[09:28] <tsdgeos> Saviq: you mentioned something about maxwidth/minwidth yesterday on my branch that for the desktop camera
[09:29] <tsdgeos> but i can't see maxwidth/minwidth being applied anywhere in the desktop stage
[09:30] <tsdgeos> mzanetti: ↑ you know anything about this?
[09:47] <pstolowski> Saviq, hey, what's the situation with unity8 landings? can i test & prepare silo 76 for landing?
[09:49] <Saviq> pstolowski, SDK is going in first with a small u8 landing
[09:49] <Saviq> tsdgeos, should be in WindowResizeArea
[09:51] <pstolowski> Saviq, what silo# is that?
[09:52] <Saviq> pstolowski, https://requests.ci-train.ubuntu.com/#/ticket/905
[09:52] <tsdgeos> Saviq: ah rihgt, totally failed at grepping
[09:53] <Saviq> pstolowski, it's at the top of QA's "ready" queue, so hopefully later today - I'd also like to drop https://requests.ci-train.ubuntu.com/#/ticket/993 on them after that
[09:54] <Saviq> obviously needs a rebuild + quick retest
[09:56] <pstolowski> Saviq, ack
[12:21] <tsdgeos> mzanetti: ping
[12:21] <tsdgeos> compare(listView.contentY, -listView.topMargin, "Launcher did not start up with first item unfolded");
[12:21] <tsdgeos> in the launcher fails if i add another item to MockLauncherModel
[12:21] <tsdgeos> in trunk
[12:21] <tsdgeos> does any of your new branches fix that?
[12:32] <tsdgeos> mzanetti: lp:~unity-team/unity8/launcher-tests-fail
[12:50] <Saviq> huhu, looks like autopilot in jenkaas was actually easier than expected
[13:01] <Saviq> now just to make it go in parallel with qml ones... need a new plugin
[13:37] <pstolowski> cimi, tsdgeos inline playback broken in single-preview silo 76
[13:37] <pstolowski> cimi, tsdgeos http://pastebin.ubuntu.com/15205625/
[13:37] <pstolowski> and dash crashes
[13:37] <tsdgeos> that is weird
[13:38] <tsdgeos> there's no reason inline playback should be affected at all
[13:38] <tsdgeos> let me try too
[13:38] <pstolowski> plus, preview models are destroyed veeery lazily, not sure if we can improve that?
[13:38] <pstolowski> e.g. i see a model gets destroyed after a few minutes
[13:39] <tsdgeos> model destroying should be as before approx
[13:40] <pstolowski> tsdgeos, that's possible, i just have never paid attention, now i have a debug message
[13:46] <pstolowski> tsdgeos, don't bother trying the silo.. works after reboot. grr
[13:46] <tsdgeos> :S
[13:46] <tsdgeos> k
[13:47] <pstolowski> something is fishy, but not this silo's fault
[14:08] <cimi> pstolowski, so all good?
[14:08] <pstolowski> cimi, yeah. there is some flakiness / race though, but i couldn't reproduce. unrelated to the silo amyway.
[14:52] <tsdgeos> Saviq: lp:~ci-train-bot/unity8/unity8-ubuntu-xenial-landing-064 has broken tags, can anyone clean them?
[14:53] <Saviq> tsdgeos, no, we should find where they come from (since we'll need to rebuild)
[14:53] <tsdgeos> Saviq: i think they come "from the silo"
[14:53] <tsdgeos> itslef
[14:53] <tsdgeos> i.e it got reused from old
[14:53] <Saviq> tsdgeos, train starts branches from scratch and then --overwrites he ones in ~ci-train-bot
[14:53] <tsdgeos> hmmm
[14:54] <tsdgeos> ok, let me run my clean branches script
[14:54] <Saviq> tsdgeos, but overwriting doesn't clear tags
[14:54] <Saviq> tsdgeos, we might ask them to delete the branches in ~ci-train-bot...
[14:54] <Saviq> but I'm afraid this will just come back again and again...
[14:55]  * Saviq wants git
[15:25] <tsdgeos> ltinkl: test for https://code.launchpad.net/~unity-team/unity8/fixTinyWindows/+merge/287173 ?
[15:27] <ltinkl> tsdgeos, sure, will add
[16:19] <tsdgeos> Saviq: mzanetti: ltinkl: you have bad tags
[16:19] <tsdgeos> http://paste.ubuntu.com/15206969/
[16:19] <Saviq> oops
[16:20]  * ltinkl on it
[16:20] <Saviq> that took a while, didn't it :)
[16:21] <tsdgeos> don't know didn't say much, but since it runs on its own may be
[16:21]  * tsdgeos times
[16:22]  * Saviq cleared
[16:27] <tsdgeos> pstolowski: ping
[16:27] <pstolowski> tsdgeos, hey
[16:27] <tsdgeos> pstolowski: how do i "favorite" the apps scope if it magically got unfavorited?
[16:29] <ltinkl> tsdgeos, ok, tags cleared and a test added (modified the unlock() test to use typeString() instead of clicking the pinpad buttons)
[16:29] <pstolowski> tsdgeos, oh, how did that happen?...  gsettings reset com.canonical.Unity.Dash favorite-scopes
[16:29] <pstolowski> tsdgeos, alternatively,  gsettings set com.canonical.Unity.Dash favorite-scopes "['scope://clickscope']"
[16:30] <tsdgeos> pstolowski: i guess at some point i uninstalled it (desktop)
[16:33] <Saviq> uninstalling anything shouldn't touch the favourites IMO
[16:38] <tsdgeos> Saviq: i guess the scope isn't there anymore
[16:39] <tsdgeos> so it can't be a favorite
[16:39] <pstolowski> hmm i think it should, otherwise this list would grow indefinitely.. upgrade is a different case (which we handle)
[16:39] <Saviq> how often do you think this would happen?
[16:39] <Saviq> that someone uninstalls a favourited scope
[16:40] <Saviq> and what impact does it have if there's even 100 things you need to skip in the list of favourites
[16:43] <pstolowski> Saviq, i hear you, but keeping a list which cannot be purged feels wrong
[16:43] <Saviq> pstolowski, that's actually how the launcher items behave
[16:43] <Saviq> in unity7 and unity8
[16:43] <Saviq> pstolowski, you can purge it - gsettings reset
[16:44] <tsdgeos> Saviq: it feels unexpected that something would be favorited on install
[16:44] <tsdgeos> if i installed it 1 year ago, favorited and then uninstalled
[16:44] <Saviq> it wouldn't "be favorited on install" you favorited it 1 year ago
[16:44] <Saviq> you never said that it's not your favourite now that you uninstalled it ;P
[16:44] <pstolowski> heh
[16:45] <tsdgeos> Saviq: i'm just saying it would seem unexpected to me
[16:45] <tsdgeos> the world is full of people that it's not me
[16:45] <tsdgeos> :D
[16:45] <Saviq> tsdgeos, was it unexpected on your launcher? ;)
[16:46] <tsdgeos> Saviq: yes
[16:46] <tsdgeos> it would too
[16:46] <tsdgeos> Saviq: 24 min for the clean branches run :D
[16:46] <Saviq> well, did you notice it, because that's exactly how it behaves
[16:46] <Saviq> tsdgeos, nice :)
[16:46] <Saviq> stress-testing LP are we ;)
[16:47] <tsdgeos> Saviq: i didn't notice since i have exactly 4 apps in my launcher that i always use and thus would not uninstall and install
[16:48] <Saviq> tsdgeos, saves my a$$ when I reinstall :)
[16:48] <Saviq> things on launcher just waiting for me to install them
[16:49] <tsdgeos> Saviq: so what does it do, it has "wholes"/[x] saying "this used to be an app and is no longer there?
[16:49] <tsdgeos> or it's not there and appears when you install?
[16:49] <Saviq> tsdgeos, filters, appears when I install
[16:49] <tsdgeos> s/wholes/holes
[16:50] <tsdgeos> k
[16:55] <tsdgeos> pstolowski: damn i think some changes in the scopes plugin are making the dash code hit this
[16:55] <tsdgeos> / TODO attach to sectionItem so we can accomodate to it changing its height
[16:55] <tsdgeos> ^_ ^
[16:55] <pstolowski> tsdgeos, huh? this is a qml TOODO?
[16:55] <tsdgeos> nah in the .cpp code
[16:56] <tsdgeos> it's just that the irc client ate the first /
[16:56] <Saviq> whoops :)
[16:58] <tsdgeos> i guess i'll attack that on monday
[16:59] <pstolowski> tsdgeos, hmm i'm not sure what kind of change would that be?
[17:00] <tsdgeos> pstolowski: it's the smart model updates that now reuse categories
[17:00] <tsdgeos> so if you happen to reuse a category that had section title to one that hasn't
[17:00] <tsdgeos> boom
[17:00] <tsdgeos> breaks
[17:00] <tsdgeos> because stuff should move up because the category title just disappears
[17:00] <tsdgeos> -s+ed
[17:01] <pstolowski> tsdgeos, ah.. but that's not new.. landed 2 months ago or so
[17:01] <pstolowski> tsdgeos, ah, so it's a very specific case, maybe we didn't hit it
[17:02] <tsdgeos> yeah
[17:02] <tsdgeos> i can easily reproduce on the desktop unity8-dash now
[17:02] <tsdgeos> but not on the phone
[17:02] <tsdgeos> because *things*
[17:02] <tsdgeos> :D
[17:07] <cimi> tsdgeos, approved those three branches of yesterday
[17:07] <cimi> tsdgeos, tx
[17:08] <tsdgeos> cimi: cool
[17:08]  * tsdgeos waves eow
[17:09] <cimi> so. fast.
[17:09] <cimi> :D
[17:09] <cimi> sometimes I wish he was waiting for us to wave back :D
[21:58] <Saviq> mterry, hey, if you're still here, could use your eyes on https://code.launchpad.net/~saviq/unity8/autopilot-dep8/+merge/287228 - you can see the test results in https://requests.ci-train.ubuntu.com/#/ticket/1053 - I'd like to put this into trunk directly so our CI benefits from this straight away
[21:58] <mterry> Saviq, hey
[21:58] <mterry> looking
[22:02] <mterry> Saviq, so the results of the dep8 test are in those "vivid landing-023 excuses.html" lines in the silo?
[22:02] <Saviq> mterry, yup
[22:02] <mterry> Saviq, branch makes sense though
[22:03] <mterry> Saviq, the vivid link brings me to an ubuntu-system-settings page though...
[22:03] <Saviq> mterry, xenial, sorry
[22:03] <Saviq> mterry, bug in train
[22:04] <mterry> Saviq, ah ok
[22:04] <Saviq> mterry, the silo is xenial only
[22:04] <mterry> Saviq, ah so it is, hadn't noticed.  dep8 is only good for xenial?
[22:04] <Saviq> mterry, no, this silo is xenial only, didn't want to waste britney's or builders' time
[22:05] <Saviq> mterry, unfortunately there's no way to SKIP a test based on whether it's touch or not, so the early-exit is the best we can do - also the isolation-machine requirement makes sure that it SKIPs when it's obviously wrong (like chroot or whatnot)
[22:05] <mterry> Saviq, yeah -- so hard to run outside of the builder
[22:06] <Saviq> mterry, those never run on builders anyway
[22:06] <mterry> Saviq, why do some arch's fail?
[22:06] <mterry> Saviq, right.  I guess I meant outside of britney
[22:06] <Saviq> mterry, outside of i386, amd64, armhf, missing deps
[22:06] <Saviq> mterry, on armhf, because we've not fixed our tests yet
[22:06] <Saviq> timing-wise
[22:07] <mterry> Saviq, well 2 arches is something anyway.  Even though armhf would be the real prize
[22:08] <Saviq> mterry, bug #1550543
[22:08] <ubot5`> bug 1550543 in CI Train [cu2d] "Automated tests publish results for disabled target seriee" [Undecided,New] https://launchpad.net/bugs/1550543
[22:09] <mterry> Saviq, so where are these autopilot tests actually running in this setup?  The ones at https://requests.ci-train.ubuntu.com/static/britney/xenial/landing-023/excuses.html are skipped because they aren't on the device
[22:09] <Saviq> mterry, our jenkins
[22:10] <Saviq> mterry, https://unity8-jenkins.ubuntu.com/job/test-0-autopkgtest/247/ red because I've busted the jenkins job
[22:10] <mterry> ah right, because of course britney doesn't have access to devices.  I forgot that jenkins was the gatekeeper to thos
[22:11] <Saviq> mterry, actually red because something failed completely in this case
[22:11] <Saviq> mterry, https://unity8-jenkins.ubuntu.com/job/test-0-autopkgtest/243/ a better result
[22:12] <mterry> Saviq, OK, now I'm caught up with how this works  :)
[22:12] <mterry> Saviq, took me a while to understand what was running what
[22:13] <Saviq> mterry, our jenkins runs them all - qml tests on amd64; autopilot tests on phones
[22:13] <Saviq> mterry, britney can only run qml tests
[22:13] <mterry> right
[22:13] <Saviq> at least until we make our autopilot runnable under xvfb or something
[22:14] <mterry> Saviq, but yeah the branch makes sense, let me approve
[22:16] <mterry> Saviq, allow-stderr seems like such an annoying flag.  I feel like it should be default
[22:16] <mterry> But I get that it's useful to catch surprises
[22:17] <Saviq> mterry, actually the problem there is set -ex prints on stderr
[22:17] <Saviq> mterry, and autopilot, too
[22:17] <mterry> Saviq, yeah -- lots of stuff does all the time is the annoying thing.  Only a very strict tool is so careful to keep to stdout
[22:18] <Saviq> mterry, FWIW we should be able to get rid of that for qmluitests https://unity8-jenkins.ubuntu.com/job/test-0-autopkgtest/248/label=amd64,release=vivid+overlay/artifact/output/qmluitests.sh-stderr
[22:19] <Saviq> and it actually does show useful stuff
[22:20] <Saviq> but we'll have to get rid of set -ex, which is kinda useful
[22:20] <mterry> Saviq, yeah for build warnings.  but that might be better done as an isolated "build warnings" dep8
[22:20] <Saviq> mterry, oh well, dep8 is too late for build warnings actually ;)
[22:20] <mterry> Saviq, but failing on build warnings are annoying
[22:20] <Saviq> mterry, since everything is^Wshould be built by then ;)
[22:21] <Saviq> ok /me clears silo
[22:21] <mterry> Saviq, packages that do that always cause problems on gcc upgrades or when some library deprecates some call.  Big pain in production
[22:21] <Saviq> yup
[22:23] <Saviq> ok let's see how that goes, I can see that arale is complaining already, need to find out what went wrong with the phones
[22:25] <Saviq> and GAAAH fucking jenkins for trying to re-escape quotes and failing like fuck
[22:29] <Saviq> hmm looks like krillin and devel-proposed don't mix