11:30 <popey> #startmeeting dekko
11:30 <meetingology> Meeting started Thu Apr 21 11:30:51 2016 UTC.  The chair is popey. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
11:30 <meetingology> 
11:30 <meetingology> Available commands: action commands idea info link nick
11:30 <popey> Hey hey everyone!
11:30 <faenil> o/
11:31 <DanChapman> Afternoon all!
11:31 <faenil> ho ho ho!
11:32 <popey> Hows it going?
11:32 <DanChapman> Going good thanks. Yourself?
11:32 <popey> First item of news, latest dekko was passed by QA, so time to upload!  😃
11:32 <popey> great!
11:32 <DanChapman> I'll do that straight after the meeting :-)
11:33 <faenil> \o/
11:33 <faenil> last one before the big move?
11:35 <DanChapman> Yep that's the last one. Already fixed a load of outstanding bugs with the qmf branch so it makes sense to move asap https://bugs.launchpad.net/dekko/+milestone/1.0-beta
11:35 <faenil> cool
11:36 <popey> sweet!
11:36 <popey> so another QA submission coming soon? :)
11:39 <DanChapman> Sorry postman was at the door. So yes not too long now :-) I haven't started figuring out a migration strategy yet though. That might take a little while to get right
11:40 <popey> ah, of course.
11:40 <popey> we need to test that quite heavily, needs to be robust
11:41 <popey> some manual test cases would help too for QA
11:41 <popey> but we can manually test on our devices with different accounts, right?
11:42 <DanChapman> Do you think it would be wise to also try and preserve the local cache? That's where it would get really tricky
11:42 <faenil> naaah
11:42 <faenil> I mean, ideally, sure
11:43 <faenil> but if it's 10x the effort of the whole move alone, then I think it's totally fine to tell the user and discard the cache
11:43 <faenil> (as long as he can first commit cached operations, etc)
11:43 <DanChapman> Yes I will have manual tests ready for that. But yes you will be able to test with different accounts. It's not ready for that right now though. You still have to create accounts from the command line.
11:45 <popey> ok
11:45 <popey> yeah, I would lose the cache
11:45 <popey> not worth it
11:45 <DanChapman> faenil, yeah it would be a considerable effort as it's turning trojita's data blobs into something qmf can parse and store. Which would mean shipping a trojita library to make that possible. Which isn't ideal.
11:46 <DanChapman> Ok great, so just inform the user at on first launch of the new version.
11:48 <popey> Ok.
11:49 <faenil> DanChapman: yeah, and ideally give him the chance to sync with the server (if he has any pending operation that hasn't complete yet)
11:49 <popey> I think we shouldn't be too detailed in that information, but just "you need to do this because reasons"
11:49 <faenil> so that there's no unexpected outcome
11:49 <popey> yeah, force sync first?
11:49 <faenil> (I'm referring to the undo actions you talked about a couple of weeks ago)
11:50 <faenil> (or anything like that which might be waiting for a connection to complete)
11:50 <DanChapman> faenil, ah.. that's qmf specific. That's not possible with trojita as all actions are performed directly with the server and it's synced back to the client on the response.
11:51 <DanChapman> So there shouldn't be issues there
11:51 <faenil> DanChapman: cool, one less problem :D
11:53 <popey> ace
11:53 <popey> Ok, anything else?
11:53 <DanChapman> I'll put this towards the top of my todo list then. And we can get some testing done on it then
11:54 <DanChapman> Not from my end. Got some questions for James but they can wait until he's feeling better :-)
11:54 <faenil> col
11:55 <faenil> cool
11:57 <popey> sweet
11:57 <popey> okay, well congratulations on the new release DanChapman !
11:58 <popey> Look forward to seeing it later. Cheers guys.
11:58 <DanChapman> thanks :-)
11:58 <popey> #endmeeting