[00:10] cjwatson: https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1218674 [00:10] Launchpad bug 1218674 in unity-scope-click (Ubuntu) "Need to make .click world readable before passing to click install" [Undecided,New] [00:10] alecu: ^ === Tassadar__ is now known as Tassadar_ [00:10] now, what's the least intrusive fix [00:11] I dont really want to poke download manager [00:11] note to self: slightly ridiculous to take that long to find a permission issue === chriadam|away is now known as chriadam [00:11] ah, I've seen bugs like that before - I suspect the right answer is to open it first in click and pass fds around [00:11] I don't really think it's the scope's job to fix that [00:12] ok [00:12] click's internal uid shuffling should be its own problem, basically [00:12] oh ok thanks Fishscene [00:12] NP mate. [00:14] I dont quite understand why the results were so random [00:14] but anyway [00:20] cjwatson: Hmm it's not strace messing the ldpreloading is it? [00:32] is there a tutorial anywhere for writing a scope? [00:38] cwayne: yes [00:38] cwayne: http://developer.ubuntu.com/resources/tutorials/scopes/writing-a-unity-scope/ [00:40] mhall119, wonderful, thank you [00:41] lool: thanks [00:47] rsalveti: available for a quick review? === _salem is now known as salem_ [01:37] slangasek: Hi - no special setup has been needed in the past, so there's no HOWTO document. I usually either sdh shell or ssh in, and run 'autopilot run test.id.here' === chihchun_afk is now known as chihchun [02:38] alecu: Good news, I think I have a patch [02:39] alecu: at least it worked in my basic testing [02:39] alecu: see https://bugs.launchpad.net/ubuntu/+source/click/+bug/1218674 [02:39] Launchpad bug 1218674 in click (Ubuntu) "clickpreload doesn't work on armhf" [Undecided,New] [03:00] thomi: hmm, so I did that, 'autopilot run ubuntu_rssreader_app.tests.test_rssreader.TestMainWindow.test_add_remove_feed_and_topic' copied from http://reports.qa.ubuntu.com/smokeng/saucy/image/3825/ubuntu-rssreader-app-autopilot/317522/, run as root, and I get dbus errors :/ [03:02] slangasek: so you're trying to reproduce the DBUS_SESSION_BUS_ADDRESS environment variable being missing? [03:03] slangasek: also, I'm not familiat with the rssreader test suite, so I don't know how stable is is. Perhaps run some unity8 test suite as a litmus test? [03:25] thomi: can you url me? === Namidairo`bnc is now known as Namidairo [03:38] slangasek: I think you're asking for this? http://reports.qa.ubuntu.com/smokeng/saucy/image/3833/unity8-autopilot/ === salem_ is now known as _salem [04:00] Does anyone know if the email core app is being working on yet? === jamesh_ is now known as jamesh === Guest54718 is now known as Zic [06:28] I want to run UT on a Galaxy S and I am following this guide: https://wiki.ubuntu.com/Touch/Porting I am at Device Changes SUID, but I can't find the fstab, should I look on my computer or on the phone? [06:58] good morning [07:09] I think I have found the fstab thing in the aries folder from which apparently some things are used [07:33] oSoMoN: is there anyone from apps team that could look at calendar/share/notes and friedns apps failures? [07:33] http://reports.qa.ubuntu.com/smokeng/saucy/image/3833/ [07:34] asac: me I guess, let me check [07:34] oSoMoN: :) [07:35] sorry. if you could have nailed down who busted you i would have preferred to go to those and tell them to better coordinate and test their stuff [07:35] but whatever fix helps :) [07:35] oSoMoN: so, but the webbrowser didnt change... [07:36] oSoMoN: i assume uitoolkit really changed and didnt test before landing? [07:54] asac: yeah, that’s a change in the uitk that triggered the failures, but really the responsibility is shared, as all the apps should be using the standard emulators provided by the uitk to implement their autopilot tests, and the browser wasn’t using them yet, so its tests were making assumptions that became false with the changes in the uitk (the actual functionality of the app wasn’t impacted, only the tests) [07:54] asac: I had been meaning to update the browser tests to use the standard emulators for a while, but the task kept being delayed by other more urgent ones [07:55] at least this is done now [07:59] oSoMoN: right. still testeing would have revealed this before the landing [08:00] oSoMoN: and you could have worked on this withou a firedrill if they had warned you :) [08:00] asac: true [08:08] Mirv: sil2100: do we have a spreadsheet that has all the branches we daily-release somewhere? [08:08] or a list that I can copy over? [08:09] psivaa: hi [08:09] psivaa: can you give back http://reports.qa.ubuntu.com/smokeng/saucy/image/3833/camera-app-autopilot/ ? [08:11] lool, great work on https://bugs.launchpad.net/ubuntu/+source/click/+bug/1218674 [08:11] Launchpad bug 1218674 in click (Ubuntu) "clickpreload doesn't work on armhf" [Undecided,New] [08:11] lool rocks :) [08:11] and rolls ... [08:11] ogra_: moin moin [08:11] :) [08:11] moinsen [08:12] and the show goes on :) [08:12] heh [08:12] ogra_: quick status update: uitoolkit didnt help alone, oSoMoN is now working on a webbrowser real fix and is looking in the apps autopilot failures [08:13] ogra_: further we have unity8 looping every other boot [08:13] https://jenkins.qa.ubuntu.com/job/saucy-touch-mako-smoke-camera-app-autopilot/85/artifact/clientlogs/top_before.log/*view*/ [08:13] ogra_: and we have sensorservice sometimes going wild [08:13] anything missing? [08:13] :) [08:14] e.g. all the failures explainable by those? [08:14] :) [08:14] asac: share-app is fixed in trunk, so the next release will make its tests green again, and I’m looking at calendar-app [08:14] oSoMoN: awesome!!! [08:14] ogra_: so i am pretty sure we want an early midday image [08:14] so we know if the weekend starts green :) [08:14] wow, there are quite some additions/removals http://paste.ubuntu.com/6043196/ [08:14] well, floodgates were open, FF was coming:) [08:14] ogra_: new unity [08:15] well [08:15] i dont care about the updates [08:15] lool: can you look at the package diff http://paste.ubuntu.com/6043196/ [08:15] it seems a lot of dependencies changed on top of the few seed changes we had [08:15] lool: i believe some there are changing because of click/systemupdates [08:15] well, yeah, cjohnston seeded packagekit [08:16] asac, the /beat/ goes on! (http://www.youtube.com/watch?v=PjUSlBCLIAU) [08:16] err [08:16] ogra_: libcheese? :) [08:16] cjwatson, [08:16] asac, goes well with libbacon :) [08:16] ogra_, asac: libcheese is coming from g-c-c, that's being worked on, do we need to have that discussion every day? ;-) [08:16] ogra_: didnt know about cheese [08:16] seb128: err you [08:16] seb128, nah, then it wouldnt be a real day [08:16] :P [08:16] no, no discussion neede don that one [08:17] yeah, packagekit => at least gdebi-core, packagekit-backend-aptcc, python-{apt,chardet,debian,six}, python3-packagekit [08:17] just wanted someone to double check that the changes we see are sane :) [08:17] or at leasat understood [08:17] asac, they are buggy but understood [08:17] and also => dropped python3-aptdaemon{,.pkcompat} [08:17] ogra_, groundhog day :) [08:17] probably also => dropped python3-defer [08:17] dholbach, hehe [08:17] dholbach: :) [08:18] doesn't look all that worrying anyway [08:18] well, it looks bigger than i expected from the seed changes [08:20] asac: is changes often, http://bazaar.launchpad.net/~cupstream2distro-maintainers/cupstream2distro-config/trunk/files/head:/stacks/head/ has them all [08:20] Looks fairly routine from the days when I used to read the diff between Ubuntu desktop images every day. [08:20] s/is/it/ [08:21] Mirv: do those .cfg files include the list of autopilots to run for each stack? [08:21] asac: yes [08:21] Mirv: i see: [08:21] testpackages: unity8-autopilot [08:22] http://bazaar.launchpad.net/~cupstream2distro-maintainers/cupstream2distro-config/trunk/view/head:/stacks/head/unity8.cfg [08:22] "tests:" (run after installing the packages "testpackages:") [08:22] Mirv: how do you guess the component to acutally run? [08:22] i know that we have a syntax best practice, but found that there are autopilot tests that have a different syntax styule [08:23] Mirv: so those files seem to not include the know how of which branch goes into which stack [08:23] asac: so there simply autopilot run unity8 is done (as tests: unity8) says after installing unity8-autopilot [08:23] Mirv: where is that know how? i assume a separate cfg for upstream merger has that? [08:23] asac: the branches are the items under "projects" [08:23] Mirv: oh i see the tests: field now. thankx [08:23] so each subitem there [08:24] Mirv: i cant see a URL for the bzr branch [08:24] asac: it's lp:item [08:24] http://bazaar.launchpad.net/~cupstream2distro-maintainers/cupstream2distro-config/trunk/view/head:/stacks/head/unity8.cfg [08:24] there is no lp: [08:24] http://bazaar.launchpad.net/~cupstream2distro-maintainers/cupstream2distro-config/trunk/view/head:/stacks/head/phone.cfg here neither [08:24] oh [08:24] asac: yes, but that's the branch name, so bzr branch lp:itenname works [08:25] i see what you mean :) [08:25] yuou could put in: [08:25] unity8/1.2 [08:25] as item [08:25] yep [08:25] asac, can you delay questions for a bit, I need Mirv (or sil2100) to look at http://10.97.0.1:8080/job/autopilot-saucy-daily_release/1437/label=autopilot-intel/console so we can land indicators and get gnome-control-center out of the touch image [08:25] and it would guess lp:unity8/1.2 [08:25] ? [08:25] seb128: of course, he just needs to tell me so :) ... this is certainly long term talk [08:25] asac, thanks [08:25] seb128: ok [08:25] Mirv, ^ can you look at that (I asked on #ubuntu-desktop also but not sure you saw it there) [08:26] seb128: I saw now also there [08:33] asac: ack, just did it [08:33] and will look at any other that needs re-running === Namidairo`bnc is now known as Namidairo === schwuk_away is now known as schwuk [08:57] How can HTML5 Ubuntu Touch apps access camera and microphone? | http://askubuntu.com/q/339253 === oreneeshy_ is now known as oreneeshy [09:12] psivaa: thanks! [09:12] asac: np [09:25] popey: I like this :) http://www.omgubuntu.co.uk/2013/08/ubuntu-touch-adds-first-run-intro-tour [09:25] timp: it's nice isn't it! [09:26] popey: yes. subtle and clean, but clear [09:26] i cant wait for mir to work on ubuntu touch [09:27] timp: saw a fair number of people at OSCON who picked up the phone and started randomly stabbing the screen, couldn't figure out how to unlock it or start apps. [09:27] this will help 100% [09:27] @popey can't wait to get it running on an old Galaxy S, currently midway of the porting guide [09:28] balachmar: ooh, that would rock [09:28] doesn't italready run? [09:28] https://wiki.ubuntu.com/Touch/Devices/galaxysmtd [09:36] mediaplayer-app when playing the sintel_trailer-1080p.mp4 demo asset is running really slow on todays image on my samsung galaxy nexus, it is pegging one CPU at 100%. the previous phablet image I was using a couple of days ago it was much smoother [09:37] it's as if A/V sync is completely messed up [09:38] cking, yeah, we are in the middle of changing to gstreamer decoding [09:38] until the new stuff is in it will use SW decoding [09:38] which causes such issues [09:38] ogra_, that explains it :-) [09:38] for me it actually runs very smooth until the buffer is filled the first time ... 15-30sec :) [09:39] after that its a slideshow with sound [09:39] ogra_, yep, I see the same kind of issue. so when will H/W decoding appear? I was going to power measure A/V playback [09:40] i think the plan is to land it today, but for details ask jhodapp and rsalveti once they are up, they are doing that transition === vrruiz_ is now known as rvr [09:40] ogra_: hi. would it be possible to add "mediascanner" and "unity-scope-mediascanner" to the ubuntu-touch meta package dependencies? [09:40] jamesh, with pleasure :) [09:41] jamesh, doe that have autopilot tests the QA team should know about ? [09:42] ogra_: there is an integration test between the two packages, but it doesn't go through autopilot -- it tests at the scope API level [09:43] ok, probably makes sense if someone develops some runtime test then ... [09:44] ogra_: Unity8 is currently configured to show music from the "mockmusic" scope at present, so it doesn't actually show up in the UI yet [09:45] right, but if it does it should get testing :) [09:45] asac: let's postpone your further daily release config questions to Monday, my post-FF+UDS head can't take more tasks at the moment than the ones I have now [09:45] being able to show some real data on the phone in the music scope is probably a precondition to get to a point where it can use the real music scope and those tests can be written [09:45] (i didnt say i want to see a test right now, but someone should talk to QA about runtime tests) [09:45] agreed. [09:51] Mirv: absolutely :) [09:51] asac: looking [09:52] asac: so dconf-cli is a bug, should get back on the image [09:52] asac: it is used to set an ofono settting; I dont know what that breaks [09:52] asac: see https://code.launchpad.net/~lool/telephony-service/add-missing-dconf-cli-dep/+merge/183055 [09:53] asac: the aptdaemon / packagekit stuff should be ok [09:53] lool, uh, then ofono should have a dep on it [09:53] asac: I dont know why we're pulling libcheese (camera app I guess), that probably explains all the GNOME deps [09:54] ogra_: that's what the mp adds [09:54] it's telephony-app's ofon-setup script IIRC [09:54] ah, ok [09:54] i was wondering why you didnt add it to ofono itself ... fine then [09:55] cheese comes in with gnome-c-c i think [09:55] I dont know, I just spotted it in passing in the logs in my home dir [09:55] which all the indicators depend on [09:56] diwic, available? === melmoth_ is now known as melmoth [09:56] lool: yeah. seb alreadyu explained libcheese with g-c-c which is being worked on [09:56] mamenyaka, somewhat [09:56] lool: thanks. sounds all is good and in hands [09:57] diwic, I am fresh installing 0830 to test the sound [09:58] mamenyaka, on what device? [09:59] diwic, Sony Tablet Z for starters, it has the apq80... card [09:59] mamenyaka, right [09:59] mamenyaka, chances are you need to fix up or completely rewrite the UCM files before it's working, but feel free to try [10:00] diwic, that's great [10:01] diwic, and later, I have a Note 10.1, with a card named Midas_WM1811 [10:01] seb128: regarding icon in click packages, in a hypothetical multi-app package this would need to be per-app not per-package, right? [10:03] mpt, ^ do you have an opinion? [10:03] That is, it's basically associated with each desktop file [10:03] cjwatson, I didn't know we would have multiple apps in a click [10:03] The design permits it [10:03] cjwatson, is each app having it own subdir/datas in that case? [10:04] I don't know how much we'll use it in practice but we decided to design it in from the start rather than having to retrofit it [10:04] They share an apparmor profile and I think a data directory [10:04] ogra_: I sympathise with your connection now... 8% [==> ] 4,281,428 3.90KB/s eta 31m 5s [10:05] heh, well, mine is usually around 300k :) i'm not syuffering *that* much :) [10:05] wow, what a nice tutorial with the new image! === gusch is now known as gusch|lunch [10:08] cjwatson, the design mockup is https://wiki.ubuntu.com/AboutThisDevice?action=AttachFile&do=get&target=phone-storage.png [10:08] So that's probably just per-package [10:08] cjwatson, my gut feeling is that the list should be a list of clicks, not apps, but I'm unsure ... mpt is the designer and he might have an idea [10:09] I think it will be a lot easier to implement per-package really and we should probably just do that [10:09] And make more sense given possible shared data between apps in the same package [10:09] cjwatson, if a click has app A-B-C and they are listed as 3 it becomes hard to know how much space you win by uninstalling that click for example [10:09] Exactly [10:09] diwic, is the Music app reliable? it hangs white [10:09] in settings, no sound [10:09] OK, so this is just a matter of recommending an "icon" entry at the top level of the manifest? [10:10] The existing official packages don't have anything like that yet so free to make one up [10:10] cjwatson, yes, basically ... I'm throwing the "click list --manifest" json output to a qml model [10:10] mamenyaka, we have had other people testing the music app and it's working [10:10] and using the title/icon to set the listitems text and icon [10:10] mamenyaka, on the nexus 4 [10:11] diwic, it worked for me too some images before, but now it's not working [10:11] mamenyaka, but start by testing "paplay /usr/share/sounds/alsa/Front_Left.wav" e g [10:11] mamenyaka, and see if it hangs or seems to play back but silent output [10:11] mamenyaka, i e quits after a few seconds [10:12] mamenyaka, paplay is in pulseaudio-utils [10:12] seb128: (installed-size is done in click trunk btw) [10:12] and where do I find the pulseaudio utils? [10:12] cjwatson, (great, thanks) [10:12] seb128: you'll need to get Sergio to add "icon" entries to the official click packages [10:13] seb128: would you like to suggest a one-line description of the "icon" key for the file format spec? [10:13] cjwatson, once they have it then the manifest is going to pick the info? [10:14] how do you mean? [10:14] "click list --manifest" just dumps the manifest out ... [10:16] cjwatson, ok, I'm unsure I understand what you call "official click packages"/what Sergio needs to do [10:16] the spec needs to be updated [10:16] and then the existing packages need to be updated? [10:17] diwic, # paplay /usr/share/sounds/alsa/Front_Left.wav [10:17] Connection failure: Connection refused [10:17] pa_context_connect() failed: Connection refused [10:18] seb128: Sergio maintains the code that creates the packages on http://people.canonical.com/~ubuntu-archive/click_packages/ [10:18] seb128: And yes, the existing packages need to be updated with an icon (I can't magic that out of nowhere) and rebuilt with click 0.4.x once released [10:19] (for installed-size) [10:19] mamenyaka, sounds like pulseaudio is not running. Try killing pulseaudio processes if there still are any and start it again, e g with "pulseaudio -vvvv" if you want a debug log too [10:19] Other click packages will presumably follow along in time too, but it's important to have the core apps follow current standards [10:19] cjwatson, ok [10:20] seb128: so can I have that description of icon? in particular I need to know if any image file at all will do or if there are limitations on its format [10:20] cjwatson, description ... we can copy the desktop spec one: "Icon to display in interfaces listing click packages. If the name is an absolute path, the given file will be used. If the name is not an absolute path, the algorithm described in the Icon Theme Specification will be used to locate the icon. " [10:20] (slightly adapted the "to display in..." === jibel_ is now known as jibel [10:20] OK [10:21] so no particular limitations, you'll just rescale if needed? [10:21] cjwatson, does that work for you or do you needs specifics on the format? [10:21] right [10:21] that's fine, I probably don't need to be more specific than the desktop entry spec is [10:21] just doing what is done for .desktop files seems logical to me [10:22] great [10:22] diwic, ok, now getting output with paplay, but no sound [10:22] mamenyaka, then you might need to adjust the UCM files in /usr/share/alsa/ucm/apq... to make it work with your device [10:23] seb128: actually we should refine the bit about paths [10:23] diwic, so I have apq8064-tabla-snd-card [10:23] diwic, and is this okay? cdev "hw:apq8064tablasnd" [10:23] seb128: I would suggest that the path must either be relative to the base of the click package (not absolute), or it should be icon-theme [10:24] cjwatson, that makes sense in context of the clicks yes, +1 from me [10:24] seb128: which means that the code using this will need to try to resolve it relative to the click base, and then fall back to icon-theme - it can't test for absoluteness first [10:24] seb128: we could include absolute path as well if you think that would be useful but I'd be inclined to say it's normally a bad idea [10:25] ok [10:25] mamenyaka, cat /proc/asound/cards to see [10:25] mamenyaka, also, does "pacmd list-sinks" show stuff like "[Out] Speaker" ? [10:25] but I suppose it would be weird not to support it [10:25] cjwatson, no, I don't think we need absolute path [10:25] mamenyaka, in the "ports" section [10:25] I'll just not specify it, if code happens to support it - e.g. os.path.join(base, abs_path) => abs_path - then so be it [10:25] diwic, No PulseAudio daemon running, or not running as session daemon. [10:25] wfm [10:25] hmm [10:26] mamenyaka, and the pulseaudio you started is still running? [10:26] how am I going to check the passcode / passphrase? [10:26] diwic, I'm guessing not [10:26] the shell is frozen [10:27] but it shows up in ps [10:27] maybe it works better if you have a better UCM file [10:28] diwic, killed, restarted, now ok [10:28] got all three out [10:28] diwic, [Out] Speaker: Main speaker (priority 100, latency offset 0 usec, available: unknown) [10:28] and an empty properties [10:29] mamenyaka, cool. So then the card name is probably right at least [10:29] seb128: http://bazaar.launchpad.net/~click-hackers/click/trunk/revision/216 [10:29] mamenyaka, now for the ucm file, the changes you need to make are very device specific and might need some detective work to figure out [10:29] diwic, so now where to poke the ucm file? [10:29] diwic, oh [10:30] mamenyaka, in my case (nexus 4), there was a file /system/etc/snd_soc_msm/* that contained very good hints of what the mixer should be set to [10:30] diwic, I have files there [10:30] cjwatson, great, thanks a lot! [10:30] diwic, snd_soc_msm_2x snd_soc_msm_2x_Fusion3 [10:30] seb128, cjwatson: Oh, no, not again. I had this same headache with USC. :-] [10:31] mamenyaka, btw, did you read my blog post / announce email which contained some basic info on UCM? [10:31] diwic, of course [10:31] mamenyaka, e g remove the VoiceCall stuff on tablets [10:31] diwic, but that didn't help me much [10:31] cjwatson, I guess I should wait for that to be released before pinging Sergio? [10:31] diwic, yes, did that [10:32] seb128: He can add the icon any time, but he'll need to wait for 0.4.0 before he can rebuild and get installed-size [10:32] ok [10:32] mamenyaka, ok. I'm not sure which of the two files (2x or 2x_Fusion3) that is the right one for your platform, but look through them both and look for hints on how to set up the speaker. [10:32] mamenyaka, then try to adjust the UCM file to fit [10:33] seb128, cjwatson: Is it possible to get the icons of all the apps in a package? [10:34] We don't have a specification for per-app icons yet in the manifest, except by going through the desktop file [10:34] So not really [10:34] Ted has been suggesting that I rename "hooks" to "apps" or something, which would make it more reasonable to add it there [10:35] Though it would still amount to duplication with respect to the desktop file [10:35] diwic, I have the files here: http://paste.ubuntu.com/6043594/ [10:35] mamenyaka, and restart pulseaudio when you want to test your changes [10:35] I guess a sufficiently smart tool could walk the manifest, find desktop files, parse each one, find the icon [10:36] Maybe that isn't even unreasonable [10:36] diwic, okay [10:36] diwic, thank you for all the help [10:36] SectionDevice [10:36] Name "Speaker" [10:36] Comment "Speaker Rx device" [10:37] mamenyaka, ^ that's where to start looking [10:38] mamenyaka, yw - I'll go for lunch now [10:44] ubuntu touch on nexus 4 is so nice [10:44] asac, FYI http://people.canonical.com/~ogra/touch-image-stats/ ... latest 15min after a new image was published you should have a diff log there [10:46] ogra_: awesome! [10:47] asac: were you wanting to display something like that on the dashboard? [10:47] ogra_: thanks for publishing that package change script. I think it'll be useful. :) [10:48] cjohnston: we want many things, yes. [10:48] :) [10:49] cjohnston: not only do i want to see the diff from previou image, i also would love to be able to interactively diff builds [10:49] like compare build from 1 week to todays [10:49] cjohnston: why are you awake? [10:49] :) [10:49] * asac tries to do the math [10:49] * cjohnston blames you [10:50] thats not nice. now i feel bad again :) [10:50] hehehe [10:50] its almost 7 [10:50] cjohnston: what have i done? :) [10:50] ah so you are a morning person? [10:50] ok then.. I blame school aged kids [10:50] not at all [10:50] so you slept before now :)? [10:50] thats what i was worried about [10:50] lol [10:51] Ya.. I went to bed a couple hours ago [10:51] * cjohnston wonders more about ogra_ script that may help make cjohnston's life a little easier ;-) [10:51] yeah [10:51] ogra_'s scripts are always open source [10:52] so you should be able to get it :) [10:52] cjohnston, feel free to grab, modify, integrate somewhere etc :) [10:52] its in the same dir ... a usage description is in the README [10:52] if i find the time i might integrate that stuff into cdimage directly [10:52] diwic, !!! headphones working! [10:53] so we just get some ".package.diff" filr in the publishing dir for that [10:53] *file [10:54] Thanks ogra_ [10:55] :) [10:59] oSoMoN: so your webbrowser fix is in the archive? [10:59] * asac wonders if we should try to validate that it fixed all with an image kick [10:59] asac, oh, in case you dont know it: http://paste.ubuntu.com/6039695/ seems jodh wrote something similar to the systemsettle test [11:00] kicking something sounds fun [11:00] but more focused on processes than just iterating over top output [11:00] jhodapp: hey... wanna chat about that script? :) [11:00] asac: it’s been released, it should be in the archive (allowing for the necessary latency to propagate there) [11:01] (well, it screen scrapes top too after all) [11:01] jhodapp: basically, want to understand if there is a low hanging fruit for improving automation with your script [11:01] asac, wrong jodh :) [11:01] jodh: ^^ :) [11:01] sorry [11:03] hi all - I've some ideas how a bookmakr feature could be implemented [11:04] where to post it? wiki? bug ticket? send some one a mail? === gatox is now known as gatox_bbl [11:04] asac: you mean QA testing? It would be easy for QA to run that script every 'n' seconds in a test environment and fail the test if it finds anything chewing cpu I guess. [11:05] thibaut`, whats wrong with the existing bookmark feature ? [11:06] thibaut`, well, the best would to file a whishlist bug against the webbrowser-app i think === hikiko is now known as hikiko-lunch [11:07] ah ! [11:07] well, then against the file manager [11:07] ok [11:07] thx [11:10] jodh: so what we currently do is that we run something that waits n-minutes and fails if the system never goes idle [11:10] jodh: does this script do the same or more or something different? [11:11] asac: my script just gets a snapshot of the top cpu hogs. Then, next time it is run, if those same pids are still above the specified cpu threshold, it will print an error message. [11:12] asac: I was actually hoping to get it, or something rather more sophisticated onto the images and hooked into apport since QA will never find all possible "spinners". [11:12] diwic, I have active port: <[Out] Headphones>, even when there are no headphones [11:12] well, the systemsettle test does something like that on boot [11:12] but not as fine grained as yours [11:13] so QA will catch CPU hogs as long as they happen on boot [11:14] ogra_: ok, great. However, I still feel there is value in having something like this on an image. Same rationale as whoopsie+apport: we can't test every possible combination of programs that might suddenly "go rogue" :) [11:14] yeah, i didnt debate to have it on the image too :) [11:15] i just think it might also improve the test QA already runs :) [11:16] The current feeling is to rewrite the script in C, potentially as part of whoopsie. However, we're at the "exploratory" phase atm. The other issue is what you do if you find a hog. Since such a facility should work on all devices, we don't want to kill long-running RDBMS services on a server, or gcc/firefox/etc on a desktop. [11:18] for those not following the thread, my proposal there was to leave GUI applications to the hanging application check that will eventually go into Mir [11:18] and just handle background processes with jodh's proposed script [11:19] you should talk to ricmm ... it sounds like a natural place for this to live would be the application lifecycle handling [11:22] hmm, i can't seem to flash this maguro. http://paste.ubuntu.com/6043726/ [11:22]