=== _salem is now known as salem_ | ||
=== salem_ is now known as _salem | ||
Cimi | Saviq, you know what is blocking input events under the osk in unity? | 07:47 |
---|---|---|
Cimi | otherwise I'll try to dig | 07:47 |
Saviq | Cimi, the InputArea in OSK | 07:48 |
Saviq | in OSKController that is | 07:48 |
Cimi | Saviq, so why doesn't work in wizard? | 07:48 |
Saviq | Cimi, did you even look in OSKController? did you check what doesn't work? | 07:50 |
Saviq | mzanetti, please keep an eye on © dates, in new files | 07:51 |
Saviq | there's 2012,2013 in the lights plugin.h | 07:51 |
mzanetti | meh... | 07:52 |
Cimi | Saviq, I will try to look for bugs | 07:52 |
mzanetti | I told hom twice | 07:52 |
Cimi | Saviq, anyway in Shell.qml we have deprecated code then | 07:52 |
Cimi | Saviq, the MouseArea responding to ApplicationManager.keyboardVisible can be removed | 07:52 |
Saviq | Cimi, no it can't | 07:52 |
Saviq | Cimi, that one blocks input *within* the shell | 07:53 |
Saviq | Cimi, so now that you mention this | 07:53 |
Cimi | Saviq, ApplicationManager.keyboardVisible does not exist | 07:53 |
Saviq | Cimi, ah ok | 07:53 |
Saviq | Cimi, OSKC took that over | 07:53 |
Saviq | Cimi, what do you mean "you'll try to look for bugs"? | 07:54 |
Cimi | Saviq, in the OSKController | 07:55 |
Saviq | k | 07:55 |
Cimi | is it possible that in the wizard, the surface does not change to MirSurface.Maximized? | 07:57 |
Saviq | Cimi, you should close your IRC client when working | 07:57 |
Cimi | Saviq, it's a valid question because I don't know mir | 07:58 |
Cimi | neither OSK | 07:58 |
Saviq | Cimi, it's easy to check, too | 07:59 |
Saviq | Cimi, it's just one console.debug line in that file | 07:59 |
Saviq | Cimi, which you can even edit on device | 07:59 |
Cimi | Saviq, well the answer is yes, don't have to debug that | 07:59 |
Cimi | Saviq, my question was more why this happens | 07:59 |
Cimi | just on the wizard | 07:59 |
Saviq | Cimi, you typed more to ask the question than to check yourself | 07:59 |
Saviq | mhr3, something dawned on me... how do we deal with dark/light foreground color (on cards with background)... | 08:00 |
Cimi | start reading mir code? :) | 08:00 |
Saviq | Cimi, no, check what happens in OSKC | 08:01 |
Saviq | Cimi, whether everything there behaves as expected | 08:01 |
Saviq | Cimi, if not, you'll at least know which part is responsible | 08:01 |
Cimi | I checked, this is the only thing that enables the inputarea | 08:01 |
Saviq | Cimi, how about its size | 08:02 |
mhr3 | Saviq, hm, scope can specify the overlay background as well, can§t it_ so it§s problem | 08:02 |
mhr3 | eh | 08:02 |
Saviq | mhr3, not overlay | 08:02 |
Saviq | mhr3, background | 08:02 |
Saviq | mhr3, like in scopes | 08:02 |
Cimi | that could be as well indeed | 08:02 |
Saviq | mhr3, and well, same for overlay colour, once we allow different overlay colours, we'll need to make the "dark or light" decision on the colour as well | 08:03 |
Saviq | mhr3, and this decision assumes that foreground colour is dark, and goes to white if background is dark | 08:04 |
Saviq | mhr3, but if foreground colour is light (because scope background is dark)? | 08:04 |
Saviq | we *could* check the foreground colour and fall back to white/grey accordingly | 08:05 |
Saviq | just wondering if we should supply both dark and light foreground colours | 08:05 |
Saviq | (or scope should) | 08:05 |
Saviq | mhr3, it's probably not even a question to you ;) | 08:06 |
mhr3 | Saviq, eh, give me a moment, you love pinging me during our morning meeting :) | 08:06 |
Saviq | mhr3, lol | 08:06 |
mzanetti | Saviq: re: Text field itself, back and clear buttons should be exempt from tap-to-dismiss behaviour of the popup and activate instantly. | 08:50 |
mzanetti | Saviq: how would that work? The Popover has the InverseMouseArea integrated | 08:50 |
mhr3 | Saviq, so seems you know what has to be done ultimately :) | 08:53 |
mhr3 | Saviq, glad to be your debugging duck ;) | 08:54 |
Saviq | mhr3, you mean "to ask design"? ;) | 08:54 |
Saviq | mzanetti, yeah, but I thought you could tweak it | 08:54 |
Saviq | mzanetti, we did that before | 08:54 |
mzanetti | hmm... I just copied the Popover as it was in the previous code | 08:55 |
Saviq | mzanetti, hmm seems like we lost the code | 08:55 |
Saviq | mzanetti, because the behaviour change | 08:55 |
Saviq | d | 08:55 |
Saviq | mzanetti, please check with timp whether that's possible | 08:56 |
Saviq | mzanetti, I think what we had is we had our own IMA covering more than just the popover... and maybe disabled the popover IMA? | 08:56 |
Saviq | mzanetti, don't remember now, but with the new behaviour (history as type-ahead search), we'll need that badly | 08:57 |
dednick | Saviq: could you do a an update for silo 15? lp:~nick-dedekind/unity-mir/trusted-sessions | 08:59 |
Saviq | dednick, ah you just need a rebuild? | 09:03 |
dednick | Saviq: if that pulls down the latest from the branch. | 09:04 |
Saviq | dednick, yup, it does, kicked | 09:04 |
dednick | Saviq: thanks | 09:04 |
=== vila_ is now known as vila | ||
mzanetti | Saviq: you really wanna get rid of the clear history button? | 09:10 |
Saviq | mzanetti, that's not a "clear history" button, just a "close history" one | 09:10 |
mzanetti | oh right... | 09:11 |
mzanetti | dropping | 09:11 |
=== shiznix_ is now known as shiznix | ||
mhr3 | mzanetti, you had the uitk pkgs with the necessary header changes? | 09:24 |
mzanetti | mhr3: one sec | 09:24 |
mzanetti | [15:38] <mzanetti> https://chinstrap.canonical.com/~mzanetti/qtdeclarative5-ubuntu-ui-toolkit-plugin_0.1.47+14.10.20140618-0ubuntu1_amd64.deb | 09:24 |
mzanetti | [15:38] <mzanetti> https://chinstrap.canonical.com/~mzanetti/ubuntu-ui-toolkit-theme_0.1.47+14.10.20140618-0ubuntu1_amd64.deb | 09:24 |
mzanetti | [15:38] * mzanetti hopes you use amd64 | 09:24 |
mzanetti | mhr3: ^ | 09:24 |
mhr3 | thx | 09:25 |
* mhr3 just wgetting some html :P | 09:26 | |
Cimi | Saviq, so UbuntuKeyboardInfo does not connect to the socket | 09:43 |
Saviq | Cimi, do you see error messages on its console? | 09:43 |
Cimi | "QLocalSocket::connectToServer: Connection refused" | 09:43 |
Saviq | Cimi, and is the keyboard launched along with the wizard or somewhen later? | 09:44 |
Cimi | Saviq, I see this after creating the mir surface | 09:44 |
Cimi | so I'd say at wizard startup | 09:45 |
Saviq | Cimi, no, you see this after instantiating OSKController | 09:45 |
Cimi | at wizard startup | 09:45 |
Saviq | Cimi, which uses UbuntuKeyboardInfo | 09:45 |
Cimi | since it's in main.qml | 09:45 |
Saviq | Cimi, but that doesn't answer my question | 09:45 |
Saviq | Cimi, it's the keyboard that creates the socket | 09:45 |
Cimi | it is launched in main.qml | 09:45 |
Cimi | like, OSKController is in the main.qml | 09:45 |
Saviq | Cimi, maliit is launched in main.qml? | 09:45 |
Cimi | no in the upstart | 09:46 |
Cimi | let me see when | 09:46 |
Saviq | Cimi, that was my question | 09:46 |
Cimi | Saviq, in post-start | 09:46 |
Cimi | might move to pre-start | 09:46 |
Saviq | Cimi, no you can't | 09:47 |
Cimi | ok | 09:47 |
Saviq | Cimi, as it will die 'cause it couldn't connect to mir | 09:47 |
Saviq | Cimi, you need to check if the keyboard creates the socket | 09:48 |
* Cimi builds unity-mir | 09:50 | |
Saviq | why? | 09:51 |
Saviq | Cimi, it should be in /run/user/32011/ | 09:51 |
Saviq | /run/user/32011/ubuntu-keyboard-info | 09:51 |
Saviq | if it's not there, but is in /tmp/ubuntu-keyboard-info | 09:51 |
Saviq | means maliit doesn't get XDG_RUNTIME_DIR set in time | 09:52 |
Cimi | Saviq, to debug | 09:52 |
Cimi | Saviq, it's in run/user | 09:52 |
Saviq | Cimi, I'm not getting the keyboard error messages in ~/.cache/upstart/ubuntu-system-settings-wizard.log | 09:53 |
Cimi | Saviq, grep -rn UbuntuKeyboardInfo /home/phablet/.cache/upstart/ | 09:54 |
Cimi | nothing? | 09:55 |
Saviq | Cimi, no | 09:55 |
Cimi | ugh | 09:55 |
Cimi | Saviq, so you don't have this? https://bugs.launchpad.net/ubuntu-welcome-wizard/+bug/1334429 | 09:55 |
ubot5 | Ubuntu bug 1334429 in Ubuntu Welcome Wizard "touch event allowed through OSK to what is underneath" [High,New] | 09:55 |
Saviq | Cimi, wizard just crashed for me when I tried this | 09:56 |
Cimi | Saviq, that was another bug | 09:56 |
Cimi | Saviq, https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1334203 | 09:57 |
ubot5 | Ubuntu bug 1334203 in ubuntu-system-settings (Ubuntu) "ubuntu-system-settings-wizard crashes/freezes when using the back button" [High,New] | 09:57 |
Saviq | Cimi, yeah, so seems I do have the bug, if the back button is causing it | 09:58 |
Cimi | Saviq, which means the events pass through | 09:58 |
Saviq | Cimi, yes, but the cause looks different than what you found | 09:58 |
Cimi | Saviq, in my case, UbuntuKeyboardInfo width and height is 0 | 09:59 |
Saviq | Cimi, I got one 'UbuntuKeyboardInfo - socket error: "QLocalSocket::connectToServer: Connection refused"' now | 10:02 |
Saviq | Cimi, but it retries | 10:02 |
Saviq | Cimi, and connects fine again | 10:02 |
Cimi | Saviq, if you want to add console.log("oskSurface state changed", ubuntuKeyboardInfo.height, ubuntuKeyboardInfo.width); | 10:02 |
Cimi | in /usr/lib/arm-linux-gnueabihf/unity8/qml/Unity/Application/OSKController.qml | 10:02 |
Cimi | at the last connection near end of file | 10:02 |
Saviq | Cimi, you'd have more of those warnings, and then an ultimate "failed to connect" if it really failed to connect | 10:03 |
Cimi | ah ok | 10:03 |
Cimi | so different indeed here | 10:03 |
Saviq | Cimi, btw, wizard-has-run is created too early IMO | 10:10 |
Saviq | Cimi, it should only be created after the wizard completed | 10:10 |
Cimi | Saviq, it is | 10:10 |
Cimi | Saviq, post-stop | 10:10 |
Saviq | Cimi, does it check the exit code? | 10:10 |
Cimi | no | 10:10 |
Saviq | Cimi, it should | 10:11 |
Cimi | it's in upstart | 10:11 |
Cimi | Saviq, what about wizard keeps crashing and you cannot boot? | 10:11 |
Cimi | I understand both ways | 10:11 |
Saviq | Cimi, a bit extreme, but maybe | 10:11 |
Cimi | Saviq, we can echo the exit code in the wizard-has-run | 10:12 |
Saviq | Cimi, oh and also, SIM page should probably be optional on devices where there isn't a modem at all | 10:13 |
Cimi | Saviq, and like redo the wizard at boot if contains just 1 invalid error code | 10:13 |
Cimi | Saviq, I know this | 10:13 |
Saviq | k | 10:13 |
Cimi | Saviq, but need to hack on ofono for that | 10:13 |
Cimi | (and add api to expose number of modems) | 10:14 |
mhr3 | Saviq, once 004 finishes migrating, am i green to rebuild and land 005? or do you have more landings and would like to combine them? | 10:15 |
Saviq | mhr3, maybe we should indeed | 10:16 |
mhr3 | Saviq, fine with me, it's nicely isolated change | 10:16 |
Saviq | Cimi, I added some qml debugs and everything works as expected | 10:17 |
Saviq | in OSKController at least | 10:18 |
mhr3 | Saviq, so, feel free to add more mps to 005... or if you want to merge it with like 009, then perhaps the other way around | 10:19 |
Cimi | Saviq, UbuntuKeyboardInfo is still 0 0 width height | 10:19 |
Saviq | Cimi, here's my OSKController | 10:19 |
Saviq | http://paste.ubuntu.com/7741264/ | 10:19 |
Cimi | Saviq, and what does it print for you? | 10:20 |
Saviq | Cimi, http://pastebin.ubuntu.com/7741273/ | 10:21 |
Saviq | dednick, btw the package built in silo 15 | 10:21 |
dednick | Saviq: thanks! | 10:21 |
=== zequence_ is now known as zequence | ||
Saviq | mzanetti, .length will complain if imageSource is not a string | 10:30 |
Saviq | mzanetti, result the same, but less warnings | 10:31 |
mzanetti | mhm... so given that imageSource is defined as property string... is there a difference? | 10:31 |
Saviq | hmm right | 10:31 |
Saviq | mzanetti, probably no diff, maybe better readable | 10:32 |
mzanetti | no problem anyways | 10:32 |
mzanetti | yeah. imo the .length is the more readable one... | 10:32 |
mzanetti | but really no strong opinion | 10:32 |
mzanetti | was just curious if I really miss a technical detail | 10:32 |
Saviq | mzanetti, I probably come from python, where you rely on boolean infers a lot | 10:32 |
mzanetti | yeah, I don't like that at all... but again, no problem at all. I can deal with both | 10:33 |
Cimi | Saviq, it is not positioned correctly | 10:33 |
Cimi | Saviq, it seems like it has space at the bottom | 10:33 |
Saviq | Cimi, panel height maybe? | 10:33 |
Cimi | more | 10:33 |
Cimi | I think | 10:33 |
Cimi | let me see in unity | 10:33 |
Cimi | Saviq, I put a nice Rectangle on the input area :) | 10:33 |
Saviq | Cimi, yup | 10:34 |
Cimi | Saviq, and I disccovered a bug in unity :) | 10:36 |
Cimi | Saviq, the very bottom edge of the keyboard passes through | 10:36 |
Cimi | in unity as well | 10:36 |
Saviq | Cimi, you can pat yourself on the back, well done | 10:36 |
Saviq | ;P | 10:37 |
Cimi | it is just not as bas as the wizard | 10:37 |
Cimi | Saviq, I tried moving a dash app at the bottom of the screen, like gallery app | 10:37 |
Cimi | Saviq, open a search | 10:37 |
Saviq | Cimi, you don't have to explain the bug to me | 10:37 |
Cimi | and tap between the spacebar and the edge | 10:37 |
Cimi | ok | 10:37 |
Saviq | Cimi, either file it as a bug or, better yet, fix it! | 10:38 |
Cimi | Saviq, I need some thoughts on this | 10:38 |
Cimi | so let me ask | 10:38 |
Cimi | the input area should be taller than the keyboard, otherwise we cannot dismiss by swiping down, right? | 10:39 |
mzanetti | gosh... we have a very bad mix of onClicked, onTriggered, onPressed | 10:42 |
mzanetti | all doing the same thing | 10:42 |
Saviq | Cimi, not sure, I think the keyboard only has something like 1gu near the top | 10:43 |
Saviq | Cimi, doesn't actually go out of the keyboard rectangle | 10:43 |
Cimi | I see those info coming from ubuntu-keyboard | 10:44 |
Cimi | on portrait, it is misplaced of 1 panel height up | 10:44 |
Cimi | (it is roughly 1 panel height) | 10:44 |
Cimi | on landscape, is fine | 10:44 |
Cimi | before digging, I need to understand one thing | 10:45 |
Cimi | - is this filter area supposed to extend also to the dismiss area on top of the keyboard? | 10:45 |
Cimi | (I think so, but currently it is not) | 10:46 |
Cimi | because uses UbuntuKeyboardInfo that has keyboard dimensions | 10:46 |
Cimi | ubuntu-keyboard provides the area, now there could be a bug there or in mir itself | 10:47 |
mzanetti | Saviq: look at that :D | 10:49 |
Saviq | mzanetti, k, one minor thing in new-header | 10:49 |
mzanetti | if (searchTextField.text) { | 10:49 |
mzanetti | this doesn't work ^^ | 10:49 |
mzanetti | while this does: if (searchTextField.text.length == 0) { | 10:49 |
Saviq | mzanetti, you're missing a ! | 10:49 |
mzanetti | lol | 10:49 |
Saviq | if you're comparing to 0 | 10:49 |
mzanetti | I guess I need a break | 10:50 |
mzanetti | btw. still fixing the non-inline comments | 10:50 |
Saviq | mhm | 10:50 |
Saviq | mzanetti, the dash failures are weird, they seem to come and go on jenkins | 10:50 |
mzanetti | yeah, will look into that | 10:51 |
Saviq | Cimi, I don't think there's a dismiss area "on top" of the keyboard | 10:53 |
Saviq | Cimi, there's only the 1GU padding you have above the keys | 10:53 |
Cimi | Saviq, ok right indeed | 10:54 |
Cimi | Saviq, so the y in the socket is wrong in portrait | 10:54 |
Saviq | Cimi, here's the MP that dealt with that | 10:55 |
Saviq | https://code.launchpad.net/~dandrader/ubuntu-keyboard/improve_kbd_info_ipc/+merge/190418 | 10:55 |
Cimi | I'm looking at ubuntu-keyboard | 10:55 |
Saviq | Cimi, not sure I can confirm your findings, though - I added a rectangle here, too, and it covers the whole bottom of the screen correctly | 10:56 |
Cimi | Saviq, in portrait? | 10:56 |
Saviq | Cimi, yes | 10:56 |
Cimi | on mako> | 10:56 |
Cimi | not here | 10:56 |
Saviq | Cimi, on flo actually, let me try on mako | 10:56 |
Saviq | Cimi, looks fine on mako, too | 10:58 |
Saviq | Cimi, maybe you're seeing http://bazaar.launchpad.net/~mir-team/unity-mir/trunk/revision/238 ? | 10:59 |
Saviq | Cimi, or well I meant https://code.launchpad.net/~afrantzis/unity-mir/fix-1332624-input-area/+merge/224768 | 10:59 |
Saviq | Cimi, that was released yesterday | 11:00 |
Saviq | Cimi, https://bugs.launchpad.net/unity-mir/+bug/1332624 | 11:00 |
ubot5 | Ubuntu bug 1332624 in unity-mir "cant' swipe the keyboard away anymore" [Critical,Fix released] | 11:00 |
Cimi | let me upgrade | 11:00 |
Saviq | Cimi, hah | 11:02 |
Saviq | Cimi, it moved after typing | 11:02 |
Saviq | Cimi, and yeah, it's panel-height | 11:04 |
Cimi | it was moving immediately for me | 11:05 |
Cimi | anyway | 11:05 |
Cimi | doesn't matter | 11:05 |
Cimi | in wizard is like double panel height | 11:05 |
Saviq | Cimi, was one panel height for me | 11:05 |
Cimi | that's why it is possible to tap on back button underneath | 11:05 |
Cimi | Saviq, weird - with the today image I have no bug | 11:06 |
Cimi | let me see in wizard | 11:06 |
Cimi | Saviq, did the rectangle move after typing in unity8? | 11:06 |
Saviq | Cimi, no | 11:06 |
Cimi | Saviq, unity8 is fine for me with 111 image | 11:07 |
Saviq | Cimi, yeah, so that but above | 11:07 |
Cimi | Saviq, wizard is 1 panel height shifted | 11:07 |
Saviq | Cimi, while in wizard panel isn't taken into account | 11:08 |
Cimi | but what is setting the panel | 11:08 |
Saviq | unity-mir | 11:08 |
Saviq | but OSK surface should be fullscreen | 11:08 |
Saviq | panel shouldn't matter | 11:08 |
Cimi | status is maximised | 11:09 |
Cimi | not fullscreen | 11:09 |
Saviq | yeah | 11:09 |
Saviq | Cimi, | 11:09 |
Saviq | there's an explicit topMargin in OSKController in Shell.qml | 11:09 |
Saviq | Cimi, that should've been a hint | 11:10 |
Cimi | Saviq, ok yeah sounds like it | 11:11 |
Cimi | Saviq, it is hardcoded or what? | 11:11 |
Saviq | Cimi, yes it is | 11:11 |
Cimi | in unity mir? | 11:11 |
Cimi | I could get the panel height from ApplicationManager in Unity.Application | 11:11 |
Saviq | Cimi, there, and in Panel.qml | 11:11 |
Cimi | (maybe) | 11:12 |
Cimi | but wizard should not look into unity8 | 11:12 |
Saviq | Cimi, panelHeight: units.gu(3) + units.dp(2) | 11:12 |
Saviq | Cimi, with a FIXME that this shouldn't be hardcoded | 11:12 |
Cimi | I can copy that as topMargin? ;\ | 11:12 |
Saviq | Cimi, yeah | 11:12 |
Saviq | Cimi, I agree OSK should communicate global coords | 11:13 |
Saviq | Cimi, and that's worth a bug | 11:13 |
Saviq | Cimi, but a quick fix is just to copy that | 11:13 |
=== MacSlow is now known as MacSlow|lunch | ||
Cimi | Saviq, thanks | 11:13 |
mhr3 | Saviq, wrt SettingsModelInterface, shouldn't it have a isEmpty / rowCount prop? so you can do showSettingsButton: !scope.settingsModel.isEmpty | 11:17 |
Saviq | mhr3, yeah, count, forgot that's not Q_PROPERTY by default... | 11:18 |
mhr3 | Saviq, or do you consider scope.settingsModel being null enough? | 11:18 |
Saviq | mhr3, I'd probably like both | 11:18 |
Cimi | Saviq, so where is the bug then? unity mir or ubuntu keyboard?> | 11:19 |
Saviq | Cimi, keyboard | 11:19 |
Saviq | Cimi, and unity8 | 11:19 |
Cimi | unity8?? | 11:19 |
facundobatista | Holas | 11:24 |
Saviq | o/ | 11:26 |
Saviq | Cimi, unity8 should not have the margin applied there | 11:26 |
Cimi | ah sure | 11:26 |
Cimi | yes | 11:26 |
Saviq | /food | 11:26 |
* mzanetti thanks init for cleaning up his window mess | 11:27 | |
=== alan_g is now known as alan_g|lunch | ||
=== MacSlow|lunch is now known as MacSlow | ||
=== _salem is now known as salem_ | ||
Saviq | mzanetti, remember we'll need to bump the UITK dependency | 12:53 |
Saviq | mzanetti, but we need to wait until silo 4 lands & merges finally | 12:54 |
mhr3 | Saviq, are there plans to resurrect grid-see-more branch? i'd kinda need that so we can have the new section header queries | 12:58 |
Saviq | mhr3, yeah, there are | 13:00 |
Saviq | mhr3, the string needs to come from scope, though | 13:00 |
Saviq | mhr3, and we need to add a sticky footer | 13:00 |
mhr3 | Saviq, what string? | 13:00 |
Saviq | mhr3, "show more" / "show all" | 13:00 |
mhr3 | oh? | 13:00 |
mhr3 | why? | 13:00 |
greyback_ | l | 13:01 |
greyback_ | f*cking multi-monitor | 13:01 |
Saviq | mhr3, because it's meant to be "show more" or "show all", depending on what it will actually do | 13:01 |
Saviq | greyback_, ah so that's why we know your password? | 13:01 |
mhr3 | Saviq, it will be just all for now, no need for more imo | 13:01 |
mhr3 | Saviq, unless you decide to chunk it | 13:01 |
Saviq | mhr3, tell Mike that | 13:01 |
greyback_ | Saviq: you know an old password... | 13:01 |
Saviq | mhr3, no no, it's more because it can't display all | 13:01 |
Saviq | greyback_, I know I know | 13:01 |
Saviq | mhr3, not because pressing it will display less than show all would | 13:02 |
greyback_ | gnome-terminal's cursor flashes even when its window does not have focus. So annoying | 13:02 |
Saviq | greyback_, hmm? | 13:02 |
Saviq | greyback_, not here it doesn't | 13:02 |
greyback_ | does here. | 13:02 |
Saviq | greyback_, I get a hollow rectangle when unfocused | 13:02 |
* greyback_ digs into settings | 13:03 | |
Saviq | greyback_, I'm using byobu + bash, maybe zsh doesn't deal with that well? | 13:03 |
greyback_ | Saviq: maybe | 13:03 |
mhr3 | Saviq, afaik there are no plans to be able to do offsetted queries, so i'm not sure now | 13:03 |
Saviq | greyback_, same behaviour without byobu | 13:03 |
Saviq | mhr3, again, it's not about what we can display in dash | 13:04 |
Saviq | mhr3, it's about what we can't | 13:04 |
greyback_ | Saviq: for me same bad behaviour with bash | 13:04 |
Saviq | mhr3, if the max you can display in dash is not "all" (because of, say, API), then "show more" is meant to be shown | 13:04 |
Saviq | mhr3, let's confirm with Mike | 13:05 |
mhr3 | Saviq, yea, that's the first time i hear this tbh | 13:05 |
mhr3 | Saviq, in spec? | 13:05 |
Saviq | mhr3, yeah... no | 13:05 |
Saviq | mhr3, well, in multiple specs | 13:05 |
mhr3 | sigh | 13:05 |
Saviq | mhr3, some say "more", some say "all" ;) | 13:05 |
=== alan_g|lunch is now known as alan_g | ||
Saviq | NOW MIGRATE ALREADY! | 13:14 |
* Saviq can die now | 13:22 | |
Saviq | dednick, please merge trunk into lp:~nick-dedekind/unity8/move-indicator-qml | 13:36 |
dednick | Saviq: doing | 13:39 |
Saviq | dednick, thanks | 13:39 |
dednick | Saviq: done | 13:48 |
Saviq | dednick, tx | 13:49 |
mhr3 | Saviq, is u8 trunk not merged with ubuntu? | 13:52 |
Saviq | mhr3, it's not | 13:53 |
Saviq | mhr3, I'll prep an MP shortly | 13:53 |
Saviq | mhr3, why? | 13:53 |
mhr3 | landing bot complaining | 13:53 |
mhr3 | https://ci-train.ubuntu.com/job/landing-013-1-build/55/console | 13:53 |
Saviq | mhr3, ah you're landing that separately? | 13:54 |
Saviq | mhr3, I added it to my silo | 13:54 |
Saviq | mhr3, lemme prep the mP | 13:54 |
mhr3 | Saviq, imo should just push the fix to trunk, we said the two should be always in sync | 13:57 |
Saviq | mhr3, yeah ok, doing | 13:57 |
mhr3 | Saviq, when was the last time you just pushed to trunk anyway, doesn't it sound exciting? :) | 13:58 |
=== pete-woods is now known as pete-woods-lunch | ||
Saviq | mhr3, done | 13:59 |
mhr3 | ty | 13:59 |
greyback_ | Cimi: hey, where does the welcome wizard print its logs? I can't find it in .cache/upstart or /var/log/upstart | 14:05 |
Cimi | greyback_, ~/.cache/upstart/ubuntu-system-settings-wizard.log | 14:14 |
greyback_ | Cimi: ah, forgot the ubuntu- prefix. thanks | 14:15 |
greyback_ | l | 14:15 |
Cimi | greyback_, what are you looking? | 14:16 |
greyback_ | Cimi: welcome wiaard needs some changes to make compatible with QtCOmp | 14:16 |
greyback_ | Cimi: I've them done, just wanted to check things are ok | 14:16 |
greyback_ | Cimi: I'm not seeing any wifi networks in the list, any idea how I could debug? | 14:17 |
Cimi | greyback_, we had this issue once, was due to initialisation errors of the indicators | 14:18 |
Cimi | maybe mterry knows | 14:18 |
mterry | Cimi, I don't recall *why* it was happening -- I believe it went away for us by starting indicators differently -- instead of just the network-indicator, we did the general indicator startup | 14:18 |
Cimi | greyback_, do you hae wireless in Ireland? | 14:19 |
Saviq | lol | 14:21 |
mhr3 | Saviq, could you jump into the meeting 5 minutes earlier? got few Qs | 14:27 |
mhr3 | the one in 30ish minutes | 14:27 |
Saviq | mhr3, prolly yes | 14:31 |
=== salem_ is now known as _salem | ||
=== alan_g is now known as alan_g|tea | ||
kgunn | greyback_: did the wifi list show up after you touched it ? | 14:40 |
greyback_ | kgunn: reboot fixed it | 14:40 |
kgunn | i saw the same bug last week...any interaction populated the list | 14:40 |
kgunn | ah | 14:40 |
greyback_ | Cimi: mterry: ever seen this crash? http://pastebin.ubuntu.com/7742229/ - I reproduce reliably with my code, when I go to list of networks (one already selected), then hit back | 14:42 |
mterry | greyback_, someone filed a bug with that, yea | 14:42 |
greyback_ | mterry: ok | 14:42 |
mterry | greyback_, haven't looked at it yet | 14:43 |
seb128 | somebody needs to do a landing with https://code.launchpad.net/~nick-dedekind/qmenumodel/unitymenuaction.lp1334203/+merge/225063 | 14:43 |
seb128 | mterry, ^ I think that might be the fix for that issue | 14:43 |
seb128 | https://bugs.launchpad.net/qmenumodel/+bug/1334203 | 14:43 |
ubot5 | Ubuntu bug 1334203 in ubuntu-system-settings (Ubuntu) "ubuntu-system-settings-wizard crashes/freezes when using the back button" [High,New] | 14:43 |
greyback_ | mterry: no worries, just checking it wasn't consequence of my changes | 14:43 |
mterry | ah indeed | 14:43 |
mterry | seb128, thanks | 14:43 |
seb128 | who is lander for qmenumodel? | 14:43 |
mterry | dunno | 14:43 |
seb128 | thostr_, ? | 14:44 |
=== alan_g|tea is now known as alan_g | ||
thostr_ | seb128: yeah, let us have a look at it... | 14:46 |
seb128 | thostr_, thanks, https://code.launchpad.net/~charlesk/qmenumodel/lp-1282282-dangling-UnityMenuAction-pointer/+merge/211211 is also approved/would be good to land | 14:47 |
Cimi | greyback_, yes that is the fix | 14:53 |
Cimi | https://code.launchpad.net/~nick-dedekind/qmenumodel/unitymenuaction.lp1334203/+merge/225063 | 14:53 |
greyback_ | Cimi: good to know | 14:54 |
Cimi | greyback_, I am not aware of any issue atm | 14:54 |
Saviq | mhr3, ready when you are | 14:54 |
mhall119 | I don't like the Apps scope changes, why did we split the "available" functionality to a different screen? | 15:03 |
Saviq | mikenagle, ↑ | 15:03 |
=== _salem is now known as salem_ | ||
greyback_ | Saviq: suru icons landed? | 15:12 |
=== greyback_ is now known as greyback|post | ||
Saviq | greyback|post, yes | 15:14 |
=== pete-woods-lunch is now known as pete-woods | ||
pete-woods | Saviq, Cimi: hey guys, just checking what the status on the infographics is? | 15:18 |
pete-woods | I tested it (to success) last week, but wasn't sure if there were more changes coming | 15:18 |
Cimi | pete-woods, we're waiting a release of the sdk | 15:19 |
pete-woods | Cimi: okay, cool, thanks for the infos! :) | 15:19 |
pete-woods | presumably that's for your fixed cross fade image thingy | 15:19 |
Cimi | y | 15:20 |
Cimi | actually branching right now | 15:21 |
Cimi | argh got infected by tags | 15:32 |
Cimi | fixing | 15:32 |
Cimi | Saviq, I updated the infographics branch with merge from trunk and removed crossfadeimage | 15:35 |
Saviq | Cimi, thanks, I'll rebuild the silo after the unity8 landing | 15:36 |
Cimi | Saviq, I am on holiday from tuesday afternoon, so need fix until monday evening pls :P | 15:37 |
Saviq | Cimi, why is there only 4 dots at the bottom of wizard when there's 5 pages? | 15:37 |
Saviq | Cimi, yeah, I plan to land this asap | 15:37 |
Cimi | will have a look at the wizard in a bit | 15:38 |
sil2100 | Saviq, rsalveti, seb128: what do you think of such output? https://ci-train.ubuntu.com/job/landing-000-1-build/7/console | 15:39 |
sil2100 | I had problems with finding the right wording, so I invented the whole 'twin package' thingy | 15:39 |
dandrader | Saviq, mzanetti: trying out latest trunk. I'm unable to drag in the launcher (left-edge drag). do you guys experience the same? | 15:39 |
seb128 | sil2100, looks good to me | 15:39 |
sil2100 | You can override this of course | 15:40 |
=== gatox is now known as gatox_lunch | ||
seb128 | sil2100, you can directly use ""their -gles counter-parts" instead if you want to avoid new vocabulary there though | 15:40 |
rsalveti | sil2100: sounds good | 15:40 |
rsalveti | yeah | 15:40 |
dandrader | Saviq, mzanetti right edge also broken for me | 15:40 |
sil2100 | Yeah, I wanted to get ready for the future, in case we have new cases of such duplicated packages | 15:41 |
sil2100 | Anyway, we can change the wording as we want even later | 15:41 |
Saviq | dandrader, sounds like your edges got stuck | 15:41 |
Saviq | dandrader, works fine here | 15:41 |
* dandrader reboots his N4 | 15:41 | |
dandrader | hmm, I though "sudo restart lightdm" would be enough to get rid of any bogus state. let's see if a full reboot does it | 15:42 |
Saviq | mhall119, if it's gonna help you, the installed category is not going to be collapsible soon | 15:43 |
Saviq | mhall119, so you'll basically have a page of your installed apps, and then a button to get more | 15:43 |
dandrader | yeah. rebooting solved it. maybe a had a rogue usc running.... | 15:44 |
=== greyback|post is now known as greyback | ||
dandrader | *I had | 15:44 |
mzanetti | dandrader: hmm... haven't tried that in the last few hours | 15:44 |
dandrader | mzanetti, anyway. false alarm. rebooting did solve it. :/ | 15:45 |
mzanetti | ah ok | 15:45 |
Saviq | sil2100, I wonder if this check could only be ran on publish, and not on build? | 15:45 |
mhall119 | Saviq: I'm not sure that's what I wanted | 15:46 |
mhall119 | but, I'll wait and see | 15:46 |
Saviq | mhall119, it was a design decision to split it out | 15:46 |
mhall119 | I know it's not your decision | 15:46 |
sil2100 | Saviq: we could do that, my original plan was that as well, but then I thought that actually the earlier the better | 15:46 |
mhall119 | it was probably going to stop being useful when we have more apps anyway, it was just a nice feature | 15:46 |
mhall119 | even having it just show results in response to a search would be nice | 15:47 |
Saviq | sil2100, yeah, but that means you'll need to keep the second package updated as you rebuild | 15:50 |
Saviq | sil2100, your call really | 15:51 |
Saviq | sil2100, maybe you're right, all should be tested together | 15:51 |
sil2100 | Saviq: I'm not entirely sure myself, it's easy to move it one way or another, so I'm open to propositions anyway ;) | 15:51 |
Saviq | mzanetti, hmm PageHeader.qml missing Themes 1.1? | 15:52 |
mzanetti | Saviq: ? | 15:57 |
Saviq | mzanetti, most recent autopilot failure in new-header branch | 15:59 |
mzanetti | Saviq: did uitk land? | 15:59 |
Saviq | mzanetti, yes, but maybe that run didn't get it yet | 16:00 |
* Saviq looks | 16:00 | |
Saviq | mzanetti, ah stupid | 16:00 |
Saviq | mzanetti, no, the new uitk is in silo | 16:00 |
mzanetti | ah ok | 16:01 |
mzanetti | phew :) | 16:01 |
mhr3 | Saviq, so how do we do the sign in button-cards? | 16:18 |
mhr3 | Saviq, feels like it should special, like new category-layout or something | 16:18 |
Saviq | mhr3, it could be a completely custom category | 16:19 |
Saviq | mhr3, like you say category-layout: "sign-in" | 16:19 |
Saviq | mhr3, and the shell takes over | 16:19 |
mhr3 | Saviq, do we just ignore the component mapping completely then? | 16:19 |
mhr3 | or pretend that it's a card with summary and title or something? | 16:20 |
Saviq | mhr3, depends, do we want to put multiple buttons in one category, or category per button? | 16:20 |
mhr3 | designs have the latter | 16:21 |
mhr3 | generalization would suggest the former :) | 16:21 |
Saviq | mhr3, well, the designs can be achieved either way | 16:21 |
Saviq | mhr3, we'd just maybe need an account type id to be able to display the logo and colours or something | 16:21 |
Saviq | mhr3, and open the right account page straight away when possible | 16:22 |
mhr3 | Saviq, *plus* we could misuse it for "This category requires internet connectivity [REFRESH]" :) | 16:22 |
Saviq | mhr3, no we couldn't | 16:22 |
Saviq | mhr3, that thing needs to be above departments | 16:22 |
mzanetti | now that's interesting... the screen is off but tapping it gives me haptics feedback in some places | 16:22 |
Saviq | mhr3, in header | 16:22 |
=== alecu is now known as alecu_ | ||
mhr3 | Saviq, it *should*, api doesn't support such hints though | 16:22 |
Saviq | mhr3, yet ;) | 16:23 |
mhr3 | just saying, would be easy fix and make it more reusable | 16:23 |
Saviq | mhr3, you mean multiple buttons per category? | 16:24 |
mhr3 | i mean thinking about the no-internet case | 16:24 |
mhr3 | will generalize the sign-in card | 16:25 |
Saviq | ah so sign-in to the internets? | 16:25 |
Saviq | mhr3, so we'll have category-layout: special | 16:26 |
mhr3 | dejavu | 16:26 |
Saviq | ;) | 16:26 |
Saviq | mhr3, at least it'd be sourced by the scope still | 16:28 |
mhr3 | Saviq, but yea, what i'm suggesting - special category-layout, and mentioning in docs that in that case of using it, you can only map summary and title? | 16:28 |
mhr3 | s/?// | 16:28 |
Saviq | mhr3, so card-layout: {sign-in,no-internets} | 16:29 |
Saviq | mhr3, and then a custom set of components | 16:29 |
Saviq | mhr3, do you have the component names hardcoded, or if we said that "account-type" is a component, would that work? | 16:29 |
mhr3 | Saviq, it's pretty much hardcoded for the json merging | 16:30 |
Saviq | mhr3, I wouldn't like to abuse current component names for this | 16:31 |
mhr3 | Saviq, well, it's also because of the model, it's roles aren't dynamic | 16:32 |
mhr3 | its* | 16:32 |
Saviq | mhr3, ah right, model | 16:32 |
Saviq | mhr3, well, they could be ;) | 16:32 |
mhr3 | doesn't mean they should :) | 16:32 |
Saviq | mhr3, well, otherwise we'll end up saying subtitle is account type or some such :| | 16:33 |
mhr3 | Saviq, honestly, i don't find it an issue, it's still a card | 16:34 |
mhr3 | action card, but still a card | 16:34 |
mhr3 | and yes, it does have subtitle | 16:34 |
Saviq | mhr3, no but I mean | 16:35 |
mhr3 | doesn't matter that the subtitle will be used for account type | 16:35 |
Saviq | mhr3, that we'd end up with a account type id (like facebook, xmpp etc., for accounts consumption) | 16:35 |
Saviq | mhr3, in some random component, just because it's not used | 16:35 |
=== gatox_lunch is now known as gatox | ||
mhr3 | ah, right, what does uoa actually need? | 16:36 |
Saviq | mhr3, not sure really | 16:36 |
mhr3 | Saviq, but.. we've been already discussing that this actionable card could be maybe used even for "Location unavailable, please allow access in scope settings [Settings]" | 16:37 |
Saviq | mhr3, but I imagine account type id is a must | 16:37 |
Saviq | mhr3, yeah, well, would that be a different type of special card though? | 16:37 |
mhr3 | i'd say ideally no | 16:37 |
Saviq | I'd say yes, TBH | 16:38 |
Saviq | because they will differ in design for sure | 16:38 |
mhr3 | hm, and i'd also open doors for more misuse | 16:38 |
mhr3 | so maybe you're right | 16:38 |
Saviq | yeah | 16:38 |
Saviq | they need to be as locked down as possible | 16:38 |
=== dandrader is now known as dandrader|lunch | ||
Saviq | mhr3, I'd think the location one would be closer to how the internet one looks | 16:40 |
mhr3 | eh, more and different api :/ | 16:40 |
Saviq | mhr3, what's more, it needs to be dismissable somehow, if you forbade the scope to get access to network or location | 16:40 |
mhr3 | forbidding access to network? | 16:41 |
mhr3 | hope that's not something for rtm :) | 16:42 |
Saviq | mhr3, just thinking ahead | 16:43 |
mhr3 | Saviq, stop scaring me! | 16:43 |
Saviq | mhr3, just means that shell needs to know what type of card that is | 16:44 |
Saviq | s/card/banner/ | 16:44 |
Saviq | whatnot | 16:44 |
mhr3 | right | 16:44 |
=== alan_g is now known as alan_g|EOD | ||
=== dandrader|lunch is now known as dandrader | ||
=== salem_ is now known as _salem | ||
=== _salem is now known as salem_ | ||
Saviq | dandrader, btw, you know we have inline comments in launchpad reviews now? | 18:24 |
dandrader | Saviq, yes, but I don't like them as you cannot select a stretch of code to comment, only a specific line | 18:26 |
dandrader | Saviq, and sometimes it | 18:26 |
dandrader | it's clearer to comment on the resulting code, not the diff | 18:26 |
Saviq | dandrader, mhm, was just asking | 18:26 |
Saviq | dandrader, and I agree with the first point, wonder how the second could be solved | 18:27 |
dandrader | Saviq, it's sad that lp doesn't properly display unformatted text (i.e. code) though | 18:27 |
dandrader | Saviq, I recall I've used a review system where you could select a stretch of code and comment on it | 18:28 |
Saviq | dandrader, I think that's just a missing feature, not something that couldn't be done | 18:29 |
dandrader | ah, for the second, I think sticking with copy-paste is fine as long as the comments could display pasted code properly | 18:29 |
dandrader | inline diff comments is still in beta, right? so I also don't trust it :D | 18:30 |
dandrader | my precious comments could get lost! who knows! :) | 18:30 |
Saviq | dandrader, you'd be surprised | 18:37 |
Saviq | dandrader, they survive closed tabs, closed browsers etc. | 18:37 |
Saviq | dandrader, they are saved in LocalStorage AFAICT | 18:37 |
=== dandrader is now known as dandrader|afk | ||
=== salem_ is now known as _salem | ||
=== _salem is now known as salem_ | ||
=== salem_ is now known as _salem | ||
greyback | tedg: hey, is there a nice way to start a click app under gdb? | 20:14 |
Saviq | greyback, zbenjamin was working on a launcher that would allow that, but afaict for now | 20:15 |
Saviq | greyback, just read up the Exec line in the .desktop file | 20:15 |
greyback | Saviq: I know /how/ to do it manually. But I wondered if a nicer way existed | 20:15 |
Saviq | greyback, you probably need to do it in Path from the .desktop file as well | 20:15 |
Saviq | greyback, k, so yeah, nothing there yet, but zbenjamin's working on it | 20:16 |
Saviq | oh mhr3 got r1000 on unity8 | 20:19 |
Saviq | nasty | 20:19 |
cwayne | oh cool, we'll get a blue led now when there's a message? | 20:22 |
Saviq | cwayne, yup :) | 20:24 |
cwayne | will different notifications have different color led flashes? | 20:24 |
=== robru_ is now known as robru | ||
Saviq | foook tags | 20:45 |
Saviq | greyback, if you merged trunk, you need to strip tags in your branches | 20:47 |
=== _salem is now known as salem_ | ||
Saviq | mzanetti, you're potentially infected, too ↑ | 20:48 |
greyback | Saviq: dammit | 20:48 |
Saviq | greyback, indeed, my bad for not checking :| | 20:48 |
mzanetti | hehe | 20:49 |
* mzanetti uses a modded commit hook to print tags on every commit :) | 20:50 | |
mzanetti | should I propose that for merging? | 20:50 |
Saviq | mzanetti, that would give us false sense of security ;) | 20:51 |
Saviq | mzanetti, assuming people would actually run it | 20:51 |
mzanetti | Saviq: how about a test that simply does bzr tags | grep "?" or similar? | 20:52 |
Saviq | mzanetti, same thing | 20:52 |
mzanetti | jenkins would fail on it | 20:52 |
Saviq | mzanetti, and actually we have one ? tag | 20:52 |
Saviq | mzanetti, jenkins doesn't deal with branches | 20:52 |
mzanetti | well, I'm sure there would be a regexp that would fit | 20:52 |
Saviq | mzanetti, it just exports a tarball | 20:52 |
=== salem_ is now known as _salem | ||
Saviq | mzanetti, I even have a sh snippet to drop branches to that print out the number of tags... | 20:53 |
Saviq | and I should use it... but forget :| | 20:53 |
Saviq | mzanetti, we already have a "did you check..." in the checklist... but apparently people don't actually read the question :| | 20:54 |
mzanetti | lol, yeah... the checklist | 20:54 |
mzanetti | also the design question... | 20:54 |
mzanetti | there *never* is ux change | 20:55 |
Saviq | yeah, I'm inclined to drop the checklist, it's not really working | 20:56 |
mzanetti | well, tbh the tag question is working for me | 20:56 |
Saviq | does for me, too | 20:56 |
Saviq | that's how I found out we have them back again | 20:57 |
mzanetti | I constantly find myself to alt-tab back to the terminal and recheck the output from the commit hook in the history when I get to that question | 20:57 |
Saviq | but design and core dev review... | 20:57 |
mzanetti | yeah... | 20:57 |
Saviq | MPs are too granular for that | 20:57 |
mzanetti | yes | 20:57 |
Saviq | we get core dev review when publishing anyway | 20:57 |
Saviq | and we ask design for review here and there, too | 20:57 |
Saviq | it's gonna be better with airline I think | 20:58 |
Saviq | where we'll be able to ask for design review on a whole silo (aka ticket), not per-MP | 20:58 |
Saviq | and it'll be easier for everyone to pick a silo up | 20:59 |
mzanetti | yep | 20:59 |
cwayne | the new icons in system-settings are bitchin' | 21:04 |
Saviq | heh :|, rewrote the strip-tags script in python... down to under a second locally, but didn't really help for the remote branch | 21:31 |
Saviq | at least it's not trying to remove tags that are not tehre | 21:31 |
=== dandrader|afk is now known as dandrader | ||
=== _salem is now known as salem_ | ||
=== salem_ is now known as _salem | ||
dandrader | bug 1332598 was fixed by unity8!? I don't get it | 21:52 |
ubot5 | bug 1332598 in Unity 8 "make testShell crashes" [Critical,Fix released] https://launchpad.net/bugs/1332598 | 21:52 |
dandrader | Saviq, ^ | 21:52 |
Saviq | dandrader, https://code.launchpad.net/~aacid/unity8/no_model_reset_from_destructor/+merge/225127 | 21:53 |
Saviq | dandrader, there were three components that were involved in that (or well, were found to be broken during investigation for that bug) | 21:53 |
dandrader | Saviq, including that Qt 5.3 fix? | 21:54 |
Saviq | dandrader, yes, there were qtbase, qtdeclarative and unity8 components for that bug | 21:54 |
dandrader | ok | 21:54 |
Saviq | dandrader, the branches are linked (but bug didn't actually have the relevant affects: set) | 21:55 |
=== mterry is now known as 18VAAOQDS | ||
Saviq | oh well, 3:13 vs. 20:44... significant improvement after all... | 21:59 |
greyback | dandrader: hey, before you go, please don't forget to push your key sending branch, so we can continue working on it | 22:00 |
dandrader | greyback, sure. don't worry about it | 22:01 |
dandrader | greyback, I'm back at it now. who knows I get lucky and get it finished before my EOD? | 22:02 |
greyback | dandrader: you have holidays, go enjoy them! | 22:02 |
dandrader | says the guy that's still online at 23:03 local time :) | 22:04 |
greyback | Saviq: unity8's changelog has a rather verbose entry from Albert. Think it could be edited down | 22:04 |
greyback | dandrader: I don't have holidays. If I had, I wouldn't be online! | 22:04 |
Saviq | greyback, yeah... that's what he put in the commit message (not the first time) | 22:04 |
Saviq | greyback, we could, if you MP, I'll ACK, but can't be bothered really ;) | 22:05 |
greyback | Nor I | 22:05 |
greyback | easier to moan | 22:05 |
Saviq | @unity: rewrote the strip tags script in py, cuts down the time to 1s locally, 3mins remotely (from a minute, 20 minutes, respectively): http://people.canonical.com/~msawicz/unity8/strip-u8-tags.py | 22:09 |
18VAAOQDS | oh nice | 22:09 |
18VAAOQDS | woah | 22:10 |
=== 18VAAOQDS is now known as mterry | ||
mterry | weird | 22:10 |
Saviq | lol | 22:10 |
greyback | bregma: hey, I've hit the old problem of Mir GLES and Qt's desktop OpenGL | 22:14 |
bregma | greyback, is that Qt 5.3? | 22:15 |
greyback | bregma: your solution was to split qtubuntu into 2 separate packages, where in one you call eglBindAPI ... | 22:15 |
greyback | bregma: yes | 22:15 |
greyback | bregma: note: qtubuntu still works fine. I have a separate project which has hit the same problem | 22:15 |
bregma | I though they were doing proper runtime autodetection in Qt 5.3, but I admit I haven't been paying close attention | 22:17 |
greyback | I'm just curious if you considered other options like checking the GL string for "Mesa" - or doing a test eglBindAPI call to see if desktop GL would be accepted by the driver | 22:17 |
greyback | bregma: maybe they have. I'm only learning about this now. Should I ping rsalveti? | 22:17 |
bregma | the problem was that _both_ EGL and GL were accepted, but Qt did not accept multiple values so it defaulted to EGL | 22:18 |
bregma | er, GLES | 22:18 |
greyback | accepted by what? the driver? Where does it specify that? | 22:18 |
bregma | I'm forgetting all my EGL now, but as I recall there's a query function and it reported both GL and GLES were supported, but Qt itself uses an enumeration | 22:19 |
greyback | hmm ok, that gives me somewhere to start | 22:20 |
bregma | it was after an eglChooseConfig() call, but for the life of me I can't remember where | 22:28 |
greyback | bregma: platformsupport/eglconvenience/qeglconvenience.cpp ? | 22:32 |
bregma | greyback, yep, that's the one | 22:49 |
greyback | bregma: dynamic gl/gles switching seems to have landed in 5.3 anyway | 22:59 |
greyback | but not turned on in our packaging | 23:01 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!