**** BEGIN LOGGING AT Wed Oct 11 02:59:56 2006 Oct 11 04:54:18 hi Oct 11 04:54:20 anyone awake? Oct 11 04:54:49 anyone here try compiling the 2005 konqueror for openzaurus? Oct 11 04:55:02 the one currently in the oz feed is 2003 Oct 11 04:59:05 ~seen minipanda Oct 11 04:59:24 minipanda was last seen on IRC in channel #oe, 26d 17h 21m 50s ago, saying: 'crap, in $OESTUFF/build/tmp/rootfs/ have /dev, but in bootstrap-image.tar.gz, no!'. Oct 11 06:50:00 Morning Oct 11 06:53:52 yo :) Oct 11 06:59:04 Can you help me starting my X? Oct 11 07:01:02 goooood morning fooks Oct 11 07:04:20 goooood morning Mr. President Oct 11 07:18:12 I would like to submit the following packages to .dev: lzma (non-native), mksquashfs-tools 3.1r2 (cross, native, both with and without lzma compression). Someone around to commit? Oct 11 07:23:20 03lenehan 07org.oe.dev * ra6c7cd41... 10/ (1 packages/havp/havp_0.82.bb): Oct 11 07:23:20 havp 0.82: Inherit update-rc.d so that the initscripts actually get Oct 11 07:23:20 registered. Fix up the naming of initscript variables now that they are Oct 11 07:23:20 actually getting used. Oct 11 07:25:25 03lenehan 07org.oe.dev * r2fb56192... 10/ (7 files in 4 dirs): Oct 11 07:25:25 clamsmtp 1.8: Add clamsmtp, an smtp proxy with virus scanning support Oct 11 07:25:25 cia clamav (clamad-daemon). It has optional support transparent Oct 11 07:25:25 proxying as well. Oct 11 07:27:03 likewise: I'm here... Oct 11 07:27:43 v8jlene: ok, tnx Oct 11 07:28:13 likewise: Were you going to bugzilla them? Oct 11 07:28:27 v8jlene: is that preferred? Oct 11 07:28:49 koen: hi koen Oct 11 07:28:57 likewise: Yeah, that's what is normally done... Oct 11 07:29:17 v8jlene: I'll submit to bugzilla, then point you guys to it so you can commit. Oct 11 07:29:31 good morning all Oct 11 07:29:51 likewise: Genesis made a .bb for antigrain yesterday Oct 11 07:30:32 koen: morning Oct 11 07:30:39 koen: yippie. Oct 11 07:30:53 koen: someone else is pushing the antigrain backend into cairo. Oct 11 07:31:16 w00t Oct 11 07:31:17 NOTE: package glibc-2.5-r0: task do_build: completed Oct 11 07:31:45 koen: I am finishing up lzma (non-native), mksquashfs-tools 3.1r2 (cross, native, both with and without lzma compression). MyFirstRecipe(tm). Oct 11 07:32:34 cute Oct 11 07:32:46 it's me :D Oct 11 07:34:11 morning Oct 11 07:34:17 koen: I'll give glibc 2.5 a go for sh4 tomorrow (and if that works i486 as well). Oct 11 07:37:38 morning Oct 11 07:38:49 hey hrw|work Oct 11 07:38:55 morning all Oct 11 07:40:07 likewise: nice work on squash/lzma Oct 11 07:40:12 hey ade|desk Oct 11 07:44:33 koen: so when can you tell us glibc2.5 fixes EABI problems :-) Oct 11 07:44:47 XorA: ehm.... Oct 11 07:48:44 hrw|work: tnx Oct 11 07:49:20 I published the squashfs{,-lzma}{,-native} stuff here: http://openembedded.sidebranch.com/ Oct 11 07:50:20 Need some testers. I am not sure if I should bring the kernel squashfs support up to date, looking into that now. Oct 11 07:50:28 cu Oct 11 07:57:18 likewise: looks like I'll be unable to test it: ../LZMA/LZMADecoder.cpp:188: internal compiler error: Segmentation fault ;( Oct 11 07:58:14 v8jlene: report which compiler that is. Most probably, disable an optimization flag to work around this. Oct 11 07:58:40 likewise: You'll want to run your .bb through contrib/oe-stylize.py at some point Oct 11 07:59:14 v8jlene: The kernel squashfs 3.1 support patch (from the official squashfs release) also gives me an internal compiler error on armeb-linux-gcc 4.1.1 if I compile with -Os instead of -O2. Oct 11 07:59:28 v8jlene: Is it that bad? :-) Oct 11 07:59:33 Good morning Oct 11 07:59:42 likewise: That's was with gcc 3.4.4 for sh4. Oct 11 07:59:46 good morning B_Lizzard, mithro Oct 11 08:00:01 Is the command I need "INHERIT =+ rm_work" ? Oct 11 08:00:23 The command that goes in local.conf Oct 11 08:01:06 INHERIT += "rm_work". It sometimes stoppes the build in mid-air. Restarting continues. Oct 11 08:01:13 s/stoppes/stops Oct 11 08:01:42 So, it's not that good of an option to add? Oct 11 08:01:51 I've had no problems till now... Oct 11 08:01:54 B_Lizzard: just see if it works for you Oct 11 08:02:30 I used it for building world, but I deleted my local.conf by accident, and I couldn't remember the command. Oct 11 08:02:36 Anyways, thanks a lot Oct 11 08:04:04 v8jlene: Yes, LZMA seems to compile with -Os on, but I have not figured out who sets the "-Os", either OE or the makefile itself. Oct 11 08:05:12 v8jlene: mksquashfs is mostly useful for -native only (when you build your rootfs), maybe check if the two native packages compile correctly? Oct 11 08:05:28 likewise: Works for me if I add: SELECTED_OPTIMIZATION = "" Oct 11 08:06:10 likewise: SELECTED_OPTIMIZATION = "-Os" still segfaults. Oct 11 08:06:14 v8jlene: ok, where exactly? In the lzma bitbake recipe? Oct 11 08:06:27 likewise: Yeah, in the lzma .bb Oct 11 08:06:30 v8jlene: -O2 maybe? That often solves it for me Oct 11 08:06:44 likewise: -O2 was the default. Oct 11 08:06:52 v8jlene: ok Oct 11 08:07:41 likewise: For me -O1 and no optmisations works. Oct 11 08:07:41 sweet: http://bash.org/?696997 Oct 11 08:09:10 likewise: That should be BUILD_OPTIMIZATION not SELECTED_OPTIMIZATION. The later is change based on if you do a debug build or not. The former is set to -O2 by default (see conf/bitbake.conf) Oct 11 08:10:25 v8jlene: thanks for pointing that out. I am still learning my way into this system :-) Oct 11 08:11:09 v8jlene: Something like BUILD_OPTIMIZATION_sh4 = "-O1" ?? Oct 11 08:11:11 likewise: No... I'm going mad ;( FULL_OPTIMIZATION is the one used for the target... Oct 11 08:11:26 v8jlene: ok, third time right Oct 11 08:12:22 likewise: No problems for i486 with gcc 4.1.1 Oct 11 08:15:14 likewise: No problems on sh4 with gcc 3.4.4 and FULL_OPTIMIZATION_sh4 = "-O1" Oct 11 08:15:59 v8jlene: OK, it probably makes sense to add that line then? Oct 11 08:19:48 likewise: hmm, the lzma packages are empty... no do_install... shouldn't you also be creating an some lzma packages? Oct 11 08:24:02 likewise: As for the sh4 line... leave it out for now. c++ is pretty useless with gcc 3.x on sh4. I'd probably change the default optmisation globally for sh4 before putting sh4 specific optmisations all over the place. Oct 11 08:26:08 v8jlene: only the native do not have do_install Oct 11 08:27:20 v8jlene: checking Oct 11 08:27:33 hi Oct 11 08:27:53 1.2G in work machine Oct 11 08:27:58 need to rebuild kernel :( Oct 11 08:28:12 hrw|work: welcome to x86 Oct 11 08:28:36 likewise: "find tmp/work/lzma-4.17-r0/install" is empty. If I built lzma I would have expected an lzma package I could install ;) BTW, should there be some header files staged as well? Oct 11 08:29:01 v8jlene: you are taking care of pushing squashfs/lzma into .dev? Oct 11 08:29:25 first we are cleaning up my faults :-) Oct 11 08:29:28 hrw|work: Just testing it at the moment... Oct 11 08:30:24 ok Oct 11 08:34:23 likewise: squashfs-tools is ok for sh4 and i486. Probably needs a DEPEND = "zlib" since it links to zlib. Apart from that it looks ok. Oct 11 08:34:39 v8jlene: something like this? Oct 11 08:34:41 install -m 0755 7zip/Compress/LZMA_Alone/lzma ${D}${exec_prefix} Oct 11 08:34:43 oe_libinstall -a -C 7zip/Compress/LZMA_Lib liblzma ${D}${libdir} Oct 11 08:35:09 v8jlene: and some header file... Oct 11 08:36:42 03koen 07org.oe.dev * r1f11e315... 10/ (1 packages/glibc/glibc_2.5.bb): glibc 2.5: re-enable zecke-sane-readelf Oct 11 08:36:59 likewise: exec_prefix? I don't think that exists, it'd be bindir. But yeah, that's the idea. Remember to install -d the directories beforehand. Oct 11 08:37:33 i wonder want delights glibc 2.5 brings Oct 11 08:37:39 s/want/what Oct 11 08:37:48 ade|desk: new and interesting breakages probably Oct 11 08:37:57 yay Oct 11 08:38:14 RP: what csl version fixes the shm problems? Oct 11 08:38:47 4G for rootfs starts to be too small ;( Oct 11 08:39:06 ... says a collie owner Oct 11 08:40:45 ;D Oct 11 08:41:20 2.6.19-rc1 is comiing to this machine Oct 11 08:42:32 4G for / wow Oct 11 08:42:51 /dev/hda5 4,0G 3,3G 450M 89% / Oct 11 08:43:17 ade|desk: wow as 'so small?' or as 'big it is'? Oct 11 08:43:27 wow as in big Oct 11 08:43:50 ade|desk: no /usr /var /boot ones Oct 11 08:43:57 ah Oct 11 08:45:17 I hope tomas finishes his iwmmxt layer for kdrive soon Oct 11 08:46:18 ade|desk: I have: / /home /home/hrw/devel/build/ Oct 11 08:46:24 brb - reboot time Oct 11 08:49:37 re Oct 11 08:50:28 total used free shared buffers cached Oct 11 08:50:28 Mem: 1295848 277004 1018844 0 24916 130060 Oct 11 08:50:31 finally Oct 11 08:53:30 koen: The more recent one Oct 11 08:54:42 hey peoples Oct 11 08:55:11 morning all Oct 11 08:55:40 morning all Oct 11 08:55:52 hey RP Oct 11 08:56:28 RP: good news, only some cosmetic changes for pxa2xx stuff. Oct 11 08:56:50 lrg: excellent :) Oct 11 08:58:36 eeeks Oct 11 08:58:38 http://rafb.net/paste/results/OOUmxF14.html Oct 11 08:59:40 * koen tries to find 999 friends that want to buy a cm270 ;) Oct 11 09:00:13 Oo Oct 11 09:01:44 998 ;) Oct 11 09:05:07 koen: They're nice boards :) Oct 11 09:08:30 I just don't buy they 'below $50' argument Oct 11 09:20:56 someone know how expensive is omap5912osk? Oct 11 09:21:07 or other devboard with usb+cf? Oct 11 09:27:27 hrw|work: you can get pcmcia with the cm270 Oct 11 09:28:01 one lab in Poland want to equip robots with kind of machines with wifi and usb Oct 11 09:28:10 and asked me about zaurus for it Oct 11 09:28:27 nslu2 + usb wifi would be the cheapest solution I guess Oct 11 09:28:52 yeah Oct 11 09:30:00 it has 2 usb ports? Oct 11 09:30:23 yes Oct 11 09:30:31 2x *real& usb2 host Oct 11 09:30:35 s/&/*/ Oct 11 09:48:07 VoodooZ does a robot with an nslu2 board Oct 11 09:48:22 he does bluetooth communication with it via a usb dongle Oct 11 09:49:43 http://voodooz.no-ip.org:8000/robotics/srx2/index.html Oct 11 09:57:57 <__law__> hi whats the differts between the oz-dev/oz-dev.summery image? Oct 11 09:57:57 <__law__> s/summery/summary Oct 11 09:58:23 __law__: +10meg faster mounting for summary Oct 11 09:58:55 summary image is actually too big to load on my c7x0 Oct 11 09:59:36 <__law__> XorA, hmm i?ll try summary on my c760 perhaps it works.... Oct 11 10:42:05 hi kerwood Oct 11 10:44:36 gm Oct 11 10:44:54 hrw|work: They're about $US300: http://focus.ti.com/docs/toolsw/folders/print/tmdsosk5912.html Oct 11 10:44:54 hrw|work: Also, TI seems to have a number of distributors in PL: http://focus.ti.com/docs/general/distribcountryresults.jhtml;?countryCode=PL®ionId=3 Oct 11 10:44:54 hrw|work: I don't have one myself...yet... Oct 11 10:45:48 hi Crofton Oct 11 10:46:04 hi Oct 11 10:48:46 Crofton: hiya ... did you notice that eth0 isn't built into the new build ? Oct 11 10:48:55 um Oct 11 10:49:03 are you sure? Oct 11 10:49:45 i updated my org.oe.dev tree this morning, cleaned along the bb file, re-built. flashed to board and there's no eth0 Oct 11 10:50:01 defconfig shows ethernet enabled, tho ... from what i can see. not sure what changed Oct 11 10:50:27 I am port scanning at school looking for the osk Oct 11 10:50:41 ok Oct 11 10:50:58 I started doing some real work, and didn't check if I could ssh in Oct 11 10:50:59 could just be me, but figured i'd check Oct 11 10:51:03 yeah Oct 11 10:51:28 ~lart people using mysql-isms Oct 11 10:51:28 * ibot steals people using mysql-isms's mojo Oct 11 10:51:46 koen: such as auto_increment or worse? Oct 11 10:52:43 it seems postgres doesn't like INSERT INTO foo SET bar=baz Oct 11 10:52:54 ah. Oct 11 10:56:29 In other news: mr worth discards all previous profiling runs and runs his own benchmark on a not-so-representative system Oct 11 11:30:45 someone remember which license mickey|ttdevdays use for presentations? Oct 11 11:32:56 the one where you send me and zecke money ;) Oct 11 11:35:15 ;PP Oct 11 11:37:05 are there any boards in oe that are using the PXA270 and also have uboot working with them? I'm looking for a good starting point to get uboot working with my pxa270 board. Oct 11 11:39:20 Gerrath: I'd use the mainstone patch for a start porting it to a custom pxa270 board Oct 11 11:41:11 Gerrath: zauruses cxx00 has uboot but out of OE patches Oct 11 11:41:59 florian_kc, hrw|work Thanks, I will check out both of these options. Oct 11 11:56:07 Crofton: got my custom kernel build with lcd support working :) Oct 11 11:57:10 bigd0g: great Oct 11 11:58:26 special thanks to everyone for helping me along the way ;) Oct 11 11:58:34 hi Oct 11 11:58:41 ~seen zecke Oct 11 11:59:02 zecke was last seen on IRC in channel #oe, 1d 14h 46m 27s ago, saying: 'mickey|ttdevdays: cya in munich tomorrow'. Oct 11 11:59:22 koen: add full when you adding images on your blog Oct 11 11:59:33 koen: otherwise they are missing in RSS Oct 11 11:59:55 hm so mickeyl and zecke are in munich Oct 11 12:00:23 hrw|work: full? Oct 11 12:01:15 koen: Oct 11 12:01:24 not like you do Oct 11 12:01:57 actually I don't insert the images, but a plugin does Oct 11 12:02:50 bigd0g, great, network up? Oct 11 12:02:56 negative on the network Oct 11 12:03:07 I'll work on that later Oct 11 12:03:13 yeah, no worries Oct 11 12:03:20 i've got serial connection for now Oct 11 12:03:28 heh Oct 11 12:03:31 masochist Oct 11 12:03:37 >:) Oct 11 12:04:02 food time Oct 11 12:04:05 bbl Oct 11 12:06:52 <[lala_]> hi guys Oct 11 12:07:02 hi lala Oct 11 12:10:17 <[lala_]> is there something special about the git fetcher? it doesn't pick up the tarball from stash for xcalibrateext ... Oct 11 12:11:16 <[lala_]> who is able to help me with that? rp? Oct 11 12:19:26 <[lala_]> i noticed that there is no separate machine config file for borzoi anymore and it tells you to use spitz Oct 11 12:19:46 You want the tar.gz file? Oct 11 12:19:47 <[lala_]> but if i try to update my borzoi i geht an error: invalid model Oct 11 12:20:21 <[lala_]> im building from .dev Oct 11 12:21:23 [lala_]: spitz == spitz, borzoi, terrier now Oct 11 12:21:51 <[lala_]> ok, but why is the updater complaining then? Oct 11 12:22:40 <[lala_]> zaurus cf updater: model:sl-c3200, error: invalid model! please reset Oct 11 12:24:13 <[lala_]> hrw|work: any idea? Oct 11 12:24:57 if I want to generate the preproccessed files of a program in OE how do I do that? I tried CFLAGS_append = " -save-temps" but it did not work. Oct 11 12:24:58 [lala_]: I beleive I used the scipt from oz354x Oct 11 12:25:10 [lala_]: there must be differences Oct 11 12:25:44 <[lala_]> it looks like it :-) Oct 11 12:27:49 Laibsch: ah.. script last 3200 Oct 11 12:28:13 hrw|work: hi! Oct 11 12:28:17 hrw|work: ? Oct 11 12:28:28 hrw|work: What are you trying to tell me? Oct 11 12:28:49 s/ Laibsch / [lala_] / Oct 11 12:28:53 Or where you trying to message [lala_] Oct 11 12:28:56 OK Oct 11 12:29:00 <[lala_]> :-) Oct 11 12:29:35 I hate nicks with [ at the front since I need to press the shift button when calling them. Oct 11 12:30:41 Laibsch: if you use irssi it will complete them for you Oct 11 12:30:50 la -> [lala] Oct 11 12:31:00 I have gaim Oct 11 12:31:07 ~lart mtn log for slowness Oct 11 12:31:07 It does not complete them. Oct 11 12:31:07 * ibot keeps mailing mtn log free America Online CDs until he drowns for slowness Oct 11 12:31:34 <[lala_]> hrw|work: it seems that 3200 is missing in the updater.sh script Oct 11 12:31:46 <[lala_]> can you fix that? Oct 11 12:31:47 [lala_]: I know - will push fix Oct 11 12:31:51 <[lala_]> cool Oct 11 12:33:37 mtn: connecting to monotone.openembedded.org Oct 11 12:33:37 mtn: network error: failed to connect: Connection refused Oct 11 12:33:40 ;( Oct 11 12:34:19 ok. works now Oct 11 12:37:01 pushed Oct 11 12:37:31 hey guys , i need a bit help on a compile failed , weird problem with ld/zlib , i past the bug here http://code.bulix.org/43yjhj-19940 Oct 11 12:38:29 Genesis: you'd have to ask the o-hand dudes Oct 11 12:38:38 hm zlib Oct 11 12:38:41 as much as we like poky, we do not support it Oct 11 12:38:52 but poky is same Oct 11 12:38:58 no, it is not Oct 11 12:39:05 it's the same bitbake Oct 11 12:39:20 if it was the same, I wouldn't have spent so much time trying to merge the differences Oct 11 12:39:21 i'll migrate to OE when i'll have some time Oct 11 12:39:39 Genesis: EXTRA_LDFLAGS = " -L${STAGING_LIBDIR} " and rebuild Oct 11 12:39:49 what differences it have that would make my compilation error ?! Oct 11 12:39:56 i just want to understand Oct 11 12:40:36 for me poky was just an oe with some extra bbclass and bbfile like oh and oh-images , running on a subversion instead of a monotone Oct 11 12:41:12 i have no pb cp -a OE bbfile in my poky tree Oct 11 12:41:25 should i try angstrom , or you consider it's same pb as poky ? Oct 11 12:42:07 angstoem is a distribution Oct 11 12:42:14 oe and poky buildsystems Oct 11 12:42:30 poky is both the buildsystem and the distribution Oct 11 12:42:36 oki Oct 11 12:42:46 ( since we have a poky in distro conf ) Oct 11 12:42:53 isn't it ? Oct 11 12:43:01 *sigh* Oct 11 12:43:38 :/ Oct 11 12:44:38 oki it has been deleted or i was dreaming Oct 11 12:44:48 03hrw 07org.oe.dev * r75f97628... 10/ (1 packages/zaurus-updater/spitz/updater.sh): zaurus-updater: add SL-C3200 support from .oz354x Oct 11 12:44:53 03koen 07org.oe.dev * racbcf196... 10/ (1 packages/zaurus-updater/spitz/updater.sh): Oct 11 12:44:53 zaurus-updater: add c3200 support Oct 11 12:44:53 * please keep fixes from your private branches in sync with .dev... Oct 11 12:48:28 <[lala_]> hrw|work: if i read the updater.sh correctly it doesnt support bz2 images, right? Oct 11 12:49:09 yep Oct 11 12:49:25 <[lala_]> so why are the images created as bz2 as default? Oct 11 12:49:53 you have such image_fstypes set? Oct 11 12:50:17 <[lala_]> i only took a default local.conf and set openzaurus-unstable and spitz - nothing fency Oct 11 12:50:21 spitz has tar.gz set Oct 11 12:50:27 <[lala_]> hmm Oct 11 12:50:33 <[lala_]> let me double check Oct 11 12:52:58 <[lala_]> that's pretty weird Oct 11 12:53:13 <[lala_]> in the default local.conf you have: IMAGE_FSTYPES = "jffs2 tar" Oct 11 12:53:21 <[lala_]> thats also in my local.conf Oct 11 12:54:00 <[lala_]> ah, and in spitz.conf is: IMAGE_FSTYPES ?= "tar.gz" Oct 11 12:54:07 <[lala_]> so it doesn't get used Oct 11 12:54:13 <[lala_]> and tar means bz2? Oct 11 12:54:30 tar != bzip2 Oct 11 12:54:45 it's probably tar'd and then bzip'd Oct 11 12:55:05 <[lala_]> sure, but who tells oe to do that? Oct 11 12:55:16 IMAGE_FSTYPES = "tar" == tar.bz2 Oct 11 12:55:32 ^ there's your answer ;) Oct 11 12:55:48 what type of rootfs do you need generated? Oct 11 12:55:52 [lala_]: there is no "default" local.conf, there is an example local.conf Oct 11 12:56:05 <[lala_]> hrw|work: ok, so the default compression method changed, i see Oct 11 12:56:31 03koen 07org.oe.dev * rb17e00ba... 10/ (3 files in 2 dirs): Oct 11 12:56:31 glibc 2.5: remove some hardcoded references Oct 11 12:56:31 glibc-intermediate 2.5: add it Oct 11 12:56:41 [lala_]: no. 'tar' years ago was tar.bz2 and when we added tar.gz and tar.bz2 we left tar to be tar.bz2 Oct 11 12:57:52 <[lala_]> well, if the distros/machines set images_fstypes this should be reflected in the example local.conf Oct 11 12:58:26 <[lala_]> and tar.gz is not mentioned there as valid type Oct 11 12:59:28 patches welcome Oct 11 13:02:53 <[lala_]> the other fix would be to allow bzip2 in updater.sh as well ,-) Oct 11 13:03:14 <[lala_]> but i guess we dont have bzip2 in that environment Oct 11 13:03:26 <[lala_]> right? Oct 11 13:03:37 [lala_]: it could be added same way as gnu-tar is Oct 11 13:03:45 [lala_]: a project for you Oct 11 13:03:51 <[lala_]> hehe Oct 11 13:03:57 <[lala_]> fair enough Oct 11 13:04:54 <[lala_]> btw: what did i do wrong if bitbake keeps telling me OE_REVISION = ? Oct 11 13:05:24 [lala_]: you have newer monotone then bitbake know Oct 11 13:05:54 <[lala_]> hmm, its 0.28 Oct 11 13:06:00 <[lala_]> not that new Oct 11 13:06:42 <[lala_]> bitbake-1.6.0 Oct 11 13:09:17 or you have BBCOLLECTIONS Oct 11 13:09:34 <[lala_]> koen: thats it! Oct 11 13:10:10 <[lala_]> i have added BBCOLLECTIONS today because i had to apply a local patch to a packet Oct 11 13:10:31 <[lala_]> and i didnt want to mess around with the original tree Oct 11 13:17:38 hm Oct 11 13:18:46 why, if i have a package that provides virtual/${TARGET_PREFIX}gcc/g++ and virtual/libc, and set that package as PREFERRED_PROVIDER, oe still wants to build gcc-cross and glibc? Oct 11 13:21:26 bitbake also warns the package depends on itself, eventually, but there are no DEPENDS in the .bb file of that package Oct 11 13:24:34 okay, "inherit cross" helps Oct 11 13:26:23 morning Oct 11 13:35:13 hey chouimat Oct 11 13:36:05 hi chouimat Oct 11 13:49:55 bigd0g, I have a network interface that works ... Oct 11 13:50:36 hrm Oct 11 13:50:43 Crofton: recommendations? Oct 11 13:51:28 http://pastebin.ca/197724 Oct 11 13:51:32 check bootargs? Oct 11 13:54:57 <[lala_]> is BBCOLLECTIONS a feature from .dev or does it work in .oz354x as well? Oct 11 13:55:15 will work Oct 11 13:55:28 [lala_]: its BitBake thing rather then OE iirc Oct 11 13:55:41 <[lala_]> ok Oct 11 13:55:52 Anybody knows if the NPTL in latest glibc works fine for ARM? Oct 11 13:55:53 <[lala_]> is bitbake-1.6.0 ok for the branch as well? Oct 11 13:55:56 Mem: 1295848k total, 1256556k used, 39292k free, 83820k buffers Oct 11 13:56:00 [lala_]: yes Oct 11 13:56:07 I'd like to make use of it for the slug. Oct 11 13:56:17 slugos (ARM) Oct 11 13:56:24 never mind how much memory will linux machine get it will use all Oct 11 13:56:38 VoodooZ: no idea here Oct 11 13:56:39 <[lala_]> hrw|work: i was only wondering, because in the branch i don't see the unknown revision ... Oct 11 13:57:41 Crofton: hrm. i have 'ip=on' and everything set in bootargs Oct 11 13:57:43 hrw|work, Ok. I know last time I checked (~glibc-2.3.5) it was stilll using the old threading support as the ARM stuff wasn't quite there yet. Oct 11 13:57:56 they haven't changed since i updated u-boot to 1.1.4 in september Oct 11 13:58:07 <[lala_]> hrw|work: i switch to oz354x for now, because the opie-image from .dev is somehow broken it seems Oct 11 13:58:53 Crofton: during boot i get "IP-Config: No network devices available." ... i think that's the issue Oct 11 13:59:12 <[lala_]> hrw|work: it boots just fine but if the system will switch to opie itself the screen remains blank Oct 11 13:59:25 I'm sure this is a simple question: how do I start matchbox-dm on top of kdrive? The xserver is up and running, I set the DISPLAY system variable to the correct display but matchbox still says can't connect to display, I also tried setting the -display option with matchbox.. Oct 11 13:59:30 VoodooZ: yes it's working Oct 11 13:59:52 VoodooZ: I've told NAiL at least 3 times how to get it working, but it seems he won't listen Oct 11 14:00:00 Crofton: from root consdole, 'ifconfig' only returns 'lo', the local loopback. no eth0 in sight Oct 11 14:00:03 koen: really? Oct 11 14:00:11 bigd0g: 'ifconfig -a' Oct 11 14:00:13 VoodooZ: yes Oct 11 14:00:22 Perhaps I'll re-iterate to him Oct 11 14:00:37 'ifconfig -a' also only display the 'lo' Oct 11 14:00:38 ;) Oct 11 14:00:45 as a few people were starting to look into updating slugos's outdated glibc. Oct 11 14:00:48 VoodooZ: http://www.angstrom-distribution.org/repo/?action=details&pnm=libc6 <- has nptl Oct 11 14:00:55 Cool. Oct 11 14:01:06 you could also try glibc 2.5 :) Oct 11 14:01:09 Do you have a summary of your discussion with Nail? Oct 11 14:01:23 2.4 or 2.5 is fine as long as it's stable "enough"! :) Oct 11 14:01:24 yes 'look at the angstrom distro configs' Oct 11 14:01:36 cool. I'll pass it on to the guys... Oct 11 14:01:56 I'm kind of tired of dealing with linuxthreads un-holy signal support. Oct 11 14:02:12 I hope there's rwmutex support too. Oct 11 14:02:54 koen, are you running 2.5 yourself or is that a for suicidal maniacs? ;-) Oct 11 14:03:06 [lala_]: looks like I will need to build opie-image from .dev and check it Oct 11 14:03:41 <[lala_]> hrw|work: that would be cool. /sbin/getkey seems to missing as well, so i can not stop starting opie at boottime Oct 11 14:03:51 VoodooZ: I only built glibc 2.5, haven't tried it yet Oct 11 14:04:09 [lala_]: getkey was addon to busybox - probably no one ported it to newer busybox Oct 11 14:04:30 [lala_]: I suspect qte-mt problem Oct 11 14:04:39 koen, cool. I'll try stuff... Thanks. Oct 11 14:05:51 koen, do you think I can simply install the arm ipk file on top of my existing setup or am i crazy to even consider that? :) Oct 11 14:06:05 crazy :) Oct 11 14:06:13 I don't think slugos is using EABI yet Oct 11 14:06:51 koen: got info from Debian guys about eabi progress witAngstrom rootfs? Oct 11 14:07:26 koen: PR etc.. ;D Oct 11 14:07:26 Good Morning ;) Oct 11 14:08:09 hrw|work: not yet Oct 11 14:08:09 morning B_Lizzard Oct 11 14:08:09 I'm having a problem with ion, which I can't really solve Oct 11 14:08:09 hrw|work: we didn't reach consensus at OEDEM about the press release Oct 11 14:08:09 koen: I remember Oct 11 14:08:09 I do the normal "bitbake ion", and the job commences Oct 11 14:08:17 B_Lizzard: ion which is in OE is very old Oct 11 14:08:19 Ah? Oct 11 14:08:46 OE: ion3_0.0+ds20041104.bb Debian: 20060524-3 Oct 11 14:08:47 1.5 year Oct 11 14:08:58 B_Lizzard: I suggest work on adding newer one Oct 11 14:09:04 Even though I have all lua packages installed, configure still complains about it missing... Oct 11 14:09:06 I'll look into it Oct 11 14:09:14 thx Oct 11 14:09:22 I'll see if I can get it work without much modification Oct 11 14:09:26 Thanks a bunch Oct 11 14:09:56 B_Lizzard: if you will get problems then look at Debian patches etc Oct 11 14:10:06 Sure thing Oct 11 14:10:23 Just one thing, are those in the repos? Oct 11 14:10:27 Debian repos? Oct 11 14:10:42 B_Lizzard: http://packages.debian.org/ion3 Oct 11 14:10:52 Thanks a lot, hrw Oct 11 14:10:52 howdy Oct 11 14:11:12 hi siriusnova Oct 11 14:11:17 hrw|work - do you think i should try to compile konq 2005 ? Oct 11 14:11:29 would it work on oz 3.5.4 > Oct 11 14:11:30 ? Oct 11 14:12:06 oh and also just to let you know there is a bug somewhere in hostmap-modules in the upgrade feed for Collie, it makes ethernet show up as eth0 and causes iwconfig to segfault Oct 11 14:12:12 ;) Oct 11 14:12:17 bigd0g, can you try configuring locally, or looking at log_read to see why you do not have it? Oct 11 14:12:28 siriusnova: with some tweaking you will be able to run konq2005 in oz354 Oct 11 14:12:52 err i meant wireless wlan0 to show up as eth0 Oct 11 14:12:56 siriusnova: that collie issue... argh Oct 11 14:13:12 hrw|work - yeah there is a bug report hehe Oct 11 14:13:15 Crofton: networking support appears to be enabled in my kernel... posting my .config to a pastebin Oct 11 14:13:41 did you have networking when you booted the images I posted? Oct 11 14:13:46 hrw|work - a straight bitbake konqueror-embedded-2005 would work? Oct 11 14:13:57 or would i have to tweak something? Oct 11 14:14:14 Crofton: http://rafb.net/paste/results/09gMHZ62.html << see anything obvious that's wrong? Oct 11 14:14:15 siriusnova: lot more rather - cant help now Oct 11 14:14:34 okay np - would you mind posting a small guide on oz Oct 11 14:14:37 the wiki i mean? Oct 11 14:14:45 as in what is required so we can follow it Oct 11 14:14:48 'config_net_ethernet' is set to 'y' Oct 11 14:14:59 like whenever you are free Oct 11 14:17:20 siriusnova: it's a wiki, you can add it yourself Oct 11 14:17:47 koen - ill try ;) Oct 11 14:18:51 brb rebooting into linux Oct 11 14:22:49 VoodooZ: http://www.openembedded.org/repo/org.openembedded.dev/conf/distro/angstrom-2007.1.conf would be a better link Oct 11 14:24:00 Thanks mate. Oct 11 14:24:33 btw, I was on the phone for a while: what's EABI? Oct 11 14:24:51 VoodooZ: http://wiki.debian.org/ArmEabiPort Oct 11 14:25:17 thanks. Oct 11 14:26:37 hmm, how is embedded ubuntu going on ? Oct 11 14:26:54 what? Oct 11 14:27:17 WILL PEOPLE JUST STOP ADDING THE WORK EMBEDDED JUST BECAUSE ITS COOL Oct 11 14:27:30 XorA-embedded Oct 11 14:27:44 or Openoffice-embedded Oct 11 14:28:03 bigd0g, I have CONFIG_MII and CONFIG_SMC91X set to y not m Oct 11 14:28:04 XorA, yeah, it's the newest buzz word it seems. Oct 11 14:28:15 Crofton: ok, i'll give that a go Oct 11 14:28:20 They obviously have no clue what it means. Oct 11 14:28:21 XorA: embedded deep fried mars bars Oct 11 14:28:26 hehe Oct 11 14:28:27 Crofton: i'm not sure why they would've changed :-/ Oct 11 14:28:31 koen: still never braved one Oct 11 14:28:36 bigd0g, try modprobing it while you rebuild Oct 11 14:28:49 XorA: I learned from Tomas that it isn't an urban legends Oct 11 14:28:53 -s Oct 11 14:29:08 koen: he eats them? Oct 11 14:29:24 XorA: he has to get one to show to mallum iirc :) Oct 11 14:29:32 Crofton: ok Oct 11 14:29:52 deep fried twinkies Oct 11 14:30:06 Bloody Debian website taking all day Oct 11 14:33:11 In order to make OE patches out of debian ones, do I just download the whole .diff.gz, unpack it, and refer to it in the .bb file, or do I have to create seperate files for each patch that exists in the .diff? Oct 11 14:33:42 If you catch my drift Oct 11 14:33:45 ;) Oct 11 14:33:46 B_Lizzard: you can do the first, the better way would be to evaluate them Oct 11 14:34:00 Ah, thanks Oct 11 14:34:14 For now, I'll just do the first, just to evaluate Oct 11 14:34:29 I'll delete debian-specific parts later Oct 11 14:34:31 thanks Oct 11 14:39:53 Even the latest ion complains about lua missing Oct 11 14:40:09 XorA, wow, Indeed, I hear that ubuntu was planning a new project named "embedded ubuntu". here is it : https://wiki.ubuntu.com/EmbeddedUbuntu Oct 11 14:40:32 XorA, ;) Oct 11 14:40:37 I have lua, lua-native, lua-gtk2 installed on OE, and liblua-dev on my host machine installed Oct 11 14:40:49 bah Oct 11 14:41:07 I've seen "Embedded [gentoo|debian|foobarbaz]" pages like that for years Oct 11 14:41:12 leoncamel jesus Oct 11 14:41:13 call me when you actually deliver something Oct 11 14:41:18 cosmic yeah Oct 11 14:41:29 you could make embedded debian within OE :) Oct 11 14:41:35 we basically are Oct 11 14:41:51 well - you Oct 11 14:41:58 I haven't contributed for a while now Oct 11 14:42:28 CosmicPenguin, :). so, that is now a *news* now. Oct 11 14:42:58 hah Oct 11 14:44:13 CosmicPenguin, OH. s/first now/not . Oct 11 14:45:04 This is such a "what have you done for me lately" community :) Oct 11 14:46:37 I have a major issue to pick with all of you in here Oct 11 14:46:58 WHY IS NETHACK NOT IN THE REPOS AFTER SO MANY YEARS :P Oct 11 14:47:20 <[{(o_O)}]> Oct 11 14:49:00 B_Lizzard: ask andersee Oct 11 14:49:15 heh Oct 11 14:49:31 I bet this is a part of OE folklore, eh? Oct 11 14:49:31 who among us hasn't started a nethack .bb? Oct 11 14:49:39 ~seen andersee Oct 11 14:49:49 Embedded OpenEmbedded Oct 11 14:49:58 andersee was last seen on IRC in channel #uclibc, 4d 21h 32m 52s ago, saying: 'morning'. Oct 11 14:50:14 NdaEmbedded Oct 11 14:50:39 The ever-evading Nethack .bb? Oct 11 14:50:46 Why is it so hard to make? Oct 11 14:50:56 have you ever built nethack? Oct 11 14:50:58 its a msess Oct 11 14:51:01 I thought the code was fairly stable by now... Oct 11 14:51:03 Oh Oct 11 14:51:13 I've built worse Oct 11 14:51:18 hi ph5 Oct 11 14:51:19 CosmicPenguin: the nethack in gnome cvs is autotooled Oct 11 14:51:36 :). it should have a filter something like "s/[Ee]mbedded//". the word "embedded" would soon be filtered. ;) Oct 11 14:51:40 Try to build Vulture's Eye Oct 11 14:51:44 * chouimat just looked at nethack ebuild in gentoo and he's crying Oct 11 14:51:46 koen: yeah - think andersee was part of that - that dude love nethack Oct 11 14:52:05 good evening Oct 11 14:52:14 morning Oct 11 14:52:14 Makefile editing, makefile here, makefile there...It's as if they've never heard of ./configure Oct 11 14:52:23 g'mornin' Oct 11 14:53:15 Well, those of you with ARM CPUs can use old debian packages Oct 11 14:53:21 anybody working on powerpc platform? Oct 11 14:53:38 03koen 07org.oe.dev * rf45c2bc2... 10/ (1 packages/linux/handhelds-pxa-2.6.inc): hh-pxa.inc: use hx4700 machine from now on Oct 11 14:53:54 Crofton: yep... it was those to 'm' settings. should've been 'y' Oct 11 14:53:57 B_Lizzard: not every program is new enough for these new fangled autotools ideas :-) Oct 11 14:53:58 s/to/two Oct 11 14:54:06 anyone here run edgy and use oe on it? Oct 11 14:54:53 Autotools is for pussies Oct 11 14:55:03 Real users write all commands by hand Oct 11 14:55:29 ~hail koen for giving us MACHINE="hx4700" back Oct 11 14:55:31 * ibot bows down to koen for giving us MACHINE="hx4700" back and chants, "I'M NOT WORTHY!!" Oct 11 14:56:23 pH5, Hi! Oct 11 14:56:44 B_Lizzard: in brainfuck Oct 11 14:56:57 :) Oct 11 14:57:03 pH5, do you think that hx4700 corgi_bl after you last tweaking is ok for commit? Oct 11 15:00:50 psokolovsky: with anton's latest fixes, yes. I'm going to wait for todd to look at it. Oct 11 15:01:46 pH5, ok. Anton wanted the same. Just keep in mind he doesn't have commit. Oct 11 15:03:46 psokolovsky, pH5: FWIW, I did see the changes to corgi and I think I was basically ok with them. You still left corgi_bl.h as pxa specific and I'm not sure that's the case though Oct 11 15:05:10 RP: looks like sooner or later corgi_bl will be common_pda_bl Oct 11 15:05:12 ;d Oct 11 15:05:15 hi rp Oct 11 15:05:33 pH5: you should like the next commit Oct 11 15:05:34 RP: but it only show how good it is ;) Oct 11 15:05:44 hrw|work: :) Oct 11 15:05:49 hi woglinde Oct 11 15:06:04 03koen 07org.oe.dev * ra3f6da3e... 10/ (1 conf/machine/hx4700.conf): hx4700.conf: drop task-bootstrap support, add task-base support Oct 11 15:06:56 koen: acx in machine-rrecommends or will we add new feature? Oct 11 15:08:20 RP, thanks, we'll note that Oct 11 15:08:35 RP: you're right, I missed the Kconfig change. Should I make a patch with only the corgi_bl changes and send it your way? Oct 11 15:08:37 hrw|work: in machine_rrecommends Oct 11 15:08:45 koen: nice Oct 11 15:09:05 pH5: The header is still in the asm/arch as well though? Oct 11 15:09:20 pH5: it needs some improvement, but the start is there Oct 11 15:09:29 hm Oct 11 15:09:30 koen: I would prefer OSS emu in distro Oct 11 15:09:41 maybee its time to work a little bit on the simpad Oct 11 15:09:47 hrw|work: right, as I said :) Oct 11 15:10:05 03koen 07org.oe.dev * rf1f7a72c... 10/ (1 conf/machine/h2200.conf): h2200: convert to task-base Oct 11 15:10:09 koen: hidp as part of bluetooth Oct 11 15:10:09 heh Oct 11 15:10:15 RP: what would be the right place for the header - include/video? Oct 11 15:10:20 Looks like that ol' 512 byte MMC patch is finally making headway Oct 11 15:10:31 CosmicPenguin: it is in 2.6.18 Oct 11 15:10:53 CosmicPenguin: Excellent :) Oct 11 15:10:59 And Greg KH posted an update on the LKML too Oct 11 15:11:07 CosmicPenguin: One down, several MMC issues to go :) Oct 11 15:11:10 heh Oct 11 15:11:19 * CosmicPenguin mutters something about a ex-maintainer under his breath Oct 11 15:11:20 koen: pcmcia-manager is distro not machine Oct 11 15:11:31 pH5: Probably even just in include/linux :-/ Oct 11 15:11:38 Has rmk dropped mmc? Oct 11 15:11:48 yeah Oct 11 15:12:00 as well as serial. Interesting... Oct 11 15:12:04 MACHINE_EXTRA__RRECOMMENDS is wrong too... Oct 11 15:12:14 RP: soon rmk will drop arm? Oct 11 15:12:38 koen: just in case, conversion of h4000 to task-base is on my todo ;-). feel free to leave it for me, I'd like to look closer at that stuff (just keep wrapping up other loose ends for now...) Oct 11 15:12:59 psokolovsky: want h4000 changed? Oct 11 15:14:21 hrw|work, ideally, would like to learn how to do that myself ;-) but if waiting would cause problems with building snapshots, and someone else can do that, it's ok of course Oct 11 15:15:34 task-bootstrap* should die as soon as possible Oct 11 15:15:57 koen yes Oct 11 15:18:45 koen: h2200 comming too Oct 11 15:19:48 hrw|work: cool, I should see if the pcmcia problem is solved with the udev rule properly Oct 11 15:20:28 I got a feeling that most of task-base stuff is by me... argh Oct 11 15:20:58 03hrw 07org.oe.dev * r689904fb... 10/ (1 conf/machine/h4000.conf): h4000: converted to task-base Oct 11 15:21:05 03hrw 07org.oe.dev * r15420257... 10/ (1 conf/machine/h2200.conf): Oct 11 15:21:05 h2200: cleaned task-base switch Oct 11 15:21:05 - do not depend on udev - kernel26 feature take care of it Oct 11 15:21:05 - do not list usb gadget modules - feature them Oct 11 15:21:05 - fixed typo: s/__/_ in MACHINE_EXTRA_RRECOMMENDS Oct 11 15:22:25 hmm.. we got oss emu by default Oct 11 15:23:58 03koen 07org.oe.dev * r9620e56c... 10/ (1 conf/machine/h2200.conf conf/machine/hx4700.conf): Oct 11 15:23:58 hx4700, h2200: fix typo Oct 11 15:23:58 * hidp and oss still need to get split out Oct 11 15:25:14 koen: s/to get split out/to get removed Oct 11 15:26:37 17:45 hrw@work:machine$ mtn dif .|diffstat h2200.conf | 1 - h4000.conf | 2 -- hx2000.conf | 3 +-- hx4700.conf | 5 +---- ipaq-pxa270.conf | 3 +-- 5 files changed, 3 insertions(+), 11 deletions(-) Oct 11 15:26:53 i spend my day trying to compile gnash ;( , some one can answer to my question on AGG packaging http://bugs.openembedded.org/show_bug.cgi?id=1464 to have a positive end of day ? Oct 11 15:28:44 koen: beware next commit Oct 11 15:29:27 03hrw 07org.oe.dev * r74c29ae4... 10/ (6 files in 2 dirs): h2200/h4000/hx2000/hx4700/ipaq-pxa270: drop OSS Emu (featured by alsa) and HIDP (featured by bluetooth) modules Oct 11 15:30:05 Upgrading libc6 on root from 2.4-r10 to 2.5-r0... Oct 11 15:30:09 * koen crosses fingers Oct 11 15:30:18 hehe Oct 11 15:30:24 Genesis: s/bbclass/recipe Oct 11 15:30:26 * XorA waits for a koen disconnect Oct 11 15:30:27 hm so maintainer field is really gone Oct 11 15:31:36 Genesis: run recipe through contrib/stylize script, remove maintainer (provide patch to /MAINTAINER), include AGG License at end of recipe Oct 11 15:31:57 hrw|work, : where should i install the example ? Oct 11 15:32:18 Genesis: if it has "agg-2.4" in S then which version it is? 2.4+cvsSRCDATE or 0.0+cvsSRCDATE? Oct 11 15:32:45 2.4 is the directory on the cvs server Oct 11 15:32:48 Genesis: PACKAGES += "agg-examples" FILES_agg-examples = "path/to/examples/*" Oct 11 15:33:08 oki i'll try to do that tommorow Oct 11 15:33:13 thx a lot for help Oct 11 15:33:56 Genesis: May 8, 2006 AGG v2.4 is released! - so its 2.4 not 0.0 Oct 11 15:34:21 HO oki Oct 11 15:34:23 Genesis: it does not need any dependencies listed? Oct 11 15:34:29 no Oct 11 15:34:33 it's AGG way Oct 11 15:34:37 no dependancy at all ): Oct 11 15:34:40 ;) Oct 11 15:35:12 ok Oct 11 15:35:24 Genesis: add AUTHOR Oct 11 15:36:07 koen@hx4700:~$ uptime Oct 11 15:36:07 17:53:07 up 31 days, 19:11, 3 users, load average: 1.43, 1.46, 0.84 Oct 11 15:36:08 a lot of changement lol Oct 11 15:36:45 Genesis: add 2.4 version AND cvs version Oct 11 15:37:03 Genesis: but cvs one only if you need it. Oct 11 15:37:26 Genesis: generally try to use archives rather then SCM if upstream provide such ones Oct 11 15:37:41 May 8 is a bit old Oct 11 15:37:50 but i'll provide both Oct 11 15:37:53 thx Oct 11 15:38:33 hm whats BOOTSTRAP_EXTRA_RDEPENDS now? Oct 11 15:38:33 koen: will you rebuild images for h2200/hx4700/h4000/hx2200 to test task-base stuff? Oct 11 15:38:42 woglinde: nothing - has to die Oct 11 15:38:54 BOOTSTRAP* has to die Oct 11 15:39:03 yes Oct 11 15:39:08 thats why I am ask Oct 11 15:39:10 bye Oct 11 15:39:34 MACHINE_FEATURES ? Oct 11 15:39:47 what fonts are commended to build in OE for matchbox/kdrive? Oct 11 15:39:48 evil must die Oct 11 15:39:49 woglinde: what does machine has Oct 11 15:40:01 Gerrath: I propose ttf-dejavu Oct 11 15:40:08 looking at the simpad confs Oct 11 15:40:15 woglinde: I will write documentation on it during weekend Oct 11 15:40:45 hrw|work, thanks. Oct 11 15:42:18 when i 'bitbake python', it says it drops things into /usr/lib/python2.4 ... but when i look in the rootfs in 'tmp' i don't see it ... am i missing something? Oct 11 15:43:07 hm okay Oct 11 15:43:12 zaurus files help Oct 11 15:43:16 will convert simpad ones Oct 11 15:44:32 w00t - psplash on my hx4700 Oct 11 15:44:49 * koen makes a note about changing the logo Oct 11 15:45:57 hmm.. Simpad... good point, we should create Simpad images :-) Oct 11 15:46:09 florian aehm Oct 11 15:46:39 oh, i see. 'bitbake python' puts the necessary things in 'tmp\staging' Oct 11 15:46:44 florian but what images? Oct 11 15:46:52 opensimpad gone Oct 11 15:47:05 angstroem dont ready becaus of eabi Oct 11 15:47:12 how, then, do i get the necessary libraries 'installed' into the rootfs image? Oct 11 15:47:18 openzaurus is only one which works now Oct 11 15:47:24 koen: I tried figuring out how to select 2.4 for slugos, but I didn't get it. Oct 11 15:48:34 woglinde: yes, oz would be fine Oct 11 15:48:39 cu guys Oct 11 15:48:53 later, hrw|work Oct 11 15:49:15 woglinde: use .oz354x, create opensimpad images for 3.5.4.1 ;) Oct 11 15:49:38 woglinde: just like koen did with h2200, hx4700, netbook and florian_kc with h3900 Oct 11 15:49:40 ;D Oct 11 15:50:35 florian maybe gpe works Oct 11 15:50:35 RP: glibc 2.5 doesn't seem to have the shm troubles Oct 11 15:55:16 hm why is PREFERRED_PROVIDER_xserver = "xserver-kdrive" in some machine.confs? Oct 11 15:55:22 should that be in distro? Oct 11 15:56:12 or GPE_EXTRA_INSTALL += "gaim Oct 11 15:57:20 woglinde: the machine knows which Xserver works Oct 11 15:57:32 woglinde: the distro can choose to ignore it Oct 11 15:58:12 hm I would prefer to handle it strict Oct 11 15:58:35 no userspace things in machine.confs only the image depend on it Oct 11 15:58:38 how would the distro know that a c700 needs xw100? Oct 11 15:58:52 or that the netbook needs Xepson? Oct 11 15:59:00 hm Oct 11 15:59:50 hm yes as I said Oct 11 16:00:00 only things needed for the machine to run Oct 11 16:00:50 woglinde: it only says '*if* you want X, use this one' Oct 11 16:00:59 jupp Oct 11 16:01:08 but what about the gaim think? Oct 11 16:01:11 thing Oct 11 16:01:15 that should die Oct 11 16:01:24 or others GPE_BLAS Oct 11 16:01:51 question: so i've built python and python-gtk using bitbake. now what do i do to get them onto my board? Oct 11 16:02:33 hm Oct 11 16:02:34 they seem to have placed themselves in the 'staging' folder... i don't see them listed in the rootfs bitbake makes Oct 11 16:03:21 koen how is montone handle umask thinks? Oct 11 16:03:37 netbook.conf has execution bit set Oct 11 16:03:42 on my side Oct 11 16:03:45 woglinde: iirc via attrs, but I'm not sure Oct 11 16:05:18 03koen 07org.oe.dev * r9eed2ca2... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: use glibc 2.5 to (seemingly) solve the shm problems Oct 11 16:06:33 ^ anyone have an idea about my question above? Oct 11 16:08:43 bigd0g: are there any ipkg's in the distribution dir? Oct 11 16:09:33 hvontres|poodle: a 'distribution' dir? i'm not seeing that anywhere... Oct 11 16:09:40 bigd0g, look in tmp/deploy/ipk Oct 11 16:10:04 Crofton: thanks...I was suffering from bitrot in the brain... Oct 11 16:10:14 :) Oct 11 16:10:21 yeah, sorry ... sometimes i can just get lost ;) Oct 11 16:10:23 anyway... yes Oct 11 16:10:27 i have the ipk Oct 11 16:10:31 what do i do with them? ;) Oct 11 16:10:41 get them on the osk and install them Oct 11 16:10:51 bigd0g: can you use ipkg on your target? Oct 11 16:10:58 I don't know how that works though Oct 11 16:11:01 it should have ipkg Oct 11 16:11:16 hvontres|poodle: is there an easy way to check? Oct 11 16:11:27 type ipkg? Oct 11 16:11:33 bigd0g: do you have a console???? Oct 11 16:11:38 well, on the osk that yields nothing Oct 11 16:11:42 my osk is booting now Oct 11 16:11:53 It used to have ipkg ... Oct 11 16:11:57 yeah, i have a console... i've never used ipkg's before tho... so Oct 11 16:12:11 ip and hit the tab key .... Oct 11 16:12:21 nothing pops up Oct 11 16:12:36 hmm Oct 11 16:12:59 ran a few finds on normal locations, not seeing 'ipkg' Oct 11 16:13:02 looks like we need to add ipkg support Oct 11 16:13:11 * bigd0g claps ? Oct 11 16:13:17 bigd0g: I think you can set your local.conf to generate .tar archives as well.. Oct 11 16:13:29 how do we tell bootstrap-image to include ipkg support? Oct 11 16:13:41 hvontres|poodle: yep, i could ... but how would that help? :-/ Oct 11 16:14:11 oh... Oct 11 16:14:21 you mean set .tar archives for build stuff Oct 11 16:14:23 I am off too lunch, I'll look into modifying the openomap distro to include ipkg Oct 11 16:14:35 ok. thanks, Crofton Oct 11 16:14:41 you caould build a tarbal of ipkg and get ipkg support that way Oct 11 16:14:55 long term I want ipkg support Oct 11 16:14:57 bigd0g: yes...and then just untar on the target :) Oct 11 16:15:01 anyone we are leaving Oct 11 16:15:45 hvontres|poodle: yeah, sorry... i thought you were referring to filesystem type since i just got through that debaucle on an unrelated issue Oct 11 16:16:09 Crofton: try looking at the Zaurus bootstrap image .bb's. I think they just add it as either a DEPEND or an RDEPEND Oct 11 16:17:41 hvontres|poodle: are those Zaurus bootstrap-image .bb's in the org.oe.dev tree or the zaurus tree ? Oct 11 16:20:44 hrm... they must be in the other tree. the zaurus package files in org.oe.dev don't reference bootstrap-image Oct 11 16:23:45 hms Oct 11 16:24:20 handheld-common has boostrap too Oct 11 16:25:31 bigd0g: I think they are restructureing things. This might be the way to go: http://www.openembedded.org/repo/org.openembedded.dev/packages/images/essential-image.bb Oct 11 16:25:44 * bigd0g clicks Oct 11 16:26:10 'inherit image_ipk' ... ah, isee Oct 11 16:26:14 bigd0g: but beyond this point I am afraid I can't help you... Oct 11 16:26:31 * hvontres|poodle is still a OE noob... Oct 11 16:26:33 no worries. i appreciate the help thus far Oct 11 16:26:54 koen or hrw would probably be your best bet.. Oct 11 16:27:47 yeah, it's all frustrating. been working for 10-weeks off and on to get a build that i need... finally get, than i have to jump through hoops to get the frame buffer support i needed, then ethernet failed... and now getting software onto the board is turning into a project Oct 11 16:27:47 heh Oct 11 16:28:13 it's one of those 2-steps forward, 1-step back projects... i guess Oct 11 16:29:53 I guess it helps when you can stand on the soulders of the others... builing my first bootstrap for poodle was easy...well once I had my machine dialed in Oct 11 16:30:35 the OE community's been a huge help, for sure Oct 11 16:31:02 * bigd0g peeks at OpenZaurus bb files Oct 11 16:37:42 * koen is mightily happy with glibc 2.5 Oct 11 16:37:58 it will blow up for sure, but it looks to be functioning Oct 11 16:38:19 That was released a week ago or something like that, eh? Oct 11 16:38:23 Nice Oct 11 16:38:27 koen hm how much space is wasted now? Oct 11 16:40:29 1352 morgue/libc6_2.4-r10_armv5te.ipk Oct 11 16:40:38 1364 libc6_2.5-r0_armv5te.ipk Oct 11 16:40:48 so it grew 12kB Oct 11 16:42:22 hm I will try a simpadimage Oct 11 16:42:24 now Oct 11 16:42:52 bitbake essential-image? Oct 11 16:43:22 where is task-base defined Oct 11 16:43:35 Crofton: packages/tasks ? Oct 11 16:43:39 just a wild guess Oct 11 16:44:24 no Oct 11 16:44:53 koen@bitbake:~/OE/monotone/org.openembedded.dev$ ls packages/tasks/ | grep base Oct 11 16:44:53 task-base.bb Oct 11 16:47:15 crap Oct 11 16:47:24 I was in the wrong directory Oct 11 16:49:28 how do I add ipk support Oct 11 16:50:15 inherit image_ipk Oct 11 16:50:15 inherit package_ipk Oct 11 16:50:22 both? Oct 11 16:50:57 openompa inherits package_ipk and debian Oct 11 16:51:23 INHERIT += "package_ipk debian multimachine" Oct 11 16:51:47 hmm Oct 11 16:52:06 just inheriting package_ipk won't get ipk stuff on the fs image? Oct 11 16:52:25 what does debian and multimachine do for me? Oct 11 16:53:14 debian gives you shared lib renaming Oct 11 16:53:29 multimachine give you a way to build for multiple machines from the same tree Oct 11 16:53:56 Crofton: add ipkg and ipkg-collateral to your image Oct 11 16:54:17 I am already inheriting debian and package_ipk, so I am confused why the ipk toools didn't get onto the image Oct 11 16:54:20 ah Oct 11 16:54:22 koen any news about eabi on arm4? Oct 11 16:54:37 so I need to explicity add them the image? Oct 11 16:54:45 woglinde: gcc is still broken afaik Oct 11 16:55:22 should that be a distro feature handled by task-base? Oct 11 16:56:54 it could be Oct 11 16:57:10 but should it be? Oct 11 16:57:30 it seems like the way to do it? Oct 11 16:57:37 does angstrom use ipk? Oct 11 17:00:20 I find this NOTE a bit awkward: "NOTE: no buildable providers for foobar-image". What exactly does it mean? Shouldn't it read "unbuildable providers for foobar-image" instead or do I miss something? Oct 11 17:00:27 and if angstrom uses ipk, how does it get into bootstrap-image? Oct 11 17:00:48 likewise: maybe it really is FUBAR? Oct 11 17:08:15 koen, who is the task-base guru? Oct 11 17:08:15 RP and hrw|gone Oct 11 17:08:16 how does ipk get into angstrom machines? Oct 11 17:09:41 Crofton: http://www.openembedded.org/repo/org.openembedded.dev/packages/angstrom/task-angstrom.bb Oct 11 17:09:50 ah Oct 11 17:09:56 and http://www.openembedded.org/repo/org.openembedded.dev/packages/angstrom/angstrom-bootstrap-image.bb Oct 11 17:10:11 note that's being switched to task-base Oct 11 17:10:11 so you build that as the base iage for angstrom machines, not bootstrap-image Oct 11 17:10:13 ok Oct 11 17:10:39 I will talk to RP and hrw|gone about packaging formats in task-base Oct 11 17:10:54 I like saying a distro is ipk, rpm, deb or whatever Oct 11 17:19:01 * Crofton detests omniorb Oct 11 17:19:14 well the build structure Oct 11 17:28:31 KRISTOFFER! Oct 11 17:32:31 hm Oct 11 17:32:39 glibc-2.5 is building Oct 11 17:32:44 *fingercross* Oct 11 17:32:50 I have matchbox running but everything takes a long time to load: couple of minutes. After something loads it runs at a good speed, I have run strace on the apps but I do not see anything that stands out.. Oct 11 17:42:35 03koen 07org.oe.dev * r389fcce8... 10/ (6 files in 3 dirs): Oct 11 17:42:35 angstrom-2007.1.conf: add some extra utils to task-base Oct 11 17:42:35 angstrom-bootstrap-image: don't depend on task-angstrom Oct 11 17:42:35 angstrom-gpe-image: don't depend on task-angstrom Oct 11 17:42:35 task-angstrom-x11: don't depend on task-angstrom Oct 11 17:42:37 task-angstrom: die, you have been replaced by task-base Oct 11 17:51:51 aaargh Oct 11 17:51:56 ~lart spurious diffs Oct 11 17:51:56 * ibot pours gasoline all over spurious diffs, ignites the fire, and then enjoys some toasty marshmallows with the glorious blaze Oct 11 17:51:57 koen? Oct 11 18:00:44 is it normal for matchbox to take 1-2 minutes to load on a PXA270 system running at 500 MHz? Oct 11 18:00:58 'matchbox' is a bit vague Oct 11 18:01:14 do you mean matchbox-wm, matchbox-desktop, matchbox-panel? Oct 11 18:01:19 koen, yes, sorry. Oct 11 18:01:43 koen, it runs well once its loaded. Oct 11 18:04:34 what it is the typical load time on other pxa OE machines? Oct 11 18:05:32 a few seconds Oct 11 18:06:32 koen, something must not be clean in my setup yet then.. Oct 11 18:06:41 nothing obvious when I run strace.. Oct 11 18:07:45 I'm trying create a .bb package for compiling pocketsphinx (which requires sphinxbase) for the NSLU2. It uses the typical configure mecasnism. Any good barebone package to recommend for cloning. Oct 11 18:08:14 pocketsphinx being a light embedded version of the Sphinx speech recognition tool. Oct 11 18:09:04 VoodooZ: http://www.openembedded.org/user-manual&dpage=ch02s04 Oct 11 18:10:00 what is the best way to write a staging method these days? Oct 11 18:10:24 Crofton: vi :) Oct 11 18:10:55 I am at war with omniorb right now Oct 11 18:12:42 hi Oct 11 18:13:32 hrm. has anyone had success making/restoring a NAND backup, on Cxxx0? Oct 11 18:13:46 koen, thanks. It looks good. Oct 11 18:15:22 I've seen oesf posts that suggest some people have had success; but the sharp menu reports corrupted image on restore of every combination I've tried Oct 11 18:16:06 is there something special I'm missing? Oct 11 18:20:00 hey when im building for openzaurus 3.5.4 do i use 3.5.4.1 in my conf or 3.5.4.2 ? Oct 11 18:24:03 * XorA|gone persuades mplayer to open a pxafb_overlay Oct 11 18:25:15 XorA|gone: cool Oct 11 18:25:22 XorA|gone: did you get my mail about glibc? Oct 11 18:25:28 koen: yes, good news Oct 11 18:25:44 koen: now I just need to teach mplayer to write to the overlay :-) Oct 11 18:26:08 XorA|gone: you taught it to keep the fb open between setup and playback? Oct 11 18:26:21 Heard that 2.5 was released :D any goodies? Oct 11 18:26:44 Kristoffer: it *seems* to fix the shm problems people were seeing Oct 11 18:27:46 koen: yes Oct 11 18:27:59 koen: Im doing a custom driver for pxafb Oct 11 18:28:25 koen: I dont think the fbdev/fbdev2 drivers will ever work as they are designed to ask for current mode, then attempt to play in that mode Oct 11 18:28:33 right Oct 11 18:28:50 we had a discussion about pxa overlay and sram last weekend Oct 11 18:29:07 :D Oct 11 18:29:41 mickey 'eyecandy on the a780' lauer is the latest victim of the overlay hype ;) Oct 11 18:31:13 koen: the best method would be Xv support in kdrive :-) Oct 11 18:31:50 XorA|gone: tf has written a iwmmxt powered xserver Oct 11 18:32:04 XorA|gone: so iwmmxt + overlay = neat Oct 11 18:32:54 anyone ? what do i use for DISTRO if im using openzaurus-3.5.4 Oct 11 18:33:27 i tried openzaurus and openzaurus-3.5.4.1 but it says "unsupported version" although the distro shows up as a conf file in the distro directory Oct 11 18:33:39 i mean openzaurus-3.5.4 instead of openzaurus-3.5.4.1 Oct 11 18:34:57 koen: there is something wierd about overlay :-( Oct 11 18:36:38 XorA|gone: we also want to abuse hardware cursor for ts calibration :) Oct 11 18:37:26 koen: overlay is not turning up where and size I expect :-( Oct 11 18:38:15 anyone know? :( Oct 11 18:39:27 siriusnova: oz 3.5.4.1 is only avaiable in oz354x branch Oct 11 18:39:50 XorA - i know im using the oz354x branch Oct 11 18:40:02 im using openzaurus 3.5.4 Oct 11 18:40:04 on my collie Oct 11 18:40:20 but openzaurus-3.5.4 and just openzaurus give me a warning about "unsupported distro" Oct 11 18:40:23 XorA|gone: hwo are you liking the terrier? Oct 11 18:40:46 koen: machine is nice, kernel seems a little iffy Oct 11 18:41:49 koen, thanks for the help I'll report back later. Oct 11 18:42:22 and now something is trying to destroy the microdrive Oct 11 18:43:49 * koen looks at 'readahead' Oct 11 18:44:52 Xfbdev is also unstable as hell Oct 11 18:45:09 right Oct 11 18:45:51 how "good" is savannah.nongnu.org? Oct 11 18:46:34 * XorA|gone has to magic sysreq in the end Oct 11 18:47:35 omniorb stages a butt load of stuff ..... Oct 11 18:47:53 I really do not want to do it by hand, but it looks like I must ... Oct 11 18:49:41 Crofton: good in the sense of what? Oct 11 18:49:54 works, useful .... Oct 11 18:50:09 are there better places to host projects Oct 11 18:50:15 Crofton: works, useful, but not much more than sourceforge.ne Oct 11 18:50:23 +t Oct 11 18:51:29 Crofton: there is also http://www.berlios.de/ Oct 11 18:51:51 I'd forgot about them Oct 11 18:51:59 also the google stuff Oct 11 18:52:12 Crofton: I'ld stick with sf.net or savannah though... Oct 11 18:52:27 k Oct 11 18:54:54 Does anyone work with gdbserver on their target? Oct 11 18:55:08 likewise not for ages Oct 11 18:55:21 likewise: I run full gdb on my targets nowadays Oct 11 18:55:43 ahah, I see, fbdev layer doesnt seem to take account of halving the y pixels in YUV modes Oct 11 18:55:48 because I recently noticed the .debug directories and wonder how I should gdbserver to understand all this. Oct 11 18:57:34 likewise: the binaries say "my debug symols are in " Oct 11 18:57:45 likewise: and gdb(server) understand that Oct 11 18:57:55 gdbserver as well? nice! Oct 11 18:58:26 koen: Hmm, but I think gdb on the dev host only needs the .debug stuff locally. Oct 11 18:58:42 koen: anyway, I'll go play with it a little. Oct 11 18:59:05 koen: submitted squashfs tools to bugzilla, fyi. Oct 11 18:59:16 I noticed that Oct 11 19:03:39 ok, wierd shit, scale to 480x640 which should fill the screen in RGB and overlay doesnt :-( Oct 11 19:04:18 XorA|gone: but is it *fast* weird shit? Oct 11 19:04:24 lol Oct 11 19:05:18 koen: no as Im running in RGB for now, there will be no speed increase, but I am stumped on why it doesnt display anything Oct 11 19:07:35 right Oct 11 19:07:35 left Oct 11 19:07:35 oe_libinstall is only useful for installing .la files? Oct 11 19:07:50 what if I ahve the .so and the .a Oct 11 19:07:50 Has anyone built gcc for his device? Oct 11 19:08:04 B_Lizzard: yes, 'bitbake gcc' Oct 11 19:08:06 gcc seems to be missing a file in mine Oct 11 19:08:21 that makes installation fail Oct 11 19:08:23 the .c code for it exists, though Oct 11 19:09:01 It configures and compiles ok, but when installing, it complains about a missing file and fails Oct 11 19:09:16 the file is called "fix-header" Oct 11 19:10:03 Since it might be a shell script, I thought if someone would be as kind as to look for it and give it to me, if it really is a shell script Oct 11 19:10:42 It's in here ../tmp/work/sh3-linux/gcc-4.1.1-r6/image/usr/libexec/gcc/sh3-linux/4.1.1/install-tools/fix-header Oct 11 19:10:55 Or something of the like Oct 11 19:17:25 Can I run a read-only filesystem like cramfs or squashfs from FlashROM instead of RAM? Oct 11 19:18:07 * XorA|gone gives up Oct 11 19:18:48 hrmm Oct 11 19:19:04 likewise: yes Oct 11 19:19:29 likewise: you have to make sure everything has been setup before making it RO Oct 11 19:21:02 koen: ok, yes, the image is OK. But I cannot find how I can make the kernel use the fs image from FlashROM rather than first copying the (compressed) image from FlashROM to RAM, then decompressing individual files when they are referred. Oct 11 19:21:30 koen: earlier victim report: http://www.mail-archive.com/linux-arm@vger.rutgers.edu/msg03974.html Oct 11 19:23:01 likewise: iirc sharp uses cramfs or something like that for / on the zaurus lines Oct 11 19:25:43 koen, it seems I'm getting a ton of kernel: Alignment traps with matchbox and apps loaded on it, is that normal on the PXA systems? Oct 11 19:26:12 alignment traps indicate bad code Oct 11 19:26:47 koen, Has bootstrap rdepends changed any? Added some programs but they never turn up inside built opie-image Oct 11 19:27:03 bootstrap is deprecated Oct 11 19:27:31 what to use instead of bootstrap_extra_rdepends? Oct 11 19:27:55 the appropriate replacements in task-base Oct 11 19:39:22 koen, If pocketsphinx requires a shared (with other spinx variations) tarball called sphinxbase, Oct 11 19:39:40 do you recommend I create a seperate .bb package or bundle them together under a single one. Oct 11 19:40:08 koen, the alignment traps is, I think, why matchbox is taking so long to load.. although if it is bad code I would think all pxa (xscale) systems in OE would have the same issus. Oct 11 19:40:19 The reason I'm asking is that they need to be compiled together. Oct 11 19:40:40 VoodooZ: a single recipe can produce multiple packages Oct 11 19:40:58 ummm. Oct 11 19:41:25 I'll try to figure out how to do it... Oct 11 19:42:13 morning Oct 11 19:42:26 gm Oct 11 19:42:39 hrw, you are the task-base guru? Oct 11 19:42:52 Crofton: looks like I'm person which hacked it most Oct 11 19:42:59 Crofton: lack of ipkg on device? Oct 11 19:43:02 ah, an expert :) Oct 11 19:43:08 yeah Oct 11 19:43:46 Crofton: DISTRO_FEATURE = 'ipk' can be solution Oct 11 19:43:57 but add it to package_ipk? rootfs_ipk? DISTRO configs? Oct 11 19:44:05 does that work? Oct 11 19:44:10 heh Oct 11 19:44:23 Crofton: not yet Oct 11 19:44:41 when might it work? Oct 11 19:44:49 Crofton: on one machine I added it into rootfs_ipk but this can be not good for ro filesystem.. Oct 11 19:45:31 maybe I should go ahead and hack it into openomap for now Oct 11 19:46:27 hi, how long does it usually take to get oesf forum account approval? Oct 11 19:46:36 um Oct 11 19:46:41 pushing Oct 11 19:46:46 oesf just stole the name from this channel Oct 11 19:46:46 ssvb: oe != oesf Oct 11 19:46:49 ~oesf Oct 11 19:46:50 :) Oct 11 19:46:58 extra, extra, read all about it, oesf is newer name of ZUG - it provide forum for Zaurus/Ipaq/Simpad/etc users (see zug). It has nothing common with OpenEmbedded (except first two words), or http://www.oesf.org/ Oct 11 19:48:02 special forces Oct 11 19:48:19 also OESF = OpenEmbedded Special Forces Oct 11 19:49:03 eh, been looking at task-base stuff. I just want a suitable replacement for bootstrap-extra-rdepends so that extra programs can be included in opie-image Oct 11 19:49:29 Kristoffer: what kind of them? Oct 11 19:49:50 hrw, konqueror Oct 11 19:50:01 03hrw 07org.oe.dev * re7879128... 10/ (1 classes/rootfs_ipk.bbclass): Oct 11 19:50:01 rootfs_ipk.bbclass: Add ipkg packages into rootfs to unbreak task-base Oct 11 19:50:01 - When we used task-base then ipkg/ipkg-link/ipkg-collateral was not installed Oct 11 19:50:01 into image. Oct 11 19:50:01 - Maybe it is not best sollution but it works and it duplicate solution which Oct 11 19:50:01 was added years ago. Oct 11 19:50:12 Kristoffer: its a matter of root-flash-size rather Oct 11 19:50:32 Crofton: thats solution for you Oct 11 19:50:43 ok, so rebuild boostrap-image? Oct 11 19:50:52 hrw, Got no special size restrictions Oct 11 19:50:54 Crofton: rebuild task-base and bootstrap-image Oct 11 19:50:58 thanks Oct 11 19:51:07 Crofton: package types are a setting for the distro and totally separate from task-base Oct 11 19:51:15 yeah Oct 11 19:51:22 Kristoffer: we do not want konqueror on collie, poodle, tosa as they have small flash size Oct 11 19:51:27 we get the ipks, just no tools to install on OSK :) Oct 11 19:52:05 Crofton: how goes work on working git kernel recipe? I have local changes for omap kernel building Oct 11 19:52:07 hrw, I know, But before I used bootstrap-extra-rdepends in my distro file, to make sure some programs were included by default. Now that thats deprecated, (and release is in like 10hours) I want something similiar Oct 11 19:53:01 Kristoffer: DISTRO_EXTRA_RDEPENDS = "konqueror-embedded" Oct 11 19:53:10 hrw, big thx man Oct 11 19:53:24 Kristoffer: or create opie-kristoffer-image Oct 11 19:53:50 hrw, we are building kernels Oct 11 19:53:54 that work Oct 11 19:54:10 what are your local changes Oct 11 19:54:25 hrw, using .dev tree for development. I've not set any local database yet :D so that would mean adding it to the .dev tree Oct 11 19:55:55 RP, btw been using my spitz for like 2 weeks now. Generally satisfied with it. Works well. The only problems Im having is that the mediaplayer doesnt react to sound changes (+/-) So I need to use alsamixer to set good volume Oct 11 19:56:22 03koen 07org.oe.dev * rb3f5af29... 10/ (1 packages/tasks/task-base.bb): task-base: bump PR Oct 11 19:57:09 Crofton: 1. move pref_ver of kernel from machine to distro Oct 11 19:57:26 hmmm Oct 11 19:57:28 why? Oct 11 19:57:40 I though kernel followed machine? Oct 11 19:57:59 Crofton: machine set which kernel. distro sets which version of that kernel Oct 11 19:58:08 yeah Oct 11 19:58:15 leet-ricer-distro.conf will use git Oct 11 19:58:31 as in bigdog wants a kernel with display enabled Oct 11 19:58:33 and conservative-distro.conf will use 2.6.12 Oct 11 19:58:38 and mine is displayless Oct 11 19:58:53 can we call for kernel features in distro? Oct 11 19:59:02 Crofton: thats a matter of modules for fb Oct 11 19:59:03 OSK is funny in that display is optional Oct 11 19:59:07 true Oct 11 19:59:14 Crofton: 2.6.12-rc has fb in modules Oct 11 19:59:18 make the lcd drivers modulear? Oct 11 19:59:24 yeah Oct 11 19:59:26 lol Oct 11 19:59:28 I thought of that to Oct 11 19:59:32 bad example :) Oct 11 20:00:19 Crofton: my kernel recipe changes can be omitted. git set to do http instead of git (proto change), added compilefix.patch for !dsp but you changed defconfig iirc Oct 11 20:00:36 yeah Oct 11 20:00:50 Kristoffer: From the kernel perspective that sounds fine :) Oct 11 20:00:52 eventually i would the git recipe to work from the defconfig Oct 11 20:00:56 Just need to fix userland :) Oct 11 20:01:07 and then create stable bb files for real use Oct 11 20:01:24 I will do that as soon as the default config compiles Oct 11 20:01:39 of course I am hoping dspgw goes in to the default :) Oct 11 20:02:12 RP, yeah Know. :D Just wanted to say good work Oct 11 20:02:31 glibc 2.5 still hasn't crashed the hx Oct 11 20:02:41 there *has* be soemthing wrong Oct 11 20:02:45 hehe Oct 11 20:02:49 koen: stop teasing Oct 11 20:02:51 simpad image still works Oct 11 20:02:52 * pH5 resists Oct 11 20:02:55 aeh compiles Oct 11 20:03:02 I don'thave a problem changing to http from git Oct 11 20:03:23 that would be a proto=http added to the k source line? Oct 11 20:04:17 pH5: http://www.angstrom-distribution.org/unstable/feed/armv5te/base/libc6_2.5-r0_armv5te.ipk Oct 11 20:05:30 :) Oct 11 20:06:27 hrw, I suspect the other omap kernel recipes will need to have the entry points changed also Oct 11 20:07:23 koen, if I need to have OE configure & compile the sphinxbase tarball first (because the pocketsphinx one refers to it in it's configure line) Oct 11 20:08:19 Is the only way via an override using something like do_compile (is there a do_configure?) Oct 11 20:09:42 Crofton, by the way, is it hard to develop code for dspgw? Oct 11 20:10:31 are there any good samples of working real world code available? Oct 11 20:12:20 not sure et Oct 11 20:12:39 there are some examples, but haven't got the basic stuff working yet Oct 11 20:13:16 If someone gets super interested, I'll add the user space stuff to OE before I need it Oct 11 20:14:23 I don't know if I'm really interested in user space stuff for OE (I'm Nokia 770 user), but what do you think, is it possible to implement efficient image scaling (for displaying video in realtime) using C55x? Oct 11 20:14:32 hrw, doesnt seem to do anything when using DISTRO_EXTRA_RDEPENDS = "..." Oct 11 20:15:03 I don't really know Oct 11 20:15:52 http://www.internettablettalk.com/forums/showpost.php?p=23273&postcount=220 Oct 11 20:16:44 I'm sorry if it is offtopic here, I'm trying to get MPlayer optimized for Nokia 770, and making use of DSP core at least for some stuff is interesting Oct 11 20:17:00 right Oct 11 20:19:28 I would love for people to screw with the dsp gateway so they can tell me how it works Oct 11 20:22:06 does task-base get run with every image build? Will I need to do -c clean task-base to ensure that my changes get into it? Oct 11 20:28:57 how would one go about specifying which of two tarball to configure/compile is done first (and how to specify unique configure arguments for each) Oct 11 20:29:05 in a package. Oct 11 20:30:19 setting DISTRO_EXTRA_RDEPENDS = "konqueror-embedded" .. doesnt do anything for the build Oct 11 20:30:32 I've cleaned out task-opie task-base opie-image and rebuilt and nothing Oct 11 20:30:35 its driving me nutters Oct 11 20:30:45 I need to have the second source folder do "./configure --with-sphinxbase=`pwd`/../" Oct 11 20:33:18 Is there such a thing as oe_runconfigure ? Oct 11 20:33:32 oe_runconf Oct 11 20:34:24 good. I guess I'll have to do override the default actions and do everything manually right? Oct 11 20:36:03 re Oct 11 20:36:15 hrw: wb Oct 11 20:36:32 Kristoffer: task-base should be enough Oct 11 20:36:35 when calling run oe_runconf though does it still fills in all the host/crosstool specific for me? Oct 11 20:36:46 Crofton: create linux/linux-omap.inc with common stuff Oct 11 20:37:06 ok Oct 11 20:37:12 I'll look at this tomorrow Oct 11 20:37:16 need to run now Oct 11 20:37:19 cu Crofton Oct 11 20:37:27 there are kernels like this already? Oct 11 20:38:21 Crofton: linux-openzaurus Oct 11 20:38:29 ls *inc :) Oct 11 20:38:56 I'll try and get one kernel built from the kernel source+patch Oct 11 20:39:07 the git tracking Oct 11 20:39:13 hrw, oki will try it. will add something else to see if it grabs that, if so then theres something with the konqueror package that Im doing wrong Oct 11 20:39:22 and something post 2..618 with the latest dspgw Oct 11 20:39:56 Crofton: kind of 'booting, working' one + devel one? Oct 11 20:40:10 yeah Oct 11 20:40:25 I want to keep the git tracker so I can detect breakage on git Oct 11 20:40:28 :) Oct 11 20:41:38 Crofton: normal ARM machines are tracked by KAutoBuild ;) Oct 11 20:42:53 evening all Oct 11 20:43:26 args Oct 11 20:43:36 again why opie-images build me perl native Oct 11 20:43:48 woglinde: bitbake -nv opie-image Oct 11 20:43:52 that must be some stupid-thing somewhere Oct 11 20:44:00 wow bitbake nano takes so long Oct 11 20:44:08 i have 1.5gb in /stuff/ Oct 11 20:44:11 :X Oct 11 20:44:21 hrw will check after the build of the image Oct 11 20:44:31 woglinde: ok Oct 11 20:44:43 hrw - is it normal for compiles to take an hour? Oct 11 20:44:50 or is it just the first time Oct 11 20:44:53 siriusnova: only hour? Oct 11 20:44:56 and later on its quicker Oct 11 20:44:59 lol Oct 11 20:45:10 well its been an hour and stuff is still compiling Oct 11 20:45:13 ccache Oct 11 20:45:18 and all i compiled is nano Oct 11 20:45:22 siriusnova: first build have to build whole toolchain and other stuf Oct 11 20:45:30 how long does that take? Oct 11 20:45:38 depends on the speed of your computer Oct 11 20:45:41 siriusnova: depends on machine Oct 11 20:45:45 1.8 ghz P4 Oct 11 20:45:52 thinkpad laptop Oct 11 20:46:02 siriusnova: my DL_DIR is 4.4GB Oct 11 20:46:06 shouldn't be too long...hours only Oct 11 20:46:10 hmm k Oct 11 20:46:21 right now its generating Locales Oct 11 20:46:26 there is a big list Oct 11 20:46:34 hrw how was your trip back to poland? Oct 11 20:46:39 siriusnova: it will take 2-3 hack Oct 11 20:46:42 then you're getting pretty close Oct 11 20:46:51 2-3 hack ? Oct 11 20:47:00 ok good Oct 11 20:47:08 lets hope it doesnt error out at the end Oct 11 20:47:08 :X Oct 11 20:47:10 woglinde: no problems except that Germans lost one wagon from train (hopefully not mine) Oct 11 20:47:19 uh Oct 11 20:47:46 woglinde: so train was ~20 minutes later at Hbf then had to be Oct 11 20:47:48 hrw, do I need to add the task-base package to my DISTRO_EXTRA_RDEPENDS?.. I tried to add mc to distro_.... and it didnt get installed Oct 11 20:48:51 Kristoffer: create own image and add that stuff into IPKG_INSTALL - will be faster now Oct 11 20:49:15 hrw, hmm yeah probobly.. Oct 11 20:49:23 Kristoffer: I just cleaned task-base to get bootstrap-image working as it should. Still have to find time to clean oter images Oct 11 20:50:03 Kristoffer: and it takes time which I usually does not have ;( Oct 11 20:50:23 Kristoffer: during weekend I will work on getting some stuff into .dev and .documentation Oct 11 20:51:39 hrw, sounds good man. Thx for helping out tonight :) been abit stressed due to release coming up in a day or so Oct 11 20:51:52 Kristoffer: I know that feeling. Oct 11 20:52:06 Kristoffer: that triple checking does everything works on all supported devices... Oct 11 20:52:20 Kristoffer: finding brave testers for -rcLAST versions... Oct 11 20:52:41 hrw, Yeah :D and adding the last needed program Oct 11 20:52:54 Kristoffer: and last polishing of images Oct 11 20:53:24 :D Oct 11 20:54:33 hms Oct 11 20:54:50 woglinde: 23:13:20 Oct 11 20:54:53 perl is build native and than split up in arm-packages? Oct 11 20:54:56 or even 231320 Oct 11 20:55:12 woglinde: perl-native is needed for perl which split up into arm packages Oct 11 20:55:47 perl packages sucks more than glibc-locale Oct 11 20:58:58 way more Oct 11 20:59:04 python too Oct 11 21:00:15 I noticed you can do EXTRA_CONF_append_