**** BEGIN LOGGING AT Sat Dec 30 02:59:58 2006 Dec 30 04:04:42 koen|away: Any plan to bring the Angstrom ixp4xxle bootstrap image under 6.5MB? Otherwise it can't be used on an NSLU2. Dec 30 07:38:18 g'morning koen joshua_ Dec 30 07:38:43 morning Dec 30 07:38:48 * joshua_ found a bug in his IRC client Dec 30 07:53:46 good morning all Dec 30 08:01:51 good morning koen Dec 30 08:07:17 morning all Dec 30 08:59:10 koen: what is the preferred kernel version for pxa-25x machines in .dev these days? Dec 30 09:03:34 depends on the machine Dec 30 09:21:33 koen: I am working on a poodle test image. I used 2.6.18 for now and was wondering if I should be using .19 or .17 instead Dec 30 09:26:06 night all Dec 30 09:36:37 03pfalcon 07org.oe.dev * r1b4ff425... 10/ (1 packages/linux/linux-handhelds-2.6-2.6.18/hx4700/defconfig): linux-handhelds-2.6 2.6.18: Update defconfig for hx4700 to .18 version. Dec 30 09:36:41 03pfalcon 07org.oe.dev * ra8755ada... 10/ (5 files in 5 dirs): linux-handhelds-2.6 2.6.18,2.6.19: Add defconfig for htcblueangel (from HH.org CVS). Dec 30 09:36:46 03pfalcon 07org.oe.dev * rded12fd5... 10/ (1 conf/machine/htcblueangel.conf): htcblueangel.conf: Update to standard template, use linux-handhelds-2.6. Dec 30 12:26:54 morning all Dec 30 12:30:30 hey RP Dec 30 12:35:22 03koen 07org.oe.dev * rbc5647f4... 10/ (3 files in 3 dirs): qof: fix up 0.6.0 Dec 30 12:35:37 03koen 07org.oe.dev * r785b9973... 10/ (1 packages/qof/qof_0.7.2.bb): qof: update to 0.7.2 Dec 30 12:35:47 03koen 07org.oe.dev * r635b821d... 10/ (3 files in 3 dirs): gpe-expenses: add 0.0.6 Dec 30 13:21:17 xserver-kdrive-X11R7.1-1.1.0-r5 brakes while patching ?? Dec 30 13:22:06 not over here Dec 30 13:22:11 you have some local diff? Dec 30 13:22:43 nope Dec 30 13:23:24 oh whait .. it didnt find the tar.bz2 Dec 30 13:25:27 hmm fetch doesnt fetch .. Dec 30 13:25:50 it's on the angstrom mirror Dec 30 13:26:11 koen: Looks like rmk basically approves my cache fix :) Dec 30 13:36:34 RP: yay! Dec 30 13:37:15 after the pxa250 I believe every bad word people say about xscale and caches Dec 30 13:44:37 koen: This isn't an xscale issue, its just broken kernel code. Strongarm and ARMv6 are also open to the problem Dec 30 13:47:33 it's been a good months for cache and vm bugs it seems Dec 30 13:47:52 -s Dec 30 13:52:23 Koen: Chaos congres is awesome Dec 30 13:52:47 gerwinin: I believe so Dec 30 13:53:00 gerwinin: still on schedule for coming over 20070102? Dec 30 14:01:21 Koen: I think it will be a week later because I need to go to the hospital Dec 30 14:01:59 Koen : I have a goedaardige tumor in my ear that needs to be removed Dec 30 14:12:11 hi all Dec 30 14:19:52 Koen but can you mail your phonenumber to gerwin@vanderkruis.net so we can make an appointment for this Dec 30 14:20:19 koen i had only the md5 laying around but not the tar.bz2 of kdrive .. that was my problem Dec 30 14:38:50 morning Dec 30 14:40:52 hey chouimat Dec 30 14:40:57 gerwinin: mail sent Dec 30 14:41:05 lala lala lala Dec 30 14:41:06 gerwinin: and good luck at the hospital Dec 30 14:45:16 Koen: thanks Dec 30 15:05:30 yop Dec 30 15:20:35 Grrr. Mainline broke the handheld backlight drivers with general lazyness :-( Dec 30 15:36:51 Any objection to dropping linux-rp 2.6.18? Dec 30 15:46:13 do13: ping? Dec 30 15:50:27 RP: does 2.6.19 work now? Dec 30 15:50:57 XorA|sleep: work in what sense? Dec 30 15:52:50 RP: well I never managed to boot a 2.6.19 as OOM killer killed the kernel Dec 30 15:53:25 and I always go OOM on first boot Dec 30 15:53:39 XorA: I've experimented with OOM and it doesn't lock my device... Dec 30 15:54:33 RP: Ill give it another go then, now the mysterious segfaults are fixed Dec 30 15:55:18 XorA: ok. I really do want to cut down on the number of kernels we have on the go... Dec 30 15:55:29 I'm thiniking of droppping 2.6.16 from .dev too Dec 30 15:56:10 RP: I think 2.6.16 is fairly safe to drop, although it is the last version with working irda Dec 30 15:56:49 XorA: Ah, I knew I was forgetting about something. What was that problem again? Dec 30 15:58:15 RP: using Irda oopses the kernel, gimme a sec, Ill dig out the bug Dec 30 15:59:17 !oebug 1429 Dec 30 15:59:18 * * Bug 1429, Status: NEW, Created: 2006-09-21 02:21 Dec 30 15:59:19 * * dp(AT)xora.org.uk: Irda causes oops in OZ 3.5.4.1 Dec 30 15:59:20 * * http://bugs.openembedded.org/show_bug.cgi?id=1429 Dec 30 15:59:40 RP: when Ive build 2.6.19 image, Ill retest with that Dec 30 16:00:22 Ick. I don't have a device with gpe-beam on to debug :-/ Dec 30 16:01:40 RP: http://www.angstrom-distribution.org/repo/?action=details&pnm=irda-utils ? Dec 30 16:02:05 someone needs to build a mindstorms nxt test rig for Z/Ipaq machines :-) Dec 30 16:02:40 if someone donates €299 and/or a nxt set Dec 30 16:02:52 XorA: nxt works over bluetooth :) Dec 30 16:03:00 koen: ok, what do I need to put where to make a device use angstrom feeds? ;-) Dec 30 16:03:34 a super robotic button pushing regression tester :-) Dec 30 16:04:01 echo "src/gz angstrom-armv5te http://www.angstrom-distribution.org/unstable/feed/armv5te/base/" > /etc/ipkg/anstrom-armv5te-feed.conf Dec 30 16:04:17 or something similar Dec 30 16:11:41 koen: thanks, I've always wondered how to set that up but never found time to work it out :) Dec 30 16:12:45 Its now upgrading gtk on my device from 2.6 -> 2.10. This is going to end in tears :-/ Dec 30 16:19:15 RP: yeah, that will make pretty much all gtk apps segfault Dec 30 16:20:46 It reports invalid module name pxa_ir :-/ Dec 30 16:21:06 but yes, the image is totally broken now Dec 30 16:28:28 XorA: It does it under 2.6.19 too Dec 30 16:28:44 ~lart irda :-) Dec 30 16:28:45 * ibot blames irda :-) for all the evil in the world Dec 30 16:35:42 * XorA needs to get his arse round to packaged staging Dec 30 17:27:24 re Dec 30 18:17:19 RP: late pong Dec 30 18:19:08 do13: Hi, I was just wondering what the status was with mainlining Chris' patches for tosa? Dec 30 18:20:13 RP: I don't know the status. Dec 30 18:20:26 do13: Do you have an email address for him? Dec 30 18:21:24 RP: I have to dig into my mailbox Dec 30 18:22:31 do13: Ok, perhaps when you have a few minutes sometime, you could dig it out for me. I'm thinking of applying gentle pressure to try and get some of that code into mainline ;-) Dec 30 18:23:16 do13: see u @fosdem 07? Dec 30 18:24:03 RP: I'll send him a note. also to spyro. I hope we can push thing Dec 30 18:24:12 greentux: yep Dec 30 18:24:35 do13: fine. hotel is booked :) Dec 30 18:24:38 do13: Yes, I'd like to be able to. I was going to cc Ian. If you do send a mail, can you cc me please? Dec 30 18:25:40 RP: ok Dec 30 18:26:17 greentux: Are you involved into the Konnektor project? Dec 30 18:26:29 do13: :) Dec 30 18:26:59 do13: where did you worked? cant remember at the moment... Dec 30 18:27:16 greentux: thales Dec 30 18:27:35 do13: right, perhaps we do something together... how knows Dec 30 18:28:10 greentux: unlikely, I have a new job in 4 weeks :) Dec 30 18:28:57 do13: oh. these are changing times... I know many devs with new jobs... Its a secret where it would be? Dec 30 18:30:09 greentux: small company. ATM and GSM/UMTS network analysers Dec 30 18:30:39 do13: ok, nothing wit oe again ... :( Dec 30 18:32:34 greentux: marvell network processor. Migrate Vxworks apps to linux :) Dec 30 18:33:13 do13: oh fine. So its good for You I assume. same city? Dec 30 18:34:03 greentux: no. a 10km more Dec 30 18:35:23 do13: thats ok Dec 30 18:38:13 greentux: Are the jobs offers on your homepage still valid? a colleague asked me Dec 30 18:38:40 which one do You mean? some job sI know Dec 30 18:39:02 hardware devloper Dec 30 18:40:34 in dresden. yes, its open. the colleague can direct talk to me. Dec 30 18:41:11 thx Dec 30 18:43:08 greentux: your security evaluator is a very interesting idea :) we pay external companies Dec 30 18:44:02 do13: which job do you mean now? Dec 30 18:44:49 greentux: no job, the company in switzerland:) Dec 30 18:45:12 do13: aaaahhh. yes, sometimes I think about moving to switzerland Dec 30 18:45:32 greentux: very expensive Dec 30 18:45:53 do13: but will get more money... Dec 30 18:46:11 greentux: that's an argument Dec 30 18:47:03 do13: no, but I worked in Switcherland for 1 year, I love that country. Some hours, You are in Italy. You have nice nature for my hobby (gliding). Dec 30 18:47:38 cool Dec 30 18:48:48 hi Dec 30 18:49:01 hi marcin Dec 30 18:50:22 heh.. called me for food.. have to go Dec 30 18:51:15 later all Dec 30 18:51:49 RP: 2.6.16 is known to be good working on clamshells, 2.6.17 was good for poodle/tosa. I do not know about 18 and 19 ones Dec 30 18:52:07 hrw: Does .17 have issues on clamshells? Dec 30 18:52:16 RP: never tried Dec 30 18:52:33 RP: I have 18 on spitz but it has hostap problem Dec 30 18:53:16 We need to cull some kernel versions. I'm open to suggestions as to which ones but we have too many to maintain atm Dec 30 18:53:18 hope to get time for testing .19 on my devices. Dec 30 18:54:37 RP: if .18 or .19 works then .16/.17 can be dropped Dec 30 18:55:50 I'll probably end up dropping .16 and .18 first, then .17 if we can prove .19 is ok Dec 30 18:56:06 got conenction drop ;( Dec 30 18:56:13 RP: ok Dec 30 18:56:56 RP: we need^Whave^Wcan to make a call for testers Dec 30 18:57:29 but I want to have results from devels first so will know which one to choose for oz 3.5.5 Dec 30 18:57:48 now have to go - cu you probably in next year Dec 30 18:57:55 hrw: I'll remove .16 and .18, then it narrows the options a bit :) Dec 30 18:58:01 sure RP Dec 30 18:58:11 hrw|gone: cu, have a good NY :) Dec 30 19:13:12 morning all Dec 30 19:13:30 hey likewise Dec 30 19:13:41 * koen suddenly realizes it's saterday Dec 30 19:13:46 ~spell saterday Dec 30 19:13:50 possible spellings for saterday: Saturday steady sated Saturdays sturdy satiety sauteed stead seated stater steed satay stray sited study yesterday Saturday's Dec 30 19:15:16 as opposed to friday? Dec 30 19:15:35 I had a 'wednesday' feeling this morning Dec 30 20:02:28 hey Dec 30 20:03:31 hey Dec 30 20:04:41 hi Dec 30 20:06:32 zecke|not23c3: what do you use to write bitbake documentation? you are using docbook, right? do you write the sgml yourself or do you use some wysiwyg editor? Dec 30 20:08:05 Jin^eLD: vim Dec 30 20:08:20 sometimes GNU emacs Dec 30 20:09:07 I see.. Dec 30 20:09:17 I am looking into Lyx but no luck yet Dec 30 20:16:27 koen: re ixp4xxle feeds - are there meant to be machine/ixp4xx (containing kernel modules) and machine/ixp4xxle (containing kernel images) separate feeds ? Dec 30 20:37:16 rwhitby: no, it's a leftover from the arch mess ixp4xx.conf had Dec 30 20:37:37 koen: ok, thx. So which one will remain? Dec 30 20:38:57 koen: what's your opinion on removing things from the ixp4xxle rootfs to bring it under 6.5MB ? (so it will fit on 8MB flash devices) Dec 30 20:40:23 rwhitby: I'm all for slimming it down Dec 30 20:41:03 ok, I'll look at making some suggestions for what's not useful on nslu2 at least, and let you review those for general ixp4xxle usefulness. Dec 30 20:41:17 rwhitby: it's a job for task-base Dec 30 20:41:37 * koen tries to shake rwhitby out of the nslu2/slugos paradigm Dec 30 20:42:06 rwhitby: you guys were way to modest with slugos :) Dec 30 20:42:20 lots of things would have been an improvement to OE in general Dec 30 20:43:39 well, we're going to do it right this time with Angstrom Dec 30 20:44:22 start again from scratch, and build the features back up as general OE features Dec 30 20:44:32 hey mickey|unpacking Dec 30 20:44:41 cheers guys Dec 30 20:45:54 mickey|unpacking: how was 23c3? Dec 30 20:46:02 hey mickey|unpacking: did you see that nslu2 guys and openprotium guys are going to work on Angstrom NAS features? We've recruited HopsNBarley into the nslu2 core team. Dec 30 20:46:32 koen: pretty cool, although i had a problem scheduling between the congress and Sabine Dec 30 20:46:40 and working together with LaF0rge on stuff... Dec 30 20:46:50 mickey|unpacking: I can imagine Dec 30 20:46:52 rwhitby: i didn't see that yet, but it's much appreciated! Dec 30 20:46:59 rwhitby: really cool Dec 30 20:48:13 koen: current Angstrom images boot on nas100d (it has 16MB flash) and I can get networking up, install packages, and even mount an ext3 filesystem from the internal pata-artop disk or an external usb disk. Dec 30 20:48:43 (all without any slugos-specific things :-) Dec 30 20:49:03 yay Dec 30 20:49:26 rwhitby: pretty cool Dec 30 20:50:10 rwhitby: it could turn out that task-base is unfit for such images, but I think it should be able to handle it Dec 30 20:50:57 ~seen lennert Dec 30 20:51:21 lennert was last seen on IRC in channel #openjtag, 52d 10h 46m 28s ago, saying: 'probably LE as that's what other folks use'. Dec 30 20:52:49 koen: do you mean that task-base should be used for such images, but needs work, or do you mean that something other than task-base should be developed? Dec 30 20:54:33 I think we should try task-base and if needed fix it Dec 30 20:54:49 and if it is unfit, developed something else Dec 30 20:54:57 ok Dec 30 20:55:48 has there been a general way decided to handle whether or not the zImage file (not the package, the file) is included in the rootfs? Dec 30 20:55:56 I'm interested in the uclibc initramfs as well Dec 30 20:56:08 rwhitby: there was a thread about that on oe-devel about that iirc Dec 30 20:56:29 koen: yeah, I read the thread - wondering if there was a resolution? Dec 30 20:56:47 not that I know off Dec 30 20:57:32 * koen is 2m UTP short of having POE working on his loft Dec 30 20:57:34 it almost seems like there needs to be a separate package for the image file itself, and a separate package for the "virtual" image file which other packages depend on (e.g. kernel modules) Dec 30 21:05:07 03koen 07org.oe.dev * r0cfaa91f... 10/ (3 files in 2 dirs): Dec 30 21:05:07 mx[23]1-kernel: use a proper PV Dec 30 21:05:07 **** DANGER **** PV went backwards **** DANGER **** Dec 30 21:29:40 can a non eabi kernel compile with gcc like 3.4.4 boot over ( busybox) init to a filesystem which was compiled with eabi ? Dec 30 21:30:39 can this eabi thing at all cause strange init problems ? Dec 30 21:33:56 'no' and 'depends on what version of busybox you ise' Dec 30 21:35:55 i will try init=/bin/busybox now .. lets see what drops out Dec 30 21:36:30 an eabi kernel built in OE should be able to boot eabi and oabi Dec 30 21:36:38 koen, but your answer to my first question was "no" , correct ? Dec 30 21:36:43 a oabi kernel can only boot oabi Dec 30 21:36:47 rob__w: correct Dec 30 21:36:55 so Dec 30 21:43:29 XorA: It's fairly certain I'll be visiting warm and sunny Edinburgh this summer Dec 30 21:44:47 Edinburgh... Edinburgh... that rings a bell Dec 30 21:45:18 home of wolfsom micro :) Dec 30 21:46:19 and I heard a debconf is taking place there in the near future Dec 30 21:47:13 mickey|unpacking: still unpacking?!? Dec 30 21:49:25 no Dec 30 21:49:33 bzcat mickey.bag.bz2 | /dev/washingmachine ? Dec 30 21:49:33 rather reading the IRC + Mail backlog :) Dec 30 21:49:46 likewise: hehe, yeah. Dec 30 21:49:52 hmmm Dec 30 21:50:02 * mickey|unpacking waiting for wife being ready for dinner... Dec 30 21:50:06 time to do s/_cvs/_svn/g for all gnome stuff Dec 30 21:50:07 *hungry* Dec 30 21:50:12 oh finally Dec 30 21:50:13 cool Dec 30 21:50:17 bon appetit Dec 30 21:50:32 thanks Dec 30 21:50:34 mickey|unpacking: only 2 days left till release ;) Dec 30 21:50:41 *cough* yeah, right :) Dec 30 21:50:51 what is a : -hh version mismatch again ? Dec 30 21:52:34 rob_w: the .bb doesn't match the cvs checkout Dec 30 21:53:07 found it , thanks Dec 30 22:09:07 mr_mice: I add a comment in the wiki Dec 30 22:11:43 wow that nokia 770 is sweet! Dec 30 22:13:05 anyone: what is the current status of uclibc support over the range of currently supported machines? Dec 30 22:13:48 likewise: outdated Dec 30 22:13:56 * koen is pining for .29 Dec 30 22:14:27 koen: what will be the smartest way to pick it up? Look at other projects configure stages etc? Dec 30 22:14:45 likewise: try looking at the uclibc_svn recipe Dec 30 22:15:34 koen: The cross chain part of OE is still a bit too hairy for me. I am used to doing this manually, or using crosstool, at least as a reference. Dec 30 22:15:45 ...toolchain... Dec 30 22:16:08 koen: Also, I am a bit confused by the initial and intermediate libc stages. Dec 30 22:16:16 likewise: the problem with uclibc is its (bogus) dependance on kernel-headers Dec 30 22:16:40 koen: isn't that the same with glibc? Dec 30 22:16:50 to build a threaded libc you need a gcc with thread support which needs glibc Dec 30 22:17:06 likewise: linux-libc-headers is used for that Dec 30 22:17:57 so to build *libc you need gcc-cross which needs *libc-intermediate, which need gcc-intermediate which need *libc-initia; Dec 30 22:18:04 (or something like that) Dec 30 22:20:25 and gcc-cross-initial builds gcc-cross-intermediate?? Dec 30 22:21:14 yes Dec 30 22:22:51 koen: I cannot remember crosstool needing three gcc stages, so I'll first get into crosstool again. Dec 30 22:23:01 * VoodooZ is taking notes... Dec 30 22:23:06 likewise: it's only needed for nptl support Dec 30 22:23:47 koen: that might explain I never worked with 3 stages. Dec 30 22:24:00 arm and mipsel can do it in 2 stages Dec 30 22:24:05 powerpc seems to need 3 Dec 30 22:24:10 and iirc x86 as well Dec 30 22:26:21 so for arm you would use *libc-initial only? Dec 30 22:32:18 VoodooZ, koen: This explains everything very detailed and perfectionistic: http://www.diy-linux.org/reference-build/ (only PPC, x86-64 and i386). It also aims for NPTL. Yes, three GCC stages. Dec 30 22:33:04 very nice. I'm using ARM but I guess x86 is more common so worth learning. Dec 30 22:33:45 VoodooZ: However, it is *not* explaining the cross stuff... Dec 30 22:34:05 * VoodooZ nods Dec 30 22:34:28 VoodooZ: so I think I will keep http://trac.cross-lfs.org/ opened as well :-) Dec 30 22:34:47 good. I'm bookmarking both for later reference. Dec 30 22:36:40 Morning VoodooZ Dec 30 22:36:54 hey! Happy new year! Dec 30 22:37:09 03pfalcon 07org.oe.dev * r886f590e... 10/ (1 conf/machine/htcblueangel.conf): htcblueangel.conf: Has 128Mb RAM. Dec 30 22:37:23 03pfalcon 07org.oe.dev * r62b36589... 10/ (1 packages/linux/linux-handhelds-2.6_cvs.bb): linux-handhelds-2.6 cvs: CVS HEAD is at 2.6.19-hh7 now. Dec 30 22:37:50 * NAiL tries to crosscompile a toolchain manually Dec 30 22:37:56 daring!! Dec 30 22:38:27 NAiL: me too, for what target? Dec 30 22:38:54 koen : I got this output: what am I to do ? Dec 30 22:38:56 http://pastebin.com/848133 Dec 30 22:39:13 likewise: I'm trying to build a toolchain for powerpc-linux-uclibc. Not a cross toolchain, but a native one. Dec 30 22:39:32 (well, obviously this involves building a cross toolchain somehow) Dec 30 22:39:34 NAiL: ok, i want that one cross :-) Dec 30 22:39:40 koen: wicked, we can get you pissed Edin stylee Dec 30 22:40:02 likewise: I am very new here Dec 30 22:40:03 likewise: I've got one using buildroot Dec 30 22:40:03 XorA: :) Dec 30 22:40:19 yet the cross thing seems interesting Dec 30 22:40:31 likewise: download the latest buildroot, make menuconfig, make. Ta-da ;) Dec 30 22:40:32 please leave a post as to how you did it.. Dec 30 22:40:37 niv_one_three: that means opie is broken Dec 30 22:40:49 koen: thanks Dec 30 22:40:56 I am so new Dec 30 22:40:57 .. Dec 30 22:41:07 NAiL: yes, but I want to see what happens inside (which configure options, which seds, which patches) and then see how this would work in OE. Dec 30 22:41:26 koen: and take you to all the "interesting" bars :-) Dec 30 22:41:33 NAiL: but I guess buildroot or crosstool are my best bets. I am now looking into crosstool first. Dec 30 22:41:44 XorA: me + my girlfriend Dec 30 22:42:22 likewise: ah, well that's gonna take a while I guess ;) Dec 30 22:42:54 koen: :-D Dec 30 22:43:08 note to self, dont get koen dumped Dec 30 22:44:04 * koen heads to bed now Dec 30 22:44:10 sleep tight Dec 30 22:44:40 which is hard, we have 110 km/h winds over .nl now Dec 30 23:13:23 ping kergoth Dec 30 23:49:32 ~seen hrw Dec 30 23:49:51 hrw was last seen on IRC in channel #oe, 4h 52m 4s ago, saying: 'now have to go - cu you probably in next year'. **** ENDING LOGGING AT Sun Dec 31 02:59:57 2006