ximionrobert_ancell: still there? You should maybe subscribe to appstream@lists.freedesktop.org - hughsie recently proposed his JSON protocol for ratings & reviews for inclusion into the standard, and I would like to have some feedback from other SC implementors00:33
ximiontechnically, you're "just" using GNOME Software, but especially since Ubuntu has it's own Ratings & Reviews, I would welcome feedback on this00:34
ximionI don't want to favor just one distro or add things to the spec which will only be used by very few people (with the rest using alternative implementations)00:35
robert_ancellximion, I will take a look00:59
ximionreminds me, does Laney know that the AppStream mailinglist exists?01:11
* ximion missed the opportunity to make some noise about it in the AppStream release notes01:11
ximionanyway, need sleep, gn801:11
=== maclin1 is now known as maclin
=== meetingology` is now known as meetingology
=== larsu_ is now known as larsu
=== sarnold_ is now known as sarnold
=== popey_ is now known as popey
larsuLaney: tried to drink orange juice, but the straw had a hole09:04
seb128good morning desktopers!09:05
Laneyplug it with your finger!09:05
seb128hey larsu Laney09:05
* larsu contends that he has the most interesting story of the morning09:05
larsuLaney: got a new one ;)09:05
larsubonjour seb128!09:05
Laneyhey seb12809:06
Laneyooh the new paint for the wall ↑ came this morning09:06
Laneythat's an interesting story09:06
Laneycheck it out09:07
larsuhm, true09:07
larsumaybe we should put it up to a vote09:07
larsuI wonder if meetingology can conduct votes?09:07
darkxstLaney, does tha mean you have to paint the wall!09:07
Laneydarkxst: our friend is starting a business doing painting :P09:07
Laneywe are some of the first customers!09:08
darkxstLaney, so you are the guinea pig :)09:08
larsuLaney: nice color!09:10
Laneynot as bright in place unfortunately, the wall it's going on doesn't get direct light ;_;09:11
seb128is that official Ubuntu orange? ;-)09:12
seb128Laney is a true believer in the brand09:12
alexarnaudGood morning everyone !09:12
Laneynew official orange or old official orange?09:12
seb128depend of your screen and color profile :p09:13
seb128we have melting snow this morning, I wonder if that's enough of a story to compete with the orange juice and paintain09:13
larsudepends. how fast is it melting?09:14
darkxstseb128, it would melt way faster here ;)09:14
seb128larsu, now it's turning on "melting before reaching the ground"09:15
seb128but earlier it was sticking a bit on the ground09:16
seb128we would need a few °C less :/09:16
larsuwow... this story is suspenseful! I don't think the paint can compete with that09:16
* Laney mutters09:19
Laneyi'll get you next time09:19
* darkxst trained the chickens to train the dogs, not to eat them! how it that for a story ;009:19
Laneycan train chickens?09:20
Laneychickens can eat dogs?09:20
Laneyso many questions09:20
darkxstapparently and no, dogs eat chickens or ours would!09:20
Trevinhomuktupavels: so... https://code.launchpad.net/~albertsmuktupavels/compiz/gwd-marco-gsettings-v2/+merge/287836 is replacing the old branch from ~ubuntu-mate-dev?09:22
seb128hey Trevinho, how are your holidays going?09:24
Trevinhoseb128: good, but working today :)09:25
Trevinhoseb128: snow was awesome two days ago... I really enjoyed skiing09:26
seb128Trevinho, ok, the "week summary" email said your were off on thursday/friday09:27
Trevinhoseb128: yeah, I changed, since weather isn't great today, so I  prefer to work09:28
seb128k, makes sense09:28
seb128it's good that we have flexibility like that ;-)09:28
TrevinhoYeah... Like a dream :)09:28
TrevinhoAlso today there was the meeting with kylin, so I preferred to stay.09:28
* seb128 doesn't know what to do with that GtkPlacesSidebar ABI change in gtk 3.18 and old nautilus09:32
seb128that's annoying09:33
* seb128 wants snaps to be able to bundle an old gtk :p09:33
darkxstseb128, you could port the 3.18 sidebar widget to old nautilus, or you just don't like the new one?09:34
seb128I'm leaning toward copying the widget from gtk 3.16 in nautilus09:34
seb128but I then need to copy gtkbookmarks&trash as well09:34
seb128I can't easily do that no09:35
TrevinhoLaney: did you see the upstream changes related to https://bugs.launchpad.net/ubuntu/+source/pkg-config/+bug/1523508 ?09:35
ubot5Launchpad bug 1523508 in pkg-config (Ubuntu) "Building xorg-gtest fails with new pkg-config" [High,Triaged]09:35
TrevinhoLaney: not sure it's the case to backport the changes...09:35
seb128the issue is that sidebar changes to be menu based to handle popovers09:35
seb128but the new api is based on gactions09:35
seb128converting nautilus 3.14 is too much work09:36
darkxstand 3.18 nautilus is too unstable?09:37
darkxstit certainly has its issues, but 3.18.5 seems better09:38
tjaaltonseb128: infinity is out so the xserver is still not copied to xenial.. should I just reupload it all so that they'll get rebuilt in -proposed instead?09:38
LaneyTrevinho: can't we take 0.29.1?09:38
LaneyTrevinho: can you test that commit?09:38
TrevinhoLaney: we could take 0.29.1 yes...09:39
TrevinhoLaney: as for compiz side, we handle both cases now, so... It just doesn't matter09:39
Laneystill would be good to know if it fixes the original problem09:39
Trevinhobut, to  avoid other unknown breakages around, I'd go with the old-fashioned way of handling vars09:39
TrevinhoLaney: ok, yes09:40
TrevinhoI can do that09:40
seb128darkxst, we had it and reverted, the zoom levels limitation and some other things make it risky user reception for the lts09:40
seb128tjaalton, hey, sorry I didn't follow ... what's the issue? are you blocked on copy and by what?09:41
darkxstseb128, yes I am aware it was reverted09:41
TrevinhoLaney: however it seems quite likely, since it's a revert...09:41
darkxstseb128, zoom levels don't seem hard to fix, but what else is causing problems?09:42
tjaaltonseb128: the xserver transition, I've been waiting for someone to copy the binaries but infinity is out09:42
seb128tjaalton, you don't have upload rights for it?09:42
tjaaltonI do, but before they've been just copied from the staging ppa09:42
seb128well if you can upload you can copy09:43
seb128I don't understand why you need somebody09:43
seb128using copy-package09:43
tjaaltonnever knew that09:43
LaneyTrevinho: sounds like a revert + looking for quotes, so should be safe but still good to confirm - thanks!09:43
Laneynice that he fixed it09:44
tjaaltonok I have the archive-tools handy.. will give it a go then :)09:44
seb128great, let me know if you need help09:44
seb128tjaalton, should be something around the line of09:45
seb128./copy-package --ppa=canonical-x --ppa-name=x-staging -s xenial --to-primary --to-suite xenial-proposed <name>09:45
tjaaltonyep, seems to work09:48
seb128darkxst, there is a summary of the reasons in09:49
ubot5Launchpad bug 1541954 in nautilus (Ubuntu) "Revert to the previous nautilus version" [High,Fix released]09:49
seb128tjaalton, sorry, I though that you know that whoever can upload can also copy09:49
seb128I didn't really understand why you were holding, I would have told you earlier :-/09:49
tjaaltonheh, no worries09:51
tjaaltonrealized that mlankhorst couldn't copy, not all of these anyway09:51
tjaaltonCandidate copy target: https://api.launchpad.net/devel/ubuntu/+archive/primary09:52
tjaaltonlooks right?09:52
seb128where do you see that?09:53
darkxstseb128, its a balancing act, carlos has been pretty good backporting fixes to the 3.18 branch09:53
seb128I'm unsure to understand the question09:53
tjaaltonCopy candidates: xorg-server 2:1.18.1-1ubuntu3 in xenial09:53
tjaaltonCandidate copy target: https://api.launchpad.net/devel/ubuntu/+archive/primary09:53
darkxstbut some things like zoom arent consider a problem upstream09:54
muktupavelsTrevinho, only if you are ok with removing tests. why they are not enabled when building debian packages?09:54
Laneydarkxst: did you see https://csorianognome.wordpress.com/2016/02/24/nautilus-3-20-last-change-zooms/ ?09:55
darkxstseb128, why is it so hard to add the menu bar back?09:55
seb128darkxst, they are, it's being worked on on e.g https://git.gnome.org/browse/nautilus/log/?h=wip/cosimoc/zoom-levels-v2 and csoriano said he's probably going to need to distro patch something there for the next RHEL update as well09:55
darkxstno I hadnt seen that09:56
seb128tjaalton, looks fine to me, I didn't use it for a while but if you did "--to-primary --to-suite xenial-proposed" it should be right09:57
Laneymost people can't upload to the release pocket directly anyway09:57
Laneyso it'll fail09:57
seb128Laney, but tjaalton can uploaded xorg so it should work no?09:58
darkxstbut why wouldnt you cherry-pick improvement rather than revert09:58
Laneyseb128: yeah I mean the maximum harm of getting it wrong isn't very high09:58
seb128darkxst, because at the time we reverted the improvements were still not done and judged difficult to do09:58
seb128Laney, right09:58
Laneyunless you're in the release team in which case it gets accepted into xenial-release09:58
seb128darkxst, also that is one of the points09:58
Laneyand then you get killed09:58
tjaaltonok, all copied10:00
tjaaltonoh hell10:01
seb128tjaalton, ?10:01
tjaaltonguess I get to do a bunch of rebuilds10:01
tjaaltonI thought this would copy the binaries too10:02
seb128sorry my fault, you need "-b" for that :-/10:02
tjaaltonyeah missed that from the help10:02
darkxstseb128, I dont rrally have time for what if's, but I do have maybe a week before I am back at full-time work10:03
darkxstbut still I have largely no  idea whats needs to be done...10:07
darkxstand yes apparently I can't type while watching TV10:08
willcookesince I'm not doing a very good job of being off sick I might as well log in10:48
=== lool- is now known as lool
tjaaltonseb128: is there a way to artificially keep a package in proposed?10:50
tjaaltonhmm I guess xserver will remain there until rdeps have been rebuilt10:57
tjaaltonwhich is good10:58
Laneyhey willcooke11:01
Laneyare you contagious?11:01
willcookeI think so11:02
willcookeI've put a face mask over my ethernet port11:02
* Laney backs off11:02
davmor2willcooke: confirmed you modem bug11:02
willcookethx davmor211:02
willcookespoke to cyphermox last night, he's going to see what he can do for NM upgrade11:02
willcookedavmor2, if you manually run usb_modeswitch does it "fix" it?11:03
davmor2willcooke: will try that now11:03
davmor2meh that'll be why I have modemmanger removed to be able to flash phones D'oh let me do a fresh install on some hardware and reconfirm11:05
willcookeah ha11:05
davmor2willcooke: infact I can run it from a live session that will be quicker still11:06
davmor2willcooke: so for me in live session I get Mobile Broadband not enabled11:12
willcookedont follow you11:13
willcookeis that good or bad?11:13
davmor2willcooke: so it looks like the switch is happening but not connection is made11:13
willcookedavmor2, oh.  But it detects the modem ok?11:14
davmor2willcooke: well debatable I'll wait for awe and cyphermox to get online and have a debug session from live session then hopefully we can whittle it down a bit for them11:15
willcookedavmor2, nice one, cheers matey11:15
davmor2willcooke: bugger looks like it is biting us in 14.04.4 too11:27
willcookedavmor2, erk11:27
davmor2willcooke: there though I get no connection at all so is likely to be the usb mode switch11:28
* davmor2 is so glad to have all this spare hardware just lying around honest11:29
davmor2oh and it's a contract sim so not like it is out of credit either and if I test it in my mifi device is just connects so definitely and issue.11:30
willcookedavmor2, in fairness, I dont have a SIM in mine, but the HW should still show up (and indeed does when I run usb_modeswitch manually)11:31
desrtword up11:31
willcookehihi desrt11:31
desrtmorning, willcooke11:38
andyrockwillcooke: not sure what to do abouth this card: https://trello.com/c/GiWywEtG/1-work-out-what-to-do-about-desktop-files11:42
andyrockmarco's was handling the bamf/unity part11:42
andyrock*marco was11:42
* andyrock is still sleeping11:42
=== olli_ is now known as olli
=== alan_g is now known as alan_g|lunch
GunnarHjpitti: Do you have time to look at bug #1510198 (FFe)? There are still a few pending details, but it would be good to have a decision regards the FFe before spending more time on it.13:20
ubot5bug 1510198 in openoffice.org-hyphenation (Ubuntu) "[FFe] Sync (almost) with libreoffice-dictionaries in Debian sid" [Medium,Confirmed] https://launchpad.net/bugs/151019813:20
Sweet5hark1Laney: thanks for the update on the seed. two things though ...13:21
Sweet5hark1Laney: 1/ libreoffice-style-human and libreoffice-style-breeze are now _both_ on the image, I dunno why -human is still around. at least in the ubuntu-meta_1349.tar.xz, there is no reference to it.13:23
Sweet5hark1Laney: 2/ the dir in the ubuntu-meta tarball is named ...-1.350 which seems to be off for a 1.349 release13:24
=== hikiko is now known as hikiko|ln
Sweet5hark1ahh, the answer to 1/ is simple 5.1.1~rc2 is still in -proposed and 5.1.0-0ubuntu1 still deps somehow on -human, I guess.13:25
* Sweet5hark1 checks13:25
LaneyAlso the migration happened after today's image13:26
Laney2> dpkg-source doesn't care about the first component13:26
LaneyI ran ./update twice and it renamed the directory the second time13:27
Laneythen I manually changed it back13:27
Laneybut that didn't rename it back :-o13:27
Laneyuse dget :-)13:27
Sweet5hark1Laney: I just looked into it on launchpad, so np, just noticed it to be odd.13:28
Laneynod, good observation!13:30
Sweet5hark1hmmm, libreoffice-gtk 5.1.0 already recommends -breeze. so I wonder what pulls in -human still.13:30
Laneythe old seed13:30
Sweet5hark1Laney: ah, doh.13:30
Laneyand we got breeze like this http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu-touch.xenial/rdepends/libreoffice/libreoffice-style-breeze13:31
Laneydoesn't matter in that case :P13:32
willcookethanks cyphermox :)13:57
cyphermoxgood morning13:57
Sweet5hark1aannd yet another more releasish libreoffice-snap build ... failing after 50 minutes on big bertha ...13:58
willcookeSweet5hark1, https://www.youtube.com/watch?v=coZfzTcv4bA13:59
=== hikiko|ln is now known as hikiko
Sweet5hark1willcooke: oh thats a lot more cute than the stuff that resulted in that machine having the name bertha (https://en.wikipedia.org/wiki/Big_Bertha_%28howitzer%29)14:01
=== jdstrand_ is now known as jdstrand
=== Pici` is now known as Pici
mhall119Laney: so I've been able to generate: https://wiki.ubuntu.com/mhall119/AppStreamTest14:44
mhall119it's a big list, but ranked by popcon installs14:44
* mhall119 is surprised by how many xpm's are still in use14:46
pittiGunnarHj, Sweet5hark1: from an FF point of view, bug 1510198 seems fine to me15:10
ubot5bug 1510198 in openoffice.org-hyphenation (Ubuntu) "[FFe] Sync (almost) with libreoffice-dictionaries in Debian sid" [Medium,Confirmed] https://launchpad.net/bugs/151019815:10
Sweet5hark1pitti: oh, yeah, that one. Im in a call right now, will look afterwards. IIRC it wasnt hugely problematic when I last looked.15:12
GunnarHjpitti: Thanks! Can you please make a note about it on the bug report.15:13
GunnarHjSweet5hark1: Great if you don't have any big doubts. There are still some pending details to deal with. I'm about to upload a new set of proposals to a PPA.15:15
pittiGunnarHj: I did already15:18
GunnarHjpitti: Have seen it now. Thanks again!15:18
Laneymhall119: cool!15:39
Laneyhow are you planning to run this?15:40
mhall119Laney: I will write a blog post detailing the need and the steps to contribute an icon (including working with upstream), then point to a wiki page with that list for people to choose from15:46
mhall119Laney: how often does the AppStream data on http://appstream.ubuntu.com/hints/xenial/main/ get updated?15:46
mhall119oh, wow15:46
mhall119well, I might re-generate the list on a daily basis for a while15:47
mhall119or just have people manually update it when they submit15:47
Laneyyou might want a link to search for already filed bugs15:47
cyphermoxwillcooke: I got NM to build (minus a lot of patches), so I'll get back to it in a few hours, see if I can quickly get nm-applet together15:47
mhall119Launchpad should do that for us15:47
mhall119Laney: since I'm pre-populating the bug title with the "Submit" link, it will match if somebody clicks it a second time15:48
mhall119that won't stop someone from creating a duplicate, but it should at least warn them15:48
Laneyok, then people should be told to file a bug when they start work15:48
mhall119sounds reasonable15:49
mhall119I can add a "claim" column to the page where people can also indicate that they've started on one15:49
Laneythat or the bug assignee, don't know what is easier15:50
* Laney had already started to finish banshee, better delete that15:50
Laneyoh, wiki is 500ing, cool15:50
willcookecyphermox, thanks!!15:50
mhall119yeah, keep trying, it 500's about half the time15:50
Laneyit means I can't log in15:52
flocculantLaney: re g-s, is it expected for it to leave packages behind that need apt-get autoremove ?15:59
Laneyflocculant: I don't know, I'm not really working on the client side16:02
Laneywhat does software-center do?16:02
flocculantleaves configs16:02
flocculantLaney: best to talk to robert ancell later?16:03
Laneyflocculant: yeah or report a bug16:03
flocculantLaney: I'll talk to him - don't want to report something not a bug ...16:04
flocculantif it is then I will :)16:04
Laneyit's easy to close them16:05
flocculantI guess16:07
Laneywell, up to you, either works16:07
seb128flocculant, g-s is not a system upgrader, why would it clean up things?16:11
Laneywhen removing16:12
flocculantseb128: it installed something - I'd expect it to clean up after itself :)16:14
seb128it's a complex problem16:14
seb128you install firefox which brings gtk216:15
seb128then you install inkscape16:15
seb128then you remove firefox16:15
Laneyapt wouldn't mark it as autoremovable then16:15
seb128gtk2 is not linked to firefox which installed it anymore16:15
Laneyif you had some local gtk2 program though16:15
seb128right, I'm just saying that autoremove is different from "that comes from that program"16:16
seb128imho those cleanup jobs are still better handled by the upgrader component16:16
seb128g-s only knows about desktop components16:16
flocculantseb128: so not a bug as such?16:17
seb128well I guess it's valid to discuss16:17
LaneyI think it's debatable16:17
Laneyif one component is doing it then it cannot be more harmful for another one to also16:18
flocculantseb128 Laney - ok I'll report it then16:18
Laneybut not really up to me :)16:18
seb128do we know how other distros/upstream g-s handle that?16:18
seb128also you said that software-center was "leaving configs"16:19
seb128but does it autoremove?16:19
flocculantseb128: yea - or did with the experiment I did16:19
flocculantI'll add as much info as I can to bug anyway16:21
* flocculant makes alias for ubuntu-bug gnome-software ... 16:24
flocculantbug 1552792 for better or worse :)16:32
ubot5bug 1552792 in gnome-software (Ubuntu) "gnome software leaves dependencies installed" [Undecided,New] https://launchpad.net/bugs/155279216:32
=== alan_g|lunch is now known as alan_g
Sweet5hark1snappy build still running, now at ~9GB ...16:50
pittiSweet5hark1: erk16:53
pittithat's 7 times bigger than an entire ubuntu desktop iso which *contains* LibO16:53
ogra_whats 9GB compared to the size of the universe though16:53
pittiogra_: wait until his build finishes :)16:54
* ogra_ is happy that we aim for embedded devices with snappy now ... 16:54
Sweet5hark1pitti: nono no panic, thats the size of the workdir, not the final snap. this is expected to be bigger than the snap will be in the end.16:54
ogra_the ones with a 5TB disk :P16:54
pittiogra_: you don't have a 16TB SD card in your raspi yet?16:55
ogra_pitti, waiting for mediamarkt to seel them for 9,9016:55
* Sweet5hark1 uses df -h as a progress bar :/16:55
* pitti ^5s ogra_16:55
pittiSweet5hark1: hah, nice; what's 100%?16:55
pittilBuild Architecture: amd6416:56
pittiBuild-Space: 1316906016:56
pittiBuild-Time: 2242416:56
pittiDistribution: xenial-proposed16:56
pittiI wonder what unit that is16:56
pittiI suppose kB or KiB, as it's certainly not B :)16:57
pitti13 GB sounds plausible16:57
pittiman, you and your universe-heating packages :)16:57
Sweet5hark1yep. but that is without l10n, I assume?16:57
pittiyeah, I just looked at libo amd64 xenial's build log16:58
Sweet5hark1pitti: lets just say I am generous with distributing entrophy ....16:58
Sweet5hark1meh -- build looks odd though: 100% make, no childs and no output :/17:02
attenteseb128: it's ok to install debs in GS using aptdaemon, right?17:10
seb128attente, yes17:11
seb128isn't that what we use as a backend atm?17:11
attenteseb128: i wasn't sure because i didn't see aptdaemon in the GS depends17:11
attenteseb128: it doesn't look like it from what i can tell17:14
seb128attente, it should probably be a depends, debian/patches/apt-plugin.patch uses it17:14
seb128+result = g_dbus_connection_call_sync (conn,17:15
seb128+      "org.debian.apt",17:15
seb128+      "/org/debian/apt",17:15
attentehrm... interesting. i'm not sure which branch that comes from17:15
seb128I would say17:15
attenteah. ok. i didn't see that17:15
flocculantseb128: sorry - got waylaid - gs/synaptic screeny added now18:05
seb128flocculant, thanks18:08
* Laney woofs goodnight18:31
seb128woot, new gst landing18:33
seb128Laney, enjoy your evening!18:33
Laneystir fry night18:33
Laneygimme a helllllllllll yeah18:33
Laneyyou too! bye!18:33
seb128stir fry, yummy! :-)18:39
* seb128 goes for dinner as well18:39
flocculantseb128: added a new screeny after starting a vm and installing the same package in synaptic and gnome software at the same time18:51
seb128flocculant, thanks18:51
flocculantseb128: sorry if I'm getting boring about this stuff - but I'm trying to test it to distraction for xubuntu ;)18:52
seb128flocculant, no, don't worry, it's useful feedback, please keep playing with it and reporting issues!18:52
flocculantI shall ofc18:53
ximionLaney: when working on the new data generator, I realized that some parts of the current one are a bit more complex than they need to be...19:29
ximionunfortunately the icon finder can't be simplified further :(19:29
* ximion will fetch food, then review all outstanding PRs19:29
robert_ancellattente, I've removed all the non-review related changes in wip/rancell/reviews and put them in wip/ubuntu-changes20:44
attenterobert_ancell: ok, thanks. sorry i wasn't sure if you wanted me to force push those removals or not...20:45
robert_ancellgit revert it fine20:45
attenterobert_ancell: i'm trying to get the deb sideloading to work, but for whatever reason GS refreshes to an empty page after installation20:47
robert_ancellattente, we don't currently reload the apt/dpkg information - could that be the issue?20:47
attentewondering if you have any ideas why it might be doing that...20:47
attenterobert_ancell: it could be, there might be some disparity between the gsapp we generate for the deb and what's added to the db20:49
seb128robert_ancell, hey20:50
robert_ancellseb128, hello20:50
seb128robert_ancell, attente, should we make g-s refresh its index on start?20:51
robert_ancellseb128, it loads the index on start, but it never reloads it. I was going to add a file watch on the files and reload when they change20:51
attenteseb128: i guess we have to refresh the index after the deb is installed20:51
robert_ancellG-S asks for new information after an install. I originally had it loading the index every time, but that seems inefficient20:52
seb128robert_ancell, does it? it doesn't indicate it's doing so and it's not obvious...20:52
robert_ancellseb128, after you install an app, G-S asks for the list of installed apps20:53
seb128g-s seems to lack feedback on when it's doing things/waiting20:53
seb128same if the apt lock is taken by another process20:53
seb128it acts like if it was doing work20:54
=== shiznix_ is now known as shiznix
robert_ancellseb128, there is an issue with the theming where the loading bar is not shown20:54
seb128yeah, I saw that one. but even if the bar was loading, if something else is having a lock on apt you can wait for ever20:56
robert_ancellseb128, yes, but there's nothing we can really do about that...20:56
robert_ancellThat's just apt being a bit shitty20:56
seb128software-center/update-manager handle that better, they tell you they are waiting because something has a lock on the db20:58
seb128so it's doable20:59
seb128but it might be hackish, no idea20:59
attentehow does s-c detect the lock?20:59
robert_ancellattente, aptd gives status information, it's probably in there20:59
robert_ancellseb128, file a bug for a label as to why it's waiting!21:00
attentevia the dbus interface?21:00
robert_ancellattente, yeah21:00
seb128I should do that ;-)21:00
robert_ancellattente, see transaction_property_changed21:00
robert_ancellattente, see transaction_property_changed_cb rather21:00
robert_ancellattente, your new gs_plugin_filename_to_app doesn't check if the number of tokens from the strsplit is what you expect...21:01
robert_ancelland your g_spawn_sync can return FALSE but doesn't set error...21:02
attenterobert_ancell: yeah, i got a bit sidetracked with getting the app page to refresh properly..21:03
attentethanks for catching those21:03
robert_ancelldesrt, has anyone ever proposed a g_strv_new, i.e. gchar **argv = g_strv_new ("foo", "--bar", "3", NULL);21:05
robert_ancellI feel like I'm writing that sort of code all the time21:06
desrta common trick is to use split21:06
robert_ancelldesrt, yeah, but that doesn't work well if the contents are variables21:06
robert_ancellie..  gchar **argv = g_strv_new ("foo", "--bar", some_user_value, NULL)21:07
desrtwhy aren't you using gsubprocess?  :)21:07
robert_ancelldesrt, is it worth proposing or is it the dreaded glib feature creep?21:07
robert_ancellI'm super glad g_strv_contains exists now21:07
robert_ancellWas sick of writing that over and over...21:08
desrtya.. me too21:08
seb128robert_ancell, is bug #1552074 one of the issues you fixed? the git commit have no bug reference/bt so not easy to say21:08
ubot5bug 1552074 in gnome-software (Ubuntu) "gnome-software crashed with SIGSEGV in gtk_stack_set_visible_child_name()" [Medium,Confirmed] https://launchpad.net/bugs/155207421:08
robert_ancellseb128, I'm looking at that now, not fixed AFAIK21:08
desrtseriously, though21:08
robert_ancelldesrt, sure, in that case, but there are others.21:09
robert_ancellPeople end up using GPtrArray instead, which is a bit heavy21:09
=== thumper is now known as thumper-afk
desrtor strv = g_new(char*, 4); strv[0] = ...; etc.21:10
desrtwhich is kinda reasonable imho21:10
desrtbut g_strv_new() is also kinda reasonable21:10
robert_ancellwell, I'd say it's leess than ideal21:10
desrti'd be happy to review a patch21:10
desrtbut it's too late for this cycle21:10
robert_ancellthat's fine21:10
robert_ancellit's not urgent, but it makes GLib easier to use21:11
desrtimho it's not really great for this case21:11
desrtsince it will necessarily dup the input strings21:11
desrtbut meh21:11
desrthint: try to copy the valist to avoid doing multiple allocations21:11
desrtie: va_copy21:11
desrti guess you could do some speculation and alloca() as well21:13
seb128robert_ancell, https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1552918 seems to be something quite some users hit21:13
ubot5Launchpad bug 1552918 in gnome-software (Ubuntu) "/usr/bin/gnome-software:11:g_menu_model_get_n_items:gtk_application_window_update_shell_shows_app_menu:gtk_application_window_real_realize:_g_closure_invoke_va:g_signal_emit_valist" [Undecided,New]21:13
desrtbut in terms of how it will end up looking in assembly, the copy is definitely cleanest21:13
desrt(and just to avoid roundtrips, don't forget): docs section addition, sentinal attribute21:15
desrtyou'll probably also get tripped up on the missing version macros.  feel free to submit a patch to add those =)21:15
robert_ancelldesrt, missing version macros?21:16
robert_ancelli.e. defines for the latest version?21:16
desrtsince you're writing the patch against the not-yet-existent series-to-come...21:17
attenteis that the LIM/headerbar bug?21:22
ximionseb128: btw, I found out why PK was percieved as slow: When it was tested initially, Ubuntu didn't have AppStream metadata, and in that case GS emits FindFiles() calls to PK, which cause a search for which pkg provides a specific filename - and those searches are super slow, even when using APT directly.21:35
ximionwith AppStream metadata present, GS has all the package information and only rarely needs to resolve a filename to a package. For packages, it will only emit a Resolve() call, and resolve calls are really cheap21:36
seb128ximion, hey, k, so maybe that's a non issue then21:39
ximionseb128: I ran into this myself yesterday on a Debian machine which had broken metadata21:39
ximionwas an enlightening experience on that matter21:39
=== marga- is now known as marga
seb128ximion, what's the status of the appstream-glib issue/screenshots? seems like the bug is without activity since december :-/21:44
ximionseb128: hughsie won't implement it, so someone else would need to do it21:44
seb128what's the difference which makes it work on fedora?21:45
ximionI wanted to quickly submit a match, but making this work requires passing around the media_baseurl parameter down to the parsing functions, which is a more invasive change21:45
ximionso I wanted to have hughsie take a look at it first21:45
ximionseb128: fedora is using the XML data, which doesn't know MediaBaseUrl and just duplicated the full url for every screenshot21:46
ximionlibappstream implements the DEP-11 spec correctly, so the only SC which is currently affected by this bug is GNOME Software (and potentially other stuff using libappstream-glib, but the only bug report I got was from GS so far)21:47
ximionseb128: btw, technically implementing MediaBaseUrl isn't hard21:49
ximionthe "Origin" field in AppStream metadata works the same way21:49
seb128I don't understand those details enough to get a clear idea of what's going on exactly21:50
seb128but we need to get that fixed this cycle one way or another21:50
seb128what do you recommend doing?21:50
seb128is that something robert_ancell_ or attente or Laney are looking at?21:51
robert_ancell_seb128, I'm aware of it, was hoping Laney was looking at it21:51
ximionseb128: definitely fix appstream-glib to follow the spec ;-)21:51
seb128that works if something is assigned/going to do it21:52
robert_ancell_ximion, is there a bug with the information? We can carry a patch in Ubuntu if necessary21:52
seb128which doesn't seem the case atm21:52
ximiontheoretically we could make the generator thow out some different version of the data, but that would have some annoying issues on the server side, e.g. we couldn't easily change the url of the screenshots and icons anymore21:52
seb128could we just do whatever fedora is doing?21:52
* ximion doesn't like that21:52
ximionrobert_ancell_: no, the data is fine - just https://github.com/hughsie/appstream-glib/issues/70 needs to be fixed21:53
robert_ancell_ximion, ok, so it just needs someone to write the patch and convince hugshie this is the right thing to do?21:54
ximionspec here: https://www.freedesktop.org/software/appstream/docs/sect-AppStream-DEP11.html#spec-dep11-general21:55
ximionrobert_ancell_: jup - while convincing hughsie would not be hard :)21:55
ximion(I hope ^^)21:55
robert_ancell_ximion, oh regarding reviews in appstream. That doesn't make sense to me either - they seem too dynamic to be mixed in..21:55
robert_ancell_And there could be many thousands of reviews21:55
ximionrobert_ancell_: jup, I don't want that - but specifying the protocol to get them in AppStream is something I would like21:56
ximionjup, I am not sure if I misunderstood hughsie there21:56
robert_ancell_ximion, that makes sense I guess21:56
ximionwe already have services in the spec: https://www.freedesktop.org/software/appstream/docs/chap-AppStream-Services.html21:57
ximionbecause before hughsie introduced metainfo / appdata files, it was thought that every distro would simply implement the Debian screenshot server API21:57
ximionand OpenSUSE even did that :)21:57
ximionnow, the screenshot service is only a fallback21:58
ximionrobert_ancell_: on the generator side, having MediaBaseUrl allows us to store the pre-generated metadata in a database and easily change the location of the screenshots server without regenerating all metadata or parsing all YAML and string-replacing the old url with the new one when altering it22:00
robert_ancell_yeah, that makes sense if you want to mirror it I guess?22:01
ximionalso, this could later be used by distributions to have multiple mirrors for screenshots, or add a mirror pattern22:01
ximionsome people expressed interest in this, because not everyone has a powerful server or redirect service running for this22:02
ximionjup, exactly22:02
robert_ancell_desrt, could you have a quick look at https://errors.ubuntu.com/problem/23ba6db9a4c30881ade44cef107a6dd5d76a41ef. Essentially, inside GApplication a GTK settings callback is trying to access app_menu_section, which is NULL even though it shoul be defined for the lifetime of the application.22:03
robert_ancell_Was wondering if perhaps the callback has been called after dispose() somehow22:04
robert_ancell_GtkApplication rather22:04
=== thumper-afk is now known as thumper
desrtsounds likely22:11
desrtrobert_ancell_: the app_menu_section here is an attribute of the window itself -- not the app22:22
desrtit is created in _init() of the window and freed in dispose22:23
desrtso this looks like someone is trying to realize a window after it has already been disposed22:23
desrtvia gtk_widget_show22:24
desrtbug is in whatever defines gs_application_activate() i'd say22:25
desrti'm guessing it holds a ref on the window so that it can show it again, but it doesn't prevent the window from being destroyed on delete22:26

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!