/srv/irclogs.ubuntu.com/2013/04/30/#ubuntu-devel.txt

mdeslaurinfinity: could you pocket-copy xen from raring-security to saucy, pleeze?00:01
infinitymdeslaur: Maaaaaybe.00:02
mdeslaurinfinity: sure you will, cause you're nice and all :P00:02
infinitymdeslaur: Done.00:02
mdeslaurinfinity: cool, thanks :P00:02
infinitymdeslaur: To be fair, you could have done it too.00:03
=== oSoMoN__ is now known as oSoMoN
mdeslaurinfinity: oh?00:03
infinitymdeslaur: copy-package -s raring-security --to-suite=saucy-proposed -b xen00:03
infinitymdeslaur: You should be able to copy to any pocket you'd be able to upload to.00:04
mdeslaurinfinity: oh, I didn't know that...thanks, I'll try it next time00:04
infinity(At least, I believe that's the case...)00:05
ogra_RAOF, huh ? i thought it uses libinput00:06
RAOFReading from evdev00:07
ogra_we dont even remotely have support for evdev on the touch images00:07
cjwatsoninfinity: Indeed.00:07
ogra_RAOF, that will cause massive issues with the kernels00:08
ogra_we would have to patch them to actually provide evdev devices instead of talking to libinput00:09
slangasekthis seems like an architecture question for Phonedations to solve00:09
slangasek;)00:09
ogra_well, lwould have been nice to know :)00:10
ogra_we will definitely have issues with that ...00:10
RAOFogra_: So what's the actual problem?00:10
ogra_RAOF, did you guys actually test mir on touch images (vs on the nexus7 desktop one for example)00:10
RAOFBecause I'm pretty sure that we're getting input events on Mir on the phone.00:10
ogra_RAOF, that the assumption is that we will go on using the android HAL layer for all devioce interaction (including input)00:11
* ogra_ needs to relocate ... lets discuss that later (and probably in person)00:11
slangasekogra_: implying that we would be using android hal even for Mir on desktop images?00:11
RAOFogra_: So, as far as I can tell, Mir uses libinput, which reads evdev from /dev/input/*.00:12
slangasekah00:12
zygadiwic: out of curiosity, the kernel associated with bug https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1169984 is not fixing the problem on my hardware10:06
ubottuLaunchpad bug 1169984 in linux (Ubuntu Raring) "3.8.0-18 HDMI audio regression: Either oops or opening device fails with -ENODEV" [High,Fix committed]10:06
diwiczyga, okay; what problem are you having with that kernel?10:06
zygadiwic: I'm updating the bug report10:07
diwiczyga, please don't10:07
diwiczyga, you're probably having a different bug then10:07
diwiczyga, the last thing we need is all different problems in the world in the same bug10:07
zygaah10:08
zygatrue10:08
diwiczyga, or maybe you have both 1169984 and another bug10:08
zygadiwic: I'll check if actual audio got fixed10:09
zygadiwic: audio works but is distorted10:10
zygadiwic: but it's better than vanilla 13.04 kernel10:10
diwiczyga, ok, so that's a different bug then10:10
diwiczyga, still a regression though, if the audio was non-distorted on 12.10 ?10:11
zygalet me just triple check on 12.04 to ensure that audio was not distorted there10:11
zygadiwic: I don't have 12.10 here10:11
diwicok10:11
zygadiwic: just 12.04 and 13.0410:11
zygadiwic: yes, perfectly clear audio10:12
diwicok10:12
diwiczyga, please run "ubuntu-bug audio", note if both test tone sets are distorted or just the second one, and point me to the resulting bug.10:13
zygadiwic: on 13.04, right?10:15
diwiczyga, yes.10:15
zygarebooting10:15
zygadiwic: should I expect some test playback during `ubuntu-bug audio`10:20
diwiczyga, yes10:20
zygadiwic: presented with the question "in what way the sound quality is bad" I'm uncertain10:20
diwiczyga, just pick one10:20
zygaI did't hear any sound10:21
diwiczyga, ok10:21
zygadiwic: the initial dialog lists several devices10:21
zygadiwic: many have the same name10:21
diwiczyga, what same name, "HDMI" ?10:21
zygadiwic: I selected the second of the group, hoping it would be my 'hdmi 2'10:21
diwicok10:21
zygadiwic: "Digital Out, HDMI"10:21
zygadiwic: I just tried another, still no sound10:22
diwiczyga, sounds like my apport symptom could be improved...10:22
zygadiwic: hmm I suspect ubuntu-bug is broken there10:22
zygadiwic: the next dialog is shown instantly10:23
zygadiwic: maybe missing dependency? can I somehow see what is going on there10:23
zygadiwic: I suspect it never actually tried to play anything, having crashed in the attempt10:23
diwiczyga, don't bother troubleshooting the troubleshooter at this point10:23
zygaok10:23
zyga:D10:23
diwiczyga, maybe just file a bug with "ubuntu-bug pulseaudio" instead10:24
zygaok10:24
zygadiwic: bug 117469610:32
ubottubug 1174696 in pulseaudio (Ubuntu) "Regression in sound quality on 13.04 as compared to 12.04 (HDMI over DP) on Thinkpad W510" [Undecided,New] https://launchpad.net/bugs/117469610:32
diwiczyga, hmm, crackling and metallic feeling sounds like https://wiki.ubuntu.com/Audio/PositionReporting - could you try the stuff there and see if it makes a difference?10:34
zygadiwic: looking10:34
zygabrb10:35
zygauh10:50
* zyga is sick10:50
zygadiwic: testing position fix=110:57
zygadiwic: using =1 makes it worse, sound cuts off after a second and does not go back11:02
zygadiwic: trying =211:02
zygadiwic: =2 is the same as without adding that option11:04
zygadiwic: tryig the scheduling trick11:04
zygadiwic: module-udev-dectect has use_ucm=0 in my default.pa, is that oka?11:06
diwiczyga, ucm is unrelated11:07
diwiczyga, ucm is for embedded devices11:07
zygadiwic: tsched did not affect this11:07
zygadiwic: if you want I can record the playback11:08
diwiczyga, ok, not sure what's wrong with it then11:08
zygadiwic: despite using LANG=C pactl list uses some localized strings, do you know where they might come from?11:28
diwiczyga, try LC_ALL=C or LC_MESSAGES=C11:28
zygadiwic: nope11:30
zygadiwic: so stuff that pactl prints itself is in english11:30
zygadiwic: but Description fields are localized11:30
zyga        Description: Wbudowany dźwięk Analogowe stereo11:30
diwiczyga, right, that's not much to do about in pactl11:30
zygadiwic: so where does that come frome?11:31
diwiczyga, that's translated in PulseAudio, you would have to run the entire pulseaudio with LC_ALL=C to get rid of that11:31
zygaohhh11:31
zygaok11:31
zygathanks for the tip11:31
zygathat means we may not be able to use Description in any heuristic11:31
diwiczyga, but then sound settings will no longer be localized for you either11:31
diwiczyga, well, there are "names" to every description you should use instead for heuristics11:32
zygadiwic: right11:34
zygahmm11:44
zygadiwic: is the hot-plug issue affecting both hdmi and dp?11:45
diwiczyga, what hot-plug issue?11:50
diwiczyga, no dynamic re-probing on hotplug? Yes, that affects both.11:50
zygadiwic: ok11:52
zygadiwic: thanks11:52
zygadiwic: I found that I don't see hdmi, after I plug it, it is there, then if I un-plug it, it seems to stay there11:52
diwiczyga, there was a bug that ELD info is still valid after hotplug, that I fixed recently - it probably has not reached 3.2 kernel yet, if it ever will.11:53
zygadiwic: this is on raring actually11:53
zygadiwic: I have a draft plan for the test \]program11:53
zygadiwic: https://gist.github.com/zyga/548824311:54
diwiczyga, ok, I'd typically go for just choosing the "stereo" option, i e remove all that do not start with "hdmi-stereo"11:55
diwiczyga, or do you typically want to test 5.1 surround too?11:56
zygadiwic: good question11:56
zygadiwic: no, let's keep 5.1 out of it11:57
diwiczyga, btw, you only have to select the profile on the card; you don't need to set the default sink, instead specify what sink you want playback on when you play the sample sound11:59
=== jbicha is now known as Guest90087
=== Guest90087 is now known as jbicha_
zygadiwic: ah, that's good12:07
zygadiwic: thanks12:07
zygadiwic: I'd like to use dbus api so that I don't have to parse all of that12:07
zygadiwic: what is the best place to learn about this?12:08
diwiczyga, about the dbus API? It's not even enabled in 12.04.12:10
zygaoh12:10
zygaso what other choices do I have?12:10
zygaI'm good with C12:10
diwiczyga, for a tool such as yours pactl + paplay seems the simplest option IMO12:11
diwiczyga, otherwise, the native API http://freedesktop.org/software/pulseaudio/doxygen/12:12
zygadiwic: for the simple approach, I'd have to parse the output of pactl list cards; right?12:12
diwiczyga, right.12:13
zygadiwic: is it any formal thing? yaml?12:13
diwiczyga, there might be some python code in my apport symptom.12:13
* zyga mutters something about inveinting output formats12:13
zygainventing12:13
diwiczyga, can't be worse than inventing libraries, can it ;-)12:14
diwiczyga, I don't think it's formalised.12:15
zygadiwic: haha, good point12:16
diwiczyga, use the C API if you want something more guaranteed to be stable12:16
zygadiwic: ok12:16
zygadiwic: actually12:16
zygadiwic: a feature to output json would be a nice patch on top of pactl list12:16
diwiczyga, patches welcome12:16
zygadiwic: that would be a good compromise on using the c api12:16
zygadiwic: and I would rather have that with some simple python side12:16
zygadiwic: rather than doing a custom program just for that12:16
zygadiwic: thanks12:16
diwiczyga, I believe json output from pactl would be encouraged by upstream12:16
zygadiwic: yeah12:17
zygadiwic: I think that would be good12:17
zygadiwic: although with dbus api that might be seen as a useless addition12:17
zygadiwic: anyway12:17
zygadiwic: let's assume that's my plan12:17
diwiczyga, well, question is if you actually submit json patches to upstream, how do you get it into 12.04...?12:18
zygadiwic: that's a valid question but I suspect I know somoene that knows the answer12:19
zygadiwic: ara is gone now12:19
zygadiwic: I'll know in an hour or so12:19
zygadiwic: I suspect that we could backport pactl into our testing ppa, it depends on exactly what needs to be certified12:20
diwiczyga, I have upload rights for PulseAudio in Ubuntu and commit rights to PulseAudio upstream.12:20
zygadiwic: ok12:20
diwiczyga, and honestly, from an SRU point of view, it seems easier just for you to parse pactl as it is12:21
diwiczyga, I've not done that many SRUs for pulseaudio but still I think I might want more motivation than "slightly more convenient interface for an internal testing tool" to give to the SRU approvers12:22
zygadiwic: hmm, perhaps12:22
zygadiwic: is there any way to access the dbus api on 12.04 or is that just gone?12:22
diwiczyga, yes, "load-module module-dbus-protocol"12:23
diwiczyga, it's disabled because there are race conditions on card unplug that can make PA crash.12:23
zygadiwic: ah12:23
zygadiwic: ok12:24
zygadiwic: I guess a parser is in order then12:25
zygadiwic: thanks for your help12:25
zygadiwic: I'll let you know which approach we'll take12:25
diwiczyga, ok, good luck :-)12:26
hallynSpamapS: could you push the libvirt raring-proposed?12:45
=== _salem is now known as salem_
=== wedgwood_away is now known as wedgwood
zygadiwic: quick question, how crazy would it be to load the pa dbus module just for testing?14:06
zygadiwic: and unload it later14:06
zygadiwic: doable?14:06
diwiczyga, I guess14:18
diwiczyga, I don't know enough about the dbus module to know if it fits the purpose14:19
zygadiwic: ok14:21
zygadiwic: I guess parsing is really the only option then :)14:21
zygadiwic: I'll make a library for that14:21
rbasak  /win 1914:58
=== mmrazik is now known as mmrazik|afk
=== salem_ is now known as _salem
=== Ursinha-afk is now known as Ursinha
zygadiwic: we've decided to create a custom parser15:36
zygadiwic: and work upstream to file bugs15:36
zygadiwic: about pactl list being json parsable15:37
zygadiwic: about the discrepancy between x and pulse15:37
zygadiwic: and anything else we may encounter15:37
zygadiwic: we'd like to get a better upstream relationships15:37
=== _salem is now known as salem_
jibelxnox, https://code.launchpad.net/~jibel/ubuntu/saucy/python-apt/AddSaucy_lp1174562/+merge/16163115:56
stgraberxnox: 15:56 < jibel> xnox, https://code.launchpad.net/~jibel/ubuntu/saucy/python-apt/AddSaucy_lp1174562/+merge/16163115:58
jibelxnox, https://code.launchpad.net/~jibel/ubuntu/saucy/python-apt/AddSaucy_lp1174562/+merge/16163115:58
=== barry` is now known as barry_
=== barry_ is now known as barry
hallyninfinity: could you push https://launchpad.net/ubuntu/raring/+queue?queue_state=1&queue_text=libvirt (raring sru for libvirt)?16:31
hallyn(if i understand right that you're part of sru team)16:31
infinityhallyn: I might be.16:39
infinityhallyn: There are precise and quantal tasks for this too, what's up with that?16:40
cousteauChangelog for aptitude:   * Apply upstream multiarch-conflicts.patch to handle conflicts on multi-arch systems.16:47
cousteauok, it only took 2 years to solve...16:48
cousteauanyway, does this mean aptitude is safe now?16:48
hallyninfinity: yeah those will be needed too16:53
infinityhallyn: Hrm.  I like that there are 3 of the in the quantal queue...16:54
hallyninfinity: yup16:54
infinityhallyn: The precise one is missing a bug ref in the changelog.16:55
infinityhallyn: Pls to be fixing.16:55
hallynk  (the quantal one is too i think)16:55
infinityhallyn: The Q one has the bug ref a few lines up.16:55
hallynah yes i see16:56
hallynre-pushing precise16:56
hallyn(pushed)16:57
=== kentb-out is now known as kentb
=== deryck is now known as deryck[lunch]
=== salem_ is now known as _salem
=== _salem is now known as salem_
xnoxstgraber: what does "lxc-start: invalid sequence number 1. expected 2" means?17:42
cjwatsonpitti: Are you already working on the systemd/powerpc build failure?17:58
cjwatsonThe udev/systemd cluster tickles some kind of excitingly pathological path in proposed-migration so I'd like to get it sorted out fairly sharpish :)17:59
pitticjwatson: was in a meeting, but I'll sort it out now18:22
cjwatsonpitti: If it helps, the full test output is http://paste.ubuntu.com/5619996/18:30
pitticjwatson: yeah, a lot of the tests don't work in our buildd environment unfortunately; I'll just disable that one for now; we don't use the journal anyway18:31
cjwatsonOK, cool, thanks18:32
pitti(uploaded)18:32
=== cjohnston_ is now known as cjohnston
=== deryck[lunch] is now known as deryck
robert_ancellpitti, https://launchpad.net/~ubuntu-desktop/+archive/ppa/+sourcepub/3151169/+listing-archive-extra19:06
pittirobert_ancell: \o/19:06
Laneyph33r19:07
* pitti reboots19:08
=== mmrazik|afk is now known as mmrazik
pittirobert_ancell, Laney: works fine here, and now properly kills the screensaver on user switching19:13
Laneyphew19:13
pittiindicator-session still doesn't show the users, though; but that's marked as pending merge?19:13
robert_ancellpitti, nice19:13
Laneyright, try that from trunk19:14
pittior is that just for suspend/shutdown?19:14
Laneycyphermox: ^ do you know when dailies for indicators are starting up?19:14
cyphermoxthe plan is actually possibly to not start them until after things can be merged for desktop/touch19:14
Laneyhuh19:15
cyphermoxto give time to do the merges, fix things, make sure it all works everywhere to some limited degree of working19:15
Laneyso can we get indicator-session (and datetime) in for logind somehow? :-)19:15
cyphermoxyes, we should19:15
cyphermoxI think we'll do one such run, and then disable it19:16
pittiOOI, can we just do a manual upload of PS packages, or would that confuse the autolanding machinery?19:16
=== smb` is now known as smb
Laneyyes to both ;-)19:16
cyphermoxLaney: I'll come see you, we need to discuss more stuff related to logind -- I've heard many stories about suspend and such not working, but it's fine here. I just want to make sure I'm set up properly19:16
cyphermoxpitti: you can do a manual upload of PS packages, yes19:17
cyphermoxit doesn't really confuse things, but it gets detected and means we need to merge the changes in afterwards19:17
Laneycyphermox: there was a big bug with suspend that pitti fixed yesterday19:17
cyphermoxbut as above ^ I guess we'll run the indicators at least once, and then disable them19:17
cyphermoxI'm supposed to enable all the daily stuff today19:18
pittithe only broken thing known to me right now is that closing the lid doesn't lock the screen (but does suspend fine)19:18
cyphermoxor you know, at least for a few of the packages19:18
cyphermoxpitti: interesting. it's been working for me forever, didn't ever run into such a bug19:18
cyphermoxI upgraded to saucy on thursday or something19:18
cyphermoxupdated yesterday morning19:18
pitticyphermox: well, "forever" means "for the last two days?" :-)19:19
cyphermoxyes ;)19:19
Laneythe suspend bug was up until you restarted your session after dist-upgrading to saucy19:19
cyphermoxforever given the current lifetime of saucy ;)19:19
cyphermoxah19:19
cyphermoxthen I would have never seen it, yes19:19
Laneyalmost melted my bag at lunch yesterday :P19:19
cyphermoxhaha ;)19:19
cyphermoxglad to know it's just that I restarted my session at the right time19:20
cyphermoxpitti: we'd be running into a similar issue with the NM session tracking changes, too19:20
pitticyphermox: but we don't restart NM on upgrades, do we?19:21
LaneyNM is fine here and I still haven't restarted since dist-upgrading19:22
pittiLaney: want to see how far you can get with the in-memory processes without rebooting? :-)19:24
pittilunch o'clock19:24
* Laney is hardcore19:25
jbicha_robert_ancell: why does Ubuntu have liblightdm-qt-2-dev but Debian has liblightdm-qt-dev? there's only 2 rdepends but it doesn't seem worth keeping a diff with Debian for that19:29
=== Ursinha is now known as Ursinha-afk
=== G4MBY is now known as PaulW2U
=== vibhav is now known as Guest20326
=== Ursinha-afk is now known as Ursinha
cyphermoxpitti: indeed we don't restart NM; we just use the reboot notifier.20:45
=== smb` is now known as smb
Laneypitti: have you looked at porting apport?21:14
pittiLaney: not yet, but on my list21:14
pittiLaney: in fact, I'm currently working on apport anyway for slangasek's upstart logs, so I can do that right now21:15
Laneyah, OK, was going to look but go ahead21:15
pittiin_session_of_problem() ... skipped 'no ConsoleKit session'21:16
pittihah, tests even pick that up :)21:16
=== salem_ is now known as _salem
=== _salem is now known as salem_
Laneybah, reimplementing GetCurrentSession everywhere sucks21:32
pittiLaney: this does nicely for me: http://bazaar.launchpad.net/~apport-hackers/apport/trunk/revision/263021:33
pittiI'll do an upstream apport release now and upload21:33
Laneyah, very clever21:35
pittino d-bus any more21:35
pittiand dropping the requirement of GI for apport21:35
pittierr, no, ignore that (the UI uses it, of course)21:35
Laneyah, gnome-user-share is done upstream, yay21:36
pittiLaney: it's also worth checking for patches on https://git.fedorahosted.org/cgit/21:37
pittifor projects which aren't maintained upstream any more21:37
* Laney nods21:39
=== kentb is now known as kentb-out
=== jibel_ is now known as jibel
pitticjwatson: mind if I steal the packagekit merge from you? I need a newer version for proper logind support22:34
=== salem_ is now known as _salem
pitticjwatson: ah well, I just upload it, and receive the blame later; it's a trivial merge anyway22:55
cjwatsonsure, that's fine23:02
pittiLaney: seems we are about halfway through :)23:04
pittiRAOF: FYI, colord FTBFS with "debian/rules build", I'll commit a fix to collab-main23:20
pittit23:20
RAOFpitti: Oh, where?23:20
pittiRAOF: oh, it is already fixed there23:20
pittisetting DEB_HOST_ARCH_OS23:20
RAOFRight.23:20
pittiRAOF: can we sync, or does this need a merge?23:20
RAOFIs libsystemd-login-dev in main? If so: sync.23:21
pittiyes23:21
pittiRAOF: syncing, thx23:21
RAOFOk.23:22
* RAOF forgot --force on his syncpackage, so you can take it :)23:22
pittiautopkgtest FTW :)23:23
RAOFI'll get you to upload 0.1.33 to experimental sometime this week too, if you're game.23:23
pittisure; my sbuilders are yearning for something to do :)23:23
pittiRAOF: or wait until Monday and upload it to unstable :)23:23
RAOFAlso win :)23:24
pitti(unless you need exp glib or so)23:24
RAOFDon't think so.23:24
pittiassuming that Debian releases on the weekend23:24
=== wedgwood is now known as wedgwood_away

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