[08:00] Saviq: ping [08:00] gusch, pong [08:00] Saviq: what size should the app icons be, and where should they be installed? [08:01] Saviq: because the icons are not square ... [08:01] gusch, yeah, they are not square in the new Unity, either [08:03] gusch, i.e. they're not meant to be square [08:03] Saviq: so installing to 256x256 or 128x128 is not very logical ... [08:03] gusch, put them in /usr/share/icons/{64,128,256...} [08:03] Saviq: there are some in "128", but they are 144x134 [08:04] gusch, I'm not really sure what's the guideline here :/ [08:04] gusch, but I'd render them to 128x*, preserving aspect width [08:05] and put in the respective dirs [08:05] Saviq: ok - I'll do that then [08:18] sil2100: compiz sponsored on quantal, sorry it was so long, but other priorities :) Thanks! [08:18] popey: FYI ^ [08:19] \o/ [08:20] thanks didrocks [08:21] Saviq: so what approach do you want to take with the todo/fixme reduction? i see kgunn created a document for it but it's mostly skeleton so not sure how to proceed here [08:22] tsdgeos, I think we should go through them and identify those that can be gotten rid of now [08:22] tsdgeos, with little effort (and only those that won't go away soon anyway, like the bunch in Dash*.qml) [08:23] tsdgeos, and put your findings in the doc with some convention [08:23] to mark those that we don't care about and those that should be fixed sooner than later [08:24] tsdgeos, that sound sane? [08:24] Saviq: so first go through them all, try to annotate into easy/hard+reason and then once we have that go through the list and fix the easy ones? [08:25] tsdgeos, that sounds perfect [08:25] okidoki [08:25] might take a good while :D [08:25] tsdgeos, we'll then go through the remaining ones and sprout work items [08:25] tsdgeos, yeah, I was surprised by the amount of them... [08:26] didrocks: \o/ Tghanks! [08:26] sil2100: thanks to you! :) [08:26] It was a busy period, so no worries ;) [08:30] didrocks, hey [08:31] salut davidcalle! [08:31] didrocks, lp projects are here, now, we have to move branches [08:32] davidcalle: yeah! sounds a good plan, you are planning to do it today? [08:32] davidcalle: when doing that, we need to change the daily release configuration files [08:32] and redeploy [08:34] didrocks, I'll do it in a moment [08:36] great ;) [08:37] Saviq: so did you agree with ted in a way to move forward the hudclient-2 thing? anyone that uses the ./build -s thing now gets something that doesn't build [08:37] * davidcalle branches and pushes [08:37] tsdgeos, no I didn't, it's not built anywhere [08:37] tsdgeos, that I know [08:38] right [08:38] so we need to get it build asap [08:38] right? [08:38] tsdgeos, well, trunk is fine until new hud is built [08:39] Saviq: not if you use ./build_unity -u, no? [08:39] that'll get you a hud that gives you hudclient2 instead of hudclient1 [08:39] tsdgeos, fine in the CI / autolanding sense [08:39] sure [08:39] maybe we want to change build so it gives us the last rev that provides hudclient1? [08:40] i mean people that read http://www.omgubuntu.co.uk/2013/04/how-to-run-unity-next-on-your-desktop are going to get a failure :D [08:40] hmm that's a good idea [08:40] to have a "safe_rev" numbers in build [08:40] and pull those in [08:40] and update them as needed [08:41] tsdgeos, will do [08:41] didrocks, shouldn't we create a project group to have a place where we can have an overview of all our scopes? [08:42] Saviq: great :) [08:42] (branches are here) [08:43] davidcalle: yeah, sounds like a good plan :) [08:43] didrocks, I don't have rights to do that, do you mind making one? [08:44] davidcalle: I think you need to ping on #launchpad, I don't as well [08:44] didrocks, ok. Any preference for the name? Ubuntu-scopes, unity-scopes? [08:46] davidcalle: unity-scopes makes sense to me :) [08:46] didrocks, sounds good to me [08:46] didrocks: Compiz project cleanup is now *done*. Feel free to hide/freeze/whatever lp:compiz/raring [08:47] Hmm and "trunk" ? [08:47] duflu: lp:compiz/raring can be removed [08:48] duflu: I already point to lp:compiz/0.9.9 since tonight [08:48] didrocks: Cool. Though I don't have permissions to mess with the "raring" series [08:48] Everything's so fine-grained now you can't always tell who does [08:50] duflu: I don't know as well [08:51] Saviq: getting error in build setup while sshing into device after flashing device. [08:51] ssh_exchange_identification: Connection closed by remote host [08:51] * sil2100 somehow missed the switch from lp:compiz/raring to lp:compiz/0.9.9 [08:52] dednick, can you ssh to the device? [08:52] sil2100: Twas not announced. I just did it yesterday/today [08:52] Saviq: no. [08:52] dednick, did you ./run_on_device -s first? [08:52] duflu: ah, ok ;) Anyway, rebasing the merge you pointed out a few hours ago [08:52] dednick, and did you enable network even before that? [08:53] Saviq: yes. installed the ssh keys, but code resync causes the error. [08:53] sil2100: Sorry, I did ask lots of people. Forgot you. [08:53] dednick, you need to be able to ssh to the device [08:53] dednick, `ssh phablet@127.0.0.1 -p 2222` should work, IIRC [08:54] Saviq: it doesnt. ssh_exchange_identification: Connection closed by remote host [08:54] dednick, `adb shell`, `ubuntu_chroot shell` [08:54] dednick, you're in Ubuntu, make sure ssh is installed and all [08:55] dednick, you can also use `phablet-network-setup -i` to setup ssh and there's some more info spat out, too [08:59] Saviq: yeah, that was it. flashed it and didnt realise ssh wasnt installed [09:00] dednick, ./run_on_device -s installs ssh [09:00] dednick, but you need network first, so maybe that was the issue [09:02] Saviq: i forgot that flashing the device would remove the wifi access [09:04] lol, kgunn added [09:04] g_action_group_activate_action(ag, action.toUtf8().constData(), g_variant_new_double(value.toDouble())); [09:04] to the TODO list of things to fix [09:04] i was wondering why [09:04] then realized it [09:04] value."toDo"uble() [09:05] lol [09:10] Heh. Try egrep '\' instead [09:10] duflu: resubmitted both branches, since the previous one had a conflict due to the branch switch - but it now should be ok ;) [09:11] sil2100: OK, but I'm not offering to re-join the compiz effort. Although I would often like to, I am focussed on Mir. That was a once-off project cleanup to keep everyone on the right track [09:14] Even updating all the bugs I got a predictable number of emails in response... "can you help me with Compiz?" [10:08] dednick, can you do https://code.launchpad.net/~saviq/unity/phablet.protect-local-build-revisions/+merge/156793 please? [10:09] Saviq: sure [10:10] dednick, pushed a whitespace fix [10:12] Saviq: does this fix a current build issue with the hud? mhr3 was just having issues building with trunk unity/phablet. [10:13] Saviq: but his issue may have to do with something else. [10:14] dednick, yes it does [10:14] dednick, HUD bumped API version in trunk, but didn't yet build packages [10:14] dednick, so we have MPs to fix it, but can't yet merge into trunk as it won't build [10:15] mhr3: can you try 'build -s' with with lp:~saviq/unity/phablet.protect-local-build-revisions [10:15] on it [10:15] mhr3, if your error was related to hud-client-2 missing, that will fix [10:15] mhr3, we got out of sync with HUD API versions, the usual, you know... [10:15] Saviq, nope, it was about test-voice failing to link [10:16] mhr3, so failure in HUD itself? [10:16] yea [10:16] mhr3, ok, please ping if it doesn't go away [10:16] k [10:22] Saviq: new script works for me. +1 [10:22] dednick, cheers === ubot5` is now known as ubot5 [10:24] Saviq: with "move to persistent storage" you mean saving the searches (and results?) on disk? [10:24] tsdgeos, yes [10:25] ok [10:25] tsdgeos, not results [10:25] tsdgeos, just search terms [10:25] ok [10:25] so we can hint them next time? [10:25] tsdgeos, yes [10:25] tsdgeos, we should probably interface with Zeitgeist there [10:25] ok [10:36] Saviq, your branch didn't help me, disabled the voice tests and now it builds (i guess it's because i'm building on Q) [10:37] mhr3, might be, will test on Q [10:37] mhr3, thanks === mmrazik is now known as mmrazik|lunch [10:41] Saviq, ideas what /home/miso/projects/phablet/main.cpp:64:11: error: ‘class QQuickView’ has no member named ‘setTitle’ [10:41] is about? [10:41] same for setFlags [10:41] mhr3, are you using ppa:canonical-qt5-edgers/qt5-proper? [10:41] yep [10:41] * mhr3 double checks [10:42] mhr3, that was an API change before Qt5 release [10:42] mhr3, so please make sure you upgrade from that ppa [10:42] Saviq, hmm, using it and nothing to upgrade :/ [10:43] mhr3, http://qt-project.org/doc/qt-5.0/qtgui/qwindow.html#title-prop http://qt-project.org/doc/qt-5.0/qtgui/qwindow.html#flags-prop [10:43] mhr3, they're definitely there [10:44] mhr3, QQuickView inherits QWindow, btw [10:44] mhr3, /me tries Q [10:45] yea, i see it in /usr/include, guess it's picking up something older from somewhere [10:45] the question is what from where :) [10:46] mhr3, /opt/qt5? [10:46] Saviq, indeed [10:46] mhr3, make sure you have all stuff from there dropped [10:46] mhr3, and that your PATH does not include it anymore === alan_g is now known as alan_g|afk [10:47] removing everything that installed things in there [10:53] hmm. update manager has frozen on me [10:58] popey, cf #ubuntu-desktop, is that the same issue? === alan_g|afk is now known as alan_g| === alan_g| is now known as alan_g [10:59] yes, thanks [10:59] popey, did you kill stuff yet? [10:59] if not hold on [11:02] showRemoteDetails = true [11:02] showRemoteDetails = (frame.height - column.minimumHeight) >= labelRemotePostTime.height [11:02] wops :D [11:03] Saviq, yey, it all built! :) [11:03] mhr3, :) === MacSlow is now known as MacSlow|lunch === mmrazik|lunch is now known as mmrazik [11:31] Saviq, shall I proceed with components testing? [11:31] mzanetti, what can I test of PageHeader? [11:31] Cimi, mzanetti is out until Friday [11:31] Cimi, and nic-doffay was working on PageHeader tests [11:32] ok [11:33] Saviq, those responsive things then [11:33] grid and flow [11:33] Cimi, aren't there tests for them already? [11:34] Cimi, not for the flow there are not [11:34] Cimi, what's the bottom bar status then? [11:35] Saviq, yes they are already [11:35] not for flow though [11:35] Saviq, status... I thought I had to wait for SDK? [11:35] Saviq, however [11:35] Cimi, that's a status, sure, but would be good to get an update [11:35] Saviq, I might add my findings [11:36] Cimi, I saw they were chatting about it later yesterday [11:36] Cimi, so please find out what's the status on their side, and if needed we'll merge your stuff with a huge "FIXME: this should come from SDK" [11:37] Saviq, ok [11:38] Saviq, ping [11:39] Cimi, and responsive flow does not have tests [11:39] tvoss, pong [11:39] Saviq, https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1163395/comments/1 [11:39] Launchpad bug 1163395 in Ubuntu UI Toolkit "Provide abstract component for edge swipe" [Undecided,Confirmed] === alan_g is now known as alan_g|lunch [12:25] Saviq: have you seen th update? https://code.launchpad.net/~schwann/gallery-app/gallery-app-app-icon/+merge/156659 [12:26] gusch, otp, will look in a bit === MacSlow|lunch is now known as MacSlow [12:51] didrocks, good afternoon [12:52] fginther: hey! nice work yesterday :) [12:52] I saw you removed the --pin-ppa, this is not supported? [12:53] didrocks, thanks, the results could have been better :-( [12:53] oh? [12:53] --pin-ppa requres a ppa argument [12:53] ah [12:53] which wasn't passed in any of the exisiting jobs that I found [12:53] ok, well, it's the same ppa that we should ping [12:53] pin* [12:53] so easy to modify [12:54] fginther: what else did fail? [12:54] didrocks, so the --enable-ppa and --pin-ppa should use the same ppa? [12:54] fginther: right [12:55] didrocks, I was hoping for better test results :-) === alan_g|lunch is now known as alan_g [12:55] ah ok ;) === _salem is now known as salem_ [12:55] fginther: there is one more modification we need to do [12:55] fginther: we need to add a --test-package [12:55] or rather [12:55] --test-packages [12:55] we will pass unity-autopilot for unity* tests [12:55] foo-autopilot bar-autopilot for others [12:56] didrocks, ack [12:57] fginther: I'm doing the change in the daily release side to pass the "testpackages" parameter to your job [12:57] didrocks, wouldn't we just add that to 'packages'? [12:57] fginther: no, because even if we use the full "dist-upgrade from ppa" [12:57] this install needs to be done independently [12:57] and not filtered [12:57] ok [12:57] (the "packages" are redirected in a file and filtered) [13:35] fginther, who do I bring utah problems to? [13:36] mterry, sent you a message [13:43] didrocks, the testpackages changes are ready [13:43] fginther: doing my side, was debugging something else :) [13:43] fginther: once sec! [13:43] didrocks, no problem, also fixed the missing --pin-ppa [13:43] \o/ [13:45] didrocks, I have a 7.0.0 unity release ready (using 'make dist') with updated AUTHORS and ChangeLog file and release tag (no code changes)... should I push that directly to lp:unity? [13:46] fginther: http://bazaar.launchpad.net/~cupstream2distro-maintainers/cupstream2distro-config/trunk/files/140?start_revid=140 [13:47] bregma: you can get it under merge and autoprooving it [13:47] does the automerger handle tags? [13:48] bregma: the upstream merger? I doubt it though [13:48] fginther: ? ^ [13:49] didrocks, bregma I think the answer is no. the automerger just does a bzr merge and a bzr push [13:49] bregma: if you are sure nothing else is merger and that your branch is building in a pbuilder, you can directly push :) [13:49] merging* [13:50] fginther: ok, deployed on raring oif/indicator/unity [13:50] didrocks, excellent [13:50] fginther: do you want to try on something? to ensure the right parameters are sent? like oif? [13:50] didrocks, yes [13:50] what do you mean by "nothing else is merging"... as in there is no automerge in progress, or just no other changes in my merge? [13:50] bregma: no automerge in progress [13:51] bregma, unity is not currently building on jenkins [13:51] it's a maze of twisty little passages, all different [13:52] I will push right now, then upload the tarball, and wheee [14:01] MacSlow, just a quick note - have a look at BorderImage for the value bar in a confirmation notification [14:02] MacSlow, http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-borderimage.html [14:16] fginther: ok, I'm moving the head job to this format [14:16] fginther: and the 100scope ppa [14:16] sounds good to you? [14:19] didrocks, yes, I think so. I noticed a typo in the job, but it should be an easy fix [14:19] tedg, hey, what's the status of HUD packages? [14:19] didrocks, I'll rerun the oif stack after fixing it to make sure [14:19] fginther: ok, the utah-jenkins change seems good to me FYI :) [14:20] didrocks, thanks for looking === alan_g is now known as alan_g|tea [14:20] yw :) === mmrazik is now known as mmrazik|afk [14:23] Hi all, Want to get some comments for this MR. And would like to know if this unittest sounds. https://code.launchpad.net/~paulliu/unity/phablet-add_unit_test/+merge/156859 [14:28] Saviq, Not sure, alesage doesn't seem to be in yet. He was working on it last night, I don't see packages anywhere, but I'm unsure of the status. [14:28] paulliu, please make sure you update https://code.launchpad.net/~paulliu/share-app/desktop_file_tweak/+merge/156560 [14:28] Saviq: yeah..you want it released, right? [14:29] paulliu, yes, but we need the two more hints in there [14:29] Saviq: ok. [14:29] Saviq: wait. [14:30] Saviq, hm... ok... was initially searching for something else... but that might work as well... just need to think of a way to "port" the original cairo drawing-code to an image to be used as source === mmrazik|afk is now known as mmrazik [14:30] MacSlow, well, you can use a rectangle with a border, but if I see correctly there's a gradient? [14:31] Saviq, yes... there's a subtle but important gradient (to give it a bit of depth) [14:32] MacSlow, yeah, so I'd go for BorderImage and stretched Image for the filling [14:32] Saviq, but I've an idea already how this could work with such a BorderImage item [14:33] Saviq, I'll see to fix the remaining issues with the other branch by tomorrow. [14:33] MacSlow, k [14:37] paulliu, can you also do: [14:37] - in webapps-demo package: /usr/share/applications/music-player-mockapp.desktop should have "Music" as Name instead of "Music Player" [14:37] - in webapps-demo package: remove fake notepad app, especially file notepad-mockapp.desktop [14:40] ok. [14:40] Saviq: wait. [14:40] paulliu, waiting ;) [14:49] fginther: so pushed the modification on head + experimental [14:49] didrocks, does that take care of all the stacks now? [14:49] mterry: you now have some example for the phablet stacks on head ^ [14:50] fginther: what do you mean? [14:50] didrocks, I mean all all of the check jobs now using the generic autopilot job? [14:50] fginther: the stack was different set of packages and ppas, I guess we addressed both [14:50] fginther: yeah [14:50] fginther: let's wait for tomorrow [14:50] before killing the others :) [14:51] didrocks, thanks [14:52] didrocks, why are there both packages and testpackages fields? [14:53] mterry: packages are the packages we need to install by default on the system (the binary packages generated by this stack) [14:53] mterry: when we install them, we filter that we only install those [14:53] and don't pull anything else [14:53] (or we fail the job on puropose) [14:54] purpose* [14:56] didrocks, so I need to list all dependencies too then, not just the top packages? [14:56] mterry: all the binary packages that we are going to install on this stack (and their deps if they are not installed by default) [14:57] mterry: basically, we filter to avoid the following case: [14:57] indicator built with bamf3 [14:57] procude bamf 3.1 [14:57] produce* [14:57] then we build unity [14:57] it will grab 3.1 [14:57] but some of the stack will fail [14:57] and we don't know that we need to publish both at the same time :) [14:58] Saviq, with the BorderImage the bar looks pretty clean/correct now... created two SVGs based on the old cairo-code [14:58] MacSlow, render them to PNGs [14:58] MacSlow, it's cheaper on the GPU [14:59] Saviq, ok [14:59] Saviq: https://code.launchpad.net/~paulliu/webapps-demo/desktop_file_tweak2/+merge/156876 [15:00] paulliu, cheers === dandrader is now known as dandrader|afk [15:03] didrocks, well, for a start, I'm just adding platform-api to the platform stack. (my other autopilot tests will want that installable) [15:03] mterry: yeah, and so add the files for the autopilot generic job [15:04] mterry: and the autopilot job that needs to be started [15:04] didrocks, hmm? platform-api doesn't have autopilot jobs [15:05] mterry: does it need some? [15:05] didrocks, when will the head stack build again? [15:05] mterry: when you decoment #schedule: :-) [15:05] uncomment* [15:05] didrocks, it's xnox's package to shepherd, I'm not sure. But I'd guess it's not a gui thing [15:06] mterry: do we have integration tests on it? [15:06] mterry: is xnox working on finishing that? [15:06] didrocks, it's got a tests directory... [15:06] I don't want that we daily release something before meeting our criterias [15:06] didrocks, he merged a packaging branch to it. xnox, is platform-api ready for daily release? [15:07] xnox: if so, please file things in the google doc I shared with you: https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AuDk72Lpx8U5dFVVX1BOYm1qdUtyX2xUNmdwWlhTS0E#gid=0 [15:07] this is done for that :) [15:07] (and we need yes everywhere or a no should be justified) [15:07] didrocks, this is just to a PPA, so while I agree we don't want to release crap, it's less of a quality expectation [15:08] mterry: well, we should take the PPA as distro [15:08] mterry: I know how it goes then, and we'll have to flip the switch without finishing that [15:08] so we have a one-time opportunity to get things right [15:08] didrocks, :) [15:09] mterry: and remember we will release the iso from the ppa first :) === alan_g|tea is now known as alan_g [15:12] tsdgeos: hey, bregma said his team is available as soon as 13.04 is wrapped up...and we thot fixme/todo's might be a good spot to help [15:12] mterry: didrocks: ack, will add to spreadsheet. Not sure there are any tests for it, as it's a wrapper api for different plug implementations, with the only currently implemented one is for libhybris, which does not have tests yet per se. [15:13] mterry: didrocks: autopilot style tests should be done on things that build on top of platform-api, imho. [15:13] tsdgeos: he was just wanting a preview of what's to come...are you using that spreadsheet i made [15:13] ? [15:13] xnox: as long as we are clear on the state and "no test" is acceptable in your opinion :) [15:13] xnox: yeah, I do agree, did you talk about it with upstream? [15:13] xnox, is src/android/tests/ useful? [15:13] kgunn: yep [15:13] or does that need android to run? [15:14] didrocks: did not "talk to upstream about it". I should discuss if it's feasible to test it more, somehow. [15:14] mterry: will double check those. [15:14] xnox: yeah, that's part of the job for the WI ;) [15:14] =D [15:14] tsdgeos: cool..i'll just share that with him...wanted to make sure it'd be relevant :) [15:14] xnox: so please once you talk with them, file the rationale for the "Not yet" and what will be done ;) [15:15] xnox: if the cells are not enough, we can create a doc or a wiki [15:15] kgunn: the TODO part is "done", working on the FIXME one [15:15] kgunn: i've tried to be as through as possible on the evaluation, but some of them seemed like "i'd have to solve it to see if it's a 5 minute or 5 days job" [15:16] then mostly went to the "5 day" side and be happy if it took you 5 minute :D [15:16] tsdgeos: sure...totally understandable [15:17] xnox: so, to sum up, for the "no", you are working on them, right? ;) [15:17] (as you didn't mark TODO everywhere) ;) [15:18] like "missing libhybris" should be "upload libhybris to the ppa" ;) [15:18] thanks xnox ;) === dandrader|afk is now known as dandrader [15:18] xnox: you are doing libhybris, isn't it? [15:18] or is it sergio? [15:19] didrocks: I really really prefer autolanding libhybris somehow =) even in the sense that it's not a "true" auto-land. [15:19] didrocks: generally package maintainance this way is awesome =) [15:19] didrocks: I don't think I was on the hook for libhybris. [15:19] xnox: let me look at the blueprints, one sec :) [15:19] xnox: the tradeoff for autolanding is that we need good tests ;) [15:20] or it's the far far far west ;) [15:21] xnox: https://blueprints.launchpad.net/ubuntu/+spec/client-1303-ubuntu-touch-porting is telling that it's rsalveti [15:21] xnox: do you mind checking with him and take the issue to get it daily landing if needed? [15:22] as this is what is blocking everything, from what I'm hearing, mostly [15:22] xnox: we can have some "on daily releasing demand" for components we are not upstream for [15:22] * didrocks should document the procedure [15:26] paulliu, please release https://code.launchpad.net/~paulliu/unity-lens-applications/phablet/+merge/155464 too [15:37] Saviq: ok === dandrader is now known as dandrader|lunch [16:38] hey bschaefer! [16:39] didrocks, hello! [16:39] bschaefer: so I still don't get what you mean ;) what exactly has an ABI break? :) [16:39] bschaefer: do you have a stack trace? when is it crashing? [16:39] seb128: FYI ^ [16:39] didrocks, when running the unity tests yeah [16:39] * bschaefer gets log [16:39] bschaefer: we are running them everyday :) [16:39] didrocks, and looking at the changelog libgcc1 was updated march 28th [16:40] or are those tests not run during build? [16:40] didrocks, the ones that aren't built [16:40] https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1162886 [16:40] Launchpad bug 1162886 in unity (Ubuntu) "test-gtest [ FAILED ] TestIconLoader.TestGetManyIcons - segfaults at times." [Critical,Triaged] [16:40] bschaefer: do you mind checking with doko? I doubt it only impact us if it's libgcc1 [16:40] didrocks, stacktrace: http://paste.ubuntu.com/5668678/ [16:40] * bschaefer isn't sure who doko is [16:41] bschaefer: ubuntu toolchain maintainer [16:41] didrocks, full stacktrace: http://paste.ubuntu.com/5668696/ [16:41] didrocks, yeah, let me go find his email [16:41] bschaefer: IRC? [16:42] #ubuntu-devel ;) [16:42] didrocks, to many rooms! I checked a few and didn't see him :) [16:42] ;) [16:42] #ubuntu-devel is still the main one for ubuntu development [16:43] didrocks, I only recently started to log into that room! [16:43] ;) [16:44] Saviq, Uploaded! https://launchpad.net/~indicator-applet-developers/+archive/hud-phablet [16:45] didrocks, he says hes not aware of any :( [16:47] bschaefer: I saw, see #ubuntu-devel ;) [16:47] bschaefer, I doubt there is any ABI break [16:47] I would doubt as well gcc has one [16:47] * bschaefer wonders why recompiling fixes problem [16:47] we didn't see any issue anywhere else and those libs didn't change around the time your issue started [16:47] there was one last cycle due to C++11, but it was removed [16:48] bschaefer, can you run that test under valgrind? [16:48] but if it were an ABI break it should seg fault 100% of the time as well... [16:48] seb128, yeah let me get that info [16:48] bschaefer, right, ABI break don't manifest by making some % of the runs fail [16:49] seb128, yeah, I was also getting it while restarting unity sometimes as well, but not very often [16:52] * bschaefer has to rebuild unity to get it to crash again === alan_g is now known as alan_g|EOD === dandrader|lunch is now known as dandrader [17:15] Hey guys u1 local music isn't showing cover art in the music lens [17:20] here is an image of what I mean http://ubuntuone.com/59a6BHVg5sHOdGoPKtmoh9 [17:20] seb128, it seems to not want to crash going through valgrind :( [17:24] bschaefer, is there any invalid read/write in the log? [17:25] bschaefer, often valgrind will workaround the segfault but still catch errors [17:25] seb128, o alright, let me check...also the part that is calling fontconfig is running in its own thread [17:26] bschaefer, fontconfig is not thread safe in the current raring version [17:26] seb128, hmm well its called through umm gtk_load something I have to look it up really quick [17:27] bschaefer, the next major version will be thread safe (e.g not for raring) [17:27] seb128, hmm i wonder if that could be the problem, also where is libfontconfig1 vs lp:fontconfig? [17:27] seb128, as lp:fontconfig fixes my crash, where libfontconfig1 does not in raring [17:28] seb128, heres the log, and I don't seen any invalid read/writes: http://paste.ubuntu.com/5674245/ [17:28] bschaefer, oh, you use lp:fontconfig, that's upstream trunk which is thread safe [17:28] bschaefer, so yeah, that would make sense [17:28] seb128, sooo that fixes it shoot... [17:29] seb128, this is what we call: gtk_icon_info_load_icon(task->icon_info, &task->error); [17:29] with a nice comment above it: // careful here this is running in non-main thread [17:29] yeah, don't do that... [17:29] it's late to get new fontconfig in raring [17:29] haha...welllll alright Ill have to look into removing that thread then [17:29] thanks [17:29] seb128, cool though, thanks for helping me look into this! [17:30] bschaefer, np, sorry I didn't understand earlier that you were running trunk fontconfig [17:31] I though you were saying that a rebuild of the same version was fixing it [17:32] oo noo, i tried to mention lp:fontconfig, but i've been confused about the crash for a bit :) [17:33] seb128, mentioning the thread part was the real problem anyway :) [17:33] bschaefer, I'm not using to uptodate upstream import [17:33] you wrote lp:fontconfig and I read ubuntu:fontconfig [17:33] e.g the current ubuntu version :p [17:33] but yeah [17:33] glad we figured it out [17:33] ooo haha, right that makes sense [17:34] * bschaefer isn't use to all the lingo yet [17:34] or rather unaware of possibly ambiguities [17:34] possible* [17:34] it makes me regret a bit to not have updated fontconfig to the new version with the threadsafe fixes this cycle [17:34] oh, well [17:34] next cycle ;-) [17:35] seb128, yup :), looks like this is has been using its own thread since rev:2364 very strange... [17:35] * bschaefer wonders if this crash has been going on for a while === greyback is now known as greyback|eod [18:17] tedg, and will that get to ppa:phablet-team or? [18:20] Saviq, Not sure how we want that to work exactly. [18:21] Saviq, Was kinda waiting for it to build first :-) [18:21] Saviq, Perhaps making it dependent on both would make sense. [18:49] seb128, well looking into this more, removing the thread usage will cause this to regress: https://bugs.launchpad.net/unity/+bug/828582 [18:50] Launchpad bug 828582 in libunity (Ubuntu Quantal) "Dash: very high latency responding to input" [Undecided,Confirmed] [18:50] seb128, sooo its going to cause a big regression in the dash loading speed :( [18:51] bschaefer, the merge linked in this bug is from 2012-01-25 [18:52] which is 1.5 years old and before the LTS [18:52] seb128, yes, which is where the threads was introduced [18:52] do we get much real world report about that issue or is it mostly tests? [18:52] seb128, but the recent change in glib [18:52] deprecated some functions [18:53] seb128, its rare for it to crash, and I've only gotten it like twice on trunk unity [18:53] Does any of you want to help me by completing this survey? It is about clipboard usage http://goo.gl/drqfR [18:53] seb128, like 1/100 compiz --replace ccp rare [18:53] bschaefer, hum, what about glib changes? [18:53] is the issue due to changes that are to accomodate the new glib? [18:54] seb128, well glib went from 2.34 -> 2.36 causing ... let me get the change [18:54] seb128, https://bugs.launchpad.net/unity/+bug/1100658 [18:54] Launchpad bug 1100658 in unity (Ubuntu) "Unity fails to build on raring: g_io_scheduler_push_job is deprecated" [High,Fix released] [18:55] andyrock, might now know a bit more about the new changes as well...but they both are using threads [18:55] does reverting that change workaround the issue? [18:55] seb128, thats what im trying to look into, im looking at compiling gcc with glib2.34 [18:55] err [18:55] compiling unity with glib2.34 [18:56] why? [18:56] those functions are deprecated [18:56] they are still there [18:56] just build without G_DISABLE_DEPRECATED [18:56] bschaefer, compile unity with glib2.36 [18:56] but try to use the deprecated function [18:56] *s [18:56] yeah [18:56] ooo [18:56] i mis interpreted what andyrock as was saying... [18:56] glib never breaks compat [18:56] * bschaefer didn't know about that flag [18:56] bschaefer, ahahahah np ;) [18:57] so they don't drop anything [18:57] they just deprecate stuff [18:57] andyrock, haha [18:57] but that's a build flag away [18:57] right, that makes sense, I thought they removed it completely haha... [18:58] bschaefer, btw are you sure there is a problem with "my" glib 2.36 thread code? [18:58] there was little documentation on how to use it [18:58] andyrock, no i have not confirmed that, I just know this problem only started about last week [18:58] andyrock, and that other thread code has been around for 1.5 years... [18:59] not the situation should be better [19:00] andyrock, well... it could be the new version is doing threading differently [19:00] sorry my fault, i wanted to say: maybe the problem is "my" glib 2.36 thread code etc. etc [19:00] bschaefer, ^^^ [19:00] the glib stuff, causing the non-thread safe libfontconfig to crash === dandrader_ is now known as dandrader|afk [19:00] andyrock, right, we can look at the new code as well, lets see if reverting it fixes it :) [19:00] andyrock, if soo, then lets look at the new thread code to see if something is missing [19:01] ok [19:01] * bschaefer forgot to turn off errors for warnings...lets rebuild agian [19:01] again* [19:10] andyrock, seg fault [19:10] hmmm [19:11] using the old code path? [19:11] andyrock, yup [19:11] :/ [19:11] same bt? [19:11] andyrock, yeaah [19:11] * bschaefer wonders what glib is doing differently now [19:12] andyrock, hmmm well I also don't think its your code either cause lp:fontconfig fixes it [19:13] andyrock, soo I guess the decision now is how big of a problem is this out side of the test? [19:13] * bschaefer reverts changes and does a lot of compiz --repalces... [19:14] seb128, ^ using the old glib functions didn't fix it... [19:14] bschaefer, https://errors.ubuntu.com/?package=compiz is a good place to ask that question [19:14] seb128, thanks! [19:14] * bschaefer forgot about errors.ubuntu.com [19:15] the number of 13.04 users is low compared to other series though [19:16] so it's not making numbers very useful :-( === dandrader|afk is now known as dandrader [19:16] seb128, yeah, but hopefully its enough to see if this crash is common or not [19:19] bschaefer, the fallback option is to declare it's not an issue until it shows up in there and figure a SRU plan if it does [19:20] seb128, yeah, that is what im thinking, but it never hurts to double check through these things [19:24] seb128, and I don't see it...so Ill keep an eye on the errors, ill lower the priority of the test failing bug report as well [19:24] seb128, thanks for all the information! [19:24] bschaefer, thanks for investigating the issue [19:25] the good news is that it will be fixed after raring with the new fontconfig ;-) [19:25] seb128, yup! Im excited for that :) [19:29] andyrock, hmm what should we do about the GetManyIcons gtest? [19:29] andyrock, as its quite annoying to have to re-run make check when it seg faults [19:33] bschaefer, we shouldn't keep tests that segfault, they break automated testing and it takes hours to run the whole build/tests again [19:33] seb128, yup, I disable it, with a comment to revert onces we have the new libfontconfig1 [19:34] ill* [19:57] Saviq: is there a specific reason to call qmlscene from a .desktop's Exec line, instead of having a separate script in /usr/bin/ that does it? [19:58] also, isn't Architecture=any the right value for non-compiled apps? [19:58] mhall119, no, any will build it for... any arch [19:58] mhall119, all will build them for all [19:58] once [19:58] mhall119, when using bash a separate process is spawned [19:58] mhall119, that's no longer associated with the original desktop file [19:59] that doesn't really clear it up for me, in my mind any="one package for any arch", and all="one package for each arch" [19:59] mhall119, any="a package _per_ arch" [19:59] mhall119, all="one package for all arches" [19:59] completely opposite of what I thought, thanks [20:00] mhall119, our app management can only deal with a single process for now [20:00] Saviq: ok [20:00] Saviq: just wanted to check before approving these MPs [20:01] mhall119, here are the three packaged up http://ubuntuone.com/4ICFnTBsks7QTgjreBROo1 [20:02] mhall119, pmcgowan is onto the MPs, too [20:03] mhall119, you should be able to just install those and they should show up in the apps installed category (search for them if they don't fit in the two rows) [20:04] Saviq: is search working in the apps lens now? [20:04] mhall119, just for the installed ones, yes [20:05] I mean, video and music lenses gave me a search option before, but apps lens didn't [20:05] mhall119, it's new [20:05] mhall119, not ideal but does the job (i.e. all the other categories are hardcoded still) [20:06] Saviq: and you say i don't sleep [20:07] kgunn, it's only 10pm, at least I won't be back up in 7 hrs ;) [20:07] Saviq: is it enough to apt-get dist-upgrade on my device, or do I need to flash a new image? [20:07] mhall119, shell should be fine, couldn't tell for anything else [20:09] ok [20:17] dandrader: ping [20:18] kgunn, pong [20:19] dandrader: hey, was thinking you might want a little break from just unit test slog [20:19] dandrader: we have a "user story" we've targeted for end of april..."closing apps from the dash" [20:20] dandrader: i understand we have code for it === francisco is now known as Guest66331 [20:21] dandrader: basically we just need to port that code over (update as / if needed) [20:21] dandrader: test it, create a test for it & land it [20:22] dandrader: whenever you get to a stopping point with whatever you're doing now... [20:22] dandrader: cool? [20:23] dandrader, one important thing that I believe happened between that branch and now is the tablet, which added support for side stage apps in the apps Running category [20:23] kgunn, sounds good. I can start on it already first thing in the morning. [20:23] so it might require some hacking around [20:24] This afternoon I was just poking at Panel indicators code, didn't start to write its test yet === jhodapp is now known as jhodapp|afk === salem_ is now known as _salem