[07:05] <Cimi> morning
[07:09] <Cimi> Saviq, noticed a bug with the new indicators
[07:09] <Darth-Apple> Morning
[07:09] <Cimi> Saviq, the orange highlight moves and is not centered
[07:25] <Saviq> Cimi, hmm, let me flash a previous image
[07:25] <Saviq> I can't see how I might've caused that
[07:37] <Cimi> Saviq, I realised your code kinda works the same, but if you prefer you can use what I put https://code.launchpad.net/~saviq/unity8/rating-non-interactive/+merge/225568
[07:40] <Saviq> Cimi, I think mine's good enough, can't really foresee when it'd break
[07:40] <Saviq> we could probably just tap 0,0, let me see if that would work
[07:42] <Saviq> Cimi, pushed 0, 0, works just fine - to really make sure what we're tapping on, we'd need to add objectNames to the icons and find it/tap on it
[07:42] <Saviq> not sure it's worth it though
[07:43] <Saviq> mhr3_, kudos on grabbing r1000 ;)
[07:47] <mhr3_> Saviq, \o/
[07:47] <mhr3_> Saviq, i want 1024 too though :)
[07:47] <Saviq> mhr3_, hmm you'll need to be rather quick ;)
[07:47] <Saviq> we're at 1009 already ;)
[07:48] <Saviq> 1010 rather, with LP syncing translations
[07:54] <sil2100> Saviq: hey! Could you remind me how it was before in the past with LP: #1320264 ? I mean, the approvers were listed only from the list of comment-approvers or did it also include the top-approver?
[07:55] <Saviq> sil2100, it was only comment-approvers
[07:55] <sil2100> ACK
[07:55] <Saviq> sil2100, but it'd be nice to have the top-approver, too, if possible
[08:05] <Cimi> Saviq, you can approve https://code.launchpad.net/~cimi/ubuntu-system-settings/workaround_lp1334429/+merge/225458 :)
[08:06] <Saviq> Cimi, move the fixme above please
[08:06] <Saviq> Cimi, and "relative to the panel"
[08:15] <mikenagle> mhall119 - for a few reasons. When we did user research, we found that people didn't like their apps being so close to apps they didn't own. And they found it confusing. Also, as the nuber of available apps grow, users will need a way to browse them, to search and filter them and to have apps recommended. And we can't provide that depth of experience when they're displayed as a simple list below the installed apps
[09:15] <Saviq> Cimi, the orange highlight bar is a bug in trunk, made a bit more visible by my branch probably
[09:15] <Saviq> Cimi, can you file please?
[09:15] <Cimi> Saviq, sure
[09:15] <Cimi> Saviq, or just fix it
[09:15] <Cimi> let me ping dednick
[09:16] <Cimi> dednick, highlight on the indicator is not centered and moves
[09:16] <Cimi> the orange bar
[09:16] <Cimi> see it on battery for example
[09:17] <Cimi> Saviq, units.gu(0.4) is interpreted as exactly 0.5?
[09:17] <Saviq> Cimi, no, you should just no use it
[09:18] <Cimi> Saviq, I know it is wrong
[09:18] <Saviq> Cimi, it not being centered I think might be a false feeling
[09:18] <Cimi> Saviq, I was just wondering what is is interpreted
[09:18] <Cimi> no it is not
[09:18] <Cimi> look at batery
[09:18] <Saviq> Cimi, it's not interpreted, it's calculated
[09:18] <Cimi> or bluetooth
[09:19] <Saviq> Cimi, can't, now, had to downgrade to confirm I'm not at fault :P
[09:19] <Saviq> Cimi, yeah, confirmed on previous icons as well
[09:21] <Saviq> Cimi, please just file a bug
[09:22] <Saviq> Cimi, it'll be a bite-size to fix in London or so
[09:26] <Cimi> https://bugs.launchpad.net/unity8/+bug/1337771
[09:28] <karni> mhr3: Would you have a hint for me regarding proper access to search metadata class? https://pastebin.canonical.com/112989/
[09:29] <mhr3> karni, you need to #include it ;)
[09:29] <mhr3> karni, really, should just #include <unity-scopes.h>
[09:30] <karni> mhr3: huh.. I would expect some "unknown class" what not error ;P
[09:30] <mhr3> karni, yea, that's unknown class in g++ lingo :)
[09:30] <karni> mhr3: thanks :)
[09:30] <mhr3> clearly you don't speak g++
[09:30] <mhr3> :)
[09:30] <karni> worky!
[09:33] <Cimi> Saviq, is there a bug for all the new icons being in wrong aspect ratio and using svg instead png?
[09:34] <Saviq> Cimi, what do you mean wrong aspect ratio? where?
[09:34] <Saviq> Cimi, you mean they're not square?
[09:34] <Saviq> Cimi, and they have to be svg, they won't render well otherwise
[09:34] <Cimi> Saviq, aspect ratio of icons is squashed
[09:34] <Cimi> web browser icon is wrong
[09:35] <Cimi> hold on, new image
[09:35] <Cimi> I was on 112
[09:35] <Saviq> Cimi, when you say "all those icons", please say that you mean app icons
[09:35] <Saviq> Cimi, those are not "all" icons
[09:35] <Cimi> well, other icons are indicators
[09:35] <Cimi> :)
[09:36] <Cimi> so yeah all the icons in the dash
[09:36] <Cimi> and launcher
[09:36] <Saviq> Cimi, no, those are also icons, not indicators
[09:36] <Saviq> Cimi, and launcher should be fine
[09:36] <Cimi> let me check I just upgraded to 113
[09:36] <Saviq> Cimi, dash we know is wrong
[09:36] <Saviq> doesn'
[09:36] <Saviq> t matter
[09:36] <Cimi> ok
[09:37] <Cimi> so web browser is wrong
[09:37] <Cimi> we should use a png or change the rendering
[09:37] <Cimi> the icons are squashed in the dash
[09:37] <Saviq> Cimi, for app icons I agree, we should switch to pngs
[09:38] <Saviq> Cimi, and indeed something's still wrong in the launcher
[09:38] <Cimi> the highlight and inset is wrong
[09:38] <Cimi> think I saw a MP for it
[09:38] <Saviq> Cimi, yeah, there's a bunch of launcher MPs (that you could review btw)
[09:39] <Cimi> yeah can do
[09:39] <Cimi> just looping into bugs
[09:39] <Saviq> damn bzr
[09:40] <Cimi> tags?
[09:40] <Saviq> no
[09:40] <Saviq> merge http://paste.ubuntu.com/7746302/
[09:40] <Saviq> mzanetti, check this out ↑...
[09:41] <mhr3> cross merges? eek
[09:41] <mzanetti> Saviq: hmm...
[09:41] <mhr3> try with different merge alg
[09:41] <mzanetti> Saviq: did I mess this up?
[09:41] <Saviq> the original diff http://pastebin.ubuntu.com/7746308/
[09:41] <Saviq> mzanetti, no, bzr did
[09:42] <Saviq> --weave it is then
[09:42] <Saviq> better
[09:43] <mhr3> Saviq, :/ hilfe http://paste.ubuntu.com/7746311/
[09:43] <mhr3> QFATAL : qmltestrunner::test_categories_roles(Model.roles[expansionQuery]) Using un-implemented Categories role
[09:43] <Saviq> mhr3, well, implement the Categories role!
[09:44] <mhr3> but it's there!
[09:45] <Saviq> mhr3, that diff's on top of unity-api trunk?
[09:46] <mhr3> aaah, nevermind
[09:47] <mhr3> there's a big if and two switches.. didn't notice that
[09:48] <Saviq> :)
[09:48] <Saviq> mzanetti, new header is ready, right?
[09:50] <Saviq> mzanetti, oh, I meant to remove the whole line with "recent searches", not just the X button
[09:50] <mzanetti> Saviq: oh
[09:50] <mzanetti> ack, lemme drop it
[09:51] <mzanetti> otherwise, I'd say yes, its ready, unless you find some more stuff
[09:54] <Saviq> mzanetti, wonder, if the popover's gone, could we unfocus the search entry on another onPressed in the IMA?
[09:54] <Saviq> mzanetti, and shouldn't we disable it when unfocused?
[09:55] <Saviq> mzanetti, oh and hmm, it shouldn't cover the whole header
[09:55] <mzanetti> Saviq: ?
[09:55] <dednick> Saviq: by not centered, you mean it moves as you move drag down?
[09:55] <Saviq> mzanetti, just the active parts
[09:55] <mzanetti> slow down
[09:55] <Saviq> mzanetti, I'll write on the MP
[09:56] <Saviq> dednick, no, not centered to the icons horizontally (sometimes)
[09:56] <dednick> Saviq: hm. ok, i'll take a look.
[09:56] <Saviq> dednick, not now
[09:56] <Saviq> dednick, Cimi will file a bug
[09:56] <dednick> Saviq: ok
[09:56] <Saviq> dednick, and we'll tackle it somewhen
[09:56] <Saviq> Cimi, right? you will?
[09:57] <dednick> Saviq: actually, i think i've fixed it in a MP already
[09:57] <Cimi> Saviq, I did
[09:57] <Saviq> dednick, oh ok :)
[09:57] <Cimi> https://bugs.launchpad.net/unity8/+bug/1337771
[09:58] <dednick> Saviq, Cimi: https://code.launchpad.net/~nick-dedekind/unity8/indicator.dismissal/+merge/225338
[09:58] <dednick> kind of unrelated to what the mp was for, but it was a one line.
[09:59] <Saviq> k
[09:59] <dednick> and fyi, the move left to right as you're dragging is a feature,  but it should settle when you let go.
[09:59] <Saviq> dednick, still moves as you move
[09:59] <Saviq> ah ok
[10:00] <Saviq> fixeded
[10:01] <dednick> one we get to rtm we should really get some fully featured design docs for each component. everything is all over the place.
[10:02] <Saviq> dednick, LOL
[10:02] <Saviq> dednick, we'll get new ones, don't worry ;)
[10:12] <mhr3> haha, that was funny
[10:12] <mhr3> dednick, come on, it's not like this was your first week :P
[10:14] <Saviq> mzanetti, commented
[10:15] <mzanetti> Saviq: cheres, will look at it in a bit
[10:16] <Saviq> it's spelled chères
[10:16] <dednick> mhr3: :)
[10:17] <dednick> well, if we want to be able to "sell" this thing, we should probably have some documentation for customers, let alone for us!
[10:31] <MacSlow> dednick, quick question regarding UnityMenuModel... can one catch a DBus-connection going missing on the QML-side?
[10:32] <dednick> MacSlow: don't think so
[10:48] <mhr3> tsdgeos is also on holiday?
[10:51] <Cimi> mhr3, yes
[10:51] <mhr3> didn't he just get back from one?
[10:51] <mhr3> slacker :P
[10:52] <Cimi> mhr3, I am worse cause I am on holiday when you are sprinting :P
[10:52] <Cimi> love that
[10:52] <Cimi> ahah
[10:53] <Cimi> Saviq, why the indicator wobbling is a feature?
[10:53] <Cimi> it's a bug to me
[10:53] <Saviq> Cimi, because dednick said so
[10:54] <dednick> because design said so
[10:54] <Cimi> I mean the fact is moves 2 pixels
[10:54] <Cimi> it is a bug
[10:54] <Cimi> design didn't see it
[10:54] <Cimi> I did
[10:54] <Cimi> and matthieu saw it too
[10:55] <Cimi> maybe other designers didn't notice
[10:55] <dednick> 2 pixels?
[10:55] <Cimi> if you slowly pull down the menu, you see it wobbling
[10:55] <Cimi> sound and battery
[10:55] <Cimi> a couple of pixels roughly
[10:56] <Saviq> dednick, yeah, as it sticks to one icon, it also gets pulled a few pixels to either side following your finger
[10:56] <dednick> Cimi: its a specific feature requested by design for the current indicator to shift slightless left/right as you finger moves left right.
[10:56] <Cimi> dednick, the finger doesn't move
[10:56] <Saviq> Cimi, it does
[10:56] <Saviq> Cimi, finger always moves
[10:57] <Cimi> ah I see
[10:57] <Cimi> well I will open a bug
[10:57] <Cimi> we should have a dead zone
[10:58] <dednick> Cimi: huh?
[10:58] <Cimi> dednick, it stays in the center of the indicator unless you move more than some pixels
[10:58] <Cimi> this will avoid wobblyness
[11:02] <dednick> Cimi: a threshold you mean.
[11:03] <dednick> or a damper
[11:03] <Cimi> dednick, it is called dead zone on interfaces/controllers
[11:04] <Cimi> Saviq, i discovered sth yesterday
[11:05] <Cimi> Saviq, basically the OSK mouse area and input area
[11:05] <Cimi> Saviq, are always on screen
[11:05] <Cimi> but disabled
[11:06] <Cimi> do we want to put them offscreen or visible false when kdb is off?
[11:06] <Saviq> Cimi, that's actually a bug in keyboard
[11:06] <Saviq> Cimi, it should slide up and down as it opens / closes
[11:06] <Saviq> Cimi, and that would communicate the correct coords
[11:07] <Cimi> Saviq, it does, but when is closed I think UbuntuKeyboarInfo has y on screen
[11:07] <Cimi> like y is screen height - kbd height
[11:07] <Saviq> Cimi, it doesn't slide
[11:07] <Saviq> Cimi, it just goes away in one frame
[11:07] <Saviq> Cimi, if you swipe the keyboard away, the input area follows
[11:07] <Cimi> Saviq, yes
[11:07] <Cimi> Saviq, but then
[11:08] <Cimi> Saviq, when the kbd is off screen
[11:08] <Saviq> Cimi, I know, it comes back
[11:08] <Cimi> Saviq, the inputarea is on screen, as well with the mouse area
[11:08] <Cimi> saw it putting a rect and anchoring
[11:08] <Cimi> I always have the rect in every app
[11:08] <Saviq> Cimi, yeah, but isn't it because the keyboard communicates the coordinates still?
[11:08] <Saviq> when it should be 0 height?
[11:08] <Cimi> Saviq, either 0 height
[11:09] <Cimi> Saviq, or y being screen_height
[11:09] <Cimi> while y I think is
[11:09] <Cimi> y = screen_height - kbd_height
[11:09] <Cimi> when keyboard is off
[11:09] <Saviq> Cimi, no it's not
[11:10] <Cimi> as soon as it pops in, you see y starting again from screen_height
[11:10] <Saviq> Cimi, because as you slide the keyboard away with your finger
[11:10] <Saviq> Cimi, the rectangle follows
[11:10] <Saviq> Cimi, if what you say would be true
[11:10] <Saviq> Cimi, the rectangle would stay in place but shring
[11:10] <Saviq> shrink
[11:10] <Cimi> Saviq, yes
[11:10] <Cimi> Saviq, until it is outside
[11:10] <Cimi> let me do video
[11:10] <Saviq> Cimi, don't
[11:10] <Saviq> Cimi, it's 10 lines of code
[11:11] <Saviq> Cimi, you're wasting time
[11:11] <Saviq> Cimi, just find what's the issue
[11:11] <Saviq> fix it
[11:11] <Saviq> done
[11:12] <Cimi> Saviq, well it is a bug in ubuntukeyboard maybe
[11:12] <Cimi> Saviq, and I don't know how we want to fix it
[11:12] <Cimi> Saviq, either flagging visible false
[11:12] <Saviq> Cimi, file a bug against the ubuntukeyboard then
[11:12] <Cimi> or reducing height
[11:12] <Cimi> or moving y outside screen
[11:13] <Saviq> Cimi, I told you how, keyboard should slide out and be 0 height
[11:13] <Cimi> so both
[11:13] <Cimi> y = screen height and height 0
[11:13] <Cimi> ok
[11:13] <Saviq> Cimi, it needs to be a rectangle, in global coordinates, that the OSK covers
[11:13] <Cimi> https://www.youtube.com/watch?v=4QWSJxnL7FY
[11:14] <Cimi> when i slide out, you see the yellow color of the mousearea delayed following
[11:14] <Cimi> (what a bad color choice yellow over white)
[11:15] <Saviq> Cimi, yes, use red
[11:15] <Saviq> Cimi, but *yes*
[11:15] <Saviq> *I know*
[11:16] <Cimi> filing bug
[11:20] <Cimi> https://bugs.launchpad.net/ubuntu-keyboard/+bug/1337814
[11:25] <mzanetti> Saviq: re "If popover is already gone, we should unfocus the text field."
[11:25] <mzanetti> Saviq: would you expect a press on the dash content to just unfocus the textfield
[11:25] <mzanetti> or unfocus *and* do the press
[11:25] <Saviq> mzanetti, both
[11:25] <mzanetti> ok
[11:25] <Saviq> mzanetti, basically, onPressed
[11:26] <mzanetti> ack
[11:26] <Saviq> mzanetti, so that when you swipe, it dismisses as well
[11:41] <Saviq> mzanetti, review for you please https://code.launchpad.net/~unity-team/unity8/scope-customizations/+merge/225170
[11:41] <mzanetti> ack
[11:55] <Saviq> mzanetti, anything I forgot https://code.launchpad.net/~elopio/unity8/flake8-precommit/+merge/224852/comments/542929 ?
[11:55]  * mzanetti reads
[11:56] <bregma> Saviq, just for the record, Unity 8 on desktop was critically broken by a bug in Mesa earlier, it got fixed Thursday and the theme change tested out OK on desktop -- sorry for the delay
[11:57] <Saviq> bregma, great, thanks
[12:02] <mzanetti> Saviq: https://code.launchpad.net/~elopio/unity8/flake8-precommit/+merge/224852/comments/542934
[12:02] <Saviq> mzanetti, I did write "drop the second hook"
[12:02] <mzanetti> meh... sorry
[12:03] <mzanetti> Saviq: but I guess the thing about blocking vs warning is still valid
[12:03] <Saviq> mzanetti, yes
[12:09] <Cimi> Saviq, what did you mean yesterday about the dots at the bottom of the page?
[12:09] <Cimi> in the wizard?
[12:10] <Saviq> Cimi, there's 4 dots
[12:10] <Saviq> Cimi, but I had 4 pages (no SIM is an optional page?)
[12:10] <Saviq> 5 pages
[12:10] <Cimi> Saviq, ah no sim then
[12:12] <Saviq> mzanetti, re: can't add a container around them
[12:12] <Saviq> mzanetti, I'd probably just go for x/y/widht/height
[12:12] <Saviq> mzanetti, or maybe left/top anchor  + width/height from the edit fields
[12:12] <Saviq> -s
[12:13] <Saviq> mzanetti, but I understand it might be fine as is
[12:13] <mzanetti> Saviq: well, the end result is the same now actually
[12:14] <karni> Saviq: would it be possible to enable cache for the images shown in previews? basically, we have scopes that use QNetworkDiskCache to cache requests, but when we pass URLs to scope result or preview, it would always fetch the image from the network. Could we somehow improve that?
[12:15] <karni> Saviq: If not, I think I found a workaround.. fetch the images manually, using QNetworkDiskCache (with limited size), and pass a local file url to the scope.
[12:15] <karni> Saviq: it's something I imagine the framework should do, but if we can't have that, at least there would be a way to speed things up
[12:16] <mzanetti> Saviq: https://code.launchpad.net/~unity-team/unity8/scope-customizations/+merge/225170/comments/542940
[12:17] <Saviq> karni, https://code.launchpad.net/~unity-team/unity8/cache-network-data/+merge/224995
[12:18] <Saviq> karni, merged yesterday ;)
[12:19] <karni> Saviq: WOHOOO!!
[12:19] <karni> Saviq: awesomesauce!
[12:22] <Saviq> mzanetti, IMA hides search even if it's not empty - it should only hide on empty
[12:22] <mzanetti> d'oh! fixing...
[12:22] <Saviq> mzanetti, states! ;P
[12:23] <Saviq> karni, there's still a delay to actually hit up the server and check for validity
[12:23] <Saviq> karni, but if the servers are configured right, it should not even do that
[12:23] <Saviq> (too often)
[12:24] <Saviq> mzanetti, apt-cache policy qtdeclarative5-ubuntu-ui-toolkit-plugin ?
[12:25] <mzanetti> Saviq: what would be in there?
[12:25] <Saviq> mzanetti, should be 20140702.1 or so
[12:25] <mzanetti> Saviq: so it didn't have the pageheader when I tried 2 hours ago
[12:25] <mzanetti> so now I have the packages with the pageheader installed
[12:25] <Saviq> mzanetti, ah right
[12:25] <Saviq> mzanetti, that's why
[12:26] <Saviq> you have wrong icons
[12:26] <mzanetti> but aren't the icons in another package?
[12:26] <mzanetti> ubuntu-mobile-icons or similar?
[12:26] <Saviq> mzanetti, you're missing uitk changes
[12:26] <Saviq> mzanetti, that switch to the suru theme
[12:26] <Saviq> mzanetti, and the icons I used are not there in ubuntu-mobile
[12:26] <mzanetti> ah ok
[12:28] <karni> Saviq: ack, thanks :)
[12:34] <facundobatista> Holas
[12:40] <Saviq> \o
[12:42] <Saviq> oh yay, segfault
[12:46] <Saviq> ha
[12:46] <Saviq> mzanetti, ok, you need to fix testDashContent, not myself ;P
[12:50] <Saviq> mzanetti, and still Needs fixing on new-header
[12:50] <mzanetti> Saviq: ok
[12:50] <Saviq> /food
[12:57] <MacSlow> greyback, hey there
[12:57] <greyback> MacSlow: hi
[12:58] <MacSlow> greyback, is it expected to not be able to run any mir-client (and mir-server) on current radeon-based systems?
[12:58] <greyback> MacSlow: have you Catalyst installed (the non-free driver)?
[12:58] <MacSlow> greyback, OpenSource-driver of course... I'm getting this "libEGL warning: unsupported platform (null)"
[12:59] <MacSlow> greyback, I tried the mir_demo_server_basic and mir_demo_server_translucent sofar
[12:59] <greyback> MacSlow: try setting "EGL_PLATFORM=mir"
[12:59] <MacSlow> greyback, ok... trying now
[12:59] <greyback> I had thought it shouldn't be necessary any more, but maybe
[13:01] <MacSlow> greyback, nope... still getting the same error
[13:02] <greyback> MacSlow: any other errors?
[13:02] <greyback> it is crashing?
[13:02] <greyback> where are you running it from? You need to be in a VT
[13:03] <MacSlow> greyback, I am on VTs
[13:04] <MacSlow> greyback, starting mir_demo_server_basic on VT1... switching to VT2 and starting my client program (the unity-system-compositor-spinner)
[13:04] <MacSlow> greyback, with EGL_PLATFORM=mir ./unity-system-compositor-spinner
[13:04] <MacSlow> greyback, this used to work just fine a few weeks back
[13:05] <greyback> MacSlow: ok then something changed. You'd better ask the Mir guys, it's beyond me
[13:05] <MacSlow> greyback, ok
[13:14] <MacSlow> greyback, just tried other mir-clients... they work :/
[13:19] <greyback> MacSlow: you recompiled it? mir 0.4.0 landed with client-facing API changes...
[13:21] <MacSlow> greyback, I pull updates and recompiled... I think :)
[13:21]  * MacSlow checks
[13:31] <MacSlow> greyback, still not working
[13:32] <greyback> MacSlow: is it crashing? Is it hanging? Any more debug output? Did you try turning on component reports (http://unity.ubuntu.com/mir/component_reports.html)
[13:33] <MacSlow> greyback, it's just not starting... I'll try the ltt-reporting to see if that gives any clues
[13:34] <greyback> MacSlow: it exists cleanly?
[13:34] <greyback> exits
[13:34] <MacSlow> greyback, yup
[13:34] <karni> Saviq: Is it possible to uncollapse a preview text by default?
[13:34] <greyback> MacSlow: that's odd.
[13:37] <MacSlow> greyback, that's all the output I'm getting from my mir-client on the VT http://pastebin.ubuntu.com/7747141
[13:48] <dednick> Cimi: would you mind reviewing https://code.launchpad.net/~nick-dedekind/ubuntu-settings-components/transfer-menu/+merge/224672
[13:50] <Cimi> yes
[13:50] <MacSlow> greyback, hm... I did run mir-server and my mir-client with lttng but the result doesn't seem to contain much
[14:14] <Saviq> karni, will be
[14:15] <Saviq> karni, once we implement the expandable widget, text alone will be expanded by default, only if you put it in an expandable will it collapse
[14:20] <karni> Saviq: ah, cool!
[14:43] <mzanetti> Saviq: one more question:
[14:43] <mzanetti> searchfield focused, empty, popup open
[14:43] <mzanetti> user taps on the IMA
[14:43] <mzanetti> popup hides
[14:43] <mzanetti> shoudl the searchfield unfocus and hide in the same tap?
[14:43] <mzanetti> or should hiding popup and hiding search still be 2 steps?
[14:44] <Saviq> mzanetti, good question
[14:44] <mzanetti> I changed it here to do both in one
[14:44] <mzanetti> seems better to me
[14:44] <Saviq> mzanetti, probably both
[14:44] <Saviq> mzanetti, yeah
[14:44] <mzanetti> ack
[14:46] <mzanetti> Saviq: you said something about testDash?
[14:46] <mzanetti> not passing for you?
[14:46] <Saviq> mzanetti, testDashContent
[14:46] <mzanetti> ah
[14:46] <Saviq> mzanetti, 10 fails
[14:47] <Saviq> mzanetti, there's tabBar something
[14:47] <mzanetti> ack. missed that one. will fix
[14:47] <Saviq> mzanetti, I'll actually kick a full run here
[14:53] <mzanetti> at tests suites like DashContent I'm really happy about the FUNCTION= feature
[14:55] <mzanetti> although test_departments would also require a LINE= feature :D
[15:01] <MacSlow> greyback, what's the correct way to start a mir-client on a device
[15:01] <greyback> MacSlow: with unity8 running as the mir server?
[15:02] <MacSlow> greyback, no... unity-system-compositor is in control iirc
[15:02] <greyback> MacSlow: check if MIR_SOCKET=/run/mir_socket is set
[15:03] <greyback> note that unity8 is a mir server too, it is both a mir server and a client of USC (i.e. is a 'nested' server(
[15:04] <MacSlow> greyback, MIR_SOCKET isn't set...
[15:04] <greyback> MacSlow: then try setting it, or else use the -m switch - equivalent actions
[15:05] <MacSlow> greyback, just trying to do "sudo ./unity-system-compositor-spinner -m /run/mir_socket" fails similarly to what I'm seeing on the my desktop
[15:06] <Saviq> dednick, https://code.launchpad.net/~nick-dedekind/unity8/transfer-menu/+merge/224673/comments/543028, there's inline, too
[15:06] <greyback> MacSlow:  MIR_SOCKET=/run/mir_socket /usr/bin/unity-system-compositor-spinner <- works for me
[15:07] <greyback> MacSlow: also note, running as sudo, if it crashes, you don't see the "aborted" message
[15:09] <MacSlow> greyback, hm... MIR_SOCKET=/run/mir_socket /usr/bin/unity-system-compositor-spinner doesn't fail or exit... but I still don't see anything...
[15:09] <MacSlow> greyback, the greeter is unresponsive to any touch-input
[15:09] <MacSlow> has to be something else that got broken
[15:09] <greyback> MacSlow: set "MIR_CLIENT_RPC_REPORT=log " in your env & try again. Does it report once per frame?
[15:10] <greyback> MacSlow: it's possible USC isn't bringing the spinner to the front
[15:12] <MacSlow> greyback, ok... restarted my N10...
[15:13] <greyback> MacSlow: super simple test: "sudo stop lightdm" to kill USC. Then "mir_demo_server_shell&" and "MIR_SOCKET=/run/user/32011/mir_socket /usr/bin/unity-system-compositor-spinner"
[15:13] <MacSlow> greyback, with MIR_CLIENT_RPC_REPORT=log I got feedback "Invocation succeeded: id: 256 method_name: next_buffer ..." etc
[15:14] <greyback> yep that's good, means client requesting new buffers for each frame & mir is giving them to it
[15:14] <greyback> however it may be that USC simply is not making the spinner surface visible
[15:15] <greyback> I've never dug into USC code, I don't know how it decides what surfaces to show at what times
[15:15] <MacSlow> greyback, no mir_demo_server_shell on the N10... have to grab and copile it
[15:15] <greyback> MacSlow: apt-get install mir-demos
[15:15] <MacSlow> ok
[15:15] <mzanetti> Saviq: DashContent tests fixed
[15:16] <MacSlow> greyback, too badk mterry has a bank-holiday today :)
[15:16] <Saviq> mzanetti, how's testShell on your end?
[15:16] <Saviq> 3 fail here
[15:16] <Saviq> on search indicator, so kinda make sense
[15:16] <Saviq> probably just for removal
[15:17] <mzanetti> Saviq: confirmed. fixing...
[15:17] <MacSlow> greyback, ah ok... that's a way to test it... works
[15:19] <mzanetti> ok. tst_Shell fixed
[15:21] <greyback> MacSlow: only danger is that you'll not be profiling the actual scenario, but a different scenario (i.e. not with USC)
[15:21] <greyback> MacSlow: I don't anticipate and major rendering differences between USC and mir_demo_server_shell, but just be cautious
[15:21] <MacSlow> greyback, yeah... just saw that the spinner doesn not show up in the process-list of top...
[15:22] <MacSlow> but only the mir_demo_server_shell
[15:22] <MacSlow> greyback, which defeats the purpose of my intention... to bring down (and verify) CPU-load of the spinner
[15:23] <greyback> MacSlow: indeed. It was step to determine if the fact you could not see the spinner was due to your spinner code, or USC. Conclusion: USC was the blame
[15:24] <greyback> so USC must have some criterion to decide that the spinner surface should be visible
[15:24] <MacSlow> greyback, next issue that turned up is that the locally compiled unity-system-compositor-spinner segfaults while the system-wide one does not... just trying to grab a bt
[15:30] <Saviq> mzanetti, one awesome thing about our xvfb tests is that you can actually run them in parallel
[15:31] <Saviq> mzanetti, so a full suite is down to some 3 minutes for me or so
[15:31] <greyback> MacSlow: yeah USC is designed to launch the spinner itself, and only show the surface from the spinner that it launched. Other surfaces are only brought to the front when USC is told to by some third party via IPC somehow
[15:32] <greyback> whatever the heck DMConnection is
[15:32] <mzanetti> nice! trying
[15:32] <MacSlow> greyback, got the crash sorted out
[15:33] <dednick> Saviq: bulkTransferMenu is a generic "menu with button"
[15:33] <mzanetti> Saviq: just passing -j?
[15:34] <dednick> Saviq: although might switch it out for the USC.StandardMenu now.
[15:35] <Saviq> mzanetti, yup
[15:35] <Saviq> dednick, right, but it was named bulkTransferMenu ;)
[15:36] <Saviq> dednick, your call there, just pointing out, you know better
[15:36] <dednick> Saviq: :) ah. you mean from the indicator
[15:37] <dednick> Saviq: you mean the "com.canonical.indicator.transfer-bulk-action"
[15:37] <dednick> or the usage of it
[15:38] <Saviq> dednick, the ": transferBulkAction" just after
[15:38] <Saviq> dednick, although I understand the extra-label extattr might not be common enough
[15:38] <Saviq> dednick, but maybe we shouldn't even introduce transfer-bulk-action
[15:38] <dednick> Saviq: right. if we come across another menu which has the same data, i can change it.
[15:39] <dednick> Saviq: yeah.
[15:39] <Saviq> dednick, and instead tell them to use action-with-button
[15:39] <dednick> charles: ^ too late to change?
[15:41] <dednick> charles: "com.canonical.indicator.button-section" maybe
[15:44] <charles> dednick, so just renaming
[15:44] <dednick> charles: ya
[15:45] <charles> dednick, so just renaming "com.canonical.indicator.transfer-bulk-action" as "com.canonical.indicator.button-section" because the former is overly specific? Yes I can update the MP with that
[15:45] <dednick> charles: yes. that's the idea.
[15:45] <Saviq> mzanetti, ok well, I lied, 7 minutes, but still
[15:46] <mzanetti> hehe :)
[15:47] <mzanetti> Saviq: while we're running tests: https://code.launchpad.net/~mzanetti/unity8/fix-testPreview-20ppgu/+merge/225690
[15:48] <Saviq> mzanetti, I don't understand why
[15:48] <mzanetti> Saviq: I think unit.dp(1) != units.gu(0.1)
[15:48] <mzanetti> which might introduce such slight differences
[15:49] <mzanetti> meaning: units.dp() and units.gu() does not grow linear
[15:50] <mzanetti> well, linear yes, but not in parallel
[15:51] <Saviq> right
[15:51] <mzanetti> not sure I explain it right... did you understand what I mean?
[15:51] <Saviq> mzanetti, btw, timing failure in PageHeader http://people.canonical.com/~msawicz/unity8/testPageHeader.xml
[15:51] <Saviq> mzanetti, I can get it to fail like this when ran in parallel and then some
[15:51] <Saviq> mzanetti, if obvious - fix, otherwise leave bt
[15:51] <Saviq> be
[15:52] <mzanetti> I've seen this here with a wait(10000)
[15:52] <mzanetti> it failed after 16 seconds with this warning
[15:52] <mzanetti> saw it for the first time today
[15:52] <Saviq> mzanetti, I meant <failure message="function returned unexpected result    Actual   (): false    Expected (): true" result="fail"/> rather
[15:52] <mzanetti> ack
[15:52] <Saviq> mzanetti, think it just exited the test because it got blocked for too long?
[15:52] <mzanetti> need to look
[15:59] <mzanetti> what the... this doesn't make any sense
[16:19] <mzanetti> Saviq: that test was totally useless. I've dropped it and replaced it with one that tests the IMA I added lately
[16:22] <mzanetti> greyback: ouch... seen the bugs om26er_ reported?
[16:23] <mzanetti> the osk one mainly
[16:24] <greyback> mzanetti: 'below unity' - I don't understand
[16:24] <mzanetti> greyback: below the indicators drop down most likely
[16:24] <mzanetti> pull down
[16:24] <greyback> mzanetti: he's right, OSK is under the indicators. Oops!
[16:24] <mzanetti> greyback: and hes kinda right with the dash schrolling performanche. although I don't see how that is related
[16:25] <greyback> mzanetti: we're switching rendering architectures, so it probably is related
[16:25] <greyback> om26er_: good catches all, please keep it coming!
[16:51]  * greyback EOW
[16:51] <greyback> o/
[17:17] <dednick> Saviq: with the maximumHeight/preferredHeight, was your comment related only to height, or width as well?