[00:58] hey guys, whats the better way to do indicator applets, pygtk or pygi? [07:39] Wellark, well, yeah, it was top-approved, and I asked mterry whether it was ok to land... [08:40] Saviq, do we have anything to review? [08:41] alt nav but I thought you were doing shader? [08:41] Cimi, yeah, you still didn't get me the shader code though, anpok_'s solution didn't work... [08:41] Cimi, but there's plenty in https://code.launchpad.net/~unity-team/unity8/trunk/+activereviews [08:42] Saviq, there's only two [08:42] Cimi, huh> [08:42] "reviews I can do" [08:42] Cimi, there's also "I am not actively reviewing" [08:42] one is no lock during call [08:43] Saviq, but that means someone else is already reviewing [08:43] Cimi, yeah, tsdgeos's hijacking everything by just doing "conflicts" [08:43] Saviq, and we assign ourselves reviews for a reason [08:43] hah, he loves karma we know... [08:43] tsdgeos, I think you shouldn't put a vote up when reporting conflicts [08:44] Saviq: i'm not hijacking anything, i go back to abstrain when conflicts are fixed [08:44] tsdgeos, that still marks you as the reviewer out of "unity team" [08:44] tsdgeos, so there's no longer a request for the team to review, so it doesn't show up in "reviews I could do" [08:45] Saviq: i don't think i've marked any of those that wasn't already reviewed by someone and if i had i followed it up [08:45] https://code.launchpad.net/~nick-dedekind/unity8/indicator-polishing/+merge/228700 you [08:45] https://code.launchpad.net/~macslow/unity8/fix-1348092/+merge/228090 Daniel [08:45] https://code.launchpad.net/~unity-team/unity8/alt_nav_support/+merge/230782 Cimi [08:45] https://code.launchpad.net/~nick-dedekind/unity8/prompt-surface-model/+merge/230813 Gerry [08:46] https://code.launchpad.net/~mterry/unity8/interactive-while-locked/+merge/231253 me, haven't followed up yet [08:46] tsdgeos, I'm not blaming you :) [08:46] https://code.launchpad.net/~mterry/unity8/no-lock-during-call/+merge/227996 You [08:46] Saviq: well you are suggesting me a behaviour change when i don't have that behaviour [08:46] tsdgeos, just talking in general, I think there's no need for a Needs fixing vote when you're just seeing a conflict, is all [08:47] tsdgeos, since then you need to take it back... dunno [08:47] ok [08:47] btw why is the unlocker so ugly now? [08:47] tsdgeos, you mean it's pretty now? [08:47] [...] and it might cause confusion that you're actually reviewing [08:48] and besides being ugly is wrong [08:48] says "Enter your PIN" [08:48] when it's not my PIN [08:49] Saviq: i'll stop doing conflict comments [08:49] if you want [08:49] i thought it was useful [08:50] tsdgeos, *comments* are useful, sure [08:51] ok [08:51] tsdgeos, although I'd rather LP do them ;) [08:51] ok [08:51] dandrader, /build/buildd/unity8-8.00+14.10.20140825/tests/qmltests/Stages/tst_SurfaceContainer.qml: bad whitespace in line 144 [08:51] greyback, so, bad news redux [08:52] greyback, we had to pull initial surface size, it (seldom, but still) caused deadlocks [08:52] dang it [08:52] Saviq: now serious, is the new locker considered to be more beatiful? [08:52] Saviq: feck [08:53] that x in the middle of nowhere is totally weird [08:53] greyback, check thread 9 http://paste.ubuntu.com/8117691/ [08:53] Saviq, so how about alt nav? [08:53] Saviq, fixed [08:53] Saviq, why do we need a shader first of all? [08:53] Saviq, I'd like to understand which situation is broken for us now [08:53] tsdgeos, that's the new design, yes [08:54] tsdgeos, to be more in concert with the dialer [08:54] makes my brain unhappy [08:54] http://i.imgur.com/JORXscB.png [08:54] lost X in the space [08:55] tsdgeos, there's a corresponding ✓ when it's a SIM PIN [08:55] tsdgeos, but for device PIN there's autoconfirm [08:55] tsdgeos, and mzanetti put a greyed-out one in that case, but design was adamant that it shouldn't be there [08:56] woohoo 5.5GB free on krillin [08:56] that's much better [08:57] Saviq: ok, more cases me and design disagree then :) [08:58] tsdgeos, yup, not liking it [08:59] tsdgeos, what is this X? [08:59] tsdgeos, although I'm not sure the missymmetry itself was thought through [08:59] Cimi, cancel [08:59] it's design by discoverability [08:59] you have no clue, click on it, and then you see what it is [08:59] it's what i did :D [09:00] Saviq, tsdgeos so it should appear when I actually start typing [09:00] so I have a visual clue it must be to clear or delete the typed char [09:00] Cimi: no, that is "Cancel last keypress" and does show on top when you actually start typing [09:00] this is "cancel cancel" [09:00] i.e. go back [09:00] tsdgeos, clear all you mean [09:00] ? [09:00] no [09:01] i mean go back [09:01] go back where? [09:01] to the locker locker [09:01] instead of locker enterpin [09:01] Cimi, a bit extreme, but conceivable http://imgur.com/Xezpv3B :P [09:01] i.e. the round thing [09:01] tsdgeos, welcome screen [09:01] tsdgeos, greeter [09:01] that thing [09:02] let me setup a pin and see [09:02] btw it's *not* a pin [09:02] that's the other bug [09:02] it's a passcode [09:03] how do I add one? [09:03] if I tap on security in system settings nothing happens [09:03] yeah [09:03] tap on the thing below [09:03] ouch, tapping on "lock phone after 1 minute" reveals the passcode [09:03] wtf [09:03] yep [09:04] I would never had guess that [09:04] I thought it was to change the time [09:04] filing a bug === tvoss is now known as tvoss|test === tvoss|test is now known as tvoss [09:15] how do I screenshot? [09:15] Cimi, phablet-screenshot [09:16] tsdgeos, PIN vs. passcode is another one into the same bucket, kemmko was confident it should say PIN [09:17] ok [09:17] then make the system settings say ping [09:17] not passcode [09:17] you can't ask me for passcode in one place [09:18] and pin in another [09:18] i'll go crazy and return the phone [09:19] the tool says FACTOR to resize the image, while is not a factor but the width [09:19] tsdgeos, sure, that I agree with [09:19] phablet-screenshot that is [09:21] Cimi, huh? [09:21] Cimi, don't think that option is used tbh [09:25] https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1361127 [09:25] Ubuntu bug 1361127 in ubuntu-system-settings (Ubuntu) "Change phone security (passcode) is undiscoverable" [Undecided,New] [09:25] gonna file a suggestion for our lockscreen soon [09:26] Saviq: so what's going to happen when i have a SIM PIN and a passcode? [09:26] i get the enter PIN dialog twice in a row? [09:26] tsdgeos, or even three times if you have two SIMs [09:26] tsdgeos, at first boot, yes [09:27] tsdgeos, but it will say "SIM PIN" [09:27] tsdgeos, and ideally even the name you gave to that SIM [09:27] then I suppose your sim is unlocked for the session [09:27] ok [09:27] yeah of course [09:27] tsdgeos, Cimi, there's a twist... the X means "skip" in the case of a SIM PIN dialog :| [09:28] well, it means cancel [09:28] you can cancel sim pin [09:28] but not unlock pin [09:28] we should put it in the header [09:28] a back button like in our UI [09:30] tsdgeos, well, the result is different, with device PIN you're taken back to the lockscreen, with SIM PIN you're let through... [09:30] tsdgeos, you and me know the difference, it's not common knowledge [09:30] which would be pretty obvious if it said "Cancel" instead of "X" [09:31] tsdgeos, IMO for SIM PIN it should say "Skip", not "Cancel", if it allows you to go through [09:31] or that [09:31] what's the package that gives me the apps scope? [09:31] i can't seem to find it [09:31] tsdgeos, unity-scope-click [09:32] ah yes [09:32] that thing that wants to install urfkill [09:32] that's why i don't have it installed [09:33] it's very strange, on the phone there's 3 icons from the click scope [09:33] that don't work with the sourceSize Item i'm coding [09:33] i don't see why [09:34] greyback, I'm a bit confused about the structure of qtmir tests. They're all under "modules" and it's one directory per class. [09:34] s/class/class under test [09:34] obviously i don't get those on the desktop [09:34] in case i thought it'd be easy to debug [09:35] greyback, could we at least skip the "modules" subdir. It doesn't seem useful. at least as it is currently [09:39] Saviq: do you know where's the image://theme code? [09:39] tsdgeos, UITK [09:39] tsdgeos, themeimageprovider [09:40] ok [09:40] tx [09:40] looks like they have a bug :/ [09:40] * tsdgeos checks [09:41] Saviq: as long as this gets dealt with ASAP I'm fine: https://bugs.launchpad.net/unity8/+bug/1360703 [09:41] Ubuntu bug 1360703 in Unity 8 "internal Ubuntu.Connectivity module has a name clash with the public module" [Critical,Triaged] [09:41] I will postpone the announcement of the API [09:41] https://bugs.launchpad.net/unity8/+bug/1361132 [09:41] Ubuntu bug 1361132 in Unity 8 "New lockscreen design is not consistent with our UX patterns" [Undecided,New] [09:41] Wellark, I'm really not sure what's the rush? [09:41] Wellark, if it's unity8's private plugin, and we're not using Ubuntu.Connectivity [09:41] Wellark, what does that mean for you? [09:43] Saviq: once the public plugin is installed on the system [09:43] unity8 fails to load Shell.qml [09:44] as the public path becomes before the private one [09:44] when qmlengine is resolving it's plugin paths trying to find "Ubuntu.Connectivity" [09:44] it takes the first one [09:44] which is the public [09:44] which does not provide "0.1" [09:44] and fails [09:44] and does not keep searching if it could find another "Ubuntu.Connectivity" [09:44] that provides 0.1 [09:45] it's an import name clash [09:45] tsdgeos, hey, there's a conflict of your resetMeansCountChanged with my https://code.launchpad.net/~stolowski/unity-scopes-shell/favorite-scopes/+merge/230824 [09:45] Saviq: so as soon as qml-module-ubuntu-connectivity gets pulled to the image, unity8 stops working [09:45] so, let's fix this before the package gets seede [09:45] *seeded [09:45] tsdgeos, and i think that your change is not needed anymore; could you please take a look at my branch and confirm? [09:45] Saviq, https://bugs.launchpad.net/unity8/+bug/1361114 wontfix? [09:45] Ubuntu bug 1361114 in ubuntu-system-settings (Ubuntu) "Says "Enter your PIN" when i have no PIN (there's not even a SIM card on the phone)" [Undecided,New] [09:45] at it will get seeded as soon as somebody depends on it [09:46] pstolowski: it is needed [09:46] and dekko package is already working for the support [09:46] actually, dekko will use the c++ binding, so we are good there [09:46] Saviq, we should just say enter your code or sth else [09:46] but really, this needs to be fixed [09:46] pstolowski: writing why is needed... [09:46] PIN is another thing [09:46] Wellark, oh, that seems wrong, our private path should be first [09:46] tsdgeos, hmm, only if m_results is empty at the beginning of the mehtod [09:46] ? [09:47] pstolowski: no, and in the bottom too [09:47] Saviq: well, you can fix it right that. and hope no component inside unity8 will ever want to use the public module [09:47] pstolowski: count is *our* property [09:47] qabstractitemmodel knows nothing about it [09:47] Wellark, sure, I agree we should rename it anyway, just didn't know that's the fallout [09:47] so if you insert rows you have to say the count changed too [09:47] Saviq: the real fix is to place the private unity modules inside Unity.Private. module import space [09:47] pstolowski: another option is this [09:47] Wellark, it should probably be Unity.Connectivity anyway [09:48] Wellark, yeah, something like that [09:48] pstolowski: http://paste.ubuntu.com/8139241/ in the constructor [09:48] Unity.Private.Connectivity would do it just fine [09:48] or even. Unity.Private.Ubuntu.Connectivity [09:48] Wellark, they are installed in a private import path so I'm not sure prepending everything with Private is overly useful [09:48] dandrader: original idea was to distinguish the qml module tests from the QPA plugin tests [09:48] Wellark, "Unity" should be private already [09:48] tsdgeos, ah, makes sense then. ok, i'll keep Q_EMIT at the bottom [09:49] pstolowski: and in the if branch [09:49] Wellark, anything that's public should be Ubuntu [09:49] Saviq: no, we also have unity API's coming from unity-api-team [09:49] tsdgeos, good point! [09:49] Saviq: like Unity.Action [09:49] which is a public API [09:49] Wellark, well, that should be Ubuntu IMO [09:49] Wellark, which, btw, conflicts with Ubuntu.Components' Action already [09:50] Wellark, IMO things should integrate with Ubuntu, not with Unity [09:50] Saviq: no it does not. Ubuntu.Component action actually inherit Unity.Action [09:50] Wellark, then Unity.Action shouldn't be public [09:50] greyback, ah, ok [09:50] Saviq: ok, we need to have a meeting about module namespacing after the RTM. [09:50] it's a wild west right now [09:50] Wellark, yeah we do ;) [09:51] dandrader: we just don't have QPA tests yet. There is that MR by racarr though [09:51] Wellark, I'll prep a fix and land it asap [09:52] Saviq: thanks! [09:56] tsdgeos, could you ack this change? https://code.launchpad.net/~stolowski/unity-scopes-shell/favorite-scopes/+merge/232050 (it was already reviewed by Saviq) [09:56] Saviq, ^ i had to re-submit that MP to resolve conflicts with tsdgeos' branch [09:57] pstolowski: r147 has lots of diff changes beside the emits?¿ [09:59] hmm [10:02] tsdgeos, these came through your branch (trunk changes), apparently mine branch was behind trunk. the diff looks fine [10:03] pstolowski: ah, i see [10:04] pstolowski: approved [10:04] tsdgeos, thanks [10:06] Saviq: the automatic unlocking is causing troubles [10:06] Saviq: do you have any unity8 brances going in that could carry a oneliner to disable it for now? [10:11] Saviq: https://code.launchpad.net/~unity-api-team/unity8/disable_automatic_pin_unlocking/+merge/232054 [10:15] Wellark, why is it causing troubles? [10:16] mzanetti, greyback, tsdgeos, would like your opinion on bug #1361149 [10:16] bug 1361149 in unity8 (Ubuntu) "unity8's private QML import path should take precedence" [Undecided,New] https://launchpad.net/bugs/1361149 [10:16] Saviq: it was prematurely landed [10:16] dandrader, ↑↑ [10:16] Wellark, yeah, but while the interface's not there it shouldn't matter? [10:16] Saviq: the interface is there [10:16] Wellark, or are you saying that you're adding the interface but it's not good enough yet? [10:17] and they are not working properly together [10:17] Wellark, "is there" or "will be"? [10:17] Saviq: no opinion on what should take precedence, imho we should simply not collide and that's it :D [10:17] and the idea was to land the dual sim unlock dialog with that call in unity8 together so that they are also tested together [10:17] Saviq: the dbus interface is there [10:18] Wellark, ok, so comm fail [10:18] Wellark, I'll squash that change into my plugin rename [10:18] Saviq: ok, thanks! as long as it lands today [10:19] as I'm going to get a lot of heat as it seems indicator-network is totally broken atm :) [10:19] Saviq: yeah... I guess that's valid [10:19] but didn't do extensive research :) [10:20] tsdgeos, it *could* give us a way to override system-wide imports for $reason, too [10:20] Saviq: i don't see $reason [10:20] tsdgeos, while I know we shouldn't have anything like that, having a way out seems useful [10:20] everything there would be bad [10:20] tsdgeos, I know [10:23] Cimi: the two comments you did in https://code.launchpad.net/~aacid/unity8/croppedImageMinimumSourceSize/+merge/232043 are the ones for unwanted space changes? [10:23] the in line comment thing in launchpad is still a bit confusing sometimes [10:23] Wellark, https://code.launchpad.net/~saviq/unity8/rename-connectivity/+merge/232058 [10:24] mzanetti, can you ↑ [10:24] sure [10:25] Saviq: why are you disabling the unlockallmodems? [10:27] Saviq, what did I do? [10:28] Saviq: ok, the sourceSize texture image reduction is up for review, unfortunately i stumbled upon a SDK bug, so it's not easy to land [10:30] mzanetti, because that doesn't work yet [10:30] mzanetti: it's because https://code.launchpad.net/~unity-api-team/unity8/disable_automatic_pin_unlocking/+merge/232054 [10:30] mzanetti, at Wellark's request [10:30] mzanetti, undoing the Unity qmltypes change now... [10:34] Wellark, btw, put it in silo 4 already, will release asap [10:34] mzanetti, there's a new commit from dandrader on the lifecycle branch if you could have a look [10:35] Saviq: is it you that wrote "scrolling up in a scope causes a lot of clipping areas to kick in, are they all needed? (QSG_VISUALIZE=clip)" ? [10:36] tsdgeos: was me [10:36] damned colors :D [10:36] greyback: i can't see that many [10:36] tsdgeos: I'm unsure of the impact of all those clipping areas. I'm told some GPUs can deal with it easy, some not [10:36] otoh maybe i'm misinterpretting the thing [10:37] tsdgeos: the more reddish, the more overlapping clipping areas [10:37] greyback: i only see 1 clipping are to kick in on scroll [10:37] which ones did you see? [10:37] are -> area [10:38] tsdgeos: gimme a minute, I'll try on the phone, I can't recall clearly any more [10:40] tsdgeos: initctl set-env --global QSG_VISUALIZE=clip <- you use this so it turned on for both dash & shell? [10:41] greyback: i'm only having a look at the dash [10:41] greyback: and only on the desktop, didn't think clip areas would be different on the pohne [10:43] tsdgeos: ah ok. Well on phone, there's a red rectangle being shown always under the panel (might be part of shell). Then if I scroll the dash, I see up to 2 levels of red - mostly when scrolling it up [10:44] ok, two levels [10:44] "lot of clipping areas" seemeed like 10 to me [10:45] well that's lvwph [10:45] can't make the header come back without clipping the other stuff [10:46] tsdgeos: sure I know some clipping has to be done, but wanted to question the 2 levels [10:47] of all the suggestions I made, it would be one I'm least expecting a big improvement from though [10:49] now when changing from scope to scope [10:49] there's really weird stuff happening === dandrader is now known as dandrader|afk === tvoss is now known as tvoss|lunch [10:55] /food [11:05] Holas [11:07] tsdgeos, we need to limit whitespace changes [11:07] tsdgeos, I thought I pressed enter before :) [11:07] I left the reply on the entry box [11:08] Cimi: yes yes, i am asking if the two comments are just about the whitespace changes or if there's something esle [11:08] find it hard sometimes to find all the inline comments [11:08] tsdgeos, just whitespaces [11:08] oki :) [11:08] tsdgeos, grep for Cimi on the launchpad [11:08] sure [11:08] after showing diff comments [11:08] but then there's various revisions [11:08] with the combo [11:08] yeah too [11:08] anyway [11:08] and can never be sure if i got them all easily [11:09] was thinking if we can avoid two boolean for the cropped image component [11:09] maybe not though [11:09] only thing we might do is having a status instead two boolean [11:10] "none" "processing" "processed" [11:10] * Cimi wish we could have enums in qml [11:13] tsdgeos, ok when you comment for conflicts then you abstain [11:13] tsdgeos, please remark it to unity-ui-team? [11:14] Cimi: don't worry i won't do needs-fixing anymore [11:14] tsdgeos, your needs fixing is fine [11:14] tsdgeos, as long as later we see it needs review === dandrader|afk is now known as dandrader [11:29] Saviq, yeah I think way to shade the color correctly is converting to a different colorspace [11:31] or actually, from wikipedia "new intensity = current intensity * (1 - shade factor)" [11:31] intensity is each R G B value [11:32] so color.r = color.r * (1 - shadeFactor) [11:32] we should try to see how shadeFactor relates to changes in V or L for HSV and HSL colorspaces [11:33] I guess I could try a test app for that [12:05] greyback, so, any idea about the deadlock? [12:06] Saviq: not a clue. It's not the first time I use a blockingqueuedconnection [12:06] it's as if the receiving thread event loop isn't spinning, which can't be the case [12:06] greyback, yeah, and it was quite a difficult thing to pinpoint, like say it was happening reliably for 5 ap runs of a certain test [12:07] greyback, and then suddenly everything became fine [12:07] hmm yeah, as I'd tested it by hand and never hit it [12:07] greyback, so race [12:07] greyback, I wonder if autopilot has anything to do with it [12:08] greyback, like if introspection could cause it to go into the deadlock [12:08] only thing I can think of is AP sending sigstop to u8 at critical time, just while u8 is shutting down and hasn't told Mir to shut down yet [12:08] greyback, no, it's not on shutdown [12:08] so u8 event loop stopped, while Mir still running [12:08] greyback, and ap never sending sigstop, only sigterm/sigkill through upstart [12:09] greyback.ideas.pop() [12:09] err, sigterm I meant [12:09] greyback, but in any case, this was just happening in the middle of a test, not at the end of it [12:09] well I'll keep digging [12:10] but the change isn't a critical feature, so might postpone it a bit [12:10] yeah, if it doesn't have to do with the wrong surface size on emu / restart, we should [12:10] (postpone) [12:11] I'm certain that can be fixed in another way, so I don't see this blocking it [12:11] but the visual glitch on startup makes me sad [12:24] Saviq, bzr branch lp:~cimi/+junk/color-shades [12:25] qmlscene ColorShades.qml [12:26] here we see the limit of the RGB intensity play... if you cap the intensity you should reduce the saturation of the other color components imho [12:26] Cimi, yeah, and where's my shader code? [12:26] Saviq, you can do shader code by converting to hsv [12:26] Saviq, but I'd much rather do in rgb [12:27] Cimi, *I* can't [12:27] Saviq, thus I started this playground [12:27] Saviq, why you can't? it's just expensive [12:27] we don't want that, I agree [12:27] Cimi, I mean *I*, *Saviq* cannot [12:27] Saviq, what you cannot? [12:28] Cimi, do what you said in shader code :P [12:28] Cimi, and yes, would not, either [12:28] Cimi, converting every pixel to hsl/v really doesn't seem like the right thing to do [12:29] Saviq, http://gamedev.stackexchange.com/questions/28782/hue-saturation-brightness-contrast-effect-in-hlsl [12:30] Cimi, by "give me the shader code" I really meant "give me the shader code" [12:30] Saviq, ok [12:30] Saviq, first we need to find algorithm for rgb shading [12:30] Cimi, it doesn't make sense for us both chasing this [12:30] Saviq, I am on this rgb shading [12:30] without conversion [12:31] otherwise plain semitransparent bright line [12:32] Cimi, but you really should find someone who knows about glsl rather than trying to come up with the algorithm yourself [12:33] Saviq, glsl isn't like in rgb? [12:33] Cimi, there's better ways to do things in shader code than just multiplying r,g,b values or something [12:33] I see [12:33] Cimi, faster, more convenient [12:34] I will study shaders then [12:34] Cimi, no [12:34] Cimi, what part of "find someone who knows about glsl" did you not get? [12:34] Saviq, I am way less scared to study shaders than C++ [12:35] it's math after all :) [12:35] at least sth I studied in uni [12:35] Cimi, you might be less scared [12:35] Cimi, but we still do not have time right now for you to be doing it [12:35] k [12:43] dandrader, do we really consider lifecycle branches fixing bug #1359819 ? [12:43] bug 1359819 in qtmir (Ubuntu) "no splash screen when restarting app killed by lifecycle manager" [Critical,In progress] https://launchpad.net/bugs/1359819 [12:43] dandrader, it feels like we should only mark ↑ fixed after we have the desaturation and things [12:53]