13:01 <popey> #startmeeting Trojita meeting
13:01 <meetingology> Meeting started Fri May  9 13:01:24 2014 UTC.  The chair is popey. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
13:01 <meetingology> 
13:01 <meetingology> Available commands: action commands idea info link nick
13:01 <popey> Hows it going guys?
13:01 <DanChapman> popey howdy amigo
13:01 <popey> mhall119: have we made any progress on building trojita in qtc?
13:02 <DanChapman> going good thanks. You?
13:02 <DanChapman> hey jkt
13:02 <popey> Great! Nice short week.
13:02 <popey> hi jkt
13:02 <popey> handy trello link:- https://trello.com/b/73rFyG6n/ubuntu-email-client
13:02 <DanChapman> yes been pretty busy this week so havn't got as much done as i hoped
13:03 <popey> I see a comment from you about building in qtc.. but it fails to run?
13:04 <DanChapman> yes it now builds fine with the armhf kit but when it complains about paths or something. SO i'm manually creating a click package and pushing it atm :-( i saw mhall119was on to it the other day, was hoping to grab him about it today
13:05 <DanChapman> s/when/then
13:05 <popey> ok.
13:06 * DanChapman just updating trello
13:07 <popey> I'm getting an error when trying to sign into email on my desktop
13:07 <popey> file:///usr/share/trojita-ubu/qml/trojita/ImapSettings.qml:72: Error: Cannot assign to non-existent property "password"
13:08 <popey> using 0.4.1+1-0~4201~ubuntu14.04.1 from the ppa
13:08 <jkt> popey: that's because DanChapman still needs to switch to using the PasswordWatcher, which was done in order to support the password plugins
13:08 <DanChapman> popey yes it needs fixing if you don't put a password in the settings and wait for the dialog
13:08 <jkt> also, that version string looks weird; didn't we agree to use output of `git describe` in there?
13:08 <popey> ah!
13:09 <DanChapman> popey is the touch-coreapps ppa?
13:09 <popey> yes
13:10 <jkt> DanChapman: perhaps adding a --version flag might be a good idea
13:11 <DanChapman> jkt yes we will need to poke dholbach about that, he set it up i believe
13:12 * popey summons dholbach
13:12 <popey> My level 2 spell of summoning isn't very good.
13:13 <DanChapman> boren did you try getting the Math.max of say the scrollHeight, clientHeight, offsetHeight on both body and documentElement?
13:13 <boren> DanChapman: still trying
13:13 <jkt> and have you guys asked the qtwebkit folks on how to solve this?
13:14 <jkt> not that they are terribly responsive, but they might help anyway
13:14 <DanChapman> jkt I have tried a few times and I had no respone
13:15 <jkt> DanChapman: hte ML works sometimes
13:15 <jkt> but I'm still waitig for an unrelated response in there
13:15 <jkt> DanChapman: which channel did you use?
13:15 <DanChapman> #qtwebkit
13:16 <DanChapman> Ok i'll ping one off to the ML
13:17 <dholbach> hiya
13:17 <DanChapman> o/ dholbach
13:17 <dholbach> hey DanChapman
13:17 <boren> o/
13:17 <dholbach> hey boren
13:18 <popey> dholbach: there was a question about the build filename...
13:18 <popey> Installed: 0.4.1+1-0~4201~ubuntu14.04.1
13:18 <jkt> popey: that was an apparent upkeep
13:18 <dholbach> great work everyone on trojita - it's really nice to have the daily build installed and see features coming in :)
13:18 <dholbach> ok
13:18 <popey> 14:08:27 < jkt> also, that version string looks weird; didn't we agree to use output of `git describe` in there?
13:18 <popey> 14:10:26 < jkt> DanChapman: perhaps adding a --version flag might be a good idea
13:18 <popey> dholbach: ^
13:18 <popey> figured that was one for you.
13:19 <dholbach> I never knew that "git describe" was meant to be used anywhere
13:19 <dholbach> Jono just told me "please get a daily build in a ppa set up"
13:19 <popey> DanChapman: i see a few items in the "waiting for review" lane.. are they actually waiting on review, or can they be moved/updated?
13:19 <jkt> dholbach: the things is, I have no idea how to translate 0.4.1+1-0~4201~ubuntu14.04.1 to soemthing which would make any sense to me
13:19 <dholbach> so I figure I used the last upstream version "+<something"
13:19 <dholbach> ok
13:20 <dholbach> as I understood it 0.4.1 was the last released upstream version (as a tarball), right?
13:20 <jkt> dholbach: yes
13:20 <dholbach> ok
13:20 <jkt> dholbach: that said, `git describe` would give that to you for free :)
13:20 <jkt> no need to hardcode anything
13:20 <DanChapman> popey: Most are in progress for changes from the review, they need moving back one
13:20 <popey> DanChapman: Ok, can you update please?
13:20 <dholbach> jkt, well, Debian packaging normally requires a released tarball
13:20 <dholbach> ... or one you generate yourself
13:21 <dholbach> and then the version number is generated from the newest entry in debian/changelog
13:21 <dholbach> also, "git describe" won't work in daily builds, where we use bzr hosted on Launchpad
13:21 <boren> DanChapan: scrollHeight, clientHeight, offsetHeight all produce a larger height.
13:21 <dholbach> 4201 is the bzr revision number of https://code.launchpad.net/~ubuntu-touch-coreapps-drivers/trojita/git-kde
13:22 <popey> DanChapman: do you have answers for everything you needed from Andrea?
13:23 <jkt> dholbach: that's rather suboptimal :(, the build system is git-aware and does stuff like setting the trojita version (which is in turn reported through the ID command to the IMAP servers) to an appropriate string
13:23 <DanChapman> popey so far yes :-) I believe gcolurra has asked for his help improving the design of the composer view
13:23 <dholbach> jkt, I don't know how to fix that
13:24 <jkt> dholbach: can you hack the git->bzr import process so that it runs `git describe > src/trojita-version` at least?
13:25 <dholbach> jkt, no, that's automatic in Launchpad - I just gave it a git clone url
13:25 <popey> hi gcollura
13:25 <DanChapman> o/ gcollura
13:25 <gcollura> hi everybody
13:25 <boren> o/
13:25 <dholbach> (https://code.launchpad.net/~ubuntu-touch-coreapps-drivers/trojita/git-kde has details of the import)
13:26 <popey> dholbach: https://help.launchpad.net/Packaging/SourceBuilds/Recipes suggests we have {git-commit}, {git-commit:<branch id>}
13:27 <dholbach> aha
13:27 <jkt> "Replaced with the last 7 characters of the git commit that was built"
13:27 <dholbach> well, I can try that out
13:27 <jkt> if htat is correct, it's exactly the wrong way
13:28 <dholbach> jkt, could you explain what you need and what you need it for?
13:28 <dholbach> I was just pulled into this meeting and may lack some background
13:29 <jkt> dholbach: sure
13:29 <jkt> when someone reports a prbolem, or just refer to Trojita like "that version 4201 from latest ppa", I have no idea what version they're talking about
13:29 <jkt> as an upstream, I do not care about what you use for your builds
13:30 <jkt> what I want is some way of ensuring that Trojita behaves the same regardless whether it was built from our git, on the OBS' infrastructure, by a Fedora packager or through Ubuntu's Launchpad
13:31 <dholbach> right, I on the other hand just used what was most convenient to me - so if I understand you correctly, you want to translate 4201 to a git commit?
13:31 <jkt> one of these bits is making sure that the version which Trojita reports about itself is correct
13:31 <dholbach> so bzr r4201 = 214156f687b4470450f838c631201cb31ae84be7 or something
13:32 <jkt> I'm most definitely not blaming you, sorry if it sounded like that way
13:32 <dholbach> no no, that's fine - I'm just trying to figure it out
13:32 <jkt> right now, my desktop build identifies itself as v0.4.1-203-g214156f
13:32 <dholbach> what is the 203?
13:32 <jkt> what I need is a way of converting this arbitrary identifier to something I could work on, and recognizable by git
13:32 <jkt> v0.4.1 == last tag name
13:33 <jkt> 203 == number of commits since that time
13:33 <jkt> g == "git prefix"
13:33 <jkt> 214156f == start of the commit sha1
13:33 <dholbach> hum
13:33 <jkt> it's just the output of git describe
13:34 <dholbach> I wonder why there's a difference in preferences regarding "end of commit sha1" and "start of commit sha1"
13:34 <jkt> I can uise `git show 214156f`
13:34 <dholbach> ie: "git describe" and the daily build thing
13:34 <jkt> but I cannot use `git show ae84be7` which is the end of the same hash
13:34 <jkt> git just doesn't recognize that
13:35 <dholbach> I see
13:35 <dholbach> I'll talk to the LP devs and see what can be done to add the end of the commit sha1 to the version number
13:35 <dholbach> I have no idea, but can at least just ask
13:35 <popey> thanks dholbach
13:36 <popey> gcollura: how are things on the SMTP backend?
13:36 <jkt> dholbach: "beginning", please :)
13:37 <dholbach> err, sorry
13:37 <dholbach> yes
13:37 <jkt> dholbach: and there are two places, the "version of package" and the string seen by cmake
13:38 <gcollura> popey, I've asked for a review, next thing on the list is two improve the design of the compose page and add support for the attachments
13:39 <dholbach> jkt, how is the "string seen by cmake" generated or how is it visible later on?
13:40 <DanChapman> gcollura: yours will need a minor change soon for the SubmissionMethod enum. it's only minor though :-)
13:40 * DanChapman officially dispises trying to get enums to work in qml
13:40 <gcollura> popey, I've contacted Andrea del Sarto for its design
13:40 <popey> gcollura: when was that?
13:41 <gcollura> some days ago, he only sent me a mockup for the address entry
13:43 <mhall119> hey popey, sorry I'm late
13:43 <popey> np mhall119, hiya
13:43 <popey> gcollura: ok, I'll chase that up
13:43 <jkt> dholbach: as I said, cmake consults the output of `git describe`.If that is not available it will use the contents of src/trojita-version
13:44 <jkt> dholbach: I can merge a patch which adds a cmake build option or whatever, but I would not want to write it
13:44 <mhall119> jkt: cmake already does that somewhere?
13:44 <jkt> mhall119: yes
13:44 <mhall119> oh, cool, I didn'tknow that
13:44 <jkt> "cmake" as in "trojita's cmake setup"
13:44 <mhall119> jkt: right, we can do the same for the ubuntu UI builds, and just put the string into the Click manifest.json
13:44 <jkt> and again, this is about what Trojita itself knows and reports, not some external stuff like a package name
13:45 <dholbach> jkt, ok, it looks like https://help.launchpad.net/Packaging/SourceBuilds/Recipes had a documentation issue
13:45 <jkt> mhall119: that manifest.json should probably be produced by cmake at build time, then
13:45 <dholbach> jkt, it should in fact be the first 7 chars of the commit sha
13:45 <popey> \o/
13:46 <jkt> dholbach: that's good news
13:46 <dholbach> jkt, so I'll change it in build recipe and with the next upstream commit we should have a fixed version string
13:46 <mhall119> jkt: yes it should
13:46 <popey> Great, thanks dholbach
13:46 <dholbach> (and some waiting time in between ;-))
13:46 <dholbach> popey, thanks for finding it in the docs - I was unaware of {git-commit}
13:46 <jkt> dholbach: I'm OK with losing that 203, and using something like "v0.4.1-ubu-214156f"
13:46 <popey> dholbach: np
13:47 <jkt> dholbach: could you do that?
13:47 <jkt> (for the trojita's inner version)
13:47 <dholbach> jkt, inner version meaning?
13:48 <jkt> dholbach: not the package version, but the stuff which cmake knowns and puts into the binary
13:49 <dholbach> jkt, so I should write something like "v0.4.1-ubu-214156f" into src/trojita-version at the beginning of the build?
13:49 <jkt> dholbach: yes, that will work
13:49 <jkt> dholbach: make sure you do that prior to running cmake
13:49 <dholbach> it'll take me some time to figure it out
13:49 <popey> gcollura: ok, mailed andrea and cc'ed you
13:49 <jkt> dholbach: thanks
13:49 <dholbach> but yeah, I'll look into it
13:49 <gcollura> yep I've seen the email
13:50 <popey> Ok, look at the time fly. Are there any other blockers or issues we need to discuss?
13:50 <mhall119> was the sorting/old email thing figured out DanChapman ?
13:51 <jkt> mhall119: it's unfortunately a bug in Trojita
13:51 <jkt> a long-term fix depends on certain refactoring which is going to be rather intrusive, so it will take time
13:52 <mhall119> do we have a stop-gap fix available?
13:52 <mhall119> or can we revert back to wht was working before?
13:52 <jkt> maybe calling the setSearchBlahBlahBlah in response to rowsInserted
13:53 <mhall119> DanChapman: you're the one who's been working on this, does ^^ make sense to you?
13:54 <DanChapman> mhall119: yes it does
13:54 <mhall119> great
13:55 <boren> DanChapman: Also, there is a modelReset signal that need be linked to search.
13:57 <DanChapman> boren, ok thanks
13:58 <DanChapman> mhall119: have you got trojita to actually run on device yet from QtC?
13:59 <DanChapman> mhall119: I just get "Failed to start program. Path or permissions wrong?"
13:59 <mhall119> DanChapman: not from QtC no, bzoltan gave me some CMake variables to set that should make that happen
13:59 <mhall119> DanChapman: oh, I get something like "No executable specified"
13:59 <popey> mhall119: does the wiki need updating with build instructions?
13:59 <mhall119> haven't had a path or permission error
14:00 <popey> oh, it doesnt work?
14:00 <mhall119> popey: it "just works" for desktop
14:00 <mhall119> it's cross-compiling and running on device or emulator that's still buggy
14:00 <popey> yeah, desktop is least of my worries
14:00 <mhall119> I'm working with the SDK team to get that fixed
14:00 <popey> given we have dholbach on the ppa case
14:01 <popey> ok.
14:02 <mhall119> popey: the tl;dr is that our QtC plugin is built around Click projects, and since Trojita it's a Click project, some of hte assumptions that our plugin makes are wrong for it
14:02 <mhall119> zoltan and I are working to fix it in a way that will allow other existing projects to "just work" with an Ubuntu UI + Click package port
14:02 <popey> yeah, that was always going to be the way bringing in an external project
14:03 <popey> Any expectations set on when we might be able to use qtc for trojita building?
14:03 <DanChapman> mhall119: ok cool
14:04 <mhall119> popey: when I have time (damn summit work) I'll try to get a proper ETA for you
14:04 <popey> Thanks.
14:04 <popey> Ok, anything else we need to discuss? We've been going an hour already!
14:04 <popey> *hint*
14:04 <DanChapman> hah :-D nope i'm good
14:05 <DanChapman> just try and get some more of the trello board cleared
14:05 <mhall119> I'm good too
14:06 <boren> good
14:06 <popey> Ok. Lets wrap up then. Thanks, and have a great weekend all!
14:06 <popey> #endmeeting