[00:31] Hey guys, im attempting a port to the LG optimus 4x HD (tegra3 SoC) and having a bit of trouble, i wanna give a android 4.4 base a try to see if it solves some issues, anyone know how to change the branch? === chriadam|away is now known as chriadam === chihchun is now known as chihchun_afk === IdleOne is now known as io === duflu_ is now known as duflu === _salem is now known as salem_ === salem_ is now known as _salem === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === chihchun_afk is now known as chihchun === Kyle is now known as Eliz === tvoss|afk is now known as tvoss [07:32] good morning === qengho is now known as CardinalFang === CardinalFang is now known as qengho [08:20] zsombi: I got a better look at the convergence bug I had. I made a new video about it at https://www.youtube.com/watch?v=FE-4NywIfZc&feature=youtu.be [08:29] dholbach: hey. I need some help with some debian packaging. Are you free? [08:30] nik90, how can I help? [08:30] dholbach: I was helping out a friend create a deb package of his application, and it seems to spit out errors while running "debuild -us -uc" [08:30] dholbach: here's the debian packaging that I implemented for his app -> https://github.com/krnekhelesh/project-dashboard/commit/fd3beb47d41b7a09f14daa8682902ba415c0cae4 [08:31] the error that I got was http://paste.ubuntu.com/7161502/ [08:33] nik90, whatever the clean target in the Makefile is trying to do - it shouldn't be there [08:34] dholbach: here is his MakeFile -> http://paste.ubuntu.com/7161512/ [08:34] nik90, I checked out the code [08:34] try commenting that part out [08:35] dholbach: on commenting the clean, backup and restore commands, it still fails [08:35] dholbach: if I comment the check, it then works [08:36] yeah, but that's a separate issue, no? [08:36] the thing is I have had to comment out the check line in about 3 projects that I did debian packaging for. [08:37] the unit tests pass 100%..which is why I was curious why this was happening [08:38] http://paste.ubuntu.com/7161535/ is what I had to do to make "debuild" pass [08:40] ok so the check and the clean lines are incorrect [08:41] running the check thing would be nice, but I don't know what to do to enable it properly [08:42] dholbach: ok [09:16] t1mp: how should we move forward on https://code.launchpad.net/~jamesh/ubuntu-ui-toolkit/albumart-dbus/+merge/212362 ? === chriadam is now known as chriadam|away [09:17] it'd be useful to have a clear decision on where these image providers should live. [09:36] Good morning all; happy Spanish Paella Day! :-D === oSoMoN_ is now known as oSoMoN [09:58] bzoltan: you know more about the packaging and sdk dependencies than I do. What do you think we should do with https://code.launchpad.net/~jamesh/ubuntu-ui-toolkit/albumart-dbus/+merge/212362 ? [09:58] t1mp, bzoltan: the new dependency in that branch is just another binary package from an existing source package dep [09:59] jamesh: dependencies like the thumbnailer some times caused issues in backporting the UITK, that's why I am a bit worried in adding more dependencies [09:59] it was added since it has different multi-arch requirements (you want a copy of the library for each arch, but a single copy of the d-bus service can serve all architectures) [10:03] jamesh: are the dependencies back portable to 12.04 and 13.10? [10:06] bzoltan: it should be: we're making use of features from existing dependencies (dbus support in gio and Qt) [10:07] no new source package dependencies have been introduced [10:09] jamesh: why doesn't the thumbnailer provide a qtquick binding? Then people import Ubuntu.Thumbnailer instead of having it in uitk? [10:11] t1mp: I don't know what the original reasoning was: I've asked Satoris to join to see if he can provide the background [10:11] jamesh: ok, thanks [10:11] I was just thinking the same :) [10:11] (he is the primary author of thumbnailer and the other thumbnailing image provider) === dandrader is now known as dandrader|afk [10:31] satoris: hi. t1mp and bzoltan were wondering what the initial reason was for putting the thumbnailer image provider in ubuntu-ui-toolkit as opposed to a separate QML module [10:33] * cjwatson posts the next layer of his attempt to make ubuntu-sdk-libs-dev:armhf cross-installable - I do hope this will end at some point [10:33] Hopefully another nail in the coffin of the practice of having libraries depend on run-time services, in the process [10:34] oh, yikes, that's relevant to the above isn't it === dandrader|afk is now known as dandrader [10:35] jamesh: what do thumbnailer-service's own Depends look like? I'm concerned that this will cause a problem for cross-installation in development chroots [10:36] jamesh: and it's pretty questionable whether e.g. the d-bus service should be mandatorily installed during development [10:36] those are usually run-time facilities that tests would have to mock anyway [10:36] jamesh: https://code.launchpad.net/~cjwatson/qml-friends/avoid-friends-dependency/+merge/213014 and https://code.launchpad.net/~cjwatson/libfriends/recommend-friends-dispatcher/+merge/213015 for background on why I care [10:41] cjwatson: it depends on libthumbnailer + its dependencies (which is already a dep for ubuntu-ui-toolkit) [10:43] jamesh: right, but in cross-install situations that'll involve following the dep trees on both architectures [10:43] best avoided if we don't have to [10:43] cjwatson: if the service isn't available, the image provider should present a fallback image and log the problem [10:44] I guess it could be a "Recommends" in that case [10:44] jamesh: ok, so that sounds like it should be Recommends not Depends, and we should take other measures if necessary to ensure that it's on our images [10:44] this sort of thing is very difficult to track down but we're not actually that far from the whole cross-install stack working [10:44] so I'm trying to preemptively avoid regressions :) [10:45] although that doesn't necessarily settle t1mp's desire for all the libthumbnailer providers to be moved out of ubuntu-ui-toolkit [10:45] for reference the cross-install stuff is *also* rooted in SDK team requests, so hopefully incentives are lined up here [10:47] cjwatson: we were discussing this MR https://code.launchpad.net/~jamesh/ubuntu-ui-toolkit/albumart-dbus/+merge/212362 [10:47] or rather, that MR started a discussion about where to put the thumbnailer [10:47] yes I know, I saw that in scrollback [10:47] ok :) [10:47] I'm just trying to avoid it breaking "click chroot" improvements [10:47] Recommends should be good enough for that purpose, if it also meets your needs (or can be made to do so) [10:47] we had some difficulties backporting to 12.04 with thumbnailer (and other dependencies) [10:48] in general I think less dependencies means less potential difficulties [10:48] Recommends should help backporting too, I'd've thought, since it doesn't have to be satisfied [10:48] moving all thumbnailer providers out of uitk is one possibility. If you know of better ways, please educate me [10:49] I'm not qualified to comment on that [10:49] ubuntu touch .. does it have a terminal app , and can it compile software locally [10:49] (even if it hasn't got 'full desktop integration' yet) [10:49] t1mp: but the libthumbnailer portability issues have since been solved, right? [10:49] separate modules may also be easier to support (for the UITK team because there is a bit less code in UITK), and for the thumbnailer maintainer because they can have their own policies for MRs [10:50] jamesh: I am not sure. The status of this bug is not clear to me https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1237045 [10:50] Ubuntu bug 1237045 in Ubuntu UI Toolkit "Ubuntu UI Toolkit no longer builds on precise, quantal and raring" [Critical,Confirmed] [10:50] bzoltan, satoris is UITK now perfectly packaged on p, q, r? [10:50] ^see the bug above [10:52] t1mp: yes it is. [10:53] Is gdbus available on precise, though? [10:53] yes [10:54] Then we should be good. [10:55] * t1mp reading the friends MRs for background information [10:56] If the thumbnailer image provider moves from uitk to thumbnailer, can we preserve API and ABI for apps? [10:56] Looks like gdbus was added in glib 2.26, and precise has 2.32 [10:57] t1mp: I suspect most of it is in the category of "you don't want to know if you don't have to", but you never know [10:58] satoris: is ABI relevant since the UITK only provides a qml interface? [10:58] satoris: API-wise, we could keep in as deprecated in uitk for a while? [10:59] With a message to shell guys saying "transition as afast as you can"? [10:59] satoris: keeping it deprecated means it will still work, but a warning is printed that the thumbnailer moved to a separate package [10:59] satoris: yes [10:59] Works for me. [10:59] satoris: more a message for people who use it that we don't know they are using it [10:59] satoris: also for shell guys, but we can work with them to make the changes at the same time as the UITK changes are done [11:00] more relevant, the file descriptor passing support in gdbus is from 2.30, which is still earlier than precise [11:01] t1mp, satoris: only one module can install a provider with a given name, although ubuntu-ui-toolkit currently includes a try/catch block around the call [11:01] cjwatson: I read the MRs, and so far I understand that you don't want compile-time deps because that will further complicate the chroots used to build the packages? [11:02] so we could potentially copy the image://thumbnailer image provider to an Ubuntu.Thumbnailer module and have them both importable at once [11:02] well the existing problems render it impossible to build coherent ones right now [11:02] this is why click chroot is not really working right for you guys today [11:02] the app compatibility problem comes when you go to remove the ubuntu-ui-toolkit version [11:03] jamesh: oh to have them both work is cool, that eases the transitioning [11:03] run-time services shouldn't be hard dependencies of libraries because they *aren't* hard dependencies - you can reasonably have the library installed in development contexts without the run-time [11:03] jamesh: I propose initially we keep it working, but we print a warning message in UITK [11:03] jamesh: next step: keep the API and print a warning message, and don't show the thumbnail (or a thumbnail with a warning) [11:03] jamesh: only after that remove the API [11:04] t1mp, satoris: so I guess it's agreed that we add a QML module? [11:04] jamesh: +1 from me [11:04] t1mp: alternatively, perhaps it is possible to have ubuntu-ui-toolkit import the Ubuntu.Thumbnailer QML module? [11:06] jamesh: I liked your previous idea of having them both importable at once. I'm not sure how that will work though [11:06] jamesh: can we do that? or would we need a new name for example image://thumbnail for the thumbnailer instead of image://thumbnailer? [11:06] t1mp: if we had two copies of the provider, it would depend on which order you imported Ubuntu.Thumbnailer and Ubuntu.Components [11:08] jamesh: for UITK we have this http://pastebin.ubuntu.com/7162114/ [11:08] jamesh: so the try would fail if Ubuntu.Thumbnailer was already imported, and all is fine for the app? [11:09] t1mp: right. it is just logging the error and continuing. [11:09] If we do the same in the (not yet existing) Ubuntu.Thumbnailer module, it would also be importable after Ubuntu.Components [11:10] Morning all [11:10] jamesh: yes. And we can print a deprecated warning in requestImagein uitk's thumbnailgenerator.cpp [11:11] jamesh: just if the app imports thumbnailer, but after uitk, the app will still have deprecated warnings [11:11] and we'll get complaints about breaking thumbnailer on PowerPC after adding QML to the mix, but that's relatively easy to ignore :) [11:13] jamesh: how would you break thumbnailer on powerpc? wouldn't the qtquick module of thumbnailer be a separate package that doesn't have to be installed on powerpc? [11:13] t1mp: Qt's Javascript engine (V8) is not compatible with PowerPC [11:13] so the package would fail to build due to missing dependencies [11:14] the PPC guys complain whenever a package that previously built on PPC stops building there [11:14] jamesh: you would keep the thumbnailer package that you have now and works on powerpc, and add a qtdeclarative5-thumbnailer-plugin package for the qml binding [11:14] jamesh: so there will only be an additional package, no change to the old packages [11:14] which I can sympathise with, but can't do much about in this case [11:16] t1mp: we've gone though this a number of times (e.g. for the Media Scanner's QML API). The real fix is for V8 to be ported to PPC rather than fighting against QML [11:17] I don't think we've got any libthumbnailer users outside of QML apps anyway [11:17] jamesh: what will break the thumbnailer on ppc if the qml binding is provided by a new package? [11:17] t1mp: the package will have to be dropped from the architecture. [11:18] jamesh: which package? libthumbnailer or the qml plugin? [11:18] the qml plugin was never there before, so I think it doesn't break anything. Does it? [11:18] t1mp: probably everything built by the source package. [11:19] ah, a single source package builds everything [11:20] Yep, that's the problem. [11:20] can we make the qml plugin a separate source package? [11:20] I don't have a lot of experience with packaging, so excuse my questions [11:21] A separate source package requires a separate launchpad project. [11:21] So preferably not. [11:21] A separate source package is a lot of work [11:21] and having the QML module and backend library under one umbrella has its own advantages: they can be updated in tandem [11:22] jamesh: yes it sounds easier to have it together [11:22] is it possible to set a flag on ppc that prevents the qml package from being built? [11:22] or use #ifdefs to build an empty plugin on ppc [11:23] the build will fail due to missing dependencies [11:23] I don't think we want to explicitly exclude PPC though [11:23] not that I care so much about ppc. [11:23] so far only unity8 uses thumbnailer? [11:23] if we did, someone would have to go around removing all those exclusions once V8 is ported [11:24] it uses the thumbnailer image provider from ubuntu-ui-toolkit, yes. [11:26] rsalveti: ChickenCutlass: how is media-hub and emulator coming along? [11:26] Can you set per-architecture build deps and output packages? [11:26] bzoltan: ^ [11:26] jamesh: eh, the stuff you say above is no longer true with Qt 5.2 [11:26] jamesh: Qt 5.2's script engine is a new thing called v4 that's portable [11:27] I wonder how the other qtdeclarative5-* plackages do it [11:27] cjwatson: cool. I haven't tried adding a a Qt dependency to a package in the ~ 1 week we've had 5.2, so hadn't noticed [11:27] satoris: you can but as I say above it's entirely unnecessary now [11:28] maybe the other qml plugins have separate lp projects, like this one https://launchpad.net/qml-friends for qtdeclarative5-friends0.2 [11:28] there are plenty of embedded ones too [11:28] and looking at Launchpad, I have PPC binaries of mediascanner2 now [11:28] good thing we didn't manually disable it :) [11:29] e.g. content-hub: Binary: content-hub, libcontent-hub0, libcontent-hub-dev, qtdeclarative5-ubuntu-content0.1, libcontent-hub-doc [11:30] Or mediascanner2, as you say [11:31] I love it when a plan comes together before we need to start doing anything. [11:32] ogra_: did you do "Bootchart for daily iamges"? [11:32] ogra_: how and where are those done? can you expand a bit more :) [11:32] asac, i do one bootchart a day, yes [11:32] not per image though [11:33] ogra_: how and where? [11:33] asac, http://people.canonical.com/~ogra/touch-bootcharts/ [11:34] currently by running a local script by hand with my N4 with the broken screen [11:34] http://paste.ubuntu.com/7162200/ [11:35] hmm. doesn't feel like something we should highlight as DONE :P [11:35] there are still issues with pybootchartgui and detecting when the system is properly idling .... only http://people.canonical.com/~ogra/touch-bootcharts/ubuntu-phablet-trusty-255.png has proper cropping yet [11:35] ogra_: did you give up getting this into proper automation? [11:36] asac, no, but the change in the CPU core behavior causes issues with idle detection there too ... i plan to put that into automation, just not right now yet === oSoMoN_ is now known as oSoMoN [11:38] ogra_: ok i change that item to "daily bootchart prototype by ogra online" [11:38] not dail bootchart done [11:38] which item ? [11:39] * ogra_ doesnt know what you refer to [11:39] ogra_: your boss claimed this as a highlight and it read as if it was finished [11:39] now it reads better :P [11:39] well, its an ongoing thing since one release already [11:40] ogra_, is your screenshotter script still the way to take shots, or was that wrapped in something now? [11:40] ogra_, (and if it is still the way, where is it again?) ;D [11:40] and will be ongoing for a while still until it is fully automated ... we get daily bootcharts, thats the important point [11:40] ogra_: so what is the problem with getting this in proper automation? [11:40] ogra_: if its one release, we should figure how to finish this up [11:40] Saviq, hmm, i thought the phablet-screenshot fix landed yesterday [11:40] (if we want it) [11:40] * Saviq tries [11:40] ogra_: do you know what part of MMS is supposed to be delivered soon? [11:41] rsalveti: ^^ [11:41] asac, fixing pybootchartgui to properly crop after unity8 is mapped [11:41] ogra_, seems it works indeed [11:41] asac, it is in the works, but we might be blocked by distro freezes now [11:41] Saviq, great :) [11:41] ogra_: but why can't we just land it so it produces these... even if the data isn't perfect [11:41] asac, because indicator-network and ofono are desktop packages too ... as is NM [11:42] ogra_: this chart, isn't this based on data? i think that data is also valuable to have in a parsable format [11:42] we can ship the daemon, but that wont help much without all the other changes [11:42] asac, you mean the bootchart tarball ? [11:42] ogra_: i dont understand. it seems we have everything to start producing bootcharts [11:42] i can upload that alongside, sure [11:42] asac, but they are inaccurate [11:42] the charts are not perfect, but we can already start producing them in automation :) [11:42] sure. not sure why we would block landing it because of that [11:43] because they dont reflect the actual boot time [11:43] they just keep running, they dont stop when the boot is done, so you cant really compare if a boot was faster or slower than another boot [11:44] lets do it proper before we rely on the data please [11:44] ogra_: adding it to be run daily doesnt mean we rely on the data [11:44] just that the work to do that is already done [11:44] step by step [11:44] i have no issue uploading the tarball and setting up a cron job to get it automted [11:44] but i dont want to publish it until it produces reliable data [11:44] ogra_: we want to produce it from devices in the lab [11:44] not from your home device [11:45] that part will take a while, so we should start proactively imo [11:45] ogra_: but you already publish it now :) [11:45] right, i'll give the script to CI once it is ready for wider consumption [11:45] i dont promote the published charts :P [11:45] and i dont ask anyone to rely on them [11:45] ogra_: well, your boss did :) [11:46] so now i want that proper [11:46] anyhow [11:46] dont think its right to have this run as a low prio effort on your desk [11:46] we probably would like to put more folks on it [11:46] well, i dont think it is good to not have proper data [11:47] i'll work with apw tomorrow (once he is back from vac.) to see what we can do about the idle stuff ... thats already planned [11:49] ogra_: does the data give us accurate boot time? [11:50] no [11:50] ogra_: if the chart is buggy its not really a blocker [11:50] it doesnt give us accurate boot time until we get proper CPU data [11:50] ogra_: because it uses a systemsettle like mechanism to figure when a process is done starting? [11:50] for which i need the kernel teams help ... which as i said is already planned for tomorrow [11:51] k [11:51] asac, well, look at the charts ... it doesnt get proper CPU data at all [11:51] and yes, it checks for idle after the process that you name started ... [11:51] that idle was only reached once in chart 255 ... by accident i think [11:53] asac, i'll work with plars and doanac next week to get their setup prepared ... note though that this needs a dedicated device for the test run ... [11:53] ogra_: constantly allocated? or just a separate run? [11:54] separate install ... several reboots and sleeps ... takes about 15-20min [11:54] not completely dedicated, but for this test run you cant use an install where ap tests are installed ... that would taint the data === marcoceppi is now known as marco_traveling [11:56] i wouldnt run it as part of the test suite, but as something that runs when spare devices are available [11:56] (in parallel) [12:00] in any case the charts show that we need to fix ureadahead ... which i was planning to work on today ;) [12:03] davmor2, just got 263 with rather odd results in the scopes, although seems nothing changed there [12:05] davmor2, my video scope is empty and I have no remote content in any [12:05] smells like fallout of removing empty video thumbnails from the gallery app [12:12] pmcgowan: I have local video content still and online stuff. Although when I installed R250 without a wipe I lost it but then when I upgraded to 262 again it was fine and 263 this morning was fine. So maybe the R250 had the fresh gallery and knock my videos off. [12:12] davmor2, I will reboot and see if it fixes [12:12] pmcgowan: bfiller will be the man to ask [12:12] about scopes? [12:13] pmcgowan: no for the gallery fix I think to ensure that is nothing to do with it [12:13] it couldnt imo [12:13] I have 3 scopes acting wrong [12:13] apps scope has no store content [12:13] pmcgowan: the scpopes here are fine for me [12:14] davmor2, reboot fixed it [12:14] pmcgowan: works here, and gallery has no videos [12:14] davmor2, will see if I can reproduce, I had it set to use 3g only, maybe something timed out === _salem is now known as salem_ [12:15] pmcgowan: oh it might be that you had low to no signal in which case it would be blank I think for the online stuff [12:16] davmor2, yeah but videos should have shown [12:18] pmcgowan: yeah videos should of been visible. That might be due to the video data being removed from the gallery though [12:18] davmor2, just reproduced it [12:18] if I turn off wifi no content [12:19] might be low 3g data where you are [12:19] davmor2, well whats worse is I have no 3g connection even though its enabled [12:20] hmm I have an address [12:20] a route as well ? [12:20] (route -n) [12:20] pmcgowan: so if I turn off cellular and wifi and reboot I get no online stuff but I still see all my local content [12:21] ogra_, yes [12:21] how about DNS [12:21] davmor2, I am just missing videos [12:21] dns not working, cant resolve names [12:21] aha [12:21] file a bug against NM then === MacSlow is now known as MacSlow|lunch [12:26] ogra_, I lied, cockpit error [12:26] it can see hosts fine [12:29] pmcgowan: open the browser I find a reliable test for if the net is working [12:29] davmor2, yes it is, so my network is fine, but my scopes are not [12:30] davmor2, what do you see on 3G only after a boot [12:30] pmcgowan: give me 5 I'm setting up screenshots currently [12:39] pmcgowan: this is with no data connection http://ubuntuone.com/7AcO0qnGT6vjJCCsm2jBPA and a reboot [12:40] davmor2, do it with a data connection on 3g? [12:40] 3g but no wifi is what I have [12:40] pmcgowan: yeap that is what I'm doing now [12:41] pmcgowan: right so I've enabled 3g now and everything stays the same, Now I'm going to search for something and see what happens then [12:42] davmor2, so searching makes them all show up [12:42] let me search on video [12:42] pmcgowan: it refreshes the scope [12:43] didrocks: Yay got a unity8 crash search in music scope [12:43] davmor2, still no local videos === alex_abreu is now known as alex-abreu [12:43] davmor2: easily reproduceable? [12:43] pmcgowan: open file manager and search in Videos [12:43] reproducible* [12:43] davmor2, well I know they are there [12:43] didrocks: I will see if I can [12:44] gack [12:44] davmor2, they are not there === jhodapp|afk is now known as jhodapp [12:45] pmcgowan: which is what I said, I think the gallery fix might remove them but I'm forever wiping my system so it's hard to tell [12:45] davmor2, where did my videos go? [12:45] davmor2, indeed, youa re correct sir [12:45] davmor2, not much of a fix Id say [12:46] davmor2, are these both known issues then [12:48] pmcgowan: so I hit it late last night but but I only hit installing R250 so wasn't sure if it was that the only way I can think to check would be to install the old gallery click package on R262 and upgrade to 263 and see if the updated gallery app and upgrade wipes them [12:48] davmor2, that seems the case a I just went from 262 to 263 [12:49] didrocks: ^ I think that just killed my dogfooding for the day [12:51] pmcgowan: Bill told it was fixed though [12:51] pmcgowan: and we don't know about any design (if it's supposed to be shown or not) [12:51] didrocks: it fixed them showing in the gallery app [12:52] didrocks: I'm wondering if the database purge removes them [12:52] jamesh, satoris did we agree to move the thumbnailer out of UITK in its own package in the thumbnailer project? [12:52] so the videos themselves are removed? [12:52] didrocks: indeed [12:52] urgh, seems critical [12:52] and seems that they will need a test for it [12:52] davmor2: you are going to try to confirm that? [12:53] didrocks: indeed hence me saying I'm going to dig into it and that might kill my chances at dogfooding [12:53] yeah [12:53] seems a priority to me anyway [12:53] well, it is kind of fixed ... you wont have empty thumbnails if you dont have videos at all :P [12:54] didrocks: yeap. I'm going to file this unity8 crash first though and see if it matches either mine or your crashes [12:54] right :) [12:56] davmor2, could I have lost the videos going to 262? [12:56] pmcgowan: I'm going to do a range of tests [12:56] vg === dandrader is now known as dandrader|afk [13:02] t1mp: assuming all the packaging stuff etc works then that's fine by me. [13:03] ok, cool === jhodapp is now known as jhodapp|brb [13:04] satoris, jamesh I rejected the MR https://code.launchpad.net/~jamesh/ubuntu-ui-toolkit/albumart-dbus/+merge/212362 [13:05] thanks for clearing stuff up [13:05] didrocks: bug #1298370 is my unity8 crasher this time which looks different again :( [13:05] bug 1298370 in unity8 (Ubuntu) "unity8 crashed with SIGSEGV" [Undecided,New] https://launchpad.net/bugs/1298370 [13:06] now onto the gallery app issue === alan_g is now known as alan_g|lunch === MacSlow|lunch is now known as MacSlow [13:15] didrocks: okay so now I'm getting lockups/crashes on every scope opening sections so I am assuming data retrieval is to blame === jhodapp|brb is now known as jhodapp [13:23] popey, didrocks: okay so any Idea how I get the old click package? currently trying an upgrade with the fixed version in place see if it is the upgrade that broke something then I'll want to try with the older click package installed and see if the update is what breaks it [13:24] rickspencer3, I tried the video you shared with me on a 2013 Nexus 7, seems to play just fine for me (normal speed) [13:25] jhodapp, that's very weird, the voices are def. slowed down when I play it [13:25] on my desktop as well [13:25] rickspencer3, give it another try after media-hub has landed (which should be early next week) [13:26] * tshirtman waves at didrocks [13:26] davmor2: hum, you are talking about a different issue than the gallery-app one, right? [13:26] hey tshirtman [13:26] davmor2: which old click package? [13:26] :) [13:26] popey: gallery the package before the latest one [13:27] didrocks (and others :P) nice job on ubuntu-touch :) [13:27] popey: we have a horrible felling it removes local videos [13:27] jhodapp, tbh, I think I already wasted too much of your time with this [13:27] it's an interesting puzzle, I guess [13:27] but I suspect that when I play dodgey videos, I should expect dodgey results [13:27] heh, thanks tshirtman :) [13:27] davmor2: yeah, popey would know, I don't [13:28] i have a cache of old app clicks [13:28] rickspencer3, actually it's not a waste, each video is more test ammo for me [13:28] jhodapp, ok, in that case, maybe check out Rick and Morty from a legitimate source, and then compare [13:28] popey: so I need the last but one version that had the blank dates in so I can see if the update is magically removing the local videos or not [13:29] and then when someone "busts" you watching Rick and Morty at work, ... [13:29] davmor2: i have 937, 934, 931, 927, 922 [13:29] pick one [13:29] rickspencer3, lol [13:29] rickspencer3, that's why I love my job...watching movies is "testing" :) [13:30] jhodapp, when I lived in Seattle, I knew some game testers from Microsoft, and some from Nintendo [13:30] talk about a mind f*ck [13:30] ha, what an awesome role [13:30] jhodapp, I dunno, playing the same level of a game all day, everday, for a week [13:31] one may find Rick and Morty less entertaining after 50 viewings ;) [13:31] it'd be fun for a bit...you'd be the god of that game though [13:32] popey: I'll take 931 please looks like I have 934 installed now [13:32] mm, I'm pretty sure game testing would make you hate games [13:33] davmor2: http://popey.com/~alan/com.ubuntu.gallery_2.9.1.931_armhf.click [13:33] cjwatson: but I don't hate Ubuntu and I test that all day :P [13:35] * rickspencer3 high fives davmor2 [13:42] dobey: what would be a decent lightweight way for me to benchmark the scope? I don't need anything super-accurate, I just want to make sure that my conversion of Interface::get_manifests and Interface::get_manifest_for_app to libclick has actually sped things up. Maybe something like a way to force the scope to refresh things? [13:42] okay now I have that blank dates again I will now update the app and see if there are videos still [13:42] dobey: I'm also a bit confused that Interface::get_manifests doesn't seem to be called anywhere. I know that your drop-vala branch hasn't landed yet, but are we using the C++ scope now or not? === alan_g|lunch is now known as alan_g [13:45] didrocks, pmcgowan: so updating from gallery 931 to 937 delete videos from the device but you have to open the gallery app on 931 or it doesn't. [13:46] pmcgowan: so you got it because it updated gallery in the update from 262 to 263 [13:46] davmor2, ok, and I had opened it previously [13:46] pmcgowan: and did you see dates with blank entries when you opened it? [13:47] davmor2, I did not look closely but I believe so, was focused elsewhere at the time [13:47] if so then I think we nailed it. and the data being removed from the DB is removing the videos [13:48] didrocks: I will write up a bug for it then [13:48] davmor2: ok, we need a bug, ping bfiller and revert the version in the store [13:48] popey: you can revert to a previous click version in the store, right? [13:48] tvoss, greyback, got time for a hangout? [13:48] davmor2, but this is the case independent of the image? i.e. folks updatign from the store? [13:48] didrocks: I can, but why do you want to? [13:48] Saviq, gimme 5 [13:48] tvoss, k [13:48] davmor2, so will people on promoted images see this? [13:48] popey: it deletes videos on user's device [13:49] pmcgowan: yeah, they would if they update the apps [13:49] Saviq: yep [13:49] "Changed published version to 2.9.1.934" [13:49] didrocks: ^ [13:50] popey: hum, dave tested 931, how old is 934? [13:50] New version: 2.9.1.934 on 2014-03-24 16:27 - 2 days, 21 hours ago [13:50] cjwatson: yes, we're using the c++ scope now. i don't know of a way to benchmark the performance though, no [13:50] ok, so it was the one with the empty videos [13:50] popey: sounds good then, thanks! [13:50] np [13:51] dobey: ok, any clue as to why get_manifests seems to exist but be unreferenced? the scope doesn't seem slow enough to be forking/execing click once per installed app [13:51] davmor2: critical, assign it to Bill [13:51] cjwatson: have you talked to tvoss about libclick being gobject/vala/exposing json-glib? it doesn't fit terribly well with the idea of new scopes not using gobject, qtc, etc… [13:51] 116 people have gallery 237 [13:51] *937 [13:51] dobey: yes [13:52]