=== chihchun_afk is now known as chihchun | ||
=== ecloud is now known as ecloud_wfh | ||
mardy | Elleo: hi! I want to carry over the word prediction data from one Ubuntu phone to another; which files do I need to copy? | 09:45 |
---|---|---|
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
=== zbenjamin_ is now known as zbenjamin | ||
=== Zic is now known as Guest2508 | ||
Elleo | mardy: ~/.presage and .local/share/maliit-server | 10:32 |
mardy | Elleo: thanks! | 10:33 |
Elleo | no problem :) | 10:33 |
=== Guest2508 is now known as Zic | ||
mardy | Elleo: hmmm... doesn't seem to work, I still get the old bogus predictions. I also tried to restart maliit-server, but didn't help | 10:46 |
=== sil2100_ is now known as sil2100 | ||
=== sil2100_ is now known as sil2100 | ||
=== _salem is now known as salem_ | ||
oSoMoN | lpotter, FYI (and if you’re still around), I just filed https://launchpad.net/bugs/1646080 | 11:33 |
ubot5 | Ubuntu bug 1646080 in qtsystems-opensource-src (Ubuntu) "Mir-backed InputDeviceModel’s count property is not updated timely" [Undecided,New] | 11:33 |
=== chihchun is now known as chihchun_afk | ||
Elleo | mardy: you probably need to replace the files while maliit is stopped or it might get overwritten when maliit shuts down | 11:36 |
* mardy tries again | 11:50 | |
mardy | Elleo: how do I stop it? I see it restarts | 11:50 |
mardy | Elleo: nevermind, it's upstart, I got it | 11:56 |
mardy | Elleo: now it did the trick, and I realised I copied the files to a wrong location, previously ;-) so maybe stopping was not even necessary, but anyway :-) | 12:02 |
mardy | thanks! | 12:02 |
=== alan_g is now known as alan_g|lunch | ||
=== alan_g|lunch is now known as alan_g | ||
Elleo | mardy: ah, great :) | 12:14 |
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
=== JanC is now known as Guest103 | ||
=== JanC_ is now known as JanC | ||
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
Telix | hello! =D | 18:04 |
dobey | hi | 18:05 |
Telix | what is this node about? | 18:06 |
dobey | this channel is about the ubuntu for phones/tablets images | 18:06 |
Telix | Development? | 18:09 |
dobey | sure | 18:10 |
dobey | and using | 18:11 |
dobey | see the /topic | 18:11 |
Telix | thanks! | 18:12 |
=== robru_ is now known as robru | ||
mterry | tedg: do you know how to get a bunch of debug output from UAL? I'd like to investigate what it thinks the installed apps are in my VM instance | 18:37 |
mterry | and just see where it's failing to find any | 18:37 |
dobey | mterry: G_MESSAGES_DEBUG=all | 19:12 |
mterry | yeah that's already set | 19:12 |
dobey | mterry: G_MESSAGES_DEBUG=all ubuntu-app-launch-appids should give you an idea of what it thinks is installed | 19:12 |
mterry | unity8-session.ubuntu-app-launch-appids | 19:12 |
mterry | shows what i'd expect | 19:12 |
mterry | I'm trying to debug specifically the interaction between the click/app scope and UAL | 19:13 |
dobey | mterry: what's the problem? click scope is empty? | 19:13 |
mterry | dobey: yah | 19:13 |
dobey | mterry: U1_DEBUG=1 for the click scope | 19:13 |
mterry | but only on Core | 19:13 |
mterry | dobey: ooh ok thanks | 19:13 |
mterry | That is not intuitive :P | 19:13 |
dobey | well, historical things rarely are ;) | 19:14 |
mterry | no extra debug with that | 19:21 |
dobey | are you running the scope manually? | 19:21 |
mterry | no, I edited the upstart job to set that | 19:21 |
dobey | upstart job? | 19:23 |
dobey | mterry: how empty is the click scope? totally blank? does it have the BFB for the store? | 19:23 |
mterry | unity8-dash's upstart job | 19:23 |
mterry | No BFB | 19:23 |
dobey | then the scope itself probably isn't actually running | 19:24 |
dobey | even with no apps installed, the BFB should be there (assuming unity-scope-snappy is also installed) | 19:24 |
mterry | Well it is installed... But this is all in a snap, and my first guess with this stuff is that some hardcoded path somewhere is making it think it's not installed | 19:25 |
mterry | dobey: how to test if the scope itself is running? unity8-dash thinks it is... is there some helper executable to look for? | 19:25 |
dobey | mterry: scoperunner is the executable which loads the .so | 19:26 |
dobey | mterry: when you refresh the scope it should be alive for like 30 seconds or so, iirc | 19:26 |
mterry | dobey: interesting. I see the scoperunners for music and video when I go to them. But nothing after refreshing click scope | 19:27 |
dobey | mterry: hmm. and unity-scoe-click is in fact installed in the snap? :) | 19:31 |
dobey | err, unity-scope-click that is | 19:31 |
mterry | dobey: yeah | 19:31 |
dobey | hmmm | 19:32 |
mterry | I see output like: | 19:33 |
mterry | [2016-11-30:19:17:16.923] Pre-populating first scope: "clickscope" | 19:33 |
mterry | [2016-11-30:19:17:16.923] "clickscope" : results invalidated, programmatic: false , active: false | 19:33 |
mterry | [2016-11-30:19:17:16.923] "clickscope" : Dispatching search: "" "" (programmatic: true ) | 19:33 |
mterry | so it thinks it's running and tries to do stuff | 19:33 |
mterry | It just returns nothing | 19:33 |
dobey | well that suggests it started | 19:33 |
dobey | but seems like it doesn't stay running | 19:33 |
mterry | Well I get as far as | 19:34 |
mterry | [2016-11-30:19:00:36.496] "clickscope" : flushUpdates: #results = 0 finalize: true | 19:34 |
mterry | [2016-11-30:19:00:36.496] "clickscope" : processPrimaryNavigationTag(): tag is ' "" ' | 19:34 |
mterry | [2016-11-30:19:00:36.496] "clickscope" : Processing 0 filters | 19:34 |
mterry | [2016-11-30:19:00:36.496] "clickscope" : Removing all filters | 19:34 |
dobey | well if it returned nothing, there'd be another line with results: 0 or something | 19:34 |
dobey | hmm, why is indicator-transfer failing to build on zesty :( | 19:34 |
mterry | dobey: I think there's some gmock mess up | 19:35 |
dobey | mterry: oh the 1.8 stuff? | 19:36 |
mterry | ::shrug:: | 19:36 |
mterry | I've only cared about xenial+overlay this past month :) | 19:37 |
dobey | well i'm re-running in jenkins which hopefully makes it a little easier for me to poke at | 19:37 |
dobey | sure | 19:37 |
dobey | i was just asked to help get this stuff landed, though. and if it's not building on zesty, then landing is a bit hard :) | 19:38 |
dobey | mterry: so i don't know what the *exact* command line would be running inside the snap | 19:38 |
dobey | mterry: but you can run U1_DEBUG=1 /usr/lib/*/unity-scopes/scoperunner "" /usr/lib/*/unity-scopes/clickscope/clickscope.ini | 19:39 |
dobey | mterry: then you can do a refresh and see what goes wrong in the terminal | 19:39 |
dobey | mterry: i guess whoopsie/apport doesn't quite work in snaps world at the moment, in case things are crashing? | 19:40 |
mterry | oy ok I have to set up all the ld-library path and nonsense | 19:40 |
mterry | dobey: | 19:50 |
mterry | Failed to open departments db: Invalid departments database | 19:50 |
mterry | WARNING: QApplication was not created in the main() thread. | 19:50 |
mterry | (process:3837): ubuntu-app-launch-DEBUG: Initialized Click DB | 19:50 |
mterry | scoperunner: /build/unity-scopes-api-WePnyG/unity-scopes-api-1.0.7+16.04.20161115/src/scopes/internal/zmq_middleware/ObjectAdapter.cpp:670: void unity::scopes::internal::zmq_middleware::ObjectAdapter::worker(): Assertion `buf.empty()' failed. | 19:50 |
mterry | Aborted | 19:50 |
dobey | oh fun | 19:51 |
dobey | so yeah, that's why you don't have any apps | 19:51 |
dobey | no idea what that means though | 19:51 |
mterry | heh | 19:51 |
dobey | hmm, yeah, indicator-transfer is the gmock 1.8 stuff it seems | 19:54 |
mterry | dobey: I assume the departments db thing is fine? That's just for click and we don't need it here? | 19:57 |
* mterry looks at u-s-a code for why buf is empty | 19:57 | |
dobey | mterry: yeah, the departments.db is just a "warning" | 19:58 |
dobey | oh well at least fixing the indicator-transfer issue was easy :) | 20:02 |
dobey | assuming i don't need to also fix a boatload of tests anyway | 20:03 |
dobey | mterry: have any luck looking at scopes-api code? | 21:12 |
mterry | dobey: no but I also got distracted | 21:17 |
=== salem_ is now known as _salem | ||
mterry | I'm not very familiar with zmq in general. A little confused | 21:26 |
mterry | Don't know who the other side of this talk is. I guess the Dash? | 21:27 |
dobey | unity-scopes-shell is the plug-in for the dash which sends the queries to scopes and such | 21:28 |
mterry | dobey: great, it doesn't crash under gdb... :( Although it still doesn't work right, so maybe there's hope there | 21:40 |
dobey | hmm | 21:41 |
mterry | dobey: with gdb I get the warning: "[2016-11-30 21:43:45.633] ERROR: clickscope: ObjectAdapter: error unmarshaling request header (id: , adapter: clickscope-c, op: ): ZmqReceiver::receive(): socket was closed" | 21:44 |
mterry | And eventually (after finding some snaps): "clickscope: ReplyImpl::error(): ThreadPool::submit(): cannot accept task for destroyed pool" | 21:44 |
dobey | mterry: interesting. a bit beyond my knowledge of scopes api though. | 21:45 |
mterry | dobey: yeah I think I need to assign to someone with domain knowledge | 21:46 |
dobey | mterry: yeah, i'll have to defer to michi i think | 21:47 |
mterry | dobey: you're trying to land the indicator silo? | 21:52 |
mterry | dobey: there are a couple branches that didn't make it in for whatever reason | 21:52 |
dobey | mterry: yeah | 21:52 |
mterry | (into the current silo) | 21:52 |
dobey | mterry: what else? | 21:52 |
mterry | dobey: indicator-network, indicator-keyboard, and mediascanner2 (though that last one may not be your domain) | 21:52 |
dobey | hmm, keyboard tests are crashing on zesty :( | 21:53 |
mterry | They are in silo 2129, but weren't in your silo last I checkd | 21:53 |
dobey | i added keyboard | 21:53 |
dobey | but it's failing on zesty :( | 21:53 |
mterry | more gmock changes? | 21:53 |
dobey | no, it's not c++ i don't think | 21:54 |
piglit | when will there be a new ubuntu phone | 22:07 |
piglit | like they are all sold out | 22:07 |
dobey | that's a question for the manufacturers, not ubuntu. we just make the OS. they make the phones and announcements about when they'll be released :) | 22:08 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!