ubuntu ? I cannot find anything about it in Debian repo [09:19] seb128: The fix is to not continue to build libreoffice-gtk3, but not install it by default. This is what we did before. [09:20] hey alexarnaud [09:20] Laney, did you get your bike back? [09:20] NO [09:20] Sweet5hark1, oh ok ... why is gtk3 not having integrated menus? ;-) [09:20] WAH [09:20] ??!?!?! [09:20] Laney, oh? :-( [09:21] I broke the new inner tube [09:21] *or* it was already broken [09:21] but more likely that I did it [09:21] but the bike was there? [09:21] seb128: There will be some fallout from people running prerelease xenial, who will have libreoffice-gtk3 installed now. But people installing fresh from media or update postrelease should be unaffected, once the fixed package is there. [09:21] aye [09:21] Sweet5hark1, k [09:21] fixed the puncture this morning [09:21] so another trip out to the station tonight ... [09:22] "fun" [09:22] seb128: (that "fallout" is actually not too bad: We also need _some_ people to test libreoffice-gtk3) [09:22] right [09:22] so you didn't reply at the "why is gtk3 not having integrated menus" [09:23] I though you got that working at the london sprint? [09:23] seb128: I've find more things on freedesktop, it seems I need to investigate [09:23] seb128: as for the gtk3 doesnt have unity menu: yes, that was fixed in london. And apparently regressed upstream -- havent looked into that yet. [09:25] alexarnaud, it seems like that simply nobody packaged it for Debian [09:25] Sweet5hark1, ok === faenil is now known as faenil_ [09:26] seb128: OK, I see [09:26] Sweet5hark1, also, http://people.canonical.com/~seb128/libreoffice.png ... is the text on the left pane supposed to be that big? [09:26] seb128: Did you know the link with Compiz ? I see it in Compiz dep. [09:26] alexarnaud, it's used for tests [09:27] seb128: Ok, I'll learn more about it. Thanks a lot for your time [09:27] yw [09:27] alexarnaud, just curious, but what are you working on? [09:27] seb128: FWIW, I also have a MIR we should punch through: ucpp. In total, we have: 1/ fix autopkgtests 2/ build libreoffice-style-elementary too from new upstream 3/ continue building libreoffice-gtk3, build dont deploy in default 4/ MIR ucpp instead of using bundled [09:28] I'm working on porting Compiz from Ubuntu to Debian [09:28] Sweet5hark1, k, seems like a good todo ;-) [09:29] seb128: So far libreoffice-gtk3 is worked on by some core hacker, I dont think he cares too much about finer UX/design things there (yet) [09:29] alexarnaud, are you just interested by having compiz in Debian as an user? or do you plan to build something on it? wouldn't be easier to build whatever you are doing on Ubuntu? [09:29] seb128: It seems that everything work good but I've dropped gtest but I need to check if it essentiels or not to keep it in the final package [09:29] yeah, it's not a runtine thing [09:29] doesn't impact on how compiz works [09:29] Sweet5hark1, I uninstalled gtk3, see the menus are not there :p [09:30] Sweet5hark1, it's a screenshot of the gtk2 version, the sidebar text feels big [09:30] seb128: I work at hypra.fr team and we use Debian as work base === faenil_ is now known as faenil [09:31] We provide repository with old Compiz version and we would to be more in contact with community, that's why we would like to make Compiz directly in Debian [09:31] I see [09:31] seb128: argh. but yeah, looks the same on libreoffice 5.0.5 -- so is intentional. [09:31] seb128: My coleegue ksamak work on connecting Compiz with AT-SPI2 to manage a11y event [09:31] alexarnaud, good luck, it still feels like things would be easier for you if you were based on Ubuntu ;-) [09:32] Sweet5hark1, so it's not only my machine ... ok, if that's the way it supposed to be :-) [09:32] seb128: maybe yes but some of us are in Debian community and our mentor are an active Debian contributor [09:32] k [09:33] :) [09:33] seb128: the todo list is already commited, I just need to cut through the redtape for the MIRs. Is MIRing ucpp a "feature"? (as there is ff tomorrow) [09:33] Sweet5hark1, I don't think it is [09:33] seb128: good, thanks. [09:33] switching from a bundle to a system shared lib feels like a bugfix to me [09:35] seb128: and we like debian cause it's free /troll [09:35] ksamak, :-) [09:36] ksamak, sorry I didn't want to troll, it just feels like that this "packaging compiz" is taking you guys quite some work, it was an honest suggestion [09:37] but yeah, if you have the free slots to get that done it's nice for everybody [09:37] would be good to have compiz uptodate in Debian ;-) [09:54] * Laney is being spammed with apparmor notifications [09:56] what is getting denied? [09:56] ntp [09:56] just got updated [09:58] hm, I'm just getting hundreds of "powercap intel-rapl:0: package locked by BIOS, monitoring only", that sounds differently [10:10] seb128, have updated https://code.launchpad.net/~seb128/software-properties/dialog-enable-proposed/+merge/286213 with the final text from design [10:11] willcooke, is "title" the window title or the bold part of the text? do we really need the "warning:" if it's the bol text? [10:12] seb128, dialog = Gtk.MessageDialog(self.window_main, 0, Gtk.MessageType.QUESTION, Gtk.ButtonsType.NONE, _("Are you sure you want to enable updates from proposed?")) [10:12] that one ^ [10:13] k [10:13] so that's the bold text [10:13] having "Warning:" in there seems like poor form to me [10:14] that dialog is a warning one, icon and bold text convey that [10:14] seb128, yeah, fair enough. Drop the "warning" [10:15] willcooke, thanks [10:15] willcooke, also any though about Brian's comment? [10:15] the fact that "proposed" has different use on devel and stable series [10:15] do we want different messages/behaviours? [10:18] I think it's a fair point, but this warning is for "normal" users running a released version. I expect people running a pre-release will have more of a clue about proposed [10:18] so I think we just stick with a single implementation [10:19] keeps it simple for us too [10:19] i.e. no, we dont want different messages / behaviours [10:19] k, wfm [10:19] :) [10:19] I was pondering not showing the proposed box on devel [10:20] since that shouldn't be used and people who need it usually know how to deal with the command line [10:20] but that can be a later change [10:20] agreed [10:23] gstreamer FAIL [10:24] * seb128 looks at https://lists.freedesktop.org/archives/gstreamer-announce/2015-October/000359.html and is annoyed [10:24] they were supposed to have 3 pre/rc between decembre and feb and have 1.8 out by now [10:24] none of that happened [10:24] :(( [10:24] just wrote a comment on that gst/rb/crossfading/noise issue [10:25] which is fixed in git [10:25] asking what's going on with that schedule [10:50] willcooke, I pushed an update to the software-properties/proposed dialog, but I've a feeling mpt wouldn't approve of the wording. I think it's unusual to have a confirmation dialog providing choices without having it's text stating what is asking about ... like it doesn't say "you are about to enable proposed, are you sure?" it just say "bla is dangerous", which the user could take as "wth is it telling me about bla" [10:52] like nautilus doesn't say "deleting by passing the trash can lead to data loss", it tells you "are you sure you want to delete those? you are not going to be able to get those data back later" [10:53] mpt, care to comment? https://code.launchpad.net/~seb128/software-properties/dialog-enable-proposed/+merge/286213 [10:59] seb128, willcooke: Something that’s specifically for testers and Ubuntu developers shouldn’t sit like a mine in the middle of a UI that isn’t. That’s why I specced that tab shouldn’t offer -proposed at all. [10:59] Putting up a sign “WARNING: Mines can blow your leg off” is not really a solution to the problem of the mine being there. [11:01] seb128: so, is bug 1546421 relevant for ff? Its already fixed in LibreOffice (which defaults to breeze on unity if available), but the one line change to ubuntu-meta to install -style-breeze instead of -style-human isnt in yet ... [11:01] bug 1546421 in libreoffice (Ubuntu) "package libreoffice-gnome 1:4.2.8-0ubuntu3 failed to install/upgrade: trying to overwrite '/usr/lib/libreoffice', which is also in package libreoffice-common 1:4.2.8-0ubuntu3" [Undecided,New] https://launchpad.net/bugs/1546421 [11:02] Sweet5hark1, seems like more an uif item [11:03] seb128: hmm? [11:03] mpt, willcooke, we don't have the contributor console though and our documentation/SRU process relies on testers being able to enable proposed to test candidate SRUs [11:03] seb128: argh, wrong bug id [11:03] willcooke, and I already linked to that spec from the bug report that seb128 is working on, but ara changed the summary to assume the warning is a good solution. I disagree with that assumption. [11:03] Sweet5hark1, installing another style by default seems like more "ui freeze" material to me [11:04] mpt, I changed the summary after talking to willcooke [11:11] mpt, sort of false, there isn't a simple command to do that, but you could do a "echo "deb ... proposed" > /etc/apt/sources.lists && sudo apt-get update" with maybe some more scripting around to only add if if it's not enabled yet [11:11] Sweet5hark1, right, that mostly impacts the look so I would consider it ui freeze not feature one [11:12] seb128: alright, thx. [11:14] well, we're not going to do significant work on the software sources or contributor console in 16.04, or in fact ever, because U8. So we either remove proposed from there altogether or we go with a warning. Both are sub-optimal. Which is least bad? Given that all our docs talk about enabling proposed through that UI, I think the warning is better than nothing. [11:16] Those aren’t the only two options. For example, we could move that checkbox to its own tab labelled “Developer” or something like that. That way you could have warning text below it without needing a dialog. [11:18] Does “because U8” mean “We’re not going to let you install .debs any more”? [11:19] mpt, @ U8 - rather that U8 will have a new UI and so development will happen in there not in U7 [11:19] i.e software sources will be re-implemented in U8 [11:19] debs will still exist, so they'll need to implement an equivalent, which can be done correctly [11:20] I'm not sure about adding a new tab. It won't be a huge amount of work, but it will render all the docs wrong [11:27] that said, it's probably not beyond the wit of a developer to work out what to do [11:27] seb128, thoughts on adding a new tab? How much work> [11:27] ? [11:28] * Sweet5hark1 filed two MIRs today already. Working hard for peace in the world! [11:31] 'bout time [11:31] ;p [11:33] willcooke, it's probably not a lot, like half a day work [11:43] pitti, seb128, IIRC langpack-o-matic is also creating the touch language packs right? Someone is asking me about bug 1491381 as it seems the Breton translation should have reached the threshold for which language packs are created [11:43] bug 1491381 in ubuntu-keyboard "Breton language Ubuntu touch" [Undecided,New] https://launchpad.net/bugs/1491381 [11:44] morning [11:44] hey andyrock, dpm [11:44] dpm, no idea about that sorry [11:45] seb128, np, we'll see what Martin has to say [11:49] dpm, btw - I think I know why you couldn't launch apps in u8 session [11:50] dpm, try installing libpam-cgm [11:51] willcooke, tried that already, but thanks :) [11:51] dpm, bah. [11:51] there's a bug about that [11:51] and I had some help from hallyn trying to debug it last week, but I had to give up at some point [11:52] I'm planning to resume my tests some time this week [11:52] willcooke, bug 1535058 [11:52] bug 1535058 in ubuntu-app-launch (Ubuntu) "applications close instantly when launched from the launcher or dash" [High,In progress] https://launchpad.net/bugs/1535058 [11:52] dpm, so all I can offer is on a fresh 16.04 install, installing u8 session and that pam package allowed it to work for me === mfisch is now known as Guest22953 [11:53] I'll try it again in a bit, perhaps a week of updates fixed it for me [14:04] good morning, attente [14:11] good morning [14:14] hey desrt attente [14:14] good afternoon seb [14:15] hi seb128! [14:15] * desrt would like to point out that she was up at 6 but just forgot to say hello [14:16] I guess you had coffee by now then ;-) [14:16] * desrt hears the dish washer cleaning the french press already [14:16] * xnox tries to upload all the things for the feature freeze [14:17] i am not posh enough to operate a french press (well, not smart enough actually) hence i have dolce gusto pod coffee thing. [14:18] xnox, all the things include that aptdaemon pkgkit1 port, right? ;-) [14:18] * desrt gets annoyed at cgroups [14:20] seb128, haha [14:21] xnox, great, I knew we could count on you, thanks! ;-) [14:21] first time I learnt this was a "french press" was abroad TBH [14:22] so, not really popular here ;) [14:25] too bad. it's wonderful. [14:25] i think it makes my favourite kind of coffee [14:25] aka "very strong" :) [14:31] I love french-press coffee too. I grind my beans a little finer than suggested to get it a little muddy, too. More coffee is better. [14:35] is freedesktop.org down for others as well? [14:36] ye [14:36] k, thanks [14:39] mterry: wrt bug 1487148, libreoffice was already depending on fonts-stix before the bug was filed. So should the state be: "libreoffice - fix released" as it has the dep and "font-stix -- fix commited" as it is still in components-mismatch? [14:39] bug 1487148 in libreoffice (Ubuntu) "[MIR] fonts-stix -- to replace xfonts-mathml" [High,Triaged] https://launchpad.net/bugs/1487148 [14:39] mdeslaur, do you plan to merge/update freetype for xenial? it looks like the new version would be nice to have for the lts [14:39] seb128: for some reason I get that this page is immutable (https://wiki.ubuntu.com/Unity/Theming), I did change it few days ago... Do you know how is this possible? [14:40] Sweet5hark, oh ok sure. I just saw the comment that went with the fix released was an auto-close due to a "no change rebuild" so it looked like some fluke. Can set back to Fix Released [14:40] Trevinho, the wiki has been locked to deal with some spam issue [14:40] seb128: I don't plan on it, no [14:40] Trevinho, from what I read [14:40] seb128: mh, who can I ping for getting access? [14:40] As I'm able to edit other pages... [14:41] Trevinho, try #canonical-sysadmin but they have it in their topic [14:41] unsure why you can edit other pages [14:41] maybe they lock concerns only some sections/pages [14:41] " Known issues: wiki read-only for spam cleanup" [14:42] ah, thanks... [14:43] Trevinho, btw the nautilus not matching issue is still there today after a session restart :-/ [14:43] what info can I provide? [14:44] brb, retrying with a fresh session [14:44] seb128: so, could you recompile unity? [14:45] Trevinho, yeah, same on a fresh session :-/ [14:46] seb128: if you can recompile you should add some output in FileManagerLauncherIcon::IsLocationManaged [14:46] Trevinho, yeah, I can recompile, do you have a patch/some hints on what info you want printed? [14:48] mdeslaur, would you consider it if asked nicely? ;-) [14:49] mdeslaur, there is a bug asking for the update stating that it would make security fixes/updates easier for the LTS [14:49] not sure if that's a true claim though [14:50] it has improved thread-safety at least from https://sourceforge.net/projects/freetype/files/freetype2/2.6/ [14:51] seb128: hrm, maybe. will that unblock your graphite2 issue? [14:51] seb128: something like this should be fine http://pastebin.ubuntu.com/15100404/ [14:52] mdeslaur, I don't think it's needed for graphite2, that's blocked on MIR and mterry wanting a team subscribed but us and foundation not agreeing on who should maintain fonts :p [14:52] seb128: hehe [14:53] seb128, you should both agree on a third team to maintain it :) [14:54] mterry, can I just subscribe you? ;-) [14:54] * Laney applies for font team engineering manager [14:54] seb128, I'm not a team! :) [14:54] can I create a "mterry is the best" team and subscribe it? ;-) [14:54] a team of one is still a team [14:57] seb128, .... I'll allow it [14:57] Trevinho, ok, starting a build, it's going to take a while (i5 and no ccache/pre-build) ... btw unity depends on libgee-dev, can we get that changed to -0.8 or removed (Gee is used in a logger .vala source only but not mentioned in any build file so I wonder if it's even used) === Guest22953 is now known as mfisch === mfisch is now known as Guest32940 [14:59] mterry, I'm not enough in troll mode today to do that, I subscribed desktop-bugs, not that those subscriptions make much of a difference anyway... https://bugs.launchpad.net/ubuntu/+source/fonttools/+bug/1538173 should be good to be re-reviewed when you have some slot [14:59] Launchpad bug 1538173 in fonttools (Ubuntu) "[MIR] fonttools" [Undecided,New] [14:59] mterry, I keep the team idea in mind for another time though ;-) [15:00] seb128, awesome, approved [15:00] mterry, thanks [15:11] seb128: about the gee thing, I think we can remove it... [15:11] Trevinho, good :-) [15:12] seb128: also still about that, the migration I've ready doesn't get approved by britney... [15:12] seb128: https://requests.ci-train.ubuntu.com/static/britney/xenial/landing-040/excuses.html [15:12] or maybe Laney ^ ? [15:12] The thing is unity doesn't build in that arch, so I don't know how I can ignore it [15:12] I guess you can since it's not a regression [15:12] I was suggested to set unity as build-depend on s390x for unity-lens-applications, thus the error [15:13] In fact, but... Without britney approval, the train doesn't allow to get things published [15:13] what do you mean "set unity as build-depend on s390x for unity-lens-applications"? [15:13] So... I'm like blocked [15:14] https://launchpadlibrarian.net/240168318/unity-lens-applications_7.1.0+16.04.20160212.1-0ubuntu1_7.1.0+16.04.20160217-0ubuntu1.diff.gz [15:14] what is this diff? [15:15] what he just described [15:15] ... [15:15] is that to trigger a dep-wait rather than a build fail? [15:15] in any case it looks wrong [15:15] exactly... [15:15] Which I didn't want't to do... [15:15] who suggested that? [15:15] "what he just described" is not a helpful response [15:15] seems like a crazy workaround from ci-lander [15:15] But I was said it was a way to do this.. [15:15] it looks weird [15:15] ci train guys [15:15] tell them to force publish [15:15] and revert that hack [15:16] I can't believe they can't force publish over britney [15:16] I hope so [15:16] but please revert that s390x hack [15:16] who exactly told you to do that? [15:16] indeed, I hated myself in doing it.. [15:17] the old package built on s390x [15:18] log from ci-eng http://pastebin.ubuntu.com/15100523/ [15:19] Laney: yes, it builds... But then there's a missing dependency [15:19] what was the actual problem initially? [15:20] Laney: njargh Prodstack access to squid.internal, I just bumped into this *again* [15:20] Laney: unity-lens-applications has a dependency on unity because of the icons it uses.. [15:20] So not really a dependency, but... [15:20] Laney: so, once it built on that arch there was the problem that one of its dependency was not available there [15:20] Trevinho: but this part isn't new? [15:20] pitti: oh dear :( [15:20] no [15:20] Trevinho: so why did it become a problem? [15:21] Laney: the build-deps is new (now reverted), but the depends was there [15:21] If it wasn't a problem last time this got uploaded I don't see why it would be now [15:21] missing something here :) [15:22] Laney: I think the same, but... For some reason britney doesn't think that this is an old problem [15:22] Trevinho, what was the britney output before your hack? [15:22] Laney: I'm rebuilding btw, so soon we'll get back to the initial issue [15:22] it's right in what it is saying now [15:22] because the build is indeed missing, that's what you forced to happen [15:22] eh, we lost it.. But something like: unity dependency missing on that arch [15:22] well it's true [15:23] Laney: oh, there's quite some progress on your ticket [15:23] but not a regression afaik [15:24] pitti: sort of - they decided mimetype whitelisting was okay, then realised that you can't do this for https === bigon_ is now known as bigon [15:24] Laney: so you have to mark your r00tkit image/jpeg, and all will be good? *smirk* [15:24] don't really understand what proxying is for https to be honest, apart from a passthrough [15:24] pitti: indeed [15:25] I can also base64 encode it and copy&paste it in [15:25] this is a poor argument [15:25] * Laney just wants it to work at this point [15:25] what I don't get is how unity-lens-applications migrated to xenial if it has missing depends [15:25] seb128: no, not a regression... But britney didn't think that (like it does for unity not having libnux working in the same arch) [15:25] seb128: the s390x build happened in the release pocket [15:25] oh [15:25] because the arch got added later [15:25] we should maybe just delete that binary then [15:25] ? [15:25] could do [15:25] or make it a recommends [15:26] Trevinho, ^ [15:26] or split the icons in unity [15:27] pitti: what do you want it for this time? [15:27] Laney: api.github.com [15:27] Laney: I'd prefer not to touch the unity packaging... [15:27] as already I'm changing something in that front [15:27] Laney: github now makes test requests for systemd upstream PRs, but I can't tell it about their status [15:27] Laney: as for reccomends, I think it's better [15:28] Laney: so... if you remove the binary let me know... otherwise I'll go for reccommends [15:28] +1 for the recommends [15:28] you need to upload it anyway [15:31] Ok, rebuild in progress... Let's see.. [15:41] cool [15:41] willcooke, qengho: about video acceleration.... While reading the firefox bug related to that (it seems there's no hope in that front), someone mentioned that it's supported by Chromium... It seems that it needs to be patched with something like (https://github.com/fqj1994/chromium-vaapi, old but...). So, wondering what can be done on that side... [15:42] Trevinho, qengho - seems like a very early PoC more than a feature at this point. [15:43] Intel have u/s a load of VP9 patches for Cr. I think, so we might get that for free first [15:43] willcooke: well it's shipped on chrome-os [15:47] Uff, you're not supposed to be so slow, you arm64! [16:04] I've to drive back to Florence, it will take ~1.5 hr [16:14] Laney: you mention apparmor in your logs. is there a bug for it? [16:14] jdstrand: yeah, and nice hilight :-) [16:15] thanks :) [16:15] https://bugs.launchpad.net/ubuntu/+source/ntp/+bug/1546455 [16:15] Launchpad bug 1546455 in ntp (Ubuntu) "Many instances of 'apparmor="DENIED" operation="create" profile="/usr/sbin/ntpd" pid=15139 comm="ntpd" family="unspec" sock_type="dgram" protocol=0' in syslog" [Undecided,Confirmed] === Guest9335 is now known as fredp === fredp is now known as Guest59972 [16:18] Laney: did you talk to kickinz1? [16:18] jdstrand: nah, I just hoped assigning would generate mail [16:18] seb128: freetype merged [16:18] and then he would be like aaaahhh jdstrand please help me [16:18] or something :) [16:18] mdeslaur, you are a hero ;-) [16:18] seb128: you owe me INTMAX beers [16:18] indeed! [16:18] Laney: heh [16:19] Trevinho, shrug, my nautilus matching issue was my fault :/ [16:20] Trevinho, I had a nautilus build without XUbuntuOpenLocationsXids (I commented out the patch while you were refreshing it and I used a version number higher than the one uploaded so was still on that) ... which is why I couldn't find it in d-feet [16:20] Trevinho, works fine with the archive version [16:21] :D === Guest32940 is now known as mfisch [16:36] larsu, hi! [16:38] larsu, I was wondering if you have time for one last request - you might have some ideas on bug 1546276 [16:38] bug 1546276 in gnome-software (Ubuntu) "Install button doesn't show loading bar in Ambience/Radiance" [Medium,Triaged] https://launchpad.net/bugs/1546276 [16:55] mpt, hi, have you been following the GNOME Software work? [16:56] robert_ancell, no, not at all [16:56] mpt, are you interested or would know someone on design who would be? [16:58] robert_ancell, bjkeyser is the right person to ask [16:58] mpt, thanks [16:58] hi robert_ancell [16:58] how's the sprint? [16:59] Laney, hi. Going pretty good. [16:59] sweet [16:59] * Laney is sending a "fix your appstream plz" email [16:59] right... about... [17:00] Laney, to... everyone? [17:00] *gets gpg passphrase right the first time* [17:00] NOW [17:00] yeah [17:00] Laney, is the appstream stuff mostly working now? [17:01] Laney, and should I do anything to remove the stuff your shell script installed? [17:01] it should be working [17:01] actually I don't know which location is preferred [17:01] I would just delete the stuff that it downloaded [17:05] robert_ancell, hey [17:05] robert_ancell, gnome-software's MIR got acked (and gcab one as well conditional on a fix), do you think it's ready to switch the seed? if so can you do it today or tomorrow (before ff)? [17:05] Laney, so I'm going to delete /var/cache/app-info/icons/, /var/cache/app-info/xapian/, /var/cache/app-info/yaml/ - that makes sense to you? [17:05] They don't seem to be owned by any packages and haven't been updated recently [17:06] seb128, yeah, I was going to ask your opinion. I think it's ready in the sense it works, but has bugs. But I think that makes sense for FF right? [17:06] yes [17:06] And users can keep using USC so it's only new users that will get switched [17:06] right [17:06] seb128, ok, I'll do that today [17:06] thanks [17:07] seb128, thank you [17:07] don't forget to unseed oneconf as well [17:07] yep [17:07] good work btw! [17:07] I played a bit with it, feels nice [17:07] robert_ancell: I think so, the system one is /var/lib/app-info/ [17:07] I updated it yesterday to read the dpkg/apt files directly. It's ridiculously fast now [17:08] I haven't tried it with a rotary drive though, but I hope it should still read pretty fast [17:10] seb128, who's the best for GTK+ themeing issues? [17:11] * willcooke sits a little lower in his chair [17:11] robert_ancell, I can add it to my themes todo [17:11] willcooke, heh, I was thinking you'd inherited that title... [17:11] :) [17:12] If it's a fairly easy css issue, I can probably fix it [17:12] at least, I can have a crack at it [17:12] Bug 1546276 is my issue - it's CSS that doesn't work with our themes [17:12] bug 1546276 in gnome-software (Ubuntu) "Install button doesn't show loading bar in Ambience/Radiance" [Medium,Triaged] https://launchpad.net/bugs/1546276 [17:12] The CSS in gnome-software just adds a bar to the bottom of the button. [17:12] It might be our theme renders over it, or ignores such changes [17:12] oki, I can check it out [17:13] I've assigned it to me [17:13] willcooke, ta [17:13] nw [17:13] thanks willcooke ;-) [17:15] oh noes I forgot to block software-center [17:15] better unblock all the other stuff === JMulholland_ is now known as JMulholland === alan_g is now known as alan_g|afk === fredp is now known as Guest23882 [17:53] seb128: ah, I knew there was something weird 😃 === alan_g|afk is now known as alan_g [18:12] robert_ancell, I've insalled g-s from the archive on a fresh 16.04 with no PPAs etc. [18:12] is that the right thing to do? [18:12] do I still need to run laney's script? [18:13] it should work after an apt update [18:14] from the archive [18:14] * Laney is off, goodnight! [18:14] Hi guys, wrt the recent mail about gnome-software and appdata, two questions,... [18:14] one, when will gnome-software actually replace software center? [18:14] and two, I'm upstream + debian maintainer for 2 packages, can I sync them with ubuntu (with the appdata changes) even after the debian import freeze tomorrow? [18:15] within days and yes [18:15] really goodnight! [18:15] Thank you Laney, bb! :) [18:15] cheers Laney [19:51] robert_ancell, Laney - it works! \m/ [19:59] right, it's fts o'clock. [19:59] night all [21:48] Sweet5hark, aroud? [21:48] around [21:49] robert_ancell: aye [21:50] Sweet5hark, in you lo snap, you have all the dependant libraries inside the snap right? [21:50] So it just depends on Ubuntu core [21:52] yes. [21:52] Sweet5hark, and snappy handles all the linking stuff for that? Or did you need to modify LO? [21:55] robert_ancell: I got it to finish building a libreoffice, by disabling tests. The result does print a help message, when you call it with ./soffice.bin --help ... as root ... with a LD_LIBRARY_PATH. It does nothing more. [21:56] Sweet5hark, root? [21:56] Beyond that it fails in ways that suggest subtle and mean toolchain errors. [21:57] Sweet5hark, where does it write the configuration? In the snappy directories or ~? [21:57] robert_ancell: hmmm, Im not sure if it even gets to that point ... [21:57] ok :) [21:58] Sweet5hark, but you would expect it to still go to ~ unless you patched it right? [21:58] robert_ancell: yes. [21:58] Sweet5hark, are there any files in ~ that it might access outside of its configuration? Even indirectly through a library. [21:59] I guess the GTK+ theming configuration [22:01] robert_ancell: hmm, never thought much about that. would need to strace it ... [22:02] robert_ancell: also how do you define "its configuration"? because "its configuration" might include executable code (e.g. extensions stored in ~) and such fun ... [22:14] Sweet5hark, sorry, was called into a conversation [22:14] Sweet5hark, yeah, I'm trying to think along the lines of "what would happen if you couldn't access ~" or "You can only access ~/Downloads, ~/Pictures, ~/.config" etc [22:18] robert_ancell: from what I see now it seems worse than that. namely, that some symbol export foo breaks C++ stdlibs in a subtle way inside snapcraft/snappy. [22:19] Sweet5hark, the C++ libs from ubuntu-core? [22:20] robert_ancell: hang on a sec, I will try to find the conversation on that and include it in the mail. [22:43] robert_ancell: you got mail. [22:43] Sweet5hark, it's all spam [22:48] robert_ancell: you still got nothing? [22:52] Sweet5hark, I got it. It's a Futurama reference :) [22:52] Though googling I think the correct line is actually "It's not spam" [22:52] Which makese more sense [22:52] heh [22:52] * Sweet5hark is too much of an old fart for futurama [23:05] Sweet5hark, are there any services that LO relies on (e.g. D-Bus) that would stop it working if they were removed? Because you can get all the libs in the snap but not the services. [23:06] I mean e.g. a D-Bus service, not D-Bus itself. [23:06] Though perhaps that is a candidate... [23:09] robert_ancell: well, we are certainly using various services, but we usually fall back soft when missing almost everything ... [23:11] Sweet5hark, yeah, I was thinking that apps like LO are a bit of a special case in that most things are optional. Other desktop apps rely on package dependencies to ensure their services are there. [23:12] So those sort of apps would need their services to be in snaps [23:12] robert_ancell: if what remains is still a full office suite platform might be debatable. I assume there will be people having very strong opinions on that. [23:12] I have raised that point [23:15] Sweet5hark, is that because it would take much work to get them all working or is there anything fundamental that would be lost converting to a snap? [23:18] robert_ancell: there is so much stuff that might break in containers, I just dont know how bad it might be and what would be the most painful things ... [23:21] robert_ancell: for an extreme, take e.g. extensions and scripting via UNO: they are supposed to be able to do anything and everything to integrate with $whatever and run in the same process as LibreOffice. Thats not a good fit to containerization. [23:23] robert_ancell: For the rest, TBH I wouldnt know. I have no idea if LibreOffice base is able to interact with any database source (mysql, postgres) anymore etc. -- and unittests dont help a thing, as they will be inside the container. [23:26] robert_ancell: (also one person can do manual testing on a LibreOffice release for a year and not cover all corners) To have any idea of realworld impact: publish PoC, buckle up seatbelts, ask for feedback, brace for impact [23:26] Sweet5hark, ack [23:27] (and better do it with volunteer early adopters enjoying the pain) [23:39] gaah what happened to my mono font :P