**** BEGIN LOGGING AT Sun Nov 11 02:59:57 2012 Nov 11 03:00:41 Thanks... Nov 11 03:01:17 rwhitby, I should have grabbed that but never thought about that. Nov 11 11:45:36 hi Nov 11 12:51:26 ShiftyAxel: pong Nov 11 12:51:44 ShiftyAxel: sounds interesting your idea about reducing scan rate for the touchpanel Nov 11 12:51:49 morning morphis Nov 11 12:51:58 morning :) Nov 11 12:52:11 ShiftyAxel: will try to reduce it Nov 11 12:52:36 cool, let me know how it goes :) Nov 11 12:52:39 but before I have to fix this strange bug with powerd/sleepd Nov 11 12:53:10 hard to debug but makes suspend nearly unusable Nov 11 12:53:13 I'm interested to find out the default value for the scan rate also, wondering if JB uses a really high one or something Nov 11 12:53:17 ah, ouch Nov 11 12:53:44 hey you guys should add quick camera functions for OWOS (if you are working on it) Nov 11 12:53:46 my short-term dayplan: get coffee and wake up... Nov 11 12:53:55 like, a swipe, or a button on the lock screen Nov 11 12:53:57 scan rate? Nov 11 12:54:03 ShiftyAxel: yeah that could be, but as far as I know we're not adjusting scan rate anywhere yet Nov 11 12:54:12 or some obscure key combo. i thought the other day when using my pre3. just an idea for anyone Nov 11 12:54:13 Stskeeps: for the touchpanel Nov 11 12:54:14 fxspec06: feel free to start working on this :) Nov 11 12:54:18 ok Nov 11 12:54:38 no time. i barely have enough time to finish 13 workstudy hrs / week Nov 11 12:54:50 fxspec06_: we have no working camera yet sadly Nov 11 12:54:50 :) Nov 11 12:55:11 fxspec06: and sadly it has low priority Nov 11 12:55:22 well, when it gets working, that is something i've been dying for on my pre3. ever since i got the lumia i have been spoiled Nov 11 12:55:26 you'd want to have a touch sensor sampling rate being a multiple of the screen refresh rate Nov 11 12:56:04 Stskeeps: the theory is that the touchpanel's scan rate is too fast, so it's sending a metric ton of mousemove events which is causing animations that use setPos(foo) every frame to run really slowly Nov 11 12:56:18 :nod: Nov 11 12:56:25 good point. makes sense cam is low priority Nov 11 12:56:39 also, did any of you just try to link up v4l2 to it? Nov 11 12:56:40 :P Nov 11 12:56:54 you'd be surprised how much of this is fairly standardized these days Nov 11 12:56:56 Stskeeps: it's not that easy Nov 11 12:57:06 Stskeeps: it only has this omap4 TILER interface Nov 11 12:57:10 no real v4l2 Nov 11 12:57:16 interesting Nov 11 12:57:23 yes Nov 11 12:57:36 as far as I know there is a gstreamer plugin somewhere floating around Nov 11 12:57:38 omap4 was practically designed by nokia, so that surprises me :) Nov 11 12:57:53 but it needs some properitary initialization sequence as well Nov 11 12:57:59 hey, i read on forums somewhere, guy says he is getting / will be getting 2xpre3's running open webos. he may have been joking, just wondering if anyone heard anything about it Nov 11 12:58:50 i would think that such a device / build would have to come from internal at hp Nov 11 13:12:33 I'm goanna go with misinformation there Nov 11 13:13:26 out of morbid curiousity, where does the OWO kernel requirement come from? Nov 11 13:18:23 i think he might have been joking. still, i wonder because he is getting 4 pre3's and he claims/claimed only 2 would be running. i've joked w/ him bit in the past in the offtopic forums, but still i wonder how/why he would be that specific for just a joke Nov 11 13:18:48 heh. i guess it's all speculation until we see something Nov 11 13:19:23 Stskeeps: I don't know, I rather thing it what was the latest kernel for qemu in OE when they decided for the requirement Nov 11 13:20:10 ok Nov 11 13:49:07 I thought it was something to do with underlying functionality introduced in linux kernel 3.x such as power management etc Nov 11 13:49:57 though that sounds like more of a practicality than anything Nov 11 13:55:49 ShiftyAxel: not that I know Nov 11 14:08:43 so, one step closer to a working suspend again Nov 11 14:14:58 morphis: great :D Nov 11 14:15:43 rather nasty as the first version of my patch leads to a crashing sleepd which then causes luna-sysmgr not to enable the disable after suspend Nov 11 14:17:26 rather nasty when people create glib watches and do not clean them up later ... so remember: never push clenaup work aside, you will forget about it :) Nov 11 14:20:22 but it takes a lot time to mess up glib's internal list of watched fds to produce the poll error ... Nov 11 14:28:22 ShiftyAxel: I can't find any direct setting of the scan rate Nov 11 14:31:56 morphis: as in there's no interface? Nov 11 14:32:02 yes Nov 11 14:32:06 hm Nov 11 14:32:16 kernel driver does deal with any scan rate setting Nov 11 14:32:28 so it's either hard coded into the firmware or I didn't find it yet Nov 11 14:34:55 ShiftyAxel: so we may have to limit the rate in the nyx module Nov 11 14:35:02 if thats really the cause Nov 11 14:35:11 hmm Nov 11 14:35:55 I'll have a bit of a poke about and see if I can hook up a test for that inside LSM Nov 11 14:36:22 since it's best we don't do unnecessary work Nov 11 14:36:39 I need to find the FPS display as well... it'd be good to know what sort of performance we're getting Nov 11 14:37:37 yes, that would be great Nov 11 14:38:02 if I have fixed this sleepd bug really now I will continue with all the other low level bugs Nov 11 14:40:02 wooot, the font problem in the browser application seems to be fixed Nov 11 14:40:25 somebody fixed that? Nov 11 14:41:49 hm, but not on all web pages Nov 11 14:44:45 font problem? Nov 11 14:45:01 * ShiftyAxel had nothing to do with this Nov 11 14:45:57 http://issues.webos-ports.org/issues/277 Nov 11 14:51:17 assignee set to dkirker yesterday, eh? Nov 11 14:51:59 somebody did Nov 11 14:53:11 hmm, we seem to be getting reasonable framerates during the slow-ass animations Nov 11 14:53:31 it averages 40-55ish and dips occasionally into the high 30s Nov 11 14:54:33 morphis: another potential solution to this problem is switching those animations from 'do this every cycle' to using QPropertyAnimations Nov 11 14:54:54 since they seem to have built-in throttling logic Nov 11 14:56:41 hm, may be worth a try Nov 11 14:57:17 Stskeeps: btw. I will do some cleanup work for libhybris in the next days Nov 11 14:57:39 alright Nov 11 14:58:05 i am working on other side atm, a good wayland/qt5 stack and getting that ramped up Nov 11 14:58:15 can finally use work time on this stuff.. Nov 11 14:58:24 you mean based on libhybris? Nov 11 14:59:32 yeah, one of the backends Nov 11 14:59:51 great Nov 11 14:59:55 i fully intend this to be product quality :P Nov 11 15:00:04 I will add proper autotools environment if you're fine with this Nov 11 15:00:12 yep Nov 11 15:00:26 ics and up is fine for me too Nov 11 15:00:42 so we can remove gb/ics code? Nov 11 15:01:02 s/gb\/ics/jb\/ics/g Nov 11 15:01:05 gb, yes, ics and jb we should support Nov 11 15:01:20 as that's where most ODMs are at atm Nov 11 15:03:05 yes, I meant the linker code Nov 11 15:03:16 as gb code seem to work fine on ics and jb too Nov 11 15:04:27 Stskeeps: and what do you think about license? Nov 11 15:06:28 Stskeeps: you're fine with common kernel coding style? Nov 11 15:09:24 original license is aplv2 so i guess it makes sense to continue on that, as i can't do gplv3/gplv2 (and it'd hinder adaptation) Nov 11 15:09:39 er, lgplv3, not gplv2 Nov 11 15:09:55 kernel coding style is fine Nov 11 15:10:13 ok, so kernel coding style + lgplv2? Nov 11 15:10:32 didn't saw any license in the code yet Nov 11 15:10:34 you can't combine aplv2 and lgplv2 :/ Nov 11 15:10:44 from what i know Nov 11 15:11:09 ah you mean apache license Nov 11 15:11:58 I am fine with it Nov 11 15:12:06 will add header blocks and license file for it Nov 11 15:12:27 alright Nov 11 15:12:52 Stskeeps: btw. whats with the folder "testcases"? Nov 11 15:13:24 thing to test libs still work.. was thing from ages ago Nov 11 15:13:33 when i wasnt confident it was working Nov 11 15:13:40 feel free to remove Nov 11 15:14:16 ok Nov 11 15:14:23 I will remove the top level test.c as well Nov 11 15:14:27 k Nov 11 15:39:18 morphis: did you see the ML replies about the animations? looks like people not using libhybris also have this issue Nov 11 15:39:42 though i'm not 100% sure who michael tsukerman is Nov 11 15:51:54 ShiftyAxel: its warzon here on IRC Nov 11 15:52:12 I was already in contact with him about getting accelerations ready for his target device Nov 11 15:55:40 hm, we really need some crash dump daemon Nov 11 15:56:20 that would be good Nov 11 15:56:56 what is warzon targeting? Nov 11 15:57:24 some ac100 device as far as I know Nov 11 15:57:51 i can recommend sp-rich-core for crash dumps Nov 11 15:58:10 strips all useless stuff from a core and adds extra info and compresses Nov 11 15:59:31 morphis: I've setup a hacky little timer system to test the lock button, it's definately the handlePenMoveStateNormal (mousemove) event getting called too fast Nov 11 16:00:11 hm Nov 11 16:00:34 ShiftyAxel: can you implement some hack around to skip 50% of the events and try again? Nov 11 16:00:42 Stskeeps: do you know where the source is? Nov 11 16:01:38 morphis: testing now, i've put if(idx % 2 == 0) around the setPos call and added idx++ below it Nov 11 16:01:43 where idx is a static int Nov 11 16:02:27 yup, definitely more responsive Nov 11 16:02:32 though there's still some catch-up Nov 11 16:02:40 Stskeeps: ah found it: https://meego.gitorious.org/meego-quality-assurance/rich-core Nov 11 16:02:47 ShiftyAxel: hm, ok Nov 11 16:03:10 that's mostly when the unlock circle is fading in/out though, if you let it catch up when the animation's finished it's much smoother Nov 11 16:03:19 ShiftyAxel: so we really have to reduce the number of events sent by the nyx-module Nov 11 16:03:47 yeah, either that or switch all the affected animations over to Qt's animation system Nov 11 16:05:08 ShiftyAxel: does that fix the issue as well? Nov 11 16:05:32 stuff that animates with it doesn't seem to have any slowdown, since Qt takes care of the timing Nov 11 16:06:02 card management, for instance, horizontal movement is controlled by Qt Nov 11 16:06:09 but vertical is setPos Nov 11 16:06:17 only vertical has the lag Nov 11 16:07:30 can you rework it to use the qt animation framework so we can see if it really fixes this? Nov 11 16:07:39 I'll give it a shot Nov 11 16:07:40 as I don't really like to limit nyx-modules just for ui bugs Nov 11 16:07:53 yeah, that would be poor practice Nov 11 16:08:30 though it may be worth looking into the get/set functions for touchpanel refresh rate in the nyx module Nov 11 16:09:13 since they're already implemented, something should be calling them- maybe there's pre-existing control code that would fix the issue? Nov 11 16:09:45 no, there are only stubs for them Nov 11 16:09:55 they should return NYX_NOT_IMPLEMENTED Nov 11 16:11:28 hm Nov 11 16:12:12 so they're there as potential functionality, nothing uses them? Nov 11 16:12:49 yes Nov 11 16:13:27 ShiftyAxel: https://github.com/webOS-ports/nyx-modules/blob/webOS-ports/master/src/rockhopper/touchpanel/evdev/touchpanel.c#L550 Nov 11 16:13:47 yeah, i saw that last night Nov 11 16:13:56 that's what got me thinking that something might call them Nov 11 16:14:34 i'll keep messing with the animations, I might be able to reimplement the finger-tracking code that the card management uses Nov 11 16:16:33 ShiftyAxel: there is some finger tracking code in lsm as well? Nov 11 16:16:54 only the event handling of touch/mousemove Nov 11 16:17:05 card view, launcher, etc Nov 11 16:17:24 hm Nov 11 16:18:01 ShiftyAxel: it would be interesting if Roger can say something about this issue on the ML Nov 11 16:24:04 it would indeed, an internal engineer's perspective would be welcomed Nov 11 16:24:54 for sure and maybe they are already planing to switch to qts animation framework for all animations inside lsm Nov 11 16:25:18 ShiftyAxel: btw. I will work on getting lsm startup more reliable as next thing Nov 11 16:25:44 ShiftyAxel: it sucks that everytime you want to restart it you have to try a number of times Nov 11 16:32:32 yeah, especially with the pixmap scaling Nov 11 16:32:41 way increased startup time Nov 11 16:33:11 yes Nov 11 16:33:49 but the UI is already looking really great Nov 11 16:34:46 we've certainly come far quickly :) Nov 11 16:34:50 we now only need to add more functionality :) Nov 11 16:35:26 I'm looking forward to when we get all the standard stuff running, so I can start improving the UI itself :D Nov 11 16:35:48 been having all sorts of great phone-ui ideas lately Nov 11 16:38:44 :) Nov 11 16:39:00 I already in preparation for getting telephony working Nov 11 16:39:15 just awaiting the alpha release to get it implemented Nov 11 16:39:19 now that I'm looking forward to, I want to use webOS as my daily driver again Nov 11 16:39:49 also, i've rigged up a 60fps timer-based animation for the lockbutton, and it seems to be working much better Nov 11 16:40:04 though it gets skippy when fading the circle in/out Nov 11 17:02:37 i've setup a Qt anim for it now Nov 11 17:02:43 * ShiftyAxel prepares for the mushroom cloud Nov 11 17:03:32 morning Nov 11 17:03:39 ka6sox: mornin Nov 11 17:03:58 ShiftyAxel: yes, webos as daily driver is my aim as well Nov 11 17:04:14 ShiftyAxel: great (qt anim) Nov 11 17:04:48 fxspec06, if the camera driver is available or documented maybe we will work on that.... Nov 11 17:04:58 but its not an alpha release goal. Nov 11 17:06:19 morning ka6sox Nov 11 17:06:20 hiya morphis today's goal is to move the ball forward on the installer...my cracked GN is charged and ready to experiment with. Nov 11 17:06:25 hiya ShiftyAxel Nov 11 17:10:54 ka6sox: yeah thats great Nov 11 17:11:05 I am still working on this nasty suspend bug Nov 11 17:11:29 and i'm trying to make the UI stop lagging Nov 11 17:14:12 ka6sox: and the list of bugs grows and grows :) Nov 11 17:20:13 morphis, it usually does... Nov 11 17:22:30 morphis: QPropertyAnimation is working well Nov 11 17:22:30 it's better to know of bugs than not Nov 11 17:22:52 though it still lags when the ring is fading, so I may need to put them in an animation group Nov 11 17:23:19 you guys know about qml profiler, right Nov 11 17:23:40 very useful tool to speed up your qml Nov 11 17:23:54 most of this stuff isn't QML Nov 11 17:24:08 ah, ok Nov 11 17:24:14 only the system menu and certain alert popups are, though it sounds like it'd be useful for those :) Nov 11 17:24:46 whoa Nov 11 17:24:51 this is really responsive Nov 11 17:24:53 ShiftyAxel: very good! Nov 11 17:25:29 and it now animates back to it's anchor point when you release your finger :D Nov 11 17:25:35 ShiftyAxel: so it seems like every animation should use the qt animation framework to not depend on hardware speed Nov 11 17:25:41 definately Nov 11 17:25:55 i'd say that's doing it the right way(tm) Nov 11 17:26:08 whereas before it was calibrated to the TP... tut tut. Nov 11 17:26:08 ok Nov 11 17:26:14 :) Nov 11 17:26:35 thankfully there aren't too many animations that do it the wrong way Nov 11 17:26:38 ShiftyAxel: can you summarize this as followup for the discussion on the owo-general ml? Nov 11 17:26:42 :) Nov 11 17:26:56 sure thing Nov 11 17:36:38 ML pinged Nov 11 17:43:21 ShiftyAxel: already read, thanks a lot! Nov 11 17:43:29 bbiab Nov 11 17:43:40 I implemented auto-connect for wifi-network now :) Nov 11 17:43:48 thanks for your great work. :) Nov 11 17:44:11 just came back from a business trip and seems many stuff has been fixed.. Nov 11 17:44:12 woah Nov 11 17:46:26 ooh, autoconnect :) Nov 11 17:46:34 gets better eeeeevery day :D Nov 11 17:47:29 ShiftyAxel: :) Nov 11 17:54:07 excellent, it sets up and animates nicely Nov 11 17:54:15 time to make a new branch with this Nov 11 18:04:43 morphis: it looks like the issue also affects QML menus Nov 11 18:05:08 so there may be more far-reaching fixes to be done than just switching this stuff over to Qt animation Nov 11 18:18:45 ShiftyAxel: how does it affect the QML menus? Nov 11 18:18:56 afaik QML uses the animation framework already Nov 11 18:19:08 it should be, but the system menu also has the finger lag Nov 11 18:21:58 hm Nov 11 18:37:02 ShiftyAxel: the notifications are still buggy Nov 11 18:37:19 they still crash for me Nov 11 18:38:43 I tested with email notifications Nov 11 18:38:52 when clicking on them the email app startsup Nov 11 18:41:48 ShiftyAxel: http://files.gravedo.de/luna-notifications.png Nov 11 18:44:26 ShiftyAxel: is this in a branch : https://github.com/webOS-ports/luna-sysmgr/commit/1614f9ae5bf237b192b1c39bafd8ede3377d96f9 Nov 11 18:44:40 ShiftyAxel: it should be in some branch Nov 11 18:44:44 ah, therein lies my problem Nov 11 18:45:03 Ow doesn't trust the cert on my main email account Nov 11 18:45:16 hotmail I have no clue about, and gmail needs an account specific password Nov 11 18:45:31 (aka I haven't set up email) Nov 11 18:46:33 ShiftyAxel: and http://files.gravedo.de/luna-notifications2.png Nov 11 18:47:03 ShiftyAxel: ah ok, I am using no secure connection on the device for my email account Nov 11 18:47:22 owo is missing some certificate management Nov 11 18:50:54 hmm, how recent is that build? Nov 11 18:51:46 I thought I had fixed the notification icon sizing Nov 11 18:54:35 very recent Nov 11 18:54:42 I am using ${AUTOREV} for luna-sysmgr Nov 11 18:56:54 hrm, well in that case I'll try to get my gmail IMAP working with it and debug Nov 11 19:02:27 garr, I hate this email client with a passion Nov 11 19:02:37 error 3010 Nov 11 19:02:50 and then some more error 3010, on top of error 3010 :| Nov 11 20:03:54 mornin all Nov 11 20:04:34 morning EricBlade, what's the word? Nov 11 20:08:49 annoyed is my word right now. Nov 11 20:10:36 Why? Nov 11 20:11:33 sup? Nov 11 20:12:44 something i was assigned to work on that looked like a 5-minute task has turned into several days Nov 11 20:13:00 ouch :| Nov 11 20:14:13 the change to make the one thing worked required making similar changes in lots of other places Nov 11 20:16:18 the joys of refactoring Nov 11 20:17:44 and i'm out of practice with regex ... sigh Nov 11 20:31:12 You should've asked here LOL ;) Nov 11 20:31:19 I know my way around regex quite well ;) Nov 11 22:16:44 hmm Nov 11 22:16:52 the launcher code never ceases to amaze me Nov 11 22:20:30 ... Nov 11 22:25:43 though in good news... the launcher is so close to lag-free :D Nov 12 00:59:05 YEAH Nov 12 00:59:15 launcher is now using Qt animation also, hooray! Nov 12 01:06:17 woohoo! Nov 12 01:06:22 closer every day :) Nov 12 01:14:13 it feels very usable now too :D Nov 12 01:26:20 woo Nov 12 02:26:30 bedtime for me, nn all Nov 12 02:27:15 ~sleep ShiftyAxel Nov 12 02:27:16 * WoceBot uses the banhammer to put ShiftyAxel to sleep for the next 12 hours **** ENDING LOGGING AT Mon Nov 12 02:59:58 2012