**** BEGIN LOGGING AT Mon Feb 21 02:59:57 2011 Feb 21 05:49:21 cooloney_, ping Feb 21 05:49:41 danders: hey, man Feb 21 05:49:51 cooloney_, hey bud Feb 21 05:50:05 danders: you changed your nickname here? Feb 21 05:50:13 doh Feb 21 05:50:19 just reloaded a pc at the house Feb 21 05:50:25 forgot to change my nick Feb 21 05:50:54 cooloney_, hey you know a matt copeland in stitsville ? Feb 21 05:52:03 cooloney_: oh, no, don't know that Feb 21 05:52:32 cooloney_: there is a matt copeland on your linkedin.com contacts that is in the ottawa area Feb 21 05:53:10 prpplague: oh, god. actually, i might not know everyone on my linkedin.com Feb 21 05:53:19 cooloney_: hehe Feb 21 05:53:34 prpplague: did he contact with you? or make you unconfortable? Feb 21 05:54:21 cooloney_: hehe, no, its a personal matter, just a weird item that his linkedin page showed that he was connected via you to me Feb 21 05:55:09 prpplague: np, let me check my linkedin. Feb 21 05:55:22 cooloney_: see /msg Feb 21 05:55:50 cooloney_: bizarre sunday evening Feb 21 12:45:01 morning Feb 21 13:32:00 is the panda es1.0 still supported without hacks? Feb 21 13:40:45 no Feb 21 13:40:53 it never was supported in a release Feb 21 13:41:03 maverick beta should run on it Feb 21 13:41:16 but es1.0 support was dropped after beta Feb 21 13:41:18 maverick beta may be hard to find Feb 21 13:41:25 why ? Feb 21 13:41:34 should be on the cdimage server Feb 21 13:41:45 Old betas get deleted without announcement Feb 21 13:42:00 bah Feb 21 13:42:02 k Feb 21 13:42:06 yeah, its gone Feb 21 13:42:28 because I have an ES1.0 board lying around Feb 21 13:43:07 Well, if you can find a suitable kernel patch that makes that work as well as the production models, without regression.. Feb 21 13:44:14 * XorA lacks ubuntu capable hardware :-D Feb 21 13:54:09 XorA, i think rsalveti kept a x-loader and u-boot biary around soemwheer (i might misremember) but i dont know about the kernel patch Feb 21 13:55:02 ogra: x-loader/u-boot easy enough to frig up from somewhere Feb 21 13:55:21 ogra: I never used a Ubuntu generated one anyway Feb 21 13:55:26 If you have a kernel that works on it (even an old one), you can probably run for a while. Feb 21 13:55:43 there is a counter builtin ? Feb 21 13:56:00 (or why "for a while") ? Feb 21 13:56:31 janimo, did you already take a look at nux ? Feb 21 13:56:40 * ogra notices that images are still failing Feb 21 13:56:45 if its more than 1/2 a days effort I probably wont attempt it, ES1.0 is missing a lot of functionality anyway Feb 21 13:56:49 Because as the kernel drifts, userspace tends to drift to match, and there is increasing potential for incompatibilities. Feb 21 13:57:08 * XorA has a convenient b1n to file old boards into Feb 21 13:57:24 persia, oh, you talk about upgrading Feb 21 13:57:37 Saving them up to deliver to the acid-etching folk for gold recovery? Feb 21 13:57:47 ogra: No, I talk about staying current. Feb 21 13:58:03 persia: recycling place has a place for old electronic Feb 21 13:58:28 what they do after that I dont know, probably illegally ship to africa or something Feb 21 13:58:47 ogra, yes, and have a merge request with the fix Feb 21 13:58:57 * ogra hugs janimo Feb 21 13:59:00 awesome !!! Feb 21 13:59:06 * janimo hugs ogra Feb 21 13:59:15 :) Feb 21 13:59:20 When I ran a electronics reclamation agency (we'll haul by the half-pallet at no charge!), it was mostly a matter of removing the metals from the boards. The chips were worthless and the boards indestructible. Feb 21 13:59:21 however a new nux needs toi be uploaded for unity to build against Feb 21 14:01:51 * XorA amusingly still has some form of ubuntu rootfs on the zoom3 Feb 21 14:02:23 That probably has better support Feb 21 14:02:35 it does? Feb 21 14:02:44 It's omap3, right? Feb 21 14:02:54 yeah, but until 2.6.38-rcX wasnt bootable in mainline Feb 21 14:03:10 ogra, persia I also filed a merge request for ubuntu-cdimage . No idea how to test that easily, it is mostly copying what is already there for the netbook project Feb 21 14:03:13 https://code.launchpad.net/~jani/ubuntu-cdimage/preinstalled-headless-arm Feb 21 14:03:20 linux-omap from natty is mainline 2.6.38-rcX Feb 21 14:03:42 persia: ah, worth a look then, I can always file a defconfig update Feb 21 14:03:53 janimo: Testing it requires setting up a complete image building environment. My general recommendation is to leave it for the cdimage team to critique. Feb 21 14:04:03 XorA: source package is "linux" Feb 21 14:04:22 persia: board is 5 miles away and turned off at moment :-) Feb 21 14:04:28 persia, indeed Feb 21 14:04:41 persia: thanks for taking care of my friday, was awesome Feb 21 14:04:44 XorA: Well then, when you have a couple hours :) Feb 21 14:04:50 but I thought I'd let you too know Feb 21 14:04:52 I'm glad you liked it. Feb 21 14:05:16 * XorA is less likely to kill all hardware designers now :-) Feb 21 14:07:05 janimo: I'm confused by the case $PROJECT bit in etc/config: I wouldn't have expected you to add arch limitation there matching i386. (I'm not cdimage, but I've looked at that code a bunch) Feb 21 14:07:22 * janimo checsk Feb 21 14:07:52 * persia is looking at https://code.launchpad.net/~jani/ubuntu-cdimage/preinstalled-headless-arm/+merge/50568 Feb 21 14:08:25 persia, I copied the case for ubuntu-netbook, that is unless ARCHES is set default to i386 Feb 21 14:08:29 XorA: ogra: I should have the pointers to make a valid kernel for it Feb 21 14:08:45 I assume for ubuntu-netbook ARCHES is set to armel Feb 21 14:08:46 just need to find at the irc log, many asked this when the es2.0 was out Feb 21 14:09:24 janimo: Right. I probably would have defaulted to armel in that case, for this image. Doesn't matter that much. Also, it's probably worth glancing at the -server or -desktop models, as examples where it pulls from a common seed (ubuntu) Feb 21 14:09:28 rsalveti: dont hunt too hard, I was just vageuly wondering, wont spend real work on it Feb 21 14:09:28 persia, I saw not place except the germinate from seed file where I thought diverging from ubuntu-netbook would make sense Feb 21 14:09:44 Ah, OK. Feb 21 14:10:10 persia, I am sure things can be done better, but not knowing the code and being aware it is full of legacy boobytraps I am not trying to do anything smart Feb 21 14:10:38 I suppress my desire to cut and refactor stuff Feb 21 14:10:48 I don't know if things can be done better, and my understanding is that this code mostly grows organically anyway. Feb 21 14:11:10 yes, lots of organic weed too there Feb 21 14:11:15 Someone should probably refactor it at some point, but it's never a priority, and the risk of getting it wrong is fairly high. Feb 21 14:12:04 right, probably Colin and a few others know what is safe to nuke but since few people work on it few are hindered by it as it is Feb 21 14:12:14 that sounds like a job for someone with asbestos pants Feb 21 14:12:16 * persia gets excited about libasound2 1.0.24.1-0ubuntu1 Feb 21 14:12:27 * XorA worries about persia Feb 21 14:12:40 XorA: Contains support for UCM Feb 21 14:12:53 * janimo google UCM Feb 21 14:12:53 (and just hit natty very recently) Feb 21 14:13:01 ah, good old Slimlogic product that :-) Feb 21 14:13:24 Yep Feb 21 14:13:52 * XorA thinks he has zero lines of code in there though, so dont blame me Feb 21 14:14:54 * XorA has had enough headaches with DAPM over the years to avoid UCM coding :-) Feb 21 14:15:50 XorA: http://paste.ubuntu.com/570052/ Feb 21 14:15:54 this is for maverick Feb 21 14:16:24 rsalveti: cheers Feb 21 14:19:55 * XorA actually saw a mythical Pandora at FOSDEM Feb 21 14:27:11 janimo, could you have NCommander review the debian-cd bits (or possibly cjwatson), i'm traveling this week and am happy if i get unity-2d done before FF Feb 21 14:28:03 ogra, sure it was more of a heads-up not a request for review Feb 21 14:28:08 janimo, looking at it, you shouldnt put the image type into the flavour name Feb 21 14:28:26 you mean preinstalled? Feb 21 14:28:35 ubuntu-preinstalled-headless should become ubuntu-headless instead Feb 21 14:28:50 I was trying to be descriptive, some of the other project names are vague Feb 21 14:29:28 and the project naming so far does not seem to follow a guideline, some are one others two words Feb 21 14:29:41 vague is good in terms of product naming: it provides semantic flexibility Feb 21 14:29:51 what if i want a d-i ios with ubuntu-headless on it = Feb 21 14:29:55 ? Feb 21 14:29:58 *iso Feb 21 14:30:01 They are all transitioning towards two words. Feb 21 14:30:18 for-project is invoked with the type attached usually Feb 21 14:30:39 So images end up being some combination of ${BRAND} ${PRODUCT} ${TYPE} Feb 21 14:30:56 e.g. "Xubuntu Desktop Preinstalled" Feb 21 14:31:06 * persia isn't sure that one will ever exist, but as an example Feb 21 14:31:10 for-project Feb 21 14:31:12 i.e. Feb 21 14:31:33 for-project ubuntu-headless cron.daily-preinstalled Feb 21 14:32:08 or a headless live image: Feb 21 14:32:14 for-project ubuntu-headless cron.daily-live Feb 21 14:33:25 persia, well in source code I prefer descriptivness over vagueness :) Feb 21 14:33:44 the flavour naming goes beyond source code ... thats the point Feb 21 14:34:10 Indeed, and as Brands and Products get more important in terms of advocacy, this trend will only continue. Feb 21 14:34:12 if you ever create a d-i iso or a dvd with headless they would be called *-preinstalled Feb 21 14:34:25 I am happy with any name actually just to get this spec done and have headless images Feb 21 14:34:26 :) Feb 21 14:34:39 right, just remove the type Feb 21 14:34:43 You want ubuntu-headless (${BRAND}-${PRODUCT}) Feb 21 14:34:45 the rest looks actually fine Feb 21 14:35:40 you could s/preinstalled/chicken/ Feb 21 14:35:42 *g* Feb 21 14:35:49 I'll do that :) Feb 21 14:35:56 that would result in ubuntu-chicken-headless images Feb 21 14:35:57 pigeon Feb 21 14:36:01 Please don't. Feb 21 14:36:04 yeah, even better :) Feb 21 14:36:13 Please pick a vaguely descriptive product name ("Headless" is good) Feb 21 14:36:26 and chicken-headless isnt ? Feb 21 14:36:42 you all running around random like? Feb 21 14:36:43 :) Feb 21 14:36:53 vague is good, but too vague requires PR coordination, which tends to cause headaches, joint pain, and premature drowsiness in coders. Feb 21 14:37:23 (and I refuse to help with PR coordination for any product with a fowl name) Feb 21 17:03:36 rsalveti, you around? Feb 21 17:15:23 sveinse: yup Feb 21 17:15:49 What's the story about powervr-omap3 ? Feb 21 17:16:15 I'm uncertain if I should build my own or use the package Feb 21 17:16:33 sveinse: well, using the one available at our repo should be fine Feb 21 17:16:42 there's a newer version, that still didn't have time to test Feb 21 17:17:00 but the changelog didn't say much, and says it wasn't tested with x11 Feb 21 17:17:22 I glanced briefly at it... The package contains (as usual) a lot of build output files :( Feb 21 17:17:37 and repeated files Feb 21 17:17:51 yeah =\ Feb 21 17:18:20 I managed to compress a 600M tar.gz to a 150M git repo, which clearly indicates that the same file is repeated many times Feb 21 17:19:05 sveinse: and were you able to test it? Feb 21 17:19:22 No, not yet. But I will in the next two weeks or so Feb 21 17:19:24 if you say it's better I can try to update it Feb 21 17:19:33 ..on 3530 Feb 21 17:19:40 ok, should be fine Feb 21 17:20:09 But it hasn't been approved for mainline yet, have it? Feb 21 17:20:23 I remember some issues with udev permissions Feb 21 17:25:09 *sigh* it's quite an effort finding the useful/needed files from the "build junk". I have to admit I'm a little surprised by the state of the package they release Feb 21 17:28:54 sveinse: to include it at ubuntu we should be fine, but first I'd like to make sure it runs well with x11 Feb 21 17:29:03 sveinse: and it's better than what we had in the past Feb 21 17:29:15 Ought test quick: FF is Thursday Feb 21 17:30:04 FF exceptions are just paperwork though Feb 21 17:30:33 They're more than that: extra testing precautions, assertions of quality, etc. Feb 21 17:30:35 I'm not working towards X11. I'm targeting Qt/QWS, so I won't be testing X11. Sorry Feb 21 17:31:24 persia: even for universe/multiverse? Feb 21 17:31:41 sveinse: that's fine, if it works for you is already something Feb 21 17:31:45 then I can test with x11 later Feb 21 17:32:03 persia, indeed implying that you do the testing etc before asking for them Feb 21 17:32:25 rsalveti: Yep. universe and main aren't really that different. The only meaningful bit is the code review part of the MIR, in my opinion. Feb 21 17:32:39 persia: ok, that's fine Feb 21 17:32:49 is the powervr license compatible for multiverse? I know I accepted an license for downloading it, but admit not reading it... Feb 21 17:33:12 sveinse: if the newer version has the same license as the one we already have at multiverse, then we're fine Feb 21 17:33:32 but the package is total mess hehe Feb 21 17:33:49 oh indeed times 10 Feb 21 17:37:22 heh I should fix that :-) Feb 21 17:37:31 no idea who to send the patches to though Feb 21 17:38:10 XorA, to rsalveti Feb 21 17:39:07 and I need to find who to send then :-) Feb 21 17:39:22 sgx-omap 4 people are not the same from sgx-omap 3 Feb 21 17:39:30 * XorA is probably not supposed to stil have the source code though Feb 21 17:39:52 Everyone who wants to be pvr-omap3 upstream, please raise your hand Feb 21 17:39:57 Im surprised anyone still making omap3 packages Feb 21 17:40:11 with TIs history of abandonment Feb 21 17:41:50 I really hope that isn't true. We're working on a prototype product based on OMAP3 and are waiting for kernel features to be implemented Feb 21 17:42:08 Like USB OTG for AM3517 and similar Feb 21 17:43:23 sveinse: The rules are probably a little different for silicon customers compared to volunteer projects Feb 21 17:43:28 sveinse: will be different if your buying chips Feb 21 17:43:41 sveinse: but for us, omap3 is effectively dead Feb 21 17:43:57 sveinse: apart from small team at ASP who do beagle Feb 21 17:44:04 Well, at least the parts dependent on binary blobs Feb 21 17:44:40 It depends. Unfortunately we're competing (sizewise) agains the mobile industry. As an industrial mfg. we are a drop in the ocean in the big picture Feb 21 17:45:30 sveinse: The comparison we usually consider is more the difference between buying one more beagleboard or overo, so your scale is hugely larger in that sense. Feb 21 17:45:52 persia: solved that issue with X and touch Feb 21 17:45:56 it was the driver Feb 21 17:46:06 lilstevie: Cool! You have a patch? Feb 21 17:46:25 * XorA would just need TI permission to work on subject, already got the NDAs Feb 21 17:46:27 XorA, rsalveti: if you send me patches I can try to "route" them inside TI (to ASP folks, maybe). Feb 21 17:46:50 (pvr patches, that is) Feb 21 17:47:17 persia, yes. The interesting note that the factor between you and us are probably in the same order of magnitude between us and the mobile industry Feb 21 17:47:31 anyway it is time to get bus home Feb 21 17:47:38 vstehle: do you have the pointers to where I can get the omap 3 code? Feb 21 17:47:49 vstehle: I keep forgetting you still hang here :-) Feb 21 17:48:04 vstehle: would be nice if we could get some bugs fixed, like the soname and etc, but then I'm not sure if I can have access Feb 21 17:48:06 sveinse: Sure, but it's probably threshold between "spare time" vs. "low-priority funded project" :) Feb 21 17:48:07 or work on it Feb 21 17:48:09 persia: no Feb 21 17:48:16 I hang "more" since I setup some highlight keywords :) Feb 21 17:48:17 lilstevie: No patch required? Feb 21 17:48:28 persia: the evdev driver isnt compatible with the type of touch panel Feb 21 17:48:40 switching to mtev solved it Feb 21 17:49:12 on the galaxy tab we now have unaccelerated graphics, touch wifi Feb 21 17:49:15 In that case, you need to fiddle the udev rules, and it's all good. Feb 21 17:49:22 rsalveti: I am not sure where the sources are. I think folks like koen use the DDK1.6 w/o DRM, currently. Feb 21 17:49:37 koen would be my first bet :) Feb 21 17:49:53 vstehle: yeah, will try to ping him Feb 21 17:49:58 aparently sound can be achieved with a little fiddling Feb 21 17:49:58 I forget he's also TI now Feb 21 17:49:59 :-) Feb 21 17:50:12 graphics are going to be the hardest though Feb 21 17:50:37 lilstevie: From our experiences with sound, 99% of the issue is driver bugs, which often aren't solved because it's far too easy to fiddle userspace. Feb 21 17:50:47 yeah Feb 21 17:51:06 persia: from what the guy said who got sound it was fiddling that got it to work Feb 21 17:51:32 unfortunantly though graphics is SGX54- Feb 21 17:51:34 540* Feb 21 17:51:37 Once the fiddling with asound.conf and pulse is done, there's usually enough information to fix the driver. Feb 21 17:52:51 Talk to Samsung: they may be able to make a binary driver available (as TI does for the omap) Feb 21 17:53:02 we have the kernel binary Feb 21 17:53:15 No kernel source? Feb 21 17:53:26 hell if you dive into nexus s kernel you have kernel source as well Feb 21 17:53:28 you should be able to get at least the kernel source Feb 21 17:53:36 we dont have the pvr kernel module source Feb 21 17:53:40 Oh, kernel binary for graphics: yeah, you'll need a new one whenever the ABI changes though :( Feb 21 17:54:06 the soic is similar to the nexus s though Feb 21 17:54:17 and that does have it as source Feb 21 17:54:32 there is talk about samsung open sourcing the driver Q4 2011 Feb 21 17:55:02 also something I have noticed Feb 21 17:55:17 android RTC and what linux expects are different lol Feb 21 18:04:39 rsalveti, which cpu's use the gfx_rel_es2.x since it's included in the opengles-sgx-omap package? Feb 21 18:04:54 sveinse: old beagles, 3430 Feb 21 18:05:03 ah, thanks Feb 21 20:14:03 Does anyone know if the Ubuntu armel toolchain will work for the ARM chips in sheevaplugs (Marvell kirkwood)? Feb 21 20:16:13 I've been told it won't, but I'm having trouble finding anything describing what versions armel covers. Feb 21 20:17:11 ubuntu armel targets minimum armv7, kirkwood is armv5 Feb 21 20:17:22 Thanks! Good to have a definite answer. Feb 21 20:29:00 akk: http://plugcomputer.org/plugforum/index.php?topic=885 Feb 21 20:29:38 "Is there any hope that ubuntu will support armv5 in the future again?" <- lol Feb 21 20:31:26 Hmm, that says Debian still supports it. Wonder if that's true, or if they just mean "because Debian won't release a new version for a long time"? Feb 21 20:31:53 lol Feb 21 20:32:23 akk: afaik debian has no plans to drop the support for armv4t, which is the minimum they support Feb 21 20:32:26 suihkulokki: right? Feb 21 20:33:43 Cool. I had assumed they didn't because of emdebian: http://emdebian.org/tools/crosstools.html Feb 21 20:34:30 which seems to be saying that you need that repo to get the cross-compilation tools. Feb 21 20:41:25 armin76: we might to got ARMv5t, if upstream gets flimsy on ARMv4t and the openmoko people don't fill in Feb 21 20:42:24 to got -> go to Feb 21 22:28:35 * XorA|gone hates mysql Feb 21 22:44:24 fyi for anyone interested, TinCanTools will be doing a production run of RevB Trainer Boards, if anyone has feedback on changes needed please submit them by the end of the week http://www.elinux.org/BeagleBoard_Trainer , Trainer works with the PandaBoard as well **** ENDING LOGGING AT Tue Feb 22 02:59:57 2011