=== _salem is now known as salem_ | ||
=== salem_ is now known as _salem | ||
=== chihchun_afk is now known as chihchun | ||
=== chihchun is now known as chihchun_afk | ||
renatu | mhall119, hi could you review this mr? https://code.launchpad.net/~renatofilho/ubuntu-docviewer-app/no-file-hint/+merge/316865 | 11:31 |
---|---|---|
renatu | mhall119, and this one? https://code.launchpad.net/~renatofilho/ubuntu-calculator-app/snap-mir/+merge/316889 | 11:32 |
=== _salem is now known as salem_ | ||
* Mirv updates qt58 cloud part description a bit now that Qt 5.8.0 is actually released | 13:52 | |
ltinkl | kalikiana, so to continue, I branched trunk and submitted against staging, is that correct? | 13:55 |
kalikiana | ltinkl: Aha. I was about to ask that. You need to branch lp:ubuntu-ui-toolkit/staging - not trunk | 13:56 |
kalikiana | Otherwise your diff includes all the changes since the latest stable trunk | 13:56 |
ltinkl | kalikiana, that explains it :) ok, will resubmit | 13:56 |
ltinkl | kalikiana, now it looks better: https://code.launchpad.net/~lukas-kde/ubuntu-ui-toolkit/asyncDbusClipboard/+merge/316965 | 14:09 |
kalikiana | Thanks! | 14:16 |
kalikiana | artmello: ^^ | 14:19 |
kalikiana | ltinkl: Did you consider how this could be unit-tested? | 14:21 |
ltinkl | kalikiana, well hard to do imo - as it's a race condition | 14:23 |
ltinkl | kalikiana, and the MockContentService used in the test won't cut it | 14:24 |
kalikiana | ltinkl: How about making the mock service sleep when it initializes or responds? And then one could see if the test slows down when initializing the QML | 14:26 |
kalikiana | Simplifying the problem, the service shouldn't affect creating the component, right? | 14:26 |
ltinkl | kalikiana, but it does, since it's a singleton and it's used in a Connections from QML | 14:28 |
kalikiana | ltinkl: Well, I'm trying to narrow down what's affecting startup of Unity in the context of the bug report | 14:29 |
kalikiana | That's what you're trying to fix | 14:29 |
kalikiana | It can affect startup so long as initialization is synchronous | 14:29 |
kalikiana | Right? | 14:30 |
ltinkl | kalikiana, the problem is: unity8/qtmir is starting up, we have a TextField somewhere in our GUI which contains the call to UCContentHub, calls content-hub over DBUS, but at the same time content-hub is asking qtmir for authorization | 14:30 |
ltinkl | kalikiana, I assume content-hub somehow also queries the clipboard, but the implementation of QPlatformClipboard is in qtmir :) so we get a nice circular dependency and DBUS gets stuck, until one of the calls times out | 14:33 |
kalikiana | Hmmm | 14:34 |
ltinkl | kalikiana, the tricky (blocking) part is creating QDBusInterface in your CTOR, you don't even have to make any calls | 14:35 |
kalikiana | ltinkl: Maybe the mock service should actually just time out the first call? To mimic what's happening at login, regardless of the actual reason | 14:38 |
* ltinkl thinking if it would really prove anything | 14:38 | |
kalikiana | ltinkl: Do you know if the real content hub eventually returns anything useful right now? | 14:41 |
kalikiana | Or does it just timeout with an error | 14:41 |
ltinkl | kalikiana, heh ye, that's another prob - your implementation uses calls that (at least here) do not even exist on dbus | 14:41 |
kalikiana | ltinkl: How do you mean don't exist? | 14:42 |
ltinkl | kalikiana, if I introspect my content-hub, it e.g. doesn't have "RequestPasteByAppId" | 14:42 |
ltinkl | kalikiana, or the "PasteSelected" signal | 14:43 |
kalikiana | Interesting. | 14:43 |
ltinkl | kalikiana, PasteboardChanged and PasteFormats is there | 14:44 |
kalikiana | Maybe let's wait til artmello could provide some input - I don't actually know that much about contenthub | 14:44 |
ltinkl | kalikiana, yeah, me neither | 14:46 |
=== LiQiong_Lee1 is now known as LiQiong_Lee | ||
kalikiana | ltinkl: Do you mind if I copy your explanations into the MR description? | 14:48 |
ltinkl | kalikiana, sure. go ahead | 14:48 |
kalikiana | Thanks | 14:50 |
ltinkl | kalikiana, hmm, the popover is also wrong: if I select some text and do Ctrl+C, it gets copied to clipboard which I can paste using Ctrl+V; but the Paste icon is disabled (because content-hub doesn't have it) | 14:53 |
artmello | ltinkl, kalikiana: hey, sorry was afk | 14:55 |
artmello | ltinkl, kalikiana: reading. content-hub still do not return anything because silo 2179 still did not land | 14:56 |
artmello | ltinkl, kalikiana: we hoped it would land together with uitk changes but we got delayed after mir update. right now we are blocked with bug #1661704 | 14:57 |
ubot5 | bug 1661704 in mir (Ubuntu) "mir_window_request_persistent_id_sync seg faults when called twice" [Critical,Triaged] https://launchpad.net/bugs/1661704 | 14:57 |
=== LiQiong_Lee1 is now known as LiQiong_Lee | ||
artmello | ltinkl, kalikiana: so, yeah, none of those dbus calls are available on content-hub so far | 14:58 |
ltinkl | artmello, I see, thx; please have a look at my branch, it blocks u8 startup, quite severe | 14:59 |
artmello | ltinkl: sure, checking it right now | 15:00 |
kalikiana | ltinkl: You've got some unit test failures | 16:10 |
ltinkl | kalikiana, yup, looking into it | 16:11 |
=== JanC_ is now known as JanC | ||
ltinkl | kalikiana, do you have an easy way to run a single test (instead of the "make test" beast)? | 16:31 |
kalikiana | ltinkl: Sure. Change into the folder of the test and run 'make check' (assuming you're using Make). See also the file README, section Tests. | 17:13 |
=== LiQiong_Lee1 is now known as LiQiong_Lee | ||
=== salem_ is now known as _salem | ||
=== _salem is now known as salem_ | ||
=== salem_ is now known as _salem | ||
=== _salem is now known as salem_ | ||
=== salem_ is now known as _salem |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!