[05:50] <hikiko> Hi
[06:09] <pitti> Good morning
[06:11] <Fudge> morning/evening
[08:19] <flocculant> gnome-software looking better this morning - it sees the repos - I've installed from it
[08:19] <flocculant> missing featured and editors picks - but better them than the repos :)
[08:20] <flocculant> it did prompt a new bug though
[08:48] <pitti> Laney: I now fixed one race condition in the udisks2 tests, and understand the eternal hang
[08:48] <pitti> so this should hopefully work better now
[08:48]  * pitti adds a workaround for the hang too
[08:50] <seb128> good morning desktopers
[08:50] <pitti> bonjour seb128, ça va ?
[08:51] <seb128> salut pitti! oui, bien, et toi ?
[08:51] <pitti> seb128: je vais d'accord; je n'ai pas assez dormi, Annett est malade :/
[08:52] <pitti> alors, c'est l'heure de corriger les bugs d'udisks :)
[08:52] <seb128> oh :-( j'espère que ça va aller mieux aujourd'hui pour elle
[08:52] <pitti> merci !
[09:04] <Laney> hi!
[09:04] <willcooke> morning Laney
[09:05] <seb128> oh, Laney is back!
[09:05] <seb128> hey Laney
[09:05] <seb128> hey willcooke
[09:05] <willcooke> morning seb128
[09:05] <seb128> Laney, had a good double w.e? ;-)
[09:06] <hikiko> hey desktopers!
[09:06] <seb128> hey hikiko
[09:06] <willcooke> hi hikiko
[09:09] <Laney> hi willcooke seb128 et hikiko
[09:09] <Laney> was good thanks!
[09:09] <Laney> 1) tea 2) coffee 3) cake
[09:09] <Laney> 4) falafel
[09:09] <Laney> and no rain, which waited until today
[09:10] <Laney> I just got to see the hidpi greeter for the first time because my laptop didn't resume ;-)
[09:10] <Laney> oh AND i left my suit on the train /o\
[09:10]  * Laney calls edinburgh waverley now
[09:11] <seb128> :-(
[09:11] <seb128> good luck
[09:11] <seb128> Laney, how does the hidpi greeter looks? works fine?
[09:12] <Laney> it started in 1x and then shifted after about 1s
[09:12] <seb128> I guess the greeter is ok but the transitions are annoying/not nice?
[09:12] <seb128> :-/
[09:12] <Laney> but looked normal after that
[09:12] <seb128> what about after entering the password, did it get back to 1x or a grid of 2x2 images?
[09:12] <Laney> oh it got messed up
[09:12] <pitti> hey Laney!
[09:13] <Laney> but it wasn't like 4 copies of the background
[09:13] <Laney> there was some corruption
[09:13] <pitti> Laney: argh, hope you get your suit back, would be a pity
[09:13] <seb128> k, I get different variants
[09:13] <seb128> unsure what to change to fix that though :-/
[09:13] <Laney> hey pitti!
[09:13] <Laney> indeed
[09:14] <Laney> I phoned last night but the person on the other end said they couldn't get someone to go look /o\
[09:15] <Laney> station's website says
[09:15] <Laney> Onboard a train
[09:15] <Laney> Items left on trains are kept by the train company
[09:15] <Laney> train company says to phone the station
[09:15]  * Laney screams
[09:15] <Fudge> good morning Laney , who do we need to knee-cap!
[09:18] <Laney> hi Fudge
[09:18] <Laney> start with the department for transport and work down please :-)
[09:19] <Laney> pitti: good work on udisks2!
[09:20] <Sweet5hark1> Laney: this is why you brits consider leaving the EU: otherwise you would need to start yet another level higher in Bruessels.
[09:21] <Laney> Sweet5hark1: the galactic high command has received your request and will respond within 1bn working years
[09:23]  * seb128 pokes robert_ancell with a stick
[09:23] <Laney> umm
[09:23] <Laney> https://code.launchpad.net/~geonames-dev/geonames/trunk
[09:23] <seb128> good work getting gnome-software listing non-installed things
[09:23] <Laney> is something going on here
[09:23] <seb128> he converted it to bzr for some reason it seems
[09:24] <seb128> hit git mp had a comment "I'm closing this in favor of the bzr-based https://code.launchpad.net/~mterry/geonames/double-free/+merge/287359"
[09:24] <seb128> robert_ancell, if you ever read irc logs, your appstream(-glib) update losts the editor picks or whatever the banner and features softwares were called
[09:25] <seb128> Laney, I expect it's because of https://code.launchpad.net/~mterry/geonames/trainify/+merge/287362 / because CI doesn't handle git workflows
[09:26] <Fudge> Laney:  do you want to borrow my tassie devil?
[09:28] <Laney> seb128: what do you think about that?
[09:28] <Laney> anyway, I heard desrt was maintaining that now
[09:28]  * Laney runs
[09:28] <seb128> Laney, changing to bzr?
[09:28] <seb128> ideally we would have support for git in the CI tools
[09:29] <Laney> or training it at all
[09:29] <Laney> like it is going to be a dependency of libtimezonemap which is a real upstream component
[09:29] <seb128> but atm all our projects and workflows are based on bzr so I'm fine being pragmatic and converting that one if it's making work easier for whoever is maintaining it
[09:29] <Laney> sooooooooooooooooooo
[09:29] <Laney> having that in the train is a bit nasty
[09:29] <seb128> up to whoever takes over maintaining it I guess
[09:29] <seb128> maybe mterry wants to do it? ;-)
[09:30] <seb128> oh, and I didn't know libtimezonemap was going to use it
[09:30] <seb128> nice :-)
[09:31] <seb128> let's wait for Michael to be online, better to have the discussion with him
[09:31] <Fudge> lol
[09:32] <Laney> "going" in a loose sense
[09:37] <seb128> btw did we document somewhere why we needed a new lib? mterry was curious about it and I don't think I made a good case about why we couldn't just fix libtimezonemap (out of the depends on gtk part)
[09:40] <Laney> it's a feature that is wanted in a few places and the interface was very shit on the old one
[09:40] <Laney> and who cares if it's in one source package or another?
[09:41] <Laney> (I guess)
[09:42] <seb128> yeah, I was just trying to convince the touch team that they should use it
[09:42] <seb128> but I didn't feel like I had arguments to make a good case about it/how it's better
[09:43] <seb128> how well, looks like the "doesn't depends on GTK" and "trust us" was enough for mterry since he ported his changes to geonames
[09:44] <Laney> looks like he already change the project to use bzr
[09:44] <Laney> guess that decides who maintains it then
[09:44] <ricotz> hello desktopers :)
[09:44]  * Laney goes to look at fontconfig instead
[09:44] <Laney> hi ricotz
[09:44] <seb128> hey ricotz
[09:44] <ricotz> Sweet5hark1, hi, just noticed 5.1.1~rc3 :\ (after doing the backports)
[09:44] <seb128> maybe he talked to larsu
[09:45] <ricotz> Laney, seb128, hey
[09:45] <seb128> I don't see mterry doing those changes in an unilateral way without asking us first
[09:46] <ricotz> seb128, jfyi unfortunately graphite2 will require another merge
[09:47] <seb128> ricotz, why?
[09:47] <ricotz> seb128, just another upstream release
[09:47] <seb128> oh, that's not "required" then
[09:48] <seb128> or is there anything we really need in the new version?
[09:48] <ricotz> seb128, it is required imo, given it has been picked up for lo 5.1.1 https://cgit.freedesktop.org/libreoffice/core/log/?h=libreoffice-5-1-1
[09:51] <davidcalle> didrocks: ping, got time for a call now?
[09:56] <didrocks> davidcalle: yep!
[09:56] <didrocks> davidcalle: mind starting the HO?
[09:57] <larsu> seb128: he didn't :)
[09:57] <larsu> seb128: and bonjour!
[09:58] <davidcalle> didrocks: https://plus.google.com/hangouts/_/canonical.com/david-didier
[10:02] <seb128> hey larsu
[10:02] <seb128> wie gehts?
[10:05] <seb128> Laney, do you work on gnome-calendar atm/plan to work on it a bit more this cycle?
[10:06] <larsu> seb128: good thanks! Had relaxing 2 weeks off :) How are you?
[10:06] <Laney> seb128: not explicitly planned, but if there is stuff on rls-x-tracking then I would probably look
[10:06] <seb128> Laney, I was pondering doing a round of 3.19 bugfixes backports, just didn't want to step on your feet (and also that would put it out of sync at least for this cycle where you put efforts to have it in sync)
[10:06] <Laney> ...but if it's bugs about eds
[10:07] <Laney> man...
[10:07] <seb128> larsu, where did you go?
[10:07] <larsu> seb128: one week at home and one week of hiking in harz mountains
[10:07] <Laney> seb128: put it in experimental :-)
[10:07] <Laney> GTK_MIN_VERSION=3.15.4
[10:07] <Laney> ECAL_REQUIRED=3.13.90
[10:07] <Laney> EDATASERVER_REQUIRED=3.17.1
[10:07] <Laney> EDATASERVERUI_REQUIRED=3.17.1
[10:08] <seb128> Laney, they forgot to bump the gtk requirement, I tried to build trunk
[10:08] <seb128> it has new css code
[10:08] <Laney> god!
[10:08] <seb128> and the commits are not easy to revert
[10:08] <seb128> I was looking at doing 3.20 - gtk css changes
[10:08] <Laney> I'm used to losing on syncing anyway
[10:08] <seb128> k, sorry about that
[10:08] <Laney> or you could cherry-pick into Debian
[10:09] <seb128> right
[10:09] <Laney> assuming that they aren't ubuntu specific patches
[10:09] <seb128> well, atm I'm looking at bugfixes
[10:09] <seb128> but we might end up doing a "normal decoration on unity" since everything else in the default install (or almost) has those
[10:09] <seb128> but let's not discuss that
[10:09] <seb128> going to look at some cpu spin and segfaults issues first
[10:10] <Laney> haha yes let's not
[10:10] <seb128> :-)
[10:11] <seb128> Laney, thanks for looking at fontconfig btw!
[10:11] <Laney> that bug is easy to reproduce
[10:11] <seb128> I'm always unsure about the unpack/configure order
[10:11] <seb128> right
[10:11] <seb128> I was just unsure if the shlibs would be enough
[10:11] <seb128> or if pre-depends are required there
[10:11] <Laney> apt-get install fontconfig; sed -i 's/trusty/xenial/' /etc/apt/sources.list; apt-get update; apt-get install fontconfig
[10:12] <seb128> and you can't test that with dpkg
[10:12] <seb128> though I guess throwing to a ppa would be easy enough
[10:12] <Laney> laney@nightingale> dpkg-scanpackages . > Packages                                                                                                                                   ~/temp/fontconfig
[10:12] <Laney> dpkg-scanpackages: info: Wrote 5 entries to output Packages file.
[10:12]  * Laney hax0r
[10:13] <seb128> Laney, El3te
[10:15] <Laney> yeah this works
[10:28] <Laney> they have the suit! \o/
[10:28] <Laney> but they told me to come to edinburgh to collect it /o\
[10:28] <seb128> urg
[10:29] <Laney> https://goo.gl/maps/x5J1D3W5nv12
[10:29] <seb128> w.e in edingurgh it is then? ;-)
[10:29] <seb128> Scotland is nice
[10:29] <Laney> didn't know you could fly there from here :-o
[10:29] <seb128> I recommend it!
[10:33] <davmor2> jibel, willcooke: https://docs.google.com/spreadsheets/d/1lY70PWj0gTVRj_qvrGhVpNpqnXGafEvxXpsiQ-Lk0eY/edit#gid=0 doc link for the desktop testing, I will mostly be doing this inside a vm on a fresh install so no conflicts in configs, there may well be duped bugs in the list as it goes but I'll try to curb them.  LET THE GAMES BEGIN :D
[10:34]  * Laney thought that would be more exciting than it was
[10:34] <willcooke> desktoppers, please be aware and offer davmor2 whatever help he needs ^
[10:34] <willcooke> technical help that is
[10:34] <willcooke> it's too late for anything else
[10:35] <willcooke> davmor2, are the installer issues mostly fixed now?
[10:36] <seb128> no
[10:36] <seb128> I don't think lot of people looked at upgrades issues
[10:37] <davmor2> willcooke: Installer is mostly fixed I need to test a free software only install at some point to confirm they all are, updates from 14.04.4 to 16.04 are horribly broken
[10:37] <seb128> like that fontconfig bug is being worked on but the fix is not uploaded yet
[10:37] <davmor2> seb128: oh there is a new one now
[10:37] <davmor2> seb128: you don't even get that far now
[10:37] <seb128> "great"
[10:37] <seb128> but that was my point
[10:38] <seb128> the upgrades are known to be buggy atm
[10:38] <davmor2> seb128: yeap upgrades are very broken right now
[10:38] <seb128> sorry, I read "installer" as "upgrader"
[10:38] <seb128> I don't know of the ubiquity install state in 16.04
[10:38] <seb128> that's probably better than upgrades
[10:44] <liuxg> I am now running 15.04 desktop. May I know how I can upgrade 15.10? thanks
[10:46] <willcooke> liuxg, best to ask in #ubuntu
[10:47] <liuxg> willcooke, thanks :) This is xiaoguo, I met you in MWC
[10:47] <willcooke> liuxg, Hi! Have your feet recovered yet? ;)
[10:48] <liuxg> willcooke, yeah, it is fine for me now. Yesterday, I even played badminton. Yeah, it was really tiring for me. How are you now?
[10:48] <willcooke> liuxg, still tired :)
[10:48]  * willcooke <-- old man 
[10:49] <liuxg> willcooke, please take some rest ;) hopefully, you will fully get recovered soon ;)
[10:49] <willcooke> :)
[10:51] <davmor2> seb128: for beta 1 ubiquity was horribly broken, thankfully cyphermox put pay to the bulk of those issues.
[10:51] <seb128> good
[10:51] <davmor2> seb128: gnome software, should that be handling updates?
[10:54] <seb128> define updates
[10:54] <seb128> and define should
[10:55] <seb128> ideally it could(should?) replace update-manager but that's not on the todolist for this cycle
[10:55] <seb128> it's not supposed to do that in its current version in any case
[10:55] <seb128> it does handle update of desktop applications though
[10:55] <seb128> like if there is a new gedit it's going to list it in the updates section
[10:56] <seb128> it's not going to list non desktop items though
[10:56] <seb128> does that reply to your question?
[11:00] <davmor2> seb128: I opened gnome software and got a dialogue similar to update-manager but moved to the Updates page on gnome software. Where I have individual install buttons
[11:00] <seb128> davmor2, do you have a screenshot of the dialog?
[11:01] <davmor2> seb128: oh and the restart & install button which is a little disturbing :) Why would you restart before you install :D
[11:01] <seb128> because that's what e.g windows is doing
[11:01] <seb128> well, except different implementation
[11:02] <davmor2> seb128: umm no this is a button by the look of it
[11:02] <seb128> you do updates out of the session, reboot the system, apply updates, then resume boot
[11:02] <seb128> davmor2, bug #1546634
[11:02] <davmor2> seb128: oh okay
[11:03] <seb128> we are likely going to hide it this cycle, the out of session updates are not tested yet
[11:03] <seb128> it's a bit too much of a change just before a LTS
[11:04] <davmor2> seb128: fair enough, I will grab a screenshot for you one second
[11:05] <mitya57> didrocks, hi, will you mind if I add NotShowIn=GNOME-Flashback to session-shortcuts package?
[11:11] <didrocks> mitya57: hey, no, sounds fine :)
[11:11] <mitya57> Thanks
[11:12] <mitya57> didrocks, there is also bug 1544459 but I won't fix it :)
[11:12] <didrocks> mitya57: argh, yeah for generic names :p
[11:13] <didrocks> the package isn't even on the distro, should be some ppa thingy
[11:14] <mitya57> Right, I didn't notice it. Then we shouldn't care
[11:17] <didrocks> yeah ;)
[11:26] <andyrock> morning
[11:32] <seb128> andyrock, "morning" ;-)
[11:32] <seb128> how are you?
[11:33] <Sweet5hark1> ricotz: yeah -- I guess I even triggered that by pushing about the unfortunate state of l10n. But beyond than rc3 only has tdf#97739 (which is windows and -gtk3 only), so no biggy AFAICS.
[11:40] <ricotz> Sweet5hark1, I see, you mean indic fonts renames?
[11:46] <Sweet5hark1> ricotz: heh, no. the tiny detail that there is no 5.1 branch on https://translations.documentfoundation.org/
[11:51] <desrt> meow
[11:52]  * Laney puts some Deluxe Tuna Bitz out for desrt 
[11:52] <desrt> hells ya
[11:53] <desrt> my 1+x came yesterday
[11:53] <desrt> so i have to start from scratch
[11:55] <Laney> fun fun
[11:55] <desrt> no deluxe tuna bitz for me for a while :p
[11:56] <larsu> 1+x?
[11:57] <desrt> oneplus x
[11:57] <larsu> ah. enjoy!
[11:57] <desrt> got it mostly because it's dual sim and my old phone supported only edge (and sometimes only gprs) frequencies in europe
[11:57] <Laney> suc(x)
[11:57] <desrt> also has microsd slot, 3gb of ram, a bunch of other nice upgrades
[11:57] <desrt> unfortunately it's bigger (of course)
[11:58] <davmor2> seb128: hmmm can't retrigger that dialogue I'll try a fresh install later in the day and see if shows up then.
[11:58] <davmor2> seb128: also I can only seem to find a ppa for gnome-software do you happen to know that the link is for launchpad at all please
[12:08] <davmor2> nevermind found it
[12:13] <seb128> hey desrt
[12:14] <seb128> davmor2, https://launchpad.net/distros/ubuntu/+source/gnome-software you mean?
[12:15] <Laney> cyphermox: congrats ;-)
[12:16] <seb128> what do we congrats cyphermox for? ;-)
[12:17] <Laney> https://lists.debian.org/debian-devel-announce/2016/02/msg00005.html
[12:17] <davmor2> seb128: Yay found one that Robert hadn't beaten me too :D https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1551698
[12:19] <seb128> ah, nice ;-)
[12:19] <davmor2> damn it no I didn't
[12:19] <davmor2> dupe it
[12:20] <seb128> davmor2, double negations are not the best!
[12:20] <seb128> also does it work after starting update-manager or doing an apt-get update?
[12:20] <seb128> Laney knows about that I think
[12:20] <seb128> the datas are not autogenerated at install
[12:20] <seb128> they might also not be on the livecd
[12:23] <davmor2> seb128: most likely this is an install not livecd, so I think the issue is that on the initial opening of the app it doesn't trigger an update of the apt list
[12:24] <seb128> davmor2, right, I think Laney was looking at a way that the datas are generated on package install
[12:24] <davmor2> seb128: also software-sources option in gnome software is empty so that seems to be a new bug, I don't know if that is the list it is meant to use
[12:25] <seb128> but I guess g-s could also trigger a refresh when you open it
[12:25] <Laney> I don't know a fix
[12:25] <Laney> going to mail the apt people and see if they have an idea
[12:25] <seb128> thanks
[12:25] <Laney> otherwise gnome-software could do that itself and watch the directory for changes, then refresh
[12:25] <Laney> that would be some work if it can't handle that currently
[12:25] <seb128> davmor2, there is a software-sources option? I didn't see it yet, where is it?
[12:26] <seb128> oh, menu
[12:26] <davmor2> seb128: under the unity bar
[12:26] <seb128> k, I've no idea about that
[12:26] <seb128> we should probably make it open software-properties
[12:27] <seb128> can you open a bug about it?
[12:27] <davmor2> seb128: well for now I will check if that updates after an apt update
[12:27] <seb128> it's empty on my desktop as well, so not likely
[12:29] <seb128> that dialog is useless :-/
[12:29] <seb128> but it does the same on a fedora livecd
[12:29] <seb128> unsure what's the point
[12:30] <davmor2> seb128: I wonder if it is for showing sources of external apps so I will grab chrome and install that and see if it shows anything
[12:31] <flocculant> hi davmor2 seb128 :)
[12:32] <davmor2> flocculant: hey
[12:32] <seb128> hey flocculant
[12:33] <seb128> https://bugzilla.gnome.org/show_bug.cgi?id=737514
[12:33] <seb128> I still find the current message useless though
[12:33] <seb128> it doesn't give you any hint where to go/how to add sources
[12:33] <seb128> there is no point to display the item at all in that context
[12:34] <flocculant> if you're picking at g-s bugs again - now that I've used it to install something - the remove option could do with some tlc bug 1551599
[12:34] <flocculant> seb128: yea I looked at that sources thing and scratched my head
[12:36] <seb128> going for some exercice, bbiab
[12:40] <davmor2> flocculant: Testing most of the new software on the default ubuntu install today so just prodding in general :)
[12:40] <flocculant> :)
[12:41] <flocculant> I've been prodding at g-s :p
[12:41] <Laney> £27.50 to post a suit back!
[12:42] <flocculant> Laney: should have gone to oxfam - you could just take it back and have saved £25
[12:42] <Laney> heh
[12:42] <Laney> and I have to print this form out and fill it in with black ink, then scan it
[12:43]  * Laney screams
[12:43] <flocculant> :)
[12:43] <pitti> wut?
[12:43]  * ogra_ shades his ears
[12:43] <pitti> Laney: inkscape FTW
[12:44] <ogra_> good reason to buy a wacom tablet ;)
[12:44] <pitti> I usually just use evince or so to fill out a form, and paste my signature in with inkscape
[12:44] <davmor2> Laney: is it a pdf if so just use a pdf editor and type it in :)
[12:44] <Laney> I *do* have touchscreen
[12:44] <pitti> (or gimp even)
[12:44] <Laney> it specifically says fill it in using black ink :/
[12:44]  * Laney is scared some jobsworth will make me do it again
[12:44] <Laney> might do so though
[12:44] <ogra_> does it say "non virtual ink" ?
[12:45] <davmor2> Laney: well it will be black type face
[12:45] <Laney> haha
[12:45] <Laney> it says "You will find attached the couriers Terms and
[12:45] <Laney> +Conditions, Hazardous Goods and Packing List forms etc., there are 8 pages in total. All of these must be completed in black ink , signed and dated, and then E-mailed back to us in PDF format."
[12:45]  * Laney risks it
[12:46] <pitti> Laney: remember, https://xkcd.com/499/
[12:46]  * flocculant sends Laney the biscuit
[12:46] <ogra_> yeah, just blame them for not saying "physical ink"
[12:46] <Laney> pitti: hah, that's a new one for me
[12:47] <Laney> oh god I have to fill my card details in
[12:53] <Laney> desrt: are you planning to do stable releases for that gvariant parser bug?
[12:53]  * Laney is uploading for something else to wily and would otherwise take a release
[12:57] <Laney> seb128: http://permalink.gmane.org/gmane.linux.debian.apt.devel/30085
[13:16] <Sweet5hark1> willcooke: libreoffice 5.0.4/xenial on snappy PoC pushed to https://git.launchpad.net/~bjoern-michaelsen/df-libreoffice/+git/libreoffice-snap-playground/log/?h=xenial
[13:17] <willcooke> Sweet5hark1, \m/
[13:17] <Sweet5hark1> willcooke: referenced that and updated bugs. Also wrote some new bugs: https://bugs.launchpad.net/snapcraft/+bugs?field.searchtext=&orderby=-importance&search=Search&field.status%3Alist=NEW&field.status%3Alist=OPINION&field.status%3Alist=INVALID&field.status%3Alist=WONTFIX&field.status%3Alist=EXPIRED&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3A
[13:18] <Sweet5hark1> (woha, longcat is long)
[13:51] <attente> seb128: i can't replicate https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1549502, gnome software still appears in the dash when i search for it while it's still running
[14:11] <mterry> seb128, heyo, so you may have noticed I took executive action and migrated geonames to bzr.  I did this so its branches could work in the citrain.  I hope that wasn't a jerk move
[14:18] <slomo> seb128, Laney: hi :) did you get any gstreamer related bug reports with 1.7.2?
[15:11] <Laney> hi slomo, I didn't hear of anything yet
[15:11] <slomo> great :) thanks
[15:11] <Laney> I saw ricotz pinging you about something or other
[15:11] <Laney> didn't follow that though
[15:12] <slomo> ah that was minor stuff that is fixed now
[15:12] <Laney> ok cool
[15:12] <Laney> are you doing .90?
[15:12] <slomo> yes
[15:12] <Laney> ++
[15:22] <slomo> Laney: i just can't build packages because dh-python is broken in sid in pbuilder (complains about missing dependencies when configuring /o\)
[15:25] <Laney> slomo: Does anything use that?
[15:25] <Laney> kill it from your chroot :)
[15:26] <slomo> Laney: something seems to depend on it, some build dependency
[15:27] <Laney> ah
[15:27] <Laney> a plain apt install dh-python works for me
[15:27] <Laney> not that that's very helpful for you...
[15:28] <ricotz> slomo, hi, I guess this release doesn't mean you will branch 1.8 yet?
[15:28] <slomo> ricotz: no, that will happen together with 1.8.0
[15:28] <Laney> (in a sid schroot)
[15:28] <slomo> why?
[15:28] <slomo> Laney: i'll debug it in a bit :)
[15:29] <ricotz> slomo, just because for switching gnome's jhbuild target
[15:30] <slomo> Laney: installing it in a chroot works for me too
[15:30]  * willcooke in an over running meeting
[15:30] <willcooke> please stand by


[15:31] <willcooke> let's kick off, but I'm still in a meeting...
[15:31] <willcooke> #startmeeting Desktop Team Weekly Meeting - 2016-03-01
[15:31] <meetingology> Meeting started Tue Mar  1 15:31:28 2016 UTC.  The chair is willcooke. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
[15:31] <meetingology> Available commands: action commands idea info link nick
[15:31] <willcooke> Roll call:  andyrock (appt), attente, desrt,  dgadomski, fjkong (out), happyaron, hikiko (appt), laney, qengho (appt), seb128, sweet5hark, themuso (out), tkamppeter (out), trevinho(hols), robert_ancell (out)
[15:31]  * davmor2 I hate that willcooke now I have greensleeves stuck in my head
[15:31] <Sweet5hark> aye
[15:32] <seb128> _o/
[15:33] <desrt> word up
[15:33] <willcooke> #topic andyrock
[15:35] <desrt> *cough*
[15:35] <willcooke> * Reviews
[15:35] <willcooke> * HR stuff (self review and 360)
[15:35] <willcooke> * Fixed all the branches I proposed according to Marco's reviews.
[15:35] <willcooke> * Found a simple fix for the lockscreen/suspend issues. I'm looking
[15:35] <willcooke> into adding some unit tests for it.
[15:35] <willcooke> my bad, distracted
[15:35] <willcooke> #topic attente
[15:36] <attente> hey
[15:36] <attente> fix a couple minor crashers in gnome-software, make sure we ask for u1 creds for all requests where we need it, e.g. #1549508
[15:37] <attente> doing some refactoring of the gtkmenu api branch
[15:37] <attente> (eof)
[15:38] <willcooke> sorry, talking
[15:38] <willcooke> thanks attente
[15:38] <willcooke> #topic desrt
[15:38] <Laney> yak yak yak
[15:38] <desrt> hihi
[15:38] <desrt> was a 'bugs' week, more or less
[15:38] <desrt> eof.
[15:39] <willcooke> #topic dgadomski
[15:39] <dgadomski> hey
[15:39] <dgadomski> I was on a sprint last week, so nothing really to share this time.
[15:39] <willcooke> thank dgadomski
[15:39] <willcooke> #topic FJKong
[15:39] <willcooke> bug/1545910
[15:39] <willcooke> Cannot input Chinese characters into TextField in a QML application
[15:39] <willcooke> still in progress, tesing TestField
[15:39] <willcooke> pinyinsearch
[15:39] <willcooke> finding new method for converting pingyin
[15:39] <willcooke> Sogou IM
[15:39] <willcooke> verify loging problem causing by IM
[15:39] <willcooke> #topic happyaron
[15:41] <seb128> zzZZzz
[15:41] <willcooke> happyaron, please let me have you updates
[15:42] <willcooke> We need to have them on a Tuesday afternoon your time *every* week.
[15:42] <willcooke> #topic hikiko
[15:42] <willcooke> Sorry, haven't had time to summarise this:
[15:42] <willcooke> Black dots bug: I tested my theory that the problem is in Compiz and
[15:42] <willcooke> found out for sure that it's at the part that creates the textures for
[15:42] <willcooke> the windows. It can't be in gtk because then the dots would appear in
[15:42] <willcooke> every window manager but I see no issues in XFCE. It's not related to
[15:42] <willcooke> the shadows or the decorations as we initially thought because we still
[15:42] <willcooke> see the problem when the unityshell plugin is disabled. So, to find out
[15:42] <willcooke> what happens in Compiz, I modified the opengl plugin to extract and save
[15:42] <willcooke> the texture of the dot-windows at the time that glDrawTexture is called
[15:42] <willcooke> for the window and right before the texture is send to the GPU and I
[15:42] <willcooke> noticed that the image I write to the disk always has the dots on it
[15:43] <willcooke> (the window textures are created and filled by compiz so, compiz creates
[15:43] <willcooke> the dots). Then, I created and saved images with the window alpha values
[15:43] <willcooke> to see how they vary in the transparent and semi-transparent regions of
[15:43] <willcooke> the window and the images were all white (=> we had no alpha). So, I am
[15:43] <willcooke> debugging the opengl plugin texture mapping parameters and the compiz
[15:43] <willcooke> plugins that modify them to see why the A is ignored in all cases even
[15:43] <willcooke> when we use RGBA, I hope that the fix won't take more than a few days
[15:43] <willcooke> now that I figured out where the problem is and what I have to test
[15:43] <willcooke> exactly. EOF
[15:43] <willcooke> hikiko, bullet points next time please
[15:43] <willcooke> #topic Laney
[15:43] <Laney> • It's been a short week
[15:43] <Laney> • I had broken the release upgrader, so I fixed that
[15:43] <Laney> • Reviewed, travestied and uploaded a Zeitgeist branch from Trevinho to VACUUM every now and again, which speeds up the dash
[15:43] <Laney> • Got pinged by rishi to backport a patch to gnome-terminal from Fedora to fix some padding problems
[15:43] <Laney> • Some yak shaving when my internal mirror started 404ing for directories which existed (srsly)
[15:43] <Laney> • Chatted with IS about appstream screenshot problem - we found a way out!
[15:43] <Laney> • Wrote a little patch for appstream-dep11 to put a "Date" field in the header document, so we can track if server → Launchpad → apt is working right, got some comments from ximion to fix, thanks
[15:43] <Laney> • Backported GDBusProxy memleak fix for wily, just got to dput it... oh I might as well do that. OK it's done.
[15:43] <Laney> • Merged ffmpeg for some reason, I think maybe because I noticed an outstanding security fix?
[15:43] <Laney> ⭒ (https://codepoints.net/random is my new friend)
[15:44] <Laney> https://codepoints.net/U+1F060
[15:44] <seb128> 𓃀
[15:44] <Laney> https://codepoints.net/domino_tiles
[15:44] <attente> ꕥ
[15:44] <Laney> unicode is the best
[15:44] <seb128> nice
[15:45] <seb128> 🁠
[15:45] <seb128> it even works in xchat ;-)
[15:45] <willcooke> thanks Laney
[15:45] <Laney> I can't see those characters
[15:45] <Laney> :(
[15:45] <willcooke> #topic qengho
[15:45] <willcooke> * in-progress: Chromium snap. Going pretty well.
[15:45] <willcooke> * in-progress: Revisiting chromium default-browser setting. Internal xdg-settings broken again. Needs autopkgtest too.
[15:45] <willcooke> * done: grub patches and plans for zfs patches with cking.
[15:45] <willcooke> #topic seb128
[15:46] <seb128> • looked at some gtk theming issues (totem slider & checkboxes, editable listview entry, ...)
[15:46] <seb128> • new poppler (soname change -> transition/rebuilds)
[15:46] <seb128> • hit some build issue due to ghostscript, backport a fix for that
[15:46] <seb128> • reported some gnome-software & gnome-calendar bugs upstream
[15:46] <seb128> • fixed gnome-software missing translations (potfiles outdated)
[15:46] <seb128> • desktop updates (gtk)
[15:46] <seb128> • worked on nautilus sidebar/gtk issues (gtk changed abi), no easy solution so far (tried to copy the old gtkplacesidebar widget in nautilus but that turned out to be too much work)
[15:46] <seb128> • canonical hr reviews
[15:46] <seb128> • contributed to snappy discussions
[15:46] <seb128> • sponsored onboard changes to the default config (docked, stretched at the bottom of the screen)
[15:46] <seb128> • new reviews (a11y-profile-manager, libertine-scope, qclib)

[15:47] <seb128> ping?
[15:47] <seb128> (I wonder if I'm lagging today?)
[15:47] <willcooke> thanks seb128
[15:47] <willcooke> #topic Sweet5hark
[15:47] <seb128> feels like high latency, unsure if that's on my side
[15:48] <willcooke> seb128, I'm in a meeting, trying to do two things at once and failing
[15:48] <Sweet5hark> - snapcraft wrangling on xenial: filed bug count now at 6, more feedback and input
[15:48] <Sweet5hark> - build a libreoffice 5.0.4 snap PoC on snapcraft (source is here: https://git.launchpad.net/~bjoern-michaelsen/df-libreoffice/+git/libreoffice-snap-playground/log/?h=xenial)
[15:48] <Sweet5hark> -- not running, but limping along in a most basic if you smack it with LD_LIBRARY_PATHs and fun
[15:48] <Sweet5hark> - setting up snappy dimension on classic
[15:48] <Sweet5hark> - upstream TDF board call
[15:48] <seb128> I see
[15:48] <Sweet5hark> - upstream ESC call:
[15:48] <Sweet5hark> -- RedHat want to sponsor experimental xdg-app build there
[15:48] <Sweet5hark> -- created a development budget estimate
[15:48] <Sweet5hark> -- other local events admin: Ankara Hackfest, German LoCo budgeting
[15:48] <Sweet5hark> - some firefighting on lagging/missing upstream l10ns
[15:48] <Sweet5hark> - some contract/job candidate negotiations
[15:48] <Sweet5hark> - libreoffice 5.1.0-ubuntu1 bump
[15:48] <Laney> xdg-app build...
[15:48]  * Sweet5hark got throttled ...
[15:48] <Laney> sounds up your street Sweet5hark
[15:48] <Sweet5hark> -- created a development budget estimate
[15:48] <Sweet5hark> -- other local events admin: Ankara Hackfest, German LoCo budgeting
[15:48] <Sweet5hark> - some firefighting on lagging/missing upstream l10ns
[15:48] <Sweet5hark> - some contract/job candidate negotiations
[15:48] <Sweet5hark> - libreoffice 5.1.0-ubuntu1 bump
[15:48] <Sweet5hark> - libreoffice 5.1.1~rc2 bump (with bugs 1506544, 753627)
[15:48] <Sweet5hark> EOF
[15:49] <willcooke> thanks Sweet5hark
[15:49] <willcooke> #topic TheMuso
[15:50] <willcooke> oh, TheMuso I am missing your update
[15:50] <willcooke> #topic tkamppeter
[15:50] <willcooke> - Preparations for the OpenPrinting Summit 2016
[15:50] <willcooke>   o Finalized agenda
[15:50] <willcooke>   o Sent out remaining invitations
[15:50] <willcooke>   o Started organization of the trip
[15:50] <willcooke> - Updated OpenPrinting for some new Ricoh printers
[15:50] <willcooke> - Google Summer of Code 2016: The Linux Foundation got accepted!
[15:50] <willcooke>   o Spread the news to the possible mentors
[15:50] <willcooke> - Tests for possible new concept for auto-downloadable driver packages.
[15:50] <willcooke> - Bugs.
[15:50] <willcooke> #topic Trevinho
[15:50] <willcooke>  * New libdbusmenu upstream release
[15:50] <willcooke>  * New unity and compiz landing prepared
[15:51] <willcooke>  * gtk black-corners saga: Working on exporting gtk border radius for
[15:51] <willcooke> windows with a titlebox (headerbar) and tooltips. First part is done,
[15:51] <willcooke> tooltips are in WIP. Unity side of things is ready.
[15:51] <willcooke> #topic robert_ancell
[15:51] <willcooke> - GNOME Software work
[15:51] <willcooke> #topic any other business
[15:51] <willcooke> Sorry folks, still in another meeting
[15:51] <willcooke> didn't want to hold you all up
[15:51] <Laney> I hope it is a good meeting
[15:52] <willcooke> its certainly interesting
[15:52]  * seb128 does like much when things are "interesting" :p
[15:52] <willcooke> very wise seb128
[15:52]  * seb128 prefers "quiet"
[15:52] <willcooke> any one got aob?
[15:52]  * Sweet5hark mumbles "may you live in interesting times"
[15:53] <willcooke> #endmeeting
[15:53] <meetingology> Meeting ended Tue Mar  1 15:53:23 2016 UTC.
[15:53] <meetingology> Minutes:        http://ubottu.com/meetingology/logs/ubuntu-desktop/2016/ubuntu-desktop.2016-03-01-15.31.moin.txt
[15:53] <Laney> thanks!
[15:53] <seb128> thanks!
[15:54] <Sweet5hark> aob means "all order beer" doesnt it?
[15:54] <seb128> \o/
[15:54] <slomo> Laney: https://wiki.debian.org/PbuilderTricks#How_to_include_local_packages_in_the_build i think it's this not working anymore because "E: Internal Error, Pathname to install is not absolute 'libgstreamer1.0-0_1.7.90-1_amd64.deb'"
[15:55] <slomo> Laney: dh-python was just follow-up failure
[15:55] <slomo> Laney: any ideas? :)
[15:55] <Laney> slomo: not one that you're going to like, but I use sbuild and a local repository :)
[15:55] <seb128> slomo, hey, no new gst 1.7 bugs reported that I saw
[15:56] <seb128> mterry, hey, we were just discussing that git->bzr conversion this morning ... any idea if CI support for git is planned? also unsure we need CI landing for that lib
[15:56] <slomo> Laney: ok, i don't like that answer :)
[15:57] <mterry> seb128, git support is planned (being worked on now), but robru said it would be several months before done
[15:57] <Laney> slomo: mapreri seems to have done the last upload and is in #debian-gnome & other channels, could maybe ask
[15:57] <seb128> mterry, I've to admit I'm more familiar with the bzr workflow so it makes things easier, feel a bit wrong to go against the maintainer decision though (but at the same time we don't have an active maintainer now so...)
[15:57] <mterry> seb128, CI landings are very useful for me right now, to be able to match fixes / API additions to other work in silos.  But I get it if you'd rather not.  We can go back to git easily enough
[15:58] <mterry> seb128, yeah I figured there was no maintainer, except Canonical.  Who usually does bzr
[15:58] <seb128> mterry, you can dput to silos as well
[15:58] <seb128> but as said bzr feels more in line with what we know/use
[15:58] <mterry> seb128, sure.  But that's more awkward
[15:58] <seb128> that might shift at some point
[15:58] <Laney> slomo: I meant #debian-devel
[15:58] <seb128> so I'm fine with the bzr switch
[15:59] <mterry> seb128, sorry for just going ahead and doing it, but I figured this was a ask-for-forgiveness-rather-than-permission situation  :)
[15:59] <mterry> Made it easier for me to offer a silo for the geonames transition for system-settings and the oobe wizard
[15:59] <seb128> no worry
[16:00] <seb128> we need to review your changes as well
[16:00] <mterry> yeah
[16:00] <mterry> most of them should be no brainers
[16:00] <seb128> I guess larsu is not very interested in that anymore so I'm going to try to have a look
[16:00] <seb128> or maybe Laney beats me to it :p
[16:00] <Laney> mterry: what's the answer if a real upstream component wants to depend on something which is in the CI train?
[16:01] <Laney> like if and when libtimezonemap itself gets ported
[16:01] <mterry> Laney, same as today.  Wait for whatever change you want to be in the archive, then upload
[16:02] <Laney> I mean for other distributors
[16:02] <mterry> Laney, ah.  Well we can still issue tarballs if we want.  Or just point at trunk as many projects do these days
[16:03] <mterry> Laney, but given our (lack of) success with libtimezonemap, getting other distros to use it, I have little hope
[16:05] <Laney> http://koji.fedoraproject.org/koji/packageinfo?packageID=17332
[16:05] <seb128> what does use it for?
[16:05] <seb128> last time I checked g-c-c had a modified version compared to us I think
[16:05] <Laney> dunno, maybe anaconda
[16:06] <Sweet5hark1> (re) -- notebook hardfreezed, yay
[16:08] <seb128> indeed
[16:08] <seb128> I guess it's the only project using it
[16:08] <seb128> debian codesearch lists nothing
[16:08] <seb128> anyway I think CI landing process doesn't block us to roll tarballs as well
[16:09] <seb128> it just gives more agility to snapshot in Ubuntu
[16:10] <Laney> up to you
[16:11] <seb128> up to whoever is taking over that project I gues
[16:11] <seb128> assuming that larsu is not interested to keep maintaing it
[16:11] <seb128> larsu, still interested in geonames?
[16:13] <larsu> seb128: I have some pending work for it still, but I'm not interested in maintaining it further (desrt said she'd like to do that)
[16:14] <seb128> larsu, desrt, mterry proposed some changes, maybe you could give them a look? https://code.launchpad.net/~geonames-dev/geonames/trunk/+activereviews
[16:14] <seb128> well at least the double-free and expose-more
[16:14] <seb128> the third is packaging
[16:15] <larsu> thanks, will try to have a look later today
[16:16] <seb128> larsu, thanks!
[16:17] <desrt> larsu: thanks for volunteering me :p
[16:18] <larsu> desrt: I remember us talking about this in Cologne...
[16:18] <larsu> no?
[16:18] <desrt> that's not true
[16:18] <desrt> we talked about it in brussels :)
[16:18] <larsu> ha
[16:18] <larsu> fair enough
[16:18]  * Laney wants crepes
[16:18] <desrt> Laney: queen west
[16:19] <larsu> Laney: I know a *very* good place in Potsdam. Just went to that on the weekend
[16:20] <larsu> desrt: not even close
[16:20] <larsu> (different style, tough)
[16:21] <Laney> maison des crepes
[16:21] <desrt> i don't really feel too qualified to review these patches...
[16:23]  * desrt receives a patch that uses an unsigned 32bit int for population of a city and begins to ponder potential future scalability issues
[16:24]  * desrt ponders world municipal government
[16:24] <Laney> make it an array of digits
[16:24] <desrt> gvariant nearly had an arbitrary precision int type
[16:25]  * larsu is thankful it didn't make it in
[16:25] <larsu> framing offsets are arbitrary position though
[16:25] <desrt> ish...
[16:26] <desrt> but fwiw, i would indeed implement the arbitrary int type in exactly the same way
[16:26] <desrt> probably dropping only the power-of-two-size restriction
[16:26] <desrt> i was mostly considering the potential use of gvariant in cryptographic protocols... but 'ay' works well enough there
[16:27] <larsu> you'd need to specify the size as well, though
[16:27] <desrt> nope
[16:27] <larsu> which is implicit for container types
[16:27] <desrt> this is not how gvariant works
[16:27] <desrt> the size is always assumed to be known
[16:28] <larsu> ah. right. it would be for that type as well
[16:28] <desrt> storing a 4096 bit RSA public modulus, for example, would look pretty much exactly like having 4096 bits
[16:28] <larsu> I read the spec the other day and started an implementation. in javascript
[16:28] <larsu> (because why not)
[16:28] <desrt> oh my god
[16:28] <larsu> haha
[16:29] <desrt> kgvariant gets joined by jsgvariant
[16:29] <larsu> I imagined that's how you'd react
[16:29] <larsu> k?
[16:29] <desrt> kernel
[16:29] <larsu> well that doesn't exit
[16:29] <desrt> dvdhrm is also working on another independent userspace implementation, i think
[16:29] <desrt> (or maybe that's the kernel one?)
[16:29] <larsu> I know
[16:29]  * desrt has lost track at this point
[16:30] <larsu> there is no kernel one as far as I know
[16:30] <desrt> i thought kay was using one
[16:30] <larsu> if there will be, then it's probably dvdhrms (which is called CVariant)
[16:30] <larsu> nah, Kay is not working on this at all
[16:31] <desrt> well, cool
[16:31] <larsu> more implementations \o/
[16:31] <desrt> ping me if you need any help
[16:31] <desrt> i know dvdhrm is already compiling a list of errata to the spec
[16:31] <larsu> thanks. looking good so far though
[16:31] <larsu> but I only have the basics
[16:32] <desrt> mostly parts that are ambiguously-worded when it comes to the proper handling of non-normal forms
[16:32] <desrt> (i think the spec is clear, but it never hurts to be "more clear")
[16:32] <larsu> and only a reader
[16:33] <desrt> there are some philosophical questions in there... like what does it mean to attempt to load a value of a fixed-sized type with the wrong size?
[16:33] <larsu> abort(0)
[16:33] <desrt> and whether a fixed-sized value appearing at the end of an incorrectly-formed tuple with extra bytes there is incorrectly-sized, or if the container itself simply contains extra space
[16:34] <desrt> (my read: the container has extra bytes and they are ignored)
[16:34] <larsu> be liberal in what you expect?
[16:34] <desrt> ya.. i think david tries to deal with this case by loading an appropriate number of zeros and discarding the user's buffer
[16:35] <desrt> well, the trouble with being liberal in what you accept is that a design criteria of GVariant is that all implementations should always deserialise every single byte sequence in exactly the same way
[16:35] <larsu> yeah i think so
[16:35] <desrt> so the spec tries to precisely define even the broken cases
[16:35] <larsu> I know
[16:35] <desrt> and that's where some corners are not hammered-down as much as dvdhrm would like them
[16:36] <larsu> why do you have that requirement by the way?
[16:36] <larsu> you could just say certain cases are invalid and can't be deserialized
[16:36] <desrt> it comes as the absurd logical result of some perfectly reasonable design constraints
[16:36] <desrt> but basically it comes down to the following:
[16:36] <desrt> 1) for performance reasons, we cannot scan the entire buffer on loading it
[16:37] <desrt> 2) we don't want the accessor APIs to be failable
[16:37] <desrt> 3) therefore, there is no place in which we can possibly detect errors
[16:37]  * larsu was expecting (2)
[16:37] <desrt> 4) therefore we must forbid the possibility of there being errors
[16:38] <desrt> "each logical value must only serialise to a single byte sequence" is not a requirement, so that's what we throw out the window
[16:38] <desrt> (although that would be convenient for quick equality testing... equality testing serialised GVariant values is an almost-never operation)
[16:40] <larsu> why not though? seems reasonable
[16:40] <larsu> because of dicts?
[16:40] <desrt> dicts are actually the one place where we do do equality tests
[16:40] <desrt> for lookups...
[16:40] <desrt> it's just not really a common thing that you want to know if one dbus message is exactly equal to another
[16:41] <larsu> I mean, why not the requirement
[16:41] <desrt> i mean... you just don't care
[16:41] <larsu> yeah of course
[16:41] <desrt> so requirements should only come from things that we actually care about :p
[16:43] <larsu> ha fair enough
[16:59] <seb128> attente, how do you test gnome-software built from git?
[16:59] <seb128> jhbuild? start in the build tree? package?
[16:59] <attente> seb128: i'm doing it via jhbuild
[17:00] <seb128> using the aptd backend as well? or packagekit?
[17:00] <seb128> because building wip/rancell/reviews doesn't give me sources
[17:01] <attente> seb128: i'm not sure, but i guess it's using PackageKit
[17:01] <seb128> I guess maybe something missing in the apt backend then
[17:02] <attente> is there a concrete way for me to find that out?
[17:02] <seb128> I don't know
[17:02]  * seb128 looks at some code
[17:02] <attente> like could i do a strace and grep for it somehow?
[17:06] <seb128> attente, I think the backend needs to implement gs_plugin_add_sources()
[17:06] <seb128> the xdg-app and packagekit does it
[17:08] <attente> seb128: yeah, i just commented it out from the packagekit backend
[17:08] <attente> so it's using that backend to get the sources list i guess
[17:12] <seb128> attente, I think we have enough work on g-s and we are not going to replace software-properties this cycle, we should probably just make the menu item call that one, wdyt?
[17:13] <attente> seb128: that sounds totally reasonable to me
[17:13] <seb128> cool, I commented on the bug saying so, let's see what robert thinks
[17:14] <seb128> attente, does "gnome-software --local-filename=somepackage.deb" works in jhbuild?
[17:14] <seb128> where the .deb is a valid one you have on disk
[17:15] <attente> seb128: no, i get a really ugly message box saying "Sorry, this did not work" "no filename_to_app results to show"
[17:15] <seb128> k, so same as on xenial
[17:16] <seb128> bug #1551707
[17:43] <davmor2> Sweet5hark1: How do I check if LO is using the right icon set please
[17:46] <Sweet5hark1> davmor2: depends on the scenario. I assume this is about the recent "default to breeze on unity" thing?
[17:47] <seb128> Laney, bug #1551877 if you feel like approving a ffe... ;-)
[17:47] <davmor2> Sweet5hark1: nope just testing and I know it was an issue before so just checking that it is still fixed
[17:49] <Sweet5hark1> davmor2: well this is currently an area under construction ;)  -- see bug 1506544.
[17:51] <Sweet5hark1> the thing to check for would be: 1/ have a fresh install (or deleted .config/libreoffice) 2/ start LO 3/ go to tools->options->libreoffice->view->user interface->icon size and style and it should say "automatic (Breeze)" now.
[17:51] <Sweet5hark1> davmor2: ^^
[17:52] <seb128> $ appstreamcli search emacs
[17:52] <seb128> ...
[17:52] <seb128> Identifier: uxvile.desktop [desktop]
[17:52] <seb128> Name: UXVile
[17:52] <seb128> Summary: VI Like Emacs
[17:52] <seb128>  
[17:52] <Sweet5hark1> davmor2: That works only though if libreoffice-style-breeze is installed, which is a recent change on the xenial-seed and I dont know if that is on the image already.
[17:52] <seb128> that feels like somebody is trolling
[17:52] <davmor2> Sweet5hark1: hmmm automatic (human) I assume this isn't running the very latest version
[17:54] <davmor2> Sweet5hark1: nope human on my main machine that is up to date
[17:54] <davmor2> Sweet5hark1: on a plus side they both look the same so it isn't running the GTK theme I guess \o/
[17:55] <Sweet5hark1> davmor2: thats ok, the change is due with both libreoffice-style-breeze preinstalled from the seed and libreoffice-5.1.1~rc2-0ubuntu1 being promoted out of proposed ...
[17:57] <Sweet5hark1> davmor2: which I guess will happen as part of the Grand-Poppler-so-update, so stay tuned.
[18:02] <davmor2> Sweet5hark1: nice Thanks I'll keep an eye out for that then
[18:02] <sarnold> seb128: I used vile on windows years ago; that was back before my fingers knew vim-specific keybindings, but vile worked great for me ~twenty years ago
[18:03] <Sweet5hark1> davmor2: thanks for that! Good to prevent another OMG-moment like the last time we talked about this ;)
[18:03] <seb128> sarnold, yeah, I was just pointing out that searching for emacs returns a "vi like" entry ;-)
[18:03] <sarnold> seb128: ah :)
[18:06] <Laney> seb128: will have to do that tomorrow, feel free to get someone else
[18:06] <Laney> my beautiful synced package
[18:06] <Laney> see you!
[18:06] <willcooke> see ya Laney
[18:08] <Sweet5hark1> seb128: we should totally ship https://github.com/seanyeh/vibreoffice by default. People will love it.
[18:10] <seb128> Laney, have fun!
[19:00] <davmor2> willcooke, seb128: did startup disk creator loose it's ability to add persistence?
[19:01] <willcooke> umm, wasnt that a UEFI issue?
[19:01] <willcooke> don't know, laney probably does
[19:09] <davmor2> willcooke: ah right okay I'll try an i386 install then
[19:10] <flocculant> davmor2: that went with trying to make it work iirc
[19:11] <seb128> davmor2, yes, mdeslaur removedit
[19:11] <seb128> https://launchpad.net/ubuntu/+source/usb-creator/0.3.0
[19:17] <davmor2> seb128, willcooke: hmmm okay so not it does nothing more than Disks then now right?
[19:17] <davmor2> s/not/now
[19:18] <willcooke> right, but it at least works again
[19:18] <seb128> correct
[19:18] <seb128> well, I think mdeslaur said that it disks would hit some udisk issue than the usb-creator refactoring doesn't
[19:19] <seb128> it's also what is documented on the internet, so good to still have it and working
[19:19] <willcooke> plus there are a lot of docs out there which refer to it, so better it does something
[19:19] <willcooke> :)
[19:21] <willcooke> seb128, those calendar bugs which davmor2 flagged - you think a lot of them are fixed in the newer u/s version?
[19:22] <mdeslaur> davmor2, seb128: the plan was to add persistence back at some point, but lack of time, etc.
[19:22] <seb128> willcooke, a few are, but I don't think most are important
[19:22] <seb128> I also can't reproduce some of those
[19:22] <seb128> e.g clicking on online accounts works for me
[19:23]  * willcooke has a go
[19:23] <seb128> or clicking on "today" bring back to the current month
[19:23] <seb128> even if that's probably not an obvious result
[19:23] <seb128> unsure what else we would expect without a day view
[19:28] <willcooke> Today works for me
[19:28] <willcooke> online accounts does not
[19:31] <willcooke> also the high CPU usage is on my test machine too
[19:31] <willcooke> and in online accounts I have no Google Calendar option at all (which might be know)
[19:32] <seb128> willcooke, davmor2, I triaged the bugs
[19:32] <seb128> willcooke, online accounts, you should be able to add a google account, that includes access to the calendars
[19:32] <willcooke> oh, nice one, thanks seb128
[19:32] <willcooke> seb128, yeah, that's what I thought, but no
[19:32] <davmor2> mdeslaur: thanks for the clarification.  One last question should it support gvfs connections over ssh for iso collection it seems to fail currently so I assume not but thought I would get a definite answer
[19:33] <seb128> willcooke, you might lack an uoa provider deb
[19:33] <mdeslaur> davmor2: no, it doesn't...I don't believe it did before either
[19:34] <davmor2> mdeslaur: it finds the iso but then throws up an error when it tries to transfer it by the look of it
[19:34] <mdeslaur> yeah, I don't think the backend supports that
[19:34] <davmor2> mdeslaur: awesome thanks
[19:35] <davmor2> mdeslaur: good job by the way the basic tool works well :)
[19:35] <mdeslaur> heh, at least it kind of works now :)
[19:39] <willcooke> seb128, aha...  (process:2315): e-data-server-WARNING **: e_module_load_all_in_directory: Error opening directory '/usr/lib/evolution-data-server/credential-modules': No such file or directory
[19:39] <willcooke> thats trying to open u-o-a
[19:39] <willcooke> from the link in g-c
[19:40] <seb128> willcooke, do you have evolution-data-server-online-accounts installed?
[19:41] <willcooke> seb128, nope.  installing...
[19:42] <willcooke> yay, I logged out after installing it and now I can't log back in :_)
[19:43] <willcooke> reboot, and can log in again
[19:49] <willcooke> still doesn't let me open online accounts
[19:52] <willcooke> ah ha
[19:52] <willcooke> davmor2, seb128 installing gnome-online-accounts fixes it
[19:52] <willcooke> tut
[19:53] <davmor2> willcooke: right so it is pointing to the wrong thing then right?
[19:53] <willcooke> I think theres more to it
[19:54] <willcooke> since my Google calendars weren't showing up either until I installed g-o-a
[19:56] <davmor2> seb128: so regarding the colour change I can only assume because on restarting the app it showed the next bug and displayed no calendars at all at which point I had to stop
[19:56] <willcooke> I think we can get those fixed in time.  Not too worried
[19:57] <davmor2> willcooke: indeed this was more just a tyre kicking to see what shook loose
[19:58] <willcooke> davmor2, totes - and much appreciated
[19:59] <davmor2> willcooke, seb128: so on a plus side the bulk of the apps just work nautilus being backed out seems to of fixed the screen resize issue too which is nice :)
[19:59] <willcooke> nice!
[20:00] <seb128> yeah
[20:00] <seb128> willcooke, goa shouldn't be required, looks like you are missing the uoa stack
[20:01] <willcooke> lemme see what it pulled in
[20:01] <seb128> willcooke, is unity-control-center-signon installed?
[20:01] <attente> ximion: hey. is there an easy way to force gnome-software to quit when the main window is closed?
[20:03] <ximion> attente: no, it will keep running in background holding the cache for faster startup, and occasionally search for updates and download them when not on battery and on a regular wifi/network connection
[20:03] <ximion> that's on purpose
[20:03] <davmor2> seb128: ii unity-control- 0.1.8+16.04. amd64 .....
[20:03] <willcooke> seb128, yeah it is, and I dont think it was pullled in by g-o-a - so I think it was there before
[20:03] <seb128> davmor2, yeah, no doubt about you, I was asking willcooke who doesn't have a standard install/was missing some binaries
[20:03] <ximion> oh, and GS will also provide the PackageKit session interface for apps to consume, but I think that's intentionally broken (or removed?) on Ubuntu
[20:03] <seb128> willcooke, weird
[20:04] <seb128> willcooke, I need to go for dinner but I might be back later, otherwise we can debug tomorrow
[20:04] <attente> ximion: i think we might want to disable that and just use update-manager for ubuntu
[20:04] <seb128> +1
[20:04] <seb128> no process eating memory for nothing please ;-)
[20:04] <willcooke> seb128, sure thing, no hurry.  Enjoy dinner.
[20:05] <attente> ximion: i also saw that the code allows us to say that the distro manages updates
[20:05] <willcooke> seb128, I'm just procrastinating so I dont have to do the reviews anyway ;)
[20:05] <attente> ximion: by removing permissions to org.freedesktop.packagekit.trigger-offline-update
[20:06] <ximion> attente: likely - the key point is really the smarter updates via PK (with battery and wifi checks), if you don't use it, there's not that many reasons to keep the daemon running, although I am not sure if there are other features provided by it now (haven't looked at that part for a while, and GS development is advancing rapidly)
[20:06] <ximion> attente: GS was never really designed to run without PK - support for that was added to allow it to be xdg-app sandboxed, and Ubuntu with it's APT backend is really the only distro doing that
[20:07] <ximion> so: I am not sure if that's intentional ;-)
[20:07] <ximion> (and what the sideeffects would be)
[20:07] <ximion> the offline updates are supposed to be configured by the local administrator though, so removing permissions should definitely prevent users from upgrading
[20:08] <ximion> (but you could just as well drop that, since Ubuntu's PK is way too old to support this feature anyway)
[20:10] <attente> ximion: interesting... but even when those permissions are disabled, GS seems to still run as a daemon
[20:11] <attente> i can't figure out what part of the code is holding GS running
[20:11] <attente> (after the main window is closed)
[20:13] <ximion> attente: the part where it never quits the mainloop and keep GsUpdateMonitor running? ;-)
[20:13] <davmor2> willcooke: am I remembering correctly, was there not talk of removing the web apps?
[20:13] <willcooke> davmor2, yeah, I think it went in 15.10
[20:14] <davmor2> willcooke: but we still have the amazon app on the install
[20:14] <attente> ximion: yep, i've disabled the parts where g_application_hold() is called, but it still somehow sticks around. is there another hidden window somewhere i'm not aware of?
[20:14] <willcooke> davmor2, ah, different thing.  Yes, that is still there
[20:15] <davmor2> willcooke: right so it is the firefox goto facebook do you want this as an app bit that is removed then right?
[20:15] <willcooke> davmor2, should be, yes
[20:17] <davmor2> yeap not happening on plus.google.com
[20:18] <ximion> attente: I think you just disabled the timeout then and now it will never quit... You should remove the G_APPLICATION_IS_SERVICE flag from the GApplication flags
[20:18] <ximion> see if that solves your problem
[20:18] <attente> what's setting G_APPLICATION_IS_SERVICE?
[20:18] <willcooke> davmor2, the system works!
[20:19] <davmor2> willcooke: well you still can't use software calendar or upgrade but on varying levels of works yes :)
[20:21] <attente> ximion: if you're talking about the quit_activated() code path, it doesn't get called when the main window is closed
[20:22] <davmor2> willcooke: amazon is pointing to amazon.com rather than co.uk and I have a UK locale so it might just be an issue in the vm so I'll test that on hardware tomorrow and see what is what with that
[20:23] <Fudge> morning
[20:23] <willcooke> davmor2, https://bugs.launchpad.net/webapps-sprint/+bug/1523464
[20:24] <davmor2> willcooke: nice
[20:24] <willcooke> \m/
[20:24] <ry> Hey guys, is there a specific place to ask a question or two about Deja-Dup?
[20:25] <mhall119> Laney: ping
[20:25] <mhall119> about AppStream data
[20:25] <willcooke> ry, try in #ubuntu - probably the best place
[20:26] <willcooke> mhall119, he's EOD.  But ask anyway, cos there are a few people around who might know ;)
[20:27] <ry> willcooke, thanks
[20:28] <robert_ancell> seb128, are you getting a menu for GNOME Software?
[20:29] <mhall119> willcooke: Laney: so I'd like to start organizing community contributors towards fixing some of the AppStream data, it seems that missing or incorrect icons are the cause of the majority of the errors we have, and it would be an easy thing for people to get (we did this a few years ago to get hi-res icons for apps in Unity)
[20:29] <willcooke> mhall119, nice one, thanks
[20:29] <mhall119> my primary question, though, is if contributors could just provide the icon to someone on this team, or if they would have to create (and learn how to create) a package patch to include it
[20:30] <mhall119> in the previous effort I mentioned, I believe they just collected the icons and attached them to bug reports for apps that needed them, then somebody else actually updated the package
[20:30] <mhall119> obviously they would be getting the improved icon into upstream too, but for 16.04 they would need to get it into Ubuntu directly
[20:31] <mhall119> or is AppStream something that can be updated easily after release?
[20:33] <ximion> mhall119: only with hacks
[20:33] <ximion> ideally, metadata updates and bugfixes should go to upstream or the packaging
[20:34] <ximion> attente: it's indeed weird - you could quit the mainloop when the main window is closed, that should work
[20:34] <seb128> robert_ancell, hey
[20:34] <seb128> robert_ancell, @menu, yes ... not you?
[20:35] <robert_ancell> seb128, yeah, I'm not
[20:35] <ximion> attente: because since GS is supposed to continue running as daemon, it will never exit the loop
[20:35] <ximion> (usually)
[20:35] <robert_ancell> seb128, what's your "Show the menus..." setting?
[20:35] <robert_ancell> (In Appearance settings panel)
[20:35] <attente> ximion: yeah... i'm just wondering what exactly is keeping the main loop running. i feel like i've disabled everything that could possibly keep it open
[20:36] <attente> well, clearly i'm not ;)
[20:36] <robert_ancell> seb128, yeah, must be that
[20:36] <seb128> robert_ancell, I've the old "included in the unity panel" one
[20:36] <robert_ancell> seb128, they don't show when set to "in the window's title bar"
[20:36] <robert_ancell> Is there an open issue for that?
[20:36] <robert_ancell> Or is this G-S specific
[20:36] <seb128> not yet
[20:36] <seb128> but it's what we discussed yesterday
[20:36] <robert_ancell> seb128, yeah, part of the CSD stuff?
[20:37] <seb128> stop using GtkHeaderBar decorations
[20:37] <seb128> yes
[20:37] <seb128> CSD is uncompatible with local menus
[20:37] <seb128> by design
[20:37] <robert_ancell> seb128, so, should I make a patch to G-S to stop using CSD in Unity
[20:38] <ximion> attente: AFAIR it doesn't quit automatically, but you need to explicitly quit it - at least that's the case with GMainLoop classes, and last time I checked GTK+ also didn't add additional magic to that
[20:38] <seb128> robert_ancell, +1
[20:38] <seb128> that's what I was saying yesterday
[20:38] <seb128> attente, that sounds like a desrt's question
[20:38] <seb128> or a gapplication thing
[20:38] <seb128> those "act as a service" things
[20:38] <desrt> we've already been chatting =0
[20:39] <attente> :)
[20:39] <attente> ximion: ok, i'll do the explicit quit then, thanks
[20:40] <seb128> attente, thanks
[20:41] <seb128> I still think that daemon thing is stupid
[20:41] <robert_ancell> seb128, seems like a good way to waste RAM...
[20:41] <seb128> get a gtk process staying around eating memory
[20:41] <seb128> yeah
[20:41] <seb128> if you need notification get a non-UI service or the shell to check out for things
[20:41] <attente> ¯\_(ツ)_/¯
[20:42] <seb128> :-)
[20:42] <seb128> robert_ancell, btw did you notice that yesterday's updated removed the featured/picked softwares?
[20:42] <robert_ancell> seb128, yes, looking into it
[20:44] <seb128> great
[20:44] <seb128> robert_ancell, also any opinion about  https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1551702?
[20:45] <robert_ancell> seb128, That's why I asked about the menu - I had never seen that.
[20:45] <seb128> ah
[20:45] <Fudge> jdstrand:  anything I can do to help with bug #1551351
[20:45] <attente> seb128: just added a commit that calls software-properties-gtk instead
[20:45] <seb128> I've the menu but I never tried it
[20:45] <robert_ancell> Will see if it makes sense to report the apt sources, or just redirect to the standard UI
[20:45] <seb128> before somebody asked about it today
[20:45] <seb128> k
[20:45] <seb128> I don't know how much work that is
[20:45] <seb128> also attente tried in jhbuild and got sources listed
[20:46] <seb128> but those didn't seem to list anything when clicked
[20:46] <seb128> so unsure how much sense the dialog does
[20:46] <seb128> makes
[20:46] <seb128> attente, great ;-)
[20:46] <seb128> robert_ancell, ^ there you go :p
[20:47] <ximion> robert_ancell: FYI, the regular sources dialog has multiple issues, even with the PackageKit backend, and isn't really featureful (check the GNOME BT to get an idea), so I think using software-sources-gtk is a good option
[20:47] <ximion> (just by 2 ct)
[20:47] <ximion> *my
[20:47] <robert_ancell> ok
[20:48] <ximion> I also wanted to do that with GNOME PackageKit, but the PackageKit patches didn't get merged in time (now waiting for aptd to be dropped from Debian)
[21:03] <seb128> robert_ancell, do you know if g-s is supposed to handle local debs?
[21:04] <robert_ancell> seb128, not sure, see bug 1546628
[21:04] <seb128> e.g if that's something supposed to work and to fix or if that's something that you can trigger but shouldn't be able to
[21:04] <robert_ancell> seb128, Dave's bug report seems to indicate you can trigger it
[21:04] <robert_ancell> seb128, bug 1546628
[21:05] <seb128> robert_ancell, right, there is a specific .desktop claiming the mimetype
[21:05] <seb128> attente said it was doing the same in his jhbuild with the packagekit backend
[21:09] <attente> seb128: how do we explicitly deny those permissions in polkit? the one that prevents GS from doing its update thing
[21:10] <attente> org.freedesktop.packagekit.trigger-offline-update
[21:11] <attente> i wonder if something else is using that though and maybe we should just disable it manually in GS
[21:14] <seb128> attente, https://launchpad.net/ubuntu/+source/policykit-desktop-privileges
[21:15] <seb128> I guest you need something along the lines of
[21:15] <seb128> [Disable offline updates]
[21:15] <seb128> Identity=unix-user:*
[21:15] <seb128> Action=org.freedesktop.packagekit.trigger-offline-update
[21:15] <seb128> ResultActive=no
[21:17] <attente> seb128: do you think it's ok to do? does that have some effect somewhere else on the system like update-manager or aptdaemon?
[21:17] <seb128> I think it's ok
[21:17] <ximion> attente: your PK in Ubuntu is 2+y old, so there actually can't be anything using this right now
[21:18] <seb128> offline updates are new and not used anymore yet afaik
[21:18] <ximion> (the feature was introduced later)
[21:18] <attente> ximion: ok, good to know :)
[21:18] <ximion> seb128: technically, we have them for more than a year now ^^ - so not exactly new :P
[21:19] <ximion> but GNOME is the only one using them
[21:19] <ximion> (says Debian's codesearch)
[21:19] <seb128> we should maybe update packagekit
[21:19] <seb128> it doesn't seem like anyone is going to have free cycles to update aptdaemon for the abi changes though
[21:20] <ximion> seb128: I was lobbying robert_ancell a bit, because IMHO updating PK is far less pain than creating a new apt backend for GS
[21:20] <seb128> or api
[21:20] <ximion> the issues so far seem to be click support and $stuff directly using aptd interfaces - although the latter shouldn't be a problem if aptd stops providing PK interfaces
[21:21] <seb128> well "updating pk"
[21:21] <seb128> it means either porting everything away from aptd
[21:21] <seb128> or porting aptdaemon
[21:21] <ximion> or using both side-by-side until everything is ported away from aptd
[21:21] <seb128> they are claiming the same dbus namespace
[21:22] <ximion> in Debian, aptd doesn't provide PK interfaces, and the tools are co-installable
[21:22] <seb128> we would have looked into that
[21:23] <jdstrand> Fudge: I'm not actively working on that one. I fixed the apparmor problem in the other bug. if someone could isolate the cause and the fix for bug 1551351 it would help the foundations team a lot
[21:23] <seb128> but packagekit performances are poor compared to aptdaemon
[21:23] <seb128> or the apt backend to packagekit
[21:23] <seb128> you said that was being worked on but the improvements are not going to be there for the lts
[21:24] <ximion> seb128: quite the contrary, PK is really fast, since it's pure C and C++ - unless you query lots of data, then it becomes slower than aptd
[21:24] <ximion> usually though, apps don't do that
[21:25] <seb128> well you said yourself that the apt backend was doing things in a sync way
[21:25] <ximion> the APT backend in PK could teoretically be optimized to be faster, using parallelization, but that is indeed some work
[21:25] <attente> seb128: there's already a org.freedesktop.packagekit.pkla with that permission, should we modify that instead?
[21:25] <seb128> which was planned to be fixed
[21:25] <seb128> but we are not there yet
[21:25] <seb128> attente, where?
[21:25] <attente> seb128: /var/lib/polkit-1/localauthority/10-vendor.d/org.freedesktop.packagekit.pkla
[21:26] <seb128> attente, I don't have that file on my xenial
[21:26] <seb128> dpkg -S ?
[21:26] <attente> packagekit
[21:26] <seb128> yeah, we don't install that
[21:26] <seb128> so no
[21:26] <attente> ok
[21:26] <ximion> the case for speed optimizations wouldn't be GS though... Ideally, run GS with the PK backend from upstream, I had it working on multiple machines with good performance
[21:26] <seb128> policykit-desktop-privileges is our custom settings
[21:27] <seb128> it should rather be there
[21:27] <ximion> the only thing which the non-paralleluzation can affect is Resolve() calls, which GS caches - so you will likely only notice an effect if you dump all applications GS has on one page
[21:27] <attente> seb128: so what happens if that same entry exists in two files (i.e. someone installs packagekit)
[21:28] <seb128> k, I don't know, but from people who tried the aptd backend perform quite better than the pkgkit backend did
[21:29] <ximion> seb128: calls for a benchmark ^^
[21:29] <seb128> attente, good question, I don't know ... but that wouldn't happen on normal installs unless you specifically install policykit, which is then your issue
[21:29] <ximion> reminds me that I wanted to do a PK benchmark too anyway
[21:29] <ximion> *tool
[21:31] <ximion> so far, from the Debian userbase which uses PK exclusively for everything (including updates), there were never complaints about performance since we switched to the aptcc backend
[21:31] <seb128> in any case we remain on the "if somebody updated aptdaemon to work with pkgkit 1.0 we would unblock that side"
[21:31] <ximion> Debian users rather complain about needing a daemon at all ^^
[21:31] <ximion> (but Pk isn't for those users)
[21:32] <seb128> but so far nobody stepped up for that
[21:32] <ximion> seb128: I assume you just need to drop python3-aptdaemon.pkcompat from the archive
[21:32] <seb128> could be
[21:32] <ximion> that's the only thing claiming its interfaces
[21:33] <ximion> then, only stuff using aptd API directly will use aptd, the rest will use PK
[21:33] <seb128> I guess it requires somebody understanding the stack/issue to open a bug suggesting what to do
[21:33] <seb128> so it can be tested/decided on
[21:33] <seb128> we might be able to co-install those
[21:33] <seb128> but so far nobody suggested that as a working solution
[21:34] <seb128> yeah, well when we updated packagekit previous cycle it got blocked by britney because the update was dropping some enums that were used by aptdaemon
[21:34] <seb128> which made build/some tests failo
[21:34] <seb128> yeah for upstream projects breaking compat...
[21:34] <ximion> right, the aptd compat layer needs to be dropped or updated
[21:35] <seb128> but it's possible that those are only in the pkcompat part of the code
[21:35] <ximion> well, your PK in Ubuntu is more than 2 years old, so...
[21:35] <ximion> okay, you got me, I'll check for that now ^^
[21:35] <seb128> thanks ;-)
[21:36] <ximion> upgrading PK will make Elementary and some other GNOME apps really happy
[21:36] <seb128> yeah, I'm all for updating it
[21:36] <ximion> (and me as well, because I don't get bug reports from prehistoric PK versions anymore ;-) )
[21:37] <seb128> I've been trying to find somebody to look at porting aptdaemon
[21:37] <ximion> whether GS will use it or not is a secondary issue then
[21:37] <seb128> xnox was almost wanting to volunteer but they stepped back ...
[21:37] <seb128> yeah
[21:38] <ximion> at Debian, we ship aptd without any PK shim: https://packages.debian.org/source/sid/aptdaemon
[21:38] <ximion> works well
[21:38] <seb128> I don't like the idea to have an old unsupported version in the LTS
[21:38] <seb128> then what g-s does is the next question
[21:38] <seb128> but Debian installs aptd by default?
[21:39] <ximion> that could use some testing, because I'm really not sure if the performancy impact of sync transactions might be higher on less-powerful machines... It certainly doesn't matter for mine.
[21:39] <ximion> no, Debian installs PK...
[21:39] <ximion> but I thing software-properties pulls it in
[21:41] <ximion> seb128: okay, looks like we really just don't install the PK support shim
[21:41] <ximion> there's no magic or even just a patch involved
[21:42] <ximion> I can compile a package doing that on the Ubuntu version, but the main change would be dropping the pkcompat stuff from the Ubuntu seeds and replacing it with packagekit
[21:43] <ximion> and that's a policy decision ;-)
[21:43] <seb128> that would need testing
[21:43] <seb128> i'm going to try to test that this week
[21:44] <ximion> updating aptd will not be as easy, but its also a possibility - I just wonder for what gain one would do that if aptd is unmaintained anyway
[21:44] <ximion> seb128: awesome - I am running with a similar setup for ages on my Ubuntu box, but it's also running KDE, so maybe not an ideal testcase
[21:45] <seb128> updating aptd, you mean?
[21:45] <ximion> running w/o aptd
[21:47] <ximion> seb128: here's a list of things that will be impacted by this change: http://paste.debian.net/410775/
[21:47] <seb128> ximion, thanks
[21:47] <ximion> I am a bit worried about the click plugin, because there is a 100% chance that this will break
[21:48] <ximion> and click is also not fixable unless someone creates an own service for it (I think I've seen mvo do that)
[21:48] <seb128> yeah, they tried that
[21:49] <seb128> but they don't like the idea to have a new priviledged service
[21:49] <ximion> ubuntu-drivers-common is something one needs to look at, all the rest will be straightforward ports
[21:50] <ximion> and the dependency by gnome-settings-daemon should be dropped when gnome-software takes this over, otherwise we will see some race between the daemons of who triggers an update first and locks the apt cache first
[21:50] <ximion> seb128: that's strange, given that PK already *is* a privileged service...
[21:50] <seb128> did GNOME drop that part?
[21:50] <ximion> yes
[21:50] <ximion> also a long time ago
[21:51] <seb128> well, I guess it's swapping a tested/written service for a new one
[21:51] <ximion> we added it back in Debian when we didn't have GS yet, I guess Ubuntu did the same...
[21:51] <seb128> so they would need to be careful writting it
[21:51] <ximion> that's a fair point, but the risk of running unmaintained old code IMHO balances it out ^^
[21:54]  * ximion things plugin support in PK was dropped too early
[21:54]  * ximion *thinks
[21:55] <seb128> right...
[21:57] <ximion> apparently, the plugins were the no1 cause for crashes of the daemon in Fedora, so Richard was super-eager to get rid of them quickly back in the days, and I couldn't stop him (I also didn't have good arguments, bugs are bugs...)
[22:16] <willcooke> alright, enough for today.  G'night all
[23:12] <robert_ancell> attente, you should move the non review related changes to another branch
[23:12] <attente> robert_ancell: sure, sorry
[23:12] <robert_ancell> attente, I'm rebasing and will do an update into xenial with your changes
[23:13] <attente> robert_ancell: thanks
[23:13] <attente> robert_ancell: i think that one crash with the get_app_folder() is an upstream bug, i'm going to file it now
[23:13] <robert_ancell> cool