**** BEGIN LOGGING AT Thu Oct 19 02:59:57 2006 Oct 19 03:29:42 i'm confused about certain aspects of bringing a new hardware device into the OE fold... Oct 19 03:30:34 as a start, what "distro" should i use? I like openslug, but the distro definition is pretty much welded to the slug itself. Oct 19 03:30:45 any tips? Oct 19 04:28:35 hey rwhitby, got a sec? Oct 19 04:29:06 Sure Oct 19 04:29:20 * rwhitby has just landed in Chicago Oct 19 04:29:38 welcome to the states! hey, late for you, you sure? Oct 19 05:29:44 HopsNBarley: back Oct 19 05:30:26 oh, hey, mighty polite to ring back. Oct 19 05:31:24 i'm just working on putting a new machine type into the oe fold. Oct 19 05:31:45 its a freescale 8241 soc - the iomega storcenter Oct 19 05:32:12 so far so good. i should just ring back with detailed questions.... don't really have them now! Oct 19 05:45:36 HopsNBarley: I'm not the best resource to ask for a new machine addition. Oct 19 05:46:05 cool. i love openslug, btw. Oct 19 05:46:07 koen or mickey|zzZZzz or zecke or RP are far more experienced. I didn't even add the nslu2 machine (kergoth did that for us) Oct 19 05:46:47 have you seen the iom's? not bad. gigE, 64Mb, twin ide Oct 19 06:12:50 mickey|zzZZzz: i've submitted qtopia-core_4.2.0.bb Oct 19 06:28:52 darn, i woke up way too early today Oct 19 06:28:57 morning folks Oct 19 06:30:45 mickeyl: Morning :) . Oct 19 06:30:54 <_law_|iBook> morning all Oct 19 06:31:12 _law_|iBook: Morning. Oct 19 06:35:37 morning _law_ Oct 19 06:37:56 good morning all Oct 19 06:40:28 morning Oct 19 06:41:10 HopsNBarley|zzz: adding machine can be quite easy now Oct 19 06:41:44 HopsNBarley|zzz: set TARGET_ARCH, MACHINE_FEATURES and add kernel (if needed) Oct 19 06:41:57 HopsNBarley|zzz: then run build and look does it works Oct 19 06:41:59 morning koen, yo hrw Oct 19 06:42:01 koen hrw: Morning. Oct 19 06:42:30 yesterday lecture was quite good - about 20 persons but only 2 questions. Oct 19 06:42:46 they probably did not catched what OE is ;D Oct 19 06:43:03 heh Oct 19 06:43:53 the good part is what my friend from work told after: "I did not realise that this your project is so BIG. I was thinking that this is something just for palmtops" Oct 19 06:44:14 that's excellent. looks like something we need to stress even more everywhere we can Oct 19 06:44:44 * mickeyl thinking about OE myths again Oct 19 06:45:06 my recent builds are for omap and mainstone machines - nice experience it is.. building for !palmtops Oct 19 06:46:42 1. OpenEmbedded is for palmtops Oct 19 06:46:42 2. Target device needs to run X11 to be able to use Python Oct 19 06:46:47 what more we had? Oct 19 06:46:52 ah.memory Oct 19 06:47:05 4. OE gives no SDK Oct 19 06:47:57 koen: ugh....something is *really* still wrong with packaged-staging...glibc still tells me I have too old kernel headers installed... Oct 19 06:48:06 koen: what do you have in ASSUME_PROVIDED? Oct 19 06:48:14 http://www.openembedded.org/oemyths Oct 19 06:48:15 JustinP: nothing Oct 19 06:48:21 ... Oct 19 06:49:31 if I don't ASSUME_PROVIDED += "quilt-native stagemanager-native" Oct 19 06:49:41 it won't build anything Oct 19 06:49:54 JustinP: quilt-native depend on itself? Oct 19 06:50:24 * JustinP tries again to see what it says Oct 19 06:50:56 bitbake trunk says there's a circular dependency...I'm using bitbake 1.6 right now Oct 19 06:52:47 yep Oct 19 06:53:02 quilt-native depends on stagemanager-native and vice versa Oct 19 06:53:42 assuming stagemanager-native *should* fix it I think.... Oct 19 06:54:30 cu Oct 19 06:57:04 * JustinP attempts to debug.....I just have no idea how you could be using this with how it is... Oct 19 06:57:21 koen: have you tried building from scratch with packaged-staging? Oct 19 06:58:07 All the time Oct 19 06:58:24 * JustinP wonders how our environments differ Oct 19 06:58:51 dev definately has a circular dependency with packaged-staging inherited... Oct 19 07:01:54 koen: do you bitbake anything before adding INHERIT += "packaged-staging"? Oct 19 07:02:02 I'm trying to figure out what I'm missing here... Oct 19 07:02:21 I don't Oct 19 07:03:32 what machine and distro....? Oct 19 07:04:32 I'm using a vanilla local.conf with machine=spitz and distro=openzaurus-unstable, branch .dev Oct 19 07:04:38 machine is irrelevant Oct 19 07:04:49 I figured...it's just... Oct 19 07:04:58 DISTRO=angstrom-2007.1 Oct 19 07:05:12 if I try to bitbake anything it fails with a circular dependency Oct 19 07:05:54 03koen 07org.oe.dev * r550348b1... 10/ (1 packages/stage-manager/stagemanager-native_0.0.1.bb): stagemanager: remove dep on quilt-native by zapping PATCHTOOL Oct 19 07:13:23 well, it's started this time at least Oct 19 07:13:36 with angstrom and with your commit (thanks) Oct 19 07:15:02 re Oct 19 07:32:05 hrw|work: i started to add some text to the myths to clarify them. Oct 19 07:32:18 morning Oct 19 07:32:25 hey XorA Oct 19 07:33:02 hey XorA Oct 19 07:33:23 mickeyl: so now pygtk is running we gonna see you coding loads of gpe apps :-) Oct 19 07:34:53 XorA: hehe. i'm afraid not. you will see me working on gtk+ based apps though Oct 19 07:35:57 mickeyl: s/_foo_/'''foo'''/g Oct 19 07:37:38 mickeyl: thx Oct 19 07:39:08 mickeyl: yay Oct 19 07:39:19 koen: well, it's building glibc-intermediate now....seems to work Oct 19 07:39:31 koen: looks like OZ;s glibs has an issue with packaged-staging then Oct 19 07:41:01 JustinP: probably, RP and I didn't test with OZ Oct 19 07:47:13 /media/PR0N Oct 19 07:47:19 ~lart gnome Oct 19 07:47:20 * ibot does a little 'renice 20 -u gnome' Oct 19 07:48:05 shouldn't this be gone in bitbake trunk: Oct 19 07:48:09 OTE: exceptions.SyntaxError:invalid syntax (line 1) while evaluating: Oct 19 07:48:09 ${@base_read_file('/Volumes/Koen's Firewire disk/OpenEmbedded/build/tmp/angstrom/staging/h2200-angstrom-linux-gnueabi/kernel/kernel-abiversion')} Oct 19 07:48:11 ? Oct 19 07:48:23 aah Oct 19 07:48:30 the ' in Koen's Oct 19 07:48:46 :) Oct 19 07:48:53 that's on OSX btw Oct 19 07:48:53 we probably need to escape the string before handing it over Oct 19 07:49:15 for now just substitute the ' with " Oct 19 07:49:23 (both of 'em) Oct 19 07:50:02 classes/module-base.bbclass:export KERNEL_VERSION = "${@base_read_file('${STAGING_KERNEL_DIR}/kernel-abiversion')}" Oct 19 07:50:59 mickeyl: right no we can't replace them :( Oct 19 07:51:02 try Oct 19 07:51:12 hmm Oct 19 07:51:13 right Oct 19 07:51:17 hello / bonjour Oct 19 07:51:27 Genesis: bonjour Oct 19 07:51:29 koen: add it to the bugtracker, i'll think about escaping it Oct 19 07:51:50 thinking about it Oct 19 07:51:52 that's a bitbake bug Oct 19 07:52:07 * XorA boots the dreaded smdk2440 again Oct 19 07:52:09 adding this globally would probably make bitbake slower :/ Oct 19 07:54:10 mickeyl: http://bugs.openembedded.org/show_bug.cgi?id=1511 Oct 19 07:54:44 k Oct 19 07:55:41 hmm zaurus warez on amazon http://www.amazon.co.uk/FigLabs-English-conversion-Sharp-Zaurus/dp/B000CZ0RYC/sr=8-5/qid=1161245669/ref=sr_1_5/026-4983970-6164452?ie=UTF8&s=electronics Oct 19 07:56:21 i got my host cable from them Oct 19 07:57:00 ah, figlabs Oct 19 07:57:04 mickeyl's buddies Oct 19 07:57:13 ade|desk: thats ok, but I am suspecting that sale includes some copyright Sharp stuff :-) Oct 19 07:57:31 hehe Oct 19 07:58:00 mickeyl: some history? Oct 19 07:58:18 XorA: just that I had to fight with 'em to make them honor the GPL Oct 19 07:58:51 are they selling the software or a card that happens to have some random bits in some clear pattern that could be seen as written software ? Oct 19 07:58:53 hrw|work, what does the "glibc-intermediate" mean ? Oct 19 07:58:56 mickeyl: so known pirates Oct 19 07:59:14 XorA: yarrrr! Oct 19 07:59:19 leoncamel: thats one of steps needed to bootstrap toolchain Oct 19 07:59:32 leoncamel: koen will be able to tell more Oct 19 07:59:51 gcc needs glibc build, but glibc needs gcc to build Oct 19 08:00:22 koen, so what ? Oct 19 08:00:57 if you already have the software licence is it wrong to get the translated version ? Oct 19 08:01:15 koen, bitbake will build glibc-intermediate first ? then gcc ? and then glibc ? Oct 19 08:01:19 gcc-cross-initial build glibc-intermediate to build another step to build glibc to build gcc-cross - kind of Oct 19 08:01:53 ade|desk: they add pdf viewer - which means qpdf2 but does they give you source offer for it? Oct 19 08:03:54 gcc-cross-i -> glibc-i -> gcc-cross -> glibc Oct 19 08:04:02 hrw|work: ah two issues, one is copyright and the other is GPL violations, that is clearly a violation Oct 19 08:04:04 ade|desk: some of the "English" translations involved patched binaries in the past Oct 19 08:04:26 is it any different to trisoft etc ? Oct 19 08:04:50 ade|desk: do they charge 50 GBP for translation? Oct 19 08:05:01 no idea Oct 19 08:05:24 they charge for importing and modifying Oct 19 08:06:54 they don't supply the source for there nand backup images .. most companies i think are in the same boat Oct 19 08:09:04 ade|desk: it hurts my morals to charge for warez :-) Oct 19 08:11:01 koen: http://pastebin.ca/209636 any idea? Oct 19 08:12:02 koen: nm...it's earlier than that Oct 19 08:12:06 JustinP: that doesn't seem to list the actual error Oct 19 08:12:49 koen: tested rxvt-unicode on a bfs image and it still shows no pink, pink must be from legacy something we both had Oct 19 08:13:03 http://pastebin.ca/209638 Oct 19 08:13:25 XorA: I think the pink scrollbar on my hx is a nice extra :) Oct 19 08:13:56 JustinP: that looks like a bad linux-linux-headers Oct 19 08:13:58 koen: appeals to a different sector of gadget geeks :-) Oct 19 08:14:03 damn it! Oct 19 08:14:05 same as before, then Oct 19 08:14:10 koen: Im sure my wife would use a pink PDA Oct 19 08:14:55 koen: /sda1/oe/packaged-staging.angstrom.build/tmp/staging/arm-angstrom-linux-gnueabi/include/linux/errno.h Oct 19 08:15:01 koen: seems to exist fine.... Oct 19 08:16:12 koen: it's only including . and # Oct 19 08:16:12 koen: it's only including . and /sda1/oe/packaged-staging.angstr Oct 19 08:16:12 # Oct 19 08:16:15 om.build/tmp/work/armv5te-angstrom-linux-gnueabi/gcc-cross-4.1.1-r7/gcc-4.1.1/libiberty/../include Oct 19 08:17:12 something having to do with packaged-staging is causing the -I to not be set with the normal staging incdir.... Oct 19 08:19:06 does anyone know what package libstdc++6 comes from? Oct 19 08:21:40 can someone fix linuxtogo? Oct 19 08:23:14 XorA: sure, what needs fixing? Oct 19 08:23:29 koen: network is running like through damp string Oct 19 08:23:42 koen: both from wolfson and my ADSL line Oct 19 08:24:03 the carrier pigeons are tired Oct 19 08:24:31 did you forget to feed them ? Oct 19 08:27:30 XorA: I have a recipe for gnutls 1.4.4 almost ready Oct 19 08:28:35 koen: cool Oct 19 08:28:38 koen: many changes? Oct 19 08:28:42 do_configure() { Oct 19 08:28:42 oe_runconf Oct 19 08:28:42 find ${S} -name Makefile | xargs sed -i s:'-L/usr/lib':'-L${STAGING_LIBDIR}':g Oct 19 08:28:59 XorA: that's the only change I need Oct 19 08:29:16 koen: thats bad form, fixing configure.ac would be a better option Oct 19 08:29:25 right Oct 19 08:29:32 that's why I said 'almost' Oct 19 08:29:40 * JustinP ......... Oct 19 08:29:42 koen: ah ok Oct 19 08:29:48 yeah, I know, I know.... Oct 19 08:30:54 koen: does that gnutls want autoconf 2.60? Oct 19 08:31:08 koen: or is that 1.5 series Oct 19 08:31:13 XorA: it seems that way Oct 19 08:36:22 yeah....the include path is no longer being set right Oct 19 08:36:44 only change is INHERIT += "packaged-staging" Oct 19 08:36:57 there has got to be something I'm missing Oct 19 08:37:28 koen: can you pastebin your local.conf? I'd like to compare Oct 19 08:37:46 DISTRO=angstrom-2007.1 Oct 19 08:37:59 I know Oct 19 08:38:02 INHERIT += "packaged-staging" Oct 19 08:38:10 that's all I have in my local.conf :) Oct 19 08:38:16 hmmmmm.... Oct 19 08:38:20 koen: it has wrong syntax Oct 19 08:38:37 DISTRO=angstrom-2007.1 -> DISTRO = "angstrom-2007.1" should did you Oct 19 08:39:02 I've tried both angstrom and oz and both die trying to find normal include files...stdlib.h, etc..... Oct 19 08:40:16 koen: wait...no machine? do you have another conf which is being used? Oct 19 08:40:51 yeah, just those 2 lines of course don't work... Oct 19 08:40:53 JustinP: machine in auto.conf Oct 19 08:41:07 ok Oct 19 08:41:15 JustinP: site.conf for DL_DIR etc stuff, local.conf for distro/inherit, auto.conf for machine Oct 19 08:41:58 gah Oct 19 08:42:07 hrw|work: so what's in your site.conf? Oct 19 08:42:49 this isn't going to end up different... Oct 19 08:43:50 ok, it's building again.... Oct 19 08:44:01 added DL_DIR, BBFILES, and BBINCLUDELOGS Oct 19 08:45:44 I can tell you're all riveted by what I'm saying Oct 19 08:48:22 JustinP: dl_dir, cvs_dir, git_dir, bbfiles, bbincludelogs, debian_mirror, gnome_mirror, gnu_mirror, parallel_make, multimachine stuff Oct 19 08:51:41 aren't those mirrors standardized somewhere? Oct 19 08:52:12 JustinP: they are but I want local ones Oct 19 08:52:17 and shouldn't the git/cvs dirs and multimachine stuff be handled automatically by the conf files in OE? Oct 19 08:52:46 I'm rebuilding from scratch yet again but I expect it to fail in the same way.... Oct 19 08:53:01 perhaps you're all also setting a magic ENV var which I'm not? Oct 19 08:53:12 that's all I can think of which could possibly be different Oct 19 08:53:15 perhaps we should have a location where people post their current confs as a reference for people to see Oct 19 08:53:50 JustinP: I do not want to have tmp/downloads/ but prefer to keep all in ~/devel/sources/{dl_dir,cvs_dir,git_dir}/ Oct 19 08:54:05 JustinP: and multimachine in my site.conf is old thing Oct 19 08:54:37 ok Oct 19 08:54:37 ade|desk: I wrote wiki page about it months ago Oct 19 08:55:04 hrw|work: do you have some CFLAGS or such ENV var which is including your host include dirs for you? Oct 19 08:55:44 hrw|work: of course, file, gettext, bison, flex, binutils (-cross versions) all build fine for me....it's just gcc-cross or glibc failing Oct 19 08:56:02 and only when building with packaged-staging, normal builds still work fine... Oct 19 08:56:09 JustinP: no Oct 19 08:56:10 hrw|work: where ? Oct 19 08:56:17 * JustinP attempts to compare run scripts Oct 19 08:57:40 ade|desk: good question. Oct 19 08:58:22 hrw|work: we have local.conf example at http://www.openembedded.org/wiki/LocalConfExamples but that is only one and not 'confs' purely local only Oct 19 08:58:50 http://www.openembedded.org/firstproject is great Oct 19 08:59:05 XorA: http://rafb.net/paste/results/b6h4Eq29.html and http://rafb.net/paste/results/duiwEM98.html Oct 19 08:59:42 hrw|work: you mean http://www.openembedded.org/wiki/FirstProject ? Oct 19 08:59:52 koen: yes Oct 19 09:00:09 hrw|work: There is currently no text in this page, please help to fill it in. Oct 19 09:00:18 useful Oct 19 09:00:25 export includedir="/sda1/oe/packaged-staging.dev.build/tmp/cross/include" Oct 19 09:00:28 doesn't exist... Oct 19 09:01:37 ok, I'm way too tired for this Oct 19 09:03:26 koen: then why http://www.openembedded.org/firstproject == http://www.openembedded.org/wiki/FirstProject at all? Oct 19 09:04:00 hrw|work: autopath symlinking Oct 19 09:04:29 who's works on tslib ? Oct 19 09:04:56 Genesis: kergoth? Oct 19 09:05:02 thx Oct 19 09:08:15 good morning Oct 19 09:15:08 hi florian_kc Oct 19 09:19:51 hey, flatronf701C Oct 19 09:19:56 hey, florian_kc :) Oct 19 09:21:22 XorA: does your s3c24 board launches getty on the s3c ttySAC nowadays? did you have anything more to do than add the udev rule (which you did, thanks) Oct 19 09:21:46 mickeyl: just whats in the smdk2440.conf file Oct 19 09:21:57 thanks, i'll take a look Oct 19 09:22:02 mickeyl: what board you working on? Oct 19 09:22:04 i'm on a s3c2410 devel board here Oct 19 09:22:16 to be exact Oct 19 09:22:18 armzone qt2410 Oct 19 09:22:35 mickeyl: cool, the s3c stuff is a bit dire in mainstream kernel at moment Oct 19 09:22:55 XorA: we will have some updates coming in the near future.... Oct 19 09:23:08 morning all Oct 19 09:23:12 morning liam Oct 19 09:23:54 hey lrg Oct 19 09:25:51 mickeyl: The secret project still in-time? Like you and Harald came back to OpenEZX. ;) Oct 19 09:25:56 03xora 07org.oe.dev * r60663f63... 10/ (1 packages/ntp/ntp_4.2.2p3.bb): ntp_4.2.2p3.bb : source for old versions is moved into a subdir Oct 19 09:26:12 stefan_schmidt: I'm afraid we now managed to be 8 weeks over our schedule :/ Oct 19 09:26:41 stefan_schmidt: it looks like it'll keep us busy at least until the january, more likely march Oct 19 09:26:48 mickeyl: 8 weeks, isn't this normal for software projects? ;) Oct 19 09:27:03 which is really a pity considering openezx. I want to get this phone running asap, but alas I can't divide myself Oct 19 09:27:29 stefan_schmidt: ya. it could be worse. hardware manufacturing is giving us a headache as well Oct 19 09:28:00 XorA: close bug for ntp Oct 19 09:28:01 still, what we are doing will benefit openEZX in the mid run, so I'm not feeling so bad at all Oct 19 09:28:08 hrw|work: number? Oct 19 09:28:17 XorA: dont remember Oct 19 09:28:49 stefan_schmidt: I'm having a Joker in the backhand Oct 19 09:28:55 mickeyl: ok, so i keep amusing the others with small efforts in the mean time. Oct 19 09:29:13 stefan_schmidt: no kidding, the most important thing is keeing the fire burning! Oct 19 09:29:26 keeping, even Oct 19 09:29:44 * ade|desk is intrigued by mickeyl secret Oct 19 09:30:10 3 more weeks and then we can talk Oct 19 09:30:15 * den-ros too Oct 19 09:30:15 ade|desk: its obviously a handheld with 64 arm cores that can run bitbake at 1000 packages a second Oct 19 09:30:16 which will be a relief, really Oct 19 09:30:26 will do, but it's sometimes frustrating to see so less progress. Oct 19 09:30:41 I seem to be unsbale to build a few of the xorg-driver packages Oct 19 09:30:43 stefan_schmidt: you know i thought about a Sprint Oct 19 09:30:58 stefan_schmidt: locking three or four of us together for one week somewhere in germany Oct 19 09:31:02 specifically xf86-video-sis Oct 19 09:31:09 stefan_schmidt: i can hardly imagine how productive this would be ! Oct 19 09:31:44 this is what I get when I try... Oct 19 09:31:44 mickeyl: Right. Need to think about my study schedule for this. ;) Oct 19 09:31:46 http://rafb.net/paste/results/cWZdRp54.html Oct 19 09:32:10 seems to be looking in the hosts /usr/share/X11/sgml/ Oct 19 09:35:03 XorA: i'd guess something like netfiltering arm based python gadget with super OEified powers Oct 19 09:42:23 hi mickeyl Oct 19 09:43:18 * ade|desk jumps in amazement ... hark 'tis phil Oct 19 09:44:02 long time no see Oct 19 09:44:26 hey pb_ :) Oct 19 09:44:45 hi ade|desk Oct 19 09:45:23 pb_: what have you been upto lately? Oct 19 09:56:41 hrw|work: can you play 8k audio in oz3.5.4.1 on spitz? Oct 19 09:58:07 XorA: I can check it in the evening Oct 19 09:58:23 what would be the most sensible way to stop it looking for defs.ent in /usr/share/X11/sgml Oct 19 09:58:39 hrw|work: ah if you dont have it handy no problems, we might have fixed our problem by then Oct 19 09:58:53 tkp: check configure params - maybe it have --no-docs Oct 19 09:58:59 is there a trick to recompile all file depend on a file that we update ? Oct 19 09:59:01 as far as I can tell, defs.ent is used for producing documentation... which we don't need anyway Oct 19 09:59:15 hrw|work: there are no configure flags Oct 19 09:59:32 tkp: then patch it to not look there Oct 19 10:00:19 this is the offending code: Oct 19 10:00:20 http://rafb.net/paste/results/jSMOJ529.html Oct 19 10:00:45 I don't really understand the ac_cv_file_$prefix bit Oct 19 10:01:15 tkp: patch it to have $as_ac_File=no forced Oct 19 10:01:36 hrw|work: right Oct 19 10:01:47 morning all Oct 19 10:02:54 hey RP Oct 19 10:03:30 hrw|work: what do you think about nailing oz-unstable gcc to be 3.4.4 Oct 19 10:04:16 XorA: feel free to do it Oct 19 10:04:21 XorA: and binutils 2.16 Oct 19 10:05:56 hrw|work: binutils was already there Oct 19 10:12:05 03xora 07org.oe.dev * r91bf3046... 10/ (1 conf/distro/openzaurus-unstable.conf): openzaurus-unstable.conf : fix gcc version to 3.4.4 to keep us vaguely compatilible with oz release versions. Oct 19 10:13:50 thx Oct 19 10:14:15 thx Oct 19 10:20:10 hi RP Oct 19 10:20:52 XorA: That sounds like a good idea to me. It should make oz in .dev more stable Oct 19 10:23:01 RP: more importantly it stops newbies mixing gcc 4.1.1 with glibc 2.3.5 and getting wierdness Oct 19 10:24:35 hm.. I have to stop using angstrom as one of my build distros. Oct 19 10:24:53 it change too often and this machine spend too much time in building toolchains ;) Oct 19 10:25:16 RP: you seen any probs playing 8k audio on spitz? Oct 19 10:29:24 03mickeyl 07org.oe.dev * r7d9d43a2... 10/ (3 files in 2 dirs): xserver-kdrive * disable w100 by default, enable it for arm Oct 19 10:32:50 03koen 07org.oe.dev * r00032e71... 10/ (1 classes/package_ipk.bbclass classes/rootfs_ipk.bbclass): *ipk.bbclass: banish ipkg-link Oct 19 10:35:13 XorA: I think it did work in the past but I haven't tested for a while Oct 19 10:37:53 I hate my builddirs here. each one is obsolete after week Oct 19 10:45:07 03likewise 07org.oe.dev * r8e601c26... 10/ (1 conf/bitbake.conf): bitbake.conf: add support for squashfs-lzma images - close #1507 Oct 19 10:46:38 03koen 07org.oe.dev * re8682916... 10/ (3 files in 2 dirs): Oct 19 10:46:38 kdrive: remove duplicate packages. Oct 19 10:46:38 * nothing changed in the actual output, so no PR bump needed Oct 19 10:49:25 hmm.. need to bump all bluez-utils now... Oct 19 10:55:13 03crofton 07org.oe.dev * r0d91e749... 10/ (8 files in 2 dirs): Add files for Virginia Tech's OSSIE project. (http://ossie.mprg.org) Oct 19 10:55:17 03crofton 07org.oe.dev * r4d2506a7... 10/ (1 packages/omniorb/omniorb-native_4.0.7.bb): Oct 19 10:55:17 omniorb-native_4.0.7.bb : Ugly hack removes omniorb libraries from build Oct 19 10:55:17 system library directory. ossiecf needs this so Oct 19 10:55:17 libtool does not try to link against them. We only Oct 19 10:55:18 build this package so we can run omniidl. Oct 19 10:55:25 03hrw 07org.oe.dev * r179df98f... 10/ (12 files in 2 dirs): bluez-utils: enable hid2hci util - close #907 Oct 19 10:58:32 in which package are the X start session scripts usually? i.e. the one calling mb-panel, mb-desktop, etc. Oct 19 10:58:46 x11-common? Oct 19 10:59:10 xserver-common in OE Oct 19 10:59:40 ah, they're not directly in OE but used from GPE upstream? Oct 19 11:00:00 03hrw 07org.oe.oz354x * rb595a835... 10/ (5 files in 3 dirs): gnuz: remove from repository - close #933 Oct 19 11:00:04 03hrw 07org.oe.dev * rba8fff0d... 10/ (5 files in 3 dirs): gnuz: removed from repository - close #933 Oct 19 11:00:55 XorA: #1058 - can you? Oct 19 11:00:58 mickeyl: right, gpe split them of for the people that don't use gpe Oct 19 11:01:07 ok, thanks Oct 19 11:01:19 i need a dedicated one without any login stuff or so Oct 19 11:01:44 hrw|work: works in .dev Oct 19 11:01:45 mickeyl: xserver-common should be able to provide that Oct 19 11:03:08 Crofton: channeldemo_svn.bb - a bit too generic for a package name? name it (the package) ossie-channeldemo? Oct 19 11:04:18 XorA: I need to push kernel update Oct 19 11:04:34 hmm Oct 19 11:04:50 that is why I have been putting them in one dirr Oct 19 11:05:03 but I suppose that does not solve name space issues Oct 19 11:05:25 hrw|work: I dont think it will be kernel dependant, I suspect the original poster was using massive cpu load Oct 19 11:05:37 hrw|work: mp3 with floats for example Oct 19 11:05:56 maybe Oct 19 11:06:12 XorA: anyway I have kernel update for 3.5.4.1 built ~month ago Oct 19 11:06:56 hrw|work: Id close as not a bug then see if anyone reopens it Oct 19 11:07:36 ok Oct 19 11:07:51 Crofton, yes, that solved .bb namespace issues, but yes, package namespace is differrent from .bb namespace. Oct 19 11:08:20 I was wondering about that already Oct 19 11:08:28 I'll go ahead and change the names Oct 19 11:08:36 03pfalcon 07org.oe.dev * r2461ced4... 10/ (1 conf/distro/angstrom-2007.1.conf): Oct 19 11:08:36 angstrom-2007.1: Update handhelds-pxa-2.6 to 2.6.16-hh7 Oct 19 11:08:36 * Changes since hh5 (hh6 is void) include: working sound for h4000, Oct 19 11:08:36 asic3_mmc driver changes fixing possible loss of data during suspend/resume. Oct 19 11:08:36 * Approved by Koen Kooi. Oct 19 11:08:37 Crofton, just override PN and all should be ok. Oct 19 11:08:44 especially since a bunch more will appear Oct 19 11:08:44 heh Oct 19 11:08:55 I have to override that in most cases already :) Oct 19 11:09:04 we suck at getting the name "right" Oct 19 11:09:45 since we have 4K+ packages now, it seems like some thought shold go into the namespace issue Oct 19 11:11:00 Crofton, I may agree. for a example, I'm not sure that perl package are named right with "-perl" as the suffix. I'd prefer that to be prefix, so sorted package list showed that in naturally groupped. Oct 19 11:11:33 Did this come up at the OE meeting? Oct 19 11:11:58 not but you can post RFC to list Oct 19 11:12:25 I'll see what I can come up with Oct 19 11:12:42 not that I am in a position to do anything about it :) Oct 19 11:13:01 and as noted, there aren't enough bb hackers to go around Oct 19 11:19:46 03koen 07org.oe.dev * ra6f1d179... 10/ (1 packages/bluez/bluez-libs_3.7.bb): bluez-libs: add 3.7 Oct 19 11:20:02 Crofton: can you fix #90? Oct 19 11:20:15 omnievents? Oct 19 11:20:32 yep Oct 19 11:20:42 can I delete the bb file for now? Oct 19 11:21:04 Crofton: thats your recipe Oct 19 11:21:08 :) Oct 19 11:21:10 yeah Oct 19 11:21:24 I am avoiding using the CORBA event service atm Oct 19 11:22:29 maybe I will look and see if I can see an easy way to fix frst Oct 19 11:22:34 now that I am smarter Oct 19 11:23:29 hmm, but maybe I should delete since I have no way to actually test it right now Oct 19 11:25:29 koen: #297 is solved or not? Oct 19 11:25:43 * koen takes a look Oct 19 11:26:57 yep. closed Oct 19 11:33:03 gracias Oct 19 11:33:33 olla Oct 19 11:33:43 ola Oct 19 11:33:43 hi likewise Oct 19 11:33:52 hi hrw Oct 19 11:34:04 likewise: buen dia Oct 19 11:34:17 florian_kc, koen, mickeyl: #347 (New version of bl (for GPE) which supports 2.6.x kernels as well as 2.4.xx iPAQ (and other machines too hopefully) Oct 19 11:34:19 hrw|work, #90 closed as wontfix and recipe gone Oct 19 11:34:48 hrw|work: dovhy dzjenja (or something sounding like that) Oct 19 11:34:58 http://www.dailytech.com/article.aspx?newsid=4605 Oct 19 11:35:06 likewise: you mean 'do widzenia'? Oct 19 11:35:33 03crofton 07org.oe.dev * r25cf9221... 10/ (5 files in 4 dirs): Oct 19 11:35:33 omnievents : Package does not build properly. Deleted since I do not need Oct 19 11:35:33 it at the moment. I will re-add a working recipe if required Oct 19 11:35:33 in the future. This clears bug #90. Oct 19 11:35:37 hrw|work: yes :-) My girlfriend tried to spell it to me Oct 19 11:36:01 likewise: its 'goodbye' Oct 19 11:36:20 hrw|work: zendobre ?? Oct 19 11:36:21 likewise: you rather want 'dzn dodobry' Oct 19 11:36:29 much better Oct 19 11:36:29 haha Oct 19 11:36:37 arh Oct 19 11:36:40 dzien dobry Oct 19 11:36:49 ~lart this machine for lack of utf8 Oct 19 11:36:49 * ibot drops a truckload of VAXen on this machine for lack of utf8 Oct 19 11:37:14 ąąąąą Oct 19 11:37:28 hrw|work: and now she claims she can count to ten as well, and order icecream and cigarettes, but I am not going to repeat *that* :-) Oct 19 11:37:34 ~lart hrw|work for breaking irssi Oct 19 11:37:34 * ibot overclocks hrw|work until hrw|work burns out for breaking irssi Oct 19 11:38:11 likewise: trust a girl to know how to order useless stuff like icecream and not beer Oct 19 11:38:16 XørÅ Oct 19 11:38:56 hrw|work: let me see... Oct 19 11:38:57 XorA: pivo Oct 19 11:39:10 03koen 07org.oe.dev * r1fcb37f7... 10/ (1 packages/gpe-session-scripts/gpe-session-scripts_0.67.bb): gpe-session-scripts: 'matchbox' doesn't exist, remove it so bitbake doesn't choke Oct 19 11:42:28 hrw|work: okay, noone added this so far?! Oct 19 11:42:36 I think I need to make two new targets Oct 19 11:42:37 florian_kc: it is in .oz354x Oct 19 11:42:56 one to extand boostrap-image to include the base files for my software radio Oct 19 11:43:08 and the otehr to build the extra compontent it may use Oct 19 11:43:20 any suggestions on how to do this? Oct 19 11:43:26 Crofton|home: use DISTRO_EXTRA_RDEPENDS Oct 19 11:43:34 yeah Oct 19 11:43:38 Crofton|home: or create crofton-image.bb Oct 19 11:43:50 but I think there are some imge/task/meta things Oct 19 11:44:05 so I make a new image Oct 19 11:44:12 that extands task-base ... Oct 19 11:44:25 sort of like boostrap-image Oct 19 11:45:01 then I can make a meta target that build this image and some extra packages? Oct 19 11:45:07 Crofton|home: DISTRO_EXTRA_RDEPENDS extend task-base. crofton-image can DEPEND on task-crofton-radio.bb and install task-base and task-crofton-radio Oct 19 11:45:42 so the image could DEPEND on packages it does not install Oct 19 11:45:49 so I can install a subset later Oct 19 11:46:14 OK Oct 19 11:46:20 this is enought to start :) Oct 19 11:46:25 hrw|work: you wrote programs in Amiga E? (with GoldED?) Oct 19 11:47:36 likewise: yes Oct 19 11:47:50 hrw|work: Then we have a common background :-) Oct 19 11:47:54 likewise: ;) Oct 19 11:48:15 hrw|work: and thanks for the MultiView file requester, was that yours? Oct 19 11:48:19 likewise: today when I look at those sources I have laugh Oct 19 11:48:21 likewise: yes Oct 19 11:48:52 likewise: you registered it? Oct 19 11:48:55 hrw|work: This is a late "thank you". Oct 19 11:49:04 ;) Oct 19 11:49:40 likewise: http://aminet.net/search.php?query=2b_mv - source available still Oct 19 11:49:45 gah, should have been more hardcore and used Rexx Oct 19 11:50:04 hrw|work: I dunno, I registered the most common programs I used (E, GoldED, ...) Dunno about Multiview Oct 19 11:50:14 XorA: ARexx you mean? Oct 19 11:50:35 hrw|work: well that was the amiga version Oct 19 11:50:42 likewise: I never registered GED - used 4.x free ver Oct 19 11:52:23 mickeyl: the qtopia-core configures fine here, what machine/distro are you using? Oct 19 11:52:37 mickeyl: alternatively, you could try without the tslib dependency Oct 19 11:53:30 03koen 07org.oe.dev * r4f4c1c97... 10/ (1 packages/matchbox-panel/matchbox-panel.inc): matchbox-panel.inc: libx11 -> virtual/libx11 Oct 19 11:53:42 * koen finally tracked down the stray libx11 dep Oct 19 11:53:45 ~lart autoconf 2.60 Oct 19 11:53:46 * ibot frags autoconf 2.60 with his BFG9000 Oct 19 11:53:55 giel: see bugtracker comment. you added a # which broke the configure Oct 19 11:54:48 it's ok now. i commited the first version Oct 19 11:54:55 XorA: I have a scary recipe here: http://www.openembedded.org/~koen/OE/gnutls_1.4.4.bb Oct 19 11:55:05 XorA: note the do_install_append.... Oct 19 11:55:27 03giel 07org.oe.dev * r497af8d9... 10/ (4 files in 3 dirs): add qtopia-core 4.2.0 (first shot) Oct 19 11:56:12 mickeyl: any volunteers for opie maintenance in OE? Oct 19 11:56:15 koen: WTRF? Oct 19 11:56:37 XorA: exactly! Oct 19 11:56:39 koen: korovkin said something about it, but we have yet to see any real action Oct 19 11:56:45 koen: everything get arm-linux* added Oct 19 11:56:57 mickeyl: ah, ic, yeah i added that comment last-minute :S Oct 19 11:57:00 mickeyl: thanks Oct 19 11:57:03 XorA: indeed Oct 19 11:57:27 koen: I propose to move everything opie related to obsolete if no one shows up until 2007 Oct 19 11:57:34 one day I'll try to build opie-image for angstrom Oct 19 11:57:44 mickeyl: right Oct 19 11:57:49 mickeyl: send that kind of into to ml? Oct 19 11:58:14 I am making a simple task, will RDEPENDS get installed on the image, and DEPENDS just built into ipks that may be installed later Oct 19 11:58:23 hrw|work: will do that in november. i don't want to put even more pressure on the poor guys Oct 19 11:58:28 sure Oct 19 11:58:40 Crofton|home: IPKG_INSTALL get installed Oct 19 11:58:56 Crofton|home: you mostly can do IPKG_INSTALL = "${RDEPENDS}" :D Oct 19 11:59:02 ok Oct 19 11:59:19 how do I use DEPNDS and RDEPENDS (properly) Oct 19 12:00:10 how to create a private branch ? is there some useful documentation about it ? Oct 19 12:03:32 Crofton|home: task-crofton-radio.bb RDEPENDS on your soft. crofton-radio-image.bb RDEPENDS on task-crofton-radio subpackages - proper way iirc Oct 19 12:03:54 I think I have it Oct 19 12:04:06 Genesis: read monotone doc. iirc commit into own branch create it Oct 19 12:04:21 thx Oct 19 12:04:25 I RDEPENDS on everything I want built and add the base subset to ipk_install Oct 19 12:04:31 I will see what happends Oct 19 12:08:04 hrw|work, : what i need it to modify for example the dbus recipe to add/remove a option and bitbake know to cook the recipe of my own branch if it exists, else , cook recipes from the other branch Oct 19 12:08:26 http://www.openembedded.org/user-manual&dpage=ch04s03 do that ? Oct 19 12:10:33 Genesis: bitbake collections or use own branch Oct 19 12:12:54 hrw|work, : i need an example , i don't think monotone documentation helps me to config bitbake for private branch Oct 19 12:13:03 perharps someone know a project Oct 19 12:13:19 Genesis: BBFILES = "path-to-my-branch/packages/*/*.bb" Oct 19 12:14:25 food time Oct 19 12:15:44 good launch Oct 19 12:17:35 am I right in thinking that if I set OVERRIDES = "stv" in distro.conf, things appended with _stv should be honored? Oct 19 12:19:49 tkp: yes, but it would ignore all other overrides, since you overwrote them Oct 19 12:22:39 psokolovsky_: ERROR: log data follows (/data/build/koen/OE/build/tmp/angstrom/work/h2200-angstrom-linux-gnueabi/handhelds-pxa-2.6-2.6.16-hh7-r1/temp/log.do_configure.30772) Oct 19 12:22:39 | FATAL: -hh version mismatch Oct 19 12:24:17 hmmm flash 9 for linux seems to work ok Oct 19 12:24:23 for 30 seconds at any rate Oct 19 12:30:20 * den-ros Do not like flash. Oct 19 12:30:37 * Genesis 's customer wants flash Oct 19 12:31:04 pornotube only works in flash, hence its importance Oct 19 12:31:17 :) Oct 19 12:31:25 :o Oct 19 12:31:55 are you planing some package for familiar to support that? :p Oct 19 12:33:20 koen: fscken cvs. checking the tag. Oct 19 12:33:48 Marex, : help me to package gnash Oct 19 12:34:13 http://bugs.openembedded.org/show_bug.cgi?id=1480 :-) Oct 19 12:34:36 * Marex 'd like to be even able to bitbake his own palmtt or palmz71 rootfs :p Oct 19 12:35:34 Marex, drop us palmtt.conf & palmz71.conf and at least people can start looking at it ;-) Oct 19 12:35:54 psokolovsky_, first I need to put kernel sources somewhere Oct 19 12:35:58 and make kernel .bb file Oct 19 12:36:01 or I dunno Oct 19 12:36:02 :) Oct 19 12:36:18 psokolovsky_, anyway ... I'm working on merging palmtt into omap mainstream :) Oct 19 12:36:23 Marex, you don't use hackndev svn? Oct 19 12:36:28 ok Oct 19 12:36:29 nah Oct 19 12:37:35 IRQ_TYPE_EDGE_RISING Oct 19 12:37:39 hmm ... that was removed :( Oct 19 12:37:55 Marex: genirq was merged Oct 19 12:38:19 /etc is ${sysconfdir} ? Oct 19 12:39:21 koen, need to find out how to replace it Oct 19 12:39:54 koen: try now (handheld-pxa-2.6 -c rebuild) Oct 19 12:41:27 Crofton: yes Oct 19 12:42:47 thanks Oct 19 12:46:46 psokolovsky_: remove DLDIR/cvs helped Oct 19 12:47:21 psokolovsky_: but I don't think that's an option for people that have already tried to fetch it Oct 19 12:48:10 koen, I hope noone did yet ;-) Oct 19 12:49:48 koen@h2200:~$ uptime Oct 19 12:49:49 13:10:46 up 67 days, 21:07, 1 user, load average: 0.01, 0.03, 0.00 Oct 19 12:53:10 koen, woo-hoo, unbelievable! what kernel version that's? Oct 19 12:53:38 koen@h2200:~$ cat /etc/openzaurus-version ; uname -a Oct 19 12:53:38 OpenZaurus 3.5.4.1 Oct 19 12:53:38 Linux h2200 2.6.16-hh5 #1 PREEMPT Sat Aug 12 15:58:42 CEST 2006 armv5tel unknown Oct 19 12:54:00 koen, also, is uptime counted as total of "on" time or time passed since boot? I bet the second though... Oct 19 12:54:15 'on' last time I checked Oct 19 12:54:30 if you suspend it, uptime won't increase Oct 19 12:54:35 koen, does h2200 has visible PM problems? For example, when suspended, how much time it can survive? Oct 19 12:54:44 cool Oct 19 13:13:27 neat Oct 19 13:13:37 plash works without troubles on the h2200 Oct 19 13:15:25 koen, so, 2.6.16-hh7 booted ok? pls let me know, I'd like to be able start committing common buttonmap refactor, towards 2.6.16-hh8 Oct 19 13:15:44 psokolovsky_: It's in GPE now :) Oct 19 13:15:54 koen, Cool, thanks! Oct 19 13:20:09 can someone look at #59 and my recent comments? Oct 19 13:24:44 finally lrg has had the pleasure of "static" :-D Oct 19 13:27:07 * lrg calls it board noise Oct 19 13:27:33 trust a softie to blame the hardware :-) Oct 19 13:35:58 ~pastebin Oct 19 13:35:59 rumour has it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste Oct 19 13:36:43 Any thoughts on what this error is Oct 19 13:36:49 http://pastebin.ca/210036 Oct 19 13:37:03 psokolovsky_: how does psplash look on your h4000? Oct 19 13:38:32 koen, didn't yet try latest image on real device ;-) Oct 19 13:42:55 psokolovsky_, I cant calibrate ts in opie again :( Oct 19 13:43:02 it worked on 2616 Oct 19 13:43:07 and doesnt work on 2618 :( Oct 19 13:43:26 Marex, kernel? Oct 19 13:43:34 it reacts but wont let me through Oct 19 13:43:37 psokolovsky_, yup Oct 19 13:43:54 psokolovsky_: http://www.openembedded.org/~koen/psplash-h2200.avi Oct 19 13:44:30 Marex, have no idea. check all usual QTE envvars. doubt it's kernel issue per se. Oct 19 13:45:36 koen, Looks nice. gotta try h4000 this evening Oct 19 13:45:40 psokolovsky_, dunno what could have changed :( Oct 19 13:46:03 psokolovsky_, I guess it is kernel issue ... there were many changes in the driver Oct 19 13:46:11 so maybe opie doesnt like it :p Oct 19 13:46:27 gpe on 16megs of ram ... 8-Q.... Oct 19 13:46:27 :p Oct 19 13:47:03 Marex, could be. though evdev interface should be stable. anyway, hh.org on 2.6.17 still Oct 19 13:47:24 2617 was screwed too iirc Oct 19 13:47:33 maybe opie doesnt like my 320x320 screen Oct 19 13:48:18 I use h6300 rootfs Oct 19 13:48:18 Marex: opie? try tslib-tests first Oct 19 13:48:37 Marex: maybe your machine does not have ts working properly with newer kernel Oct 19 13:51:06 Marex: and add your machine config into OE so you will not have to hack rootfs from other machines Oct 19 13:51:28 * Marex is quite unfamiliar with OE Oct 19 13:52:12 Marex: like most of my lecturers yesterday were Oct 19 13:52:13 mickeyl, I have a python question ..... Oct 19 13:52:34 hrw|work: not lecturees? Oct 19 13:52:50 Crofton: go ahead Oct 19 13:53:00 http://pastebin.ca/210036 Oct 19 13:53:17 It looks like a configure script is trying to work out the python path Oct 19 13:53:21 XorA: lecturers as 'people which came to listen me' Oct 19 13:53:44 ah.. I was lecturer ;D Oct 19 13:53:48 ~lart my english Oct 19 13:53:48 * ibot runs at my english with an origami Swiss Army knife, and inflicts a nasty paper cut Oct 19 13:53:53 but it would appear the python script used by configure fails Oct 19 13:54:04 better than my Polish Oct 19 13:54:24 ah yes Oct 19 13:54:49 export BUILD_SYS := "${BUILD_SYS}" Oct 19 13:54:53 export HOST_SYS := "${HOST_SYS}" Oct 19 13:54:56 or ARCH? Oct 19 13:54:58 dunno Oct 19 13:54:58 this is pretty twisted, calling native python to work out target python path Oct 19 13:55:08 mickeyl, you are the expert :) Oct 19 13:55:11 heh Oct 19 13:55:22 take a look at the python disutils.bbclass Oct 19 13:55:34 where should I try this? Oct 19 13:55:39 this replace hack is the only way to convince distutils to build extension modules properly Oct 19 13:55:46 its much cleaner than everything i had before :) Oct 19 13:55:58 should I inherit from that? Oct 19 13:56:09 dunno, what are you building? Oct 19 13:56:12 usrp Oct 19 13:56:16 mickeyl: can you look at #59 in so called free time? Oct 19 13:56:22 /etc/init.d/rc: /etc/init.d/rc: 164: divide by zero Oct 19 13:56:25 * koen wonders Oct 19 13:56:27 it used to work :) Oct 19 13:57:18 Crofton: try to get those two variables into your env. Oct 19 13:57:20 that should fix it Oct 19 13:57:25 either export in the .bb Oct 19 13:57:32 ok Oct 19 13:57:34 or submit during configure and/or makefile Oct 19 13:57:44 ok Oct 19 13:57:58 note the bb file in the repo right now is not the one I am owrking on Oct 19 13:58:14 hrw|work: I'll try to remember Oct 19 13:58:37 hrw|work: Sorry that stupid question - i type 59 in bugtrack on OE site and got something ununderstanded (i am so stupid :) ) . Please, say - where to look? Oct 19 13:59:42 hrw|work, how to use it now? :) Oct 19 13:59:46 that ts-tests Oct 19 14:00:18 I dont have any way how to input anything and I dont want to reboot 20times Oct 19 14:00:19 Marex: run ts_print_raw and start touching Oct 19 14:00:27 only that Oct 19 14:00:30 no params? Oct 19 14:00:34 Marex: yep Oct 19 14:00:44 Marex: if it works then ts_calibrate Oct 19 14:01:05 den-ros: http://bugs.openembedded.org/show_bug.cgi?id=59 is what I got Oct 19 14:01:36 hrw|work: Thanks. I only start learn OE bugtrack system. Oct 19 14:02:18 mickeyl, now the script is complaining it cannot find Python.h Oct 19 14:02:48 is python built? Oct 19 14:02:57 den-ros: its normal bugzilla Oct 19 14:03:06 hi CoreDump|home Oct 19 14:03:11 yes Oct 19 14:03:20 hmm where does it look for it? Oct 19 14:03:24 hi Oct 19 14:03:28 python-native or python .... Oct 19 14:03:36 you need both Oct 19 14:03:38 checking for Python include path... /home/balister/oe_home/tmp/staging/include/python2.4 Oct 19 14:03:39 hrw|work: Yes. This is same link. All work. It is again my english language problems. Oct 19 14:03:40 yeah Oct 19 14:03:41 ok Oct 19 14:03:50 the path looks bad though Oct 19 14:06:48 heh Oct 19 14:07:33 bibake -c rebuild wasn't building dependencies .... Oct 19 14:17:41 argh Oct 19 14:17:52 it is building libx11 Oct 19 14:17:55 dang ... doesnt work Oct 19 14:18:02 tsprintraw Oct 19 14:18:09 it work even open ts device Oct 19 14:18:12 I forgot to ask, how do I build python without X Oct 19 14:18:15 hrw|work, ^ Oct 19 14:18:41 Crofton|home: you cant Oct 19 14:18:48 um Oct 19 14:18:55 Marex: then you know why you had problems with opie Oct 19 14:19:01 can I install python on the target without X? Oct 19 14:19:05 hrw|work, isnt there any param for tsprintraw to open certain device? Oct 19 14:19:08 Crofton|home: yes, you can Oct 19 14:19:13 wew Oct 19 14:19:18 Marex: export TS_DEVICE Oct 19 14:19:22 hrw|work, nah ... I cant just calibrate it, but it reacts on taps ... just wont let me through Oct 19 14:19:23 iirc Oct 19 14:22:07 Crofton|home: http://www.openembedded.org/wiki/OEMyths Oct 19 14:23:20 xterm seems to depend upon xpm... which doesn't exist? Oct 19 14:23:42 tkp, libxpm Oct 19 14:23:48 nicolasfr|Zzzz: the slowness is because dbus is broken for non-root users :( Oct 19 14:24:14 hrw|work, hmm ... now it screams about unability to open tslib config file Oct 19 14:24:23 and then ts_conf: illegal seek Oct 19 14:24:54 Marex: ah ok.. bb file says 'xpm' Oct 19 14:25:17 I'm just guessing ;) Oct 19 14:25:22 tkp, Oct 19 14:25:37 Marex: you are rught... Oct 19 14:25:42 the bb file is wrong Oct 19 14:29:50 hrw|work, if I do od -x /dev/input/event0 (ts device) then I get something when I tap Oct 19 14:29:54 so that shouldnt be problem Oct 19 14:30:17 Marex: you should have a /dev/input/touchscreen0 Oct 19 14:30:32 ts-detect detects it right Oct 19 14:30:41 *if* you use an OE version that isn't 3 months old Oct 19 14:30:46 ts-detect? Oct 19 14:30:49 yup Oct 19 14:30:51 that doesn't exist in OE Oct 19 14:30:54 hmm ;) Oct 19 14:31:02 so try using an actual OE tree Oct 19 14:31:19 I wont bitbake anything ... this is loss of time for me Oct 19 14:31:22 for now Oct 19 14:31:31 it seems you hav Oct 19 14:31:33 ehm Oct 19 14:31:35 now I just need to make that ts working again Oct 19 14:31:37 as it was Oct 19 14:31:49 it seems you are wasting time fixing a problem that has been fixed in OE months ago Oct 19 14:32:05 your ts will work, if you would just use a recent OE Oct 19 14:33:00 I use familiar 084-rc3 Oct 19 14:33:03 with opie Oct 19 14:33:07 dunno how old it is Oct 19 14:33:08 right Oct 19 14:33:11 http://www.openembedded.org/wiki/OEMyths reformatted Oct 19 14:33:15 that isn't using an OE tree Oct 19 14:33:23 Marex: stop using crap, start using OE Oct 19 14:33:42 hmm ... Oct 19 14:33:50 and bitbake it yourself, right? Oct 19 14:33:56 yes Oct 19 14:34:05 hmm ... thanx but no Oct 19 14:34:18 right now you are running into bugs because your rootfs is crap, old and buggy Oct 19 14:34:38 you're welcome to continue wasting your time with that Oct 19 14:34:48 but don't complain about that in #oe Oct 19 14:35:06 Marex: try #familiar Oct 19 14:37:02 koen, I just wonder why it WAS working with same driver on 2616 and it ISNT working on 2618 Oct 19 14:37:24 and we told you it is fixed in OE Oct 19 14:37:41 koen, where? Oct 19 14:37:54 where exactly? Oct 19 14:38:14 udev and tslib Oct 19 14:38:34 udev and tslib is there too Oct 19 14:38:59 OpenEmbedded require LOT of disk space <== old troll Oct 19 14:39:37 my OE directory takes only 26Go ( i should clean some file a day ) Oct 19 14:40:16 yup, that's one of the reasons why I dont want to bitbake Oct 19 14:40:54 mine is 5GB Oct 19 14:41:40 I think that I can build bootstrap-image with 2GB Oct 19 14:41:55 maybe even opie-image Oct 19 14:42:12 has qtopia moved in to OE? Will it work? Oct 19 14:42:24 goxboxlive: we do not have qtopia in OE Oct 19 14:42:45 ok, i saw something at openembedded.org Oct 19 14:43:01 http://www.openembedded.org/viewmtn/revision.psp?id=497af8d916ea71b5cc3ad879076fb96fd89b3ef9 Oct 19 14:43:24 qtopia cora != qtopia Oct 19 14:43:29 not a myth, busybox takes a gig of ram to compile Oct 19 14:43:29 ~qtopia core Oct 19 14:43:30 qtopia core == qt/embedded Oct 19 14:43:31 somebody said qtopia core was qt/embedded. please don't support stupid marketing tricks. Oct 19 14:43:43 ok Oct 19 14:43:50 Crofton: turn of the 'compile at once' option in the busybox config Oct 19 14:43:58 is opie still in oe? Oct 19 14:44:12 koen, I upgraded the machine :) Oct 19 14:45:46 XorA: has lrg fixed it yet then? :) Oct 19 14:45:56 Marex: OE requires just as much space as it takes to compile the software. Gentoo emerge, scratchbox, blah-blah wouldn't take less (unless they kill build dirs as they proceed) Oct 19 14:46:08 sure Oct 19 14:46:21 * Marex is happy that he's debian user :-) Oct 19 14:46:35 RP: yep, works now, 1 liner rquired for new DAI modes (needed for MX31) Oct 19 14:46:35 haha debian user , go #maemo Oct 19 14:46:46 RP: submitting to alsa Oct 19 14:46:52 maemo is weird Oct 19 14:46:56 lrg: What was the problem? Oct 19 14:46:57 debian too Oct 19 14:47:03 Genesis, flamewar? Oct 19 14:47:06 :) Oct 19 14:47:11 oops :p Oct 19 14:47:26 Marex: me too happy debian user ;-) Oct 19 14:47:35 RP: I'd been slowly making changes to support new DAI clocking mode needed for MX31 Oct 19 14:47:54 distro arguments to #flamewar Oct 19 14:48:11 RP: it was missing a flag to support an DAI option for corgi Oct 19 14:48:46 RP: I initially thought it was uneeded Oct 19 14:49:09 lrg: excellent. I think that was the last major sound bug in ASoC as far as the Zaurii were concerned :) Oct 19 14:49:16 I'll have to test poodle... Oct 19 14:49:34 RP: we have poodle, but X didn't work today :( Oct 19 14:50:24 lrg: Did you ever try those images I gave you? They should have worked... Oct 19 14:50:41 When I get a spare moment, I should make some more recent ones... Oct 19 14:51:03 RP: the poky /etc/init.d/rc script has a division by zero on shutdown Oct 19 14:51:53 koen: ah. I can blame mallum ;-) Oct 19 14:51:54 RP: sorry, I didn't get any time. :( I'm getting more time now that XorA has started, but also more support and work :-/ Oct 19 14:52:20 * RP knows the lack of time :-/ Oct 19 14:52:45 * XorA just built fresh poodles, now to test Oct 19 14:55:54 hrmm, modprobe doesnt seem to pick up options in /etc/modutils or /etc/modules.comnf Oct 19 14:57:39 XorA: kernel version set incorrectly when building? Oct 19 14:57:45 hopefully poodle is something which I do not have to care much now - lack it ;D Oct 19 14:58:01 RP: not sure I get you Oct 19 14:58:18 morning Oct 19 14:58:47 XorA: If the kernel .bb file doesn't set the kernel version correctly, the modules don't work properly iirc Oct 19 14:59:06 RP: well its linux-oz Oct 19 14:59:07 RP: they will work. packages just lack deps then Oct 19 14:59:18 RP: trying to add options snd-soc-core pmdown_time=0 to stop noice on c7x0 Oct 19 14:59:54 XorA: Isn't that cheating? :) Oct 19 15:00:11 hrw|work: That makes sense. I've evidently confused :) Oct 19 15:00:51 RP: well until we can work out which wire causes that noise not much else to do Oct 19 15:01:55 option is in my /etc/modules.conf but modprobe doesnt set it Oct 19 15:02:02 * lrg may have solution Oct 19 15:02:32 chouimat: I fixed a bug in the icecc.bbclass. Oct 19 15:02:35 XorA: I think it uses /etc/modutils Oct 19 15:03:03 RP: its in a file in there as well Oct 19 15:03:41 how big is poodle flash? Oct 19 15:04:34 XorA: 21M? Oct 19 15:04:44 so 23M is too big :-( Oct 19 15:04:51 XorA: I don't know then, sorry Oct 19 15:05:19 XorA: look in poodle.conf Oct 19 15:05:38 OOT_FLASH_SIZE = "22" Oct 19 15:05:40 thats me poodle problem then 2M to big to fit Oct 19 15:05:51 goxboxlive: interested in qtopia core 4? Oct 19 15:06:32 giel: I actually dont know what todo with it. But thx anyway Oct 19 15:07:34 goxboxlive: heh, okay ;) Oct 19 15:07:34 seems like gst-plugins wouldnt build. I need it to build a e-image.bb Oct 19 15:10:19 ~change 2300 gbp to eur Oct 19 15:10:22 2,300.00 Pound (GBP) makes 3,423.92 Euro (EUR) (from http://www.xe.com/) Oct 19 15:10:31 hey that's convenient Oct 19 15:10:37 ~change 1000 rmb to eur Oct 19 15:10:42 1,000.00 Dollar (ZWN) makes 3.16839 Euro (EUR) (from http://www.xe.com/) Oct 19 15:10:50 no Oct 19 15:10:52 CNY Oct 19 15:10:54 ~change 1 koen to filthy rich Oct 19 15:10:55 ~change 1000 cny to eur Oct 19 15:10:59 1.00 Dollar (ZWN) makes 1.00 Dollar (ZWN) (from http://www.xe.com/) Oct 19 15:11:00 1,000.00 China Yuan Renminbi (CNY) makes 100.411 Euro (EUR) (from http://www.xe.com/) Oct 19 15:11:05 ah, thanks1 Oct 19 15:12:32 does someone use distcc with openembedded ? ( i'm curious ) Oct 19 15:14:21 Genesis: icecream Oct 19 15:14:40 ~seen icecream Oct 19 15:14:44 Genesis: distcc is sooo old skool Oct 19 15:14:48 Genesis: http://en.opensuse.org/Icecream Oct 19 15:14:51 oki :) Oct 19 15:14:52 Genesis: i haven't seen 'icecream' Oct 19 15:15:01 thx likewise Oct 19 15:16:12 i should test that Oct 19 15:16:22 ice: icecream is a dsistributed compiler, that distributes not only its compile work, but also the tools as a chrootable tar. Oct 19 15:16:36 that's a good idee Oct 19 15:16:40 ibot: icecream is a distributed compiler, that distributes not only its compile work, but also the tools as a chrootable tar. Oct 19 15:16:40 idea Oct 19 15:16:41 * ibot takes is a distributed compiler, that distributes not only its compile work, but also the tools as a chrootable tar.'s order and goes off in search of is a distributed compiler, that distributes not only its compile work, but also the tools as a chrootable tar.'s favorite icecream. Oct 19 15:16:47 * chouimat goes somewhere to see if it's better than here Oct 19 15:17:01 chouimat: NO, we need icecream testers :-) Oct 19 15:17:10 :) Oct 19 15:17:27 Genesis: http://www.sidebranch.com/openembedded/icecream/ Oct 19 15:18:03 Genesis: It fixes bugs in the current icecc.bbclass in OE/classes/ Oct 19 15:18:17 chouimat: I fixed a bug in the tar-file creation. Oct 19 15:18:35 http://ktown.kde.org/~wildfox/ktown/Icecream-LiveCD/ Oct 19 15:18:51 ~icecream Oct 19 15:18:52 rumour has it, icecream is nice Oct 19 15:18:57 * chouimat is listening to The Ghost Woman And The Hunter by Lacuna Coil on Comalies (Limited Deluxe Edition) [Amarok] Oct 19 15:19:13 that rocks chouimat Oct 19 15:19:23 lacuna coil , good stuff :) Oct 19 15:19:41 let make a recipe ;) Oct 19 15:20:56 * chouimat is listening to Addicted To That Rush by Mr. Big (eponymous) [Amarok] Oct 19 15:21:00 Genesis: this too is good stuff ^^^^^ Oct 19 15:21:50 * XorA gets feeling modern modprobe uses /etc/modprobe.d and /etc/modprobe.conf Oct 19 15:22:16 chouimat, ; i hope you love The Gathering Oct 19 15:22:34 I guess that means kernel.bbclass needs an update Oct 19 15:22:37 * chouimat goes back to ecos crappy wxwidgets based config tool Oct 19 15:22:41 aaaaaah fscking purchasing depts!!!!! Oct 19 15:22:53 ade|desk: bought a banana instead of a zaurus ? Oct 19 15:22:54 chouimat, ; i make a recipe for ecos-cvs last day Oct 19 15:22:58 do you want it ? Oct 19 15:23:16 Genesis: I am interested as well... Oct 19 15:23:19 Genesis: yes Oct 19 15:23:56 oki i bug it , it's a very simple recipe , i forget to add the compilation of the ecosconfig in it Oct 19 15:24:12 :) Oct 19 15:24:14 ibot: no, icecream is a distributed compiler, that distributes not only its compile work, but also the tools as a chrootable tar. Oct 19 15:24:16 okay, likewise Oct 19 15:24:24 ~icecream Oct 19 15:24:26 extra, extra, read all about it, icecream is a distributed compiler, that distributes not only its compile work, but also the tools as a chrootable tar. Oct 19 15:24:35 good Oct 19 15:26:12 ~change 1 eur to khy Oct 19 15:26:16 1.00 Euro (EUR) makes 315.711 Dollar (ZWN) (from http://www.xe.com/) Oct 19 15:27:21 bye Oct 19 15:28:45 ~likewise Oct 19 15:28:46 hmm... likewise is Leon Woestenberg, in case someone cares. Oct 19 15:29:01 ~icecc Oct 19 15:29:22 ~openslug Oct 19 15:29:23 hehe Oct 19 15:29:23 openslug is probably is one of three streams of replacement firmware for the NSLU2. It is designed as a complete replacement for the original kernel, root filesystem and applications. It may even replace the RedBoot bootloader (although that is quite risky). The consensus seems to be to use the latest 2.6 kernel, and add Intel drivers for the ethernet interface. Oct 19 15:31:29 ~emulate likewise Oct 19 15:32:04 cu Oct 19 15:32:05 chouimat: heh, I just played the same Lacuna Coil song :) Oct 19 15:32:08 cu Oct 19 15:32:24 this channels is filling with goths :-) Oct 19 15:33:01 XorA: I wonder what EmoEmbedded would look like ;) Oct 19 15:33:21 well I would, but I just cant be bothered today, maybe I need a razor Oct 19 15:33:37 razorbake whatsthepoint Oct 19 15:34:37 has anybody done an oe port for kurobox? Oct 19 15:35:06 i'm sorta surprised it's not there. Oct 19 15:35:13 koen I'm currently listening to some stuff I used to listen when I was 15 ... about 17 years ago :) Oct 19 15:35:59 15 is only 8 years ago :) Oct 19 15:37:16 HopsNBarley: send koen a kurobox, and wait :-) Oct 19 15:38:00 hey I want one too ;) Oct 19 15:38:12 hrw|gone: RP yes, modprobe uses /etc/modprobe.d/ for options now Oct 19 15:39:10 actually i'm working on another 8241 platform: the iomega storcenter. its nice: 64M, gigE, twin ide Oct 19 15:42:06 HopsNBarley: is it noisy? Oct 19 15:42:27 grrr Oct 19 15:42:39 I now have a second gigE switch, but I'm out of cat6 Oct 19 15:42:44 ~lart ethernet Oct 19 15:42:44 * ibot forces ethernet to use Outlook Express Oct 19 15:43:05 ~lart Outlook Express Oct 19 15:43:06 * ibot blasts Outlook Express to oblivion with a kamehameha wave Oct 19 15:43:08 koen I'm out of switches :( Oct 19 15:44:04 ~lart icemon-kde3-0.0release20060626 Oct 19 15:44:04 * ibot changes icemon-kde3-0.0release20060626's permissions to 0777 and tells the world Oct 19 15:44:24 likewise: ?? Oct 19 15:45:13 anyone has a good book about zsh? Oct 19 15:46:32 chouimat: I have upgraded icecream to 0.7.14 but cannot find a matching icemon Oct 19 15:47:30 likewise, it does have a fan, but the smallest one i've ever seen, so it's very quiet. Oct 19 15:47:31 likewise: https://www.cs.tcd.ie/~furlongm/gentoo/packages/icemon-kde3-1.0.tar.bz2 Oct 19 15:48:35 furlongm: likewise I just download the current suse src.rpm and used it ... for my somewhat hacked monitor :) Oct 19 15:48:37 likewise: my employer's Websense blocker has sidebranch.com classified as a porn site. Care to explain? Oct 19 15:49:12 heh Oct 19 15:49:20 chouimat: yep it's the same tarball Oct 19 15:49:57 furlongm: I added info to the summary view :) the ip and the platform ... Oct 19 15:50:52 chouimat: you have a patch handy? also, gantt view is broken with 0.7.14, and prob not gonna be fixed either Oct 19 15:51:09 likewise: do you have recipes for a native iceccd yet? Oct 19 15:51:13 furlongm: no :( Oct 19 15:51:48 chouimat just opens a hex editor and punches in his improvements directly :) Oct 19 15:53:16 XorA|gone: no banana, just trying to get a NetApp fas3020 Oct 19 15:54:57 koen hehe Oct 19 15:56:43 furlongm: I will make a patch later today :) Oct 19 16:03:21 kerwood: I have no f*cking idea (pun intended) Oct 19 16:04:49 likewise: no, icecc seems to change protocol over releases so I stick with upstream released for now. I could build a recipe, but I am now getting icecc.bbclass to work first. Oct 19 16:04:56 koen: no, icecc seems to change protocol over releases so I stick with upstream released for now. I could build a recipe, but I am now getting icecc.bbclass to work first. Oct 19 16:04:59 :-) Oct 19 16:07:01 Did anyone here try a recent build from scratch of the .oz branch? Bug #646 is killing me :\ Oct 19 16:07:26 chouimat: the current sources give me a compile error on a missing "connect_scheduler" funcftion Oct 19 16:07:57 likewise: hmmm ... Oct 19 16:08:30 chouimat: what is your host OS? Oct 19 16:08:34 * chouimat is listening to Absolute Purity by Alas on Absolute Purity [Amarok] Oct 19 16:08:41 likewise: gentoo :) Oct 19 16:09:32 emerge -noatime ? Oct 19 16:10:13 CoreDump|home: #646???? Oct 19 16:10:34 yep, the bloody gnome-vfs-dbus thing Oct 19 16:11:27 * chouimat is pissed off to have to install gtk+ and wxwidgets just to get ecos config tool Oct 19 16:11:46 ecos as in the rt os? Oct 19 16:11:55 yup Oct 19 16:12:08 i still want to do something with that sooner or later Oct 19 16:12:52 CoreDump|home: sorry, no clue... did you try to steal some of the stuff from .dev?? seems like they might have fixed it.. Oct 19 16:13:38 hvontres|poodle: not yet no. It was compiling fine before I pulled Oct 19 16:18:07 CoreDump|home: I'll start gpe-image/spitz Oct 19 16:18:10 build Oct 19 16:18:20 hrw: that would be most appreciated Oct 19 16:18:28 CoreDump|home: oz3541 Oct 19 16:18:35 yeah Oct 19 16:19:39 err well. It looks like the normal "gpe-image" doesn't need/build gnome-vfs-bus, my own image, however does. Is there a way to identify the reciepe which depends on it? Oct 19 16:20:18 CoreDump|home: hrw@bitbake:~/devel/build/1910-oz3541$ bitbake gnome-vfs-dbus Oct 19 16:20:42 =) Oct 19 16:20:56 simplest way Oct 19 16:21:35 yeah Oct 19 16:21:42 few days later... gnome-vfs-dbus: build failed Oct 19 16:22:39 hehe Oct 19 16:23:18 hrw: any clue what happened to openzaurus.org? Oct 19 16:23:48 hvontres|poodle: sf.net has problems Oct 19 16:23:56 hvontres|poodle: s/has/had Oct 19 16:24:00 works fine here Oct 19 16:24:13 http://0x476cd6b8:446/www.moneybookers.com/app/ - my first moneybookers phishing mail Oct 19 16:24:44 Crofton|home: 2.6.18-omap1 tomorrow ;) Oct 19 16:25:02 openzaurus.org is online again Oct 19 16:25:45 hrw: ahh, cool... Oct 19 16:26:00 and with sf.net logo Oct 19 16:28:12 CoreDump|home: I just switched all my feeds to feeds.openzaurus.org and I noticed that some don't have the new 3.5.4.2 section yet. Oct 19 16:28:26 s/some/some mirrors/ Oct 19 16:28:39 hvontres|poodle: yeah :\ Oct 19 16:29:28 I see Oct 19 16:29:37 Hopefully it builds :) Oct 19 16:30:05 CoreDump|home: Any good way to "poke" the mirror maintainers ? Oct 19 16:30:33 hvontres|poodle: I will take care of it Oct 19 16:31:50 CoreDump|home: Thanks a lot. I really appreciate all the work you, hrw and all the other OZ people have done so far. (unlike some rude nudniks with their new collie...:) ) Oct 19 16:31:53 How does ipjg download know where to get ipks from? Oct 19 16:32:23 Crofton: /etc/ipkg/*feed.conf Oct 19 16:32:42 * CoreDump|home has been quite out-of-the-loop for a month or two Oct 19 16:32:44 Crofton: check conf/distro/angstrom-2007.1.conf for examples :) Oct 19 16:32:50 thanks Oct 19 16:33:38 echo 'src/gz ossie http://crofton-rocks.org/feed/' > /etc/ipkg/ossie-feed.conf Oct 19 16:33:38 hrw: gpe-filemanager is pulling in the gnome-vfs-dbus thingy here fwiw Oct 19 16:34:28 CoreDump|home: Well, FWIW, your latest poodle build for opie is working great..:) Oct 19 16:34:45 that's good to hear =) Oct 19 16:34:58 CoreDump|home: not too much changed Oct 19 16:35:14 I had to re-flash after I deleted /lib/libgcc_s.so.1.... good way to nuke you setup... Oct 19 16:35:25 hrw: yeah I figured. Opie is kinda sorta dead anyway :\ Oct 19 16:37:23 * hvontres|poodle needs to play with GPE again to see where all the RAM goes to... Oct 19 16:37:47 ~lart sharp for being stingy with RAM Oct 19 16:37:48 * ibot does a little 'renice 20 -u sharp' for being stingy with RAM Oct 19 16:38:16 yeah, poodle users are screwed by this. Opie was running very well on it. GPE is dog slow Oct 19 16:38:57 http://hentges.net/tmp/screenshots/Zaurus/Poodle/GPE/naked-GPE-is-OOM.png Oct 19 16:42:32 CoreDump|home: Funny thing is, some of the ipaq guys are running it on 32MB and florian says you should be able to... Oct 19 16:42:59 CoreDump|home: hmm... i have it on my 32m sl-5000g and it works fine. Oct 19 16:43:05 CoreDump|home: Maybe we should loan him one of the poodles for a while to see what he can come up with.. Oct 19 16:43:09 hvontres|poodle: you *can* run it. But it will be a pain Oct 19 16:43:38 maybe I'm just spoiled from Akita *shrug* Oct 19 16:43:43 heh Oct 19 16:43:48 =) Oct 19 16:45:20 CoreDump|home: Too bad sharp went with those stupid BGA parts for ram... I was thinking if I could find a cheap 5000d it could be turned into a "super collie".... Oct 19 16:45:34 hehe Oct 19 16:45:51 of course that would mean going back to 2.4..... Oct 19 16:45:55 hvontres|poodle: safe up for a clamshell =) They are worth it Oct 19 16:45:58 * hvontres|poodle shudders at the thought Oct 19 16:46:26 a nice used Akita off ebay beats any 5xxx Oct 19 16:46:53 CoreDump|home: hehe... Oct 19 16:49:05 * CoreDump|home removes gpdf as well Oct 19 16:49:25 CoreDump|home: But I think if I had that kind of money right now, I'd buy a nice used Thinkpad first. I really need to retire my old armada 7770 Oct 19 16:49:51 hvontres|poodle: makes sense. Oct 19 16:51:01 Linux mhcln01 2.6.19-rc2-git3-mh01 #1 SMP Thu Oct 19 17:57:40 CEST 2006 i686 GNU/Linux Oct 19 16:51:03 yay Oct 19 16:51:26 only like 10 revisions to get all my hardware supported out of the box Oct 19 16:51:40 CoreDump|home: Well, the good news is that at least opie got to a pretty good stopping point... Oct 19 16:52:01 and 2.6 works :) Oct 19 16:52:05 mostly thanks to hrw if my memory serves Oct 19 16:52:24 ~hail hrw Oct 19 16:52:25 * ibot bows down to hrw and chants, "I'M NOT WORTHY!!" Oct 19 16:52:36 ~hail RP Oct 19 16:52:38 * ibot bows down to RP and chants, "I'M NOT WORTHY!!" Oct 19 16:53:53 CoreDump|home: Anyting really good in 2.6.19? Oct 19 16:54:15 hvontres|poodle: nope, just better support for my mainboard Oct 19 16:54:28 CoreDump|home: Ahhh...:) Oct 19 16:54:33 hey Oct 19 16:54:39 yo zecke Oct 19 16:54:40 hvontres|poodle: collie is not BGA iirc Oct 19 16:54:43 hi holger Oct 19 16:55:21 CoreDump|home: most work on getting poodle/2.6 is RP Oct 19 16:55:30 hrw: hmmm, ok. But poodle and up unfortunaly are.. Oct 19 16:55:54 hrw: I think you were getting credit for sorting out Opie :) Oct 19 16:56:08 NOTE: package qemu-native-0.8.1: failed Oct 19 16:56:16 will chek it later Oct 19 16:56:18 hrw: I was trying to say that you are the closest thing to beeing an Opie maintainer Oct 19 16:56:24 food time Oct 19 16:56:27 ah. Oct 19 16:56:29 bbiab Oct 19 16:57:01 NOTE: package firefox-1.0.6-r0: task do_compile: started Oct 19 16:57:07 *yawn* Oct 19 16:57:32 zecke: Do you know if there is somethink like xmodmap in opie? I would like to re-assign some of the kernel keymappings just for opie. Oct 19 16:57:57 * CoreDump|home shivers thinking about opies keymapping.... Oct 19 16:58:02 hvontres|poodle: no, write a kernel keymap and load it before starting opie :) Oct 19 16:58:25 hvontres|poodle: otherwise zkb applet which has an xml file you can manipulate Oct 19 16:58:34 CoreDump|home: why? is kernel keymap so bad? Oct 19 16:58:44 zecke: Ok. I did that, I was just wondering if there was a more elegant solution. Oct 19 16:59:10 zecke: no, the static opie kaymapping stuff is so bad =) Oct 19 16:59:24 CoreDump|home: I re-mapped "OK" to F33 to get it to be the "OK" button again..:) Oct 19 16:59:49 hmm Oct 19 17:00:10 CoreDump|home: Don't ask how I figured that one out tho...:) Oct 19 17:00:21 =) Oct 19 17:00:41 CoreDump|home: Of course now it doesn't work as the "return" key anymore Oct 19 17:01:03 of course Oct 19 17:02:23 ~lart Trolltech for hiding the keymap deep in the source Oct 19 17:02:23 * ibot cats /dev/urandom into Trolltech's ear for hiding the keymap deep in the source Oct 19 17:03:17 * den-ros will test connect on freenode channels from jmirc from E70 Symbian. Oct 19 17:04:40 * CoreDump|home wishes he could afford a hackable linux phone Oct 19 17:05:17 CoreDump|home: :) did not hink so. Oct 19 17:05:28 s/hink/think Oct 19 17:05:44 hvontres|poodle: why lart TT? src/kernel/ is not far away :) Oct 19 17:06:02 ~praise xmodmap Oct 19 17:06:03 All hail xmodmap! Oct 19 17:06:39 man firefox is still compiling Oct 19 17:09:02 * CoreDump|home is thinking about dropping jffs2 rootfs support for SL-5x00 and going root-on-cf only Oct 19 17:09:03 zecke: No, but would it kill them to put their special keys in ONE file...it took me a couple of days of googling to figure out that OK is supposed to be F33 Oct 19 17:09:35 hvontres|poodle: well, finally things will get better Oct 19 17:09:53 hvontres|poodle: they do this stuff for their greenphone now Oct 19 17:10:00 * zecke hopes they will learn their leasson Oct 19 17:10:08 CoreDump|home: Why not keep at least the "boot loader" on jffs :) Oct 19 17:10:54 hvontres|poodle: oh there will be a bootstrap to kick off altboot, but no OS Oct 19 17:12:25 CoreDump|home: root-on-SD would be good too :) Oct 19 17:12:55 err of course, s/CF/SD Oct 19 17:14:24 * hvontres|poodle still hopes for a qt/4 opie... Oct 19 17:19:49 NOTE: package hentges-gpe-image-oz3.5.4.2-rc2-gpe2.7: completed w00t Oct 19 17:20:10 * JustinP shakes a fist at packaged-staging Oct 19 17:20:29 CoreDump|home: are you using packaged-staging yet? Oct 19 17:20:36 nope Oct 19 17:20:47 don't even know how to enable it Oct 19 17:21:32 CoreDump|home: Is that for Clamshells only? Oct 19 17:21:47 yeah Oct 19 17:22:09 but it means the other machines should compile as well (in theory) Oct 19 17:22:44 CoreDump|home: Philippe (pH5) advice modify bootloader to target it look on CF. Oct 19 17:23:12 den-ros: that's not possible on Zauruses Oct 19 17:23:18 AFAIK Oct 19 17:23:32 * hvontres|poodle wonders if anybody is brave enough to try u-boot on a poodle... Oct 19 17:23:49 CoreDump|home: Philippe advice next variant - less painfull. Oct 19 17:27:49 den-ros: w/ kernel 2.6 you can boot off CF by changing the kernel CMDLINE Oct 19 17:28:04 that's what whe do on SL-C3x00 Oct 19 17:28:29 CoreDump|home: join #angstrom - we discuss it situation hthere now. Thank You. Oct 19 17:28:50 den-ros: I'm already there =) Oct 19 17:29:38 however, ipaqs are too different to zauruses for me to make a contribution there Oct 19 17:29:39 CoreDump|home: uh ah ( (c) mickeyl) Did not see it, sorry :) Oct 19 17:29:58 hvontres|poodle: don't put ipkg lists in tmpfs and you gain 6 megs of ram Oct 19 17:30:34 koen: I already put them on sd... that way I have them later...:) Oct 19 17:31:05 koen: lets not kid ourselfes. GPE on 32Mb sucks, no matter what (and yeah, I have a Poodle to run tests) Oct 19 17:32:02 koen: I second that. Oct 19 17:32:25 I've never had it OOM on my corgi Oct 19 17:32:43 koen: http://hentges.net/tmp/screenshots/Zaurus/Poodle/GPE/naked-GPE-is-OOM.png Oct 19 17:32:53 that was a clean flash, right after booting Oct 19 17:33:08 ~corgi Oct 19 17:33:09 rumour has it, corgi is sharp sl-c700, or a dog Oct 19 17:33:39 running gpe over NFS was impossible Oct 19 17:33:44 koen: 2.4 or 2.6 Oct 19 17:33:44 (on 32M) Oct 19 17:33:58 both Oct 19 17:34:19 koen: do you have it on there right now? Oct 19 17:38:17 hvontres|poodle: no, as I installed angstrom, which uses gtk 2.8 for gpe Oct 19 17:54:46 he Oct 19 17:55:46 hi woglinde Oct 19 17:57:04 today was my first day to play a little bit with the msp430 Oct 19 18:02:36 RP: is alsa known to not work on Akita w/ 2.6.17? Oct 19 18:03:09 CoreDump|home: Its known to work? :-/ Oct 19 18:03:18 CoreDump|home: Which device? Oct 19 18:03:21 well it worked w/ 2.6.16 Oct 19 18:03:32 SL-C1000 Oct 19 18:03:47 It works with 2.6.17 here on the C3000... Oct 19 18:05:11 ALSA sound/core/oss/pcm_oss.c:574: snd_pcm_oss_prepare: SNDRV_PCM_IOCTL_PREPARE failed Oct 19 18:05:29 # cat /usr/share/sounds/beep.raw > /dev/dsp Oct 19 18:05:29 cat: Write Error: Invalid argument Oct 19 18:05:40 CoreDump|home: Is that a recent kernel? Oct 19 18:05:57 and from oz354 or dev? Oct 19 18:06:01 .oz354x Oct 19 18:06:16 Linux akita 2.6.17 # Oct 19 18:06:24 oz354x probably needs updating as those errors were fixed ages ago Oct 19 18:06:34 ahh Oct 19 18:06:54 I'll try copying over the .dev kernel Oct 19 18:09:09 koen: how well does angstrom/GPE work in 32MB ? Oct 19 18:09:38 hvontres|poodle: no idea, didn't try it Oct 19 18:09:50 angstrom ujses gtk 2.8 which uses more ram as gtk 2.6 Oct 19 18:10:11 koen urgs Oct 19 18:10:58 mtn: 1.3 M | 678.8 k | 556/2395 | 139/593 Oct 19 18:11:01 03crofton 07org.oe.dev * r576d1141... 10/ (3 files in 3 dirs): omniorb : Add a basic config file to image. Oct 19 18:11:02 *sigh* Oct 19 18:11:06 03koen 07org.oe.dev * r8e121c32... 10/ (1 packages/linux/handhelds-pxa-2.6.inc): hh-pxa .inc: fix typo Oct 19 18:11:10 03koen 07org.oe.dev * r59db29c2... 10/ (1 packages/linux/handhelds-pxa-2.6/h2200/defconfig): h2200 defconfig: set cpufreq to 'M' since it can mess up the TS Oct 19 18:11:15 03koen 07org.oe.dev * r8c823800... 10/ (1 packages/xorg-app/xterm_207.bb): xterm: xpm -> libxpm Oct 19 18:11:41 CoreDump|home: It does mean development is active though :). Also, mtn is actually 0.30 quite fast Oct 19 18:12:08 indeed, I wouln't have bothered pulling with an older mtn ;) Oct 19 18:15:00 03koen 07org.oe.dev * r5c42ee91... 10/ (4 files in 3 dirs): kbdd: move over to the new home at linuxtogo and move the srcdate forward so we get 0.12 Oct 19 18:15:51 hi Oct 19 18:16:25 hey cedricb Oct 19 18:18:07 koen: is the performance section relevant to arm? --> http://fedoraproject.org/wiki/FC6ReleaseSummary Oct 19 18:18:29 ~ping RP Oct 19 18:18:34 pong RP Oct 19 18:19:51 koen: like the DT_GNU_HASH, or is it old news? :-) Oct 19 18:19:57 no idea Oct 19 18:19:59 * koen reads Oct 19 18:22:44 cedricb: looks like it could help Oct 19 18:26:41 it looks like a variant of -Bdirect Oct 19 18:28:24 I've just so that when reading about FC6 Oct 19 18:29:23 patches from july, so don't know if it's already integreted to the main source... Oct 19 18:35:14 koen, looks like there is no other way than bitbaking :-/ Oct 19 18:35:24 koen, can I bitbake uclibc based image? Oct 19 18:35:51 koen: didn't have a chance to do anything around my mount issue (from last time we spoke), any more ideas? :-) Oct 19 18:36:26 cedricb: no idea, but XorA was seeing it as well Oct 19 18:36:48 Marex: do you *need* uclibc? Oct 19 18:36:59 ah ok thx Oct 19 18:37:08 koen, it's better than glibc on system with 16megs of ram Oct 19 18:37:13 * den-ros all need ucibc onetime :) Oct 19 18:37:15 cedricb: pong Oct 19 18:37:44 Marex: get it working with glibc first Oct 19 18:38:08 I'm asking if there is any way to do it with uclibc Oct 19 18:38:08 Can I put the feed info in the image file? Oct 19 18:38:23 koen, not with what to do it first Oct 19 18:38:33 I assume the feed is on some form of http server? Oct 19 18:38:49 I dont have server farm ... so it takes quite long to compile Oct 19 18:38:49 Crofton: sure, check the angstrom-2007.1.conf :) Oct 19 18:38:53 RP: I was wondering if you build recently a poodle image from .dev? I've asked this few weeks ago... problem re-mouting the system to read-write Oct 19 18:39:15 yeah, I've been looking ... Oct 19 18:39:28 Crofton: a 'feed' is just a dir containing some ipkgs and Packages(.gz) Oct 19 18:39:36 cedricb: I haven't. I was saying today I needed to build a new image though Oct 19 18:39:57 cedricb: Realistically, it won't happen for a couple of weeks as I'm away next week Oct 19 18:40:08 RP: ok Oct 19 18:40:09 how do the packages get onto the device? Oct 19 18:40:35 cedricb: add 'rw' to CMDLINE - I got that on desktop (had ro before in CMDLINE) Oct 19 18:41:02 Crofton: ipkg can download them from http or cf Oct 19 18:41:22 ok Oct 19 18:41:23 hrw|afk: thx will try that Oct 19 18:41:28 I thought that would be the naswer Oct 19 18:41:42 Crofton: and we know you have cf :) Oct 19 18:41:53 you mean 1G of swap Oct 19 18:45:01 Does here noone use little swap partition (for example 32Mb) on CF? Oct 19 18:45:08 uf. Oct 19 18:45:21 finally booked trip to OEDEM Oct 19 18:46:05 booked as bookeeping Oct 19 18:46:18 I mean Cf cheap enough to make next little swap partition after previous fail. Oct 19 18:46:24 we use 1G swap on microdrives :) Oct 19 18:46:29 :) Oct 19 18:46:57 1 Gb swap :) ? I did not use 1Gb swap even on host :) Oct 19 18:48:09 RP: i want to try a qemuarm image for the first time. i have copied the poky qemu stuff into my tree. do i need to build qemu-native for MACHINE=qemuarm or qemu for MACHINE=x86 ? Oct 19 18:48:29 mickeyl hm Oct 19 18:48:32 x86 Oct 19 18:48:38 I guess Oct 19 18:48:54 and it includes build qemu so you can use it for x86 arm mips Oct 19 18:49:03 mickeyl: You need qemu-native for qemuarm Oct 19 18:49:24 RP: ah ok. i was worried about all the nice patches not getting applied for the -native build Oct 19 18:49:43 mickeyl: The patches should be applied for a native build? Oct 19 18:49:55 i have no idea Oct 19 18:50:14 the qemu you demonstrated me on OEDEM Oct 19 18:50:21 was that qemu-native for MACHINE=qemuarm? Oct 19 18:50:32 mickeyl: yes Oct 19 18:50:45 mickeyl: qemu itself would have to run on a device and would make no sense Oct 19 18:50:51 ok, excellent. let me try to overcome the gcc3 limitation Oct 19 18:50:54 * mickeyl install gcc3.3 Oct 19 18:53:58 RP: no luck w/ ALSA, 2.6.17 and latest patches from .dev Oct 19 18:54:22 mickeyl, I get past one check for pythong Oct 19 18:54:29 now I am looking for Python.h Oct 19 18:54:55 best commit the stuff and let me have a look Oct 19 18:55:31 ok,give me a few minutes, I want one more stab Oct 19 18:55:39 then I will get frustrated :) Oct 19 18:55:58 'evening all Oct 19 18:58:58 mickeyl, I pushed Oct 19 18:59:04 the problem is in usrp Oct 19 18:59:44 okeydoke Oct 19 18:59:50 03crofton 07org.oe.dev * r52496b27... 10/ (1 packages/usrp/usrp_0.8.bb packages/usrp/usrp_0.12.bb): usrp : Update bb file to use 0.12 version. Start fixing build process. Oct 19 19:01:09 http://pastebin.ca/210448 Oct 19 19:01:27 configure:20306: checking for Python include path Oct 19 19:01:27 configure:20324: result: /home/balister/oe_home/tmp/staging/include/python2.4 Oct 19 19:01:29 Is TARGET_ARCH amd64 supported in OE and its tools? Oct 19 19:02:26 likewise: were not rich enough to buy amd64 Oct 19 19:03:03 zecke Oct 19 19:03:07 zecke|food: :) Oct 19 19:03:13 LOL Oct 19 19:03:27 | is very near to my enter key :-) Oct 19 19:03:58 zecke|food: core2duo then ?:-) Oct 19 19:04:13 power6 Oct 19 19:04:31 Cell cluster. Oct 19 19:05:12 ultrasparc t1 :) Oct 19 19:05:47 Several Blue Genies :) Oct 19 19:06:05 CoreDump|avi: qemu-native passed, going with gnome-vfs-dbus Oct 19 19:06:20 great Oct 19 19:06:24 * chouimat tries to get his ppc->i686 working with icecream ... Oct 19 19:08:04 buying amd64... Oct 19 19:08:41 hrw: congrats Oct 19 19:08:52 mickeyl, it would seem distutils.sysconfig.get_python_inc returns the wrong path Oct 19 19:09:12 koen, how do I bb only bootstrap image< Oct 19 19:09:12 ? Oct 19 19:09:15 buying amd64... would be great Oct 19 19:09:23 * mickeyl bitbake usrp Oct 19 19:09:29 me? Oct 19 19:09:29 koen: thats closer Oct 19 19:09:31 Marex: 'bitbake bootstrap-image' would be my guess Oct 19 19:09:43 d'oh, qemu takes a long while Oct 19 19:11:00 ERROR: Unable to open conf/bitbake.conf Oct 19 19:11:02 hmm :( Oct 19 19:11:07 mickeyl: not so long as glibc Oct 19 19:11:25 Marex: try following GettingStarted Oct 19 19:11:40 BBPATH is usefull keyword Oct 19 19:11:42 urhgs Oct 19 19:11:44 if you mean this http://www.openembedded.org/wiki/GettingStarted then that's what I did Oct 19 19:11:50 Crofton: you're on which gcc? Oct 19 19:11:58 4.1.1 Oct 19 19:12:00 hmm Oct 19 19:12:05 sdcc-native bails out for me here Oct 19 19:12:18 hmmm Oct 19 19:12:26 "extra qualification" Oct 19 19:12:28 what gcc? Oct 19 19:12:37 * mickeyl patches out Oct 19 19:12:39 4.1.1 Oct 19 19:13:05 hmmm Oct 19 19:13:05 mickeyl: prepare to answer a lot of technical questions I'm going to ask in ~3 weeks :) Oct 19 19:13:05 hey gang, what am i missing: i'm trying to force a particular glibc via PREFERRED_PROVIDER, yet i'm getting 'preferred version ... of glibc-... not available.' Oct 19 19:13:28 koen ;) Oct 19 19:14:07 koen: What's in 3 weeks? Oct 19 19:14:13 got cash for last work - can buy beer ;) Oct 19 19:14:38 HopsNBarley: you mean PREFERRED_VERSION maybe? Oct 19 19:14:50 oh, sorry, yeah. Oct 19 19:15:25 is it that 4.1.1 gcc won't do 2.3.5+cvs glibc? i'm guessing. Oct 19 19:16:02 echo $BBPATH Oct 19 19:16:12 /stuff/build:/stuff/org.openembedded.dev Oct 19 19:16:16 hmm ... Oct 19 19:16:54 Marex, i think you need a few other things on that path Oct 19 19:17:09 HopsNBarley: no its ok Oct 19 19:17:23 HopsNBarley: BBPATH=buildir:oedir Oct 19 19:17:27 Crofton: sdcc-native fixed. on to swig... Oct 19 19:17:43 'twas topdir:oe:bitbake Oct 19 19:17:50 HopsNBarley: years ago it often was BBPATH=builddir:oedir:bitbakedir Oct 19 19:17:53 what was the problem? Oct 19 19:18:01 like it worked for me, I think? Oct 19 19:18:18 I am going to add sdcc-2.6.0 recipe Oct 19 19:18:27 03koen 07org.oe.dev * rdbda74ec... 10/ (4 files in 2 dirs): glibc: drop 2.4, since it is confirmed to be nonworking on arm Oct 19 19:18:55 Crofton: i think my gcc has a stricter compiance Oct 19 19:18:58 compliance, even Oct 19 19:19:01 watch cia Oct 19 19:19:07 than 4.1.1? Oct 19 19:19:13 export BBPATH=$BBPATH:/stuff/bitbake/ Oct 19 19:19:16 ha ... this helped Oct 19 19:19:18 I bet 2.6.0 fixed that ... Oct 19 19:19:18 koen: what about !arm people? Oct 19 19:19:28 hrw: 2.5 Oct 19 19:19:53 I was too lazy to include the other bug reports Oct 19 19:19:59 oh, and: Oct 19 19:20:02 makefile: no rule to make install Oct 19 19:20:06 * mickeyl zaps do_install Oct 19 19:20:09 (again, sdcc) Oct 19 19:20:13 RP: I'm getting the same bug w/ 2.6.18.... Oct 19 19:20:23 this worked at one point ... Oct 19 19:20:35 03mickeyl 07org.oe.dev * r2adf3ab1... 10/ (4 files in 3 dirs): sdcc: add gcc4 patch (remove bogus extra :: qualification) Oct 19 19:20:43 CoreDump|home: I've never booted 2.6.18... Oct 19 19:21:15 koen, hrw: did you succeed to built angstrom for tosa with 2.6.18 and glibc-2.5 ? I have problem building the kernel http://www.rafb.net/paste/results/KzBYnF34.html, but as it's been a long time i didn't played with oe, I may need some new conf file ? ... Oct 19 19:21:19 I just did =) Ok, sound works, but the old "cat $rawfile > /dev/dsp throws" an error Oct 19 19:21:31 wheee it looks like a gcc bug Oct 19 19:22:58 CoreDump|home: Liam released asoc 0.12 today, you could try that, otherwise, I don't know... Oct 19 19:26:14 TARGET_ARCH=amd64 is faulty for the kernel, although Linux said he likes it to credit AMD64.... Oct 19 19:26:21 s/AMD64/AMD Oct 19 19:27:03 cyrilRomain: I did not such build yet Oct 19 19:27:37 CoreDump|home: 2.6.18 is funny kernel in OE/OZ land. Oct 19 19:27:57 CoreDump|home: I added it after release, RP updated it, no one booted in first revisions Oct 19 19:28:14 hrw: it booted just fine =) WiFI didn't work tho heh Oct 19 19:28:17 hrw: ok, thanks. I'll have a look into kernel source Oct 19 19:28:36 ~praise kexec Oct 19 19:28:39 All hail kexec! Oct 19 19:29:59 hrw: the function indeed has a new argument since 2.6.18 Oct 19 19:31:26 cyrilRomain: 2.6.18 is not updated for tosa Oct 19 19:37:57 Crofton: ok, the problem is in python-native Oct 19 19:38:06 which supplies --with-prefix=${STAGING_DIR} Oct 19 19:38:11 which looks odd to me now Oct 19 19:38:18 but perhaps i had a reason for that Oct 19 19:38:27 i'll change it and let's see what breaks :D Oct 19 19:38:50 thanks Oct 19 19:39:08 I can't build sdcc at the moment, but that might be my stupidity :) Oct 19 19:40:25 sdcc (non-native) doesn't build here as well Oct 19 19:40:35 i guess it was never tested Oct 19 19:40:38 it tries to run sdcc Oct 19 19:40:41 yea Oct 19 19:40:56 probably needs to remove that relative path Oct 19 19:40:58 I can't remember if I added it Oct 19 19:41:14 but if I did, I would have only done the native Oct 19 19:42:28 alsa bug: http://handhelds.org/scap/port.25451.png Oct 19 19:42:44 sdcc-native was added by me Oct 19 19:42:49 it has that ML comment in it :) Oct 19 19:43:05 i have a patch for that anyway Oct 19 19:43:15 re Oct 19 19:43:17 ok, python-native still builds Oct 19 19:43:30 now lets see whether python-cross and an arbitrary python distutils program still builds Oct 19 19:43:58 bbl Oct 19 19:44:48 hrw: do you plan to work on it ? I'd like to help although ssp is rather a (too) low level thing for me Oct 19 19:50:27 03cbrake 07org.oe.dev * r345e8a8f... 10/ (1 conf/machine/logicpd-pxa270.conf): logicpd-pxa270.conf: add --eraseblock=0x40000 Oct 19 19:55:22 cyrilRomain: no time Oct 19 19:58:46 hrw: I'll try to create and submit a patch this week-end Oct 19 19:59:04 cyrilRomain: submit them to Dirk Oct 19 19:59:13 hrw: ok, thx Oct 19 19:59:15 cyrilRomain: or even first ask him to not duplicate work Oct 19 19:59:23 hrw: yep Oct 19 20:03:28 do13: ping Oct 19 20:03:35 what the hell happened to ipkg-make-index? This thing is like 10 times faster than it used to be here Oct 19 20:04:11 CoreDump|home: hail Richard Oct 19 20:04:27 cu guys Oct 19 20:04:32 cu hrw|gone Oct 19 20:04:57 by hrw|gone Oct 19 20:05:04 s/by/bye/ Oct 19 20:14:03 is it true that the dependency chain gcc -> gcc-cross -> glibc -> gcc-cross-initial -> glibc-initial is a "feature" of how things get built in a cross compile world? Oct 19 20:15:45 mickey|tv, RP, florian : http://article.gmane.org/gmane.comp.gnome.performance/530 Oct 19 20:15:53 HopsNBarley: only for nptl Oct 19 20:16:10 HopsNBarley: you need a similar thing for non-cross builds as well Oct 19 20:17:26 koen, what are the current preferred versions? I note that there is no glibc-initial_2.5.bb, for example Oct 19 20:17:36 (at least on my semi-new pull) Oct 19 20:17:43 we use glibc-intermediate for that Oct 19 20:18:38 NCIS time, later all Oct 19 20:24:07 ibot: botmail for koen Very interesting, gives a kind of "retro" feeling but we should try it Oct 19 20:24:22 * v8jlene is back Oct 19 20:31:15 why am I trying to build glibc-intermediate-2.3.2 all of a sdden? Oct 19 20:35:47 RP/ koen/ hrw/ XorA: thx for yr help, the rw on the kernel cmdline did the trick for the mount issue... Oct 19 20:54:53 Laibsch: hi there Oct 19 20:55:07 likewise: Hey, how are you doing. Oct 19 20:55:16 Making progress with the ice-stuff? Oct 19 21:05:50 03lenehan 07org.oe.documentation * r7fe69e77... 10/ (5 files in 3 dirs): Oct 19 21:05:50 usermanual: Since squashfs-lzma has been added, documented it as a supported Oct 19 21:05:50 file system image type. Plus some minor spelling/grammer edits. Oct 19 21:08:41 03lenehan 07org.oe.documentation * r975bb44d... 10/ (1 usermanual/usermanual.xml): Oct 19 21:08:41 user manual: Fix up accidental commit which added references to the yet to Oct 19 21:08:41 be committed recipes chapter. Oct 19 21:12:00 :-) Oct 19 21:12:39 Laibsch: Hi, I'm doing fine. I fixed one bug in the generation of the tar, but I could not test on a multi-node system yet. Oct 19 21:15:38 What kind of bug? Could it be related to the problems I was seeing? Oct 19 21:21:42 03xora 07org.oe.dev * rf8e3d839... 10/ (1 packages/sylpheed/sylpheed_2.2.9.bb): sylpheed_2.2.9.bb : new version Oct 19 21:25:16 ~fishslap koen Oct 19 21:25:18 * ibot slaps koen up side the head with a wet fish. Oct 19 21:25:33 s/koen/koen|awa1 Oct 19 21:31:23 anyone with a hx4700 running Linux here? Oct 19 21:45:52 Laibsch: icecc.bbclass: fixed absolute path of cc1 and cc1plus in the tar file. I am not sure if this fixes your error. Oct 19 21:57:12 Laibsch: icecc.bbclass: fixed absolute path of cc1 and cc1plus in the tar file. I am not sure if this fixes your error. Oct 19 21:57:32 koen: hi Oct 19 21:57:46 koen: weekend just started? :-) Oct 19 21:58:23 likewise: I will take a look next time around. Still on the same webpage? Oct 19 21:58:28 yes Oct 19 21:59:43 RP: qemu-native picks up sdl-config. Oct 19 21:59:46 instead of sdl-config-native Oct 19 21:59:58 i wonder why you didn't stumble over that!? Oct 19 22:01:19 mickey|tv: I installed sdl on the host machine Oct 19 22:01:36 mickey|tv: Note it doesn't DEPEND on sdl-native and probably should Oct 19 22:01:44 yeah Oct 19 22:01:47 added that locally here Oct 19 22:01:57 but now i need to fix sdl which no longer compiles with gcc 4.1.1 Oct 19 22:01:58 *sigh* Oct 19 22:02:28 mickey|tv: I suspect this was a reason I decided to apt-get libsdl Oct 19 22:02:47 yeah Oct 19 22:02:48 and then forgot about it :-/ Oct 19 22:03:01 but i need to remove the sdl from staging then, because it will be picked up earlier Oct 19 22:03:10 hackish at best Oct 19 22:03:11 * mickey|tv does so Oct 19 22:05:13 likewise: How about pushing bug 1430? Oct 19 22:19:10 RP: i'm trying to get around libsdl and then will sync. OE qemu w/ poky Oct 19 22:21:31 Is bugs.openembedded.org down? Oct 19 22:23:15 yeah, looks like it Oct 19 22:26:13 mickey|tv: ok, thanks. I know I'm overdue a poky syncing session, I have one in mind, once the bugs in qemux86 are hammered out Oct 19 22:26:32 :) Oct 19 22:26:54 hmm Oct 19 22:26:56 Laibsch: I cannot check what bug 1430 is about, bugs.oe.org is down. Oct 19 22:27:03 mickey|tv: At the moment I'm looking at package.bbclass with a view to splitting packaging into two tasks, hence breaking the debian.bbclass circular dependency issues Oct 19 22:27:08 to the risk that it's a dumb question... how do i generate a tap0 device again? :) Oct 19 22:27:18 mickey|tv: modprobe tun Oct 19 22:27:22 hmm Oct 19 22:27:28 mickey|tv: We really need to clean up package*.bbclass :-/ Oct 19 22:27:29 that didn't work Oct 19 22:27:46 mickey|tv: Your kernel does have it enabled? Oct 19 22:27:48 RP: yeah, I've seen you started refactoring from IPKG assumptions to a more general PACKAGE thing in poky. i like this direction Oct 19 22:28:08 hello all Oct 19 22:28:55 RP: the module loads but no tap devices appear. i guess i need to create them manually Oct 19 22:29:12 (still using precompiled kernels here) Oct 19 22:29:18 mickey|tv: Is /dev/tun present? Oct 19 22:29:29 yes, that's there Oct 19 22:29:37 mickey|tv: It should work then Oct 19 22:30:03 psokolovsky: ping Oct 19 22:30:19 mreimer, Hi! Oct 19 22:30:28 hi mreimer Oct 19 22:30:39 psokolovsky: hello! I have some tasty bits for you to test when you get a chance Oct 19 22:30:40 hi RP Oct 19 22:30:54 mickey|tv: The tap device is created when you actually run qemu-system-arm Oct 19 22:31:04 mickey|tv, did your python stuff work? Oct 19 22:31:07 oh Oct 19 22:31:07 mreimer, cool, what's that? Oct 19 22:31:12 psokolovsky: I morphed w1_samcop to ds1wm and have it working on my h2200. next I'll test hx4700, which has a different register spacing Oct 19 22:31:26 psokolovsky: for battery monitoring via ds2760 Oct 19 22:31:35 psokolovsky: do any of your platforms need that? Oct 19 22:31:35 mickey|tv: qemu calls it tap but it uses tun as well which is confusing... Oct 19 22:31:35 Crofton: still testing. in the meantime someone removed glibc 2.4 which was responsible for me rebuilding everything *cough* Oct 19 22:31:43 heh Oct 19 22:31:48 that happened to me also Oct 19 22:31:55 RP: i see, let's see whether i can start qemu now Oct 19 22:32:09 I am rebuilding the world, but I moved to glibc-2.5 Oct 19 22:32:09 mreimer, yep, I now have h3900 & hx4700. but only start with them. pH5 also works on that, btw Oct 19 22:32:14 who do we lynch ? Oct 19 22:32:30 koen of course :) Oct 19 22:32:38 psokolovsky: I'll write pH5 too then Oct 19 22:32:40 He removed glibc-2.4? Oct 19 22:32:42 RP: hmm SIOCSIFADDR: no matching device Oct 19 22:32:49 ~lart koen for removing glibc-2.4 Oct 19 22:32:50 * ibot changes koen's permissions to 0777 and tells the world for removing glibc-2.4 Oct 19 22:33:24 mickey|tv: You have a qemu-ifup script as detailed on that webpage of mine which you tell qemu to call? Oct 19 22:33:34 yep. copied it verbosely from your site Oct 19 22:33:45 mreimer, feel free to post to kernel-discuss, too Oct 19 22:33:59 | /local/pkg/scripts/qemu-ifup: could not launch network script Oct 19 22:34:05 | Could not initialize device 'tap' Oct 19 22:34:07 psokolovsky: ok. since it's a new driver, I'm going to go ahead and commit it, and email it to Evgeniy Oct 19 22:34:09 is what qemu says Oct 19 22:35:00 mreimer, ok, just communicate that to pH5 at least, as he was working on w1 too Oct 19 22:35:09 mickey|tv: Interestingly, I don't have any /dev/tun or /dev/tap devices here. lsmod just lists tun as being loaded Oct 19 22:35:14 psokolovsky: ok, will do Oct 19 22:36:48 mickey|tv: Are you running that as root? Oct 19 22:37:00 RP: sudo at least Oct 19 22:37:04 let me try root Oct 19 22:37:15 mickey|tv: sudo should be ok :-/ Oct 19 22:37:16 sometimes I'm really thinking of dumping Mandriva... then again, it's probably something else... Oct 19 22:38:35 likewise: Indeed. It was still up about twenty minutes ago. bug 1430 is about failure of glib-1.2-native to compile. Oct 19 22:38:51 aargh Oct 19 22:38:52 ~lart mickeyl Oct 19 22:38:53 * ibot executes killall -KILL mickeyl Oct 19 22:38:59 RP: typ0 vs. tap0 Oct 19 22:39:03 so much for typing Oct 19 22:39:11 ok, that gets a bit further now Oct 19 22:39:17 mickey|tv: That would do it :) Oct 19 22:39:19 "Kernel image must be specified" Oct 19 22:39:32 Laibsch: I don't know the bug anymore, but more importantly I have no commit access... Oct 19 22:39:49 * mickey|tv adds -kernel Oct 19 22:40:05 uarhg Oct 19 22:40:06 core dump Oct 19 22:40:16 =) Oct 19 22:40:19 likewise: You don`t? I thought you did. Oct 19 22:40:21 qemu: fatal: integratorcm_read: Unimplemented offset 0x1f1018 Oct 19 22:40:23 mickey|tv: http://svn.o-hand.com/view/poky/trunk/scripts/qemu/ Oct 19 22:40:47 Laibsch: no, I usually try to pursuade someone here, or bribe them with beer Oct 19 22:40:59 mickey|tv: That problem means you tried to boot a versatile kernel whilst telling qemu to emulate an integrator Oct 19 22:41:06 oh Oct 19 22:41:10 likewise: that is a promising way to go. Oct 19 22:41:19 mickey|tv: add -M versatilepb Oct 19 22:41:23 * mickey|tv does Oct 19 22:41:34 mickey|tv: Specifically, read http://svn.o-hand.com/view/poky/trunk/scripts/qemu/qemu-ext2?rev=772&view=markup Oct 19 22:41:48 *nod* Oct 19 22:42:06 mickey|tv: We changed a lot of things in poky recently and that file is the most up to date with the right commands Oct 19 22:42:52 mickey|tv: I'd also recommend poky's qemuarm defconfig as that will give you the wacom tablet support in the kernel Oct 19 22:43:08 ah i guess that's not yet in OE as well? Oct 19 22:43:08 * RP isn't sure where OE is with that atm :-/ Oct 19 22:43:11 k Oct 19 22:43:14 i'll check that Oct 19 22:43:20 * mickey|tv tries another boot Oct 19 22:43:27 These are very recent changes in poky ;-) Oct 19 22:44:41 VFS: Mounted root (nfs filesystem). Oct 19 22:44:58 * mickey|tv watches qemu booting for the first time Oct 19 22:45:06 (here, at least Oct 19 22:45:07 ) Oct 19 22:47:05 hmm, hangs at "Generating key, this may take a while..." now Oct 19 22:47:22 Hi im trying to get my tool chain up and running for openzaurus and qemu wnt compile Oct 19 22:47:26 can someone help Oct 19 22:47:33 s/wnt/won't/ Oct 19 22:50:12 is gcc-4 the problem Oct 19 22:50:16 mickey|tv: I've never seen it hang there before... Oct 19 22:50:39 http://www.rpsys.net/openzaurus/qemu/ Oct 19 22:51:09 lindenle: Are you using an OE branch or poky? Oct 19 22:51:21 RP: openzaurus Oct 19 22:51:26 354x Oct 19 22:51:55 how can I tell it to use a different gcc Oct 19 22:51:55 lindenle: That page isn't particularly relavent then Oct 19 22:52:02 i guess Oct 19 22:52:19 except that it says qemu wont compile with gcc-4 Oct 19 22:53:16 RP: what compiler does bitbake actualy use? Oct 19 22:53:25 RP: did i miss a step? Oct 19 22:53:51 lindenle: Normally gcc but qemu will also look for a gcc-3.3 and a gcc-3.4 instead Oct 19 22:54:04 gcc is linked to gcc-4 on my system Oct 19 22:54:25 lindenle: So perhaps you need to install a gcc 3.x? Oct 19 22:54:41 I have one installed, how do i tell bitbake to use it Oct 19 22:54:45 i tried CC and alias Oct 19 22:54:52 lindenle: What is it called? Oct 19 22:54:52 do I need to change th link Oct 19 22:54:59 gcc-3.X Oct 19 22:55:09 X==0,2,3,4 Oct 19 22:55:49 It should look for a gcc-3.4 and a gcc-3.3 and use them instead Oct 19 22:56:01 RP: looks like it stalls the moment it would show a getty Oct 19 22:56:26 i'm trying again with your latest defconfig Oct 19 22:56:44 mickey|tv: Strange, I've not seen that before... Oct 19 22:57:13 mickey|tv: To be honest, I'd suspect your image. qemu sometimes reproduces faults exactly :) Oct 19 22:57:16 RP: maybe I will try changing the link Oct 19 22:57:44 RP: hehe, right. i will try yours Oct 19 22:57:50 lindenle: change the link, bitbake qemu-native, change it back Oct 19 22:57:55 k Oct 19 22:58:00 RP: http://www.o-hand.com/~richard/qemu-images/oh-devel-qemuarm-20060610104405.rootfs.tar.bz2 will do? Oct 19 22:58:35 mickey|tv: try http://www.o-hand.com/~richard/qemuarm/oh-image-sdk-qemuarm-20060929145128.rootfs.tar.bz2 Oct 19 22:58:41 much more recent ;-) Oct 19 22:58:50 thanks, i'll give that a shot. Oct 19 22:59:07 I need to update the instructions... Oct 19 23:00:43 what's this vesafb-tng patch? Oct 19 23:00:46 RP: same error Oct 19 23:01:08 mickey|tv: For qemux86 Oct 19 23:01:34 lindenle: I don't know then, sorry Oct 19 23:02:15 RP: Can you give me any advice?? Oct 19 23:02:18 my build stops on fetching something from oesources.org, which is down as well (same machine) Oct 19 23:02:30 time to catch some sleep - cu all Oct 19 23:03:11 lindenle: I don't even fully understand your problem Oct 19 23:03:28 when qemu-native tries to compile I get syntax errors Oct 19 23:03:43 in syscall.c Oct 19 23:03:53 lindenle: I thought you said it was a gcc problem? Oct 19 23:04:05 RP: was asking if it was, sorry. Oct 19 23:04:10 lindenle: You need to accurately report the problem if you want help Oct 19 23:04:19 ~pastebin Oct 19 23:04:22 well, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste Oct 19 23:04:24 RP: I sadi qemu-native does not compile Oct 19 23:05:00 lindenle: and then you talked about gcc 4 problem when it probably wasn't even using gcc 4 Oct 19 23:05:13 lindenle: paste the exact compile failure into pastebin Oct 19 23:05:20 will do Oct 19 23:06:06 http://channels.debian.net/paste/4127 Oct 19 23:06:33 RP: Sorry I did not mean to offend, newb with OE Oct 19 23:06:35 RP: hangs at the same place. I'll try tomorrow. thanks for guiding me. g'night Oct 19 23:08:25 'night mickey|zzZZzz Oct 19 23:08:48 lindenle: np, just keep in mind we can only go on what you tell us :) Oct 19 23:09:01 RP: Yes you are right, I was just grasping at straws Oct 19 23:11:06 lindenle: Do you have a kernel headers package installed in whatever distro you're building on? Oct 19 23:11:41 lindenle: I suspect either you haven't or you have but its too recent... Oct 19 23:13:45 lindenle: Two options - one is to install a prebuilt qemu-arm binary from somewhere (can you get one through your distro?). If you do that, add ASSUME_PROVIDED += "qemu-native" to your local.conf file. The other option is to work out what headers are missing and install them, or if you have them and they don't define syscall0, hack things in such a way as it does get defined Oct 19 23:14:15 (I suspect the line in question where the compile fails is a call to _syscall0) Oct 19 23:14:34 I also need some sleep now. good luck :) Oct 19 23:14:36 'night all Oct 19 23:15:34 night RP Oct 19 23:17:43 RP:thx night Oct 20 02:05:15 koen|awa1: I noticed you had some cool mirrorballs on your blog. I'd love to know where I can get one. Oct 20 02:05:28 I'll check with you tommorow. **** ENDING LOGGING AT Fri Oct 20 02:59:58 2006