**** BEGIN LOGGING AT Fri Aug 03 02:59:59 2012 Aug 03 03:14:24 RagingMind, yes Aug 03 03:22:44 I decided to do it a slightly different way *shrugs* Aug 03 03:25:38 got URL? Aug 03 03:32:37 http://webos-ports.org/wiki/GPU_Drivers Aug 03 03:32:58 still working on it Aug 03 03:36:05 nice! Aug 03 03:51:33 RagingMind, didn't someone say that there was an adreno 220 FOSS project? Aug 03 03:51:48 for the Snapdragon? Aug 03 04:04:23 Hi all. Just read on TheVerge that Cyanogen folks are not able to update the first gen snapdragon devices--Nexus One etc. Aug 03 04:24:50 ka6sox, yeah, freedrino (sp?) Aug 03 04:39:35 RagingMind, ya, that sounds right Aug 03 04:49:50 I'm tired so this is turning into many many small edits as I fix stuff Aug 03 04:50:13 RagingMind, take a break, it will be there tommorrow... Aug 03 04:52:29 I'm down to just omap, after this I've exhausted all the info rolling around in my noggin Aug 03 05:00:38 ka6sox, can you check it over quickly? :) Aug 03 05:19:36 sure Aug 03 05:31:40 RagingMind, looks good! Aug 03 05:33:20 RagingMind, now get some sleep! Aug 03 05:42:52 folks Aug 03 05:43:45 bhey Aug 03 05:43:47 bhuey Aug 03 05:44:18 bhuey, was there any major change in subsystems between 2.6.35 and 2.6.38? Aug 03 05:57:51 me sleep? na... not yet Aug 03 05:58:56 though I admit that placing AMD under x86 may be a mistake Aug 03 06:00:11 ka6sox: some parts Aug 03 06:00:22 instrumentation works a bit better in .38 Aug 03 06:00:34 futex doesn't need to be backported in .38 Aug 03 06:00:36 did anything functionally change? Aug 03 06:00:49 like API changes? Aug 03 06:00:50 the glibc will have to be changed to match the newer kernel Aug 03 06:00:53 no Aug 03 06:00:59 not that I know of Aug 03 06:01:06 Linux kernel is the Linux kernel Aug 03 06:01:24 well..the device system has changed how many times? Aug 03 06:01:30 the problem you have to worry about is whether or not you can get the proprietary stuff going Aug 03 06:01:45 you probably need an ABI abstraction layer for those drivers Aug 03 06:01:51 don't know Aug 03 06:02:12 since I startd paying attention at least 4 Aug 03 06:02:46 ...so that you can compile it in what ever way you want but not have it conflict with the existing ABI Aug 03 06:02:48 okay so it sounds like a Bridge Too far.. Aug 03 06:02:55 unless you can backport the futex stuff. Aug 03 06:03:01 that might not be possible if the driver pokes at an internal kernel structure Aug 03 06:03:15 futex stuff isn't important as other things Aug 03 06:03:26 the FB stuff will probabliy be a mess. Aug 03 06:03:30 msm_stat and powerd needs to be redone Aug 03 06:03:32 yeah Aug 03 06:03:40 pmem Aug 03 06:03:51 it sucks 5-10 percent of the CPU time just looping for shitty reasons Aug 03 06:04:03 that in combination with sensord Aug 03 06:04:14 oh well. Aug 03 06:04:19 unfortunately that requires intimate interaction with userspace bits Aug 03 06:04:36 spilled milk at this point on legacy devices. Aug 03 06:04:46 3.X should be better....I hope Aug 03 06:04:49 I had a kernel environment at Palm that could build such test kernels Aug 03 06:05:03 you should ask for one that can build a kernel so that you don't have to worry about ABIs Aug 03 06:05:10 3.x isn't going to be much better Aug 03 06:05:32 the problems with webOS has to do with the Webkit implementation and stuff related to it Aug 03 06:05:34 I just hate the thought of having to learn all that new pcre and frequency domain crap. Aug 03 06:05:37 it's single threaded for the most part Aug 03 06:05:54 well, that's part of the govenor stuff, should be easy to port Aug 03 06:06:13 without docs you will get whatever the vendor supplies Aug 03 06:06:17 tweaking might be hard Aug 03 06:06:20 yeah Aug 03 06:06:24 and crash happy Aug 03 06:06:45 it's why something like working with a cyanogenmod kernel might be better than a 3.x kernel Aug 03 06:06:48 if you don't set them in the right order and things....life becomes...Interesting. Aug 03 06:06:57 it's a more mature system generally minus what they've done for TP Aug 03 06:07:22 how many other Snapdragon's have they done? Aug 03 06:08:20 3.x is enterprise and Android drive integration Aug 03 06:08:29 little of that has anything to do with what the TP does currently Aug 03 06:08:38 who ? Aug 03 06:08:50 it's why I'm not that gungho about 3.x Aug 03 06:08:56 Cyanogenmod Aug 03 06:09:00 the problems are differen than what folks thing Aug 03 06:09:02 think Aug 03 06:09:16 2.4 didn't get good till 2.4.6 Aug 03 06:09:21 there should be an Android kernel that's more modern than what TP uses Aug 03 06:09:28 2.6 didn't get good till 2.6.8 Aug 03 06:09:29 I remember a 2.6.38 from coreaurora Aug 03 06:10:08 okay, I'm looking at the omapdrm source files and I don't see anything that gives an indication (that I recognize) about what omap chips are supported. grrr Aug 03 06:10:08 if were up to me, I'd work on the cyanogenmod kernels, get them fully working on TP and then port userspace bits over to that kernel Aug 03 06:10:15 use their bluetooth layer Aug 03 06:10:24 and their wireless drivers as much as possible Aug 03 06:10:40 it *is* up to you... Aug 03 06:10:49 "you have the source" Aug 03 06:10:50 :D Aug 03 06:10:56 uh Aug 03 06:11:08 I have "responsibilities" Aug 03 06:11:20 like finding a job Aug 03 06:11:42 so, with the other hand and eye hack on the kernel! Aug 03 06:12:01 the problem is that all of the legacy userspace stuff hasn't been released Aug 03 06:12:25 I can't really say much about it until it's release and the support git repo for 2.6.38 and .35 is released Aug 03 06:12:26 and can't be Aug 03 06:42:47 RagingMind, OMAP4460, khronos opengl-es Aug 03 06:43:04 available in ubuntu repos Aug 03 07:20:33 ka6sox: the multimedia thing is difficult Aug 03 07:20:43 you really need an entire team for that stuff Aug 03 07:21:57 sometimes you need help from proprietary vendors as well Aug 03 07:22:30 yup Aug 03 07:26:06 just work on the basics first Aug 03 07:26:19 make sure the graphics layer is working and maybe hijack more Android stuff Aug 03 07:26:37 for native application support which is what was kind of expected Aug 03 07:28:17 just approach it as a standalone browser at first and then add more functionality as folks do a better Android integration Aug 03 12:22:48 ka6sox-away: more cookbooks! :) https://github.com/travis-ci/travis-cookbooks Aug 03 14:24:32 So is Nokia closing down (selling) Qt? ww.phonenews.com/nokia-closing-qt-australia-office-qt-asset-sale-likely-20861/ Aug 03 14:26:41 That would be interesting Aug 03 16:10:42 Starting porting over docs to owo-lunasysmgr. I've had a couple now that have not merged in well due to changes. Aug 03 16:10:56 Seem to be pretty small so far, but it is posing a slight barrier to porting them over. Aug 03 16:11:15 Issued a pull request to them, though: https://github.com/openwebos/luna-sysmgr/pull/2 Aug 03 16:48:44 tyrok: since you're going thru the new sysmgr, are you finding much different? Aug 03 16:52:08 EricBlade: Not much yet. Aug 03 16:57:53 EricBlade: Mostly pretty similar. Most of the differences I've seen have been minor ones to add to the cross-platform compatibility. Aug 03 17:40:50 anyone know if kernel modules for 3.1.10 are compatible with 3.3? Aug 03 17:56:50 do you have the source? :) Aug 03 18:40:33 Hackers. Aug 03 18:42:35 morning all Aug 03 18:49:32 Quite a few changes from string& arguments to QString&. Aug 03 18:51:28 that sounds like a positive change, I imagine Qt works best if you use it's own data types Aug 03 18:51:47 garr, vagrant's back to the old 'take 10 minutes to boot' game Aug 03 18:56:27 * ShiftyAxel kicks vagrant in the nuts Aug 03 19:06:36 All but the last 6 documentation commits are now ported over. Aug 03 19:59:46 webos-gerrit pushed to master at openwebos/luna-sysmgr 28 minutes ago Aug 03 19:59:47 76140c2 Fixes white flickers on touchpad Aug 03 20:04:52 YOU SAW NOTHING Aug 03 20:05:28 The WebApp fix is the cooler fix. Aug 03 20:06:59 webos-gerrit pushed to master at openwebos/luna-sysmgr 28 minutes ago76140c2 Fixes white flickers Aug 03 20:07:13 :) Aug 03 20:10:20 anyone manage to get it running yet without the segfault Aug 03 20:11:12 well shifty got it in tablet mode out of the box, and the rest of us got the segfault on the button press. either way, no other action seemed to do anything Aug 03 20:12:37 on touchpad eh? Aug 03 20:12:53 no, i said nothing about the touchpad. Aug 03 20:13:19 definitely nothing about the touchpad that gerrit fixed white flickers on. Aug 03 20:13:32 gerrit is not a person Aug 03 20:13:39 buuutttt...... Aug 03 20:13:43 * halfhalo has segfault fix Aug 03 20:14:28 gerrit is not a person ? o.O Aug 03 20:14:37 that's more interesting than a segfault fix Aug 03 20:15:20 open webos is actually an artificial intelligence that is fixing itself o.O Aug 03 20:15:36 SKYNET Aug 03 20:17:38 luna-sysmgr needs an appinfo.json file in a specific location to fix segfault Aug 03 20:19:33 does the button do anything with this appinfo in place? Aug 03 20:19:39 Yes. Aug 03 20:19:59 wth. the appinfo.json is _in_ the repo, but nobody modified the scripts to use it Aug 03 20:22:29 temp fix: drop the appinfo.json from desktop-support into /usr/lib/luna/system/luna-applauncher Aug 03 20:25:08 hmm, perhaps a pull request is in order Aug 03 20:25:32 I'm going to throw a brick at someone to get it fixed Aug 03 20:32:19 yay, light maiming Aug 03 20:56:20 brick throwing! Aug 03 20:56:28 that should be an olympic event. Aug 03 20:58:42 should the other files in there be copied as well? Aug 03 20:59:58 hmm. we don't have a /usr/lib/luna Aug 03 21:01:35 i suppose maybe if we ran ~/luna-desktop-binaries/install-luna-sysmgr.sh we might have different degrees of success. Aug 03 21:02:33 which i missed that line in the readme Aug 03 21:03:20 oh hey that must be the step that Shifty did that me and hadak didn't. now i'm in desktop mode Aug 03 21:03:28 and i have that home button Aug 03 21:04:44 WHOA! it worked Aug 03 21:04:49 * EricBlade cheers halfhalo Aug 03 21:05:18 all grey screen, with tabs! Aug 03 21:05:20 yay tabs Aug 03 21:06:25 i would also note that the bar with the home button definitely acts as a gesture area now too Aug 03 21:11:35 EricBlade: sweet, how so? Aug 03 21:11:49 onscreen gesture area is definately the way forward Aug 03 21:11:52 if you left-swipe on it, the launcher closes Aug 03 21:11:57 awesome Aug 03 21:12:04 but if you right swipe it doesn't Aug 03 21:12:18 i wonder if tablets will have gesture areas, or if they'll be going with bezel swipes Aug 03 21:12:18 or if you tap on it, it doesn't Aug 03 21:13:17 now how do we get apps into it :) Aug 03 21:13:23 what about the button, does it toggle the launcher away again? Aug 03 21:13:29 yes Aug 03 21:13:36 and why do the fonts look really horrible :( Aug 03 21:13:37 hmm, i think that would require isis, i wonder if the hooks are there for it... Aug 03 21:13:52 pretty sure it built an isis Aug 03 21:14:00 oh, sweet Aug 03 21:14:11 …okay, i wanna try this now, but i have no gesturebuttonarea D: Aug 03 21:47:22 Hello Aug 03 21:47:34 how i compile source codce Aug 03 22:06:28 folks Aug 03 22:06:59 ka6sox-away: cyanogenmod kernels are in good shape. I'm running 7.2 on an ancient EVO and it's fine Aug 03 22:07:22 bluetooth and other wireless drivers work great which is the main bit of having those proprietary drivers in the first place Aug 03 22:09:02 the only bit that I might have problems with is their power management, it's non-existent Aug 03 22:09:32 wake-locks in our kernels are turned off Aug 03 22:10:21 LSM may also need some other stuff moved about to get graphics Aug 03 22:17:44 I'd ignore the grpahics for now Aug 03 22:17:47 graphics Aug 03 22:18:03 just get it working first with what ever framebuffer layout they're using and then modify as needed Aug 03 22:18:33 LSM is already using opengl. The qml stuff is just not in the right location Aug 03 22:18:42 qml ? Aug 03 22:18:44 what's that ? Aug 03 22:19:06 qt stuff Aug 03 22:19:11 the other bits are the event layer and native code interface layers for telephony, etc... Aug 03 22:19:15 oh Aug 03 22:19:19 that's right Aug 03 22:19:39 using opengl is a good thing for portability sake Aug 03 22:19:52 its muy flexible. Aug 03 22:19:53 don't like qml ? Aug 03 22:20:18 qml is like html. Its shipped with the LSM, but not linked to the right place apparently. Aug 03 22:20:20 qml is favored for speedier UI implementation Aug 03 22:20:25 ok Aug 03 22:21:30 I think they have a lot of work on their hands, moving to a 3.x kernel isn't going to cut it IMO Aug 03 22:22:08 3.5 is going to be too unstable or need additional polish that they can't afford Aug 03 22:22:49 nearly all groups or company take a raw kernel and do a lot of sustaining engineering to it before releasing it as a product Aug 03 22:23:07 there's tons of VM changes in the later kernels for KVM support and stuff Aug 03 22:23:17 file systems are still in heavy development Aug 04 00:18:53 hmmm. "Android Jellybean has hidden, rudimentary support for multiple user accounts which eventually should allow you to hand your phone over to your girlfriend, boss, or kids without that uncomfortable pause" Aug 04 01:14:03 quiet day here Aug 04 02:08:03 All docs which I've pushed to my WOCE clone are now also ported to OWO-LSM. Aug 04 02:31:30 strange article on webosnation about why touchpad not supported for openWebOS--says no money but does not consider potential value to 3rd party of having existing user base **** ENDING LOGGING AT Sat Aug 04 02:59:58 2012