[00:13] <thomi> racarr: sure, I understand the reasons. Can you let me know when there's a set of binary packages that will work together please?
[00:16] <racarr> thomi: I'm just going to build everything on the phone while I cook dinner and see if autopilot runs
[00:16] <racarr> then we can just ladn the branch
[00:16] <racarr> the longest part is fetching the build deps because my phone only gets like 100 KBs but its almost done
[00:21] <robert_ancell> kgunn, there?
[00:32] <kgunn> thomi: racarr robert_ancell sorry...went for food
[00:32] <kgunn> ok...so, i vote no merge until racarr finishes his dinner :) and has a chance to try
[00:33] <robert_ancell> kgunn, np. I think you just need to do a 'bzr revert debian/libmirserver4.install' then a 'bzr mv debian/libmirserver4.install debian/libmirserver5.install'
[00:33] <robert_ancell> BUT, I'd recommend using that script instead
[00:33] <kgunn> robert_ancell: well i've already got a branch....what if i just do a bzr add debian/libmirserver5.install
[00:33] <kgunn> won't the effect be the same
[00:34] <robert_ancell> kgunn, I don't think it would treat that as a move like git does
[00:34] <kgunn> just in lp it'll show removed and added, instead of renamed ?
[00:34] <robert_ancell> and that makes merges harder
[00:34] <kgunn> robert_ancell: ok...i'll revert....should i just pull down that mp and redo
[00:34] <robert_ancell> kgunn, it will just revert the one file
[00:34] <kgunn> robert_ancell: which script are you refering to ?
[00:35] <robert_ancell> kgunn, the one I sent you last time, I'll find it again
[00:35] <robert_ancell> I'll send it to the mailing list
[00:35] <kgunn> robert_ancell: thanks...i'm extra stupid today (series of late nights/early morns)
[00:35] <robert_ancell> np :)
[00:35] <robert_ancell> http://paste.ubuntu.com/6186191/
[00:37] <kgunn> robert_ancell: thanks...i'm gonna try to fix manually with bzr first
[00:37] <robert_ancell> kgunn, you also need to edit debian/libmirserver5.install and change the 4 to a 5
[00:38] <robert_ancell> i.e. usr/lib/*/libmirserver.so.4 -> usr/lib/*/libmirserver.so.5
[00:38] <kgunn> yes ;)
[00:41] <kgunn> robert_ancell: ok...unhappy w revert...gonna pull fresh and run the script
[00:41] <robert_ancell> :(
[00:41] <kgunn> do i run it just at the top dir of my mir branch
[00:56] <kgunn> robert_ancell: ok...just blew it away & went fresh (thanks for the script...) would you
[00:57] <kgunn> give me 2 secs...
[01:05] <kgunn> robert_ancell: ok...for reals now...can you review https://code.launchpad.net/~kgunn72/mir/bump-libmirserver5-take2/+merge/188968
[01:06] <robert_ancell> kgunn, +mir (0.0.12-0ubuntu2) UNRELEASED; urgency=low should be +mir (0.0.13-0ubuntu1) UNRELEASED; urgency=low
[01:06] <robert_ancell> otherwise good
[01:07] <kgunn> damn it...
[01:09] <robert_ancell> kgunn, oh, and you need to edit CMakeLists.txt and set set(MIR_VERSION_PATCH 13)
[01:09] <kgunn> ack...fixing it
[01:15] <robert_ancell> kgunn, you still have -0ubuntu2 not -0ubuntu1..
[01:15] <kgunn> robert_ancell: i feel i could tear up a steel ball with a rubber hammer tonight
[01:15]  * robert_ancell feels glad to be on the other side of the world
[01:16] <kgunn> robert_ancell: not angry...just capable of f'ing up something so simple
[01:17] <robert_ancell> packaging/abis/versioning just *look* simple
[01:19] <kgunn> robert_ancell: ...ok..surely...there's nothing left to have been incorrect :)
[01:31] <kgunn> robert_ancell: thanks man
[01:31] <robert_ancell> no worries
[01:33] <kgunn> racarr: any luck ?
[01:33] <kgunn> thomi: can you send another update ?
[01:34] <kgunn> at the end of your day
[01:34] <thomi> kgunn: sure, I'm waiting to hear from racarr once he's rebuilt the world and we can actually launch unity
[01:35] <racarr> kgunn: mir build at 47 percent
[01:36] <racarr> I guess lets abort building the tests
[01:36] <thomi> racarr: wait, why are you building mir? I thought we had to rebuild everything else?
[01:37] <racarr> I dont want to deal with the packages because the dev packages
[01:37] <racarr> are broken
[01:37] <racarr> apparently?
[01:37] <racarr> or
[01:37] <racarr> it tries to pull in newer versions
[01:37] <racarr> or apt insanity
[01:37] <racarr> so I dont know how I can get the build dependencies for everything else with the packages
[01:37] <racarr> so im just building the branch on phon
[01:37] <kgunn> racarr: hmmm...you mean the packages associated with dev-proposed
[01:37] <racarr> e
[01:38] <racarr> no the packages from the CI build
[01:38] <kgunn> racarr: totally agree with you...
[01:38] <thomi> right
[01:38] <thomi> the -dev packages need to pull in other things
[01:38] <kgunn> racarr: building on the phone....nuke it from orbit, its the only way
[01:38] <thomi> racarr: if you install them, then do apt-get -f install, everything workls
[01:38] <thomi> racarr: just don't do an update
[01:38] <thomi> or... you're gonna have a bad time
[01:38] <racarr> well it's too late buidling unity-mir now
[01:39] <thomi> kgunn: srsly. Can we just merge it?
[01:41] <racarr> is merging it now going to be any better than merging it in 20 minutes when its tested
[01:41] <thomi> racarr: if it really is 20 minutes, then no. I'm sceptical though
[01:44] <kgunn> thomi: is starting to learn the estimate/reality ratio :)
[01:44] <kgunn> thomi: i'd prefer to test it first
[01:45] <kgunn> if its holy hell for racarr then we'd avoid a nasty one
[01:45] <kgunn> takin' a little break brb
[01:46] <thomi> kgunn: ok
[02:00] <duflu> robert_ancell: Seems NZ DST crept up on me. Hangout?
[02:00] <robert_ancell> duflu, yep
[02:03] <racarr> kgunn: thomi: Ok it at least starts!
[02:03] <racarr> will test autopilot after I buy grociers
[02:03] <racarr> but unity runs fine so it's safe at least
[02:04] <thomi> racarr: I don't suppose you have binary packages so I can try it as well?
[02:06] <racarr> no.
[02:28] <kgunn> robert_ancell: and the trunk https://code.launchpad.net/~mir-team/mir/development-branch/+merge/188975
[02:28] <kgunn> mainly...commit msg ok to you ?
[02:28] <robert_ancell> kgunn, looking
[02:29] <robert_ancell> kgunn, you have a merge conflict
[02:29] <robert_ancell> in debian/changelog
[02:31] <kgunn> robert_ancell: ok...how to handle...
[02:32] <robert_ancell> kgunn, you need to run 'bzr merge lp:mir'
[02:32] <robert_ancell> then edit the file manually and run 'bzr resolve'
[02:32] <robert_ancell> then bzr commit and push
[02:32] <kgunn> ack
[02:32] <kgunn> thot so ...just feels a bit weird in a way
[02:33] <kgunn> robert_ancell: when i push...it'll be lp:~mir-team/mir/??? development-branch ?
[02:33] <kgunn> in order to merge back in the conflict resolution ?
[02:33] <kgunn> back in== into the dev branch
[02:34] <robert_ancell> kgunn, ah yeah. hmm. why did trunk get out of sync?
[02:34] <robert_ancell> let me try locally
[02:35] <kgunn> thanks...if you fix, lemme know...i want to know for next time...
[02:35] <kgunn> takin' a break to edit a freshman english paper
[02:36] <robert_ancell> hmm. It merges locally fine. Is LP confused?
[02:43] <robert_ancell> kgunn, manual merge did the trick
[02:53] <kgunn> robert_ancell: so are you proposing or want me to ?
[02:53] <robert_ancell> kgunn, I did the merge and the MP looks good
[02:53] <robert_ancell> I've approved it
[02:53] <robert_ancell> (not top)
[02:54] <kgunn> thnank you sir...i'll top approve
[03:07] <kgunn> racarr: any luck ?
[03:15] <kgunn> did we lose racarr ?
[03:21] <duflu> kgunn: I think he mentioned groceries?
[03:41] <robert_ancell> be back in 20 mins
[03:47] <duflu> RAOF: Please correct me where I'm wrong (again); https://bugs.launchpad.net/bugs/1234502
[03:47] <RAOF> Correct.
[03:48] <duflu> RAOF: Jetlagged much?
[03:48] <RAOF> Oddly.
[03:49] <RAOF> I got back on Saturday.
[03:49] <RAOF> And have been waking early and sleeping early since then.
[03:49] <kgunn> robert_ancell: just to verify for thomi....if he rebuilds on the phone, mir, he'll have to rebuild unity-mir & platform-api....and the order won't matter
[03:49] <RAOF> In possibly related news, Zoë reliably wakes at 5:45am.
[03:49] <duflu> How convenient
[03:50] <kgunn> robert_ancell: cause racarr 's input branch is based on dev...which is broken/updated server api
[03:50] <kgunn> not yet in archive
[03:50] <thomi> well, presumably I'll have to install those mir*-dev packages first, then rebuild unity-mir & platform-api?
[03:52] <kgunn> thomi: right
[03:52] <kgunn> thomi: so for all the effort today...did you ever run the AP test with mir enabled ??
[03:52] <thomi> ok, well, I'm re-flashing the phone now, because somehow I managed to completely bork it
[03:52] <kgunn> i was confused from some of the scrollback
[03:53] <thomi> kgunn: nope :(
[03:53] <thomi> kgunn: I was confused as well - turns out I was running SF
[03:53]  * kgunn just threw up
[03:53] <thomi> not literally I hope
[03:57] <kgunn> thomi: virtually
[03:58] <thomi> good, well, sorry about that. I guess I'll start rebuilding this stuff
[03:59] <kgunn> thomi: so...if you enabled mir...what happens (oh...nvmd..borked)
[04:00] <kgunn> right ?
[04:00] <thomi> kgunn: if I enable mir, unity won't start
[04:00] <thomi> ABI mismatch
[04:00] <kgunn> thomi: gah...right
[04:00] <kgunn> thomi: thank you for your efforts
[04:00] <thomi> kgunn: no worries, I'll rebuild this stuff. I just wish there was a faster way
[04:01] <kgunn> thomi: think there's anyone in integration team/ci/phonedations that could help build us an image quicker ?
[04:01] <kgunn> based on that branch ?
[04:02] <thomi> kgunn: not at this time of night, I doubt it. I can ask though
[04:02] <thomi> kgunn: this is why I was saying ewarlier we should just go ahead and merge it
[04:02] <kgunn> robert_ancell: ^ thots...
[04:08] <kgunn> thomi: ok...based on racarr> but unity runs fine so it's safe at least
[04:08] <kgunn> i agree...
[04:10] <kgunn> robert_ancell: i need your guru knowledge...racarr kinda did a weird one...dandrader had a branch
[04:10] <kgunn> robert_ancell: racarr originally merged ontop ...but then retargetd the whole thing to dev
[04:10] <kgunn> https://code.launchpad.net/~robertcarr/mir/1233944-addendum/+merge/188900
[04:11] <kgunn> do i just approve racarr's and then it pulls it all in ?
[04:13] <kgunn> thomi: racarr robert_ancell ...ok...whoever cares..i just top approved it
[04:13] <thomi> kgunn: I'll have a go at building *all the things*
[04:13] <thomi> kgunn: if not before, I'll TA it at my EOF
[04:13] <thomi> err, EOD
[04:13] <kgunn> alf|xmir_devel: are you seriously here?
[04:14] <kgunn> thomi: robert_ancell ok...gonna go watch some tv, try not to fall asleep then merge dev onto trunk
[04:23] <kgunn> thomi: also...can you provide a very explicit list of instructions of what should be run in terms of AP testing
[04:23] <thomi> kgunn: sure, to whome?
[04:24] <kgunn> thomi: i'd like it if you could send that to dandrader and cc me...this way, if that code getsmerged he can test his morning
[04:24] <thomi> sure thing
[04:24] <robert_ancell> kgunn, ok, we're all go :)
[04:26]  * thomi installs build-deps for unity-mir
[04:26] <robert_ancell> thomi, are you just doing a bzr-buildpackage from lp:unity-mir?
[04:27] <thomi> robert_ancell: a 'bzr bd', yeah
[04:27] <thomi> it'll take a while since I'm on the phone :-/
[04:28] <robert_ancell> oh, right
[04:29] <thomi> some decent armhf hardware wouldn't go astray. I wonder if we could set something up....
[04:30] <robert_ancell> thomi, what are the options currently?
[04:30] <thomi> options?
[04:31] <robert_ancell> thomi, for purchasable hardware
[04:31] <robert_ancell> with decent power
[04:31] <thomi> robert_ancell: oh, I have no idea. I was thinking maybe Canonical could buy and then set aside a few calxeda (sp?) boxes, for custom builds
[04:32] <thomi> another thing that I could set up that would save me a lot of time is a caching apt proxy server
[04:35] <robert_ancell> thomi, it looks like calxeda doesn't sell products directly. I guess you buy a Dell or HP server?
[04:38] <thomi> robert_ancell: yeah, or talk to them nicely :)
[04:38] <kgunn> robert_ancell: so can i just propose rcarr's branch directly to trunk? in parallel...it'd be faster
[04:39] <robert_ancell> kgunn, I guess so - it shouldn't have any major side effects (famous last words)
[04:39]  * thomi starts building unity-mir
[04:40] <robert_ancell> kgunn, it's ready to land in development-branch right?
[04:40] <kgunn> already approved...waiting for merge
[04:40] <robert_ancell> just saw that
[04:40] <robert_ancell> go jenkins!
[04:42] <kgunn> robert_ancell: just wanting it to hit trunk before didrocks gets on
[04:42] <robert_ancell> kgunn, sure
[04:42] <robert_ancell> kgunn, are you going to MP it or manually merge from development branch?
[04:42] <kgunn> was going to mp then you could approve...i mean, we have the approvals we know what we want & trust it
[04:42] <kgunn> robert_ancell: ^
[04:43] <kgunn> and a subsequent dev merge to trunk, that diff should just get ignored
[04:43] <robert_ancell> sounds good - it will take time for jenkins to merge it but safer to have the check
[04:43] <robert_ancell> I guess it's better than doing the double wait for jenkins (dev-branch merge, then merge to trunk)
[04:43] <kgunn> robert_ancell: oh yeah...wouldn't consider anything else
[04:43] <kgunn> i heart ci
[04:43] <robert_ancell> ci saves our bacon
[04:43] <kgunn> right
[04:45] <kgunn> robert_ancell: ok...can you approve and i'll ta https://code.launchpad.net/~robertcarr/mir/1233944-addendum/+merge/188983
[04:46] <robert_ancell> hey, it merged cleanly
[04:47] <robert_ancell> kgunn, approved
[04:49] <kgunn> robert_ancell: thanks...
[04:50] <thomi> robert_ancell: what's the platform-api lp project name?
[04:50] <thomi> ubuntu-platform-api or something?
[04:50]  * robert_ancell looks
[04:50] <robert_ancell> thomi, lp:platform-api
[04:50] <robert_ancell> I don't know why they didn't put the ubuntu- prefix on it
[04:54] <kgunn> ok...going to bed...night guys
[04:58] <robert_ancell> later
[04:58] <robert_ancell> I'll be back in 2 hours
[05:12] <thomi> ok, I finally have a coherent set of binary packages.
[05:12] <thomi> rebooting
[05:16] <thomi> robert_ancell: no luck - seems like nothing happens now
[05:17] <thomi> robert_ancell: I ger this: "ofono/ofono/account0 initialized" in an endless loop in the unity8 log
[05:18] <thomi> robert_ancell: and a segfault when I start unity8 manually
[05:18] <thomi> robert_ancell: BT looks bad too:
[05:18] <thomi> #0  0x409cd640 in ?? () from /lib/arm-linux-gnueabihf/libc.so.6
[05:18] <thomi> #1  0x409cf324 in malloc () from /lib/arm-linux-gnueabihf/libc.so.6
[05:26] <thomi> robert_ancell: OK, so good news! After messing about some more, it now works
[05:26] <thomi> I have mir/unity8 running, and can drag the home screen around with autopilot
[05:29] <didrocks> hey kgunn
[05:30] <didrocks> kgunn: please update the spreadhseet once the commits are in, and we'll try to get it for image 77 (or 78)
[05:30] <didrocks> we'll get it then ASAP
[05:32] <duflu> We're still using the spreadsheet idea?
[05:36] <didrocks> duflu: we do
[05:37] <duflu> didrocks: I look forward to us having separate development and maintenance branches post-saucy, to make things more clear
[05:38] <didrocks> duflu: well, we still need to have trunk always releasable.
[05:38] <duflu> didrocks: Yes, but you know what I mean
[05:38] <didrocks> yeah ;)
[05:38] <didrocks> I think there will be a lot of discussions on how to release tings
[05:38] <didrocks> things
[05:38] <didrocks> nobody (even us) aren't happy about it
[05:39]  * duflu assumes that's a mistaken double-negative
[05:40] <didrocks> duflu: yeah, it is (or you can take it as a hope :p)
[06:41] <alf|xmir_devel> kgunn: no, forgot to change nick... :)
[06:49] <alf|xmir_devel> RAOF: Hi! Do you have a moment to check the latest commit in https://github.com/afrantzis/xserver/tree/multimonitor-stabilize ? I change the code to assign root_fragments dynanically to crtcs as needed, so that when mirroring we can have a single root_fragment/MirSurface. The mirroring parts works great, but I somehow managed to cause glitches in non-mirror (extended) modes. Any ideas?
[06:50] <RAOF> alf|xmir_devel: I'll have a butcher's.
[06:50] <alf|xmir_devel> RAOF: The glitches are there only when bypassing in mir
[06:51] <alf|xmir_devel> RAOF: thanks
[06:52] <RAOF> Ah, bypass.
[06:54] <duflu> It's true, when I wrote bypass I knew nothing about DDXs and assumed XMir must work something like glamor :P
[06:55] <mlankhorst> haha
[07:03] <robert_ancell> didrocks, we have all the changes in the landing pipeline doc on row 129 except for the autopilot fix - should I append that entry or add another one?
[07:03] <robert_ancell> sorry, I just noticed the autopilot is on line 130, so all is good right?
[07:04] <didrocks> robert_ancell: yeah, all good :) (Mir is building right now)
[07:05] <didrocks> robert_ancell: you didn't add anything for the past 30 minutes, right?
[07:05] <robert_ancell> didrocks, no, and nothing planned to land
[07:05] <didrocks> (in Mir trunk)
[07:05] <didrocks> ok, great!
[07:05] <didrocks> so we're good, I merged the rest of the transitions
[07:05] <didrocks> and everything is rebuilding
[08:52] <budgee> Howdy, further to my note a couple of days ago about Synergy and Mir, please see output at: http://pastebin.com/pZVrqJqa
[08:54] <duflu> budgee: I'm not sure that log helps.... However we did land a fix for Mir exclusively locking input devices, so that sounds like the kind of fix you need. Not sure if it has been packaged yet
[08:56] <alf|xmir_devel> RAOF: trying the multimonitor-stabilize branch on intel, I don't get glitches, but I am not sure if the glitches I see on radeon are because of radeon or because on my desktop (with radeon) the system moves from clone -> extended configuration on startup. That is, because I can't mirror on my laptop, I can't tell whether after cloning we mess up bypass somehow for future extended configurations...
[09:10] <robert_ancell> bye all
[09:53] <alf|xmir_devel> RAOF: I am leaning towards the glitches being a radeon issue. With plain vladmir-upstreaming with a single monitor I still get an occasional glitch on radeon, none with intel. Plus, I get a a segfault with radeon_drv in the back trace when plugging in/out monitors http://paste.ubuntu.com/6187470/
[09:56] <budgee> duflu: thank yuou
[10:03] <alf|xmir_devel> RAOF: FYI, here is the valgrind output: http://paste.ubuntu.com/6187493/
[10:09] <davmor2> kgunn: so mir, gets very, very, very slow if you start opening lots of apps on the maguro.  Also I've noticed that memory usage is up when I enable mir over SF but I can't say for sure that, that is the cause as I'd added apps and debugging stuff etc
[10:11] <davmor2> kgunn: if I open 4 apps on mir it's like running through treacle (molasses), 4 apps on SF is noticeably fast.  It faster loading and switching and using the app.
[10:12] <davmor2> kgunn: on the plus side though I only had it lock up on me once.  So mir is definitely more stable than it was by a huge a margin.
[10:19] <duflu> davmor2: (1) kgunn is offline. (2) maguro is known to be slow with bug 1182930. (3) Multi-surface performance is known to be bad on Android due to bug 1227739. (4) There's always room for improvement later
[10:21] <davmor2> duflu: Yep I know kgunn is offline, only I'm going to be busy latter.  Thanks for the bug info though I will add myself to those.  I'm still overly impressed by the fast that mir has gone from crashing if you look at it wrong to locking up once in 16 hours use
[10:22] <duflu> davmor2: Yeah, though once in 16 hours is too much :/
[10:24] <davmor2> unfortunately I couldn't find anything that screamed crash so I'm assuming there might be a memory leak and the system just ran out of space but I couldn't get access to check, adb was out and the phone was unusable even a long hold on the power button did nothing :(
[10:24] <davmor2> I had to remove the battery in the end to reset it so I could use it :(
[10:27] <duflu> I wish more of us had maguro. Most of the mir team does not. So progress on it is slow
[10:29] <davmor2> duflu: Yeah that is somewhat daft as it is a supported device. But you can't have everything :)
[12:31] <hikiko> https://bugs.launchpad.net/mir/+bug/1234679 RAOF alf_ look at an interesting bug I found
[12:35] <hikiko> the screenshot is from monitor 2 I ll upload another one as well
[13:44] <kgunn> mzanetti: or Saviq ....so there was some prelim AP test attempts on mir, here's what psivaa got...http://pastebin.ubuntu.com/6188183/
[13:45] <Saviq> kgunn, ze Germans are on holiday today
[13:45] <kgunn> Saviq: :) i like how an english-as-a-2nd-lang can still crack a "ze" joke
[13:45] <Saviq> kgunn, "terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> >'what(): display factory cannot create fb display"
[13:46] <Saviq> kgunn, sounds like unity8 failed to start - probably an app running still holding the framebuffer
[13:47] <Saviq> kgunn, now that you mention it, though - we have the indicators_client tests that won't work, 'cause we don't have unity8 running for them
[13:47] <Saviq> kgunn, will need fixing
[13:47] <olli> keep beating on the germans
[13:48] <olli> and I'll unleash the thostr
[13:50] <kgunn> Saviq: to make sure i understand...this is the old "you need the shell up for everything"...hard to isolate problem ?
[13:50] <Saviq> kgunn, for indicators_client, yes
[13:50] <Saviq> kgunn, for now we relied on sf being there always - need to change that
[13:51] <Saviq> kgunn, the failure above, though, suggests there's a stale app running that blocks "new" unity8 from getting the framebuffer
[13:51] <Saviq> kgunn, even though unity8 itself isn't there anymore
[13:51] <kgunn> Saviq: ah
[13:52] <kgunn> what's weird he said he gets 100% failures...
[13:53] <kgunn> trying to square that with "stale app" suggestion
[14:14] <kgunn> alf_: can i ask you to sort of ready yourself to potentially help on some making mir default on touch work ?
[14:14] <kgunn> alf_: i feel we may have some digging to do...but nothing specific atm
[14:15] <kgunn> alf_: at the rate we've been going, it'd be good just to be ready...like, at least phablet-flash ubuntu-system --channel devel --no-backup
[14:16] <kgunn> to get the latest image
[14:17] <kgunn> alf_: and you know, touch /userdata/.writable_image & touch /home/phablet/.display-mir
[14:20] <alf_> kgunn: sure, I am on it
[14:35] <dandrader> alan_g, do you guys have a way of building mir packages using cross-compilation?
[14:37] <alf_> dandrader: do you care about packages in particular, or binaries will do?
[14:40] <dandrader> alf_, for binaries there are those scripts which work fine
[14:40] <dandrader> alf_, but sometime it's nice to be able to build packages to ensure the whole system is using your libraries
[14:46] <alf_> dandrader: ok, AFAIK, we don't have a straightforward way to cross-compile mir packages, at least I haven't heard of anyone doing it.
[14:49] <alan_g> dandrader: What alf_ said
[14:50] <kgunn> dandrader: alf_ actually
[14:51] <kgunn> dandrader: alf_ to replicate what _will_ be image 79....just flash the current devel-proposed....then apt-get update, install libmirserver5
[14:51] <kgunn> don't forget to make it writable first ;)
[14:51] <kgunn> and then turn on mir
[14:52] <kgunn> dandrader: alf_ then you can either follow thomi's instructions in the bug
[14:52] <kgunn> or the ci guys said they tried with this reboot, wait for Nautilus window to pop up the mounted share (connection breaks at that moment, takes about 30s after unity8 is visible), run phalet-test-run -n unity8 (after also installing unity8-autopilot)
[14:53] <kgunn> jibel: would you mind doing an impromptu run down on the bugs you've found...say in 30 minutes
[14:54] <alf_> kgunn: so that's --channel devel-proposed? Is there a way to go from devel to devel-proposed without downloading an image from scratch?
[14:54] <kgunn> alf_: i don't think so...and that's the only way to get the latest...otherwise, there's like a day lag in image creation
[14:57] <alf_> kgunn: channel devel-proposed didn't bring in any new image for me...
[14:57] <alf_> kgunn: still 78
[14:58] <kgunn> alf_: hmmm....i suppose there's a moment of promotion where they are equal
[14:59] <jibel> kgunn, sure, in 30 min.
[14:59] <dandrader> alf_, how do you tell the image number?
[14:59] <jibel> (more like 25 now)
[15:00] <alf_> dandrader: I saw messages like: INFO:phablet-flash:Pushing /home/alf/Downloads/phablet-flash/imageupdates/devel-proposed/mako/version-78.tar.xz.asc to /cache/recovery/
[15:01] <alf_> dandrader: and I noticed that using the devel-proposed channel didn't download any new files, it reused the ones I had downloaded from devel
[15:01] <dandrader> alf_, ah, ok. I'm also getting 78
[15:49] <kgunn> alf_: mind if i assign you that nasty one ? where he can't reboot to a ui
[15:51] <alf_> kgunn: np
[15:53] <dandrader> where's the mir log that comes out of unity8?
[15:53] <dandrader> I see nothing in /tmp...
[15:54] <dandrader> do I have to enable it somehow?
[15:54] <dandrader> alan_g, ^
[15:55] <dandrader> so far I've been using a modified mir_demo_standalone_input_filter to get log messages....
[15:56] <alf_> dandrader: check http://unity.ubuntu.com/mir/component_reports.html
[15:56] <dandrader> alf_, thanks!
[15:56] <dandrader> wow, real documentation!
[16:00] <alan_g> dandrader: if you want it in /tmp then add --glog/MIR_SERVER_GLOG
[16:01] <dandrader> hmmm, it seems the android-input logging is not wired to that mir log system
[16:02] <dandrader> alf_,  or maybe I should set the minimum priority level or something. is there such filtering in mir?
[16:06] <alf_> dandrader: if you are using glog you can set the level, but by default it should log everything. Using the built-in/default logging mechanism should just print everything to either stdout or stderr. Are you capturing both?
[16:06] <dandrader> alf_, I'm getting output on stdout and that's fine. but nothing from android-input. will dig further...
[16:14] <kgunn> dandrader|lunch: ok...as you try to run the AP's hopefully you can repro what plars is seeing
[16:14] <kgunn> plars can be found in #ubuntu-ci-eng
[16:14] <kgunn> basically this....<plars> kgunn: running phablet-test-run -n (for unity8) and you see unity8 go off, but never come back.
 kgunn: also, for all ap tests, we have a script that unlocks the screen and restarts unity8 in -testability mode. It doesn't seem to function now either
[16:16] <kgunn> bbiab
[16:45]  * alan_g hates it when he backs out all his changes and the problem they cause is still there
[16:54] <kdub> i like how tdd forces me to refine the dumb interfaces i sometimes invent
[17:03]  * alan_g discovers that the tests that are failing work fine if they're  run on their own
[17:03] <alan_g> ... or together
[17:06]  * alan_g|EOD knows where a nice pint of beer will clarify everything
[17:18] <racarr> kgunn: Are there images that need testing today?
[17:18] <racarr>  /bugs that need chasing
[17:19] <racarr> if not I think the up/down volume keys are my task
[17:20] <om26er> racarr, hey! would you know if we have a bug to track recent slowness under Mir ? (talking Nexus 4 here)
[17:20] <om26er> or even did anyone else report that ?
[17:20] <om26er> I am pretty sure rendering was pretty fast like 10 days ago
[17:22] <kgunn> latest image 79 supposedly up...boom
[17:22] <kgunn> racarr: did you already finish the new i/f for zanetti ?
[17:22] <kgunn> i think it was need to help hud show up
[17:23] <racarr> kgunn: the input injection yeah, proposed it yesterday
[17:23] <kgunn> ta
[17:23] <racarr> ohh and
[17:23] <racarr> 2 needs fixings
[17:23] <racarr> >iterates<
[17:24] <kgunn> :)
[17:24] <kgunn> racarr: i asked alf_ to look into this one https://bugs.launchpad.net/mir/+bug/1234609
[17:24] <kgunn> but i'm sure he'd take a second set of eyeballs...
[17:30] <kgunn> kdub: you used to have a n10 branch up...did you take it down ?
[17:30] <kdub> kgunn, yeah, put it on pause until i can get the display up
[17:31] <kdub> no real reason i shouldn't reintroduce it and land it though
[17:32] <racarr> alf_: kgunn: From this stacktrace https://launchpadlibrarian.net/152224751/Stacktrace.txt
[17:32] <racarr> the bug looks like...
[17:32] <racarr> the grand surface race.
[17:33] <racarr> we need to make it safe to hold shared_ptr<msh::Surface>
[18:05] <racarr> kgunn: alf_: This may be enough for 1234609: https://code.launchpad.net/~robertcarr/mir/hold-surface-alive/+merge/189156
[18:05] <racarr> but it needs some thought
[18:06] <racarr> im not 100% sure yet it's ok
[18:11] <kgunn> woohoo racarr love it when that happens...
[18:13] <racarr> kgunn: We'll see :D
[18:15] <kgunn> dandrader: any luck reproducing what plars was seeing...e.g. not being able to unlock the shell with args...or not having the fake shell load
[18:16] <dandrader> kgunn, still fiddling with mir code to get android-input logging out of unity8.
[18:18] <dandrader> that part is not being properly initialized
[18:18] <racarr> dandrader: Even with MIR_SERVER_LEGACY_INPUT_REPORT=log
[18:18] <racarr> ?
[18:19] <dandrader> ah, another undocumented thing. gonna try....
[18:19] <kdub> code churn thursday
[18:19] <racarr> dandrader: Technically it's in --help but I guess it's not really very clear at all
[18:20] <kgunn> kdub: its like hump day...but not as funny
[18:20] <dandrader> racarr, I was pointed to http://unity.ubuntu.com/mir/component_reports.html. but it must be outdated
[18:20] <kgunn> plars: so...if you manually unlock and run the AP tests...what do the results look like ??
[18:20] <kgunn> any passing? or is it dismal ?....any #'s ?
[18:23] <racarr> dandrader: ah yeah
[18:23] <plars> kgunn: yes, but not unity8 tests
[18:24] <plars> kgunn: those are the ones that require the restart of unity though
[18:24] <plars> kgunn: the ones that don't seem to run ok as long as I manually unlock it
[18:24] <kgunn> plars: thanks...so passing tests (other than unity8) for the most part ?
[18:25] <kgunn> plars: sorry...i'm being pestered for some #s  just to be a windsock
[18:25] <plars> kgunn: I haven't tried but a few (friends-app and notes)
[18:25] <kgunn> hence my being an irritant :)
[18:25] <plars> kgunn: np
[18:25] <dandrader> racarr, works now. thanks
[18:25] <kgunn> plars: ok...but passing ?
[18:25] <dandrader> well, wasted a good chunk of time
[18:26] <kgunn> dandrader: i'm gonna log in a bit...and write down the instructions and update the wiki
[18:26] <plars> kgunn: for those two, yes
[18:26] <kgunn> drives me crazy
[18:26] <plars> kgunn: if I get some time, I can try others, but since it's manual it will be a bit
[18:26] <kgunn> plars: ack...thanks for that
[18:26] <racarr> dandrader: :D
[18:26] <racarr> sorry :(
[18:27] <plars> kgunn: if you have any you are specifically concerned about, let me know
[18:27] <plars> I can try those first
[18:28] <kgunn> thomi: ^ ? ...he might when he comes on...
[18:28] <kgunn> Saviq: ^ maybe you have a suggestion for plars  in terms of AP tests that would be good to prioritize (besides unity8) ?
[18:29] <Saviq> kgunn, not really
[18:29] <kgunn> dandrader: so it seems like the best item to debug first is the AP run for unity8 since it seems the fake-unity8 load is unhappy....and looks like we at least can run other AP test manually
[18:29] <racarr> short lunch back in 20
[18:29] <kgunn> plars: at least gallery & camera apps would be good to run
[18:30] <kgunn> then msg
[18:37] <dandrader> racarr, need your approval here, btw
[18:37] <dandrader> https://code.launchpad.net/~dandrader/mir/mergeFakeInputReaders/+merge/188662
[18:38] <dandrader> kgunn, do you know if I still need to install that? http://people.canonical.com/~thomir/python-ubuntu-platform-api_1.1daily13.06.13-0ubuntu1_armhf.deb
[18:39] <kgunn> dandrader: it should be in image 79
[18:39] <kgunn> dandrader: i saw robru indicate he pushed it to trunk...you could install just to be on the safe side
[18:40] <dandrader> kgunn,  I have version 1.1+13.10.20131001.1-0ubuntu1 installed. but I'm too dumb to figure out for sure if this version number is higher than 1.1daily13.06.13-0ubuntu1
[18:40] <kgunn> dandrader: lemme dig for a moment
[18:42] <kgunn> dandrader: yes...it has the relevent mp merged into that one...you're good to go
[18:42] <kgunn> don't need to load that deb
[18:42] <dandrader> kgunn, ok, thanks
[18:52] <racarr> back
[18:59] <dandrader> plars, hey
[19:00] <dandrader> kgunn, told me you're able to run some autopilot tests successfully on the device with unity8-mir
[19:00] <dandrader> "notepad and friends"...
[19:00] <dandrader> plars, is that so?
[19:14] <kgunn> dandrader: so i'm on n4, image 79(latest one)...i just ran the single ubuntu toolkit ap test per thomi's bug...it worked no prob
[19:15] <thomi> morning
[19:16] <dandrader> thomi, hi!
[19:16] <thomi> o/
[19:16] <dandrader> kgunn, ok, I certainly *must* have some outdated package them. gonna re-flash
[19:17] <kgunn> thomi: \o/
[19:17] <kgunn> thomi: we're so much better than where we were....wanna quick hang out to catch up ?
[19:17] <dandrader> was using a 78 image + apt-get update&dist-upgrade
[19:18] <kgunn> dandrader: yeah...slightly untrustworthy...
[19:18] <thomi> kgunn: sure
[19:19] <kgunn> dandrader: my steps were flash, touch writable_image & display-mir, then phablet-click-test-setup, then sudo apt-get install ubuntu-ui-toolkit-autopilot...then autopilot run -v ubuntuuitoolkit.tests.gallery.test_gallery.GenericTests.test_navigation
[19:19] <kgunn> and it just worked...pretty sure its a bum img/package cfg you got there
[19:20] <thomi> kgunn: ready when you are
[19:20] <kgunn> https://plus.google.com/hangouts/_/d3c5982669f82db12c38b740a81c7e362e6bbdfe?pqs=1&authuser=0&hl=en
[19:21] <kgunn> thomi: ^
[19:21] <kgunn> dandrader: ^ if you wanna join in case you got AP queries
[19:40] <dandrader> thomi, hey, don't forget about testing autopilot tests with unity8-mir in maguro. it's certainly not working for me
[19:40] <kgunn> dandrader: no doubt...paul just verified he tested on a n4
[19:41] <thomi> dandrader: yeah, weill do
[19:41] <thomi> flashing my GN now
[19:47] <kgunn> thomi: so http://bazaar.launchpad.net/~phablet-team/powerd/trunk/view/head:/cli/powerd-cli.h
[19:48] <kgunn> according to chicken
[19:49] <thomi> kgunn: so I tried that
[19:49] <kgunn> but
[19:49] <thomi> kgunn: problem is, you need to run it as root
[19:49] <thomi> also, it doesn't appear to work
[19:49] <kgunn> ricmm: no problem...we're discussing here
[19:49] <thomi> well, it worked for a while, then my screen flashed, now it's black, and I cannot wake it up any more
[19:50] <thomi> ahhh, unity8 crashed
[19:50] <ricmm> thomi: have you managed to autopilot unity8 yet then?
[19:50] <ricmm> or only apps
[19:50] <kgunn> ricmm: only apps...
[19:50] <thomi> ahh, I see. adb bug. adb drops shell connections when you attach a second device :-/
[19:51] <kgunn> unity8 suffers the weird screen blank scenario
[19:51] <thomi> ricmm: just flashing my GN, since people said they were having input problems with it
[19:52] <ricmm> thomi: did you get around the -fullscreen issue?
[19:52] <thomi> ricmm: I'm not sure what that issue is, sorry
[19:53] <ricmm> $ unity8 -testability -fullscreen
[19:53] <ricmm> was failing for me, because -fullscreen actually made Mir try and setup ./ullscreen as the socket path
[19:53] <ricmm> it taks a file,f argument for it
[19:53] <ricmm> takes*
[19:54] <thomi> haha
[19:54] <thomi> nice
[19:54] <ricmm> kgunn: I think a fix for that should be in Mir, making the server process take such a common argument like -f or -file sounds problematic
[19:55] <ricmm> make it mir_socket_path or something instead
[19:56] <thomi> or maybe make mir understand --
[19:56] <thomi> so you could do something like ./unity8 -some -mir -options -- -some -unity8 -options
[19:57] <ricmm> or that
[19:58] <kgunn> ricmm: you need it like asap ?
[20:00] <ricmm> kgunn: I dont need it, we can probably get away with a change in unity8-autopilot itself
[20:01] <kgunn> ricmm: thomi ...back to the wakelock need....what's the best way to manage a wakelock between unity8 and fakeunity8
[20:59] <racarr> robert_ancell: brt
[20:59] <robert_ancell> bus rapid transit?
[21:02] <thomi> kgunn: just confirmed that mir does not recognise the AP-created device on a maguro
[21:02] <thomi> filing a bug now
[21:13] <thomi> kgunn: racarr: what do you think? https://bugs.launchpad.net/mir/+bug/1234956
[21:14] <thomi> kgunn: seems like that should be 'critical', but I'll let you decide :)
[21:21] <kgunn> racarr: can you have a look ? ^
[21:39] <racarr> I have no maguro
[21:40] <racarr> kgunn: ^
[21:40] <kgunn> dang ti
[21:40] <kgunn> and it
[21:40] <racarr> ok heres the thing though do we even know
[21:40] <racarr> that autopilot finger is the issue as the bug implies
[21:40] <racarr> because don't we get that output on the nexus 4 too
[21:40] <racarr> while it works
[21:40] <kgunn> thomi: ^ ?
[21:41] <thomi> racarr: I don't think so, but I can test
[21:41] <thomi> one second.. have to replicate mysetup :-/
[21:41] <thomi> s/second/10 minutes/
[21:41] <kgunn> turns into a day
[21:42] <racarr> ok. ill start theorizing...?
[21:42] <racarr> not sure how else to help
[21:49] <racarr> thomi: Is there anything after
[21:50] <racarr> "Device added id=8, etc" in the log
[21:50] <thomi> racarr: no
[21:50] <thomi> there's the WW line, then the II line that you just referred to, then nothing
[21:51] <racarr> thomi: ok I think I see the problem
[21:51] <racarr> event hub.cpp l315
[21:51] <racarr> (('ABS_X', 0L),
[21:51] <racarr>                    AbsInfo(value=0, min=0, max=0, fuzz=0, flat=0, resolution=0)),
[21:51] <racarr>                   (('ABS_Y', 1L),
[21:51] <racarr>                    AbsInfo(value=0, min=0, max=0, fuzz=0, flat=0, resolution=0)),
[21:51] <racarr> min = max
[21:51] <racarr> so mir detects it as an invalid axis
[21:53] <thomi> any ideas why it works on the mako?
[21:53] <thomi> interestingly, that's not what autopilot sets
[21:53] <thomi> we set screen resolution as the max value
[21:54] <racarr> perhaps that code has a problem
[21:54] <kgunn> racarr: so you think its in autopilot code?
[21:54] <kgunn> racarr: or eventhub ?
[21:54] <thomi> easy enough to print out what capabilities we're setting
[21:54] <thomi> in autopilot i mean
[21:54] <racarr> I mean
[21:55] <racarr> the python log from checking the device
[21:55] <racarr> without mir involved
[21:55] <racarr> is returning this device with no valid x or y axis
[21:55] <racarr> so I think the problem is elsewhere
[21:55] <racarr> I think it would work, is ABS_X/ABS_Y had min or max values
[21:56] <racarr> or there was a
[21:56] <racarr> REL_X and REL_Y
[21:56] <thomi> that's true, I wonder if that's because autopilot is setting it invalid'ly, or the uinput system is changing it on us?
[21:56] <kgunn> thomi: ok..i feel stupid asking this, because its been such a _thing_ but the "unlocking shell" automation....where would that bug be ?
[21:56] <racarr> we should see what
[21:56] <racarr> evdev says about the device on uh
[21:56] <racarr> nexus 4
[21:56] <racarr> though that doesn't tell us that much really
[21:56] <racarr> im checking it out
[21:56] <thomi> kgunn: ugh. That's a topic for a hangout call
[21:57] <kgunn> now ?
[21:57] <thomi> kgunn: if you like
[21:57] <kgunn> https://plus.google.com/hangouts/_/463beb0eb5cb65ab4981da5ea7d6cbcf45202c13?pqs=1&authuser=0&hl=en
[22:02] <racarr> on nexus 4 the axis are set http://pastebin.ubuntu.com/6190026/
[22:02] <racarr> just using python and creating a device
[22:53] <thomi> ahhh, the problem is that the upa hack/fix returns 0,0,0,0 on the maguro
[22:54] <thomi> easy to fix, at least
[23:09] <lool> racarr: ^
[23:14] <thomi> racarr: robert_ancell, do either of you have a Nexus 10?
[23:14] <robert_ancell> thomi, no, I think only duflu has one int the team
[23:14] <robert_ancell> I mean kdub
[23:14] <thomi> kdub: awake still?
[23:16] <kdub> still awake!
[23:16] <kdub> i have a nexus 10
[23:18] <thomi> kdub: can you please tell me what this returns?
[23:18] <thomi> adb shell getprop ro.product.device
[23:19] <kdub> probably manta... checking
[23:19]  * thomi hopes it returns "manta\\n"
[23:19] <kdub> yes, manta
[23:19] <thomi> cool. Could I get someone to approve this please https://code.launchpad.net/~thomir/python-ubuntu-platform-api/fix-1234956/+merge/189193
[23:19] <thomi> and robert_ancell, could you please add that to the landing SS?
[23:20] <thomi> it's super-duper-critical.
[23:20] <robert_ancell> thomi, sure
[23:20] <thomi> like.. the world will 'splode otherwise
[23:20] <thomi> thanks :)
[23:20] <kdub> thomi, looks okay to me.. do i have approve powers?
[23:21] <robert_ancell> thomi, is ro.hardware an obsolete key?
[23:21] <kdub> i think its just a key that not everything sets
[23:21] <thomi> well, the problem was that on maguro it returned "tuna"
[23:22] <thomi> robert_ancell: I'm not sure what you mean by an "absolute key"
[23:22] <kdub> ah, a hardware variant key
[23:22] <thomi> kdub: if you approve it I can really approve it
[23:22] <robert_ancell> thomi, huh, it appears I can't edit that spreadsheet
[23:22] <robert_ancell> kgunn, ^
[23:22] <kdub> thomi, i did +1
[23:22] <thomi> kdub: thanks
[23:23] <thomi> that worked
[23:23] <thomi> canonical-ps is in ~python-upa-team
[23:23] <thomi> so, \o/
[23:23] <thomi> robert_ancell: paste me the link? I might have edit rights still
[23:24] <robert_ancell> I guess there's a whitelist of editors?
[23:26] <kgunn> robert_ancell: the landing sheet ?
[23:26] <robert_ancell> kgunn, yes, but thomi has got access to it
[23:27] <thomi> OK, I added it. not sure I did it right
[23:27] <kgunn> robert_ancell: shared with you again...
[23:27] <thomi> kgunn: do I need to ping anyone?
[23:27] <thomi> or is adding it to the SS enough?
[23:27] <kgunn> robert_ancell: it was the other google avatar :)...so you have access twice
[23:28] <robert_ancell> ah, that makes sense
[23:28] <robert_ancell> I *thought* I had access
[23:29] <kgunn> thomi: no, you might just update the status when its merged....but if you put waiting for merge & "its for mir" :) then it usually fast tracks it
[23:49] <kdub> oh boy, found the ioctl i needed in a package we already depend upon