/srv/irclogs.ubuntu.com/2015/08/20/#ubuntu-unity.txt

=== FJKong_afk is now known as FJKong
guest42315𝓐𝓱, 𝔀𝓱𝓪𝓽 𝓪 𝓯𝓲𝓷𝓮 𝓭𝓪𝔂 𝓯𝓸𝓻 𝓢𝓬𝓲𝓮𝓷𝓬𝓮!05:19
guest42315is OTA6 landing today?06:17
=== mardy_ is now known as mardy
mzanettiguest42315, no08:19
guest42315mzanetti, 𝓸𝓱 𝔀𝓮𝓵𝓵, 𝓶𝓪𝔂𝓫𝓮 𝓷𝓮𝔁𝓽 𝔀𝓮𝓮𝓴09:01
mzanettididn't even know one could change font in here09:01
guest42315mzanetti, http://antglove.com/erger/09:01
* guest42315 𝖎 𝖓𝖊𝖊𝖉 𝖇𝖑𝖔𝖔𝖉!09:20
=== guest42315 is now known as drk|aw
cimitsdgeos, ping10:05
tsdgeoscimi: hi10:06
cimitsdgeos, just realised we might have a bug in card anyway, if you set cardData["background"]["type"] to "color" you still have a gradient10:06
tsdgeoscimi: in the current card?10:06
cimitsdgeos, yup10:06
tsdgeosgive me a sec i destroyed my dev env10:09
tsdgeoscimi: even if there's a bug doesn't mean you should not fix the other bug :D10:14
cimitsdgeos, yeah, just saying10:15
tsdgeosi may have broken that10:15
tsdgeoslet me log10:15
cimitsdgeos, because I changed my code a little in the branch, and something is still wrong10:15
cimitsdgeos, backgroundMode: cardData && typeof cardData["background"] === "object" && cardData["background"]["type"] === "gradient" ? UbuntuShape.VerticalGradient : UbuntuShape.SolidCol10:15
tsdgeosor was broken for a long time :D10:17
tsdgeossince last year at least10:17
cimiyeah10:19
=== greyback_ is now known as greyback|shops
tsdgeoscimi: does that work?10:26
ciminope10:31
=== greyback|shops is now known as greyback
cimitsdgeos, I tried putting console.log("card type", cardData["background"]["type"]); \n\ inside the getColor function of that card, and always says cardData["background"] is undefined10:35
cimitsdgeos, I am wondering if it ever worked10:35
tsdgeoswell you get gradient or not10:35
tsdgeosdo you?10:35
cimitsdgeos, but is always the same color10:37
tsdgeoswht do you mean it's always the same color?10:37
cimitsdgeos, and to be honest, if cardData changes the color won't change10:37
cimitsdgeos, the function getColor(0) (1) is not a binding, so is only evaluated at creation10:37
tsdgeoscorrect10:38
tsdgeosthat's enough10:38
cimitsdgeos, I am not sure when cardData is set10:38
cimitsdgeos, is it set before the creation?10:38
tsdgeosi don't know10:39
ciminevermind let me see10:39
tsdgeoscimi: so you're using the tests?10:40
tsdgeosbecause in the tests the color is now driven by cardData but by the template10:40
cimitsdgeos, I just realised the test use background "elements"10:41
cimilet me see10:41
cimi5 mins10:41
cimitsdgeos, I pushed, test pass now10:46
tsdgeostests were passing before too, does it actually work?10:47
tsdgeosseems it does :)10:49
ltinkltsdgeos, hiho, could you please review https://code.launchpad.net/~lukas-kde/unity8/volumeSliderNotification/+merge/268518 for me? should be straightforward10:49
ltinkltsdgeos, there's a bug report attached to it10:50
ltinkltsdgeos, for more explanation10:50
tsdgeossure10:50
tsdgeosltinkl: commented on it11:25
ltinkltsdgeos, ty11:25
mterrymzanetti, We have an EdgeDraggable (or something) class that I remember you(?) mentioning doesn't work with mice?13:22
mzanettimterry, DDA13:23
mzanettimterry, well, all of them13:23
mzanettimterry, ah nice. You looking into the emulator issue?13:23
mterrymzanetti, yeah kgunn mentioned it.  I assume it's not an easy fix?13:23
mterrymzanetti, (other solution is to disable demo in emulator if we can detect it...)13:24
mterryBut that's gross13:24
mterryI think we disabled the edge demo in Desktop mode13:24
mterryFor that same reason13:24
mzanettiyeah... would prolly go for that, given the timeline13:24
mzanettiyeah13:24
mzanettibut.13:24
mzanettiwill still be a problem with bringing in the launcher13:24
mzanettior the right edge13:24
mzanettiwe might need to run unity8 with -mousetouch in the emulator13:25
dandradergreyback, mzanetti, about the mousePointer branches: we need a lot of its code in any case as it has the code that makes window borders change the cursor shape and adds the needed API to enable QML to select cursor shapes not available when using Qt's API13:25
mzanettidandrader, yep, saw the comment on the MP13:25
mterrymzanetti, how do we handle launcher/right edge on Desktop?13:25
dandradergreyback, mzanetti, is the guy making the mir cursor work also doing qtmir & unity8 stuff?13:25
mzanettimterry, launcher can be revealed with mouse hovering... I just remember13:25
mzanettimterry, the phone spread is not there on desktop, there's the desktop stage13:26
mzanettimterry, but I assume we want the phoneStage on the emulator13:26
greybackdandrader: I don't have good answers for that yet, we need to figure that out next week13:26
mterrymzanetti, honestly, -mousetouch sounds like a good solution.  The emulator is emulating the phone, were things are touch events13:26
mzanettimterry, proper fix would really be to make the emulator convert mouse to touch13:26
greybackdandrader: you're right, lots of the code there is useful still13:26
mzanettimterry, yeah, -mousetouch should do13:26
mterrymzanetti, so I guess if we can detect it, our upstart job should pass that13:28
mzanettimterry, yes13:28
mzanettimterry, oh, also:13:28
mterryrsalveti, do you know how an upstart job (or anything) might detect it's running in an emulator?13:28
mterryrsalveti, specifically, our Touch emulator13:28
mterryNot just any emulator  :)13:28
mzanettimterry, I heard that it starts in windowed mode. we probably want to override usage-mode with Staged13:28
mzanettimterry, just like we do windowed override for desktop13:29
mzanettimterry, as we can't know how many mice keyboards people have plugged, we always want it in staged13:29
mzanettiI guess13:29
mzanettifor now at least that would do13:29
mzanettiuntil we ship pocket desktop13:29
mterryHrm.  That's not a command line option, so that would need detection in C/qml13:30
mterryIt looks like maybe the emulator sets ANDROID_PRODUCT_OUT13:37
scriptmonkey00wondering if anyone can give me some advice on the best way to turn a shell script into an icon on my Unity desktop...13:38
scriptmonkey00as-is when I doubleclick on it, it just opens it to edit13:38
mzanettimterry, getprop should hel13:39
mzanettip13:39
mterrymzanetti, ah fair13:41
mterrykgunn, I just went through the emulator with a mouse... and it seemed to work?13:46
mterrykgunn, on vivid13:46
mterry+overlay13:46
kgunnmterry: no change ?13:47
kgunnso vivid+o+silo913:47
kgunn?13:47
mterrykgunn, oh didn't grab silo9...  does that make it worse?13:47
kgunnmterry: that's interesting, well, there is a mir fix in there for "screen config reporting"13:48
kgunnand there's some uitk fix in there13:48
kgunn(for shader stuff i think)13:48
mterrykgunn, will try with 9.  where is the report of it not working?13:49
kgunnmterry: davmor2 irc13:50
kgunngreyback: so the qtmir proposed fix, check for children....2 ques, are tp's always and forever children ? (wondering if there is some odd case where we did a custom thing and they are not)13:52
kgunnsecond, do we not have children in cases other than tp ?13:52
kgunnit's qwidget call basically right?13:52
kgunnor sorry qabstractitemmodel13:53
greybackkgunn: TPs are always children to a session. TP allow data to be shared from one thing to another, so TP need to have a "thing" to share data with13:53
greybackkgunn: atm, the only children are TPs13:53
kgunngreyback: browser tabs ?13:55
greybackkgunn: nope13:56
kgunngreyback: do you have silo16 on? try browser tabs....specifically open 2, then try to get it into the mode where they are both shown (as in you can choose from what you see)13:58
kgunnif that makes sense13:58
kgunnhaving trouble here13:58
kgunnon arale13:58
mterryoh I was testing old emulator image13:58
kgunnmterry: yeah, wonder if mir might have "regressed" something13:58
kgunn?13:58
kgunnlike you were on an old mir13:58
mterrykgunn, old ubuntu-touch image/channel13:59
kgunnhmmm...but if you loaded silo9 you would get the new one13:59
mterrykgunn, testing rc-proposed now13:59
greybackkgunn: working for me, the bottom swipe up shows me the 2 tabs, and I can select as usual13:59
kgunngreyback: so, launched clock...tp comes up, don't select, put clock to background, for like a long time (my phone might have even gone into suspend)14:00
kgunnwake, bring clock fwd14:00
kgunnselect accept14:01
kgunnclock app goes blank(white) and eventually crashes14:01
kgunndamn have a 1on1...will check for crash log tho14:01
greybacktrying14:02
greybackkgunn: ok, similar steps, clock did vanish when I tapped on it. Logs indicate it crashed.14:14
greybackdunno why yet. Would be flabbergasted if due to that patch, but let's see if I can repro without14:16
mterrykgunn, ok can reproduce with latest everything + silo914:19
kgunngreyback: are you testing on mako or arale ?14:38
kgunncurious14:39
greybackkgunn: arale & krillin14:39
kgunngreyback: are you really not having any trouble with the browser tabs ?14:39
kgunnbottom swipe14:39
greybackkgunn: nope, it's fine14:40
kgunngreyback: what image are you on ?14:40
kgunni'm on 9514:40
=== dandrader is now known as dandrader|afk
rsalvetimterry: what do you want to check?14:40
greybackkgunn: 9514:40
rsalvetiif you're running on the emulator?14:41
mterryrsalveti, I think I got it...  getprop ro.hardware == goldfish?14:41
mterrymzanetti, -mousetouch doesn't fix the issue for me...14:41
mzanettiouch14:43
rsalvetimterry: yes, you can check if the hardware name starts with goldfish, or by checking if you have 'qemu=1' at /proc/cmdline14:43
rsalvetithat's how we check for the emulator in a few scripts we have14:43
mterryrsalveti, starts with goldfish, not equals?14:47
greybackkgunn: have reproduced that trust-helper+clock-app crash on vanilla image14:48
rsalvetimterry: do you have the emulator running? I need to check if the x86 version has _x86 appended to it14:49
rsalveticreating the image here to check14:49
rsalvetiif you have it around, just call getprop and paste me the output14:49
mterryrsalveti, it doesn't14:49
mterryrsalveti, ok, will pastebin14:49
mzanettimterry, so if -mousetouch isn't doing, what's the issue then?14:50
mterryrsalveti, paste.ubuntu.com/1213596114:50
mterrymzanetti, not sure...  so it's not even letting me drag from left in tutorial mode14:50
mterrymzanetti, which is even worse than the DDA issue we had when we disabled parts of the greeter on the desktop14:51
rsalvetimterry: yeah, hardware will be goldfish for both14:51
mzanettiodd14:51
rsalveti[ro.build.product]: [generic_x86]14:51
mterrymzanetti, in that situation, we could drag from left, but not do the more fancy right-edge tutorial14:51
kgunngreyback: this is what i'm seeing14:51
rsalvetithat would be the one that has _x86, which is usually the one we use when checking for different devices14:51
kgunnhttp://www.youtube.com/watch?v=fL3Z8oCtKkY14:51
mterrymzanetti, I also can't swipe greeter away if I force it to be in phone usagescenario14:51
rsalvetimterry: so ro.hardware should be all good14:51
mzanettibut dragging from left (launcher at least) is a touch-only DDA  too14:51
mterrymzanetti, is it?  maybe it didn't used to be?  but ok14:52
rsalvetiand in case you don't want to use getprop for that, you can just check for 'qemu=1' at /proc/cmdline14:52
mterryrsalveti, ok thanks14:52
greybackkgunn: weird14:52
mterrymzanetti, so we can check for emulator using goldfish.  If we pass -devicename=goldfish or whatever, we can have a different DeviceConfiguration for it or do whatever we want in qml14:53
mterrymzanetti, so I'll try skipping the whole tutorial instead of just part of it14:53
mterrymzanetti, but i'm not sure we want it to be in phone usageScenario since we can't swipe greeter away14:53
mzanettibut that's the whole point of it... emulating the phone14:54
greybackand that used to work14:54
kgunnbut it's more phone like than being absolutely useless atm :)14:55
mzanettikgunn, what's that browser thing?14:55
mterrymzanetti, I'm not sure why -mousetouch isn't working14:56
mterrymzanetti, presumably that would fix this14:56
mzanettiyeah... we should try to find that...14:56
kgunnmzanetti: see video link ^^^14:57
mzanettiyeah, I watched the video, and can see the tabs are not animating smooth, but what's the context?14:57
mzanetticannot repro the browser thing btw14:59
kgunnmzanetti: iiuc, the tab should stall on a short swipe, showing all tabs open14:59
kgunnto choose14:59
kgunnbut good news14:59
kgunnvirgin image does the same thing for me14:59
kgunnand it might be the user14:59
kgunnor maybe something landed overnight14:59
mzanettiyep. that works here, both with virgin o krillin and with silo 16 on mako14:59
=== dandrader|afk is now known as dandrader
kgunni flashed rc-prposed....wonder, that is likely not the same as our ota6 phone image15:00
kgunnsince they took a snapshot of ppa15:00
kgunnhmmm anyhoo...maybe it's just my thumb it doesn't like15:01
tsdgeoscimi: another regression in our shapes13 branch15:05
mterrymzanetti, kgunn: ok, have branch to skip tutorial.  will file MP now15:06
mzanettimterry, how about the right edge then?15:06
mterrymzanetti, I haven't fully looked into why drags aren't working15:08
mterrymzanetti, https://code.launchpad.net/~mterry/unity8/fix-emulator-tutorial/+merge/26860715:08
mterrymzanetti, that is just a very dumb "skip tutorial entirely in desktop mode" MP15:08
mterrymzanetti, if we fix drags in emulator and can put it back in phone mode, presumably the tutorial will also work again15:09
mzanettiyeah, very likely15:09
mterryoh gosh, I don't know if tests pass with that branch, let me check15:10
mterrymzanetti, presumably the desktop tutorial was also broken then without this MP15:10
mterrymzanetti, since it seems we grew the usage of the problematic drag code15:11
mzanettiat least I didn't hear any complaints15:11
mzanettibut might well be, yes15:11
kgunnmzanetti: bfiller says "like silo 17 is building for the fix to qtmir for the ghosting apps that we think is causing the content-hub/MMS crash of https://launchpad.net/bugs/1486890"15:12
ubot5Ubuntu bug 1486890 in messaging-app (Ubuntu) "Phone crashes if add more than one photo to MMS" [Critical,Confirmed]15:12
mzanettikgunn, yep, I'm building that one15:12
kgunnk15:12
kgunndidn't know if he jwent and did that himself15:12
mzanettikgunn, I've tried reproducing their isse, I couldn't see the crash they're talking about, but the ghost surfaces are definitely involved there15:13
kgunnmzanetti: are you and gerry ready to say +1 on silo 16 ?15:17
kgunni'm seeing 1 or 2 issues, but i think they're rotation related...unrelated to tp change15:18
greybackkgunn: I'm happy to +1 it15:18
kgunnwtf.... ubuntu-touch/rc-proposed/ubuntu-developer doesn't have shell rotation ?15:22
tsdgeosmterry: do we realy want to turn on the screen when getting notifications?15:22
kgunnbut ubuntu-touch/rc-proposed/ubuntu does...15:22
greybackI've no idea what ubuntu-touch/rc-proposed/ubuntu-developer is, but it seems ancient/broken15:23
mterrytsdgeos, I'm not 100% sure, but couldn't think of why we'd show something we don't want the user to see15:23
kgunngreyback: we advertise it on ubuntu public wiki as "good for app devs" cause it has all the extra stuffs15:23
tsdgeosmterry: because he's not looking?15:23
mterrytsdgeos, ?15:24
tsdgeosmterry: the screen is off, means that most probbly the person is not looking at the phone15:24
greybackkgunn: really? Ok, we'd better get someone to look at it15:24
tsdgeoshow often are you looking at your phone when the screen is turned off?15:24
mterrytsdgeos, ... I mean the point is that your phone beeps or vibrates and the screen comes on.  Like you get an SMS or something15:25
mzanettikgunn, yes, +1 here too15:25
mterrytsdgeos, like how all phones work  :)15:25
mterrytsdgeos, we do this today15:25
tsdgeosmterry: really?15:25
mterrytsdgeos, that branch is about consolidating code branches15:25
tsdgeosthe phone i use doesn't do that15:25
tsdgeosone for phone calls15:25
tsdgeosthe rest just beeps and that's it15:26
mterrytsdgeos, for example, we don't do it on telegram message right now (or any push messages) but we should treat them the same as SMS15:26
tsdgeoss/one/only15:26
mterrytsdgeos, well android and ios do it.  We do it for SMS and calls15:26
mzanettitsdgeos, the phone you use has a notification LED that makes everyone blind when looking at it15:26
tsdgeosmterry: that was my other question, you say "unify"15:26
tsdgeosbut there's no more branches?15:26
tsdgeosi thought unification meant there would be code removed from somewhere else?15:27
mterrytsdgeos, once we land this we could go and add branches to the places that currently turn the screen off and remove them15:27
kgunngreyback: one thing, did you see why clock crashed ?15:27
mterrytsdgeos, but it's harmless to do it twice in the meantime15:27
mterrytsdgeos, I suppose I could make the branches now15:27
tsdgeosmterry: ok, you meant "turn the screen on and remove them" and not off, right?15:27
mterrytsdgeos, I just wanted to get opinions on this first15:27
mterrytsdgeos, right...  not on-then-off.  I meant turn the screen on and then go and remove the duplicate lines of code15:27
tsdgeosok15:28
greybackkgunn: something weird going on. Here's my stacktrace: https://errors.ubuntu.com/oops/7ec19558-4730-11e5-a7eb-fa163e707a72 - crashed somewhere in dbus of all places15:28
greybackkgunn: hoping not clash between libdbus-cpp & qtdbus15:28
tsdgeosmterry: and why we don't do it for confirmation? because they are "ok i did something you asked" and thus not that important?15:29
mterrytsdgeos, right.  That's like volume change notifications.  For which we don't turn the screen on today intentionally.  I wanted to keep that behavior15:29
tsdgeosoki15:29
tsdgeosmterry: i agree that unifying on one place makes sense15:30
tsdgeosi'll comment on that and we can wait for more comments15:30
kgunngreyback: minor nit, do we expect a TP to hold the screen on eternally  ?15:30
greybackkgunn: no15:30
kgunni see that15:31
kgunnand....15:31
kgunni left it there for many minutes15:31
kgunn> 5min15:31
kgunni finally hit accept15:31
kgunnand the recording does not continue15:31
greybackthat's USC's job to timeout the screen15:31
kgunn....and now camera app seems frozen15:31
greybackah, camera app may be keeping dsplay on15:31
greybackwhich is thinks it is justified to do, as it is focused15:32
kgunnbut hmmm...15:37
kgunnthis is messed up15:37
kgunnshell responsive but camera not15:37
kgunnat all15:37
mzanettikgunn, now stop breaking things :D15:38
mterrymzanetti, ok I think I fixed the only test broken by my emulator-tutorial branch15:39
mzanettiack15:39
mterry(it's hard to tell, we seem to have multiple failures in trunk, right?)15:40
* mterry wants to make sure it's not just my crazy machine15:41
mzanettimterry, we shouldn't have15:41
mzanettimterry, but ci being broken atm makes it a bit hard to tell (there's ongoing efforts to fix ci)15:42
dandradergreyback, and easy one for you: https://code.launchpad.net/~dandrader/qtubuntu/ensureDBusClipboardReady/+merge/26861215:42
greybackdandrader: ok. you got some comments on ghostApps iin the mean time15:44
dandradergreyback, yeah. saw the e-mail. guess this has higher priority15:45
dandrader(ghostApps, that is)15:45
dandradergreyback, "or crashed of its own accord" <- how can an app crash of its own accord? :)15:46
greybackdandrader: yes yes15:46
dandraderI see it can end/stop/close of its own accord, but crashing...15:46
greyback:)15:46
kgunn@unity anyone know how to "reset" say camera app to reprompt for trusted prompts15:47
dandradergreyback, but I get your point15:47
greybackkgunn: delete ~/.local/share/CameraSomething/trust.db15:47
kgunnta15:47
greybackCameraService15:48
mzanettikgunn, actually, to repro this bug you want also ~/.local/share/PulseAudio/trust.db15:48
greybackdandrader: cool15:48
greybackI tend to nuke ~/.local/share/UbuntuLocationService/trust.db too15:48
greybackthen reboot15:48
mzanettikgunn, and to verify if we dodn't break other trusted sessions you might want ~/.local/share/LocationService/trust.db15:48
mzanettiyeah.... that: find ~/.local -name trust.db -exec rm {} \;15:49
kgunnmzanetti: greyback so on either the audio or camera TP within camera app, if you leave for a while...then hit accept15:56
kgunncamera app becomes unresponsive and camera service spins ~50%15:57
greybackkgunn: I think I got similar with vanilla image tho15:57
greybackit's as if they timeout if they don't get a reply within certain time15:58
kgunngreyback: yes...16:00
kgunnconfirmed16:00
kgunnso mzanetti greyback i am now +1 also16:00
kgunnon the silo 1616:00
kgunni see same bugs on virgin image i see16:01
mzanettikgunn, cool.16:01
mzanettikgunn, maybe we should write those bugs down tho16:01
greybackdandrader: mzanetti found ghost with ghostApps, I confirmed, I added steps to repro in the MR16:03
kgunnack, will log16:04
dandradergreyback, related to those 3 bugs or is it a different, forth, use case?16:12
greybackdandrader: the symptoms are the same, and steps to repro very close to one of the bugs16:12
greybacksince i can't repro on arale, there still may be a race in there somewhere16:13
=== drk|aw is now known as o
o_O16:35
=== o is now known as Guest83769
=== alan_g is now known as alan_g|EOD
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!