**** BEGIN LOGGING AT Mon Aug 28 02:59:56 2006 Aug 28 06:19:21 morning all Aug 28 06:45:14 hello Aug 28 07:26:11 good morning all Aug 28 07:36:32 hey mikearthur Aug 28 07:41:06 03koen 07org.oe.dev * rf91914d2... 10/ (1 conf/distro/include/angstrom.inc): angstrom: tidy up angstrom.inc Aug 28 07:41:11 03koen 07org.oe.dev * rf44d275b... 10/ (1 conf/machine/h5xxx.conf): h5xxx: add apprpriate MACHINE FEATURES Aug 28 07:49:10 03koen 07org.oe.dev * r48b6fb9f... 10/ (1 conf/machine/ipaq-pxa270.conf): ipaq-pxa270: add appropriate MACHINE FEATURES Aug 28 07:57:00 morning Aug 28 07:57:45 morning all Aug 28 07:58:50 hey XorA & Dirk Aug 28 07:59:09 ~lart gdm for having 550MB of logfile Aug 28 07:59:10 * ibot pulls out a ClueBat (tm) and thwaps gdm for having 550MB of logfile Aug 28 07:59:25 hi koen, XorA Aug 28 08:06:25 http://tslib.berlios.de/ Aug 28 08:08:26 kergoth: yay! Aug 28 08:09:21 :) Aug 28 08:09:47 hi Aug 28 08:11:37 ~lart display manager Aug 28 08:11:37 * ibot turns display manager into a lifesized tux doll Aug 28 08:15:54 kergoth: do we want .la files in the -dev packages? Aug 28 08:23:11 kergoth: what do you think about this: http://www.openembedded.org/~koen/OE/tslib-release.diff ? Aug 28 08:24:53 koen: looks fine Aug 28 08:25:29 kergoth: I am a bit stumped by the fact I had to remove the CONFFILES Aug 28 08:25:40 which is why I didn't commit it Aug 28 08:26:34 hmm, what happens if you dont? Aug 28 08:27:25 it bombs out on packaging because /etc/ts.conf doesn't exist Aug 28 08:43:47 RP: I think I found a dependancy bug in bitbake 12.6.0 Aug 28 08:43:51 1.6.0 Aug 28 08:43:58 crap my net connection is laggy today Aug 28 08:47:25 Yeah, I have a 7Mb/s hookup here Aug 28 09:00:03 linux/drivers/input/tslibdev.c is obsolete ? isn't it ? Aug 28 09:05:17 morning all Aug 28 09:05:54 morning :) Aug 28 09:05:58 hey Liam Aug 28 09:14:47 hey lrg Aug 28 09:22:15 hey lrg Aug 28 09:23:01 meh Aug 28 09:24:33 ~lart network in damp string mode Aug 28 09:24:34 * ibot readies the nuke launcher and fires some rounds at network in damp string mode Aug 28 09:30:20 03koen 07org.oe.dev * r0fea0578... 10/ (1 packages/tslib/tslib_1.0.bb): Aug 28 09:30:20 tslib 1.0: use tarball instead of svn and fix packaging Aug 28 09:30:20 * CONFFILES = /etc/ts.conf has been removed Aug 28 09:34:11 morning all Aug 28 09:34:30 hey RP Aug 28 09:35:07 hey Aug 28 09:36:09 XorA: dependency bug? Aug 28 09:36:47 hey RP Aug 28 09:37:09 hey Richard Aug 28 09:41:34 morning Aug 28 09:41:41 hi hrw Aug 28 09:41:46 RP: nice changess during weekdn Aug 28 09:42:28 hrw: Its a start :) Aug 28 09:42:48 * T0mW thinks he needs another gig of SDRAM to build oe Aug 28 09:44:32 morning Aug 28 09:46:05 hi CoreDump|home Aug 28 09:47:47 T0mW: hey :) Aug 28 09:47:51 meh Aug 28 09:48:13 kergoth: hey, figured I'd give it a try now that you've worked on it awhile. Aug 28 09:48:14 RP: note to self: careful with scripts that use rm -rf Aug 28 09:48:22 * kergoth finally got most of his homedir recovered :P Aug 28 09:48:56 kergoth: plus, buildroot just couldn't cope with the new project. Aug 28 09:48:58 kergoth: I have a rule of avoding using that in scripts ;-) Aug 28 09:49:22 http://www.openembedded.org/oedem reformatted Aug 28 09:49:49 heh Aug 28 09:49:56 RP: its a good rule Aug 28 09:50:01 "Buy Beer for Marcin." Aug 28 09:50:18 T0mW: aye, buildroot is nice for some things, but it isnt for everything Aug 28 09:50:23 heh Aug 28 09:52:12 kergoth: I stopped at the point that I was trying to merge the .config stuff for busybox, packages, and apex into a "grand unified config file" hehheh Aug 28 09:52:24 RP: xmms RRECOMMANDS xmms-mad, xmms-mad DEPENDS on xmms, bitbake attempts to build xmms-mad first when in my mind it should be other way round Aug 28 09:52:46 XorA: bitbake trunk just errors out on a dep-loop Aug 28 09:53:25 Right, that changes in trunk and causes an error as debian.bbclass makes it impossible :-/ Aug 28 09:53:35 T0mW: guh, been there, done that :) Aug 28 09:53:45 nod Aug 28 09:54:00 koen: that situation is solvable though as an R* doesnt have to be built before another package Aug 28 09:54:15 kergoth: so far, it looks interesting Aug 28 09:54:15 oe and bitbake make maintainance a great deal easier on a project in my opinion, and the shared metadata amongst all these projects is a great boon Aug 28 09:55:01 * RP notes he's becom a chair at oedem :) Aug 28 09:55:04 just need to remember to either lock down your own known working copy of the metadata repository, or create a seperate minimal metadata repository for your project with just the bits you need from upstream Aug 28 09:55:17 kergoth: it is a bit disconcerting not to have makefiles to paw around in inside the package tree Aug 28 09:55:24 hehe Aug 28 09:56:58 kergoth: is this like: ${PN}-clean, ${PN}-install, ${PN}-dirclean sort of approach, a command target naming for the makefile targets? Aug 28 09:57:13 s/command/common Aug 28 09:57:49 RP: is today a holiday in the uk? ohand seems awfully quiet this morning Aug 28 09:58:22 kergoth: yes, it is. Aug 28 09:58:25 ah Aug 28 09:58:51 well, guess i've got some time to finish reconstructing my dev env then :P Aug 28 09:58:59 * kergoth plays with scratchbox some more Aug 28 09:59:32 mikearthur: how goes touchscreen fixing? Aug 28 09:59:45 RP: http://tslib.berlios.de/ :) Aug 28 10:00:08 RP: want to do me a favor and announce the release on arm-linux? i havent been on those lists in ages Aug 28 10:00:12 heh Aug 28 10:02:13 kergoth: ok, anything in particluar you want me to say or just make something up? Aug 28 10:02:45 kergoth: you know that scratchbox is able to use OE toolchains now, right? Aug 28 10:02:59 nothing special, you can paste the description from the site for anyone that doesnt know what it is, just say its finally released for the first time Aug 28 10:03:06 koen: yeah, i've seen the pages on the subject Aug 28 10:03:39 congratulation :) Aug 28 10:03:53 koen: I did some more testing, and it works really well - except for c++ side of things :-F Aug 28 10:04:40 suihkulokki: c++ sucks Aug 28 10:05:50 hrw: it's going Aug 28 10:06:18 koen: I agree. In this particular case because g++ doesn't obey -isystem Aug 28 10:06:25 hrw: problem is, I'm a bit of a noob, and the tosa_fb/lcd doesn't seem to have many of the necessary hooks, so its taking a while Aug 28 10:07:05 ~lart xfs Aug 28 10:07:05 * ibot pulls out a ClueBat (tm) and thwaps xfs Aug 28 10:08:43 CoreDump|home: ~lart linux-2.6 as is Aug 28 10:09:20 2.6 is just fine, but the 2.6.17.6 bug cost me dearl Aug 28 10:09:26 dearly Aug 28 10:09:32 mikearthur: good - I delayed OpenZaurus 3.5.4.2-rc2 for it Aug 28 10:09:36 CoreDump|home: I have xfs too Aug 28 10:10:11 RP, CoreDump|home, koen, kergoth: we got another poodle offer - who can use it? Aug 28 10:10:13 hrw: think lrg is gonna give me a hand, so it shouldn't be too long, sorry :s Aug 28 10:10:48 * CoreDump|home dunnos Aug 28 10:11:06 hrw: new uni year is starting, so I won't have much time the next two months Aug 28 10:11:20 do13_: are you around? Aug 28 10:11:47 koen: I have one poodle already - this one had to go to mickeyl on oedem Aug 28 10:13:52 i think i still have the poodle jason perlow sent me Aug 28 10:14:01 hrw: oh well, at least boostrap builds again =) Aug 28 10:14:01 the one that says not yet approved for public release or whatever :P Aug 28 10:14:36 RP: are you around? Aug 28 10:14:42 mikearthur: I am Aug 28 10:15:01 RP: do you know anything/much about the tmiofb driver? Aug 28 10:15:07 and/or tosa_lcd Aug 28 10:15:23 mikearthur: I know nothing about tmiofb Aug 28 10:15:39 RP: sucks to be me Aug 28 10:15:58 I guess just wait for dirk to pop his head up Aug 28 10:16:14 Is tosa_lcd based on corgi_lcd? Aug 28 10:16:30 If not, I know nothing about that either! Aug 28 10:17:14 mikearthur: Just imagine having to work all this out from the Sharp 2.4 code - believe me, you have it easy ;-) Aug 28 10:17:41 RP: its vaguely basd Aug 28 10:17:51 RP: vaguely based Aug 28 10:18:15 iirc the source for the production tosa kernel wasn't released Aug 28 10:18:17 mikearthur: yep Aug 28 10:18:17 RP: problem is, I'm trying to get the hsync_len, so I need to pass devices around, but they aren't available like in the corgi_lcd Aug 28 10:18:33 do13_: pm, as this may get lengthy... Aug 28 10:18:55 mikearthur: You'll probably need to add them, or find them Aug 28 10:19:00 aye Aug 28 10:19:11 kergoth: prerelease poodle? Aug 28 10:23:06 good morning Aug 28 10:23:11 I have problems w/ compiling gconf-dbus_cvs.bb in the oz354x branch. configure is looking for some packages on the host: http://rafb.net/paste/results/VrEUij15.html Any idea what those packages might be? I tried all obvius dbus and orbit -dev stuff :\ Aug 28 10:24:44 CoreDump|home: check config.log Aug 28 10:25:12 hi florian_kc Aug 28 10:25:13 hrw: yeah, it was one of the ones before the fcc approved it :P Aug 28 10:25:20 from when jason worked for sharp, back in teh day Aug 28 10:25:59 too bad that it has only 32m ram.. Aug 28 10:26:09 with 64 it would be collie killer. Aug 28 10:26:24 hrw: solder some more in Aug 28 10:26:55 yeah no doubt Aug 28 10:27:17 XorA: not with my solder iron Aug 28 10:27:42 hrw: make it upto 128MB for a super PDA Aug 28 10:27:57 i kinda miss having a pda like device to hack on Aug 28 10:28:00 i should pick up a 770 Aug 28 10:28:14 XorA: more ram with broken cache? Aug 28 10:28:18 hehe Aug 28 10:28:22 XorA: pxa250 :( Aug 28 10:28:29 koen: he might be lucky and have the unborken Aug 28 10:28:29 kergoth: read RMK reply on tslib release Aug 28 10:28:39 XorA: I have pxa250 poodle Aug 28 10:28:44 shitty slow Aug 28 10:28:54 kergoth: I see that you've moved away from using my CVS server - would people Aug 28 10:28:54 mind keeping me informed of such events? Aug 28 10:28:54 Can I kill the old tslib cvs repository and remove the accounts used Aug 28 10:28:54 to access it? Aug 28 10:29:08 ah. you got copy ;D Aug 28 10:29:28 heh Aug 28 10:29:39 ~lart Polish gov laws.. Aug 28 10:29:39 * ibot resizes Polish gov laws..'s terminal to 40x24 Aug 28 10:29:43 just reading that mail Aug 28 10:29:49 creating own company takes too much time Aug 28 10:30:26 over here you have to pay duties and stuff for the complete year Aug 28 10:30:33 even if you would found it in december.... Aug 28 10:30:35 havent gotten a mail from him yet, and i'm not on the arm linux lists anymore Aug 28 10:30:38 need to resubscribe Aug 28 10:31:20 kergoth: subscribe, tell mailman to not deliver mails and subscribe to the gmane group Aug 28 10:31:59 I use that setup to work around thunderbirds inability to search in imap folders Aug 28 10:32:21 is there a decent linux newsreader nowadays? i ahvent used usenet in ages Aug 28 10:32:30 i think the last news reader i used was tin :P Aug 28 10:32:50 I use thunderbird for that :) Aug 28 10:32:54 koen: http://hentges.net/tmp/config.log I don't see anything obvious :\ Aug 28 10:32:57 ah. hehe Aug 28 10:32:57 (on OSX, mind you) Aug 28 10:33:02 how is it? Aug 28 10:33:10 i havent tried thunderbird in a while Aug 28 10:33:10 tin was nice Aug 28 10:33:31 kergoth: thunderbird, knode (kde), pan (gtk), slrn (slang curses) Aug 28 10:33:31 :) Aug 28 10:34:11 I remember time then I build newer pine and tin and send it to root@localhost to install. and they did it. Aug 28 10:34:22 hello, is this the right place to enquire about OpenZaurus bugs? Aug 28 10:34:22 noone asked me how did I build it with 1.5M quota ;D Aug 28 10:34:29 hahah Aug 28 10:34:43 good ole shell accounts, i miss getting online with just one of those Aug 28 10:34:48 slip/ppp Aug 28 10:34:58 kergoth: thunderbird is pretty nice, if you don't mind it being slow when using imap Aug 28 10:35:17 i'm used to using mutt for mail, maybe i'll try slrn Aug 28 10:35:43 kergoth: it was my account when I was study. 80x25 terminal (HP 2623A)... Slowaris Aug 28 10:38:06 i liked solaris :) had some nice stuff Aug 28 10:38:08 later all Aug 28 10:38:10 cant say i'm a fan of cde Aug 28 10:38:24 tip kicked ass for serial stuff Aug 28 10:38:40 besides, gotta love the sparc keyboards :) fuck capslock Aug 28 10:38:47 * kergoth gets nostalgic Aug 28 10:39:18 the type6 keyboards are pretty neat Aug 28 10:39:28 I want a type 7.. Aug 28 10:39:37 too bad it's hooked up to a 110MHz sparc5 Aug 28 10:40:06 heh.. sun machines is something which I never used directly Aug 28 10:40:11 which isn't supported by openslowaris Aug 28 10:40:15 I could really use the extra two columns of keys on the left side of the keyboard Aug 28 10:41:04 uhm, I was wondering if someone more knowledgeable about OpenZaurus bugs knew if there was a bug opened yet on the resume not working as reported on the OESF forums Aug 28 10:41:23 if not, I'll gladly open a bug Aug 28 10:41:37 ehr, suspend not working Aug 28 10:42:34 freaknl: on which machine/distro? Aug 28 10:43:07 C3200 / OZ 3.5.4.1 Aug 28 10:43:13 OZ/GPE Aug 28 10:43:51 freaknl: ask on #openzaurus Aug 28 10:43:59 thanks Aug 28 10:51:06 I'll go out - cu Aug 28 10:51:18 time to gain new infos on creating own business.. Aug 28 10:51:34 its suxx that I have to take free day to find all needed infos.. Aug 28 11:10:21 The OE bugtracker won't send me a new account password (or it is taking a very long time) Aug 28 11:10:38 is it supposed to take more then fifteen minutes? Aug 28 11:22:00 I understand what conf/machine is about, but what is all those choices in conf/distro about??? Aug 28 11:22:54 T0mW: distribution based choices Aug 28 11:23:04 I keep getting a "/usr/bin/ld: cannot find -lc" for the module-init-tools-cross build Aug 28 11:23:34 koen: is there that much difference? Aug 28 11:23:51 T0mW: a device is hardware, a distro is software Aug 28 11:24:40 T0mW: compare redhat to debian and you'll get an idea what goes into conf/distro Aug 28 11:24:54 I can see package installations differing as to specifics, but, what else? I guess uClibc vs. glibc, maybe busybox init vs. sysvinit? that sort of thing? Aug 28 11:25:45 debian renaming or not, package format, devfs or udev, etc Aug 28 11:26:09 package versions, etc Aug 28 11:26:31 would there be something like a "class viewer", ala object languages, where you can get a broader view of what a machine type or distro choice encompasses? Aug 28 11:27:12 something like bitbake-tree Aug 28 11:28:21 I'm just getting started with OE and am looking to place the buildroot stuff for my project into OE Aug 28 11:29:28 the h1940 is pretty close to the board that I have, processor and such. So, I thought that I would build that using task-essential-to-boot Aug 28 11:30:05 but each distro I try comes up with a newer error Aug 28 11:30:11 ehm Aug 28 11:30:22 if you switch distro you have to wipe tmp Aug 28 11:30:37 since different distros use a different floating point format Aug 28 11:30:47 and a different gcc, glibc, binutils, etc Aug 28 11:30:50 oh no! and rebuild the whole locale pile again?! Aug 28 11:30:56 yes Aug 28 11:31:07 * T0mW mumbles something Aug 28 11:31:33 choose one distro (generic is a sane choice) and report errors with that Aug 28 11:31:50 we can't help you if you change your config every 5 minutes Aug 28 11:32:04 ok, say I stay with one distro, and I get an error. How does one find where / what bitbake was doing when it broke? Aug 28 11:32:27 nm Aug 28 11:32:36 I'll goto the bitbake docs Aug 28 11:32:42 pastebin the error and show us :) Aug 28 11:32:46 ~pastebin Aug 28 11:32:48 rm -rf tmp is ok? Aug 28 11:32:56 i heard pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.com/ (BROKEN AND SUCKING NUTS), or http://pastebin.ca, or http://channels.debian.net/paste, or for #oe use http://oe.pastebin.com, or http://bzflag.pastebin.ca/, or http://paste.lisp.org/ for the lisp/scheme nuts, or http://paste-it.net Aug 28 11:33:01 T0mW: yes Aug 28 11:33:03 k Aug 28 11:33:05 thanks Aug 28 11:33:27 T0mW: http://www.openembedded.org/user-manual is also a good read Aug 28 11:33:34 (the sections that aren't a stub) Aug 28 11:33:36 I don't expect to get this working today. Maybe I should take the zarus example to build that. Aug 28 11:33:51 btw, if you change machine do you have to blow tmp away too? Aug 28 11:34:21 not if your distro inherits multimachine.inc Aug 28 11:35:11 I figured that the only penalty that I would have in changing distro would be just a larger set of precompiled binaries laying around. Not a big deal since I have 185G free Aug 28 11:35:53 I have TMPDIR=/data/build/koen/OE/build/tmp/${DISTRO} Aug 28 11:36:07 that allows me to switch distro without running into API/ABI problems Aug 28 11:36:25 ok, I read through that user-manual. Aug 28 11:36:34 at the cost of taking more space Aug 28 11:37:25 no big deal then, I have tons of storage space. Aug 28 11:38:13 koen: thanks Aug 28 11:38:37 one more thing, how stable is generic-uclibc ? Aug 28 11:40:05 no idea Aug 28 11:40:22 * koen is still waiting for a new uclibc release that support EABI Aug 28 11:40:34 I'm developing for arm920t, samsung s3c2410 Aug 28 13:16:46 <[TEHb]> Hallo Guys Aug 28 13:17:02 <[TEHb]> eFfeM-log: ping Aug 28 13:23:15 <[TEHb]> Guys, error happens when building glibc 2.3.6 with gcc 4.1.0 Aug 28 13:23:17 <[TEHb]> ../linuxthreads/sysdeps/unix/sysv/linux/raise.c: In function 'raise': Aug 28 13:23:17 <[TEHb]> ../linuxthreads/sysdeps/unix/sysv/linux/raise.c:32: error: '__pthread_raise' undeclared (first use in this function) Aug 28 13:23:20 <[TEHb]> ../linuxthreads/sysdeps/unix/sysv/linux/raise.c:32: error: (Each undeclared identifier is reported only once Aug 28 13:26:05 is there any way of telling oe NOT to generate the modules-*.tgz files in the deploy directory? Aug 28 13:26:15 I don't use them, and its a waste of time :) Aug 28 13:26:40 sure, edit the recipe or include Aug 28 13:26:50 but generating them only takes a few secs Aug 28 13:46:16 morning Aug 28 13:48:24 hi koen, can you tell me which bitbake is recommended for .dev building (distro=familiar or distro=angstrom), version 1.4.2 or 1.6.0? Aug 28 13:49:36 familiar is broken in .dev Aug 28 13:49:42 no version of bitbake can fix that Aug 28 13:50:03 right, and what about angstrom? Aug 28 13:50:19 that should work, regardless of bitbake version Aug 28 13:50:26 ok, thanks Aug 28 13:50:26 I'd recommend the latest stable release Aug 28 13:51:06 which is 1.6.0? Aug 28 13:51:21 * koen dunnos Aug 28 13:51:26 I'm using trunk Aug 28 13:51:38 let's see.. I never actually manage to finish an angstrom build :) Aug 28 13:52:10 uhm... doesn't angstrom set TARGET_OS? bitbake complains about that Aug 28 13:54:23 it does over here Aug 28 13:57:14 ah, I just found my error, I was still using angstrom-2006.9 as DISTRO, but it should be angstrom-2007.1 Aug 28 14:03:00 would you be so kind to add h1940 to 'devices without a keyboard' in the GPE section of angstrom-2007.1.conf? Aug 28 14:26:12 how can i see why a certain package is going to be installed into the image? some -dbg packages end up there, and i wonder why. Aug 28 14:34:21 rwhitby: ping Aug 28 14:38:16 is there any other good tutorial on using bb other than the oe wiki ? Aug 28 14:38:34 i have it set up but dont want to bother people here really Aug 28 14:38:41 ibot message to rwhitby: I'm running OpenSlug 3.10-beta w/ kernel 2.6.16/PREEMPT. Do you have any idea why my average load is 1.0 in idle mode? Aug 28 14:39:19 vom: i'm afraid there isn't much. we have the oe wiki, the bitbake manual, and the OE documentation branch. everything else is forthcoming ;) Aug 28 14:39:30 ibot botmail to rwhitby: I'm running OpenSlug 3.10-beta w/ kernel 2.6.16/PREEMPT. Do you have any idea why my average load is 1.0 in idle mode? Aug 28 14:40:04 mickey|work: hey mickey. I was wondering if you had any thoughts on trying to improve dependancy handling in the python packages. It would be nice if each package knew which other pyhton modules it depends on. Aug 28 14:40:20 mickey|work: okay thanks, maybe ill try #openzaurus Aug 28 14:42:53 hvontres|poodle: i have been thinking about this. It's hard. To do it really good, you'd need to keep a database of providers and run the python scripts, because there are so many ways to import packages. We could do perhaps a less perfect solution by just parsing for 'import' and 'from foo import' which would be a first step. That leaves us with the problem of a database for the providers. Aug 28 14:43:58 mickey|work: nothing a bit of python and sqlite can't solve ;) Aug 28 14:44:45 mickey|work: ok , thanks... I just wanted to see if there were plans to improve the "install module, run prog, see what fails now.." method :) Aug 28 14:45:08 heh. of course it's not impossible, but it's a lot of work. one problem is that I manually grouped modules to form packages. i don't want to lose this handcrafted selections, because they really make sense. doing it one module per package (like perl does it) would be just insane. Aug 28 14:45:38 hvontres|poodle: it's on my Todo, but i'm afraid not very near the top ;) Aug 28 14:46:01 mickey|work: thats ok.... we all gotta eat :) Aug 28 14:46:34 *nod* I'm still looking for someone who needs python in a commercial project, so that i can get some paid time for python in OE Aug 28 14:47:13 being employed at university was much better... spare time wise ;) Aug 28 14:48:11 mickey|work: well, we do use Python here at work....but only as a scripting tool to test machines...running on Windblows. Aug 28 14:48:26 good luck with your freelancing :) Aug 28 14:48:36 thanks :) Aug 28 14:49:22 ~lart oesf server for being down agarin !!! Aug 28 14:49:22 * ibot breaks out the Hoover and sucks up oesf server for being down agarin !!! Aug 28 14:50:08 if all goes well i'm accepting just very few offers in Q1 and Q2 2007 to get writing the Embedded Linux + OE book done. I imagine having a good amount of time to work on OE during these days. Aug 28 14:51:01 lrg: http://xkcd.com/comics/sandwich.png Aug 28 14:51:18 hehehe Aug 28 14:51:43 heh Aug 28 14:53:55 ~praise internet radio Aug 28 14:54:01 All hail internet radio! Aug 28 14:54:08 i almost stopped hearing local radio stations because they are so crappy Aug 28 14:54:16 s/hearing/listening to/ Aug 28 14:54:37 can somebody explain me how do_split_packages works in combination with sorting out -dbg files? Aug 28 15:00:04 Chris Humbert: who are you and are you here? :D Aug 28 15:17:45 mikearthur: Chris Humbert isn't here Aug 28 15:19:11 is there a reason that do_split_packages doesn't package links? ("if not (stat.S_ISREG(mode) or (allow_dirs and stat.S_ISDIR(mode))): continue") Aug 28 15:20:47 hi all Aug 28 15:25:43 03tmbinc 07org.oe.dreambox * r74c1e0e6... 10/ (1 classes/package.bbclass): package.bbclass: also allow links to be packaged in do_split_packages Aug 28 15:25:49 03tmbinc 07org.oe.dreambox * r917569da... 10/ (3 files in 3 dirs): initscripts-opendreambox: work around busybox 'mount -a' bug Aug 28 15:26:35 hi mr_nice_slacker Aug 28 15:26:57 anyone here know much/anything about dual-channel memory? Aug 28 15:28:00 looks like dbus-glib is broken in Angstrom 2007.1 http://pastebin.ca/152119 Aug 28 15:28:34 do13: the tmiofb seems to be broken :( Aug 28 15:29:37 pH5: hi Aug 28 15:30:40 Hi all! Aug 28 15:30:53 Been away from OE waay too long. Aug 28 15:31:00 time to start work again :) Aug 28 15:31:17 lrg: pH5 is interesting in bugreports about dbus Aug 28 15:31:57 ehm Aug 28 15:32:00 interest Aug 28 15:32:01 that looks like an old OE tree Aug 28 15:32:45 i.mx32!?!? Aug 28 15:32:49 bastard! Aug 28 15:34:50 do13: shout when you're about Aug 28 15:35:12 lrg: dbus-glib-native compiles just fine here, nobody looking for dbus/dbusdbus/dbus-shared.h Aug 28 15:36:00 mx32 ? wah i'm out with my 21 :') Aug 28 15:36:15 s/32/31/ Aug 28 15:36:24 ho 31 , oki :) Aug 28 15:36:29 03koen 07org.oe.dev * rb27820d7... 10/ (1 packages/dbus/dbus-glib_0.71.bb): dbus-glib: depend on dbus to get dbus.h Aug 28 15:36:59 some document related to teamworking with OE ? Aug 28 15:37:03 koen: I suspect dbus-native_0.92 should be re-disabled, as it doesnt supply introspect data Aug 28 15:37:47 i'm thinkin' about that ( TW with OE ) Aug 28 15:37:57 XorA: it is needed for dbus-glib-native to generate it Aug 28 15:38:36 Genesis: teamworking in what sense? Aug 28 15:39:33 pH5: versions less than 0.92 stage the .xml files, 0.92 doesnt Aug 28 15:39:50 mickey|work: so do you still have to "Register" your radios over in germany? Aug 28 15:40:03 koen, : many people on a same OE tree Aug 28 15:40:14 hvontres|poodle: of course. they GEZ will come after you, if you don't :/ Aug 28 15:40:18 s/they/the/ Aug 28 15:40:29 Genesis: as long as they have a seperate TMPDIR Aug 28 15:40:32 pH5: was causing me confuson earlier Aug 28 15:40:35 XorA: because dbus-glib-native stages the xml file, but it needs the dbus-daemon in staging Aug 28 15:40:44 koen, : yeap but i've some issue with that Aug 28 15:40:46 to generate it Aug 28 15:41:02 Genesis: as far as the same metadata: we have monotone for that :) Aug 28 15:41:19 RP: Is there a "formal" tinderboxing schedule in place now? Or people just report whatever they compile? Aug 28 15:41:22 i want to share the cross tool , the work & staging etc Aug 28 15:41:28 in a clever way Aug 28 15:41:39 depend what task man want to do ... Aug 28 15:41:50 pH5: something was wrong in the order everything was running then, this was a clean gpe-image build btw Aug 28 15:42:06 Laibsch: people just report what they compile Aug 28 15:42:25 I see. I'll continue to do that. Aug 28 15:42:37 pH5: and you missed an install -d ${STAGING_DIR}/${BUILD_SYS}/share/dbus/ Aug 28 15:43:36 XorA: "install -d ${STAGING_DATADIR}/dbus" in dbus-glib-native_0.71.bb Aug 28 15:44:11 pH5: missng in dbus-native_0.92.bb Aug 28 15:44:27 oh, that's right. thanks! Aug 28 15:44:30 p Aug 28 15:45:45 * XorA needs to run for train Aug 28 15:46:06 look at this tomorrow and see if I was just being a spanner first time round Aug 28 16:03:13 a+ Aug 28 16:08:35 pH5, XorA|gone: thanks Aug 28 16:25:10 lrg: still liking the i.mx? Aug 28 16:26:44 Does EVERYONE get that tosa screen corruption thingy? It seems I hear less and less about it, but it's still there at least on my Z. Aug 28 16:28:10 I remember investigating about it. Buffering the screen in kdrive fixed it (although it makes it slower). I recall testing some direct writes to fb0 and seeing that it happened with unaligned writes IIRC Aug 28 16:29:24 koen: yep, it's fast ! Aug 28 16:29:34 I guess I get to work on it :) (am I the only one that gets it though?) Aug 28 16:29:53 http://bugs.openembedded.org/show_bug.cgi?id=490 Aug 28 16:30:09 koen: 530 bogomips Aug 28 16:32:24 cute Aug 28 16:32:36 marcan: everyone gets it. Aug 28 16:32:59 koen: we need an nbp with a i.MX31 :) Aug 28 16:33:03 marcan: tosa users are not much and they're not very vocal Aug 28 16:33:11 Good (sorta) then. I'll dig into it. Aug 28 16:33:16 excellent! Aug 28 16:33:16 yeah, that's what I thought. Aug 28 16:34:15 I suspect something as simple as ensuring writes are always 16 or 32 bits (and doing read-modify-write if appropriate) may fix it. Aug 28 16:34:25 ya, that may be it Aug 28 16:34:56 mickey|work, marcan : mikearthur is vocal at times on tosa, he should be back tomorrow if you want to ask him Aug 28 16:35:12 good :) Aug 28 16:35:45 oh BTW, awesome work on 2.6. I pretty much stopped messing with the 2.4 kernel due to its crappiness, but I should be able to work much better now. Aug 28 16:37:08 03tmbinc 07org.oe.dreambox * r28bef9f1... 10/ (1 conf/distro/opendreambox-1.4.conf): opendreambox-1.4: use swig 1.3.29 Aug 28 16:37:14 03pH5 07org.oe.dev * r1d92faee... 10/ (1 packages/dbus/dbus-native_0.92.bb): dbus-native-0.92: do_stage: create dir before putting files to it Aug 28 16:37:19 03mickeyl 07org.oe.dev * r842867b8... 10/ (4 files in 2 dirs): angstrom|familiar|openzaurus.inc: s/DISTO_FEATURES/DISTRO_FEATURES/ :D Aug 28 16:50:40 koen: you did accidentally remove the dbus-glib-native dependency from dbus-glib. Aug 28 16:51:19 I did? Aug 28 16:53:43 x29a_: ping Aug 28 16:54:16 re Aug 28 16:55:26 totally. b27820d75a034508419598a68bcbe1342bcf2ee1 was a s/dbus-glib-native/dbus/. shall I re-add it? Aug 28 16:56:11 please do Aug 28 16:56:18 * koen wonders how that happend Aug 28 16:56:22 hey hrw Aug 28 16:57:06 did I said already that Polish laws and rules suxx? Aug 28 16:57:15 you did Aug 28 16:57:34 uf Aug 28 16:59:05 zecke: hi Aug 28 16:59:09 03pH5 07org.oe.dev * reffb96b2... 10/ (1 packages/dbus/dbus-glib_0.71.bb): dbus-glib-0.71: depend on dbus-glib-native for binding header generation Aug 28 17:00:24 zecke: alive? Aug 28 17:00:36 hehe Aug 28 17:00:48 zecke: I need login/pass to OE survey Aug 28 17:01:58 right Aug 28 17:02:14 hrw: I couldn't find mine, otherwise I would have sent them a while ago Aug 28 17:02:47 koen: got it Aug 28 17:02:53 grrr Aug 28 17:02:59 ~lart php exif handling Aug 28 17:02:59 * ibot moos at php exif handling Aug 28 17:03:14 ~hail system("exiftool $foo |grep Lens"); Aug 28 17:03:16 * ibot bows down to system("exiftool $foo |grep Lens"); and chants, "I'M NOT WORTHY!!" Aug 28 17:06:15 koen: system not always is available.. Aug 28 17:06:25 I know Aug 28 17:06:34 and there are dozens of php exif libs Aug 28 17:06:49 but this is just a quick shellscript in php :) Aug 28 17:13:27 ~lart gnumeric for lack of ODS Aug 28 17:13:28 * ibot executes killall -KILL gnumeric for lack of ODS Aug 28 17:14:24 and they call it opendocument support... Aug 28 17:15:21 hrw: Do you know of any .bb files used to build standalone 2.6 kernel modules? I would like to try to get fbvncserver working again. Aug 28 17:15:33 hvontres|poodle: hostap-modules? Aug 28 17:15:43 orinoco-modules, rt2570 Aug 28 17:15:49 hrw: no, kbdsim.o Aug 28 17:15:57 hvontres|poodle: I told 3 examples Aug 28 17:16:06 hrw: ok..thanks Aug 28 17:20:00 koen: how to edit fieldvalue in gnumeric without usage of mouse? Aug 28 17:20:08 f2? Aug 28 17:21:53 koen: thx - weird usage Aug 28 17:22:42 gnumeric uses largely the same keybinding as excel Aug 28 17:22:52 I do not use them Aug 28 17:26:20 pH5: '-Wcast-align' might be a usefull addition to cflags Aug 28 17:28:30 koen: sounds like a good idea Aug 28 17:30:22 In your opinion, is it easy to contribute to OpenEmbedded?: 55% Yes, 45% No Aug 28 17:30:38 hm Aug 28 17:30:40 it is easy Aug 28 17:30:45 eh, its quite easy Aug 28 17:31:10 the "what's the best way to add patches back to OE" questions every know and then kinda indicated that Aug 28 17:31:14 I'm processing OE survey results Aug 28 17:33:15 The problem is that patches usually come under bugzilla and that they are forgotten if not continually pointed out Aug 28 17:33:51 they are not forgotten Aug 28 17:34:07 they are not on top of people's TODO Aug 28 17:34:16 03pH5 07org.oe.dev * r6b6561dd... 10/ (4 files in 2 dirs): libnss-mdns: remove avahi-daemon recommendation to avoid circular dependency Aug 28 17:34:17 look at teh bugday we had a while back Aug 28 17:35:29 commit the patch yourself (just kidding) Aug 28 17:39:42 koen: we had to have them each month... Aug 28 17:40:33 hrw: sure Aug 28 17:55:23 * [TEHb] successul build glibc-2.3.6 for ARM by gcc-4.1.0 Aug 28 17:58:00 [TEHb]: with OE? Aug 28 18:02:23 ???? ?? ??????? Aug 28 18:05:11 ~lart spreadsheets and their messing with formulas after sorting Aug 28 18:05:12 * ibot eats spreadsheets and their messing with formulas after sorting's liver with some fava beans and a nice chianti Aug 28 18:05:36 =B6/sum(b3:b11) changes to crap after sorting... Aug 28 18:05:38 mr_nice_slacker: pong Aug 28 18:06:02 hrw: heh Aug 28 18:06:07 all because it change sum(b3:b11) to sum(four rows before:11 rows after) Aug 28 18:06:28 hrw: I could tell you stories about the police department and screwing up their order by sorting excell wrong Aug 28 18:06:28 koen: ? Aug 28 18:06:42 koen: I suppose Aug 28 18:07:04 koen: that does not change sucking of spreadsheets Aug 28 18:07:15 yeah Aug 28 18:08:23 ~fsck gnumeric Aug 28 18:08:27 e2fsck /dev/gnumeric : warning! filesystem contains dickheads! Aug 28 18:08:41 does anyone have a *working* mirror that i can get uboot-0.0+cvs20060828-r2 from ? Aug 28 18:09:05 the sourceforge and oesources.org mirrors appear to be down Aug 28 18:13:37 bigd0g: oesources.org is likely to not be current Aug 28 18:13:50 bigd0g: and try another SF mirror :) Aug 28 18:14:14 zecke: for cvs fetcher? Aug 28 18:14:42 err, how do I set default branch and pull site? Aug 28 18:14:55 hehe... Aug 28 18:16:49 nevermind, big duh Aug 28 18:20:23 03kristoffer 07org.oe.dev * r3ffb499d... 10/ (1 packages/glib-2.0/glib-2.0_2.12.0.bb): packages/glib-2.0/glib-2.0_2.12.0: Fixes broken URL Aug 28 18:25:37 zecke: can i get a list of sf mirrors somewhere? Aug 28 18:25:47 zecke: for the cvs pull of uboot, that is Aug 28 18:25:52 * koen waits for the script to crunch through all the images Aug 28 18:26:02 bigd0g: sf.net only has cvs, no cvs mirrors Aug 28 18:28:02 koen: ok, so if it doesn't have what i need... where should i look for it? Aug 28 18:28:20 koen: bitbake tried oesources.org, but they also didn't have it. then it failed Aug 28 18:28:42 try the checkout by hand? Aug 28 18:28:55 but why are you trying to build an obsolete version of uboot? Aug 28 18:29:34 that's the version bitbake is trying to build when i run 'bitbake linux-omap1' Aug 28 18:29:41 is there a newer version i can point it to? Aug 28 18:29:52 probably have to put a preferred_version tag in my conf file? Aug 28 18:29:55 try the git version Crofton added some weeks ago Aug 28 18:30:35 i pulled the db after crofton made the omap update and it's still looking for the old file. Aug 28 18:30:36 hrm Aug 28 18:30:39 i'll play with it, then Aug 28 18:37:56 <[TEHb]> Guys, anyone use Intel Strata Flash NOR's ? Aug 28 19:15:40 bigd0g, I am back Aug 28 19:16:13 bigd0g, I think you want the kernel bb file that builds current omap git kernel Aug 28 19:17:00 koen, I think I need to commit an updated omap kernel, I'll check my diffs from dev Aug 28 19:18:05 03pH5 07org.oe.dev * rdf5459c3... 10/ (1 classes/rm_work.bbclass): rm_work.bbclass: stop cleaning out before populate_staging Aug 28 19:22:13 koen: do we plan to have SVN mirror of OE? Aug 28 19:22:46 zecke: survey results are interesting read Aug 28 19:22:59 survey results available? Aug 28 19:23:18 Crofton: not yet Aug 28 19:27:57 hrw: http://projects.linuxtogo.org/plugins/scmsvn/viewcvs.php/?root=openembedded Aug 28 19:28:18 hrw: mickeyl wanted to hook up it to ciabot, but he's too busy too look at it right now Aug 28 19:29:10 koen: too many people do not liked monotone <0.26 Aug 28 19:29:29 too many people are whining complainers Aug 28 19:31:04 koen: from my discussion with some people wanted to use OE I can point few things Aug 28 19:31:13 1. lack of STABLE branch Aug 28 19:31:15 2. lack of STABLE branch Aug 28 19:31:18 we had daily snapshots and diffs on ewi for a long time Aug 28 19:31:18 3. lack of docs Aug 28 19:31:44 which according to apache were massively unpopular Aug 28 19:31:59 'I built something month ago and today it is unbuildable' Aug 28 19:32:22 "it hurts when I poke in my eye" Aug 28 19:32:43 'WTF? build stopped on fscking anoncvs after one hour of build and I wasted whole night o cpu time' Aug 28 19:33:07 I also read the survey results Aug 28 19:33:34 I'm going to create kind of summary Aug 28 19:33:37 It strengthened my opinion we need to have a page on how to use google Aug 28 19:33:57 hrw: so you accept being session chair for session #1 ? Aug 28 19:34:04 not that errors should happen, but people can try searching for fixes themselves Aug 28 19:34:10 RP: is it ok for you to be session chair for bitbake? Aug 28 19:34:11 mickey|work: yes Aug 28 19:34:26 zecke: did you have a look at OEDEM schedule? please comment when you have a chance Aug 28 19:34:30 hrw: excellent Aug 28 19:34:40 ah, I'd like to join you at OEDEM. anything to consider in advance? Aug 28 19:34:43 at what time do we start? Aug 28 19:35:05 10, 12, 14 ? Aug 28 19:35:24 I still need to find place to stay Aug 28 19:35:39 koen: 10-11 would be good Aug 28 19:35:59 mickey|work: looked yes Aug 28 19:36:07 pH5: just have a look at the OEDEM wiki page and prepare yourself if you have some important points to add to the discussion Aug 28 19:36:09 mickey|work: nothing unrealizable Aug 28 19:36:12 that would give us time to have cafe und kuchen Aug 28 19:36:14 zecke: ok, cool Aug 28 19:36:21 mickey|work : Evening. Did you find sometime to update the keys ? i am trying to push something and i get "remote public key hash XXXXX unknow" errors Aug 28 19:36:48 Ifaistos: sorry, not yet, but within the next 4 hours for sure. I'll answer your mail when you're in. Aug 28 19:36:53 * mickey|work doing 5 things at once atm. Aug 28 19:37:06 I'd like the sessions to be roughly 1.30 Aug 28 19:37:11 with 30min break after each Aug 28 19:37:20 that would give us 8 work hours for SAT and SUN Aug 28 19:37:20 mickey|work : NP its not that critical :) Aug 28 19:37:22 no Aug 28 19:37:25 Fri and Sat Aug 28 19:37:38 mickey|work: marked itself as chairman Aug 28 19:37:41 Thursday and Sunday should be open Aug 28 19:37:42 hrw: thanks Aug 28 19:37:54 and have a IM/IRC/VOIP link to kergoth and other non-attending core members? Aug 28 19:38:08 hmmm Aug 28 19:38:14 we could use gobby for the notes Aug 28 19:38:15 koen: while being neat, I guess we're already at the limit of people to have a constructive discussion Aug 28 19:38:30 * koen packs a hammer Aug 28 19:38:31 zecke : I have setup 5 machines but i found out that i had no HD. expect the get them withing the week Aug 28 19:38:40 i did not agree :) with that Aug 28 19:38:42 "order, order" *thwap* Aug 28 19:39:19 koen: does gobby run under gpe already? Aug 28 19:39:32 hrw: yes Aug 28 19:39:37 Xora tested it Aug 28 19:39:41 how about setting a live stream and recording the whole thing for those that could not attend ? Aug 28 19:39:54 koen: good. then I will have a machine which be able to run it. Aug 28 19:40:07 mickey|work: did you read oz-devel? we got another poodle offer Aug 28 19:40:13 VLC can do that very easily. Aug 28 19:42:05 hrw: hehe, yeah i have seen that and i loughed Aug 28 19:42:12 years over years no chance to get a poodle Aug 28 19:42:15 now we have more than we can use Aug 28 19:42:17 *g* Aug 28 19:43:26 yep Aug 28 19:43:52 honestly, i have no idea who would be interested Aug 28 19:43:58 I think that if it is pxa255 poodle then we should think about getting it (all our ones are pxa250) Aug 28 19:44:02 that's right Aug 28 19:44:03 ! Aug 28 19:44:09 oedem? Aug 28 19:44:32 OpenEmbedded Developers European Meeting Aug 28 19:45:02 ah Aug 28 19:45:07 hrw: http://helmsdeep.homelinux.net/~pat/gobby/ Aug 28 19:45:23 should a call a bb file linux-omap1_2.6.17+git Aug 28 19:45:36 koen: hmm.. not org.openembedded.dev/packages/gobby/?? Aug 28 19:45:36 hrw: coredumps one is pxa255 Aug 28 19:45:36 or linux-omap1_2.6.x+git.bb Aug 28 19:45:49 hrw: I just noticed it's missing from there Aug 28 19:45:55 * koen builds and will push Aug 28 19:46:17 * kergoth wishes he could go to oedem.. stupid being on different continents Aug 28 19:46:34 This continent thing is overrated Aug 28 19:46:39 makes it hard to get around Aug 28 19:47:03 koen: o.. did not know about that pxa255 one Aug 28 19:47:27 kergoth: now, when you have r/w.... irssi description is fscked.. Aug 28 19:48:39 kergoth: you could try to get to FOSDEM Aug 28 19:48:47 kergoth: we could have a strategy meeting there as well Aug 28 19:48:53 well, it's EU as well though Aug 28 19:49:25 gurus, I need a quick answer on the kernel bb file name question :) Aug 28 19:50:29 Crofton: go ahead... Aug 28 19:51:02 linux-omap1_2.6.17+git verus _2.6.x+git Aug 28 19:51:20 for a bb file that gets source from the linux-omap git repo Aug 28 19:51:39 the openzaurus git kernel uses 2.6.17+git Aug 28 19:51:44 Crofton: or set PV inside of recipe Aug 28 19:51:48 but I think that will get dated in time Aug 28 19:52:10 right now, I just want a file name :) Aug 28 19:52:11 i'm always in favour of explicit namings Aug 28 19:52:18 but since mv is cheap, go with what you like Aug 28 19:52:31 version is tricky for kernel build :) Aug 28 19:52:34 ok Aug 28 19:52:45 I am going to use x+git Aug 28 19:53:42 It isn't quite perfect, but hopefully it will help bigd0g Aug 28 19:55:31 mickey|work: you guys should see about finding a conference phone, either land line or voip, and i could call in ;) Aug 28 19:55:41 "OE Developers appear to be too limited in time available to make use of donated hardware." Aug 28 19:56:13 TRUE TRUE TRUE. Aug 28 19:56:22 atleast in my situation Aug 28 19:56:36 well if hardware == fully functional clone Aug 28 19:59:29 * mickey|work adds his idea of time schedule to sessions to oedem page Aug 28 19:59:35 ping if you think it's too tough Aug 28 19:59:45 but if we don't start early we won't get anything done :) Aug 28 20:00:22 is there going to be bw for a live stream ? Aug 28 20:00:29 Ifaistos: no Aug 28 20:00:32 mickey|work: could you login to the site before editing? Aug 28 20:00:46 mickey|work: that way you aren't credited as 'anonymous' :) Aug 28 20:00:49 oops - perhaps i could :) Aug 28 20:00:55 mickey|work: you should Aug 28 20:01:33 ok, done Aug 28 20:02:25 mickey|work: and if your account is still 'mickeyl' then change it to 'name surname' - looks better Aug 28 20:02:27 zecke: do you know which equipment we can use in the rooms? e.g. do we have beamer or whiteboard? Aug 28 20:02:40 hrw: i think koen added myself as Michael Lauer Aug 28 20:03:11 i wonder why it still shows the [login] when i logged in Aug 28 20:03:15 it should hide it then Aug 28 20:03:32 it should in theory, but it's a static link Aug 28 20:03:40 mickey|work: the location depends on if we have office space rented for october Aug 28 20:03:50 one might have noticed the website has some rough edges Aug 28 20:04:07 mickey|work: either it will take place at the BMCO where I have to check permissions (e.g. "Wachschutz") Aug 28 20:04:15 mickey|work: but I'm sure we will net+beamer+whiteboard Aug 28 20:04:36 zecke: ok, great. btw., will you be at home in the 1st or 2nd week of september? it looks like i'm coming to berlin for 2 or 3 days Aug 28 20:05:09 mickey|work: one of the two weeks I will be there Aug 28 20:06:23 ok, it will probably be rather 2nd week Aug 28 20:06:30 i'm having a meeting with Harald. Aug 28 20:06:39 well, _will have_ Aug 28 20:07:11 would be cool, if we could spend an evening together Aug 28 20:08:23 sure Aug 28 20:08:50 mickey|work: you could drop by at the BMCO Forum/Office as well Aug 28 20:09:53 zecke: yeah i'd like to, e..g meet stevie in person Aug 28 20:10:02 hehe, why? Aug 28 20:10:20 just because. i like to have a face in mind when I'm talking to or about persons Aug 28 20:10:31 hehe Aug 28 20:10:54 besides... business terms, maybe there are opportunities to work together :) i need to start thinking like a business guy :) Aug 28 20:11:14 at least a bit Aug 28 20:11:15 heh Aug 28 20:12:10 mickey|work: yes, you could sell your ass :) Aug 28 20:12:24 right. there's a lot to sell there Aug 28 20:12:26 * mickey|work chuckles Aug 28 20:12:33 why sell your donkey? Aug 28 20:12:48 to get flithy rich Aug 28 20:12:54 Ifaistos: you forget disks? Aug 28 20:13:01 cu all Aug 28 20:13:14 time to sleep Aug 28 20:14:23 03crofton 07org.oe.dev * rc6503f65... 10/ (3 files in 3 dirs): Aug 28 20:14:23 linux-omap1_2.6.x+git, omap5912osk: Add bb file to build kernel from omap git Aug 28 20:14:23 * Build kernel from omap git repository, use most recent version Aug 28 20:14:23 * Change omap5912osk machine to use this kernel Aug 28 20:14:23 * The depmod command is failing at the moment Aug 28 20:14:46 zecke : Though i had a number of 40Gig ones from a "retired" raid but it seems my colegues found other uses for them... i.e car mp3 players Aug 28 20:16:34 zecke : so i called a pc reclying store and got 10 pcs 6Gig for 5Euro a pc :) Aug 28 20:17:54 zecke : i think they should be more than enough for a swap and a minimal root fs Aug 28 20:20:28 Ifaistos: you got mail Aug 28 20:20:34 * mickey|work needs to sleep now Aug 28 20:20:37 g'night Aug 28 20:20:42 03clarson 07org.oe.dev * r66118fca... 10/ (7 files in 6 dirs): Aug 28 20:20:42 Merge from poky: Aug 28 20:20:42 Rework the way patches are handled. There are now two abstract base classes, Aug 28 20:20:42 initialized in patch.bbclass. One for patchset operations on a directory, and Aug 28 20:20:42 another for patch failure resolution. Currently includes 'patch' and 'quilt' Aug 28 20:20:43 concrete PatchSet classes, and a 'user' resolver class, which simply drops you Aug 28 20:20:47 into a shell in the source tree to fix the rejects. Aug 28 20:21:59 let me know if there are any issues with that. i know the api of the resolver class sucks, need to fix that Aug 28 20:26:02 kergoth: cool :) Aug 28 20:26:05 looks cute Aug 28 20:26:28 03ifaistos 07org.oe.dev * r5972bffd... 10/ (1 conf/machine/dht-walnut.conf): Aug 28 20:26:28 Machine config for the dht-walnut (ppc) board Aug 28 20:26:28 Machine config for the dht-walnut ppc based board Aug 28 20:26:28 Machine config file for the ppc based dht-walnut board Aug 28 20:26:36 03ifaistos 07org.oe.dev * r14f068bc... 10/ (1 conf/machine/rb500.conf): Machine config file for the mips based RB5XX family of boards also knows as Routerboards Aug 28 20:27:42 * Ifaistos hopes he did not brake anything Aug 28 20:28:52 Ifaistos: Adding machines should be safe enough :) Aug 28 20:29:49 RP: I though so, I am worried with monotne. i am still trying to de-svn... Aug 28 20:29:57 heh, i suck at mtn still too Aug 28 20:30:03 i'm quite impressed with its speed though Aug 28 20:30:13 mtn st is faster than i expected, as are commits and such Aug 28 20:30:39 mtn is a definite improvement on monotone :) Aug 28 20:32:05 0.30 will be at least 40% faster on initial pull as 0.29 Aug 28 20:32:13 so even less reason to complain Aug 28 20:33:39 now to close #1327 and #1328 i should mention the mtn id also ? Aug 28 20:33:51 that would be helfpull Aug 28 20:34:01 and mention the bugnumbers in the commit message as well Aug 28 20:34:57 i always loved the way mtn uses the hash instead of arbitrary numbering Aug 28 20:35:00 just makes sense Aug 28 20:35:20 You'd like git then, its even worse for it :) Aug 28 20:35:47 mtn is faster than bitbake :) Aug 28 20:36:07 * RP ignores Crofton Aug 28 20:37:17 hehe Aug 28 20:37:29 bitbake is continually getting faster. The C parser is probably our next speed gain (hopefully) Aug 28 20:37:38 isn't mallum rewriting bitbake in iwmmxt asm? Aug 28 20:38:01 hahah Aug 28 20:38:08 * Crofton knows, he was waiting on the parsing at the time :) Aug 28 20:39:38 RP: mtn will use the git asm core if all goes well Aug 28 20:41:00 koen: That's interesting to know and doesn't surprise me. git was written for speed :) Aug 28 20:41:14 by kernel developers who know who Linux works :) Aug 28 20:41:26 * koen has blind faith in nico Aug 28 20:41:32 (no pun intended) Aug 28 20:42:43 :-) Aug 28 20:44:08 03Patrick 07org.oe.dev * rbba107bc... 10/ (7 files in 2 dirs): gobby: add gobby + supportlibs Aug 28 20:44:45 kergoth: new toyy for you to play with: gobby Aug 28 20:49:34 03koen 07org.oe.dev * re61f227f... 10/ (3 files in 3 dirs): libxml++: move to correct subdir Aug 28 20:50:47 bigd0g, bootstrap-image should build now Aug 28 20:50:51 is usbnet working on tosa / OZ3542? I load g_ether and ifconfig up the interface, but it complains it is "not ready" and the PC doesn't detect any USB device. Aug 28 20:51:42 03crofton 07org.oe.dev * r924dc0c3... 10/ (1 conf/machine/omap5912osk.conf): Aug 28 20:51:42 omap5912.conf: Update required module list to match kernel config from Aug 28 20:51:42 omap5912osk_defconfig in kernel source. Aug 28 21:04:05 03ifaistos 07org.oe.dev * rfd010f8f... 10/ (1 packages/libesmtp/libesmtp_1.0.4.bb): Aug 28 21:04:05 New package for libesmtp 1.0.4 Aug 28 21:04:05 Closes bug #1346 Aug 28 21:10:48 I having problems getting the depmod command to work in my git kernel bb file Aug 28 21:11:17 the OZ kernel does a KERNEL_RELEASE-2.6.17, but it doesn't seem to work for me Aug 28 21:11:20 Crofton: still? Aug 28 21:11:30 DEPMOD="2.6"? Aug 28 21:11:53 yeah Aug 28 21:11:59 I though I had fixed it ... Aug 28 21:12:28 in kernel.class, it uses the include/linux/version.h file to get the kernel version Aug 28 21:15:48 the depmod depends is based off $PV Aug 28 21:15:56 bitbake world is funny with bitbake-mt Aug 28 21:16:11 then the cde that runs the depmd is based off KERNEL_MAJOR_VERSION Aug 28 21:16:41 whcih comes from the version.h file Aug 28 21:19:31 03ifaistos 07org.oe.dev * rbea143d9... 10/ (3 files in 3 dirs): esmtp 0.5.1 patch for uclibc Aug 28 21:19:40 What is the difference between = and ?= Aug 28 21:21:03 ?= only sets when unset Aug 28 21:21:13 "weak assignment" Aug 28 21:21:37 hmmm Aug 28 21:22:26 I think KERNEL_VERSION in kernel.bbclass should be ?= then Aug 28 21:22:45 so we can override from outside in cases where version.h does not exist ? Aug 28 21:29:02 night all ! Aug 28 21:43:30 03koen 07org.oe.dev * rfc1c689a... 10/ (1 packages/scummvm/scummvm_0.9.0.bb): Aug 28 21:43:30 scummvm 0.9.0: enable all engines to get rid of undefined symbols* Aug 28 21:43:30 * and be able to play more games Aug 28 21:45:41 hmm, OE/bitbake is building makedevs for the target, but is failing to run the host version of makedevs. Do I need to install makedevs into the host? Aug 28 21:46:42 fakeroot is being built ok, I would have thought that makedevs would be there as well in the staging dir Aug 28 21:51:18 can someone point me to the syntax for the 'patch' line in a .bb file ? Aug 28 21:51:20 I think all the smart guys are gone now Aug 28 21:51:31 im trying to pass patch -p Aug 28 21:51:34 vom, look in other bb files Aug 28 21:51:38 Crofton: heh Aug 28 21:51:49 yeah i was getting ready to Aug 28 21:51:53 vom: pnum= Aug 28 21:52:01 i have this one patch thats like 4 levels deep Aug 28 21:52:08 * T0mW admits that he is one of the dummies Aug 28 21:52:08 koen_: thanks Aug 28 21:52:16 file://leet.patch;patch-1;pnum=0 Aug 28 21:52:21 s:-:=: Aug 28 21:52:32 03koen 07org.oe.dev * r14917049... 10/ (1 packages/scummvm/scummvm.inc): scummvm.inc: remove Troll technology Aug 28 21:55:39 * T0mW takes makedevs binary from buildroot an puts in into bitbake/bin dir Aug 28 21:55:51 heh Aug 28 21:56:11 I am going back through my OE work and removing hacks like that :) Aug 28 21:56:15 heh, it works fine for buildroot: makedevs == makedevs, right. Aug 28 21:57:11 tomorrow I need to resolve my arm-linux-dpemod headache Aug 28 21:57:30 works fine if you have a version.h in your kernel tree, but I am running from git Aug 28 21:57:32 hmmm Aug 28 21:57:41 Where does version.h come from? Aug 28 21:57:51 nope, "invalid option -- r", Aug 28 22:11:50 bah! "genext2fs: command not found" Aug 28 22:16:43 serves you right Aug 28 22:31:27 vom: The SRC_URI stuff is in the user-manual: http://www.openembedded.org/user-manual&dpage=ch07s09#id2527963 Aug 28 22:33:40 T0mW: fakeroot-native should get built, and fakeroot should then end up in staging (tmp/staging/x86_64-linux/bin/makedevs for me). Aug 28 22:39:53 03tmbinc 07org.oe.dreambox * r4fb48716... 10/ (3 files in 3 dirs): enigma: update rotor_fix.diff Aug 28 22:39:58 03tmbinc 07org.oe.dreambox * r6cfc1d86... 10/ (4 files in 3 dirs): enigma: remove old versions Aug 28 22:40:04 03tmbinc 07org.oe.dreambox * rd4333ebe... 10/ (1 packages/dreambox/dreambox-dvb-modules.bb): dreambox-dvb-module: update dm600pvr drivers to 20060829 Aug 28 22:50:31 Can anyone help me a bit with debugging a problem with the initscript .bb? I get this error when trying to build: http://pastebin.com/778365 Aug 28 22:50:47 This is the do_install script: http://pastebin.com/778364 Aug 28 22:54:47 (it build on debian etch i386 IIRC, but not on ubuntu x86_64 for some reason I can't quite figure out) Aug 28 23:01:23 * NAiL notes psyco won't run on x86_64 Aug 28 23:03:13 NAiL: Dunno. Maybe trying running run.do_install with tracing on (sh -x) to see what it's actually doing? Aug 28 23:03:21 v8jlene: I've been using buildroot for a number of years. This OE is a puzzle as to why the makedevs source is there, it gets built for the target, but the native i686 binary doesn't. I need to delve deeper into the whole OE build system to see how / why that would be the case. Aug 28 23:04:09 v8jlene: you would think that there would be a dependancy someplace for the i686 version. Aug 28 23:06:49 T0mW: There is, classes/image_ipk.bbclass adds a dependency on makedevs-native and it's this class which calls makedevs to generate filesystem images. Aug 28 23:07:14 ok, will start looking from there. Aug 28 23:09:01 T0mW: Not sure what you were trying to do, but when asking bitbake to build an image (which will use image_ipk) it should then build makedevs-native - the "native" versions are the ones for your host. Aug 28 23:14:46 v8jlene: for some reason, this test breaks it: if [ "${DISTRO}" == "openzaurus" ]; then Aug 28 23:14:47 it doesn't like the "==" there... Aug 28 23:17:23 NAiL: Hmm, isn't == a bash builtin which isn't support by the test command? So it probably should just be a single = for test. Aug 28 23:23:43 yeah, it worked when I replaced it with "=" Aug 28 23:25:09 the other problems were caused by some leftover links Aug 28 23:25:14 thanks for the help ;) Aug 29 01:55:39 does anyone have a local.conf file for thier project that they can give me? I seem to be having the devils own time of getting OE to give me any results build a downstream build error. Aug 29 01:55:53 I'd like to get it build "something" Aug 29 01:59:05 this is typical of what I run into (the current problem): http://rafb.net/paste/results/dry8y950.html Aug 29 02:00:49 this is my local.conf: http://rafb.net/paste/results/gkyup224.html Aug 29 02:01:33 I assume, that if I run "bitbake essential-image" that OE would make a tarball filesystem image? **** ENDING LOGGING AT Tue Aug 29 02:59:56 2006