[00:04] robert_ancell: here is a recent output of libreoffice ./configure --help. I assume every second or third --with-foo or --enable-foo option might break: http://pastebin.com/NpkmD4BR [00:05] --disable-coinmp Disable compilation of the CoinMP solver [00:07] robert_ancell: oh, yes: we will also bundle dictionaries for all languages with the snap, I assume? [00:08] Sweet5hark, Can they download on demand? [00:12] robert_ancell: so far we provided them as proper packages installed along with lang support. But yes, if they find an language dict on extensions.libreoffice.org they might install it (to ~) just like windoze users would. Except that many dictionaries arent published on that page, but spread all over the internets, so many will do some easter egg hunt for their dicts ... [00:12] ... and then end up installing something malicious from an untrusted source. [00:12] yep [03:27] Laney: can you refresh the upload rights for ubuntukylin packageset? [05:29] does lightdm auto-adjust the scaling factor now? [05:32] scaling in unity doesn't seem to work at all (xenial) [06:07] Hi [07:04] Hi, so, to check how our packages appstream data shows up, we're supposed to install gnome-software, right? [07:04] It's so bare... it always lists apps as "coming from a 3rd party", it doesn't list the version or license or source or size of the programs etc [07:04] It's not even in the gnome menu, I have to launch it from the command line [07:05] Am I testing it wrong? [07:06] I search for "gedit" and it shows "seahorse", it doesn't have an index of available programs... is it supposed to be so broken or am I looking at the wrong application? [07:12] * alkisg just ran `apt-get install gnome-software`, hope I'm not missing any suggested packages that make it less functional... [07:40] good morning desktopers [07:42] Good morning [07:43] hey pitti! [07:43] wie gehts? [07:45] morning amigos [07:45] it's not my morning, I was just checking back on IRC before passing out. [07:46] ha ha [07:47] hey cyphermox [07:47] hey seb128 ! terribly tired, I hardly slept at all for some reason [07:47] lol I was going to ask [07:47] cyphermox, have a good night ;-) [07:47] pitti, oh :-( [07:47] pitti, strong tea time! [07:47] good day ;) [07:47] seb128: dentist time too! [07:47] good luck [07:47] cyphermox, thanks ;-) [07:50] moin! [07:57] hey Sweet5hark1! [08:01] seb128: heya [08:04] going to grab a coffee and something to eat, bbiab [08:25] morning [08:29] Hi guys, let me ask again, gnome-software currently is quite empty (no index in the inital page, no categories, no working search etc), is that to be expected or is it an issue with my local installation? [08:36] (I'm asking here because a mail instructed us to use it to check the appstream data in our packages, and to direct questions to #ubuntu-desktop) [09:01] morning all. Still no SSL on Freenode for me :( [09:03] hey andyrock willcooke Laney [09:03] hello! [09:04] alkisg, did you apt-get update since you installed it? do you have the current appstream version? [09:04] Laney, doing good today? [09:05] seb128: I've apt-get updated, `apt-cache search appstream` shows nothing, is that a package? [09:06] alkisg, dpkg -l | grep appstream-util [09:07] Whoops sorry I ran the previous one in a 12.04 chroot. So, I have "appstream 0.9.1-1ubuntu" but not appstream-util. [09:07] Installing... [09:08] you might not need util [09:08] and what is after that 1"ubuntu" [09:08] It didn't make any changes [09:08] you cut the part which is interesting [09:08] 0.9.1-1ubuntu1, sorry small term [09:08] that's old [09:08] we are at 1ubuntu6 [09:09] urg [09:09] no, sorry [09:09] what about libappstream-glib8? [09:09] * alkisg apt-get updated a few minutes ago... [09:10] alkisg, do you get any result if you "appstreamcli search ."? [09:11] Yup, I get a lot of results there [09:11] dpkg -l | grep gnome-software? [09:11] libappstream-glib8:i386 0.5.8-1ubuntu6 [09:12] what version do you have for that one? [09:12] gnome-software 3.19.5+git20160212.a64f331-0ubuntu1 [09:12] I'm out of ideas sorry [09:12] maybe Laney has a clue [09:13] seems to be an issue on the gnome-software side [09:13] Do you guys see the software in categories? [09:13] can you run it with --debug and share the log? [09:13] E.g. I have a freshly installed xenial vm, if I just run `apt-get install gnome-software; apt-get update` then I'm supposed to see the software categories etc? [09:15] Ah, much better [09:15] gnome-software was running as a pid even without a window [09:15] So all my tries were relaunching the already running, hidden version of it [09:15] * alkisg sees categories etc now... [09:16] (after pkilling it ) [09:16] good [09:16] Thank you seb128 [09:16] yw [09:20] hey, sorry, had a few emails to read ;-) [09:20] don't like that you have to update another time after installing appstream [09:20] hey Laney ;-) [09:20] hey Laney [09:20] right [09:20] * Laney wonders how to fix that [09:20] can we make that update a trigger [09:20] not really [09:20] or in the postinst or something [09:20] it's a config file to apt [09:21] Apart from that, there was also the problem of gnome-software not really closing when its window is closed [09:21] to download some more files from the mirror [09:21] yeah that's a gnome thing, apps do that now [09:21] alkisg, right, that's worth a bug report [09:21] I wonder if that's a new gnome thing [09:21] got similar problems with calendar [09:21] Laney, do you know why? [09:22] it means things keep using resources even after being closed [09:22] that has to be a bug [09:22] or are we going phone mode and never closing anything and hopping the oom does the job? [09:23] I don't really know much about it [09:23] If ran from the command line, when I close the window, I don't get a prompt, and when I press ctrl+c, the "background process" closes normally, so I'm guessing it's a gtk quit handling issue... [09:23] or a "feature" [09:23] but please report a bug about it [09:23] but it's about single instance [09:25] hey pitti! [09:32] hey desktopers [09:32] Laney, is gnome-software supposed to use the same screenshots data than software-center? [09:32] hey darkxst, how goes? [09:33] Laney, e.g https://screenshots.debian.net [09:33] ? [09:33] hey darkxst [09:33] I think it can fall back to that but upstream are supposed to supply them but also ximion mentioned that he needs to fix something to do with screenshots [09:33] hey willcooke, Im good again now! [09:34] darkxst, glad to hear it :) [09:34] Laney, what are the icon requirements for appstream? must have 64x or anything higher? [09:34] higher is ok [09:35] Laney, gnome-weather has 256x but is still failing [09:36] willcooke, the side effects from the miracle drugs are crap! [09:39] darkxst: lemme see [09:40] Laney, thanks [09:43] seb128, Laney - I've noticed some lag in the loading of the icons in categories. Is that normal? Oh, actually - no, today its fine. Ignore [09:44] willcooke, I didn't notice it [09:45] seb128, seems ok today. Might be a first run thing I guess. I'll see if I can reproduce it, but I don't think its a big issue [09:45] the UI feels responsive and looks ok [09:46] I miss being able to sort categories by rating though [09:46] I *really* like it so far [09:46] I wonder if that should be the default? [09:46] I think alphabetical should be the default [09:47] but a rating sort order would be really handy [09:49] hum [09:49] for me it doesn't get added to the unity dash "most recently used" [09:49] even after killing the service [09:49] I wonder if that has to do with the service mode [09:55] same here [10:33] oh [10:37] * Sweet5hark1 is feeling so snappy today. dunno why. [10:37] schnappi [10:37] * Sweet5hark1 thows some red meat in the air for Laney to fetch ... [10:43] lol [10:44] Sweet5hark1, breakfast wine is the answer [10:44] is that a thing? [10:44] it is in my house today [10:44] * seb128 would rather go for the whisky in the coffee [10:45] :D [10:45] :-) [10:46] wine? Ok, Im going with a Tullibardine 228 Burgundy Finish. There is burgundy in it, so it qualifies as wine on my book. [10:46] it's all just fruit juice [10:47] * willcooke goes for a top up [10:49] darkxst: haha [10:49] " # FIXME: Do we support icon names which contain a dot? [10:49] " [10:52] Laney, oh right, wtf! [10:52] "ummmmmmmmmmmm, yes please" [10:55] Laney, I'll send them a whipping, though tomorrow, sleep now [10:55] darkxst: looking at the code [10:55] oh [10:55] ximion is here! [10:59] ximion: looking at https://appstream.debian.org/html/sid/main/issues/gnome-weather.html and http://appstream.ubuntu.com/xenial/universe/issues/gnome-weather.html - seems you have a fail-fast path if the icon filename contains a '.' [10:59] Laney, so screenshots, I guess it has some support because it displays the widget just says the image is invalid when there is/should be one [10:59] which breaks this case [10:59] it does [10:59] where for some others it doesn't display the widget [10:59] just needs some kind of fixing [10:59] yeah [10:59] no worries [11:00] also it seems there is still the bug about it considering packages third party non free [11:00] robert had an upload to fix that I think, I guess it doesn't work/needs more fixing [11:03] Laney: heh, I even knew about that when writing the code, but assumed there were no apps having dots in their icon filename... So, again, more complexity is needed in the icon handler [11:03] I'll fix that in a few minutes [11:04] ximion: can you make the check smarter or just delete it? [11:04] ximion: also another check needed for ubuntu contents files, pr coming up :) [11:04] and hey, :) [11:05] how's deutschland today? [11:05] Laney: I need to make that check smarter, removing it will break a lot more other apps [11:05] the icon stuff is really dumb ^^ [11:06] great and sunny :) - I'm not at the lab today, which is why you find me online :D [11:06] (at this time) [11:06] \o\ [11:10] Laney: what would be more important for Ubuntu in terms of metadata: having font components in the AppStream metadata, or having information about which software is localized in which language to what percent? [11:10] ximion: what does gnome-software know how to deal with? [11:11] the latter is so complex on Debian based systems, I am not sure if the additional processing effort would be worth the potential gain - on Fedora, this is a bit easier to implement, since they don't split the software into so many binary packages [11:11] (unless it's LaTeX, there they seem to go nuts) [11:11] you would parse the .mo files or something? [11:12] Laney: yes and possibly other formats - problem is, I need to find them too, which means another call to the Contents list and extracting even more packages, and the package extraction is the really expensive thing [11:12] I am actually thinking about rewriting some of those performance-critical things in C and have Python call that code... Not 100% sure about that yet, though, since python-apt also hooks into what is mostly C++ code [11:13] GS can handle both [11:13] but only uses the l10n data for sorting, and I think it also sometimes displays an "available in your language" badge [11:13] I guess you would want to factor out the contents handling [11:13] there is scope for modularisation here [11:14] problem is, that if a binary uses 10% of it's srings from its own gettext domain, but the other 90% come from shared libraries, it's impossible to give an accurate measurement of localization [11:14] willcooke, seb128: there is one blocker with nvidia still holding back the new xserver, so maybe I need to file a FFE bug? [11:15] Laney: reading any .deb file data is by far the most expensive task, followed by reading the Contents.gz files, followed by reading other Packages.gz and supplimentary data - that are the big cost items, at least, there are a few smaller ones which sum up since they're called often [11:15] but not as much as .deb file reading, which is both expensive and called always [11:17] that's unavoidably expensive [11:17] Laney: btw, when you boostrap a new architecture or suite using the generator, you can now use the prepopulate-cache action, which will at least set the packages which definitely have no interesting data to ignore [11:17] that speeds up the initial processing run massively [11:18] nice [11:18] but requires that your contents file is up2date, otherwise you will loose information [11:18] exists for a while now, but I didn't document it yet, to search for issues [11:18] I reckon we could make use of a copy-suite action too [11:18] since our new releases start out as copies [11:18] * desrt yawns [11:19] Laney: that should work out of the box, since the database doesn't know about suites at all [11:19] only about package versions, so when a new suite is added, it will only process the package versions it doesn't know yet and skip the rest [11:19] (= take the very same data) [11:19] why do you need the prepopulate thing then? [11:20] only when you deleted the cache or when you add a new architecture, which brings in tons of new package/version/arch combinations [11:22] (when I want to reprocess everything, it's really useful) [11:23] a delete-by-tag command will also be nice in future [11:23] I wrote a script to do that [11:23] * Laney is cool [11:28] * Laney PRs [11:33] tjaalton, seb128 - ack [11:43] Laney: do you still remember the bugzillla # about "double-click in terminal doesn't select URLs any more"? [11:43] AFAIR upstream considered that a feature, not a bug [11:44] pitti: my awesome bar seems to have forgotten about it [11:44] mine too :/ [11:44] I use right click -> copy url [11:44] Laney: I found https://bugzilla.gnome.org/show_bug.cgi?id=640727, but that's older and the inverse problem [11:44] Gnome bug 640727 in general "Links having a colon in the middle are not selected completely" [Normal,Resolved: fixed] [11:48] tjaalton, wfm, should be easy to get [11:53] Laney: oh, bug 1501250 and that links to the upstream bug [11:53] bug 1501250 in Gnome Virtual Terminal Emulator "double clicking on a URL drops the protocol from the URL" [Medium,Confirmed] https://launchpad.net/bugs/1501250 [11:54] ah [11:54] I was searching for 'select' [11:55] willcooke, seb128: another update I'd like to get in is mesa 11.2, which will release -rc1 tomorrow and final around Mar 11th. then it would still get first or maybe even second bugfix pointrelease before final freeze [11:56] i'll push rc1 in the staging ppa [12:00] Laney: I dropped the dot-check afterall, there is no way to implement this safely if we assume that dots in icon filenames are common [12:01] btw, the Contents.gz of unstable should be included on Ubuntu even when main is processed, AFAIK [12:02] ximion: that's what I thought, either that or do a check for the filename + extensions which would be expensive and this is supposed to be a fail fast case - thanks! [12:02] ximion: universe? [12:02] it is - but we can't read it because of that bug [12:03] ah, then it's good that this is fixed now, thanks! [12:03] tjaalton, thanks [12:04] well, if we check against a list of filenames, we still get and error, because ".Application" isn't recognized as allowed icon filetype [12:04] interesting: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=815006 [12:04] Debian bug 815006 in iceweasel "Renaming Iceweasel to Firefox" [Normal,Open] [12:07] good morning everyone [12:08] willcooke: looks mostly like compromise from the debian side to me [12:08] hey hey desrt [12:08] hi desrt, still suffering with jetlag? [12:08] suffering? [12:08] THIS IS AWESOME [12:08] hehe! [12:08] if i knew what waking up at 6am was like i would have been doing it all along! [12:08] (...until next week when i finally give out) [12:08] haha [12:08] if only waking up at 6am was like that normally [12:09] the key to waking up at 6 is going to bed at 10 [12:09] which, so far, has not been too big of a problem [12:09] once that starts slipping, i'm doomed [12:11] " In case of derivatives of Debian, Firefox branding can be used as long as the patches applied are in the same category as described above. " [12:11] that's a pretty weak statement from mozilla, and it's surprising to see debian change their position based on that [12:11] considering "in the same category" in this case means "architecture-specific portability patches" [12:15] good to see debian back off a bit on this issue, in any case... some moderation here is probably not a bad thing [12:17] what's the argument [12:18] it's okay because if we needed to rename we could do it? [12:18] the argument is (seemingly) that mozilla released a friendly (albeit weak and vague) statement toward debian so we should all feel nice === hikiko is now known as hikiko|ln [12:20] indeed it would have been nice if they had included something along the lines of "we guarantee that we will never ask you to change the branding of firefox in a stable release or after feature freeze" or something like that [12:21] but they didn't do that [12:22] but this was never the original argument (at least as far as i understood it)... it was about the fact that even if mozilla was happy about debian's usage, it's clear that they may have become unhappy about derivative usage if that usage made more substantive changes to firefox... which was a restriction that debian saw as out of step with the DFSG [12:23] this remains completely unaddressed except by that one sentence about "derivatives can do the same stuff that you've been doing"... which is not much... [12:25] of course mozilla is no bully here... they need to protect their mark, and there are real cases of people abusing firefox's reputation to harm users in the real world... it's always been about how "morally flexible" debian is willing to be with the letter of the law in order to do something that is clearly moral in spirit [12:25] and that, more than anything else, seems to be what is changing here. an interesting development. [12:27] there was a specific problem originally about some patches Debian made - Mozilla themselves explicitly revoked trademark permission [12:27] 10 years ago-ish [12:28] outright, or as an ultimatum? [12:28] it was a bug report like "hi, I'm from mozilla, please rename your package" [12:28] it was thunderbird, fwiw [12:28] and it was mike connor. nice. [12:29] Don't see why this couldn't happen all over again [12:29] I can see the arguments that it is in line with the DFSG [12:29] but seems annoying nonetheless [12:30] mozilla has said almost nothing that changes the situation [12:30] presumably a part of this is that the maintainer works for mozilla now [12:30] perfect. put a mozilla employee in charge of packaging and move it to non-free. problem solved :D [12:42] * Sweet5hark1 spills some oil into the discussion with: https://wiki.documentfoundation.org/TDF/Policies/Trademark_Policy#Use for comparision. [12:44] * ogra_ checks if he accidentially ended up in #ubuntu-legal [12:44] we're about as well informed as debian-legal :P [12:46] lets retroactively establish IANAL as an acronym for "I am not a layman" and sue everyone based in common law countries over legal advise! [12:46] seb128: looks like we managed the webkit switch in one image [12:46] want to do some hardcore demoting? [12:52] * desrt renames her fork to Liberated OpenLibreOffice === alan_g is now known as alan_g|lunch === maclin1 is now known as maclin === hikiko|ln is now known as hikiko [13:16] Laney, great, well done! sure I can do that ;-) [13:17] woot [13:17] now something weird has happened in this glib release [13:17] /usr/bin/install -c ./libglib-gdb.py /<>/debian/install/deb/usr/share/gdb/auto-load/usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.4706.0-gdb.py [13:17] nice path [13:18] but looks like it is actually right! [13:18] !!! [13:38] * willcooke late lunches === alan_g|lunch is now known as alan_g [14:33] Laney, sphinx didn't migrate yet so not much demotion [14:42] mitya57, ^ are you looking at that one? [14:50] seb128, not yet, will look now [14:50] mitya57, thanks [15:05] happyaron, hey, do we need fcitx to recommends fcitx-frontend-all? could we limit the recommends to (gtk2)gtk3&qt5? [15:07] seb128, Laney: uploaded pyresample with fixed tests which should unblock sphinx [15:07] Mirv, mitya57, Laney, willcooke, what do you think about making indicator-appmenu stop recommending appmenu-qt and unseeding sni-qt but make those rather a recommends from some qt4 binary? it looks like we could get qt4 out of the iso/default installation [15:07] mitya57, great, thanks! [15:09] seb128, +1 to this idea [15:12] Though, from Qt we can recommend sni-qt but it would be a bit ugly to recommend appmenu-qt (which is the server side of SNI spec) [15:12] * willcooke checks rdepends on qt4 [15:20] willcooke, on the iso it's only fcitx/appmenu/sni [15:20] oh, cool [15:20] which are integration with qt4 softwares, so not needed if there is no qt4 applications [15:20] sorry, I got distracted and forgot I was supposed to be looking at that [15:20] yeah, I sort of forgot too [15:20] :) [15:20] then +1 [15:20] thx [15:21] but we got a part of the qt4 stack uninstalled today with the software-center drop since that took off the ubuntuone package that was still using qt [15:22] which made me look at the remaining items [15:22] mitya57, yeah, I don't have a good idea, we don't have a way to state "install that if unity is installed" [15:23] we could let to users to install those but most wouldn't and wouldn't have correct integration [15:23] (have to go pick up somebody, bbiab) [15:23] OK, I guess it's not a problem to add a recommends on appmenu-qt, at least for LTS [15:24] If it's recommends, it's easy for users to avoid installing it if they don't want to. [15:25] right [15:45] anyone managed to install gnome-software from sources? [15:45] i used to be able before [15:46] but now i cant' install apps with my own g-s [15:46] i get "installation failed" even if the app has been installed [15:48] seb128: getting qt4 out would be very nice, I agree [16:07] wow, so software-center fell off the image? nice, congrats! [16:08] :) [16:08] good work team! [16:12] it was more pushed [16:13] (seed change [16:13] ) [16:14] seb128: g-s is going to use aptd to install/remove/etc things right? [16:21] seb128, should g-s be added to the launcher by default? (I dont think it is atm) [16:21] Laney, do you know how the "editor picks" work? [16:23] I think they are hardcoded [16:23] so they'll always be the same? [16:24] I wonder if we can mix that up, or indeed become "the editor" ourselves on a rotational basis :) [16:24] I would imagine that you would want to make it possible to change it externally [16:25] I'll open a bug with some thoughts === charles_ is now known as charles [16:30] Looks like plugins can influence this list [16:30] ah nice [16:30] https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1547095 [16:30] Launchpad bug 1547095 in gnome-software (Ubuntu) "Featured and Editors Picks are hard-coded" [Undecided,New] [17:07] Sweet5hark, is bug 958345 resolved for LO? [17:07] bug 958345 in ttf-indic-fonts (Ubuntu) "ttf-indic-fonts packages are outdated" [High,Confirmed] https://launchpad.net/bugs/958345 [17:07] I *think* it is [17:17] robert_ancell: sorry, will look at in 60 minutes -- in a call now === rickspencer3_ is now known as rickspencer3 [17:32] andyrock, yeah, it's using aptd, but better to ask robert_ancell about details [17:32] willcooke, yeah, good point for the launcher [17:34] robert_ancell, hey, g-s still seems to think that all softwares are third-party/non-free, did your fix still work for you? also is that normal/known that none of the "translated/documented/integrated with the system" seem ever acked? [17:34] seb128, I'll open a bug [17:35] robert_ancell, also the "updated" and "license" are unknown [17:35] seb128, yes, the third party thing is known - we need to mark things that come from main/universe. I have a couple of solutions, not sure how reliable they are. The non-free is because we don't know the licenses. That could come from appstream or we could just mark main/universe as "Open Source". The translated/documented/integrated come from appstream. [17:35] They're all fixable I think, though the solution might just be to hide them. [17:35] Please file bugs if not there already. [17:35] k [17:36] robert_ancell, do you know if there is anywhere in g-s to show prices? [17:36] I assume not since most things will be assumed to be free [17:36] willcooke, G-S has no concept of prices. [17:36] robert_ancell, willcooke, is that only me who find the "was this review useful" and "report abuse" buttons to be given too much importance? [17:37] seb128, Yeah, I see that too - ask aday [17:37] like they are quite noticable but I would assume they are not the most used thing in comments [17:37] They were designed / added by GNOME [17:37] k [17:37] would you usually ask on IRC or file a bug? [17:37] seb128, bug then IRC/email [17:37] k [17:37] thanks [17:37] aday seems very busy, so I never see him online [17:38] robert_ancell, oh, also do you know if it's normal that gnome-software doesn't exit on close? [17:38] seb128, yes, it seems to run some sort of service for noticing upgrades. We should probably disable that (there's a bug on LP relating to that dialog) [17:38] k [17:39] it's especially annoying because it makes the unity dash not list it [17:39] well I assume it's due to that, at least if it's running the dash stop listing it [17:39] but it's not in the most recently used even after closing so there might be something else [17:41] seb128, yeah, I thinking these feature can all be fixed between FF and UI Freeze [17:41] right [17:41] even after uif [17:41] those are mostly bugfixes [17:45] seb128, did you have any info on the desirability of the offline updates? [17:46] robert_ancell, I think it would be considered a good improvement [17:47] how easy would it be to enable? [17:47] seb128, do you know if it works on Ubuntu [17:47] seb128, it's there, but I don't know if it works [17:47] willcooke might know better about partner/design opinion on the topic [17:47] I've no idea [17:47] I can try to play with that tomorrow though [17:47] I figure we either fix/support it or disable the button [17:47] maybe Laney knows [17:47] It does somewhat overlap with update-manager [17:48] it's quite a big change [17:48] unsure it's safe to include now for the LTS [17:48] but it could be nice to have it working on possible to turn on [17:48] maybe some oem images would do that [17:49] sorry, what is offline updates? [17:49] willcooke, windows style, apply them on shutdown/reboot [17:49] it uses some systemd thing [17:50] don't know if that is set up for us [17:50] pitti, ^ do you know if that should work on deb/ubuntu? [17:50] but the idea is a sound one imo [17:50] willcooke, https://fedoraproject.org/wiki/Features/OfflineSystemUpdates [17:51] willcooke, that basically solve all the issue of doing in session updates/replaces things being in use under their feet [17:52] just reading it now [17:52] so the issue is that the UI which handles this is not Ubu-themed? [17:53] I guess that's one [17:53] I would be more concerned about the fact that it's quite a workflow change and a new untested stack handling your updates [17:54] thinking about it, that feels risky to do by default [17:54] willcooke, the issue is A: there is a button in GNOME Software, but not in update manager (Do we add one). And B: Don't know if there's work required to make it work [17:54] you don't want to have LTS updates handled by some new untested stack [17:54] ohhhh [17:54] sure why not [17:54] seb128, yeah, that was my feeling, unless someone has more experience to tell us otherwise [17:54] JOKE!!! [17:54] I crack myself up. [17:54] lol [17:55] robert_ancell, I would say to not bother for ff and not do it by default [17:55] willcooke, I heard your stress levels were too low, opt in for some more features! [17:55] Sounds like something for 16.10 to me [17:55] robert_ancell, :D:D:D [17:55] but getting it to work for those who want to opt in would be nice [17:55] see the conversation about breakfast wine from earlier on [17:56] willcooke, robert_ancell, I can see that as a nice thing oems install might want to opt in [17:56] seb128, yeah [17:56] Perhaps a gsettings key will be enough [17:56] well maybe it get enough testing and it's something we feel safe tell them they can use by LTS .3 or something [17:56] like get opt in from tech users [17:56] see how it goes [17:56] right, is it something which we could sru? sounds like yes [17:57] and it if it goes well we can decide to enable later [17:58] sounds wise to me [17:59] ugh, offline updates mean users don't get timely security updates [18:00] as opposed to thinking they have them because they've installed by not restarted [18:00] offline updates mean you know you have all updates applied [18:00] and waiting for your laptop to shutdown like windows is a pretty awful user experience [18:00] Laney: pff [18:01] * Laney just yesterday did a stupid update and restart services dance for glibc [18:01] ohhhhhh [18:01] thats what I was supposed to do this morning [18:05] now that every other platform has live app updates, let's try and not regress to 1999 with an update on shutdown scenario :P [18:06] it's for system components, which other platforms do offline [18:06] so it only suggest it when you have a need-reboot package? [18:06] and it's restart & update, not update & shutdown [18:06] ugh, even worse [18:06] whatever [18:06] bye [18:06] hehe [18:06] I hate it on windows as well [18:07] last time I want to play a game I ended up having to wait for windows to finish applying updates [18:07] and 1h30 later it was too late to play [18:07] FAIL [18:07] have you tried this implementation? [18:07] either of you [18:08] no, that was the start of the question, asking if it's supposed to work/if we can try it [18:08] isn't it what fedora does? [18:08] or is it different? [18:08] also I was commenting on the concept of blocking the computer to do updates [18:08] I don't need to try it to know that it's time I can't use the machine :p [18:09] but yeah, maybe it takes less than 15s and it's fine [18:09] need to play with it to see ;-) [18:13] robert_ancell, Laney, willcooke, reading the fedora wiki, there is at least a part which is implemented in packagekit so unsure it's going to work with aptdaemon [18:14] I bet it requires pk 1.0 [18:14] it's like -> tell systemd to boot into some target [18:14] wiki says 0.8.1 [18:14] -> in that target do the update [18:14] -> restart [18:15] nice [18:15] :-) [18:15] thing is we can't co-install packagekit and aptdaemon [18:15] and I guess aptdaemon doesn't implement those functions [18:15] so not going to work [18:15] it's probably just a dist-upgrade in that target [18:15] guessing [18:16] ...pkcompat must do that... [18:16] I'm going to give it a try after ff ;-) [18:16] seb128, I think we *can* install aptdaemon and PackageKit 1.0, as long as we remove the PackageKit compatibility layer from aptdaemon [18:16] rrrrrrrright, laters [18:16] Laney, bye [18:17] Laney, cya [18:17] Laney, nice work on the appstream stuff btw! [18:17] robert_ancell, right, which start sounding like too risky/much work for this cycle [18:17] seb128, exactly [18:17] so yeah [18:17] 16.10 thing then [18:17] yes [18:17] Unless fairies and unicorns solve the issues [18:18] (I'm out of those at the moment) [18:30] re [18:30] wb [18:37] robert_ancell: thx wrt bug 958345 -- resolved now. [18:37] bug 958345 in ttf-indic-fonts (Ubuntu) "ttf-indic-fonts packages are outdated" [High,Confirmed] https://launchpad.net/bugs/958345 [18:37] (for libreoffice) [18:38] seb128, can you remove from the archive? [18:38] Could you look at ^^ and remove those old font packages [18:39] robert_ancell, ok [18:40] willcooke, if you ever find yourself writing "should" in a bug title I'd say you need to reword it :) [18:41] that's what mpt said iirc [18:41] not sure I agree with that :p [18:42] seb128, it makes the description unnecessarily longer (harder to parse) and it implies you have already decided the solution rather than stating the problem [18:42] well, sort of [18:42] the first part is right [18:43] but like "shouldn't segfault on invalid password" is not really something you decided the solution on [18:43] robert_ancell, ack :) [18:43] sure it could be worded "segfaults on invalid password" [18:43] In that case the solution is (probably) obvious, but the later description is better [18:43] The "should" discourages thinking about alternative solutions [18:44] I tend to word things with "should" because I find it more "polite" I think [18:44] but that's probably a personal/non native speaker issue :p [18:45] seb128, yeah, I was thinking it might be that. To me "should" can be taken aggressively / be seen as arrogant [18:45] For example "GNOME Software should support the display of prices for purchasing software" says we have to modify GNOME Software in a specific way. The issue is "Support purchasing software" which could potentially be solved in other ways / codebases. [18:46] to me should says "it would be nice" [18:46] where the direct form is more an order [18:46] like "do that" [18:46] seb128, that's the literal translation, but if I said to you "You should leave the room" it's quite an assertive statement. [18:46] right [18:46] I'm not really giving you a choice. [18:47] but people don't use should that way in bug titles [18:47] or that's not the most common case [18:47] they tend use it for e.g "should give better feedback about foo" [18:47] to say "having better feedback would be nice" [18:47] My interpretation is the more agressive one. Of course that will vary. But also Engineers aren't known for having the best communication skills. [18:47] but yeah, I see your point [18:47] right [18:48] point taken, you are probably right ;-) [18:48] language is hard! :p [18:48] Yes. [18:48] It's not as tightly defined as code. [18:57] robert_ancell: hi, does lightdm do some autoscaling in xenial, or is it unity-greeter? [18:58] tjaalton, that would be unity-greeter [18:58] alright, thx [18:58] tjaalton, you are going to say that the transition to the desktop doesn't look right? ;-) [18:59] seb128: that too, everything looks tiny [18:59] and scaling in the session doesn't work [18:59] how doesn't work? [18:59] moving the slider does nothing [18:59] wth? [19:00] wfm [19:00] is it logged somewhere? [19:01] what logged? unity? [19:01] .cache/upstart [19:01] yeah [19:01] of course.. [19:01] hum, I tried rhythmbox-plugin-alternative-toolbar to reproduce a rb segfault (just uploaded a fix for that) [19:01] their toolbar layout is actually nice [19:03] * willcooke installs [19:04] willcooke, you can enable it in the tools->plugins [19:04] but there are segfault when you try to play (fixed with the libdmapsharing upload I just did) [19:05] changes the plugins window too [19:05] yeah [19:05] no seg fault here on play [19:05] but it's more the top bar which I like [19:05] yeah [19:05] oh, you probably need daap enable for that to matter [19:05] ah [19:05] that's where the segfault is [19:06] anyway the default toolbar feels clumsy [19:06] like different height for the widgets [19:06] that one looks nice, though the artwork is a big small [19:06] yeah [19:06] although I just hit a situation where I couldnt open the plugins window without a restart [19:07] right, if you enable/disable it seems to go wrong [19:07] like probably something unrefed uncorrectly on unload [19:07] just added the vol slider in options too [19:07] nice [19:08] I wonder if we should do that by default [19:08] feels much nicer to me [19:08] and disable "enhanced plug ins dialogue" [19:09] not so keen on the "modern" option [19:09] yeah [19:10] what I like is just to have everything in the toolbar at the same height [19:10] yeah [19:10] me too [19:10] ship it [19:10] feels much better [19:10] :-) [19:10] going to have a look to that before uif [19:10] :) thx [19:10] seb128: looks like display scaling not working might be an xserver 1.18 thing, need to double check that :/ [19:11] tjaalton, urg, good that you catched it then! [19:11] caught even [19:11] that might actually also be why nvidia prime fails [19:11] if something broke in randr [19:12] does gnome have a similar scaling thingy? [19:13] gsettings should work [19:13] I'll test that too.. [19:14] you can also GDK_SCALE=2 gedit [19:14] haha, ouch [19:14] good to know :) [19:14] that seems to work with 1.18 [19:15] the unity scaling is not relying on gtk [19:15] well at least the unity custom elements [19:15] like panel, indicator, etc have some gtk [19:15] Trevinho might know better if you need specific/that's something you try to debug [19:16] ok I'll ask tomorrow [19:17] seb128, @ alternative tool bar - looks a bit odd when playing a track with a long title: http://i.imgur.com/iW9jKc4.png [19:17] Sweet5hark, libreoffice bug #958345 is not fixed, http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.xenial/rdepends/ttf-indic-fonts/ttf-indic-fonts-core [19:17] bug 958345 in ttf-indic-fonts (Ubuntu) "ttf-indic-fonts packages are outdated should be removed" [High,Triaged] https://launchpad.net/bugs/958345 [19:18] Sweet5hark, e.G libreoffice-l10n-bn recommends ttf-bengali-fonts which is a binary from ttf-indic-fonts [19:18] willcooke, right, should probably ellipsize [19:19] seb128, also, I miss the "View All" button from the normal toolbar [19:19] so if I drill down in to an artist, then I have to scroll back to the top click click on all again [19:20] willcooke, ctrl-t? [19:20] BOOM! [19:21] ta [19:22] yw! [19:23] if we do change, I'd vote for that to be on by default (if poss) [19:24] it's not? [19:24] I need to try on a fresh install [19:25] anyway, that's for another day [19:25] going to have a look in the fridge and think about dinner ;-) [19:25] cheers seb128 [19:46] seb128: k, will fix that tomorrow. [19:48] eod [20:06] urgh bluetooth device window is uuuuugly [20:13] willcooke: on the phone? [20:13] cyphermox, nah, u7 desktop [20:13] when you add a device [20:13] ah, yeah, I suppose [20:13] no padding around anything by the looks of things [20:13] willcooke: as in the wizard? [20:14] we get all that straight from gnome, in any case. [20:14] Bluetooth settings -> Click the + bottom left -> my eyes my eyes [20:15] In other news: John McAfee looks like he sleeps in his car. http://www.businessinsider.com/john-mcafee-ill-decrypt-san-bernardino-phone-for-free-2016-2?IR=T [20:16] via Trevinho [20:16] willcooke: oh, the wizard steps are messed up [20:17] willcooke: that applies to the wizard for mobile data connections in nm-connection-editor too [20:17] this is fun, is that theming or straight gtk? [20:18] oh yeah [20:18] dunno, could just be themeing [20:18] I'll ask Seb tomorrow [20:20] willcooke: hmmm so the new bar for gnome is the thing that causes the ugly corners right? I see them on the calendar app :( [20:20] davmor2, ugly corners are client side decorations issues. We have a fix in progress right now [20:21] \o/ [20:21] oh, well, actually, looking at my test machine, it might be fixed [20:23] willcooke: I just full-upgraded it isn't here on gedit/calendar but I still need to reboot for the new kernel so I assume I'm still running on old stuff [20:24] davmor2, ah yeah, you're right [20:24] I'm looking at older apps === doko_ is now known as doko [20:57] night all