| tedg | Saviq, Hmm, I thought it did because that's how Click checks to see if the app is running and shuts it down, no? | 00:49 |
|---|---|---|
| tedg | Saviq, Yeah, this calls UAL, so it needs to have access to the session bus. http://bazaar.launchpad.net/~click-hackers/click/trunk/view/head:/lib/click/database.vala#L267 | 00:53 |
| === fginther is now known as fginther` | ||
| === _salem is now known as salem_ | ||
| === salem_ is now known as _salem | ||
| === Mirv_ is now known as Mirv | ||
| tsdgeos | Saviq: interestingly the backtrace is different | 07:46 |
| tsdgeos | so indeed seems a different thing :S | 07:46 |
| tsdgeos | or not... | 07:46 |
| Saviq | tsdgeos, I assume you're talking about the last crash in DashContent? | 08:07 |
| tsdgeos | Saviq: yes | 08:07 |
| tsdgeos | i think it's a different crash | 08:07 |
| tsdgeos | because if i run from qtdeclarative git it doesn't crash (or hasn't for the last 20 min) | 08:08 |
| Saviq | mzanetti, did you notice tedg's reply yesterday about click hooks? | 08:08 |
| tsdgeos | let me try to see if the incubate patch we spoke yesterday actually fixes this one | 08:08 |
| mzanetti | Saviq: no... fell asleep like instantaneously | 08:08 |
| Saviq | <tedg> [02:49:58] Saviq, Hmm, I thought it did because that's how Click checks to see if the app is running and shuts it down, no? | 08:08 |
| Saviq | [02:53:25] Saviq, Yeah, this calls UAL, so it needs to have access to the session bus. http://bazaar.launchpad.net/~click-hackers/click/trunk/view/head:/lib/click/database.vala#L267 | 08:08 |
| Saviq | mzanetti, ↑ | 08:09 |
| mzanetti | err... | 08:09 |
| mzanetti | but... | 08:09 |
| mzanetti | it doesn't :D | 08:09 |
| mzanetti | Saviq: my hook is a script that echoes those variables to a file, and they are all empty | 08:10 |
| mzanetti | so probably something wrong with the hook still | 08:10 |
| mzanetti | Saviq: see anything wrong? http://paste.ubuntu.com/8086809 | 08:11 |
| Saviq | mzanetti, not really | 08:13 |
| * Saviq tries it out | 08:13 | |
| mzanetti | Saviq: http://paste.ubuntu.com/8086819 | 08:13 |
| Saviq | mzanetti, yeah I just went env > | 08:17 |
| Saviq | mzanetti, but confirmed it doesn't work | 08:18 |
| Saviq | or well, that the env is empty / weird | 08:18 |
| Saviq | on hook install it was sane | 08:19 |
| Saviq | but on app install it was just 4 random values | 08:19 |
| mzanetti | Saviq: so what you suggest? | 08:35 |
| mzanetti | going the session route? | 08:36 |
| mzanetti | I mean the upstart thing we discusses yesterday | 08:37 |
| Saviq | mzanetti, no, we'll need to talk to folks later today | 08:49 |
| Saviq | mzanetti, let's find out what's their approach, unfortunately cjwatson isn't around until later this week AFAIR | 08:49 |
| sil2100 | Saviq: hi! I noticed you did some reviews for MacSlow's unity-notifications | 09:17 |
| sil2100 | Saviq: could you (or find anyone else from your team) have a take on reviewing his branch? It's a blocker fix: https://code.launchpad.net/~macslow/unity-notifications/fix-1354406/+merge/231005 | 09:18 |
| Saviq | sil2100, will do | 09:27 |
| Saviq | @unity, who wants to take ↑ please? | 09:27 |
| sil2100 | Thanks! | 09:27 |
| Saviq | MacSlow, just one clarification on the above | 09:27 |
| Saviq | MacSlow, I'm worried that the push notifications backend might send multiple notifications in the name of different services | 09:28 |
| Saviq | Chipaca, can you relate to ↑ for us? | 09:28 |
| Chipaca | what's that again? | 09:29 |
| * Chipaca reads | 09:31 | |
| Chipaca | I'm not sure what that does | 09:32 |
| Chipaca | just skimming the mp, i mean, i can't tell if it's adding senderName to the dbus api, if it's adding it to the libnotify api, or if it's deducing it from the dbus connection | 09:33 |
| Chipaca | Saviq: ? | 09:33 |
| Chipaca | none of those don't break push notifications, fwiw | 09:35 |
| Saviq | Chipaca, yeah, that's what I wanted clarification on | 09:35 |
| Chipaca | Saviq: which that is that? | 09:35 |
| Saviq | Chipaca, with MacSlow ;) | 09:35 |
| Chipaca | the breaking of push, or the thing that's being implemented? | 09:35 |
| Saviq | Chipaca, don't you send more than one notification from the push service if they come in at the same time? | 09:35 |
| Chipaca | yes | 09:36 |
| Chipaca | not more than one per application at the same time though | 09:36 |
| Chipaca | although they could be in quick succession | 09:36 |
| Saviq | Chipaca, yeah, that's the point, per application, but I'm worried the change in the above MP actually looks at the dbus sender name | 09:38 |
| Saviq | Chipaca, which would mean that you'd be prevented from sending more than one in quick succession | 09:38 |
| Saviq | tsdgeos, could you please look through https://code.launchpad.net/~marcustomlinson/unity8/lp-1356410/+merge/231326 | 09:39 |
| Saviq | tsdgeos, I think we might need to do onGotoScope there too | 09:39 |
| tsdgeos | Saviq: sure | 09:39 |
| Cimi | where goes the qml log of autopilot? | 10:05 |
| Saviq | Cimi, same as everything | 10:06 |
| Saviq | Cimi, ~/.cache/upstart/unity8{,-dash}.log | 10:07 |
| Cimi | Saviq, I have AP failure on scope settings | 10:07 |
| Cimi | Saviq, fixing that... | 10:07 |
| Cimi | Saviq, currentIndex related issues | 10:07 |
| Cimi | Saviq, hah, found | 10:09 |
| Saviq | Cimi, good boy | 10:09 |
| Saviq | :P | 10:09 |
| Cimi | Saviq, are you sure of this? > + readonly property int count: status === Loader.Ready && item.hasOwnProperty("count") ? item.count : 0 | 10:12 |
| Cimi | item && item.count || 0 | 10:12 |
| Cimi | mmm | 10:13 |
| Saviq | Cimi, why not? | 10:14 |
| Saviq | Cimi, if item is null, 0 | 10:14 |
| Saviq | Cimi, if item.count evaluates to false (null, undefined, 0), 0 | 10:14 |
| Saviq | otherwise item.count | 10:15 |
| Cimi | Saviq, yeah indeed... | 10:16 |
| Cimi | Saviq, there is sth weird or I am stupid | 10:17 |
| Cimi | I'm in autopilot test for open preview, this code http://paste.ubuntu.com/8087581/ | 10:17 |
| Cimi | preview_list.currentIndex seems 1 here, but log in qml shows 0 | 10:18 |
| Saviq | Cimi, onCurrentIndexChanged: console.debug(currentIndex)? | 10:24 |
| Cimi | 0 0 0 | 10:24 |
| Cimi | starts from -1 and becomes 0 | 10:24 |
| Cimi | not 1 | 10:24 |
| Cimi | I just forced it to 0, let's see | 10:25 |
| Cimi | forcing to 0 in the loader is actually 0 | 10:26 |
| Cimi | test pass | 10:26 |
| Saviq | Cimi, sure you get the correct object in autopilot? | 10:27 |
| Cimi | Saviq, works with readonly property int currentIndex: status === Loader.Ready && item.hasOwnProperty("currentIndex") ? item.currentIndex : -1 | 10:29 |
| Cimi | Saviq, but doesn't with readonly property int currentIndex: item && item.currentIndex || -1 | 10:29 |
| Saviq | Cimi, why do you even need to pass currentIndex up? | 10:32 |
| Cimi | Saviq, tests I think | 10:32 |
| Saviq | Cimi, just get into the loader in tests instead | 10:32 |
| Cimi | Saviq, ok but doesn't answer my mental question "what's wrong in that code" | 10:33 |
| mzanetti | soo... I tried to switch over to unity and increased the scale factor for high dpi to 1.5, which causes icons to be 1.5 but fonts to be scale factor 3 | 10:35 |
| mzanetti | 'cause X11 reports a dpi of 220 which changes fonts already | 10:35 |
| mzanetti | anyone else here running unity on a high dpi screen? | 10:35 |
| Saviq | Cimi, I'd say the binding is broken somewhere, and you're not really looking at the same property | 10:36 |
| Cimi | Saviq, but console.log I put both inside the previewlist and the loader, they all change at the same time | 10:36 |
| Cimi | Saviq, unless this is a timing issue | 10:37 |
| Cimi | Saviq, having -1 intead 0... | 10:37 |
| Cimi | even this works item && item.hasOwnProperty("currentIndex") ? item.currentIndex : -1 | 10:43 |
| dandrader | mzanetti, ping | 10:43 |
| Cimi | but not item && item.currentIndex || -1 | 10:43 |
| Saviq | Cimi, well | 10:43 |
| Saviq | Cimi, 0 evaulates to false | 10:44 |
| Saviq | Cimi, so when currentIndex is 0 | 10:44 |
| Saviq | Cimi, it would end up -1 in this case | 10:44 |
| Cimi | right!!!!!!!!! | 10:44 |
| Saviq | Cimi, it only made sense when it was || 0 | 10:44 |
| Cimi | Saviq, that must be it, thanks!! | 10:44 |
| mzanetti | dandrader: pong | 10:47 |
| Cimi | Saviq, ok pushed | 10:47 |
| === MacSlow is now known as MacSlow|lunch | ||
| greyback_ | Saviq: think you'd have a chance to re-visit https://code.launchpad.net/~gerboland/unity-api/surfaceSizerCallback/+merge/230270 ? | 11:02 |
| greyback_ | if not, please point someone else to it | 11:02 |
| Saviq | greyback_, yeah, will do | 11:04 |
| dednick | Saviq: fix for the indicator bug - https://code.launchpad.net/~nick-dedekind/unity8/lp1328646/+merge/231335 | 11:14 |
| Saviq | dednick, great, thanks | 11:15 |
| facundobatista | Holas | 11:17 |
| === MacSlow|lunch is now known as MacSlow | ||
| === alan_g is now known as alan_g|lunch | ||
| greyback_ | dandrader: I replied to you in https://code.launchpad.net/~gerboland/unity8/initialSurfaceGeometry/+merge/230490 | 12:13 |
| dandrader | greyback_, !? http://bazaar.launchpad.net/~gerboland/unity8/initialSurfaceGeometry/revision/1147 | 12:14 |
| dandrader | ah right. nevermind that. you're adding it back again | 12:15 |
| Cimi | Saviq, finally something that looks sane https://code.launchpad.net/~cimi/unity8/infographics-august-merge/+merge/231344 | 12:15 |
| Cimi | took me a while :D | 12:16 |
| === _salem is now known as salem_ | ||
| === Pici` is now known as Pici | ||
| === alan_g|lunch is now known as alan_g | ||
| mzanetti | tedg: ping | 13:16 |
| tedg | Good morning mzanetti | 13:16 |
| mzanetti | tedg: good morning! | 13:16 |
| mzanetti | tedg: so, I saw your reply regarding the session and click hooks | 13:17 |
| dednick | Saviq: re the ref/deref thing. I couldn't get it working properly with being able to share the object. If i give the menumodel i create no context, it doesn't seem to delete until the app is close, not when the last ref is released. | 13:17 |
| mzanetti | tedg: doesn't work for me :D and thinking about it, it kinda makes sense that the hooks don't have the session exported | 13:17 |
| mzanetti | tedg: because there could be multiple sessions running | 13:17 |
| dednick | Saviq: and if i give it the createObject context, it just deletes the menumodel regardless if there are other things referencing it. | 13:18 |
| tedg | mzanetti, The click user hooks should always be run as the user that they're referencing. | 13:18 |
| mzanetti | tedg: yes, its ran as the user phablet, that's fine | 13:18 |
| mzanetti | tedg: but it doesn't have DBUS_SESSION_BUS_ADDRESS exported | 13:18 |
| Saviq | dednick, hmm wonder if you forced the garbage collector to run... | 13:19 |
| dednick | Saviq: howto? | 13:19 |
| tedg | mzanetti, Asking cjwatson about it, worst case you can get it from the file, but it seems like other hooks would use it as well. | 13:19 |
| * mzanetti is confused about the channel := | 13:20 | |
| mzanetti | :) | 13:20 |
| Saviq | dednick, http://qt-project.org/doc/qt-5/qtqml-javascript-qmlglobalobject.html gc() | 13:20 |
| mzanetti | tedg: right... we figured the way to get it though the upstart session file and then get-env for the dbus address | 13:20 |
| Saviq | dednick, that object needs to be parentless when passed to QML for it to be garbage collected though | 13:20 |
| mzanetti | tedg: but that seems a bit... | 13:20 |
| Saviq | mzanetti, I talked with cjwatson briefly today | 13:21 |
| dednick | Saviq: yup. | 13:21 |
| mzanetti | ah | 13:21 |
| mzanetti | Saviq: outcome? | 13:21 |
| Saviq | mzanetti, and he said this should really work, as those are run *from* the upstart session | 13:21 |
| josharenson | If I have a branch in +junk and I need to move it to a 'team branch' so it can be merge elsewhere, can I just push it to ~unity-team ? | 13:21 |
| mzanetti | Saviq: ah, so you mean the hook would be executed multiple times in case of multiple sessions? | 13:22 |
| Saviq | mzanetti, but then I lost his attention | 13:22 |
| Saviq | mzanetti, would have to be yeah (not that we do support multiple sessions yet) | 13:23 |
| tedg | mzanetti, Saviq, what do you guys do for legacy apps, a directory watch? | 13:23 |
| mzanetti | not sure yet... | 13:23 |
| Saviq | tedg, no, we'll only do startup | 13:23 |
| Saviq | tedg, as those can only change on OTA | 13:23 |
| Saviq | tedg, and if you make your device writable... your fault, reboot :P | 13:23 |
| tedg | Saviq, I can write something to .local/share/applications anytime I want :-) | 13:23 |
| Saviq | tedg, hmm right forgot .local | 13:24 |
| tedg | (or delete) | 13:24 |
| Saviq | tedg, sounds like directory watch then, nothing else | 13:24 |
| tedg | Saviq, If you make a dbus interface for "changed" I'd recommend setting up an Upstart job with the directory watch that then signals the changed. | 13:24 |
| Saviq | tedg, but *later* | 13:24 |
| tedg | Saviq, Keeps the code simpler. | 13:24 |
| Saviq | tedg, yeah we do | 13:24 |
| Saviq | have a dbus interface | 13:24 |
| Saviq | sounds like a good idea | 13:25 |
| dednick | Saviq: hm. no, doesnt seem to do the trick either. | 13:25 |
| mzanetti | I guess I would just do the filesystem watch in the launcher though | 13:25 |
| mzanetti | why another process? | 13:25 |
| tedg | You can do something like: start on (file FILE=~/.local/share/applications/*.desktop) | 13:25 |
| Saviq | dednick, thing is I think they need to get out of scope to get gc'd | 13:25 |
| Saviq | dednick, there's no refcounting IIRC | 13:26 |
| dednick | Saviq: they should be oos. if the UnitMenuModel cache is destroyed. | 13:26 |
| Saviq | actually no, it should be refcounted | 13:27 |
| dednick | Saviq: yet they don't really have scope, since they are created without context | 13:27 |
| dednick | so the engine is probably their scope | 13:27 |
| dednick | well, without parent i mean | 13:28 |
| dandrader | greyback_, mzanetti all fixed in the unity8/lifecycle mp | 13:28 |
| mzanetti | dandrader: ok... I think we're good code wise now... need to have some little more testing before I feel confident enough to approve | 13:29 |
| mzanetti | dandrader: but I don't expect much more breakage | 13:29 |
| Saviq | dednick, yeah, so they should be ref'd after all | 13:29 |
| tedg | Saviq, Anyway, where I was going, if you make your own directory of links and watch that you'll get the change when the links are updated and a distinction between legacy and click apps. | 13:29 |
| tedg | Saviq, You can also use the short appids | 13:30 |
| Saviq | dednick, can you read http://qt-project.org/doc/qt-5/qqmlengine.html#ObjectOwnership-enum and see if anything you're doing applies and results in cppownership? | 13:30 |
| mzanetti | tedg: ooh! that's awsome | 13:30 |
| mzanetti | didn't know that | 13:30 |
| mzanetti | tedg: will that still work when we switch to systemd? :) | 13:30 |
| Saviq | tedg, mzanetti, but do we care actually? we need to go through all locations regardless | 13:30 |
| Saviq | ? | 13:30 |
| mzanetti | Saviq: well, we need to get triggered somehow | 13:31 |
| tedg | Saviq, Yes, but knowning click or not would be useful for you. And so would be using the short ids because they'd change less. | 13:32 |
| Saviq | mzanetti, yeah, we get triggered by the hook, whether the Pattern: is different for us than for click hook | 13:32 |
| Saviq | tedg, but then if we monitor .local/share/applications, we'd get notified about any changes there anyway | 13:32 |
| Saviq | tedg, short or not | 13:32 |
| mzanetti | Saviq: not sure I understand. I don't think the click hook will execute when a legacy app is installed/removed, is it? | 13:33 |
| Saviq | mzanetti, exactly, no, but that's not something that tedg was referring to | 13:33 |
| Saviq | mzanetti, unless you mean the upstart notify ;) | 13:33 |
| mzanetti | yes, I meant that | 13:33 |
| mzanetti | ok | 13:33 |
| Saviq | mzanetti, because there seem to be two conversations going on in parallel :) | 13:33 |
| mzanetti | confusion solved | 13:33 |
| mzanetti | there seem to be more than 2 :D | 13:34 |
| mzanetti | Saviq: so I'm still not sure what to do now.. | 13:37 |
| Saviq | mzanetti, first we need to understand why your hook does not get the right env | 13:38 |
| Saviq | mzanetti, join #ubuntu-devel and talk to cjwatson please | 13:38 |
| === dandrader is now known as dandrader|afk | ||
| dednick | Saviq: it seems to be working, but don't seem to have any control over garbage collection. | 13:46 |
| dednick | Saviq: so don't really know if the models are being deleted. | 13:47 |
| dednick | Saviq: which means they will be left running | 13:47 |
| Saviq | dednick, well, qDebug() << "deleted!"; to try? ;) | 13:47 |
| Saviq | in ~ | 13:47 |
| dednick | Saviq: yes, well that's what i mean. it's a bit random | 13:47 |
| Saviq | dednick, yeah, that's gc for you | 13:47 |
| Saviq | dednick, it's not an exact science | 13:47 |
| dednick | Saviq: well, i'm not convinced! | 13:48 |
| dednick | Saviq: we could be left with models running until app close | 13:48 |
| Saviq | dednick, until shell close, which it never does... | 13:48 |
| Saviq | dednick, and because it's a cache... | 13:48 |
| dednick | Saviq: presicely. | 13:49 |
| Saviq | dednick, it's caching things | 13:49 |
| Saviq | dednick, but do you mean we'll be leaking them? | 13:49 |
| MacSlow | dandrader|afk, Saviq: whoever has time... https://code.launchpad.net/~macslow/unity8/fix-1348092/+merge/228090 is passing all AP- and qml-tests again... so another look is welcomed. | 13:49 |
| Saviq | dednick, the cache should probably be a factory is all? | 13:50 |
| dednick | Saviq: possibly. if we're switching from different profiles | 13:50 |
| Saviq | dednick, meaning it keeps a handle on those it created | 13:50 |
| Saviq | dednick, and gives the same up if they exist in the cache? | 13:50 |
| dednick | Saviq: that's what it does | 13:51 |
| Saviq | dednick, then at worst we'll have $number_of_indicators of them hanging around | 13:51 |
| Saviq | dednick, what's your worry? mem? cpu usage? | 13:51 |
| Saviq | dednick, don't we have them all around anyway for the panel? | 13:51 |
| dednick | Saviq: yes. responding to changes in indicators. | 13:51 |
| Saviq | dednick, or are those only used in the pages themselves? | 13:52 |
| dednick | Saviq: i think we are supposed to be switching indicator profiles when changing from greeter mode to phone mode. which means we'll have indicators*2 models | 13:53 |
| dednick | to "normal" mode i mean | 13:53 |
| Saviq | dednick, can't the cache be switched to a different profile? | 13:54 |
| Saviq | dednick, and really I don't think it's so bad to have them around for a while, they will get gc'd at some point | 13:54 |
| Saviq | dednick, but if you're going in and out if them, it's even better to keep them around | 13:55 |
| dednick | Saviq: ok, well if you think it's ok then i'm not fussed. | 13:56 |
| mzanetti | Saviq: so we're going for a workaround/hack atm, that's for sure. The 2 possibilites we have atm is changing click-scope to execute click hook run-user after package operations or the session file thing from yesterday | 13:58 |
| mzanetti | I guess I vote for the session file | 13:58 |
| Saviq | dednick, as long as you can see them really getting destroyed | 13:58 |
| Saviq | mzanetti, "atm" meaning until we get an update from cjwatson? | 13:58 |
| mzanetti | Saviq: yes | 13:59 |
| dednick | Saviq: i can see them being destroyed on app shutdown. | 13:59 |
| Saviq | mzanetti, or should we just do the upstart thing | 13:59 |
| Saviq | dednick, oh well, that sounds like they're not deref'd then on menus closed | 13:59 |
| mzanetti | Saviq: atm means, cjwatson can look at it at earliest in september | 13:59 |
| Saviq | dednick, so not gc'd | 13:59 |
| mzanetti | Saviq: after rtm afaiu | 13:59 |
| Saviq | mzanetti, should we just go for the upstart thing anyway | 13:59 |
| mzanetti | Saviq: +1 | 13:59 |
| Saviq | mzanetti, we wouldn't need the click hook at all then | 14:00 |
| mzanetti | huh? | 14:00 |
| mzanetti | Saviq: why not? | 14:00 |
| mzanetti | Saviq: aaa | 14:00 |
| Saviq | mzanetti, because we'll be monitoring ~/.local/share/applications already | 14:00 |
| mzanetti | you mean tue upstart file watcher | 14:00 |
| Saviq | yes | 14:00 |
| mzanetti | that actually sounds even better | 14:00 |
| mzanetti | Saviq: ok. /me doing that | 14:00 |
| mzanetti | thanks! | 14:01 |
| Saviq | tedg, can you point us at the upstart directory watcher thingy? | 14:01 |
| tedg | Saviq, One sec, OTP | 14:01 |
| mzanetti | Saviq: iafau its justa job file with "start on (file FILE=~/.local/share/applications/*.desktop)" | 14:02 |
| mzanetti | and the dbus call in exec | 14:02 |
| Saviq | ah so it's a bridge | 14:03 |
| Saviq | mzanetti, yeah, sounds good, make it start on /usr/share/applications/*.desktop too | 14:04 |
| mzanetti | Saviq: yeah, sure | 14:04 |
| Saviq | mzanetti, and on unity8 startup, but that one probably internally in LauncherBackend | 14:04 |
| mzanetti | yep | 14:04 |
| Saviq | MacSlow, btw, please put some context in your branch names, is difficult to find them when they're just named after bugs | 14:06 |
| tedg | Cool, looks like you guys got it. Yeah, just using he file bridge. | 14:09 |
| tedg | Send the dbus message then cat all the files into /dev/null so the FS loads them into cache :-) | 14:11 |
| Saviq | dednick, is the panel not using the cache models? | 14:12 |
| dednick | Saviq: everything should be | 14:14 |
| MacSlow | Saviq, hehe... I put the bug-# in the name to lead right to it... anyway branches for unity8 and unity-notification with the name "fix-1348092" are the visual-updates for notifications. | 14:14 |
| Saviq | MacSlow, yeah, it's fine if I actually go to that page | 14:15 |
| Saviq | MacSlow, but when I try to use my browser history, that doesn't work any more ;) | 14:15 |
| Saviq | MacSlow, and then the branches are linked to bugs | 14:15 |
| dednick | Saviq: everything that gains a reference to a model should be getting it from the cached model, which is part of IndicatorBase (which both Page & Widget inherit from). So as sson as base goes OOS, then the cached model should be destroyed and references released. | 14:16 |
| Saviq | dednick, so how can they ever get deleted if the panel uses it? | 14:16 |
| Saviq | dednick, isn't it the same cache both Page and Widget use? | 14:16 |
| Saviq | *cached model, that is | 14:17 |
| MacSlow | Saviq, true... but seeing (just from looking at branch-names) what is a actual bug-fix vs a test/feature/proof-of-concept branch saves time skimming through a lot of them... imo | 14:17 |
| Saviq | MacSlow, I didn't say *do not* put bug numbers in there | 14:17 |
| Saviq | MacSlow, just add some context too | 14:17 |
| Saviq | MacSlow, and remember we have bugs that are features too ;) | 14:17 |
| MacSlow | Saviq, yeah... and odd concept ;) | 14:18 |
| dednick | Saviq: yes. if we lose an indicator (remove a file for example) the indicator manager will remove the entry from the indicators model. This should then remove both page & widget from the panel. | 14:18 |
| Saviq | dednick, ok, and that doesn't happen? | 14:18 |
| dednick | Saviq: as far as i know it does. trying to figure it out now. | 14:18 |
| Saviq | dednick, or e.g. when you disable the bluetooth indicator (or should that even cause the removal?) | 14:18 |
| dednick | Saviq: disabling not necessarily. that might just switch off the visibility. | 14:19 |
| dednick | Saviq: which we still need the connection to the model for. | 14:19 |
| Saviq | dednick, in that case IMO it's really a moot point to be trying to deleting them, when would an indicator ever disappear? | 14:19 |
| Saviq | in real life that is | 14:19 |
| dednick | Saviq: well, it's not the removals i was concerned about. was the profile changing. | 14:20 |
| Saviq | dednick, ah ok, so yeah, can't we reuse the cache for a different profile, or do we explicitly *not* want to do that | 14:21 |
| Saviq | dednick, aaand anyway, tedg, mterry, is the plan to actually switch profiles between greeter/touch in the short term, or just the touch profile going into a stripped down mode when greeter locked? | 14:21 |
| dednick | Saviq: we want to allow any number of profiles to exist, but don't really want the ones we're not using to continue running. (or this is what i thought) | 14:22 |
| Saviq | dednick, yeah, can we *switch* a cached model to a different profile, not destroy and create one with the different profile? | 14:22 |
| mterry | Saviq, last time I spoke to tedg, he suggested the easiest way was for the touch profile to enter a stripped mode itself rather than us switching. But maybe that implementation preference has changed since then | 14:24 |
| Saviq | yeah that's kinda my question | 14:24 |
| dednick | Saviq: well, at the moment the cached model will fork. leave the other models as they want and create a new model. | 14:24 |
| mterry | Saviq, yeah I guess I'm saying I don't know ;) | 14:24 |
| dednick | for the new profile | 14:24 |
| Saviq | mterry, that's I pung tedg on this, too :) | 14:25 |
| Saviq | why | 14:25 |
| tedg | Saviq, mterry, I think having the stripped profile in teh session makes sense, as the greeter profile. | 14:26 |
| tedg | The issue was before is that we were "cheating" in making them the same, just the data was stripped. | 14:26 |
| tedg | But I think we should have both in the session and have the stripping go on further down the path. | 14:26 |
| Saviq | dednick, so yeah, I would actually think switching the profile on a pre-existing model should be faster than destroying and creating a new one... | 14:26 |
| Saviq | dednick, and now we're getting to a "why do we need a cache again?" situation... | 14:26 |
| tedg | The actions will still be shared between them. | 14:26 |
| dednick | Saviq: does the same thing. needs to wipe the menu and redcreate. | 14:26 |
| === alan_g is now known as alan_g|tea | ||
| Saviq | dednick, well, at least it doesn't recreate the object itself, but I agree negligible | 14:27 |
| dednick | Saviq: well, the cache is just because multiple things point at the same model. widget/menu, etc. | 14:27 |
| Saviq | dednick, just two - page and widget, right? | 14:28 |
| Saviq | dednick, so in theory we could pass just one object between them | 14:28 |
| === dandrader|afk is now known as dandrader | ||
| dednick | Saviq: well, there is also another thing that sorts out the visible/invisible items | 14:29 |
| Saviq | dednick, but sure, the cache gives us some encapsulation | 14:29 |
| Saviq | dednick, ok, so for now just have a look please whether switching profiles actually gc's them as expected | 14:29 |
| dandrader | MacSlow, did you fix that Image.visible property expression I commented about? | 14:32 |
| MacSlow | dandrader, don't recall seeing that... looking again. | 14:35 |
| === alan_g|tea is now known as alan_g | ||
| MacSlow | dednick, you turn | 14:38 |
| dandrader | MacSlow, it was a comment in the diff, easy to miss | 14:41 |
| MacSlow | dandrader, it is in ~macslow/unity8/fix-1348092, right?! | 14:42 |
| MacSlow | dandrader, ah... found it... was in the first revision of that branch | 14:43 |
| greyback_ | Saviq: FYI I've stuff in silo16 - and unity8 is in there. Lemme know if I'm risking a conflict | 14:46 |
| Saviq | greyback_, you are risking it, but when d'you plan to land? | 14:47 |
| greyback_ | Saviq: was hoping today | 14:47 |
| Saviq | greyback_, no worries then, I'll rebase | 14:47 |
| greyback_ | ok thanks | 14:48 |
| anpok | on applicatin start, what is drawing the application name and rotating dots? | 14:50 |
| === om26er is now known as om26er|away | ||
| greyback_ | anpok: unity8 | 14:51 |
| mzanetti | vesar: hey, quick question: | 14:56 |
| mzanetti | vesar: we're about to land the app lifecycle stuff | 14:56 |
| mzanetti | vesar: which allows keeping screenshots in the spread if an app is killed because of low memeory | 14:56 |
| mzanetti | vesar: so that seems to work fine, but takes a few seconds to respawn the app when selected in the spread | 14:57 |
| mzanetti | I told you about this in our last weeks meeting | 14:57 |
| mzanetti | vesar: should we land it as is for now and wait until you come up with something to fix it? or have an idea already what we could/should do? | 14:58 |
| mzanetti | i.e. spinner on top or whatever | 14:58 |
| mzanetti | we can refine it later in any case | 14:58 |
| Saviq | greyback_, fyi, I've a lp:~mir-team/qtmir/gles-sync branch that I reuse for sync MPs | 15:01 |
| Saviq | greyback_, only problem if there's more than one sync in progress, but that should be rare | 15:01 |
| greyback_ | Saviq: noted | 15:01 |
| Saviq | greyback_, good thing it just pulls trunk fine with bzr pull, no conflicts or anything | 15:02 |
| greyback_ | Saviq: yep, my u8 changes were small enough | 15:02 |
| Saviq | greyback_, how's u8 related to what I wrote above? :D | 15:04 |
| Saviq | mzanetti, let me hit up vesar about this here | 15:04 |
| mzanetti | ack | 15:04 |
| mzanetti | Saviq: thanks | 15:04 |
| greyback_ | Saviq: oh misread, sorry | 15:05 |
| greyback_ | context switch fail | 15:05 |
| Saviq | @unity any design inquiries as I'll be going over to that part of the office? | 15:05 |
| Saviq | greyback_, kk, must be the underscore | 15:05 |
| mterry | none that they probably don't know about | 15:05 |
| Saviq | mterry, sure they know, maybe they need reminding? ;) | 15:05 |
| kgunn | Saviq: there was one about how to handle restarting app for vesa | 15:06 |
| Saviq | kgunn, yup, that's what mzanetti just asked | 15:06 |
| kgunn | oops :) | 15:06 |
| Saviq | @unity or design reviews maybe? I'm still here until tomorrow after lunch, so can impose myself on them :) | 15:07 |
| tsdgeos | Saviq: are we ok in the "see less" stuck on bottom decisions? | 15:07 |
| Cimi | I've some annoying headache... I'll afk for a bit... | 15:09 |
| Saviq | tsdgeos, will confirm | 15:11 |
| mzanetti | Saviq: right... and that pull up thingie in scopes | 15:11 |
| mzanetti | but I guess that's a more general discussion | 15:11 |
| Saviq | mzanetti, you mean the hint? | 15:11 |
| Saviq | mzanetti, what about it? | 15:11 |
| mzanetti | yeah... that it hides the activityindicator | 15:11 |
| mzanetti | and in general shouldn't be there (go away) I guess | 15:12 |
| mzanetti | we talked about that last week | 15:12 |
| Saviq | mzanetti, as it is it's the current desired design | 15:12 |
| Saviq | mzanetti, will need a larger wave of rethinking to change this | 15:12 |
| Saviq | mzanetti, I'll get the lockscreen a design review too then | 15:13 |
| mzanetti | yeah | 15:13 |
| mzanetti | Saviq: don't :D | 15:13 |
| Saviq | lol | 15:13 |
| Saviq | @unity: ah just remembered (or did I say that before?), please make sure to do newline after a brief commit message in case there's more to describe, this way we get concise changelogs but still verbose commit msgs | 15:14 |
| Saviq | s/newline/empty line/ | 15:14 |
| dandrader | Saviq, that didn't work for qtmir | 15:15 |
| Saviq | dandrader, this only happened recently | 15:15 |
| Saviq | dandrader, like last week | 15:15 |
| dandrader | Saviq, the changelog added the brief line *and* the long description | 15:15 |
| Saviq | dandrader, yeah, one of the reasons why it happened last week ;) | 15:15 |
| Saviq | dandrader, because I complained | 15:15 |
| dandrader | :D | 15:15 |
| === dandrader is now known as dandrader|lunch | ||
| === salem_ is now known as _salem | ||
| Saviq | greyback_, in case you'll be rebuilding unity8, please pull in all the top-acked branches into your silo, I'll help testing | 15:22 |
| greyback_ | Saviq: I wasn't planning on a rebuild, but in case I do, ok | 15:22 |
| MacSlow | dandrader|lunch, fixed that visible-thing too | 15:22 |
| Saviq | greyback_, you still might need to, pending results from my -api review ;) | 15:24 |
| greyback_ | Saviq: oh a power play, how naughty | 15:25 |
| Saviq | or at least your reactions to it by noew | 15:25 |
| Saviq | -e | 15:25 |
| Saviq | greyback_, I like to think that's not what happened, the review was in earlier ;) | 15:25 |
| greyback_ | bah, missed that | 15:27 |
| === gatox is now known as gatox_lunch | ||
| greyback_ | Saviq: "why not Binding?" - how do you propose it could work? | 15:49 |
| greyback_ | syntactically I mean | 15:49 |
| === dandrader|lunch is now known as dandrader | ||
| mterry | Saviq, regarding the changelog endline thing... shouldn't we just fix the tool to enforce a newline? | 16:14 |
| AlbertA2 | kgunn: adb shell cat /usr/share/ubuntu-touch-session/usc-wrapper | 16:21 |
| kgunn | ta | 16:21 |
| === gatox_lunch is now known as gatox | ||
| === _salem is now known as salem_ | ||
| Saviq | mterry, enforce it where? | 17:01 |
| Saviq | greyback_, callback: function boo() { } | 17:01 |
| Saviq | dandrader__, greyback_, you around? | 17:02 |
| Saviq | I got a phone in the edge-lock state | 17:02 |
| dandrader__ | Saviq, yes | 17:02 |
| === dandrader__ is now known as dandrader | ||
| Saviq | and it's WEEEIRD | 17:02 |
| Saviq | mumble? | 17:02 |
| dandrader | ok | 17:02 |
| greyback_ | Saviq: that syntax is valid? | 17:03 |
| === alan_g is now known as alan_g|EOD | ||
| Saviq | greyback_, why wouldn't it be? a function is just an object | 17:17 |
| Saviq | greyback_, another possible syntax: | 17:18 |
| Saviq | function blah() { } | 17:18 |
| Saviq | foo: blah | 17:18 |
| Saviq | greyback_, you wouldn't be able to assign it to a property if it wasn't an object | 17:18 |
| greyback_ | Saviq: I guess, my JS-foo ain't too high. But AppMan is a singleton, so have to use Binding{} somehow | 17:19 |
| Saviq | greyback_, sure, Binding { target: ApplicationManager; property: "callback"; value: function foo() { } } | 17:20 |
| greyback_ | not the most readable line IMO | 17:21 |
| Saviq | greyback_, not one line ;) | 17:25 |
| Saviq | greyback_, didn't want to span multiple lines | 17:25 |
| greyback_ | Saviq: ok you can use my silo, just don't screw it up. I'm not ready for a rebuild, but add your crap to it | 17:39 |
| === dandrader is now known as dandrader|doctor | ||
| Saviq | greyback_, I thought you'd do that ;) | 17:44 |
| Saviq | greyback_, aanyway tomorrow | 17:44 |
| greyback_ | sabotage | 17:44 |
| === om26er|away is now known as om26er | ||
| mterry | Saviq, enforce it in the tool that concatenates changelog entries (sorry, was out for a doctor's appt) | 18:47 |
| kgunn | greyback_: i'm +1 on the qtmir testing | 18:50 |
| kgunn | tested on n4, n7, n10 | 18:50 |
| kgunn | greyback_: actually...i just noticed, clock app specifically on the n10 the clock does shift up just a tad, but all other apps look ok | 18:56 |
| kgunn | wonder if its just clock app on a mainstage layout isn't quite ready for primetime ? | 18:57 |
| greyback_ | kgunn: not sure , will look into it tomorrow | 19:00 |
| === dandrader|doctor is now known as dandrader | ||
| dandrader | Saviq I think I found the variable that could disable mouse event emulation for good if not reset properly | 20:55 |
| dandrader | It's QQuickWindowPrivate::touchMouseId. | 20:56 |
| === salem_ is now known as _salem | ||
| dandrader | mterry ping | 21:04 |
| mterry | dandrader, hello | 21:04 |
| dandrader | mterry when you are on the swipe-to-unlock lockscreen, should you be able to access the launcher and indicators or not? | 21:05 |
| mterry | dandrader, you should be able to if you have no passphrase. You may be experiencing either bug 1358340 or bug 1357230 | 21:06 |
| ubot5 | bug 1358340 in Unity 8 "[Indicators] Complete greeter profiles" [Critical,In progress] https://launchpad.net/bugs/1358340 | 21:06 |
| ubot5 | bug 1357230 in Unity 8 "After booting, indicators and launcher can't be pulled in over Greeter" [Critical,In progress] https://launchpad.net/bugs/1357230 | 21:06 |
| dandrader | mterry thanks. just asking because it's not obvious to me what's the expected behavior | 21:07 |
| kgunn | popey: you on? | 21:59 |
| popey | kgunn: yo | 21:59 |
| kgunn | :) | 22:00 |
| kgunn | popey: so trying to repro bug 1327547 | 22:00 |
| ubot5 | bug 1327547 in mir (Ubuntu) "mirscreencast crash on flo #73" [Undecided,New] https://launchpad.net/bugs/1327547 | 22:00 |
| popey | oooh, olde | 22:00 |
| kgunn | i killed powerd | 22:00 |
| kgunn | i verified unity8 is running | 22:00 |
| kgunn | how do i start an app ?(i assume some incantation with desktop file...which i don't do often) | 22:00 |
| kgunn | @olde ...yeah, house cleaning | 22:01 |
| popey | adb shell sudo -u phablet -i start application APP_ID=foo | 22:01 |
| popey | where foo is the app id, com.ubuntu.appname | 22:01 |
| popey | I think | 22:01 |
| kgunn | got it | 22:01 |
| popey | thats what my manky script does anyway | 22:02 |
| kgunn | hmmm, i just get a "failed to start" | 22:03 |
| popey | balls | 22:04 |
| popey | lemme try here | 22:04 |
| popey | ah hang on | 22:04 |
| popey | phablet@ubuntu-phablet:~$ start application APP_ID=com.ubuntu.calendar_calendar_0.4.402 | 22:05 |
| popey | that worked | 22:05 |
| popey | so basically get the .desktop file name from ~/.local/share/applications and strip .desktop off it | 22:05 |
| popey | thats what my script does | 22:05 |
| kgunn | :-/ | 22:05 |
| kgunn | i'm sure it's me somehow | 22:05 |
| popey | doing it as phablet user, right? | 22:06 |
| kgunn | yeah for sure, via phablet-shell...so logged in as user not via adb | 22:07 |
| popey | same here | 22:07 |
| kgunn | popey: if that matters | 22:07 |
| popey | kgunn: while you're here.. ☻ | 22:08 |
| kgunn | oh boy | 22:08 |
| popey | I have been talking to someone about porting their game framework to Ubuntu touch | 22:08 |
| popey | which may have dependencies somewhere down the line on x which needs porting to mir | 22:08 |
| popey | do we have a list of things that are non-toolkits which need porting | 22:08 |
| popey | or is it "as and when they come up" and do we give them help? | 22:08 |
| popey | (i realise you guys are super busy) | 22:09 |
| kgunn | popey: prob is x got to be a monster...so yeah, we kinda gotta look at stuff in a one off fashion of a) do we wanna add some extension | 22:09 |
| kgunn | and b) then figure out where/when to fit that in | 22:09 |
| kgunn | but we're open for discussion | 22:09 |
| popey | thats what I figured | 22:09 |
| popey | I'll see how far he gets, the good news is all his stuff is open | 22:09 |
| popey | and runs on android, ios and others already | 22:10 |
| kgunn | oh yeah...in that case, we're more like android/ios | 22:10 |
| kgunn | i'm simplifying...but just saying, we're a little hesitant to add x extension if they're not really needed | 22:11 |
| kgunn | popey: no matter what i can't seem to start the app from the command line, is there some precondition i'm missing? | 22:20 |
| popey | kgunn: not that I'm aware of | 22:22 |
| popey | whats the full output of you running the command? | 22:22 |
| kgunn | popey: phablet@ubuntu-phablet:~/.local/share/applications$ start application APP_ID=com.ubuntu.calendar_calendar_0.4.393.desktop | 22:23 |
| kgunn | start: Job failed to start | 22:23 |
| popey | ah no | 22:26 |
| popey | omit .desktop | 22:26 |
| popey | start application APP_ID=com.ubuntu.calendar_calendar_0.4.393 | 22:26 |
| kgunn | ta | 22:26 |
| popey | do that | 22:26 |
| kgunn | yep | 22:26 |
| kgunn | popey: ok to use phablet-screenshot for that portion ? | 22:29 |
| pdo_fn14 | I think HUD doesn't work properly for another language, it's only designed for English. | 23:27 |
| pdo_fn14_ | Sorry suddenly disconnected. | 23:30 |
| pdo_fn14_ | Give a try With Trusty and Utopic still same problem who can't fulfilled menus in some query non-english language. | 23:33 |
| pdo_fn14_ | Example, try to type "inf" when rhythmbox playing music and Using Indonesian as Primary Ubuntu Language, it's won't show "Info Program". | 23:35 |
| pdo_fn14_ | Trying to search these Unity Bug in LP, but no result. | 23:36 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!