[09:06] <Cimi> easy way yo deploy ubuntu-ui-toolkit branch on the phone?
[09:14] <nic-doffay> MacSlow, I've added examples of the scrollable selectors to that branch, thanks for your feedback.
[09:17] <MacSlow> nic-doffay, np yw... thanks
[10:48] <nic-doffay> mzanetti, in the PageHeader.qml which exposed variable means the searchbar has been lowered (not searchFieldActive). Any idea? Been trying to find it.
[10:54] <mzanetti> nic-doffay: what do you mean with searchbar has been lowered?
[10:55] <nic-doffay> mzanetti, the white search bar?
[10:55] <nic-doffay> It gets lowered on to the screen?
[10:56] <nic-doffay> When you click on the area with the scope name.
[10:57] <mzanetti> nic-doffay: when I click on the area with the scope name, nothing happens. but I think you mean searchFieldActive
[11:00] <nic-doffay> mzanetti, searchField.activeFocus is true when you select the search to begin typing.
[11:00] <nic-doffay> I'd like to dismiss the entire search field if it has been lowered irrespective of whether the searchField is active or not...
[11:01] <mzanetti> nic-doffay: here searchField.active and the position are always related...
[11:01] <mzanetti> no focus in the search field means the search field is dismissed
[11:04] <nic-doffay> mzanetti, hmm ok.
[11:15] <dednick> Saviq: ping
[11:15] <Saviq> dednick, kinda pong
[11:16] <dednick> Saviq:  :) you know if there are any design requirements for the search swipe thingo?
[11:17] <dednick> Saviq: or, was there somebody specific in design who you talked to about it. Oren doesnt seem to be around.
[11:18] <Saviq> dednick, yeah, it just needs to be a progressive gesture - the search entry should slide in following your finger
[11:18] <dednick> Saviq: also about the design requirement for the icon overflow.
[11:18] <dednick> Saviq: yeah, bit slide from where? right, left, up down?
[11:18] <Saviq> dednick, yeah, I think we need to ask them what's the current plan
[11:18] <dednick> s/bit/but
[11:18] <Saviq> dednick, from up, like it does now
[11:18] <dednick> Saviq: slide to dismiss as well?
[11:19] <Saviq> dednick, so you swipe through SEARCH, and the search entry follows your finger
[11:19] <Saviq> dednick, no, I think only slide down, there's no handle that you could use to slide up
[11:19] <Saviq> dednick, there's one trick there, I'm afraid, you'll need to handle the page header as well, in case it's not on screen
[11:20] <Saviq> (+fully)
[11:20] <Saviq> dednick, sync with nic-doffay, too, he's been hacking on the page header quite a bit, so it might be good to wait for his stuff to land
[11:20] <dednick> Saviq: ok
[11:22] <dednick> too many swipes... next we're going to swipe to click a button.
[11:22] <Saviq> dednick, well, it's the same swipe as for the indicators...
[11:23] <dednick> oh right, i was thinking swipe left to right across "Search" rather than down like the indicators.
[11:25] <dednick> gona have to have very small fingers to get any meaningfull progress feedback :) that bar is very thin.
[11:30] <mzanetti> Saviq: do you have a minute to help me?
[11:38] <dednick> nic-doffay: ping
[11:43] <Saviq> mzanetti, didn't get to the office yet, in ~ an hour ok
[11:43] <Saviq> ?
[11:46] <mzanetti> sure
[12:58] <nic-doffay> dednick, what's up
[13:00] <Saviq> mzanetti, pong
[13:00] <mzanetti> Saviq: hi
[13:01] <Saviq> mzanetti, mornin'
[13:01] <mzanetti> so... I'm trying to mock the AccountService class in order to be able to test the update-on-config-change feature
[13:01] <mzanetti> but I'm failing to create a QDBusArgument
[13:01] <Saviq> nic-doffay, bug #1213034 fix committed?
[13:02] <Saviq> nic-doffay, btw, you should use `bzr commit --fixes lp:1213034` or similar
[13:02] <Saviq> nic-doffay, to mark the commit as fixing that issue, and then bug statuses will get handled automagically
[13:03] <mzanetti> Saviq: I pushed to lp:~mzanetti/unity8/launcher-update-on-as-change
[13:03] <Saviq> nic-doffay, also, I'd rather you have a separate branch for the bug fixes, instead of putting them in the filters branch (unless there'd be big conflicts somehow?)
[13:03] <mzanetti> Saviq: the code works, but the launchermodeltest doesn't...
[13:06] <Saviq> mzanetti, looking
[13:06] <mzanetti> Saviq: realized I forgot to push the mocks. pull again
[13:09] <mzanetti> Saviq: as you can see I did quite a lot of experimenting in tests/plugins/Unity/Launcher/mocks/AccountsServiceDBusAdaptor.cpp
[13:10] <mzanetti> Saviq: the place where this is read again is launcherbackend.cpp :215
[13:10] <mzanetti> and I just don't seem to be able to produce a QVariant that is the same as the one we get from DBus in the real accountsservice
[13:16] <dednick> nic-doffay: hey. you working on the page header?
[13:23] <mzanetti> mterry: hi... may I ask you to use "auto" a bit less please :)
[13:23] <mzanetti> mterry: it decreases readability quite a bit
[13:23] <Saviq> mzanetti, I can't build, even
[13:23] <mterry> mzanetti, oh, I thought people liked auto
[13:24] <mzanetti> Saviq: yeah... let me comment away that line
[13:24] <Saviq> mterry, mzanetti, auto has advantages and disadvantages - easier to write, harder to read
[13:25] <Saviq> and QtCreator's code completion doesn't grok autos
[13:25] <mzanetti> yeah... /me loves easy reading
[13:26] <mterry> Saviq, mzanetti: OK, I am actually not a huge fan of auto myself, I just thought that it was in vogue
[13:26] <mzanetti> mterry: no... so far I think you're the only one using it in our code
[13:26] <nic-doffay> dednick, yeah a lot.
[13:26] <mzanetti> mterry: which is not a problem. but I see you're _only_ using auto in .cpp files
[13:26] <nic-doffay> I'd say it's doubled in size almost to trunk.
[13:26] <dednick> nic-doffay: any idea when it's going to be ready?
[13:27] <mterry> mzanetti, I must have heard c++11 and mentally thought all new stuff was all preferred  :)
[13:27] <mzanetti> no way
[13:27] <mterry> mzanetti, what do you mean about using it only in cpp files?
[13:27] <nic-doffay> dednick, not sure at the moment. There's an autopilot issues which prevent a branch it's reliant on from landing.
[13:28] <mzanetti> mterry: for example in AccountsService.cpp there isn't a single variable with its real type written :)
[13:28] <nic-doffay> dednick, I'd say next week.
[13:28] <mzanetti> mterry: which caused me browse header files quite a lot today
[13:28] <mterry> mzanetti, yar, like I said, I'm not an advocate of them.  I just mistakenly got the impression they were preferred
[13:28] <mzanetti> mterry: ok, cool
[13:29] <mzanetti> Saviq: pushed btw...
[13:29] <Saviq> mzanetti, k
[13:30] <mzanetti> Saviq: but I bet you'll end up uncommenting that line soon :D
[13:31] <nic-doffay> dednick, just confirmed, it's still broken. I'm off tomorrow so I can def say next week.
[13:32] <nic-doffay> Is it a big problem?
[13:32] <Saviq> Cimi, standup?
[13:32] <Saviq> dednick, standup?
[13:40] <dednick> nic-doffay: nope. just didnt want to start working too much on the indicator search bar if it's changing alot
[13:40] <dednick> nic-doffay: so i'll just hold out
[13:44] <nic-doffay> dednick, cool
[13:44] <mzanetti> Saviq: tests/plugins/Unity/Launcher/mocks/AccountsServiceDBusAdaptor.cpp
[13:44] <mzanetti> Saviq: the place where this is read again is launcherbackend.cpp :215
[13:44] <nic-doffay> dednick, oren's in a meeting but I'll let him know you want to chat when he's out.
[13:45] <mzanetti> Saviq: and the real thing that comes from DBus is in plugins/AccountsSErvice/AccountssServiceDBusAdaptor.cpp
[13:45] <dednick> nic-doffay: dont worry, i'll just send him an email
[14:04] <Saviq> mzanetti, TBH it feels to me like we're trying to mock this on too low a level...
[14:19] <jbicha> could https://code.launchpad.net/~jbicha/unity/recommend-telepathy-indicator/+merge/186392 be reviewed?
[14:22] <jbicha> otherwise as a workaround, I could have mcp-account-manager-uoa recommend it but that's not the best place for that relationship
[14:23] <Saviq> mzanetti, why do we even need to mock around the whole DBus adapter, and not just deal with QVariants?
[14:24] <cwayne> mterry: heyo
[14:24] <mterry> cwayne, hi
[14:24] <mzanetti> Saviq: yeah... I tried that too...
[14:24] <cwayne> mterry: im still seeing the old hardcoded greeter.. do you know if your branch ever landed in the image?
[14:24] <mzanetti> Saviq: thing is, we couldn't keep the AccountsServiceDBusAdaptor as generic as it is right now
[14:25] <mterry> cwayne, (A) my branch landing will still give you the old hardcoded greeter
[14:25] <mterry> cwayne, (B) my branch didn't land, it got delayed, potentially until mir does land
[14:25] <mterry> cwayne, sergiusens saw a performance hit with my branch and I didn't have time to track it down
[14:25] <cwayne> mterry: i thought the whole point of that branch was that it fallsback to the background gsetting?
[14:25] <Saviq> mzanetti, yeah, but then that means the AccountsService plugin is tied to DBus completely
[14:25] <mterry> cwayne, oh oh oh
[14:26] <mterry> cwayne, you're talking about the background branch
[14:26] <mterry> cwayne, that landed
[14:26] <Saviq> mzanetti, which is not great
[14:26] <mterry> cwayne, but maybe didn't get released to saucy yet, let me check
[14:26] <mzanetti> Saviq: yeah... I started to refactor that too. but reverted at some point because it seemed to get worse
[14:27] <mterry> cwayne, nope.  It landed on 9/16, right after that day's unity8 release, and we haven't had another since
[14:28] <cwayne> mterry: ah, alright so it's not in the image then... any idea when we'd have another release?
[14:28] <mzanetti> Saviq: but I guess I'll just introduce  additional methods in the AccountsServiceDBusAdaptor called [set]launcherItems() which handle that dbus stuff in there
[14:29] <mzanetti> mterry: what is your opinion? you wrote that stuff
[14:29] <mterry> cwayne, no, I'm not sure what's holding up the release.  we have started doing more measured releases, with this new landing spreadsheet and stuff
[14:29]  * mterry reads scrollback
[14:29] <mzanetti> mterry: so I'm in the need to mock the accountsservice stuff
[14:29] <mzanetti> mterry: but I'm just not able to fake the launcher items struct
[14:30] <Saviq> mzanetti, actually
[14:30] <mzanetti> mterry: and we kinda think the dbusadaptor shouldn't expose QDBusArguments to the outside
[14:30] <ChrisTownsend> jbicha: I was asking my team about that MP and we need someone from distro to ack as well.
[14:30] <Saviq> mzanetti, shouldn't the adapter cough up "flat" QVariants already?
[14:30] <Saviq> mzanetti, instead of going qdbus_cast in LauncherBackend?
[14:30] <mterry> mzanetti, I didn't think the dbusadaptor was exposed at all.  I thought we only exposed the Qml-friendly AccountsService class
[14:31] <Saviq> mzanetti, at that point mocking the adapter would become trivial
[14:31] <mzanetti> mterry: no... you created it for the launcher :)
[14:31] <mterry> mzanetti, oh in the launcherbackend code it is
[14:31] <mzanetti> mterry: and afterwards you only wrapped the QML stuff around
[14:31] <mzanetti> Saviq: yes... I agree
[14:32] <mzanetti> Saviq: but that means the accountsservicedbusadaptor needs to know about the launcher's config format.
[14:32] <mzanetti> which somewhat sucks too
[14:33] <Saviq> mzanetti, there's no way of QDBusArgument > "flat" QVariant without knowing the format?
[14:33] <mterry> mzanetti, Saviq: I'm not opposed to that class being smarter about what it exposes.  But I was envisioning all smarts to belong in the AccountsService class.  Can we push the smarts there and change launcherbackend to use the "wrapper" class?
[14:33] <Saviq> mzanetti, there's at least currentSignature on QDBusArgument that suggests it has all the knowledge it needs to flatten itself
[14:33] <mzanetti> Saviq: that's the thing... you need to know what's in the QDbusArgument to be able to unmarshall it
[14:33] <Saviq> mterry, yeah, that was another of my points
[14:33] <Saviq> mzanetti, yeah, but it knows what it is, does it not?
[14:34] <Saviq> mzanetti, it has currentSignature() which holds what it is
[14:34] <Saviq> aa{sv} in our case
[14:34] <mzanetti> ah right...
[14:34] <Saviq> so it'd be braindead that it could not unmarshal itself
[14:34] <Saviq> mzanetti, but yeah, what mterry said
[14:34] <mzanetti> I agree with that too
[14:35] <Saviq> mzanetti, the AccountService is where this should probably happen
[14:35] <Saviq> mzanetti, and that's what we should mock
[14:35] <mzanetti> ack
[14:35] <mterry> I think we have a mock for that already
[14:35] <mzanetti> Saviq: but it still annoyes me that I didn't manage to fake that damn QDBusArgument
[14:36] <mzanetti> mterry: yep, we do
[14:36] <Saviq> mzanetti, heh ;)
[14:36] <mzanetti> ok then... I know how to proceed
[14:36] <mzanetti> thanks guys
[15:13] <mzanetti> Saviq: are we ok with scripts for manual testing in /tests/tools/ ?
[15:13] <Saviq> mzanetti, I've nothing against that
[15:13] <Saviq> mzanetti, but id -u
[15:13] <Saviq> mzanetti, not id -g
[15:13] <mzanetti> right...
[15:13] <mzanetti> thanks :)
[15:52] <mzanetti> does ctrl+/ in QtCreator work for you? It never worked in the Qt5 edition for me and I assumed its a temporary bug. but it doesn't seem to go away for me.
[15:53] <Saviq> mzanetti, seems to work here
[15:53] <mzanetti> hmm... interesting... any hints on what it could be? works fine with the Qt4 qtcreator
[15:54] <mzanetti> I tried remapping that to something else. still nothing
[15:54] <mzanetti> all the other combos seem to work
[16:10] <Cimi> mzanetti, do we have any tests for the bottom bar?
[16:10] <Cimi> I finished my branch...
[16:10] <Cimi> https://code.launchpad.net/~cimi/ubuntu-ui-toolkit/panel_export-position
[16:10] <Cimi> https://code.launchpad.net/~cimi/unity8/unity8.hud-2_hint-reveal-commit
[16:11] <mzanetti> Cimi: yes, we have. I think you wrote them
[16:12] <Cimi> nope
[16:19] <Cimi> mzanetti, I don't think there are any
[16:20] <mzanetti> Cimi: is the bottomBar something different than the DashBar?
[16:21] <Cimi> mzanetti, yes
[16:21] <mzanetti> Cimi: ah ok... so we probably don't have any
[16:21] <Cimi> mzanetti, it's the hud revealer
[16:21] <mzanetti> ah
[16:21] <mzanetti> Cimi: we have some of them in the hud tests
[16:21] <mzanetti> Cimi: and autopilot tests for that too
[16:21] <Cimi> but we changed the behaviour
[16:21] <Cimi> I am changing the behaviour
[16:21] <Cimi> so I need to test
[16:22] <mzanetti> Cimi: just adding the visual behavior or the stuff that gets communicated?
[16:22] <mzanetti> -adding +changing
[16:22] <Cimi> mzanetti, the hud stays on screen
[16:22] <Cimi> hud button
[16:23] <mzanetti> Cimi: yeah... we have autopilot tests for that
[16:23] <Cimi> so I need to test that
[16:23] <Cimi> and might need to test the dismiss as well
[16:23] <mzanetti> which probably should be qmltests actually, but that's what it is right now
[16:23] <mzanetti> Cimi: yeah... the test is quite precise in that
[16:23] <Cimi> I can do qmltest maybe
[16:23] <mzanetti> Cimi: I'm sure autopilot will fail with your branch
[16:24] <mzanetti> Cimi: if you'd convert that to be a qmltest, that'd be great
[16:24] <Cimi> mzanetti, where's the autopilot test? how can I test it?
[16:24] <mzanetti> Cimi: ./build
[16:24] <mzanetti> cd builddir
[16:24] <mzanetti> make install
[16:25] <mzanetti> cd ../tests/autopilot
[16:25] <mzanetti> export `dbus-launch`
[16:25] <mzanetti> autopilot run unity8
[16:25] <mzanetti> Cimi: the export thingie needs to be done only once per shell you want to run the tests
[16:26] <mzanetti> Cimi: do "autopilot list unity8" to see all tests
[16:26] <mzanetti> Cimi: and give the full path at autopilot run to run only a single test
[19:32] <mterry> mzanetti, speaking of AccountsService in unity8, can I get a review for https://code.launchpad.net/~mterry/unity8/statsWelcomeScreen/+merge/184153 ?
[19:32] <mterry> mzanetti, ick
[19:32] <mterry> mzanetti, let me update it's naming scheme to fit proper style
[19:33] <mzanetti> mterry: ah right... I already started on this one and was interrupted. sorry. will do tomorrow
[19:33] <mterry> mzanetti, no worries.  I'll have it fit and shiny tomorrow