[06:22] <hikiko> hi
[06:44] <pitti> Good morning
[06:47] <hikiko> hi pitti
[07:47] <om26er> Hi! is there a way to delay the startup of unity-settings-daemon ? one my system seems to be too fast to wait for themes/font to load correctly. I have to logout and login again to make things correct.
[07:47] <pitti> if you log out, u-s-d will be stopped as well, and restarted on login
[07:50] <om26er> pitti, hmm, so maybe I am looking at it the wrong way. I am actually trying to workaround bug 1369251
[07:51] <om26er> if I wait for the password screen like 2-3 seconds and then enter my password, things show up nicely. Its only when I enter my password as soon as lightdm shows that issue happens.
[07:54] <Trevinho> Morning
[08:56] <willcooke> morning all
[08:58] <willcooke> hey Trevinho, good news on the launcher at the bottom then :)
[09:02] <Laney> hi!
[09:02] <willcooke> morning laney
[09:07] <Laney> what is up
[09:11] <Laney> mhall119: I tried to say in the mail - file a bug, attach the icon, tag it appstream, subscribe ubuntu-sponsors, *send it upstream*
[09:17] <willcooke> grumble grumble.  My RasPi 3 order remains "unfulfilled"
[09:18] <davmor2> willcooke: steal popey he won't mind
[09:18] <willcooke> :)
[09:50] <seb128> good morning desktopers!
[09:50] <willcooke> hey seb128, how goes?
[09:51] <seb128> willcooke, hey, good! how are things here?
[09:51] <willcooke> seb128, so far, so good
[09:51] <seb128> I had a banck appointment but that took a bit longer than I though
[09:52] <willcooke> ain't no thing, you were still here when I left last night
[09:52] <willcooke> anything more on the PK discussions?
[09:54] <seb128> not really no
[09:54] <seb128> would be nice to get the update in
[09:54] <seb128> but we are not changing our course of action for g-s
[09:54] <willcooke> cool
[09:54] <willcooke> makes sense to me
[09:58] <Laney> hey seb128
[09:58] <Laney> how was the bank robbing?
[09:58] <Laney> oh shit this isn't our encrypted line
[10:02] <seb128> Laney, I told them to stop robbing me!
[10:02] <seb128> he needs to ask his boss what they can do
[10:02] <Laney> is this france?
[10:03] <seb128> but basically I was complaining about the credit card charges when travelling out of the eurozone
[10:03] <seb128> yes
[10:03] <seb128> and the cost of the card
[10:03] <Laney> get didrocks on your side
[10:04] <seb128> yeah, I told the bank that I would probably get an account at the same online bank if they couldn't make an effort
[10:04] <seb128> those bank give you a visa premium for free and charge half of what the normal banks do
[10:04] <seb128> that was the banking news of the day ;-)
[10:05] <Laney> you badass
[10:06] <seb128> looks like robert_ancell had another busy g-s day
[10:07] <seb128> no love for my gnome-calendar ffe :-/
[10:09] <davmor2> seb128: oh is it a bug if so I bet I can make it more popular with one comment
[10:11] <seb128> thanks but don't worry, I'm sure it's going to be reviewed ;-)
[10:22] <Trevinho> willcooke:  yeah great news
[10:26] <seb128> hey Trevinho!
[10:33] <popey> davmor2: sadly my rpi3 isn't currently being used because we don't have a snappy image for it yet.
[10:33]  * popey glares at ogra_ 
[10:33] <ogra_> blame DHL
[10:33] <ogra_> :P
[10:34] <ogra_> i ordered two yesterday, should be here before the weekend i guess so i can take a look then
[10:34] <ogra_> if the changes arent to massive i should have something by monday
[10:35] <davmor2> ogra_: pfff poor excuse ;)
[10:36] <ogra_> haha
[11:42] <seb128> Laney, where is appstream getting it packages descrptions translations from?
[11:42] <seb128> is the name/subtitle the info from the .desktop?
[11:46] <seb128> Laney, thanks for the calendar ffe
[11:48] <Laney> yeah think so
[11:49] <Laney> or the appdata
[11:50] <seb128> Laney, is it likely that it doesn't play well with our build tools stripping the inlines translations from .desktop to use gettext?
[11:50] <seb128> I'm trying to figure out why many important pacakges don't have their title translated
[11:50] <seb128> eog to take an example
[11:50] <seb128> they show translated in the dash and in software-center
[11:50] <Laney> probably
[11:51] <seb128> :-/
[11:51] <Laney> That's an X- keyword in desktop files isn't it
[11:51] <seb128> X-Ubuntu-Gettext-Domain=eog
[11:51] <seb128> yeah
[11:51] <Laney> So you have to teach tools about it
[11:52] <seb128> what $tools would be?
[11:52] <seb128> I've no idea about that stack
[11:52] <seb128> are the datas collected on package build time?
[11:52] <Sweet5hark1> desrt: https://www.google.com/trends/explore#q=%22move%20to%20canada%22&geo=US&date=today%2012-m&cmpt=q&tz=Etc%2FGMT-1 <- dunno why that happens there in the us, but "blame canada" seems to be replaced by "move to canada" in recent weeks for some reason ...
[11:52] <seb128> could we get it from the .desktop before it's stripped down?
[11:52] <seb128> Sweet5hark1, lol
[11:53] <seb128> Sweet5hark1, Canada is the new U.S ;-)
[11:54] <Laney> https://github.com/ximion/appstream-dep11/blob/master/dep11/parsers.py#L78
[11:54] <Laney> so you probably have to load up the langpacks and then try to translate the string using the domain given
[11:59] <seb128> thanks
[12:00] <seb128> I'm going to start by filling a bug I guess
[12:03] <desrt> moin
[12:05] <willcooke> hi desrt
[12:05] <Sweet5hark1> desrt: moin
[12:06] <desrt> good morning, willcooke, Sweet5hark1
[12:15] <willcooke> andyrock, are these two fixed now with the changes you made?
[12:15] <willcooke> https://bugs.launchpad.net/ubuntu/+source/hud/+bug/1368896
[12:15] <willcooke> https://bugs.launchpad.net/ubuntu/+source/hud/+bug/1354134
[12:15] <willcooke> fixed/otherwise resolved
[12:16] <Sweet5hark1> seb128: how do I find out which commit of the xenial seed a daily image was build upon?
[12:23] <seb128> hey desrt
[12:23] <seb128> Sweet5hark1, no idea, ask on -devel maybe?
[12:23] <Laney> d€$rt
[12:23] <Laney> Sweet5hark1: look at the manifest
[12:23] <Laney> then you can see which version of ubuntu-desktop it contains
[12:25] <desrt> morning, seb
[12:25] <desrt> and £4n3y
[12:26] <andyrock> "morning"
[12:26] <andyrock> willcooke: i don't think so
[12:32] <Sweet5hark1> Laney: thanks.
[12:35] <Laney> Sweet5hark1: manifest is on cdimage next to the iso fwiw
[12:38] <Sweet5hark1> looking at the manifest, I find ubuntu-desktop 1.348 there, uploaded by robert_ancell on 2016-02-17, so it cant have my changes to the xenial seed at rev. 2418 on 2016-02-22. Seeing that ubuntu-meta 1.348 was released on -02-17 and before that 1.347 was on -02-01, Im having hope this will soon be updated.
[12:42] <Laney> Sweet5hark1: it would have been usual for the person pushing to also do the meta
[12:42] <Laney> I wouldn't hope, because you could wait an arbitrary length of time :)
[12:42]  * Laney will do that for you
[12:42] <Sweet5hark1> Laney: oh, thanks a lot!
[12:43] <Laney> it's mostly just running ./update
[12:45] <Laney> uuuuuuuuuuuuuuuuuh it didn't get the new thing
[12:46] <Laney> probably needs to be promoted
[12:49] <Laney> seb128: want to promote libreoffice-style-breeze to make Sweet5hark1 happy?
[12:50] <seb128> Laney, Sweet5hark1, done
[12:50] <Sweet5hark1> seb128: thx
[12:50] <seb128> yw!
[12:50] <Laney> ♥
[13:36] <willcooke> Laney, I've got the greyscales now too: https://bugs.launchpad.net/ubuntu/+source/ubuntu-wallpapers/+bug/1551208
[13:36] <willcooke> Assuming it's you that uploads them, do you want the greyscales as separate attachments or the single zip file I got from design?
[14:10] <Laney> willcooke: umm zip file is a bit less work to download but doesn't really matter
[14:10] <Laney> is there a wallpaper contest?
[14:16] <willcooke> Laney, there is, still on going.  nhaines is running it
[14:17] <Laney> ok
[14:18] <willcooke> Laney, uploaded the greyscale
[14:18] <Laney> not sure about adding the huge ones
[14:27] <Laney> debdiff --auto-ver-sort ♥
[14:50] <mhall119> thanks Laney
[15:27]  * desrt steps out to a doctor's appt. back in a bit.
[15:57] <mhall119> Laney: is there a machine-parsable version of the data on http://appstream.ubuntu.com/xenial/universe/issues/index.html ?
[15:58] <mhall119> I'd like to make a list of apps needing icon work
[15:59] <Laney> mhall119: http://appstream.ubuntu.com/hints/
[16:00] <Laney> should be what you want
[16:03] <mhall119> Laney: indeed it is, thanks :)
[16:39] <seb128> mvo, so, need any help with the snappy MIRs?
[16:51] <flexiondotorg> Trevinho, I've updated https://code.launchpad.net/~ubuntu-mate-dev/compiz/marco-gsettings/+merge/282882
[17:04] <mvo> seb128: https://bugs.launchpad.net/ubuntu/+source/golang-websocket/+bug/1548887 I just need to go over it point by point I think
[17:17] <Sweet5hark1> seb128: wrt -breeze -- how is it that 1:5.1.1~rc2-0ubuntu1 says its in universe, but 1:5.1.0-0ubuntu1 says its in main?
[17:20] <Trevinho> flexiondotorg: thanks, I'll check thata
[17:21] <flexiondotorg> ty
[17:31] <Trevinho> flexiondotorg: alberts' comment seems reasonable, can you fix that too?
[17:32] <flexiondotorg> Oh, erm.
[17:32] <flexiondotorg> I have no idea what it means.
[17:33] <muktupavels> flexiondotorg, add gboolean is_mate_desktop in private struct just like you did with GSettings *marco;
[17:33] <muktupavels> then you will just use 'priv->is_mate_desktop' not 'priv->marco && is_mate_desktop ()'
[17:34] <flexiondotorg> Sorry, just seen there is a new comment :-)
[17:35] <seb128> Sweet5hark1, hum, because the change-override command act on a pocket and I did it for xenial and forgot there was another version stucked in proposed, I'm going to fix that one
[17:35] <seb128> mvo, ok, let me know if I can help you with something
[17:37] <flexiondotorg> muktupavels, Any chance you could help me out here.
[17:37] <flexiondotorg> muktupavels, You've requested something beyond my XP ;-)
[17:38] <muktupavels> flexiondotorg, what help do you want from me?
[17:38] <muktupavels> Trevinho, can you look at my new merge proposals?
[17:39] <Trevinho> muktupavels: yeah, I was monitoring it already... I'll check that in bait
[17:39] <Trevinho> bit*
[17:40] <flexiondotorg> muktupavels, Would you be able to refactor my merge proposal to meet your requirements?
[17:41] <muktupavels> flexiondotorg, change should be simple... 1) if (priv->marco && is_mate_desktop ()) will change to if (priv->is_mate_desktop). 2) You add gboolean is_mate_desktop to private struct. 3) you initialize priv->is_mate_desktop in gwd_settings_storage_gsettings_new.
[17:46] <muktupavels> flexiondotorg, I could create branch, but then it will be untested...
[17:46] <Sweet5hark1> seb128: ah ok, thanks! I was unsure if that was just something taking time to update or something.
[17:46]  * Sweet5hark1 feels guilty for not checking components earlier.
[17:47] <flexiondotorg> muktupavels, I can build and test. But the modification you're requesting, which might be simple for you, are complicated for me.
[17:47] <seb128> Sweet5hark1, I fixed it, should be good after next publisher
[17:50] <Laney> man
[17:50] <Laney> ximion: the 'keywords' desktop file parser block hurts my head :-)
[17:54] <ximion> Laney: I wash my hands in innocence ^^ - I only touched the .desktop file parser once or twice so far
[17:54] <Laney> heh
[17:55] <ximion> I am working on a rewrite of the generator right now, which will also work for more distributions - unfortunately, until it gets to the same level as dep11-generator, we are stuck with Python slowness
[17:55] <muktupavels> flexiondotorg, check gwd-marco-gsettings
[17:55] <ximion> I can already read more .deb files much faster, and parsing the TagFile for suites takes just a few seconds
[17:55] <Laney> nice
[17:55] <Laney> what are you using?
[17:57] <flexiondotorg> muktupavels, Thanks. I'll test that tomorrow and get back you :-)
[17:57] <flexiondotorg> Thanks for helping. I appreciate it.
[18:03] <ximion> Laney: right now, D - I felt adventurous ;-)
[18:03] <Laney> O_O
[18:03] <ximion> I also did a small Go implementation, but D is much better for that particular task
[18:04] <ximion> I am using libappstream's parser internally, and D interfaces well with C and - to some extend - C++ code
[18:05] <ximion> goal is to move all parsing and metadata-generating into libas, and have the generator do the dirty work, which is finding icons, rendering fonts, downloading screenshots, ...
[18:05] <ximion> that way, I only need to implement a feature once
[18:06]  * Laney goes mad
[18:06] <ximion> Laney: I suppose you don't like C? :P
[18:07] <Laney> no, C is great, I'm going mad at this python :)
[18:07] <ximion> Laney: ah, that's a relief ^^
[18:07] <Laney> especially new style glib code
[18:08] <ximion> yeah, I got pretty annoyed by Python's multiprocessing - that is a constant source of incredibly hard to debug issues in the generator, and I really want to get rid of it... Both D and Go have excellent ways of handling paralleluzation
[18:08] <ximion> *parallelization
[18:09]  * Laney whispers Haskell
[18:11] <ximion> Laney: that's something I didn't look at... I might still be a bit traumatized from doing Haskell transitions...
[18:13] <Trevinho> muktupavels: any reason why your MP is still set as "in hold"?
[18:13] <muktupavels> Trevinho, which?
[18:13] <Trevinho> muktupavels: https://code.launchpad.net/~albertsmuktupavels/compiz/gwd-marco-gsettings/+merge/287820
[18:14] <muktupavels> I pushed it as example, it is untested...
[18:15] <ximion> Laney: oh, btw, when you're bored: can you sync Limba (pkg:limba) from Debian unstable to universe? That package FTBFSes due to a change in AppStream...
[18:15]  * ximion can also file an FFe / sync request, if that's easier
[18:16] <Laney> ximion: probably tomorrow
[18:17]  * Laney is trying to finish this feature, taking longer than expected :)
[18:19] <ximion> Laney: which feature?
[18:19] <ximion> oh, and I should merge your pull-request, looked fine when I looked at it
[18:20] <Trevinho> flexiondotorg: so, once your MP is ready again ping me... And muktupavels too (a review of you too, would be nice).
[18:20] <ximion> I also want to add a status command to appstreamcli, so we can get an overview of the appstream data, and in which shape it is
[18:24] <Laney> ximion: do you want to know or do you want it to be a surprise? :)
[18:24] <Laney> aaaaaaaaaaaaaaaaaaaah shit, pickling error
[18:27] <ximion> classic
[18:27] <ximion> maybe we should really use dill...
[18:27] <Laney> I'm adding support for langpack translations
[18:29] <seb128> Laney, thanks ;-)
[18:29] <Laney> you're not welcome :|
[18:29] <Laney> :)
[18:29] <seb128> !
[18:30] <mvo> seb128: feel free to pick any of the tasks in this bug :)
[18:31] <seb128> mvo, going to try to do my best to help ;-)
[18:32] <ximion> Laney: as in "read .mo files and determine their translation status" or as in "add metainfo for language-packages"?
[18:32] <ximion> both is cool, btw :)
[18:33] <ximion> (although the latter isn't even specified yet, I need to come to an agreement with hughsie)
[18:33] <seb128> mvo, I guess the team to subscribe is one I'm not admin for that can't be me doing it
[18:33] <Laney> ximion: reading mo, yeah
[18:34] <ximion> neat - that's the larger task and more work ^^
[18:34] <mhall119> Laney: what does 'Ignored': True mean in the AppStream hints yaml file?
[18:34] <ximion> mhall119: means we dropped the component because it has fatal flaws
[18:35] <ximion> (like violations in against the spec, e.g. a .desktop app without icon)
[18:35] <seb128> ximion, Laney, what's the easiest way to determine why a specific package is not listed? e.g emacs, "/usr/share/applications/emacs24.desktop 	emacs24" ... should that be enough to have it?
[18:35] <mhall119> ximion: ok, so if I'm trying ot get people to help fix the spec, those should be targetted?
[18:36] <ximion> seb128: go to the issues page of the respective suite/component and search for the package name - if it has hints of type error, it won't be included: http://appstream.ubuntu.com/xenial/universe/issues/index.html
[18:36] <ximion> I find http://appstream.ubuntu.com/xenial/universe/issues/emacs24-lucid.html
[18:37] <Laney> it's probably in main
[18:37] <ximion> mhall119: jup - ideally point them to the HTMl output, that is easier to read and contains some detailed explanations on what went wrong, while the hints YAML file is rather compact and designed for machine-reading
[18:37] <ximion> jup: http://appstream.ubuntu.com/xenial/main/issues/emacs24.html
[18:37] <ximion> no icon, as usual
[18:38] <seb128>       cid: emacs24.desktop
[18:38] <seb128>     tag: gui-app-without-icon
[18:38] <seb128>       icon_fname: /usr/share/icons/hicolor/scalable/apps/emacs24.svg
[18:38] <seb128>     tag: icon-not-found
[18:38] <seb128> ximion, thanks
[18:40] <seb128> the icon is in the -common
[18:40] <seb128> I guess that's a standard/frequent issue
[18:40] <seb128> is there any way we could workaround it?
[18:42] <seb128> I don't know how the parser works, in what env it tries to look for the icon, but could we make whatever pulls the package try to pull "$name-common" as well?
[18:54] <mhall119> ximion: I'm planning on making a separate listing of just package with missing or incorrect icons
[18:56] <ximion> seb128: no, that's not feasible - although it should find the icon in -common automatically, if it's listed in Contents.gz
[18:56] <seb128> ximion, http://packages.ubuntu.com/search?searchon=contents&keywords=emacs24.svg&mode=exactfilename&suite=xenial&arch=any
[18:56] <seb128> unsure why that doesn't work then :-/
[18:58] <mhall119> can AppStream use SVG, or do they have to be PNG?
[18:58] <ximion> smalls like a Contents issue...
[18:59] <ximion> mhall119: can be SVG, those icons will be automatically rendered to pixmaps on the server
[18:59] <ximion> ideally, the icons should be - if they are pixmaps - available in at least 64x64px
[19:00] <ximion> usr/share/icons/hicolor/scalable/apps/emacs24.svg	    editors/emacs24-common
[19:00] <ximion> Laney: ^ - could this be a remnant from when the generator was parsing the Contents file improperly?
[19:00] <ximion> maybe trigger emacs24 for reprocessing...
[19:01] <muktupavels> Trevinho, are you ok with removing tests from gwd? I created merge proposal for it.
[19:10] <ximion> mhall119: for automatic processing, the YAML is perfect, of course ^^ - for an - also machine-readable - description of the tags, you can look here: https://github.com/ximion/appstream-dep11/blob/master/data/dep11-hints.yml
[19:32] <Laney> ximion: worth a try
[19:32]  * Laney has almost won
[19:32] <Laney> going to send this PR in a second, will do fixes tomorrow
[19:33] <ximion> Laney: the D code is insanely fast! I am really excited! and: real threads and no picking ;-)
[19:33] <ximion> the .deb package format is a bit bad for performance - I have to basically exract it to get a file-list, which is and always will be slow
[19:34] <ximion> I could extract it into memory, but for large packages this might be an issue, unless you have loads of ram
[19:36] <ximion> Laney: thanks for working on this - I postponed this task for a long time ^^
[19:36] <Laney> it's probably not exactly what you have in mind
[19:36] <ximion> (it had a high boringness factor, and wasn't visible in any UI for a very long time anyway)
[19:36] <Laney> this is an ubuntu specific thing
[19:38] <ximion> ah, eww, that's trickier then
[19:39] <Laney> what is your thing?
[19:39] <ximion> does the validator validate the result?
[19:39] <Laney> if it validates translations
[19:39] <Laney> it just gets them from mo files as well as inline in the desktop file
[19:40] <mhall119> ximion: dpkg --contents extracts the whole thing?
[19:42] <ximion> mhall119: need to look at the internals, but yes, it needs to open up the ar file and get to the actual tarball and then read through the whole tarball contents
[19:42] <ximion> it won't actually extract the data from the tarball though
[19:42] <ximion> (fortunately)
[19:42] <ximion> Laney: ah, that's uncontroversial and a neat feature :)
[19:43] <Laney> except for the bit where I have to extract all language-pack-* and deal with X-Ubuntu-Gettext-Domain key
[19:43] <Laney> bit annoying
[19:44]  * Laney probably finds the last known bug
[19:44] <ximion> can't be uglier than what is already in there
[19:57] <Laney> ok there we go
[20:02] <Laney> night!
[20:10] <mhall119> Laney: is there any kind of priority for AppStream apps? There are a lot of them missing icons and it's generally better to present a smaller "wishlist" for people to choose from
[20:17] <willcooke> davmor2, can you see if this has regressed for you as well:  https://bugs.launchpad.net/ubuntu/+source/usb-modeswitch/+bug/1498805
[20:17] <attente> robert_ancell: any idea if there's a better way to get the deb info besides spawning dpkg-deb? (for the sideloading)
[20:18] <willcooke> davidcalle, specifically my ZTE modem is not recognised again
[20:19] <ximion> mhall119: priorize by "stuff in main", then go by the popcon values Ubuntu has collected. The higher the popcon index of an app, the more likely it is that people will miss it
[20:19] <ximion> (that's what I'd do ;-) )
[20:20] <robert_ancell> attente, not off the top of my head, but the dpkg/apt libraries are a bit crap when I've tried to find them
[20:20] <ximion> attente: use apt_inst's DebFile and TagFile classes
[20:21] <attente> ximion: that's a python module? is there a c library i can interface with?
[20:22] <mhall119> ximion: do we have popcon values? IIRC, that service was pretty hit or miss (usually miss)
[20:22] <ximion> attente: it's a C++ library
[20:22] <robert_ancell> ximion, yeah, they all seem to be C++ :(
[20:22] <ximion> mhall119: is http://popcon.ubuntu.com/ still alive?
[20:23] <ximion> robert_ancell: shouldn't be a problem though, you can mix C++ and C easily
[20:23] <ximion> that's what we do in PackageKit and also AppStream
[20:24] <ximion> https://anonscm.debian.org/cgit/apt/apt.git/tree/apt-inst/deb/debfile.h is the thing
[20:24] <ximion> if you read the contents via TagFile it's pretty fast
[20:26] <attente> ximion: mixing c++ and c in GS?
[20:26] <ximion> attente: well, it will work...
[20:26] <ximion> I don't think it's the best idea ever, but it will work
[20:27] <mhall119> ximion: ah, it is alive (and well,surprisingly)
[20:27] <attente> between that and parsing the output of dpkg-deb, which is preferable? dpkg-deb does have a show-format option that allows us to specify exactly what fields we want
[20:28] <ximion> attente: I think a sane way would be calling "dpkg --field" and parsing the output
[20:28] <ximion> (if you want multiple fields)
[20:29] <attente> ximion: does that require multiple spawns though?
[20:29]  * attente should just try it
[20:32] <attente> dpkg-deb --showformat is a bit better because --field also includes the field name in the output
[20:32] <attente> i'm ok with doing it this way since GS is just a front-end for dpkg in this case
[20:35] <ximion> attente: as it happens, I need almost the same solution right now on the genertaor side
[20:35] <ximion> a benchmark has shown that using the dpkg cli is fast enough, if it's just one call.
[20:36] <attente> ximion, yeah, i agree. dpkg-deb doesn't seem too bad for this
[20:39] <ximion> for your case, the overhead of spawning a process should be neglectable
[20:40] <ximion> for my case, I am not yet sure, since I do that 1000 times
[21:48] <willcooke> gnight all
[22:59] <robert_ancell> attente, can you move those non-review changes into another branch?
[22:59] <robert_ancell> branche(s)