Friday, 2012-01-20

*** faenil has joined #nemomobile00:32
*** tibs has quit IRC00:38
faenilanyone awake?00:38
*** tibs has joined #nemomobile00:51
*** M4rtinK has quit IRC00:51
phaerondepends ..00:57
faenilgoing to be now :D00:58
faenilcya ^_^00:58
*** faenil has quit IRC00:59
*** tibs has quit IRC01:07
*** Anssi138 has quit IRC01:15
*** tibs has joined #nemomobile01:20
*** vgrade has quit IRC01:22
*** vgrade has joined #nemomobile01:22
*** mikhas has quit IRC01:24
*** jluisn has joined #nemomobile01:44
*** jluisn has quit IRC01:50
*** arcean has quit IRC01:52
*** tibs has quit IRC01:53
*** jluisn has joined #nemomobile02:02
*** KaIRC has quit IRC02:05
*** tibs has joined #nemomobile02:06
*** phaeron has quit IRC02:08
*** jluisn has quit IRC02:26
*** kimitake is now known as kimitake_idle02:47
*** phaeron has joined #nemomobile03:01
*** beford has quit IRC03:02
*** beford has joined #nemomobile03:10
*** beford has quit IRC03:16
*** beford has joined #nemomobile03:18
*** berndhs has joined #nemomobile03:21
*** kimitake_idle is now known as kimitake03:49
*** phaeron has quit IRC04:48
*** beford has quit IRC05:50
*** tibs has quit IRC05:54
*** singler has joined #nemomobile06:02
*** tibs has joined #nemomobile06:07
*** beford has joined #nemomobile06:08
*** berndhs has quit IRC06:35
*** vakkov has joined #nemomobile06:49
Stskeepsmorn sage07:10
Stskeepsok, so what did we conclude yesterday -- systemui failures?07:10
Sagesystemui fails, it is stuck in some loop somewhere and eats all cpu07:11
Stskeepsyes, and that the secure sharing of pixmap may be faulty07:11
Stskeepsi'm brewing on the new prerelease07:12
SageStskeeps: on panda :P07:14
Stskeepshmm :P07:14
Stskeepswell not really surprised07:14
* Sage ponders what that does on wetab07:19
Stskeepsit tests nokia extensions doesn't it07:20
Sagewell works better on x86 than panda :D07:21
Sage wetab07:21
Sagemainly interested that on Panda the compositor is slow as hell and on x86 it works nicely07:22
Stskeepseglinfo on both could be interesting07:23
Sagethe package was again? :)07:23
Stskeepsegl-utils or something07:23
SageWeTab EGL info:
StskeepsEGL_nok_swap_region? that's surprising07:24
Sagetwo nokia extensions actually07:25
Stskeepswell that explains a lot07:25
Stskeepsmcompositor will work normally on that, but fall back to slow on panda07:26
* Sage look his n950 and looks worried as nothing comes to screen even when powered on07:29
Stskeepscan you grep glEGLImageTargetTexture2DOES /usr/lib/libGLESv2* on panda?07:29
Sageyes, a moment07:33
*** kimitake is now known as kimitake_idle07:33
Sagefor some reason pandaboard isn't getting ssh server up on boot everytime07:33
SageBinary file /usr/lib/ matches07:34
SageStskeeps: ^07:34
Stskeepsi have one last idea, but i need to see eglinfo07:35
SageThe following NEW packages are going to be installed: egl-utils glew libXxf86vm mesa-libGL mesa-libglapi mesa-libGLU07:35
Sagethat is still a bit of a problem07:36
Sagedoesn't override the GLES though07:36
SagePandaBoard EGL Info:
Stskeepsyeah, hence why we should package our own07:37
Stskeepswtf :P07:37
SageEGL_NV as in NVidia?07:38
Stskeepsthat's sometimes ok07:38
Stskeepswe can agree it has all those?07:39
* Sage tries to start mcompositor again after a long time 07:41
* Sage tries to remember how to get the mouse to the screen07:45
Stskeepsuxlaunch remove nocursor?07:45
SageM_USE_SHOW_CURSOR env var07:45
Stskeepsor that07:46
Sagewell did both :)07:46
Sagedidn't help :/07:47
Sageok, so something else needed as well07:47
Sagehmmp. some xorg option was needed07:48
Sagemaybe it was Option "HWCursor" "off"07:50
Stskeepsdo you have Xorg.0.log somewhere?07:50
Stskeepslooks correct07:53
*** veskuh has joined #nemomobile07:54
Stskeepsi wonder why the composition is so slow though07:55
Stskeepswell, besides the fact it'd be compositing 1920x1280..07:55
*** Anssi138 has joined #nemomobile07:57
Sage        Option          "SWCursor"  "on"07:57
SageUnder "Screen" section in /usr/share/X11/xorg.conf.d/99-pvr.conf07:58
Sageactually this works much better than last time I tested07:58
Sageyou can actually use this now to some extend07:59
Sageon application grid 1508:00
Sageapp grid 4 (with widgetsgallery showing)08:00
*** norayr has joined #nemomobile08:00
Sageon widgetsgallery 9-1508:00
Sageyes it is slow, but still much better than last time08:00
Stskeepsok, but isn't systemui still going crazy?08:01
SageI don't have that running :)08:01
Sagestarted only, mdecorator, compositor and home08:01
Sagewithout any apps home is 18FPS08:02
*** tibs has quit IRC08:03
Stskeepsok, try to kill compositor and then widgetsgallery on it's own08:03
Stskeepsi don't presume you can set resolution to something lower?08:06
SageXorg and widgetsgallery taking 100% cpu constantly08:07
Stskeepsthat's not good08:07
Sagelet me try with plain xterm08:07
Sageyup, with plain xterm no cpu usage08:08
Sagelol, cpu usage comes from FPS meter08:09
Sagewithout that no cpu usage08:09
Stskeepsok :P08:09
Sageso that is probably slowing the ux down as well08:09
Stskeeps-log-fps instead?08:09
* Sage tries08:09
Sagewhere does it put it?08:10
Stskeepsstdout or $PWD08:10
Sagesame 2208:11
Sageand huge cpu usage08:12
w00tfps meter is not a surprise08:14
Stskeepsanyway, see if you can somehow drop to 800x60008:14
w00twhen measuring fps, naturally, the application is forced to redraw constantly (even if it doesn't need to) and as fast as it can08:14
*** tibs has joined #nemomobile08:15
*** beford has quit IRC08:37
* Stskeeps preps new prerelease08:38
w00tso was there any consensus about 4.8 problems?08:40
Stskeepsi'm not confident it was related to it08:40
Stskeepswe're downgrading input proto at least08:40
* w00t cancels the ticket to columbia08:40
*** tibs has quit IRC08:42
SageStskeeps: did cp N900.conf Default.conf and got smaller display generated, however Xorg still pushes the full HD out and fps about the same08:43
*** slaine has joined #nemomobile08:44
Stskeepsyeah, i was thinking xrandr style08:45
*** DocScrutinizer has quit IRC08:47
*** DocScrutinizer has joined #nemomobile08:47
Sagewell, tried xorg.conf style but didn't succeed08:49
veskuhHow old resource engine we have? There is definitely some issues with it or with the way we use it. the occasinal dialer issues seem to be related to resource use08:50
veskuhother possible quilty party might be audio pipeline since the problems are related to ringtone08:50
Sageveskuh: resource engine? As resource policy?08:51
Stskeepsn900 or n950?08:52
*** araujo has quit IRC08:53
veskuhOn n950. the issue might be in the policy or the engine08:53
veskuhI'm assuming the engine is libresource?08:53
SageStskeeps: we don't have xrandr in mer?08:55
*** tbf has joined #nemomobile08:55
Sagewe have lib, but where should the binary be?08:56
Stskeepsok, it might be that we don't have the tool08:56
Sageseems like that yes08:57
Sage <-08:57
SageStskeeps: armv7l08:58
MerBotStskeeps: Error: "$" is not a valid command.08:58
SageMerBot: hi08:58
MerBotSage: Error: "hi" is not a valid command.08:58
SageMerBot: quit08:58
MerBotSage: Error: You don't have the owner capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified.08:58
Stskeepsit logs, i think08:58
StskeepsSage: so pvr standard is 800x600 ..?08:58
SageStskeeps: no idea of that just tried :)08:59
Sageat least in xorg log it says that08:59
Stskeepsso it does 22fps at 800x600? hmm08:59
Sagewell, it outputs the full hd still to screen09:02
Sageso that xorg conf did nothing09:02
Sageany thoughs what is wrong? I would have expected that Xorg wouldn't start09:02
Stskeepsxorg has a tendancy to ignore everything in it's conf09:02
Sagegot hd ready tv for quick test09:04
Sagewidgetsgallery 35fps09:04
Sageappgrid with widgetsgallery open 7fps09:04
Sageappgrid 21fps09:05
Sagehome without apps 26FPS09:05
Sagewith 1360x768 resolution09:05
Sageor what ever tha tis09:06
Sageso notable improvement there09:06
Sagethat is the same resolution as wetab though, right?09:06
Sagechecked this tv is 1366 x 768 so same as wetab09:08
Sageadded xrandr -s 800x600 call to /usr/bin/mcompositor and well now it is a slideshow09:15
Sageeven the mouse09:16
Sageglestest goes from 166 to 177 fps when changing to 800x60009:17
Sageand widgetsgallery gives 49fps without mcompositor09:18
Sagen950 gives 44FPS with all the crap running in the background on same screen09:19
*** singler has quit IRC09:22
* Sage updates his n950 to latest nemo release09:27
*** admiral0_ has quit IRC09:28
*** mikhas has joined #nemomobile09:31
StskeepsSage: how long are you approx staying today?09:32
Sage~5hours more I would say09:33
*** vakkov has quit IRC09:41
*** smoku has joined #nemomobile09:46
*** ronoc has joined #nemomobile09:49
*** singler has joined #nemomobile10:06
*** rcg has joined #nemomobile10:14
*** KaIRC has joined #nemomobile10:22
Sagehmmp... I guess we should drop the "testing" from the image names10:24
Sagelegacy leftover from meego ce times10:24
Sageveskuh, Stskeeps: opinions?10:25
Sagealso probably we can drop the meego .ks files from our git tree as well10:27
Sageor can someone see need for those? Repos are broken already.10:27
veskuhthe ks files are anyway in our old releases archive If I remember correctly10:27
Stskeepswell, we will switch to inheriting from mer-kickstarter-configs i presume?10:28
*** arcean has joined #nemomobile10:29
SageStskeeps: well completely different thing10:29
SageI have been dragging meego based .ks files in see 1.2 and .1310:30
Sagethose are broken already and do not work10:30
SageNemo images do the same but better10:31
Sageso in general duplicated broken information in there10:31
Sagethose can be found from gitorious backlog anyway if someone really wants them at some point10:32
StskeepsCOBS is getting the new prerelease now10:39
Stskeepslet's see if that helps the regressions10:39
*** vgrade1 has joined #nemomobile10:43
*** vgrade has quit IRC10:43
*** vgrade1 is now known as vgrade10:44
*** lizardo has joined #nemomobile10:45
Sageok, it takes at least 2-3 hours to compile I guess10:48
Sageor well actually if version numbers do not change much then probably not10:49
Sagedepends of libmeegotouch gets recompile10:49
the-bossphaeron lbt sage stskeeps SR#4158 waiting for review at
the-bossphaeron lbt sage stskeeps SR#4158 Accepted promotion request10:59
w00tveskuh: just close bug 4, I haven't been able to verify so far as I have an absolutely terrible memory, it may well be gone11:01
* Sage fixes his copy-kernel mic patch11:02
veskuhw00t: cool11:06
*** niqt has joined #nemomobile11:11
*** jluisn has joined #nemomobile11:15
*** xruxa has joined #nemomobile11:28
Sage2000-01-01 00:03:17 [libcontextsubscriber] CRITICAL [provider.cpp:165:void ContextSubscriber::Provider::constructPlugin()] Error loading plugin "/usr/lib/contextkit/subscriber-plugins/liblocation" : "Cannot load library /usr/lib/contextkit/subscriber-plugins/liblocation: (/usr/lib/contextkit/subscriber-plugins/ undefined symbol: _ZN16LocationProviderD1Ev)"11:30
Sage^ I've got tired to that one will check it11:30
Stskeepsi'm fairly sure we don't have that patch11:35
Stskeepslooking at the secure drawable sharing stuff11:35
*** Anssi138 has quit IRC11:37
Sage <- it should fail then?11:37
veskuhStskeeps: Isn't that related to AEGIS? shouldnt bother us11:37
Sageprint "can't send pixmap sharing message"11:37
Stskeepsveskuh: .. ish11:38
* Stskeeps checks xorg source11:38
Sage/usr/share/contextkit/providers/location-skyhook.context <- ring bells anyone? Do we need this?11:40
Stskeepsweren't that the guys that google screwed over?11:41
veskuhskyhook has/had somekind of wifi based location db11:41
Sagewell, we should use geoclue anyway I guess or something?11:42
Stskeepswell, technically it should be liblocation or qt mobility's11:43
Sageok, I'll drop that then11:44
Sagewe don't use it anyway11:45
*** ronoc has quit IRC11:46
*** singler has quit IRC11:46
* Stskeeps has some worry that that atom is being set by policy framework11:47
*** norayr has quit IRC11:47
*** ronoc has joined #nemomobile11:51
*** araujo has joined #nemomobile11:56
*** araujo has quit IRC11:56
*** araujo has joined #nemomobile11:56
Sagepah, all misunderstanding and invalid packaging11:56
Sagedoes something use contextkit-meego liblocation (geoclue)?11:57
Stskeepsright, pmo's acting up so i can't check the xorg fbdev sgx driver for hints11:57
Stskeepsi don't think so11:57
Stskeepsmaybe qt mobility, but that would be odd11:58
Sageok, so I'll just drop all of those from the packaging11:58
Sagethey will be there in separate pacakge11:58
Sagebut not in the default mess that they are atm.11:58
Sagealso picked up the latest contextkit-meego upstream changes as patches11:59
*** dazo_afk is now known as dazo12:00
Sagedoesn't fix the actual error but drops the liblocation stuff that shouldn't be there in the first place atm.12:04
Sage%{_libdir}/contextkit/subscriber-plugins/*.so in the default package matched a bit too much :)12:04
*** KaIRC has quit IRC12:06
*** smoku has left #nemomobile12:16
*** KaIRC has joined #nemomobile12:18
*** smoku has joined #nemomobile12:19
*** smoku has left #nemomobile12:20
*** smoku has joined #nemomobile12:20
*** phaeron has joined #nemomobile12:32
SageStskeeps: btw, noted that on pandaboard the same pid increas happens12:35
SageXorg has pid 1974 on boot12:35
Sageso it is most probably more generic issue than only on N90012:38
Stskeepsthat linking looks totally messed12:39
Stskeepscan you force a rebuild?12:39
phaeronextlinux on mer is behaving weirdly12:41
*** norayr has joined #nemomobile12:42
phaeronthe one I installed on my usb hdd from my laptop works. the one in mer extra doesn't find the config file12:43
phaerondoes it need to support ext4 .. ?12:43
Stskeepsand btrfs maybe12:43
Stskeepsbtw, instead of doing %ifnarch %{arm} for make check, etc, we can do %if ! 0%{?qemu_user_space_build} instead12:44
Stskeepswill be in next core (ie, not the one being preleased atm)12:45
*** singler has joined #nemomobile12:45
phaeronbut it does find the kernel when I type the command line12:50
phaeronis there a usable dri for i956 somewhere12:53
Stskeeps965? should be along with normal x86 adaptation12:54
phaeronconnman really needs a command line interface12:57
Stskeepsusr/lib/connman/tests (connman-tests)12:57
phaeronwhich is not installed .. and I need network to install it :D12:58
phaeronbrb switching the cable12:59
*** phaeron1 has joined #nemomobile13:01
SageStskeeps: same failure again :/13:01
Stskeepsit should fail all the time13:02
*** phaeron has quit IRC13:02
Stskeepsit tries to link to nonexisting stuff13:02
Sagewhy it happens with core next only?13:03
Stskeepsvery good question13:03
Sagegcc changes went in already13:03
*** arcean_ has joined #nemomobile13:03
*** arcean has quit IRC13:04
Sagepreinstalling qemu-usermode-static-arm... ?13:04
Stskeepsproject-wide qemu13:04
*** tbf_ has joined #nemomobile13:05
Sageinstalling cross-armv7tnhl-platformfile-0.0.1-1.2 on armv7hl ?13:05
Stskeepsit just means 'up to'13:05
*** tbf has quit IRC13:08
Sagedid your python change?13:08
*** smoku has quit IRC13:10
Sage-j3 vs -j413:10
Sage-j3 fails -j4 succeeds13:11
Stskeepslovely, and -j1 ?13:11
SageI mean that there is -j3 and -j4 difference in those13:11
Stskeepsit should fail all the time, i would have hoped13:11
Sagethose libs come from
Sagesee line 19913:12
Sagedoesn't mer-release get installed?13:12
Sageto build root that is13:13
Sageit should have link to /etc/meego-release13:13
Stskeepsyes, mer-release gets installed13:13
Stskeepsalso, wtf13:13
Sageit is there13:14
Sagethat python script is a WTF moment yes :)13:14
Stskeepsyes, but what does it contain..?13:14
Stskeepscontains Mer ;)13:15
Sage  58 ln -s mer-release $RPM_BUILD_ROOT%{_sysconfdir}/meego-release13:15
Stskeepsecho "%{release_name} release %{version} (%{release_name})" > $RPM_BUILD_ROOT%{_sysconfdir}/mer-release13:15
Stskeepsthis warrants a patch to applauncherd13:16
Stskeepsi don't suspect you have time to make a image build with new prerelease?13:17
*** smoku has joined #nemomobile13:17
SageI have time I think13:19
Stskeepsjust upload somewhere13:19
Sagefunny thing is that i586 builds ok :)13:20
Sageprobably would nee more fixing13:20
Sageno time for that so this have to do13:21
Sageok, builds ok. Pushing forward13:22
the-bossphaeron lbt sage stskeeps SR#4159 waiting for review at
the-bossphaeron lbt sage stskeeps SR#4159 Accepted promotion request13:27
Sagealso .ks and .packages if needed13:44
Stskeepsthanks, grabbing13:45
*** nicola_ has joined #nemomobile13:46
Sageflashing it on my device atm.13:46
*** nicola_ has quit IRC13:46
*** faenil has joined #nemomobile13:56
phaeron1ok extlinux wanted /boot/syslinux/syslinux.cfg13:57
phaeron1debian's was ok with /boot/extlinux/extlinux.cfg13:57
faenilheya people :)13:57
Sagephaeron1: intel did probably some odd stuff there13:58
SageI recall founding some oddities from there when doing images as well13:59
phaeron1anyway let's see if I can get X ..13:59
SageStskeeps: didn't fix the touch issue :)13:59
Sageerr :(13:59
Sagedrag is ok, but actual touch isn't14:00
Sagealso it seems that touch is ok on quick launchbar but not in application menu14:00
Stskeepsdoes xterm work?14:00
Sagegestures work now though14:00
Sagedid the qt upgrade touch to the touch patches we had there?14:01
Sagesystemui is still acting up as well14:01
Stskeepsyeah, but systemui acts up on old core too14:02
Sagenot like this14:02
Sagewith old core systemui crashes at times yes. But it never goes to loop and takes 100% cpu14:02
Sageor at least I haven't seen that happening with curren core14:02
SageThe core it self seems ok, but qt seems to be acting up14:05
Sageno crashes, no odd messages14:05
Sagewell, timed seems to crash at times, but not sure if it does that also with current core14:06
Stskeepsok, i'll have a look14:06
Stskeepscould you do a x86 image too for comparison purposes?14:06
*** jsv has quit IRC14:07
w00tI don't think there were any conflicts with the XInput patch code14:07
w00tthere was one conflict with configure14:08
Stskeepsw00t: got a n900 along?14:08
* Sage ponders if timed is build with qt14:08
w00tStskeeps: no14:09
w00tmy n900 is an ex-phone14:09
w00tit went from occasionally refusing to boot to just being dead14:10
Sageyes, timed is also created with qt.14:10
*** berndhs has joined #nemomobile14:13
Sagew00t:,sidebyside,312,1,qt-4.8.0-xinput2-support.patch intentionally dropped stuff after ~2200 line?14:14
* Sage ponders if someone forgot git add14:14
Sageerr.. nm. it is ~ file14:15
w00tSage: qwidget_p.h~14:15
w00twould it be quicker to just revert to 4.8-rc (from the non-next repo) and see if it suddenly works?14:16
Sagethat can be tried sure14:18
SageI'll do that and see what happens14:18
Sageneed to go soon though14:18
Stskeepshave a good weekend :)14:19
* Sage installs rc1 qt rpm's 14:22
Sage Stskeeps w00t14:22
Sagejust links if you want to try -rc114:23
* Sage reboots14:23
Sageok, seems as bad as before14:25
Sagew00t: you are probably safe here :)14:25
* w00t phews14:25
*** phaeron has joined #nemomobile14:25
*** phaeron1 has quit IRC14:26
Sageok, have to go now. Will be at IRC at some point during the weekend.14:27
faenilcya .)14:27
*** singler has quit IRC14:28
faenilwould you be so kind to try help me with this?14:29
faenilit's been on forum and on qt-qml mailing list for a while answer :(14:29
*** norayr has quit IRC14:30
w00topengl needs FullViewportUpdate14:31
w00tas for the FPS, give profiling a shot14:31
faenilit needs fullviewport?14:32
faenilwhy :(14:32
faenilthere's no need to draw the UI every time...14:33
faenilabout profiling, it doesn't work on Qt for Android, it says the peer has disconnected, and doesn't start14:33
w00topengl is not like software rendering14:34
w00tit isn't necessarily always a performance hit14:34
w00ttry profile it on a desktop14:34
faenilit doesn't work on desktop for a bunch of reasons...14:35
w00tthen you're in trouble14:35
faenil(all related to my laptop)14:35
faenilyes I am...but is that a Qt thing? or OpenGL?14:35
w00tyou can either figure out a way to get some kind of performance metrics, or take random stabs in the dark and hope you find things that improve performance14:35
faenilthe fullviewportupdate14:35
faenilthe need for it14:36
w00tI don't know where the requirement comes from, I just know it exists14:36
w00tI'd guess it's probably opengl-related though14:36
faenilI see...strange, strange..14:36
w00topengl is very much about a scene where as with software rendering you can just clip all your drawing to one section of the screen fairly easily14:37
w00t</uneducated guessing>14:37
faenilthat would sound new to me, but I'm no expert, so could be :)14:37
faenilone alternative could to redraw things myself14:39
faenilset NoViewportUpdate14:39
faenilproblem is, how do I tell qml components to repaint?14:39
faenilso you're basically telling me I can't set one context and draw a part of the screen14:41
faeniland then set another context and draw another part?14:41
faeniland draw the first part using the first context just once in a while?14:41
mikhasfaenil, how about trying QML2 first before wasting time optimizing old QML?14:43
faenilmikhas: you've already told me that, haven't you :D14:43
mikhasI guess.14:44
mikhasIt bears repeating.14:44
faenilsure ;)14:44
faenilbut I wanted to know why it's not working this way :)14:44
mikhasI dont think that current stable QML gives you enough hooks to really get optimal performance.14:44
mikhasIf you want to get best perf out of QGraphicsView, do it in C++, not QML :-/14:45
faenilI thought a mix was doable via QDeclItem... :)14:45
mikhasthen one can disable all the Qt "convenience", like painter state saving and such14:45
faenilbut also loses all components14:46
mikhasfaenil, as a rule of thumb: count the amount of items that update during a paint event, and get that number down.14:46
faenilbut still, that's not a problem, I just wanted to know the tech reason behind the current failure :)14:46
mikhasthe claim that QGV easily supports a million items is only true for certain conditions (one of them being "runs on desktop")14:47
faenilbut ui items do not update if you don't touch them14:48
faenilwhy are they redrawn14:48
mikhasyou need to call update on the item that is scheduled for redraw14:48
mikhasyou can work with FullViewport updates14:48
*** phaeron has quit IRC14:49
mikhas*if* you carefully enable device coordinate caching for the items that dont change14:49
mikhasbut that's QGraphicsItem API, not exposed to QML14:49
faenilI see...14:49
mikhasyou can expose it manually14:50
faenilI also hoped SmartViewportUpdate would help me :D14:50
mikhasI think I also explained that some time ago, somewhere.14:50
mikhasforget it, there is no silver bullet14:50
mikhasfaenil, you can read "The Strategy" part in
faenilthanks :)14:51
faenilinteresting, interesting, thanks :)14:53
mikhasI also have code examples14:53
*** phaeron has joined #nemomobile14:53
faenilif this helps, my gl view doesn't have a fixed size..14:54
mikhasthat should not matter, although the reconfigure event for the widget itself will usually be expensive (or rather, lead to expensive repaints)14:56
faenilone thing is not clear14:58
faenilwhen you press the key it's drawn over the idle as overlay14:58
faenilbut when you release the finger, you have to repaint the idle keyboard (the whole one), don't you?14:59
faenilcoz that's one item, as you stated14:59
*** himamura has quit IRC15:00
mikhasfaenil, but it is redrawn from Qt's internal pixmap cache15:03
mikhasso no paint event is emitted15:03
faenilI see15:03
mikhasand no item redraws15:03
faenilcouldn't I paint a transparent background instead of the black one?15:04
faenilso that the UI is still shown but not repaint?15:04
mikhasyou can, but drawing transparent pixels is also expensive15:04
mikhasQGraphicsView does not handle transparency well15:05
faenilI see...15:05
mikhasthe patches are all there, but it was decided to only fix that for Qt515:05
mikhas(I think it's a pure marketing stunt, the fix could have been done in Qt4 without public API changes)15:05
faenilwell, now that everything's somewhat clear, I'll wait for QML2 :D15:05
mikhasmission accomplished15:06
faenilahahaah :D15:07
faenilit should be out in April, right? or has there been any news about any delay?15:07
w00tthere is no date firmer than "sometime in summer (hopefully)"15:17
faeniloh, summer hopefully :O15:19
faenilmikhas: would you mind providing an explanation to another issue I'm having ? :D15:22
faenilI need a velocity attribute because I need to handle a pan gesture15:25
faeniland qt doesn't seem to offer anything like that..15:25
faenil(which is strange)15:25
*** emmynet has joined #nemomobile15:25
lbtdoes nemo dial the #* phone numbers?15:25
Stskeepsi'm not sure the dialer has that functionality15:26
lbteg yes... #1345#15:26
lbtor *#1345#15:26
faenilso I was trying to calculate the velocity by myself15:26
*** lizardo has quit IRC15:26
lbtOK, thanks Stskeeps15:26
faenilI get deltaTime, deltaTransl, and calculate velocity at each touch event15:26
faenilbut deltaTime and deltaTransl are incoherent, thus resulting in unreal velocity values...15:27
faenildeltaTime is 1-2-3ms most of the times (80-170ms some times)...15:27
*** lizardo has joined #nemomobile15:28
*** tbf_ has quit IRC15:32
*** tbf_ has joined #nemomobile15:35
*** tbf_ has quit IRC15:43
*** arcean__ has joined #nemomobile15:45
*** M4rtinK has joined #nemomobile15:45
faenilhas anyone ever used QPanGesture's acceleration? I'm reading sourcecode and it seems that the acceleration is only handled on MACOS! Which gives it for free!15:47
faenilIt's written on the docs, but the only thing I see here is acceleration = 0 in the initialization15:47
*** arcean_ has quit IRC15:48
*** veskuh has quit IRC15:49
faenilcan't believe there's no velocity property in qml or even in qpangesture...16:00
*** xruxa has quit IRC16:00
*** himamura has joined #nemomobile16:04
*** lizardo has quit IRC16:11
*** phaeron has quit IRC16:16
*** tbf_ has joined #nemomobile16:17
*** arcean__ is now known as arcean16:26
*** niqt has quit IRC16:26
*** emmynet has quit IRC16:33
*** ronan22 has quit IRC16:48
*** faenil has quit IRC16:49
*** faenil has joined #nemomobile16:51
*** lizardo has joined #nemomobile16:54
*** araujo has quit IRC17:03
*** tbf_ has quit IRC17:12
*** rcg has quit IRC17:14
*** AIN2 has joined #nemomobile17:16
*** berndhs has quit IRC17:21
*** smoku has quit IRC17:42
*** beford has joined #nemomobile17:48
*** slaine has quit IRC17:59
*** vakkov has joined #nemomobile18:14
*** ronoc has quit IRC18:22
*** faenil has quit IRC18:48
*** araujo has joined #nemomobile19:01
*** araujo has quit IRC19:12
*** faenil has joined #nemomobile19:26
*** dazo is now known as dazo_afk19:33
*** AIN2 has quit IRC19:46
*** luke_dirtwalker has quit IRC19:52
*** jluisn has quit IRC20:01
*** admiral0 has joined #nemomobile20:01
*** jluisn has joined #nemomobile20:04
*** faenil has quit IRC20:07
*** tibs has joined #nemomobile20:15
*** phaeron has joined #nemomobile20:21
*** AIN2 has joined #nemomobile20:27
*** phaeron has quit IRC20:30
*** AIN2 has quit IRC20:31
*** AIN2 has joined #nemomobile20:32
*** phaeron has joined #nemomobile20:35
*** araujo has joined #nemomobile20:51
*** araujo has quit IRC20:51
*** araujo has joined #nemomobile20:51
*** lizardo has quit IRC21:04
*** norayr has joined #nemomobile21:14
*** araujo has quit IRC21:17
*** tibs is now known as mrtibs21:22
*** jluisn has quit IRC21:23
*** faenil has joined #nemomobile21:26
*** npm_ is now known as npm21:34
*** stroughtonsmith_ has joined #nemomobile22:41
*** stroughtonsmith has quit IRC22:42
*** stroughtonsmith_ is now known as stroughtonsmith22:42
*** stroughtonsmith has quit IRC22:48
*** vakkov has quit IRC22:50
*** araujo has joined #nemomobile23:01
*** araujo has quit IRC23:01
*** araujo has joined #nemomobile23:01
*** araujo has quit IRC23:02
*** araujo has joined #nemomobile23:06
*** araujo has quit IRC23:06
*** araujo has joined #nemomobile23:06
*** Anssi138 has joined #nemomobile23:43
*** Anssi138 has quit IRC23:48

Generated by 2.11.0 by Marius Gedminas - find it at!