**** BEGIN LOGGING AT Tue Dec 08 02:59:56 2009 Dec 08 06:06:53 that was annoying Dec 08 06:12:25 morning Dec 08 06:16:01 heyo Dec 08 06:16:05 its 1am here :P Dec 08 06:35:35 03Martin Jansa  07org.openembedded.dev * r51f6b034a1 10openembedded.git/conf/distro/micro.conf: Dec 08 06:35:35 micro.conf: remove preferred providers as the same are defined in compatibility-providers.conf Dec 08 06:35:35 Signed-off-by: Martin Jansa Dec 08 06:35:36 03Martin Jansa  07org.openembedded.dev * r0daead504d 10openembedded.git/conf/ (bitbake.conf compatibility-providers.conf): Dec 08 06:35:38 bitbake.conf: create compatibility-providers.conf from micro, minimal, angstrom, shr preferred providers Dec 08 06:35:40 Signed-off-by: Martin Jansa Dec 08 06:35:42 03Thomas Zimmermann  07org.openembedded.dev * rfb55f52daa 10openembedded.git/ (4 files in 3 dirs): Dec 08 06:35:45 epdfview: add icon for shr and recipe for svn version Dec 08 06:35:47 Signed-off-by: Klaus Kurzmann Dec 08 06:35:49 Signed-off-by: Martin Jansa Dec 08 06:35:51 03Martin Jansa  07org.openembedded.dev * r9b641bbfed 10openembedded.git/ (10 files in 4 dirs): Dec 08 06:35:54 opkg: add provides for update-alternatives and set opkg as default u-a provider in sane-feed-ipk Dec 08 06:35:56 Signed-off-by: Martin Jansa Dec 08 06:36:00 03Martin Jansa  07org.openembedded.dev * ra8a0773d69 10openembedded.git/conf/distro/minimal.conf: Dec 08 06:36:05 minimal.conf: remove preferred providers as the same are defined in compatibility-providers.conf Dec 08 06:36:07 Signed-off-by: Martin Jansa Dec 08 06:36:09 03Martin Jansa  07org.openembedded.dev * r6f22f90f27 10openembedded.git/conf/distro/angstrom-2008.1.conf: Dec 08 06:36:14 angstrom-2008.1.conf: remove preferred providers set already in compatibility-providers.conf Dec 08 06:36:16 Signed-off-by: Martin Jansa Dec 08 06:36:18 03Martin Jansa  07org.openembedded.dev * r8b332a7d14 10openembedded.git/conf/distro/shr.conf: Dec 08 06:36:21 shr.conf: remove preferred providers set already in compatibility-providers.conf Dec 08 06:36:23 Signed-off-by: Martin Jansa Dec 08 06:37:49 ~ugt Dec 08 06:37:50 methinks ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Dec 08 06:37:53 Aditya__: 7:37 here Dec 08 06:39:50 03Martin Jansa  07org.openembedded.dev * r408c47697d 10openembedded.git/conf/compatibility-providers.conf: Dec 08 06:39:50 compatibility-providers: update frameworkd-config Dec 08 06:39:50 Signed-off-by: Martin Jansa Dec 08 06:45:14 see you in 3h Dec 08 07:20:38 could someone tell me how i can check out a specific revision of bitbake via git? Dec 08 07:20:49 eg ffdba52362372efc5f18fd5d42697483c0109b4f, or 1.8.18 Dec 08 07:23:24 erm nm. Dec 08 07:23:40 i just cloned the regular bitbake.git, and then did 'git checkout ffdba52362372efc5f18fd5d42697483c0109b4f' Dec 08 08:00:17 is there someway I can get the licenceinfo for an image? like a packages-installed.txt but with their license also? Dec 08 09:38:27 03Martin Jansa  07org.openembedded.dev * r298787c63e 10openembedded.git/ (3 files in 3 dirs): Dec 08 09:38:27 e-tasks: bump srcrev, update recipe from c_c, fix docs dir Dec 08 09:38:27 Signed-off-by: Martin Jansa Dec 08 09:47:40 XorA|gone: ping Dec 08 09:47:41 pkg_parse_from_stream_nomalloc: missing new line character at end of file! Dec 08 09:47:57 we get this a lot from the newer opkg Dec 08 09:48:11 the status file is corruped Dec 08 09:50:28 Does anyone know what sort of state the canadian SDK stuff is in these days? Got a request to do a windows native SDK/toolchain and I am not in the mood to battle with MinGW ;). Dec 08 09:50:51 good morning Dec 08 09:53:34 Morning florian Dec 08 10:28:27 florian: good morning Dec 08 10:28:31 he pb Dec 08 10:28:35 mickey|office: good morning Dec 08 10:28:36 woglinde: good morning Dec 08 10:29:12 hi pb_ Dec 08 10:29:20 hi florian Dec 08 10:29:26 hi woglinde Dec 08 10:30:24 uh Dec 08 10:30:26 http://www.linuxfordevices.com/c/a/News/Kontron-and-ADL-Microspace-MPCX28/ Dec 08 10:30:35 poulsbo with 256 mbram Dec 08 10:32:42 poulsbo sucks badly Dec 08 10:33:07 florian hm Dec 08 10:33:14 with iegd its okay Dec 08 10:33:29 but thats not everybody knows Dec 08 10:37:45 woglinde: quite big for carpc Dec 08 10:37:50 morning Dec 08 10:38:18 hrw yes Dec 08 10:38:22 woglinde: sounds like a time consuming hacking session to get it working with ubuntu Dec 08 10:38:27 hi hrw Dec 08 10:38:33 hi cbrake Dec 08 10:38:34 florian hm nope Dec 08 10:38:44 only download building kernel modules Dec 08 10:38:51 and installing libs Dec 08 10:39:02 latest revision works fine with ubuntu 9.05 Dec 08 10:39:18 I have it running for project Dec 08 10:39:24 3d dont works Dec 08 10:39:27 woglinde: sounds good Dec 08 10:39:39 but I didnt check they latest version Dec 08 10:39:47 it even ships libvaapi Dec 08 10:39:55 but most stuff is binary only Dec 08 10:39:57 as expected Dec 08 10:41:12 woglinde: gma500 is disaster nevermind what intel will tell about it Dec 08 10:41:37 hrw sure for free software its disaster Dec 08 10:42:24 oh my god why 2.6.32 so badly behave with postgres Dec 08 10:42:34 hm buts only phoronix test Dec 08 10:42:47 and ext4 Dec 08 10:44:59 morning pb Dec 08 10:45:00 hi florian, woglinde Dec 08 10:45:43 yes... updating the binary crap for every ubuntu update will be no fun. apart from mentioned disasters... Dec 08 10:45:49 hi mickey|office Dec 08 10:48:45 woglinde: I use 32-rc7 here Dec 08 10:49:13 hrw I am going with debian kernel releases Dec 08 10:49:23 and I dont use ext4 Dec 08 10:54:58 03Henning Heinold  07org.openembedded.dev * rcc665bd8ea 10openembedded.git/ (3 files in 3 dirs): libcli: introduce cisco ios command line like library Dec 08 10:54:59 03Henning Heinold  07org.openembedded.dev * r6676586580 10openembedded.git/ (3 files in 3 dirs): libdessert: introduce framework to develop routing protocols Dec 08 11:18:41 quit Dec 08 11:19:53 I have a trouble when I was trying to compile ffmpeg, i get some errors with libxcb and X11 libraries, I paste here the log : http://pastebin.com/m5a0c776 . Anyone may help me, please? Dec 08 11:36:24 I have a trouble when I was trying to compile ffmpeg, i get some errors with libxcb and X11 libraries, I paste here the log : http://pastebin.com/m5a0c776 . In the log discover the error but i don't know how to fix, the error is that: The compile tried to make a link with a system file (xproto.xml) and not the file that is inside the folder libxcb. How may i change that? Dec 08 11:37:48 03Martin Jansa  07org.openembedded.dev * r7679546068 10openembedded.git/ (3 files in 3 dirs): Dec 08 11:37:48 shr-launcher, e-tasks: bump srcrev, remove sed from recipe as fix applied upstream Dec 08 11:37:48 Signed-off-by: Martin Jansa Dec 08 11:39:16 roliveira hm libxcb Dec 08 11:41:37 woglinde: I was looking for the Makefile and founded this line: XCBPROTO_XCBINCLUDEDIR = /usr/share/xcb . I think this is the error, what do you think? Dec 08 11:43:05 hm yes Dec 08 11:43:14 that would search on host Dec 08 11:43:23 okay I have to go now Dec 08 11:46:46 Anyone know what I have to do to compile the ffmpeg recipe whitout X11 libraries and gtk+? Dec 08 12:15:21 morning Dec 08 12:15:21 does anyone know which package provides xmodmap and friends? Dec 08 12:15:55 xmodmap is provided by.... xmodmap Dec 08 12:16:17 lol Dec 08 12:16:26 is that all I need to get decent keymap settings on a laptop? Dec 08 12:18:49 oh I just found I don't have xserver-common maybe that's why ... Dec 08 13:21:15 hmm Dec 08 13:21:39 I have tried loading xserver-common it just makes things worse and xmodmap /etc/X11/default.xmodmap makes the keyboard even more wacky Dec 08 13:21:54 any idea what I can do, to make the keyboard sane on a laptop? Things like page up/down and arrow keys aren't working Dec 08 14:02:48 03Sebastian Spaeth  07org.openembedded.dev * re2b09f5910 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Dec 08 14:02:48 sane-srcrevs: bump shr-related elementary-* themes Dec 08 14:02:48 Signed-off-by: Sebastian Spaeth Dec 08 14:31:52 03Sebastian Spaeth  07org.openembedded.dev * r30a9cb7e62 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Dec 08 14:31:52 sane-srcrevs: bump shr-settings Dec 08 14:31:52 Signed-off-by: Sebastian Spaeth Dec 08 15:25:45 someone built qt3 recently? Dec 08 16:36:12 could someone give me a hint as to where i'm going wrong here: http://pastebin.com/d22386c63 Dec 08 16:36:24 i'm trying to add a vr5500 mipsel target. that's from gcc-cross-initial Dec 08 16:38:17 http://pastebin.com/d408fd8e is tune-vr5500.inc Dec 08 16:39:43 FLAGS_FOR_TARGET = -march=mips32 -mcpu=vr5500 -mabi=n32 -mhard-float -isystem/home/matt/devel/rockhopper/tmp/staging/vr5500-angstrom-linux/usr/include -B$(build_tooldir)/bin/ -B$(build_tooldir)/lib/ -isystem $(build_tooldir)/include -isystem $(build_tooldir)/sys-include Dec 08 16:45:58 hrm maybe i just have to use march=vr5500? Dec 08 16:59:47 Anyone may help me, please? I'm trying to put mkfs.ext2 and mkfs.jffs2 apps inside my image. I wrote "e2fsprogs" and "mtd-utils" in IMAGE_INSTALL, but doesn't work, when I look in the rootfs /bin, there aren't this apps. What may i do to solve it? Dec 08 17:00:13 is this a custom image? Dec 08 17:00:16 well. Dec 08 17:00:27 i mean, are you modifying one of the pre-existing ones? like base-image? Dec 08 17:00:28 mt4: yes, it is a custom image Dec 08 17:00:47 there are two sections you probably need Dec 08 17:01:08 DEPENDS and IMAGE_INSTALL Dec 08 17:01:16 mt4: yes, i changed the minimal image, and configured the distro conf Dec 08 17:01:43 the actual .ipk for those two are in tmp/deploy/*libc/machine? Dec 08 17:02:45 can unpack with 'ar x pkg.ipk' and check what data.tar.gz looks like... Dec 08 17:02:52 really they should be in there. Dec 08 17:03:13 mt4: The Depends olny have "task-base" and in IMAGE_INSTALL there are task-base and the variable EXTRA_INSTALL, and there i put the e2fsprogs and mtd-utils. Dec 08 17:03:29 hrm i dunno what EXTRA_INSTALL does. Dec 08 17:03:52 for the one image i've made i added them to DEPENDS and IMAGE_INSTALL Dec 08 17:04:12 might not be proper. but they all were included in the cpio i used. Dec 08 17:04:14 m4t: The Extra_Install is a varible with several apps to install Dec 08 17:04:46 roliveira , my only suggestion to you is ditch EXTRA_INSTALL and use the two i mentioned Dec 08 17:04:56 m4t: i'm going to put in the DEPENDS flag to, and see what happens Dec 08 17:05:07 m4t: thanks Dec 08 17:05:22 btw, is this angstrom? Dec 08 17:05:26 you *might* need ANGSTROM_EXTRA_INSTALL. Dec 08 17:06:18 -march=vr5500 seems to be working, regarding my earlier question Dec 08 17:06:36 i mightve been looking at some old redhat documentation when i pulled out those targets Dec 08 17:06:42 gcc 3.x Dec 08 17:46:13 m4t: vr5500 is not a supported option for gcc but for gas Dec 08 17:46:47 hmm -march=v5500 'worked' Dec 08 17:46:51 is my toolchain going to broken Dec 08 17:47:28 i got an old rockhopper vr5500 board for $0.99 Dec 08 17:47:43 no that should be ok Dec 08 17:48:04 cool Dec 08 17:48:14 its nice. gold headers etc. Dec 08 17:48:17 '-march=mips32' is not compatible with the selected ABI Dec 08 17:48:28 yea. n32 didn't work either Dec 08 17:48:32 just -march and hard fpu Dec 08 17:48:33 is also a problem and specifying march should fix it too Dec 08 17:49:01 n32 would have worked I believe Dec 08 17:49:07 with march=vr5500 Dec 08 17:49:11 said incomaptible abi Dec 08 17:49:28 even with -march=vr5500 ? Dec 08 17:49:42 -mabi=n32 Dec 08 17:49:43 yea Dec 08 17:49:57 strange I think it should not happen Dec 08 17:50:27 i dunno Dec 08 17:50:28 -march=mips2 -mabi=n32 are incompatible but not -march=vr5500 -mabi=n32 Dec 08 17:51:47 i wonder what it defaults to Dec 08 17:52:50 ive got a lot of work ahead on this one. the rockhopper board support has long been deprecated, but only recently was improved/proper vr5500 support added for the emma2h multimedia soc's Dec 08 17:53:06 so my goal is to take the old board support. which specified vr5432 i think Dec 08 17:53:21 and bring it up to speed with config_vr5500 and 2.6.32 Dec 08 17:54:00 hmm cool Dec 08 17:54:09 yea Dec 08 17:54:21 i want to find an 800mhz cpu card too, i think this one is only 3-400mhz Dec 08 17:54:33 i havent booted it yet though, need a 20pin atx Dec 08 17:55:25 oh there is 800mhz one too ? Dec 08 17:55:46 yup so says the bbrochure. its a bit newer Dec 08 17:56:32 hmm ok Dec 08 17:57:42 http://www.am.necel.com/docs/files/50398-2.pdf is the brochure Dec 08 18:15:06 03Liam Girdwood  07org.openembedded.dev * rff16088a34 10openembedded.git/ (2 files in 2 dirs): Dec 08 18:15:06 libomx-bellagio: added version 0.9.2 Dec 08 18:15:06 Signed-off-by: Liam Girdwood Dec 08 18:42:17 see you tomorrow Dec 08 19:05:48 Any thoughts on why building poky-image-sato with L23.i3.1 hangs after "external-csl-toolchain-2008q3-72: completed" (more log here: http://pastebin.com/m4f19d698) - no errors though. Thanks. Dec 08 19:14:20 Magdalena: you might force bitbake to spit more verbose messages so you know whats going on Dec 08 19:29:38 khem: thanks, will try digging more into it Dec 08 19:29:54 hmmm, where is ant hiding Dec 08 19:38:55 in colonies :) Dec 08 19:39:13 hi XorA Dec 08 19:40:29 hey khem Dec 08 19:41:15 mickeyl: good evening Dec 08 19:41:24 * pb__ returns from reindeer visiting night Dec 08 19:42:03 heya Dec 08 19:42:06 aw, cute Dec 08 19:42:09 did she like it? Dec 08 19:42:41 (or why did you go? ;) Dec 08 19:43:26 heh Dec 08 19:43:26 yeah, it was fun, I think Emma enjoyed it Dec 08 19:46:07 managed to do almost all of our remaining Christmas shopping at the same time, which was a bonus. Dec 08 19:53:16 re Dec 08 19:56:03 wow, quite in time then Dec 08 19:56:17 * mickeyl still has nothing yet Dec 08 19:56:46 * XorA will have a mad 2 days buying Dec 08 19:57:06 unless I find time to shop in Utah Dec 08 19:58:21 wb Dec 08 19:58:32 hi dth-oops :-) Dec 08 19:58:42 oops Dec 08 19:59:50 hi florian Dec 08 20:00:19 i am ready with the shitlist Dec 08 20:00:28 :-) Dec 08 20:00:36 [x] done :-) Dec 08 20:00:58 alrighdy Dec 08 20:01:23 * florian going downstairs to get some food and will start to collect kernel updates for multiple devices then Dec 08 20:01:34 lets make some stimulating work Dec 08 20:01:41 :) Dec 08 20:01:59 gh Dec 08 20:02:19 like... merging 2.6.25 kernel patches into 2.6.32 ;) Dec 08 20:02:20 brb Dec 08 20:02:58 any bitbake folk mind reviewing the 3 new patches on the list when they get a chance, before i see about pushing them to master? they're the last 3 non-mvl6-specific changes we have to bitbake at MV Dec 08 20:03:52 now.. syncing the metadata is not going to be fun, at all :| Dec 08 20:11:09 ~ugt Dec 08 20:11:10 somebody said ugt was Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Dec 08 20:13:35 03Graeme Gregory  07org.openembedded.dev * rcd99ea7115 10openembedded.git/contrib/angstrom/omap3-mkcard.sh: contrib/angstrom/omap3-mkcard.sh : add script for formatting cards for omap3 Dec 08 20:15:29 mickeyl: yah, my wife is more organised than me :-} Dec 08 20:16:38 Ahem... Good evening Dec 08 20:16:42 guh, i still need to buy just about every christmas present.. someday i'll do it with time to spare.. someday Dec 08 20:16:42 hehe Dec 08 20:17:10 PE: good evening Dec 08 20:17:32 i think trying to sync these repositories with oe is going to drive me absolutely insane. Dec 08 20:17:48 * XorA is only really stumped with one present to buy Dec 08 20:17:51 * kergoth thinks about how we could improve the system to make overlay management easier Dec 08 20:18:02 kergoth: what sort of diffs do you have? Dec 08 20:18:20 as in, which sorts of changes are currently not working well with overlays? Dec 08 20:18:22 I'd be happier if people would actually give me a christmas list, so i have some clue what to get them, otherwise they'd better be happy with a gift certificate Dec 08 20:18:55 kergoth: hehe that would make the world easier Dec 08 20:19:05 pb__: just in general, keeping track of what changes exist between the repositories for each file.. maintaining a more minimal repository means you aren't based on the upstream oe git branches, so a regular git log isn't exactly helpful Dec 08 20:19:15 personally I prefer people come to pub with me and ignore presents Dec 08 20:19:31 I dont feel like such a bad citizen of the world then Dec 08 20:19:41 Okay, I am an openembedded newbe looking for advice: I just want o confirm what I have read for my project: I want to run owfs (One-wire File System) on a NSLU2 Dec 08 20:19:47 thus far the best I've found is to generate a -s ours merge commit from the date i created the collection in OE, and then look at logs in both directions restricted by the files that actually exist in the repository Dec 08 20:20:49 hmm Dec 08 20:22:40 what if i looked at a given file, took the blob hash of its initial version and traversed OE history to find the commit that changed it to be that blob, then I'd know the exact branch-off-point for that file.. Dec 08 20:22:49 course if the initial version already had local changes against upstream, hosed Dec 08 20:26:10 re Dec 08 20:26:25 wb Dec 08 20:29:11 hi Dec 08 20:31:54 Anyone knows what packet i have to compile the jffs2 progs in my distro? Dec 08 20:39:49 roliveira: mtd-utils Dec 08 21:13:58 03Klaus Kurzmann  07org.openembedded.dev * rd3dd1db4be 10openembedded.git/recipes/e17/eve_svn.bb: Dec 08 21:13:58 eve: bump PR to pickup new EFL_SRCREV Dec 08 21:13:58 Signed-off-by: Klaus Kurzmann Dec 08 21:14:28 jo effem Dec 08 21:24:23 woglinde, hi, at what point are you with ooo? Dec 08 21:24:38 someone told me that with 256M of ram it ate all the ram Dec 08 21:24:54 and koen told me that under 2GB ***[censored] Dec 08 21:25:01 gnutto hm **** BEGIN LOGGING AT Tue Dec 08 21:26:51 2009 Dec 08 21:26:54 problem right now glibc and eglibc dont work at my works computer Dec 08 21:27:04 or better arnet compilable Dec 08 21:27:13 glibc as an ubscure udev error Dec 08 21:27:45 what about eglibc Dec 08 21:27:55 eglibc maybee default binutils of angstroem are to old Dec 08 21:28:06 khem some UNWIND compile erros Dec 08 21:28:15 ah Dec 08 21:28:24 give me more details Dec 08 21:28:28 man.. i really don't know how best to deal with this situation where recipes are incompatible amongst collections due to config & class changes Dec 08 21:28:28 decidedly nontrivial Dec 08 21:28:36 * kergoth sighs Dec 08 21:28:47 khem on thursday I will be at work again Dec 08 21:29:05 ok Dec 08 21:29:35 unwind errors could be toolchain misconfig or as you said old gas Dec 08 21:31:27 kergoth: collections should adapt always imo Dec 08 21:31:53 adapt in what way? Dec 08 21:32:16 oe recipes are changed to match oe configuration metadata, but try to bring poky recipes in, or take those oe recipes into poky, and the shit hits the fan Dec 08 21:32:28 (just as a theoretical example) Dec 08 21:33:51 may be there should be versioning of config metadata Dec 08 21:34:05 to which a recipe is developed on Dec 08 21:34:07 yeah, I've thought about it for classes, but the config metadata is more painful Dec 08 21:34:10 hrm Dec 08 21:34:57 i could see an argument for moving variable defaults into the classes that use them, and then version those classes.. but +=/=+ are problematic then Dec 08 21:35:55 yeah Dec 08 21:36:54 what would really be nice is if recipes could *declare* config variables, and those get pushed up and used as defaults, so += in a distro .conf would still function Dec 08 21:37:02 make things less line-by-line imperative Dec 08 21:38:28 something like includes and two pass parsing Dec 08 21:39:24 yeah. though perhaps everything could be parsed at once, just evaluation / expansion / whatever postponed until certain other things get done.. split the evaluation Dec 08 21:39:35 * kergoth shrugs, can't think of an ideal solution for this mess Dec 08 21:40:18 montavista customers want to be able to pull recipes from community repositories to use with the product, but compatibility issues are cuasing problems with that. its not a new problem, poky/oe can hit it too.. just.. ugh Dec 08 21:41:21 yeah hmm Dec 08 21:42:04 everything that comes to mind involves a complete redesign of bitbake & oe :) Dec 08 21:42:04 * kergoth shakes head Dec 08 21:43:20 kergoth written with qt? Dec 08 21:45:21 kergoth: may be maintain backward compatibility at any cost Dec 08 21:45:28 like symbol versioning in glibc Dec 08 21:47:07 khem: heh, yeah, i guess we could try to retain compatibility in config files & classes indefinitely, but somehow i doubt people would want to do that, would keep building up more and more compatibility cruft Dec 08 21:47:08 hmm Dec 08 21:48:21 yeah Dec 08 21:49:31 * kergoth can't imagine trying to go back and restore compatibility across layout_* removal, sdk rework, etc Dec 08 21:49:58 XorA: pong? Dec 08 21:51:03 course, this kind of goes back to the discussions we've had in here about recipe responsibility.. if the distro/packaging/etc was more in the distro hands and out of the recipes, recipes would become more portable Dec 08 21:53:07 yeah Dec 08 21:53:15 that wud be nice Dec 08 21:55:46 hms someone should fix oe commit ml stuff Dec 08 21:55:52 ant__: were you seeing issues where your linux-kexecboot never did do_deploy? Dec 08 21:55:53 it makes unacessary traffic Dec 08 21:55:57 killing staging was a good step in the recipe simplification direction.. if we ripped the packaging/files bits into the distros, thatd take it even further.. wouldn't be much left but descriptive metadata, configure/compile/install Dec 08 21:56:00 hmm Dec 08 21:56:10 ant__: unless you specifically did bitbake linux-kexecboot Dec 08 21:56:16 I would look at it if I would get the admin rights Dec 08 21:56:32 XorA: not that I remember Dec 08 21:56:46 l-k is built when building from scratch Dec 08 21:57:07 at leas for Z machines Dec 08 21:58:10 XorA: last line of zaurus-2.6.inc Dec 08 21:58:50 ant__: yeah but that has nothing to do with whether it deploys Dec 08 21:59:14 ? Dec 08 21:59:15 EXTRA_IMAGEDEPENDS += "linux-kexecboot" Dec 08 21:59:21 what I see here is my recipe builds, but never deploys Dec 08 21:59:39 unless I specifically call bitbake linux-keecboot Dec 08 22:05:39 hmm Dec 08 22:06:50 * kergoth gives thought to adding a distro specific path to FILESPATH, then using a distro amend.inc to inject the packaging information into the recipe without being part of the recipe Dec 08 22:21:32 * kergoth mulls over reintroducing a manifest file for packaging description, rather than using bitbake variables Dec 08 22:26:10 03Klaus Kurzmann  07org.openembedded.dev * re6a4fd3e02 10openembedded.git/conf/distro/include/sane-srcrevs-fso.inc: Dec 08 22:26:10 sane-srcrevs-fso.inc: bump rev for libframeworkd-glib Dec 08 22:26:10 Signed-off-by: Klaus Kurzmann Dec 08 22:45:33 jo gnutoo Dec 08 22:45:58 I'll go to sleep Dec 08 22:46:06 I was just trying 2.6.32 on my desktop Dec 08 22:46:43 nite Dec 08 22:52:28 besser :) Dec 08 22:52:40 oops Dec 08 22:58:11 florian: woglinde: ping Dec 08 22:58:26 one doubt about head.S Dec 08 22:58:40 ant__: pong Dec 08 22:58:40 on armv5te (Zaurus) Dec 08 22:58:51 --- linux/arch/arm/boot/compressed/head.S~pxa-linking-bug Dec 08 22:58:51 +++ linux/arch/arm/boot/compressed/head.S Dec 08 22:58:51 @@ -1064,5 +1064,9 @@ Dec 08 22:58:51 reloc_end: Dec 08 22:58:51 . Dec 08 22:58:52 <-----><------>.align Dec 08 22:58:54 +#if defined(CONFIG_MACH_POODLE) || defined(CONFIG_PXA_SHARPSL) Dec 08 22:58:56 +<-----><------>.section ".stack", "aw" Dec 08 22:58:58 +#else Dec 08 22:59:02 <-----><------>.section ".stack", "w" Dec 08 22:59:04 +#endif Dec 08 22:59:06 user_stack:<-->.space<>4096 Dec 08 22:59:08 is necessary in 2.6.32? Dec 08 23:00:02 this patch was in all previous linux-RP Dec 08 23:04:52 hmm .stack should be allocatable Dec 08 23:06:03 I'm flashing now...let see... Dec 08 23:06:06 YES! Dec 08 23:06:13 now can boot from nand! Dec 08 23:07:48 ant__: we expect linux-kexecboot to be 2.6.32 2 weeks ago, get to it :-D Dec 08 23:08:11 that was the missing key-patch :=) Dec 08 23:08:22 dunno why has status:unmergeable doh Dec 08 23:08:38 ant__: I guess khem has some ideas on that Dec 08 23:09:17 well..it crashes..probably I need another patch Dec 08 23:09:21 :P Dec 08 23:10:14 ah..no Dec 08 23:10:24 touchscreen works, keyboard not Dec 08 23:10:30 wow Dec 08 23:10:51 should'nt be difficult to unbreak Dec 08 23:14:02 ant__: did you also change vmlinux.lds.in Dec 08 23:14:18 .stack (NOLOAD) : { *(.stack) } Dec 08 23:14:25 you should get rid of NOLOAD Dec 08 23:14:44 and make the section "wa" in .S file Dec 08 23:17:12 absolutely not... Dec 08 23:17:27 but I think it should be some linker problem that triggerred this patch Dec 08 23:17:29 ~lart developer of AC_PATH_XTRA Dec 08 23:17:29 * ibot pushes the wall down onto developer of AC_PATH_XTRA whilst whistling innocently Dec 08 23:17:46 I'm/I was desesperately trying patches to boot 2.6.32 frm nand Dec 08 23:18:06 the suxxer boots from kexec, btw, without patch Dec 08 23:18:22 now I ask the experts about RP patch wrt jffs2: Dec 08 23:19:50 http://fr.pastebin.ca/1707591 Dec 08 23:19:57 ant__: I think the patch must have been a workaround to bypass faulty linker or asm Dec 08 23:20:35 ah Dec 08 23:20:58 about jffs2 our Zaurus images are built -m favourlzo Dec 08 23:21:12 so I suppose the patch is not needed Dec 08 23:21:19 anymore Dec 08 23:21:49 khem: just faulty mtd routines? Dec 08 23:22:21 could be Dec 08 23:22:30 ant__: may be you have some bad blocks in your nand Dec 08 23:22:41 he..now I have a dilemma: the size of kexec static binary is too big.. Dec 08 23:22:45 that dont get used when you use kexecboot Dec 08 23:22:48 khem: yes, 2 Dec 08 23:22:59 badblocks Dec 08 23:23:30 can I try -Os with klcc? Dec 08 23:23:37 you could Dec 08 23:23:46 woglinde tried dietlibc but the binary was bigger :) Dec 08 23:24:05 or I could revert to older kexec-tools, way smaller Dec 08 23:24:08 in size Dec 08 23:24:09 * khem 's console image with uclibc nptl finished Dec 08 23:24:55 ..otherwise I have to strip down ubifs from kernel :/ Dec 08 23:25:45 ~lart kernel tendency to bloat in the years Dec 08 23:26:38 amazin it boots first time around Dec 08 23:26:51 minimal-uclibc .dev-snapshot-20091208 omap5912osk Dec 08 23:27:05 size? Dec 08 23:28:06 ant__: I am doing something else Dec 08 23:28:15 ant__: size of my root file system you mean ? Dec 08 23:29:13 just for curiosity Dec 08 23:29:32 12M Dec 08 23:29:35 -rw-r--r-- 1 andrea users 9.0M Dec 8 15:01 Angstrom-console-image-glibc-ipk-2009.X-test-20091208-c7x0.rootfs.tar.gz Dec 08 23:30:25 what's in minimal more than in A. ? Dec 08 23:30:53 hmmm no idea for now Dec 08 23:31:19 what all ipks are installed on your box Dec 08 23:31:33 I did compare the size one year ago...glibc was not bigger than uclibc :/ Dec 08 23:31:49 was a gettext issue iirc Dec 08 23:32:11 http://khem.pastey.net/129970 Dec 08 23:32:26 thats my list Dec 08 23:32:41 glibc is bigger when you compare the libraries itself Dec 08 23:33:21 e.g. my libc is 355.3K Dec 08 23:33:25 http://fr.pastebin.ca/1707608 Dec 08 23:34:37 hmm I have bluez4 Dec 08 23:34:40 I'll post the sizes if you need Dec 08 23:34:48 sure Dec 08 23:34:54 can opkg do that Dec 08 23:35:05 I'm looking in testlab Dec 08 23:35:54 http://fr.pastebin.ca/1707616 Dec 08 23:36:50 ok Dec 08 23:36:58 I would change kernel-module-nls-iso8859-1 with kernel-module-nls-iso8859-15 Dec 08 23:37:00 you seem to have more packages than me Dec 08 23:37:04 but mine seems to be fatter Dec 08 23:37:04 because of € Dec 08 23:40:40 woglinde: your libc size is 1.5M as compared to uclibc which is under .5M Dec 08 23:42:07 mine? Dec 08 23:44:06 ugh meant ant__ Dec 08 23:44:30 I have gstreamer as fat ipk Dec 08 23:44:37 I wonder why it is in console-image Dec 08 23:45:22 bluez Dec 08 23:46:40 ant__: ah you reporte size of compressed image I reported size of tar file :) Dec 08 23:46:54 if I gzip the tar file it comes down to 5.5M Dec 08 23:47:08 so that is close to half of what you have Dec 08 23:47:42 compare unpacked Dec 08 23:47:45 or jffs2 Dec 08 23:48:30 unpacked is 14M for me Dec 08 23:49:14 I use nfsboot so I dont generate jffs2 usually Dec 08 23:51:55 I'll build for uclibc soon Dec 08 23:52:01 I'll do some checks Dec 08 23:52:32 atm I'm beating head against klibc... Dec 08 23:52:45 again... Dec 08 23:54:50 khem: btw du -sh -> 24M Dec 08 23:55:03 he Dec 08 23:58:52 way bigger than mine :) Dec 09 00:00:15 lol Dec 09 00:01:39 hm..for RP patches should I refer to ${RPSRC} or can I just copy that patch in file:// ? Dec 09 00:05:50 RPSRC if applicable Dec 09 00:06:40 currently it points to http://www.rpsys.net/openzaurus/patches/archive/ Dec 09 00:06:57 see if the patch there applies cleanly to your version of kernel Dec 09 00:07:10 no, I did edit it iirc Dec 09 00:07:39 khem: I applied both Dec 09 00:07:50 vmlinux.lds.in too Dec 09 00:08:03 sorry, I see now the question Dec 09 00:08:32 patch has no headers :/ Dec 09 00:08:34 # Patch managed by http://www.holgerschurig.de/patcher.html Dec 09 00:24:50 03Andrea Adami  07org.openembedded.dev * r97ac254aec 10openembedded.git/recipes/kexecboot/linux-kexecboot_2.6.32.bb: linux-kexecboot_2.6.32:[Zaurus] unbreak boot from nand (add old patch). Dec 09 00:33:01 scheint zu bauen Dec 09 00:33:19 das ist gut Dec 09 00:33:28 heh Dec 09 00:33:50 this really looks like its time to get some sleep :-} Dec 09 00:33:58 hehe Dec 09 00:34:01 sleep well Dec 09 00:34:03 wrong info Dec 09 00:36:52 'nite all Dec 09 00:37:58 good night Dec 09 00:38:28 good night Dec 09 00:57:44 Hello, I have a quick question: in SRC_URI, is there an option to get a svn repo with module=XXX; but to have the download in ${WORKDIR}/mypath/XXX instead of ${WORKDIR}/XXX? Something like: ;module=trunk;location-to=src/project/;rev=432;proto=http. I took a look a little bit everywhere but I didn't find any such trick. I know that I could do a mv in do_fetch_append but it's more complex for some other reasons. If not, where is the code in classes tha Dec 09 01:17:26 hmm Dec 09 01:17:40 if I uninstall ccache would that break anything on debian? Dec 09 01:17:48 (I am using this VM solely to build OE Dec 09 01:18:31 Aditya_, you should be fine Dec 09 01:18:37 cool Dec 09 01:18:57 going to try and follow this guide to the letter this time :P Dec 09 01:20:13 oh Dec 09 01:20:16 also Dec 09 01:20:38 do you guys use /stuff as your working directory? I usually try to keep away from doing anything in / Dec 09 01:20:51 the guide makes it sound like using your home dir is actually a bad idea Dec 09 01:22:40 i use $HOME/oe Dec 09 01:22:48 i think most people do Dec 09 01:22:50 just trying to make sure Dec 09 01:23:12 the manual should probably be changed to avoid hanging things off / Dec 09 01:23:43 it probably encourages people to build as root, when they shouldn't Dec 09 01:24:29 yeah Dec 09 01:24:36 which is actually what threw me off Dec 09 01:24:44 cuase I remember people saying I should build as my user Dec 09 01:24:46 not as root Dec 09 01:24:58 and the entire guide seems to run in root Dec 09 01:25:05 well or as a folder given access to the user by root I guess Dec 09 01:25:38 most likely the person who wrote it had a separate partition mounted on /stuff Dec 09 01:39:20 so to answer my own question, it's not implemented and it would require a patch in bitbake itself. Feasible but a lot of pain. Dec 09 02:03:14 gregoiregentil1: what's that? Dec 09 02:05:01 read the log of the chatroom to see my question Dec 09 02:10:47 if i was in the channel 5 minutes ago, i would have Dec 09 02:10:54 i guess I'd better pull out the crystal ball Dec 09 02:12:56 Hello, I have a quick question: in SRC_URI, is there an option to get a svn repo with module=XXX; but to have the download in ${WORKDIR}/mypath/XXX instead of ${WORKDIR}/XXX? Something like: ;module=trunk;location-to=src/project/;rev=432;proto=http. I took a look a little bit everywhere but I didn't find any such trick. I know that I could do a mv in do_fetch_append but it's more complex for some other reasons. If not, where is t Dec 09 02:12:56 he code in classes th Dec 09 02:13:12 .. Dec 09 02:13:15 message ends Dec 09 02:13:42 message ends is: If not, where is the code in classes that does the svn checkout? Dec 09 02:18:41 ahh, right. do_fetch in base.bbclass calls the functions in the bb.fetch python module to do the work. don't think bb.fetch supports that directly, as you seem to have already determined :) would be trivial to add, if its determined that it satisfies a real need Dec 09 02:19:00 * kergoth grumbles Dec 09 02:19:33 trivial to add ;-) Not complex for sure but painful because you have also to deal when the tar.gz exists in downloads folder Dec 09 02:20:00 I'm stuck here because my recipe has multiple svn and they have the same module but should be put in different subfolders... Dec 09 02:20:23 anyway, I have developed a custom home-made do_fetch_append which fixes my problem Dec 09 02:20:47 ouch. yeah, that's probably your best bet, short of creating an svn repository on your own server that uses svn:exernals to construct the tree ;) Dec 09 02:20:53 blah, can't type Dec 09 02:21:02 yup Dec 09 02:28:07 kergoth: so I have written a home-made bash-based function "custom_cached_svn" in the recipe. How can I declare such function in the recipe so that I can call it multiple times in do_fetch_append? Dec 09 02:28:27 the same way you define a shell task Dec 09 02:28:44 shell tasks are just shell functions with flags set instructing bitbake about what to do with them Dec 09 02:28:48 custom_cached_svn () { Dec 09 02:28:49 echo foo Dec 09 02:28:50 } Dec 09 02:29:08 OK. strange. It's what I did and I get an error. let me check Dec 09 02:33:18 Typo. Sorry. Thanks Dec 09 02:33:22 :) Dec 09 02:34:35 03Graeme Gregory  07org.openembedded.dev * rc178abb294 10openembedded.git/ (conf/checksums.ini recipes/dvnixload/dvnixload_0.2.6.bb): Dec 09 02:34:35 dvnixload_0.2.6.bb : add tool to download to davinci chips Dec 09 02:34:35 Using BBCLASSEXTEN = "native" to give a native version too Dec 09 02:55:42 kergoth: I still think that I need your help ;-) What do I get an error with this: http://pastebin.com/d546b604c ? NameError: global name 'custom_cached_svn' is not defined Dec 09 02:57:04 gregoiregentil1: you're trying to call a shell function from python. bitbake doesn't provide a native interface between them Dec 09 02:57:25 if you want to avoid python, I'd do something like: Dec 09 02:57:31 addtask fetch_post Dec 09 02:57:34 do_fetch_post () { Dec 09 02:57:45 custom_cached_svn v8.googlecode.com/svn/trunk ${S}/v8 3431 Dec 09 02:57:46 } Dec 09 02:57:55 or rather, addtask fetch_post before do_unpack after do_fetch **** ENDING LOGGING AT Wed Dec 09 02:59:57 2009