/srv/irclogs.ubuntu.com/2013/06/17/#ubuntu-mir.txt

=== marlinc is now known as marlinc|away
=== marlinc|away is now known as marlinc
robert_ancellRAOF, my bzr productivity has gone up so much since you showed me init-repo - thanks!04:52
RAOF:)04:52
tvoss_RAOF, robert_ancell good morning :)05:13
RAOFtvoss_: Good morning!05:13
robert_ancelltvoss_, hello05:14
tvoss_robert_ancell, hey there :) How is it going?05:14
tvoss_robert_ancell, for the brittle dep on libmirserver: could we increase the patch version of mir at least? trying to parse bzr743 vs. bzr 745 is a little annoying :)05:16
robert_ancelltvoss_, what does that acheive?05:18
RAOFAnd isn't that going to be pretty much bumping the patch level for each bzr commit anyway?05:19
tvoss_RAOF, sure, it's just more readable :) which is the only benefit robert_ancell :)05:19
robert_ancelltvoss_, how is it more readable?05:19
tvoss_robert_ancell, 0.0.4 is more readable to me than 0.0.3bzr{revision}, but that's admittedly quite subjective05:20
RAOFI think it sounds more readable to you because you're thinking it'll be mir 0.0.6, mir 0.0.7, … Pretty quickly it's going to be mir 0.0.743 and mir 0.0.745, and I don't think you'll find that too much more readable :)05:20
robert_ancelltvoss_, yeah, there's a conflict between "old naming" and "new naming". Really the bzr number is the only one that matters. We just bump the main version number to close out bugs at the moment05:22
tvoss_robert_ancell, okay, cool05:22
robert_ancellIt should be set to what RAOF said, but no-one has done the work afaik05:22
RAOFOooh, that reminds me.05:23
RAOFTime to construct a symbols file for libmirclient!05:23
dufluRAOF: Umm, "make" ? :)05:24
dufluA symbols file is usually just a binary that's not stripped :)05:24
dufluOr you mean symbols deb? (!)05:24
RAOFI mean a file listing the exported symbols.05:30
RAOFThat gets compared to the *actual* exported symbols when the package is built, and causes the build to fail if you've changed the ABI (in that way).05:30
tvoss_robert_ancell, still about?05:38
tvoss_damn :)05:38
tvoss_thomi, ping05:46
tvoss_RAOF, unity-system-compositor fails to install right now due to unity-system-compositor : Depends: libmirserver0 (= 0.0.3bzr743saucy0) but 0.0.3bzr745saucy0 is to be installed05:49
RAOFYeah, just noticed.05:49
tvoss_RAOF, how can I help in pushing the correct version to the ppa?05:49
RAOFHm. And the last Mir build was 2013/06/15. The autorebuilder should have rebuilt unity-system-compositor.05:50
RAOFtvoss_: I think we need to prod the autobuilder, but I have no idea how to do that.05:51
RAOFtvoss_: I'm leary of manually uploading because I'm not sure that I can chose a version later than what's currently in there that won't confuse the autobuilder.05:52
tvoss_RAOF, ack, by autobuilder, you are referring to good ol' jenkins?05:53
RAOFtvoss_: I believe so, but it's not quite the usual CI infrastructure - each Mir build *should* provoke a new unity-system-compositor build.05:54
tvoss_RAOF, I think that's the issue, forward build dep propagation seems not to be working as expected05:54
RAOFOnce I've checked that it works I'll push a change to unity-system-compositor to disable the cursor, which should cause a rebuild.05:55
tvoss_RAOF, \o/05:56
tvoss_RAOF, how can people see that they are running XMir then?05:57
RAOFBy grepping /var/log/Xorg.0.log05:57
tvoss_RAOF, okay, something more user-friendly might be a good idea, but we can leave that for later :)05:57
tvoss_RAOF, can you ping me once the changes land? if that is eta today05:57
RAOFIt should be, but that does depend on my understanding of the code being correct :)05:58
RAOFAnd this stupid laptop being less slow. Argh!05:58
* tvoss_ remembers RAOF posts about a laptop refresh :)06:03
RAOFIndeed.06:03
RAOFA 240 GB SSD + a Haswell i7 should make things much more fun.06:03
tvoss_RAOF, I can only recommend that setup, although I don't have a Haswell, it's a core i7 vpro and an ssd in my x22006:06
tvoss_RAOF, quite fast it is :)06:06
RAOFPlus a terabyte of rotating rust to store all the other things.06:07
tvoss_RAOF, yup :906:08
tvoss_mlankhorst, ping06:11
tvoss_mmrazik, ping06:16
mmraziktvoss_: pong06:16
tvoss_mmrazik, good morning :) quick question: Do we have a solution for forward-build-dep propagation, yet?06:16
tvoss_mmrazik, aka the meta-build-system?06:17
mmraziktvoss_: should be there06:17
mmraziktvoss_: do you miss something in particular?06:17
mmrazikI think thomi/fginther were enabling something for mit06:17
mmraziks/mit/mir/06:17
tvoss_mmrazik, installing unity-system-compositor from the mir staging ppa fails right now due to unity-system-compositor : Depends: libmirserver0 (= 0.0.3bzr743saucy0) but 0.0.3bzr745saucy0 is to be installed06:18
tvoss_mmrazik, seems like unity-system-compositor has not been rebuild06:19
mmraziktvoss_: oh. I think its not dputing into PPA. It only rebuilds in jenkins to check for build failures I believe06:19
mmrazikthe PPA should go away fairly soon anyway06:19
mmrazikor not?06:19
tvoss_mmrazik, hm, any eta when we will land in universe?06:19
mmrazikdon't know06:20
mmrazikI guess didrocks knows more06:20
tvoss_mmrazik, just pinged him06:20
didrockstvoss_: mmrazik: I'm waiting kgunn to be back (which should be today, right?) to discuss about it with him06:21
didrocksI want Unity 8 (without indicators) + Mir to be in the next couple of weeks into distro06:21
tvoss_didrocks, right, should be back today06:21
tvoss_mmrazik, seems like the ppa will be with us for some time then :)06:22
mmraziktvoss_: I'll drop francis an e-mail. It might be that the answer will be to use an archive located on private API06:22
mmraziksince copenhagen (I believe) there was a push towards killing staging PPAs06:23
mmraziks/private API/private IP?06:23
tvoss_mmrazik, okay, my goal is to make installation from the ppa as straightforward as possible06:23
mmraziktvoss_: ok. so not just for canonical people...06:23
tvoss_mmrazik, right06:24
mmrazikthen the private archive is not an option06:24
mmraziktvoss_: sent an e-mail to francis. I think it should not be very hard06:25
tvoss_mmrazik, ack06:25
RAOFdidrocks: Is our autolanding infrastructure robust enough to handle “each Mir upload breaks unity-system-compositor, so don't publish it until unity-system-compositor has built against -proposed”?06:31
didrocksRAOF: we don't build enough into -proposed. For detecting ABI break and upstream not following the guide in such a case, you should provide some integration tests (those are run, and so, when failed, block the publishing)06:32
didrocksRAOF: FYI, this is what needs to get done when you are breaking the ABI: https://wiki.ubuntu.com/DailyRelease/FAQ#I_need_to_break_an_API.2BAC8-ABI06:32
didrocksbasically, it's about "bump the version in debian/changelog" and bump the build-dep on the project dep on this one06:33
didrocks(just to trigger the 2 rebuilds)06:33
didrocksthis in a coherent piece, before 00 UTC06:33
RAOFBah.06:34
RAOFWe can't just have Depends: libmirserver0 (= $THE_VERSION_BUILT_AGAINST_INCLUDING_DEBIAN_REVISION)?06:35
RAOFBecause we bump ABI (for mirserver) approximately each commit, and don't intend to stop doing that for a while.06:35
RAOFI thought the -proposed → archive transition checked that it didn't make the archive uninstallable.06:37
didrocksRAOF: -proposed -> archive does, but you don't want to break proposed everyday as well I think :)06:39
didrocksRAOF: what does mirserver has as build-deps?06:40
RAOFIt's built by mir06:40
RAOFI thought we could happily break -proposed every day?06:40
RAOF:)06:41
didrocksRAOF: well, everything that will dep on MIR will break then06:41
RAOFIs that a problem?06:41
RAOFThere's not that much that depends on mir.06:41
didrocksRAOF: in fact, thinking about it, if the packaging forces latest version, we won't be able to install all MIR-related package06:41
didrocksRAOF: this can even be a test!06:41
didrocksand so we block the upload to -proposed06:41
didrocksRAOF: unity8 will I guess, isn't it?06:42
RAOFYeah, once it's rebased on Mir.06:42
RAOFBut it, too, will be broken by every Mir commit :)06:42
didrocksRAOF: when do you plan to stabilize the ABI?06:42
RAOFOnce Unity 8 is substantially complete, I believe06:43
didrocksRAOF: if MIR is going to be broken everyday, there is no real point to push it to distro :p06:43
RAOFI thought that was the decision, yes :)06:43
didrocksany rough date idea?06:43
didrocksthe decision to not push it or to push it?06:43
RAOFYou'd need to talk to the shell team for that.06:43
RAOFThe decision *not* to push it.06:43
didrockshum, interesting, last time, the decision was to push it :)06:44
didrocksRAOF: anyway, I think the best to block publishing would be to have an integration test with a dep06:44
didrocksand this will block the publication as the test will fail06:44
didrocks(an autopilot tests as we standardized on it)06:45
RAOFOk06:45
tvoss_alf_, ping07:10
alf_tvoss_: pong07:11
duflummrazik: Is there a bug for "Fix committed at revision None..." ? As in: https://bugs.launchpad.net/mir/+bug/1191636/comments/207:40
ubot5Launchpad bug 1191636 in Mir "setup-android-dependencies.sh hung at "Do you want to continue [Y/n]?"" [Medium,Fix committed]07:40
dufluI keep seeing various bugs say "at revision None"07:41
mmrazikduflu: yes. I've seen it before but was too lazy too look it up. Let me check now.07:41
dufluI was too lazy too :)07:41
mmrazikduflu: mhm. Might be difficult to fix. It looks like some timing issue. The branch is already merged but launchpad seems to be reporting it is not (for a while).07:47
mmrazikThe only fix I can think of is sleep07:47
mmrazikbut obviously I'm not too excited about it07:47
duflummrazik: Yeah suspected as much. Thanks anyway07:47
duflummrazik: Is the a bug for it yet?07:47
mmrazikI'll create one07:48
* tvoss_ reboots :) 08:03
=== 13WAAK4RM is now known as tvoss
tvossRAOF, ping08:21
RAOFtvoss: Pong.08:38
tvossRAOF, did you push the update removing the orange pointer?08:38
RAOFNot yet. Turns out that it really wants a little bit more mir infrastructure, and I'm just doing that.08:39
tvossmmrazik, seems like psjenkins has just dput'ed to the ppa, leaving it in a broken state right now08:40
RAOFI should really file a “btrfs snapshots cause catastrophic performance degredation over time” bug08:51
mmraziktvoss: what was dputed? I can dput something manually to trigger a rebuild in PPA if you want (unity-system-compositor ?)08:51
tvossmmrazik, nope, ignore that08:51
mmrazikok08:51
xnoxRAOF: there are a couple, I believe..... known upstream issue as well08:51
=== hikiko_ is now known as hikiko
alf_tvoss: https://code.launchpad.net/~afrantzis/mir/platform-enablement-docs/+merge/169754 , let me know if you need more information or any refinements09:52
tvossalf_, thanks, looking09:53
arssonunity-system-compositor : Riippuvuudet: libmirserver0 (= 0.0.3bzr745saucy0) mutta 0.0.3bzr746saucy0 on merkitty asennettavaksi09:55
arsson?09:55
tvosskatie, ping11:44
kgunndidrocks: i'm in today ;)11:50
* alf_ upgraded to saucy and is not happy about the tedious gmock related warnings that g++-4.8 spits out11:55
=== greyback is now known as greyback|lunch
=== alan_g is now known as alan_g|lunch
tvossmmrazik_, any idea where this issue originates from: https://launchpadlibrarian.net/142643349/buildlog_ubuntu-saucy-i386.xorg-server_2%3A1.13.3%2Bxmir1-0_FAILEDTOBUILD.txt.gz12:12
tvoss=12:12
tvoss?12:12
tvossdidrocks, ping12:16
=== greyback|lunch is now known as greyback
tvossseb128, ping12:54
=== mmrazik_ is now known as mmrazik
didrockstvoss: pong12:59
tvossdidrocks, hey there :) I'm looking at https://launchpad.net/~mir-team/+archive/system-compositor-testing/+packages12:59
tvossdidrocks, trying to understand why the build failed for 38612:59
mlankhorsttvoss: looks like a broken builder tbh12:59
didrockskgunn: tell me once you have some time for a short meeting around mir and unity 8 in saucy12:59
didrockstvoss: what mlankhorst says12:59
didrockstvoss: did you ask the launchpad guys?12:59
seb128tvoss, hey13:00
tvossdidrocks, nope13:00
seb128tvoss, just retry the builds13:00
kgunndidrocks: got 30 min right now13:00
didrockskgunn: if you don't care seeing me with wet air (just back from some exercice), let me starts a hangout :)13:01
didrocksstart*13:01
kgunnsure13:01
=== alan_g|lunch is now known as alan_g
=== mmrazik is now known as mmrazik|afk
didrockskgunn: https://plus.google.com/hangouts/_/3c9650ce706154e7b24b272d2c14df4d5bc36fff13:02
alf_alan_g: Is there a reason TemporaryBuffers (in compositor/) assume that their buffer bundles may be destroyed while they are still alive? Which scenario does this assumption support?13:20
alan_galf_: sorry, not sufficiently familiar with the current code to answer OTTOMH13:21
alf_alan_g: np, for now I will assume that we don't need this scenario, and let Kevin prove me wrong :)13:22
alan_galf_: OK. FWIW I'd expect the bundle to outlive the buffers too13:23
tvossmlankhorst, ping13:29
mlankhorstpong13:29
tvossmlankhorst, hey :) not sure you are the right person to ask, but does fglrx support radeon 8xxx series?13:29
mlankhorstradeonhd or the ancient ones?13:30
tvossmlankhorst, the binary driver that is13:31
mlankhorsti know but https://en.wikipedia.org/wiki/Radeon_8000_Series or https://en.wikipedia.org/wiki/Radeon_HD_8000_Series :P13:33
katietvoss pong13:34
tvossmlankhorst, http://www.amd.com/us/products/desktop/graphics/8000/Pages/8000-series.aspx#213:35
mlankhorsti dont see it in their official notes, so perhaps not13:37
mlankhorsthttp://support.amd.com/us/kbarticles/Pages/AMDCatalyst13-6LINBetaDriver.aspx13:38
tvosskatie, hey there. did you see my comments on the surface mgmt doc?13:39
katietvoss, hi. no, haven't looked at that doc today13:40
katietvoss, but now you mention it, I will have a look later on :)13:40
tvosskatie, first wave of feedback, stay tuned :)13:40
katietvoss, cool, thanks13:41
=== mmrazik|afk is now known as mmrazik
tvossdidrocks, can we have apport integration for packages not in main?13:57
didrockstvoss: sure, this has nothing to do with main or not main13:58
tvossdidrocks, great :)13:58
didrockstvoss: what do you mean by apport integration? additional questions/files to grab?13:58
tvossdidrocks, yup, like being able to say ubuntu-bug xmir13:59
didrockstvoss: ah, so you mean apport integration for packages not even in distro13:59
didrocksbut in a ppa13:59
tvossdidrocks, for example, yes13:59
didrockstvoss: by "main", you could have infer "in universe"14:00
didrockstvoss: however, I think you want stacktrace retracing?14:00
tvossdidrocks, yup, it should mostly be a convenient way for users/early adopters to report issues14:00
didrockstvoss: they won't be able to retrace automatically though14:01
didrockstvoss: as the dbgsym are not published in ddebs.ubuntu.com14:01
tvossdidrocks, can we get them there somehow for ppa packages?14:02
didrockstvoss: that would be a question for pitti I guess, not sure14:02
didrockstvoss: having dailies will enables to push stuff quickly and have this kind of feedback loop14:03
tvossdidrocks, ack14:03
tvoss_alf_, ping14:17
alf_tvoss_: pong14:17
=== alan_g is now known as alan_g|tea
=== alan_g|tea is now known as alan_g
kdub_good morning15:04
racarrMorning!15:05
kdub_oh, so android/mir users, you should upgrade to the latest saucy-based images, the mir fixes for saucy have landed :)15:08
kdub_ ^ricmm15:08
alan_g\o/15:09
* alan_g is not so happy about the compiler diagnostics when compiling on saucy15:10
kdub_yeah :/15:10
racarrkgunn: unity mir sync? https://plus.google.com/hangouts/_/8015a964c25dbb4a92528d82f94676c27eb14a7a?authuser=115:31
kgunnracarr: greyback go for it...15:32
kgunni'm gpu boy today15:33
greybackyep, fighting with SSO15:33
racarrfun fun15:33
kgunnricmm: ^15:33
=== mmrazik is now known as mmrazik|afk
racarrkdub_: I only saw your mouth moving XD16:13
kdub_racarr, greyback yeah, was muted, just trying to see how the qt bits tie into the compositor16:13
racarrkdub_: Oh you know, with an API16:14
racarr;)16:14
alan_gkdub_: my by definition of "precondition" there's no promise about behavior - conversely if you promise behavior then you've not got a precondition16:22
alan_gAnd a test is for promised behaviour16:22
* alan_g prepares to move desktop to saucy16:24
kdub_alan_g, good luck! made for an interesting friday afternoon for me16:24
kdub_alan_g, yeah, i guess its just about the definition of 'precondition'16:25
alan_gkdub_: my netbook finally looks stable on saucy, so I'm risking it (my laptop however is staying or raring)16:25
alan_ghttp://en.wikipedia.org/wiki/Precondition16:27
alan_gI guess I've been hanging around the DBC crowd.16:27
kdub_alan_g, so i guess if there's a test that will throw under certain conditons, thats not a 'precondition' its just behavior16:28
=== francisco is now known as Guest20683
ricmmkdub_: racarr hey guys, trying out the XMir dogfooding guide17:41
racarrhttps://github.com/eMir-server/eMir17:58
racarroO17:58
racarreMir was forked from http://launchpad.net/mir as the '-ng' version of Canonical's Mir.17:58
kdub_hah :P18:09
kdub_ricmm, hopefully going well18:22
=== marlinc is now known as marlinc|away
racarrapparently my consumption of thai food from the place downstairs is high enough it motivated them to switch from plastic containers to compostable cardboard19:04
racarrkilling the planet one curry at a time :(19:04
=== marlinc|away is now known as marlinc
=== marlinc is now known as Marlinc
robert_ancellracarr, hey, do you think it's likely we'll have u8 on Mir usable by the end of the month? At least from our side?20:23
racarrrobert_ancell: Yes!20:24
robert_ancell\o/ :)20:24
racarrextremely20:24
racarrrobert_ancell: we should have images in parallel CI today. then this morning gerry and I came up with a plan to keep all of the shell20:25
racarrin one surface, like they are doing now20:26
racarrthat is just on top of application surfaces20:26
robert_ancellah, cool20:26
racarrwhich means they can implement the existing applicationmanager API they are exposing out to QML (what gerry is working on now)20:26
racarron top of libmirserver20:26
racarrand20:26
racarrBOOM!20:26
racarrI am adding support for shaped surfaces today to mir (at leat on the the input side, for output they can just use transparency and clipping is an optimization later)20:27
racarrthis is kind of how they handle it on the phone images so will be easy20:27
racarroh I wanted to ask you, i heard some stuff about20:27
racarrlightdm on the phone, and upstart, and launching applications20:27
racarrand...20:28
racarrwell, that's basically what I heard XD. what does it all mean? I know there is an upstart job for launching applications20:28
racarrwhere does lightdm come in to the picture?20:28
robert_ancellracarr, otp, will get back to you on that. In short we're not involved20:29
racarrok20:30
racarrI didn't think so much, just trying to get an idea of when the pieces20:30
racarrwill fall in to place20:30
mterryracarr, so upstart will start lightdm; lightdm will start a greeter, and then once user is authenticated, lightdm will start the session20:35
racarrand you use the session with upstart to launch application?20:36
racarrs?20:36
mterryracarr, you could yeah.  I think that's the plan, that the launcher could use upstart to launch apps20:37
mterryracarr, that could happen today though, eh?  That's not a lightdm thing20:39
mterryracarr, although maybe there isn't a user upstart session today?20:40
mterryI think unity8 is launched via a system upstart session20:40
mterryBut I haven't read it closely to see whether it also starts a user upstart session20:40
racarrmterry: I guess. I'm just confused because I heard it connected to lightdm20:43
racarrand don't want to confuse you with my confusion XD20:43
racarrI find20:57
racarrstd::chrono to be the most infuriating API20:57
mterryracarr, well, lightdm will definitely have user sessions21:03
mterryracarr, there may not be user upstart sessions today, I don't know.  They *could* exist, just depends on how the session-starter script is written21:03
robert_ancellkgunn, reuse same hangout?21:06
kgunnhttps://plus.google.com/hangouts/_/3d242d63010d29a68a15623903d47cd73b7e85a121:07
kdub_RAOF, ping21:14
robert_ancellkgunn, sorry, was mixed up about which bug you meant - I've assigned the plymouth one back to me21:29
kgunnrobert_ancell: np21:32
kgunnso should chris get the other ?21:32
=== marlinc is now known as marlinc|away
=== marlinc|away is now known as marlinc

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