**** BEGIN LOGGING AT Thu Jan 21 02:59:56 2010 Jan 21 06:49:17 03Martin Jansa  07org.openembedded.dev * rc85f3b3957 10openembedded.git/recipes/cairo/cairo-native_1.8.0.bb: Jan 21 06:49:17 cairo-native: depend on pixman-native instead libpixman-native Jan 21 06:49:17 Signed-off-by: Martin Jansa Jan 21 06:49:20 03Martin Jansa  07org.openembedded.dev * r7747113703 10openembedded.git/recipes/xorg-lib/pixman_git.bb: Jan 21 06:49:21 pixman_git: add BBCLASSEXTEND=native (pixman-native used in cairo-native) Jan 21 06:49:23 Signed-off-by: Martin Jansa Jan 21 06:49:25 03Michael Pilgermann  07org.openembedded.dev * r5a390f90ec 10openembedded.git/ (conf/checksums.ini recipes/openmoko-3rdparty/rtmom_0.1.3.bb): Jan 21 06:49:28 rtmom: elementary based gui for RTM (Remember the milk) Jan 21 06:49:30 Signed-off-by: Martin Jansa Jan 21 06:49:32 03Martin Jansa  07org.openembedded.dev * r51b1288c48 10openembedded.git/recipes/libpixman/ (6 files): Jan 21 06:49:35 libpixman: move all to obsolete directory (replaced with pixman recipes) Jan 21 06:49:37 Signed-off-by: Martin Jansa Jan 21 06:49:39 03Martin Jansa  07org.openembedded.dev * r86dacb28c2 10openembedded.git/ (15 files in 10 dirs): Jan 21 06:49:42 xorg: update recipes to latest releases 2010-01-18 Jan 21 06:49:44 Signed-off-by: Martin Jansa Jan 21 06:49:50 03Martin Jansa  07org.openembedded.dev * re725b1c530 10openembedded.git/recipes/xorg-lib/pixman_0.16.2.bb: Jan 21 06:49:53 pixman_0.16.2: drop DEFAULT_PREFERENCE = -1 and add BBCLASSEXTEND=native (pixman-native used in cairo-native) Jan 21 06:49:55 Signed-off-by: Martin Jansa Jan 21 06:50:26 03Martin Jansa  07org.openembedded.dev * r52391e0ef8 10openembedded.git/recipes/tasks/task-shr-feed.bb: Jan 21 06:50:26 task-shr-feed: add stopwatch and rtmom Jan 21 06:50:26 Signed-off-by: Martin Jansa Jan 21 07:12:01 good morning Jan 21 07:18:53 I've deleted tmp/work/* , now I can't rebuild . Any idea to proceed? Jan 21 07:25:11 recalcati: bitbake world ? Jan 21 07:25:37 recalcati: actually I'd suggest you delete all of tmp so also stamps, staging etc Jan 21 07:26:09 eFfeM: I need to preserve all downloads in order to create a mirror for my suppliers Jan 21 07:26:26 where are stamp only for compilations? Jan 21 07:27:47 stamps are in tmp/stamps :-) Jan 21 07:28:24 do you need to preserve downloads or packages, downloads are for me outside of tmp, generated packages are in deploy Jan 21 07:28:27 rm tmp/stamps/* doesn't give problem about fetching .. ? Jan 21 07:28:51 it will grab from downloads when it is there Jan 21 07:29:11 so tmp/downloads contains packages and their md5 . Jan 21 07:29:14 but in case of doubt i suggest you make a copy first so you can always set it back Jan 21 07:29:23 for me downloads is outside tmp Jan 21 07:29:24 eFfeM: already dobe Jan 21 07:29:26 eFfeM: already done Jan 21 07:30:52 http://pastebin.com/m59464a97 after rm -rf tmp/stamps/* Jan 21 07:31:40 rm -rf tmp/cache/* Jan 21 07:31:44 no result again Jan 21 07:32:33 you kept tmp/saved_tmpdir ? Jan 21 07:32:39 rm -rf tmp/cross/* Jan 21 07:32:45 yes Jan 21 07:32:50 if not I guess you will get the moved msg Jan 21 07:33:38 if I want to start from scratch I typically do an rm -rf tmp, but as I said my downloads dir is outside tmp Jan 21 07:34:02 forgot how you specify that, might be if you redirect tmp Jan 21 07:34:11 eFfeM: where I configure to get download? Jan 21 07:34:31 not sure will check Jan 21 07:36:52 good morning Jan 21 07:38:15 moorning Jan 21 07:39:07 eFfeM: sorry, pc restart Jan 21 07:40:16 recalcati: np, Jan 21 07:40:35 recalcati: location of downloads dir is specified in DL_DIR var Jan 21 07:41:08 my site.conf says .../conf/site.conf: DL_DIR = "${OE_TOPDIR}/downloads" Jan 21 07:41:40 default is DL_DIR ?= "${TMPDIR}/downloads" (see bitbake.conf) Jan 21 07:42:05 then you can nuke all of tmp Jan 21 07:42:24 export PATH=$OE_HOME/opt/bitbake/bin:$PATH Jan 21 07:42:30 I'm in stalbe branch Jan 21 07:43:47 DL_DIR not defined in opt/bitbake/conf/bitbake.conf Jan 21 07:45:03 sorry... it IS defined Jan 21 07:45:30 i suggest NOT to change it there but define it e.g. in site.conf or local.conf or so Jan 21 07:45:40 i always keep bitbake.conf untouched Jan 21 07:47:16 done Jan 21 07:47:21 rm -rf tmp Jan 21 07:48:11 recalcati: try using KaeilOS method and I'm sure you will feel better http://www.kaeilos.com/?q=download Jan 21 07:49:21 mckoan: thx, I look at .. Jan 21 07:54:00 recalcati: gl Jan 21 07:55:30 eFfeM: ? Jan 21 07:56:24 mckoan: I'll integrate you setup afterwards. first of all I need to recompile all akita and benchmark its compilation time Jan 21 07:57:07 eFfeM: moving downloads works .. thx very very much! Jan 21 07:57:26 now I have to stop forefox in order to have a clean benchmark Jan 21 07:57:35 see you later Jan 21 08:40:49 is there a world equivalent for a specific distro? I'd like to be able to specify a list of packages that should be available in a feed and build them with one command Jan 21 08:42:50 no Jan 21 08:42:57 what people usually do is to create a task package Jan 21 08:43:04 that RDEPENDS on everything you want to see in the feed Jan 21 08:43:11 and then you can just bitbake task-my-super-feed Jan 21 08:44:11 ah Jan 21 08:44:17 thanks :) Jan 21 08:44:35 would task-jlime-repo recipe be accepted into OE git? Jan 21 08:45:47 sure Jan 21 08:45:55 see recipe/tasks Jan 21 08:46:01 there are quite a few distro-specific recipes Jan 21 08:46:59 cool Jan 21 08:56:33 benchmark finished Jan 21 08:57:14 time bitbake zaurus-updater : requires ---- real 49m28.810s Jan 21 08:57:51 after removing rm -rf tmp/* but keeping downloads outside as eFfeM told me Jan 21 09:07:57 recalcati: very impressive, wich PC is it? Jan 21 09:37:01 good morning Jan 21 09:37:32 morning flo Jan 21 09:38:02 hi mickey|office Jan 21 09:47:35 hi lrg, hi XorA Jan 21 09:49:10 morning florian XorA Jan 21 09:50:57 hey florian Jan 21 09:51:02 morning pb__ Jan 21 09:51:05 lrg: WTF you doing out of bed? Jan 21 09:51:20 XorA: back in UK timezone now ;) Jan 21 09:55:28 03Martin Jansa  07org.openembedded.dev * re0f660d44b 10openembedded.git/recipes/mokomaze/mokomaze_0.5.5.bb: mokomaze: bump PR after libsdl-ttf update Jan 21 09:57:55 fsck git-svn isnt in backports Jan 21 09:59:53 hmmm, libdrm2 broken :-( Jan 21 10:01:01 how can I have python code to set a variable in a .conf? I get "unparsed error" from my "python __anonymous" stanza Jan 21 10:04:20 JaMa|Off: seeing any libdrm2 problems? Jan 21 10:06:28 after running bitbake -c menuconfig virtual/kernel. Do I need to so anything else than running a bitbake virtual/kernel to recompile it with my new configuration? Jan 21 10:07:45 XorA: no Jan 21 10:07:53 03Graeme Gregory  07org.openembedded.dev * r7c5dd2b13d 10openembedded.git/: Merge branch 'org.openembedded.dev' of git+ssh://git.openembedded.org/openembedded into org.openembedded.dev Jan 21 10:07:54 03Graeme Gregory  07org.openembedded.dev * r3d1fd2e7f1 10openembedded.git/ (conf/checksums.ini recipes/gnome/libgdata_0.5.1.bb): Jan 21 10:07:54 libgdata_0.5.1.bb : add newer libgdata Jan 21 10:07:54 The only 0.4.0 version wasnt buildable for me. Jan 21 10:08:07 JaMa|Off: curses, intel_bufmgr_fake is failing here :-( Jan 21 10:08:49 XorA: but drm tested mostly only on spitz glamo has own branch.. Jan 21 10:10:24 JaMa|Off: looks like it happens on gentoo as well Jan 21 10:10:38 they have indicated a fix Jan 21 10:10:52 well disable intel on arm :-) Jan 21 10:12:03 XorA: well we have --disable-gallium-intel in mesa-dri_git.bb Jan 21 10:12:16 JaMa|Off: sounds safe then Jan 21 10:12:18 XorA: yep --disable-intel if the buildarch is ARM fixes libdrm 2.4.17 (or seems to) Jan 21 10:12:31 DJW|Home: cool I shall fix and push then Jan 21 10:13:07 Or are you not talking about libdrm, ahh, you are ;-). Just trying it a little more. Jan 21 10:14:29 I assume we arent using intel KMS on arm :-D Jan 21 10:15:01 XorA: despite the number of people who seem to think we are I think it's safe to say, no ;-) Jan 21 10:21:37 yo z.! Jan 21 10:21:44 morning pb_ Jan 21 10:23:38 hi hi Jan 21 10:29:17 to answer my own question I solved it by adding a class and adding that to INHERIT+= in the config Jan 21 10:41:30 DJW|Home: BTW when are you going to send an ssh key to admins :-D Jan 21 10:42:40 XorA: Just leave the box unadminned for long enough and you can use a remote root exploit. Jan 21 11:07:53 mckoan: sorry .... I was away Jan 21 11:08:09 mckoan: impressive because is too much time? Jan 21 11:14:32 recalcati: very fast Jan 21 11:19:01 i need to install apache on an embedded system and i've seriously no idea how tro procede Jan 21 11:19:11 i've already downloaded the source, should I compile them with a cross compiler and copy it to my embedded system ? Jan 21 11:19:33 you should bitbake apache on this channel Jan 21 11:23:34 mckoan: now I move to SSD disk and retry. the pc is Sony Vaio VGN-AW11XU/Q Jan 21 11:24:46 mckoan: I have to go away again to do a clean test Jan 21 11:47:15 hej, can someone explain to me, what 'bbclassextend' does? (it's not in the docs) Jan 21 11:48:20 I hate people that wait less than 30s for an answer Jan 21 11:48:35 XorA: mobile modem ;/ Jan 21 11:48:39 mlip: it automatically does inherit for each class listed in BBCLASSEXTEND :D Jan 21 11:48:47 mlip: no hasles Jan 21 11:48:50 :D Jan 21 11:49:03 thx Jan 21 11:49:07 so you get the basic recipe, then recipe- Jan 21 11:49:19 so recipe and recipe-native in its normal use Jan 21 11:49:29 mlip: please patch the docs :-D Jan 21 11:50:28 oke, so just saving space and generating an additional native recipce Jan 21 11:50:43 XorA: "maybe" (;p) i will Jan 21 11:50:44 XorA: do you know how should bitbake -c clean -b some/recipe.bb with native in BBCLASSEXTEND? I got something like Jan 21 11:50:57 mlip: space and parsing time Jan 21 11:50:58 "no such package virtual:native:some/recipe.bb" Jan 21 11:51:13 JaMa|Off: dont use -b Jan 21 11:51:35 XorA: yes I used without -b later.. but it's a bit slower :) Jan 21 11:52:05 JaMa|Off: I think with RP new cool stuff -b is going to die Jan 21 11:52:20 XorA: and as I'm using this -b usually for fast cleaning of recipe I'm changing atm.. it was quite handy Jan 21 11:52:42 XorA: OK, I can live with that.. Jan 21 12:02:21 Hi, i have a usb-to-ethernet converter, what options should i add in kernel configure to enable it? Jan 21 12:03:04 man gnome stuff is in a bad state Jan 21 12:31:50 03Liam Girdwood  07org.openembedded.dev * r7adc0d6855 10openembedded.git/MAINTAINERS: Jan 21 12:31:50 MAINTAINERS: update my email and add interests + recipies. Jan 21 12:31:50 Signed-off-by: Liam Girdwood Jan 21 12:42:09 03Guo Hongruan  07org.openembedded.dev * r3713a07653 10openembedded.git/conf/distro/micro.conf: Jan 21 12:42:09 Set preferred glibc to 2.10.1 in conf/distro/micro.conf Jan 21 12:42:09 * the preferred version of binutils of conf/distro/micro.conf is 2.19.51 Jan 21 12:42:09 * the preferred version of glibc is 2.6.1 which is defined in sane-toolchain.inc Jan 21 12:42:11 * the ld produced by binutils-2.19.51 reports that it can not find the __begin symbol when linking with glibc-2.6.1 Jan 21 12:42:16 Signed-off-by: Guo Hongruan Jan 21 12:52:37 XorA: Added the patch for libdrm 2.4.17 on ARM to my tree if you want to cherry pick it in http://git.openpandora.org/cgi-bin/gitweb.cgi?p=openembedded.git;a=commit;h=f3244b236955716de3fb73f9ca4dc80c1bad7402. Jan 21 12:53:52 DJWillis: I already got past that now Jan 21 12:54:28 XorA: I forgot to push it this morning, yep, past all that now myself. Was not sure if you had been busy hacking. Jan 21 12:54:46 DJWillis: yours looks neater though, push it to OE :-D Jan 21 12:55:37 Can't right now, i'll get on it later. Jan 21 12:55:44 coolio Jan 21 12:55:58 this chat is not connected to same machine as my OE Jan 21 12:58:05 Can i ask here a question about gkermit and minicom ? Jan 21 13:00:12 melchior, how to use, or about a recipe? Jan 21 13:01:27 how to use to do a file transfer Jan 21 13:17:36 melchior: please explain exactly what do you want to do Jan 21 13:19:37 i'm connected from my pc to an embedded system via minicom and i'd like to transfer a file from my pc to the embedded device Jan 21 13:20:02 so i've tried in the minicom console ctrl-A then R to receive a file Jan 21 13:20:49 then i launch kermit on my pc and tried gkermit -s 'filepath' Jan 21 13:21:13 but nothing happened except gkermit wrote 9 S~' @-#Y3~*!J*0+++J"U1@O Jan 21 13:25:49 melchior: no lan available? Jan 21 13:26:48 i can ping from my pc to the devie but not in the other way and i've no ssh ser/client on the embedded device for the moment as i can't upload files on it Jan 21 13:26:56 melchior: why don't you use minicom on the PC too ? Jan 21 14:02:46 Anyone else who sees a lot of alighment trap messages on armv5te? Jan 21 14:03:11 * JaMa|Off too Jan 21 14:04:17 florian: sometimes Jan 21 14:05:29 I just built with gcc 4.4.1 an eglibc based image for armv5te using arm instructions. And I see really many. Jan 21 14:09:00 * florian tries 4.4.2 Jan 21 15:04:32 morning Jan 21 15:05:50 hey hrw Jan 21 16:25:26 03Marco Cavallini  07org.openembedded.dev * r21b699c064 10openembedded.git/recipes/xinput-calibrator/xinput-calibrator_0.5.0.bb: xinput-calibrator_0.5.0.bb: added line for new patch to original package Jan 21 16:25:37 03Marco Cavallini  07org.openembedded.dev * r4dfad05b99 10openembedded.git/recipes/xinput-calibrator/xinput-calibrator-0.5.0/0002-gui_x11.cpp-Load-font-fixed-when-9x15-fails-in-GuiCa.patch: xinput-calibrator-0.5.0/0002-gui_x11.cpp-Load-font-fixed-when-9x15-fails-in-GuiCa.patch: Load font fixed when 9x15 fails in GuiCalibratorX11 Jan 21 16:28:35 * kergoth grumbles Jan 21 16:29:17 heh Jan 21 16:29:24 not to early over there Jan 21 16:34:37 heh Jan 21 17:36:29 Laibsch: will you be around later this evening? Jan 21 17:36:39 maybe Jan 21 17:36:44 what's up? Jan 21 17:37:08 ok, just to chat about legacy Sharp headers Jan 21 17:37:44 injecting that crap on the top of jffs2 is a legacy of updater.sh Jan 21 17:38:00 I mean jffs2 images Jan 21 17:38:40 ant_work: hey Jan 21 17:38:54 ant_work: were you having issues with pstage lately? Jan 21 17:39:00 hey denix Jan 21 17:39:03 yes Jan 21 17:39:24 some staging-ipks are defaulting to host arch Jan 21 17:40:17 was talking with Tartarus about this and abut moving the 'pstage' dir out of tmp Jan 21 17:40:44 so gave a link to your blog ;) Jan 21 17:40:49 the base class is still not fixed for that Jan 21 17:41:13 I've seen Jan 21 17:41:22 I believe someone was having issues with that change, but it worked for me... Jan 21 17:41:34 just kernels and bootloaders have probs Jan 21 17:41:50 perhaps we should fix hacking that few recipes... Jan 21 17:43:13 ant_work: why do we have that legacy? Jan 21 17:48:54 Laibsch: I think original kernel was checking for that header? Jan 21 17:49:30 for sure, updater.sh cut the header Jan 21 17:50:11 so probably was at the time when updater.sh was a black-box Jan 21 17:50:36 I think hrw|gone would know better than me Jan 21 17:50:43 ah, yea Jan 21 17:50:59 If you make changes, I'd like to have backcompatibility if possible Jan 21 17:51:25 the idea was to make this 'custom' jffs2 only for sharprom-compatible Jan 21 17:51:31 we don't need to continue to support flashing OpenZaurus Jan 21 17:52:00 I'm not sure we need to support flashing images for sharprom-compatible Jan 21 17:52:22 we can flash any image from 2.6 console Jan 21 17:52:26 anyway Jan 21 17:52:44 the reason to have (or at least the reason I'm interested) is to be able to build packages that can be run on an existing SharpROM-based installation Jan 21 17:52:54 If that continues to work, I'm happy Jan 21 17:53:18 no issues, then Jan 21 17:54:03 but if we decide to still use updater sh, this one need to be patched Jan 21 17:54:13 or the new one renamed Jan 21 17:54:39 ok, let's follow on zaurus-devel ML Jan 21 17:54:46 I'll post a couple of patches Jan 21 17:54:50 bbl Jan 21 17:58:14 who was the one setting up the git mirror on repo.or.cz originally? Jan 21 18:04:56 god, debugging problems like this with bitbake is a nightmare Jan 21 18:05:00 we need to do something about htis Jan 21 18:08:58 also, bitbake's messages when encountering a recursion in variable expansion are not very helpful Jan 21 18:09:12 and it should really catch it before hitting the python maximum recursion depth Jan 21 18:34:49 mlip2: did you manage to get the libxml recipe working? Jan 21 18:37:35 eFfeM1: sort of, but no completelly; configure runs smoothly; compile failed due to libxml2 not being build with dockbook and catalog support (fixed now); now i am fighting some weird typemap error ;) Jan 21 18:37:46 (still in compile) Jan 21 18:38:19 ah ok, Jan 21 18:38:32 i'm probably afk most of the evening Jan 21 18:38:55 i will try to get it running later on this evening Jan 21 18:39:01 need some food now ;) Jan 21 18:39:25 i'm rebuilding hte other two as verification, if they are ok do you want me to commit them for you ? Jan 21 18:40:26 sure; did rebuild them already a few times (worked fine); stylizer added some blanks too Jan 21 18:40:32 should i send them again ? Jan 21 18:41:41 yeah if you made style changes that is the best Jan 21 18:42:08 I'll use the email address your are sending it from as committer ok Jan 21 18:42:27 sure Jan 21 18:46:21 eFfeM1: sent, thx Jan 21 18:54:24 mkip2 there you go Jan 21 18:54:30 mlip2 there you go Jan 21 18:54:32 03Michael Lippautz  07org.openembedded.dev * r9de70ca57a 10openembedded.git/recipes/perl/libxml-namespacesupport-perl_1.10.bb: libxml-namespacesupport-perl_1.10.bb: create recipe Jan 21 18:54:33 03Frans Meulenbroeks  07org.openembedded.dev * r33762d41c0 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Jan 21 18:54:44 03Michael Lippautz  07org.openembedded.dev * r6cd92d1545 10openembedded.git/recipes/perl/libxml-sax-perl_0.96.bb: libxml-sax-perl_0.96.bb: new recipe Jan 21 18:55:11 Laibsch: not sure why I still get the merge message, I did set the rebase flag Jan 21 18:55:49 maybe the flag isn't working Jan 21 18:56:53 you could try an explicit "git pull --rebase" Jan 21 18:57:01 probably, actually i thought about using rebase, but then thought the flag would take care of it, Jan 21 18:57:10 should have checked the git log, sorry Jan 21 18:57:48 not a big deal. really Jan 21 18:58:07 recent git all you have to do is "git status" Jan 21 18:58:11 andi t says "1 commit ahead of master" Jan 21 18:58:26 incredibly easy to see at a glance if you'll be pushing something you didn't intend to push Jan 21 18:59:40 kergoth: and is "merge commit" counted in git status? IIRC no Jan 21 18:59:58 hmm, thought it was, but maybe not Jan 21 19:00:20 either way, good policy to never, ever push unless you've checked what you're pushing Jan 21 19:03:17 hmmm gcc 4.4.3 released. anybody want it in OE ? Jan 21 19:04:22 its a bug fix release only I could apply the diff only on top of 4.4.2 Jan 21 19:04:41 who is using 4.4.2 atm" Jan 21 19:04:42 ? Jan 21 19:04:50 minimal and micro Jan 21 19:04:55 oh bite me, bitbake Jan 21 19:04:55 I suspect it would be better to make it seperate Jan 21 19:05:18 yeah, separate is always best, imho Jan 21 19:05:25 makes it easier to see when something was fixed in gcc Jan 21 19:05:29 right so separate it would be Jan 21 19:06:32 kergoth: Did you have a chance to look at the bitbake vim extensions I posted on bitbake-de Jan 21 19:06:41 bitbake-dev Jan 21 19:08:11 haven't had a chance. if it works, and improves it, i'd say we should just check the thing in. i'll add it to my todo, should get to it tonight or this weekend Jan 21 19:08:17 unless rp or someone else does Jan 21 19:09:47 ok Jan 21 19:10:02 * kergoth 's swamped lately Jan 21 19:11:24 khem: and SHR is using that through sane-toolchain.inc Jan 21 19:13:57 re Jan 21 19:14:12 JaMa: cool thx Jan 21 19:14:32 JaMa: I need to keep in mind next time sane-toolchain versions need change Jan 21 19:22:38 03Aleksey Makarov  07org.openembedded.dev * reaf451fc3e 10openembedded.git/classes/kernel.bbclass: Jan 21 19:22:38 kernel.bbclass: fix kernel build, version >= 2.6.33-rc1, arch x86 Jan 21 19:22:38 * Apply the same change as c0a3e91845fad1cdda723bdcf962de8bc01f8a2d for Jan 21 19:22:38 x86 path Jan 21 19:22:38 * A check was missed in the x86 path of the stage method. Jan 21 19:22:41 Signed-off-by: Martin Jansa Jan 21 19:22:54 03Dallas Foley  07org.openembedded.dev * r728ce9aa3d 10openembedded.git/recipes/dri/libdrm_2.4.17.bb: Jan 21 19:22:54 libdrm_2.4.17: disable intel for non-x86 Jan 21 19:22:54 * older version 2.4.11 builds OK on arm, but not 2.4.17 Jan 21 19:22:54 * someone with x86 and intel, please test Jan 21 19:22:54 Signed-off-by: Martin Jansa Jan 21 19:24:16 I am interested in embedded system developing specifically using powerpc architecture and any embedded linux distro. How to efficiently learn related topic . I am quite good in programming and I understand embedded issues but I am newbe in ppc and rather less experienced linux user/programmer. Jan 21 19:24:16 have you any suggestions? Jan 21 19:25:06 www.openembedded.org Jan 21 19:26:10 XorA: image built :) Jan 21 19:31:45 for example I have started anylasing kernel source code using very old release early 2.2 because I have assuned that it is simpler to begin( because of less code amount). Is this good approach? Jan 21 19:33:43 kergoth: thanks for the git status suggestion; actually I hardly ever double check what I push since i always push immediately after a commit, unpushed cmmits only can give merge problems, and if the patch is not yet good enough to be pushed, it is not good enough to be committed Jan 21 19:33:52 03Chris Larson  07org.openembedded.dev * r3192d261f8 10openembedded.git/recipes/konqueror/konqueror-embedded_20030705.bb: Jan 21 19:33:52 konq-e 20030705: kill the exec_prefix var hack Jan 21 19:33:52 Signed-off-by: Chris Larson Jan 21 19:34:08 eFfeM1: strongly, strongly disagree. Jan 21 19:34:46 much better to commit early and often, to keep changes separated, and retain history, and then prep for upstream before you push or send-email Jan 21 19:34:56 can always rebase -i and squash em or whatever Jan 21 19:35:39 kergoth: typically my changes are single file/recipe so they are stand alone and can be pushed when done Jan 21 19:35:59 which isn't an excuse to not learn to use the scm tool well :P Jan 21 19:36:07 :-) Jan 21 19:36:53 Question: Are there any known good collie rootfses? Jan 21 19:37:54 keeping changes separate ofc is good, pushing soon (of course without hampering quality) is also good as it avoids that others invent the same wheel Jan 21 19:38:12 but for complex changes I fully agree with you Jan 21 19:40:22 pushing soon doesn't necessarily require that the changes not be complex, since you can push to another branch for review and then rebase onto master when ready.. you'll hear no argument from me on pushing early Jan 21 19:40:53 :-) Jan 21 19:41:23 I'll try to improve :) suffer with me ;-) Jan 21 19:41:28 we all do :) Jan 21 19:41:35 try to improve that is, not suffer Jan 21 19:41:39 though maybe suffer too, come to think of it.. Jan 21 19:41:40 some days.. Jan 21 19:44:03 son of a bitch.. Jan 21 19:44:21 * kergoth has a task level dependency loop that bitbake isn't being very helpful with Jan 21 19:44:29 Aborted dependency loops search after 10 matches. Jan 21 19:44:32 Jan 21 19:52:27 eFfeM1: I started to commit every single change, just after I decided few times to split already commited stuff to more commits (sqashing together is much more simplier or I did something wrong for split) Jan 21 19:53:42 eFfeM1: btw does that rebase flag or git pull --rebase work that it first fetches all new stuff and then apply your changes on top of that + you have to solve merge conflicts? Jan 21 19:54:08 eFfeM1: or does it say, that there are conflicts that you have to solve and then it rebase? Jan 21 19:54:18 splitting isn't too terrible if you use reset (or rebase) + git add -p, but it gets annoying on occasion since add -p only adds hunks.. you can split them sometimes, but it can't add line by line into the index Jan 21 19:54:27 JaMa: yes, thats what rebase is. Jan 21 19:54:53 03Simon Busch  07org.openembedded.dev * r0854ce6638 10openembedded.git/recipes/preboot/linux-preboot.inc: preboot: remove defconfig requirement Jan 21 19:54:53 it applies commit by commit, if one has problems, drops you back to your shell to fix them, then git add the fixed files and git rebase --continue Jan 21 19:55:44 kergoth: the first one, right? ahh I need that.. I always git pull, then I cannot git rebase -i upstream_branch unless I resolve conflicts (somehow) and then I call rebase and have to solve conflicts (from the other side) Jan 21 19:56:15 yikes, thats way too much work, and duplicated effort :) Jan 21 19:56:36 kergoth: I know how rebase works when called on already pulled branch (that's what I do after every pull :)) Jan 21 19:56:50 * JaMa feels a bit stupid for not checking this before :) Jan 21 19:57:03 pull is fetch + merge Jan 21 19:57:08 if all you want is fetch, run fetch Jan 21 19:57:16 git fetch origin; git rebase origin/org.openembedded.dev Jan 21 19:57:20 is basically pull --rebase Jan 21 19:57:26 :) Jan 21 19:58:53 kergoth: and I have to check that split as you said ( I used git rebase -i, marked commit for split as edit, then git reset HEAD^1 for files I want to move to new commit and git commit --amend first one and after that commit un staged changes as new one) Jan 21 19:59:16 kergoth: which was usable for splitting changes in separate files Jan 21 19:59:49 yeah, that'd work fine if the files are separate Jan 21 20:00:02 git add -i / add -p lets you add individual hunks of changes to the index, instead of teh whole file Jan 21 20:00:06 really really handy at times Jan 21 20:00:50 but still squashing is way faster for separate and not separate changes :) Jan 21 20:01:01 indeed Jan 21 20:01:07 thanks for tips! Jan 21 20:01:33 np.. downside & upside to git.. some things are obscure, but extremely powerful Jan 21 20:04:21 yeah once I was really surprised how clever rebase was, I had patch changing few files and later another one moving them to different directory, then I decided to move them first and then change them.. so I changed order of patches in rebase -i and expected conflict when 2nd patch try to change those files in old directory.. but git was clever and changed them right in new place Jan 21 20:05:27 nice Jan 21 20:08:59 what would be helpful is form someone to write an article explaining how to commit something in git, and send an email with the commit to the list Jan 21 20:16:11 * Laibsch strongly suggests the use of tools like qgit, etc. to look at how rebase and friends change history Jan 21 20:16:24 a picture says more than a thousand words ;-) Jan 21 20:19:32 * JaMa too often runs git without X (from work without X11Forwading because slow connection) Jan 21 20:31:59 qgit is really helpful Jan 21 20:33:02 stgit remains the thin thread that keeps me tied to reality Jan 21 20:42:23 03Martin Jansa  07org.openembedded.dev * rb6f9a84fd9 10openembedded.git/recipes/tasks/task-shr-minimal.bb: task-shr-minimal: add xinput-calibrator Jan 21 20:44:11 JaMa: do understand what git rebase does it's sufficient to look at qgit three or four times when doing a rebase Jan 21 20:44:35 once you've done that, you'll start checking it from time to time, though ;-) Jan 21 20:45:39 :) Jan 21 21:01:47 * khem uses giggle for visual gitting Jan 21 21:04:20 does anyone know if openembedded recipes can be used to create dpkgs (debian packages for ubuntu)? Jan 21 21:04:38 awozniak, I tried applying the patch from patchwork, but it is corrupt Jan 21 21:04:52 Crofton_|work: which one? Jan 21 21:04:56 or do we have to move from *.bb recipes to make files? Jan 21 21:05:01 I applied one you sent me with "patch", if olsrd builds I'll push that Jan 21 21:05:23 the one closest to the top Jan 21 21:05:30 Crofton_|work: I used it few times before and twice today and it worked ok Jan 21 21:05:52 the patch from patchwork for olsrd? Jan 21 21:06:05 or the script? Jan 21 21:06:26 I suspect a problem with the specific patch Jan 21 21:06:32 what is olsrd? :) Jan 21 21:06:46 a routing thingy for mobile ad hoc networks Jan 21 21:06:54 git-am.sh script worked and patches were ok for me Jan 21 21:06:55 I know people that knwo what it does :) Jan 21 21:07:00 yeah Jan 21 21:07:11 I suspect a problem with this speciofic patch Jan 21 21:07:19 ah ok, sorry Jan 21 21:07:40 we really should write a guide on creating the patch and sending it to the list Jan 21 21:07:49 so stuff in pw just works via the script Jan 21 21:08:07 I suspect this one was linewrapped or something Jan 21 21:08:41 yes.. even syntax highlight in patchwork is broken on we lines.. Jan 21 21:08:57 it's wrapped for 80 cols Jan 21 21:09:33 he sent me something in email that applied, hopefully I do bork the author bit Jan 21 21:10:09 I just want to get it in, since I sort of care about MANET stuff a little Jan 21 21:11:13 heh my PhD theme was "Multicast routing in ad-hoc wireless networks" so maybe I should care too.. but as I quit PhD study I also quit ad-hoc interest .. Jan 21 21:13:32 Which school? Jan 21 21:13:56 Czech Technical University in Prague Jan 21 21:15:31 JaMa: hmmm Pilsner comes from Czech isnt it :) Jan 21 21:15:55 khem: yes, czech most famous product Jan 21 21:16:18 Not familiar with them, I'll have to ask my freinds if they have heard of it Jan 21 21:16:34 khem: maybe hockey team few times.. but beer is known always :) Jan 21 21:17:26 Crofton_|work: well PhD wasn't organized there as well as it should be :/ Jan 21 21:17:41 and also JaMa oh soccer too :) Jan 21 21:18:21 Crofton_|work: or at least it didn't satisfy me at all Jan 21 21:18:23 and skoda Jan 21 21:18:48 khem: great, you have 3 points :) Jan 21 21:18:49 my czech friend in munich wud only buy skoda Jan 21 21:19:03 and I would only buy BMW .. Jan 21 21:19:39 heh not me anymore since alternator died in the middle of night and it was snowing Jan 21 21:19:50 bad memories Jan 21 21:21:02 jo Jan 21 21:21:17 my friends bmw can't go in the snow, even if it is thinking of snowing it slides off the road Jan 21 21:21:21 gm Jan 21 21:21:38 heh real wheel drive ? Jan 21 21:21:48 rear I meant Jan 21 21:21:49 I think so Jan 21 21:21:50 crofton haha yes heckdrive suckz badly Jan 21 21:21:53 in snow and ice Jan 21 21:22:12 * khem is happy with VW Jan 21 21:22:14 hi khem Jan 21 21:22:14 but BMW has also great xDrive :) Jan 21 21:22:20 hi woglinde Jan 21 21:22:26 Crofton_|work: I think you mean its "fun" in the snow ;-) Jan 21 21:22:29 jama with x5? Jan 21 21:22:36 Dad's Mazda is like that... Jan 21 21:22:37 all suv suckz Jan 21 21:22:42 hi rp Jan 21 21:22:48 hi woglinde Jan 21 21:22:51 * JaMa not having BMW now still saving money for semi-new 535 Jan 21 21:23:13 jo jama Jan 21 21:23:13 * RP just has the rear wheel drive MGB Jan 21 21:23:24 * woglinde dont save for money for car Jan 21 21:23:26 which is also interesting one snow Jan 21 21:23:32 house is better Jan 21 21:25:20 woglinde: I'm just waiting for incomming monthly instalments, because I borrowed most of money to sister for flat.. Jan 21 21:29:57 * pb__ just has the broken Audi, doh Jan 21 21:30:04 why isn't -v the default for git cherry? Jan 21 21:30:11 rofl Jan 21 21:30:25 pb__, the Audi is still broke? Jan 21 21:30:28 yeah Jan 21 21:30:32 bummer Jan 21 21:30:49 they reckon it should be fixed by next week though, just waiting for a new throttle body Jan 21 21:30:53 pb bah dodnt fixed until now? Jan 21 21:31:02 woglinde: no, this german engineering is the pits Jan 21 21:33:09 in the meantime the garage has loaned me a toyota avensis, which is possibly the worst car I have ever driven. Jan 21 21:33:28 pb hm why? Jan 21 21:33:40 and which avensis the newest? Jan 21 21:33:56 no, it's quite an old one. 1998-ish I think. Jan 21 21:34:06 oh Jan 21 21:34:11 yeah Jan 21 21:34:16 they arent that good Jan 21 21:34:29 latest and prevoius model are quite okay Jan 21 21:34:34 it steers roughly like an oil tanker, except that you need to hold the wheel at about 2 o'clock in order to go down the road in a straight line. Jan 21 21:35:03 zero to 100 km/h in something like six weeks Jan 21 21:35:29 haha Jan 21 21:36:14 but yeah, I can imagine the newer ones are probably okay Jan 21 21:37:05 the only advantage is that it has loads of luggage space: I could probably fit my house into the boot with a bit of room to spare. Jan 21 21:37:59 anyway, fingers crossed, I should have the audi back next week. Jan 21 21:38:20 just have to hope nothing else goes wrong with it :-} Jan 21 21:39:52 pb__: Sounds like someone has bounced the steering off a few kerbs Jan 21 21:40:16 * RP has just found he needs a new inner tube for the CRM and that the frame might not be bent, the forks may be broken instead Jan 21 21:40:34 rp *g* Jan 21 21:40:49 motocross is a nice hobby Jan 21 21:41:58 woglinde: I do not do MX, that is just dangerous ;-) Jan 21 21:42:13 woglinde: Trail riding - all public roads and the odd rally Jan 21 21:42:21 *g* but your crm looks like you did Jan 21 21:42:39 woglinde: The public roads have no tarmac ;-) Jan 21 21:43:05 woglinde: The bike wouldn't look out of place on a MX track admittedly apart from the lights Jan 21 21:43:15 hm everyday a new word Jan 21 21:43:17 tarmac Jan 21 21:43:19 hey RP Jan 21 21:43:22 will remember it Jan 21 21:43:37 RP: has anyone ever tried making it so that some tasks can be kicked off and not count against BB_NUMBER_THREADS ? Jan 21 21:43:43 was funny this morning to see how the quick fixed holes in the tarmac here Jan 21 21:43:52 s/the/they Jan 21 21:45:39 RP: I've just noticed that I have one more simple patch for bitbake (for bzr pull) http://www.mail-archive.com/bitbake-dev@lists.berlios.de/msg00621.html (tested with recipes/openmoko-3rdparty/babiloo-efl_bzr.bb) Jan 21 21:46:33 Tartarus: It would be trivial to do Jan 21 21:46:54 RP, ah good Jan 21 21:46:57 Tartarus: The scheduler is actually controllable as a plugable module Jan 21 21:47:21 Oh? Hmm Jan 21 21:47:29 JaMa: right, I still have that in my queue Jan 21 21:47:29 That leaves open more complex changes then Jan 21 21:47:39 RP: ok, thanks Jan 21 21:47:53 I was thinking of just doing some stuff, like do_configure_qa and some other always quick tasks "free" Jan 21 21:47:53 Tartarus: It was always intended that we'd have a nicer scheduler, just nobody has written one Jan 21 21:48:37 Tartarus: The only tricky bit is ensuring we don't teach bitbake something too specific about OE :/ Jan 21 21:48:44 Well Jan 21 21:48:50 If it's a plugin, why not? Jan 21 21:49:46 RP: and is it possible that env variable PYTHONPATH is now filtered a bit different in master than was in 1.8? I had to add export to few places to build packages with new bitbake (it tried to use x86_64 python from buildhost instead staging) Jan 21 21:49:49 Tartarus: The configure_qa thing assumes you have insanity running Jan 21 21:49:58 RP: yes, that too Jan 21 21:50:04 But even do_install Jan 21 21:50:17 Tartarus: I'm just saying it needs some thought :) Jan 21 21:50:24 RP: yeah, that too :) Jan 21 21:50:49 JaMa: Recent 1.8 should be the same Jan 21 21:50:50 Just coming up with some ideas to throw a person at Jan 21 21:50:59 JaMa: but I've seen that in Poky too, its worrying Jan 21 21:51:20 Tartarus: Did you try the -nativesdk stuff? Jan 21 21:51:27 Tartarus: and BBCLASSEXTEND? Jan 21 21:51:57 I think I read it at some point, it sounded good and then I had to get back to work work stuff Jan 21 21:52:01 push it :) Jan 21 21:52:11 Tartarus: I did ;-) Jan 21 21:52:20 RP: it was missing also in distutils, so I'll try revert my patch and test it with 1.10 Jan 21 21:52:20 RP: Have we moved to 1.10 in OE yet ? Jan 21 21:52:22 Tartarus: OE hasn't quite converted yet though, Poky has Jan 21 21:52:51 khem: We don't have 1.10 yet Jan 21 21:53:07 khem: Current plan is to make master rock, expose to people, fix the bugs, call it 1.10 Jan 21 21:53:19 RP: right but we have branch in bitbake Jan 21 21:53:24 khem: and as a measure of confidence, Poky now uses bitbake master Jan 21 21:53:34 khem: Yes, I'm wiping that, it was a mistake Jan 21 21:53:46 bitbake master all the way now Jan 21 21:53:47 RP: ok that will be good Jan 21 21:53:52 it confused me a bit Jan 21 21:54:00 khem: sorry, I have changed plans a bit Jan 21 21:54:11 khem: I like the new plan more though :) Jan 21 21:54:14 1.8 is good for BBCLASSEXTEND? Jan 21 21:54:22 khem: yes Jan 21 21:54:27 I see Jan 21 21:54:28 khem yes Jan 21 21:54:34 we should start converting Jan 21 21:54:36 I have been using master for quite long Jan 21 21:54:41 then started using 1.10 Jan 21 21:54:41 rp did it at all for poky Jan 21 21:54:57 khem: Please do use master, I need to know how its looking Jan 21 21:54:59 I thought that would be where next release coming from Jan 21 21:55:11 no worries I will do that Jan 21 21:55:18 and jama has is first package with BBCLASSEXTEND done too Jan 21 21:55:25 his Jan 21 21:55:26 args Jan 21 21:55:32 damn typos Jan 21 21:55:42 I'd love to see BBCLASSEXTEND used more in OE, particularly with nativesdk Jan 21 21:55:56 rp I have gettext working Jan 21 21:56:01 I think I should commit it Jan 21 21:56:16 woglinde: wow do you mean with BBCLASSEXTEND ? Jan 21 21:56:17 woglinde: ported from Poky? Jan 21 21:56:26 rp yes Jan 21 21:56:35 woglinde: cool - did it need any changes? Jan 21 21:56:37 and split out libintl into own package Jan 21 21:56:40 no Jan 21 21:56:44 just worked Jan 21 21:56:52 woglinde: please push :) Jan 21 21:57:05 hm its only tested here Jan 21 21:57:05 woglinde: also, I added some checks to insanity.bbclass in poky for gettext Jan 21 21:57:10 woglinde: Just do it ;-) Jan 21 21:57:12 RP: IIRC you also said, that you would like to see checksum check in fetcher code in bitbake, right? Do you expect it to be compatible with that syntax? SRC_URI[rc4.md5sum] = "checksum" Jan 21 21:57:16 rp okay Jan 21 21:57:22 JaMa: yes Jan 21 21:57:57 woglinde: You might be interested in those sanity checks although OE will be missing loads of gettext dependencies :/ Jan 21 21:58:36 JaMa: I want to see it using python modules for the checks, not the standalone tools if they're available Jan 21 21:58:56 * JaMa would like to use in-recipe checksums more.. but it would be nice to have some policy for SRC_URI naming like SRC_URI[archive.md5sum] named the same in every recipe with just one archive? Jan 21 21:58:58 rp gettext.bbclass is not in best shape Jan 21 21:59:00 but it works Jan 21 21:59:57 because if someone write checksums.ini -> in-recipe checksums script later.. then would be great to have consistent names if possible Jan 21 22:00:20 woglinde: Poky has some changes to it to improve it Jan 21 22:00:33 rp oh Jan 21 22:00:38 I will look at it Jan 21 22:00:52 rp do you have some spare time to put it in oe? Jan 21 22:01:03 my time for oe will be short the next months Jan 21 22:02:02 woglinde: It varies. At the moment I'm putting it into creating one bitbake Jan 21 22:02:15 * woglinde wished we could find some money to hire RP for oe work from intel Jan 21 22:02:15 woglinde: merging in zecke's parsing work and so on Jan 21 22:02:37 woglinde: At least Intel are letting me work with Poky ;-) Jan 21 22:04:38 okay I will push my gettext stuff now Jan 21 22:04:57 hm I have to look not to brake 0.14 Jan 21 22:05:57 Hi Jan 21 22:07:00 gnight all Jan 21 22:07:54 is there a way to compile for a 64 bit target ? Jan 21 22:09:04 sure Jan 21 22:09:16 pwgen: sure, you can set TARGET_ARCH=alpha (or x86-64, or sparc64, or...) in your local.conf Jan 21 22:09:46 x86-64 is now possible ... cool Jan 21 22:10:11 actually, I think it's spelt "x86_64" now I come to think of it. but yeah, I think it is supported. Jan 21 22:12:50 but there is no MACHINE definition for an x86_64 bit target Jan 21 22:13:36 there have been patches posted to the list; you could use those. but a MACHINE definition is not mandatory anyway. Jan 21 22:14:10 i will try a look after the qemux86 image hopefully builds .. Jan 21 22:15:09 hi likewise Jan 21 22:15:43 03Paul Menzel  07org.openembedded.dev * rb84a8334ff 10openembedded.git/recipes/libcap/libcap_1.10.bb: (log message trimmed) Jan 21 22:15:43 libcap: Append `-fPIC` to `CFLAGS` to fix compilation on newer binutils/ld. Jan 21 22:15:43 Doing `bitbake libcap` with Jan 21 22:15:43 PREFERRED_VERSION_binutils = "2.20" Jan 21 22:15:45 PREFERRED_VERSION_binutils-cross = "2.20" Jan 21 22:15:47 PREFERRED_VERSION_binutils-cross-sdk = "2.20" Jan 21 22:15:49 PREFERRED_VERSION_gcc = "4.4.2" Jan 21 22:16:28 uh who pushed paul's patch? Jan 21 22:17:31 ~hail pb Jan 21 22:17:32 * ibot bows down to pb and chants, "I'M NOT WORTHY!!" Jan 21 22:24:25 03Khem Raj  07org.openembedded.dev * ra92eb1effa 10openembedded.git/recipes/uclibc/ (3 files in 2 dirs): Jan 21 22:24:25 uclibc-nptl: Fix Global dynmaic TLS loads. Jan 21 22:24:25 * Update to latest tip of branch. Jan 21 22:24:25 * Delete the patches which are now upstream. Jan 21 22:24:25 * Hide __libc_errno and __libc_h_errno and __libc_res. Jan 21 22:24:27 Signed-off-by: Khem Raj Jan 21 22:28:26 jo ant Jan 21 22:28:49 hey Jan 21 22:28:52 khem: I'm moving to uclibc-nptl :D Jan 21 22:29:16 ant__: Good you wont be disappointed Jan 21 22:29:29 hey ant__ Jan 21 22:29:37 and I will be happy to fix any issues you run into Jan 21 22:29:42 he I suppose yesterday TLS was still an issue Jan 21 22:29:58 no its only for mips Jan 21 22:30:02 arm works fine Jan 21 22:30:07 ah, ok Jan 21 22:30:13 pwgen: hello there Jan 21 22:30:44 khem whats 0.9.30.2 status? Jan 21 22:31:34 getting the angstrom-gnome packges to run now i have a ~2G image .... Jan 21 22:31:40 hi woglinde, fooks Jan 21 22:31:52 woglinde: I have the port but not tested Jan 21 22:32:04 woglinde: I want to boot on arm/mips/x86 qemu Jan 21 22:32:07 pwgen: ahem..2GiB ? Jan 21 22:32:08 pwgen hihi yes Jan 21 22:32:13 ant sure Jan 21 22:32:16 OMG Jan 21 22:32:20 and run uclibc regression Jan 21 22:32:32 hi likewise Jan 21 22:32:35 long time Jan 21 22:32:36 khem: hi khem Jan 21 22:32:48 khem: what is uclibc regression, a built-in self-test? Jan 21 22:32:54 likewise: Do you know if some powerpc works well on qemu Jan 21 22:32:55 ant some angstroem tasks nearly pull in all recpies Jan 21 22:33:00 ant__ angstrom gnome + illume + sdk + many other tools . a nice useful akita ... Jan 21 22:33:07 khem: been busy with device driver stuff, little OE lately Jan 21 22:33:07 likewise: its the uclibc testsuite Jan 21 22:33:09 I meant Jan 21 22:33:24 pwgen: have you followed the buzz aboit touchscreen calibration? Jan 21 22:33:33 no Jan 21 22:33:44 khem: can I help test uclibc on ppc? Jan 21 22:33:46 can we gather any progress? on c7x0 default pointercal is a bit wrong... Jan 21 22:33:50 likewise: sure Jan 21 22:33:57 -nptl that is Jan 21 22:34:07 likewise: i would like to have qemu working for ppc too some day Jan 21 22:34:16 ant__ now i have a way to emulate the right mouse click to access some clasw features ...(:-(( Jan 21 22:34:24 very good Jan 21 22:34:29 .. looking for a way ... Jan 21 22:34:35 ahh Jan 21 22:34:49 likewise: hmmm I havent done any ppc work but there was someone who posted some ppc nptl work for uclibc Jan 21 22:35:08 likewise: right now its known to work for arm,sparc,mips,sh Jan 21 22:35:17 pwgen: xinput-calibrator Jan 21 22:35:17 ant__ maybe the cal/address/mail buttons may be helpful Jan 21 22:35:19 and x86 to a certain extent Jan 21 22:35:31 pwgen: libgtkstylus or evtouch (urgg) or xf86-tslib and a hal/xconf hack for right click emu.? Jan 21 22:35:53 hi djwillis Jan 21 22:35:57 khem: it has been a long road, but I must say kudos to those involved Jan 21 22:36:07 hey woglinde Jan 21 22:36:20 or i use my old stoway userspace driver tool ... Jan 21 22:36:39 likewise: yeah too few dev too much work Jan 21 22:37:00 likewise: .31 should see nptl on some arches Jan 21 22:37:15 ppc is one big one missing Jan 21 22:37:30 but I think it could also be in Jan 21 22:39:06 ppc is too often big-iron-only, not much interest in true embedded at first point. Jan 21 22:39:16 right Jan 21 22:39:31 ant__ 2320 packages installed Jan 21 22:39:33 but some guys have it in TVs Jan 21 22:41:15 it is on the outer ends of the spectrum: either embedded, or 4096 cpu systems :-) Jan 21 22:42:36 there a lots of PPC boards http://penguinppc64.org/embedded/howto/ Jan 21 22:43:47 I did not know there are still Z80 in production... Jan 21 22:44:07 ant haha Jan 21 22:44:16 there are lot more in production Jan 21 22:44:17 8051 Jan 21 22:44:21 *FG* programmed in a GAL ? Jan 21 22:44:23 for instance Jan 21 22:45:14 Z80 TRS80 model 4 TRSDOS 0.004 BogoMips ...... Jan 21 22:47:46 ant__ 2.6.32+rcX doesn't boot. maybe i set MACHINE to "akita" instead of "spitz" and there is no defconfig for akita Jan 21 22:48:44 oh, I just added akita fotr the linux-kexecboot, not yet for linux Jan 21 22:50:46 JaMa: btw, worth to move to -rc4 ? Jan 21 22:51:14 I think the real change will be with 2.6.34. lot of pending patches. Jan 21 22:51:23 (for arm) Jan 21 22:52:41 khem: (aside: git send-email can accept format-patch/rev-list args, at least in recent versions (not sure when it was added, or if i just didnt notice it). can do git send-email HEAD~3.. or what have you, instead of having to send the files) Jan 21 22:52:46 quite handy Jan 21 22:53:24 kergoth: yes indeed Jan 21 22:54:08 * kergoth is finally happy with his patch sending setup, set up multiple identities in the ~/.gitconfig for his gmail and work addresses, use one send-email command and off it goes :) Jan 21 22:55:02 yep [sendmail] --to == Jan 21 22:55:05 is handy Jan 21 22:55:50 03Eric BĂ©nard  07org.openembedded.dev * r3020a58f3e 10openembedded.git/recipes/barebox/ (barebox.inc barebox_2009.12.bb barebox_git.bb): Jan 21 22:55:50 barebox: Add recipe for this new bootloader Jan 21 22:55:50 * barebox is "a bootloader that inherits the best of U-Boot and the Linux kernel" Jan 21 22:55:50 * add recipe for v2009.12.0 (last stable) Jan 21 22:55:52 * add recipe for git repository Jan 21 22:55:54 Signed-off-by: Eric Benard Jan 21 22:55:56 Signed-off-by: Khem Raj Jan 21 22:56:34 ant__: I moved already.. but it's quite the same as rc3.. /me sleeping Jan 21 22:56:34 mmm machine.conf for eee701(I will commit as soon as possible) should have a MACHINE_DRI_MODULES,should I put i915 or intel? also is how is mesa-dri choosen over mesa-xlib? Jan 21 22:56:34 I'm trying to build with intel now Jan 21 22:56:40 I built with i915 before Jan 21 22:56:58 JaMa: as I thought, thx Jan 21 22:57:40 ah..Barebox (formerly known as u-boot-v2) Jan 21 22:57:44 I see Jan 21 22:58:49 uboot-v2 is used in production? Jan 21 22:59:30 nite jama Jan 21 23:01:03 15s of lag...ouch Jan 21 23:01:04 ping RP Jan 21 23:02:37 03Nicolas Ferre  07org.openembedded.dev * ra0833a8385 10openembedded.git/recipes/xserver-common/ (files/Xserver-at91.patch xserver-common_1.25.bb): Jan 21 23:02:37 xserver-common: add SAM9G10/G45/9RL to Xserver file Jan 21 23:02:37 Adding some Atmel AT91 chips to Xserver file. Jan 21 23:02:37 Signed-off-by: Nicolas Ferre Jan 21 23:02:37 Signed-off-by: Khem Raj Jan 21 23:02:41 03Nicolas Ferre  07org.openembedded.dev * r9f7e313d91 10openembedded.git/conf/machine/ (5 files): (log message trimmed) Jan 21 23:02:44 Update of ATMEL board configuration files Jan 21 23:02:46 Correction of at91sam9260ek and at91sam9g20ek configuration files: Jan 21 23:02:48 Xserver is not needed. At91sam9rlek configuration file: no usbhost. Jan 21 23:02:50 Adding of two new boards: at91sam9g10ek and at91sam9g45ekes. "es" stands Jan 21 23:02:52 for Engineering Sample and correspond to the marking that is on the board. Jan 21 23:02:54 Signed-off-by: Nicolas Ferre Jan 21 23:02:56 03Nicolas Ferre  07org.openembedded.dev * r233e56f1a4 10openembedded.git/conf/distro/include/sane-feed.inc: Jan 21 23:02:59 sane-feed.inc: adding some ATMEL boards Jan 21 23:03:01 Add some armv5te entries for ATMEL -EK boards. Jan 21 23:03:03 Signed-off-by: Nicolas Ferre Jan 21 23:03:05 Signed-off-by: Khem Raj Jan 21 23:04:15 having trouble building ti-dsplink-module - tinderbox says a newer version than whats in git was built - how does that make sense? perhaps I don't understand what tinderbox is Jan 21 23:04:55 tharvey: is the newer version also in recipes/... ? Jan 21 23:05:15 tharvey is automagic log upload when something fails Jan 21 23:05:24 tharvey: tinderbox collects builds from users. If a user locally has a newer module version (not committed yet) you will see it in tinderbox. Jan 21 23:05:39 gah, i need to stop switching to every irc channel with activity every few seconds even if i'm not reading it, just to make the little indicators go away Jan 21 23:06:18 kergoth: information-savy, aye? Jan 21 23:06:26 http://tinderbox.openembedded.org/packages/ti-dsplink-module/ shows denix built version 1_64 the other day, yet recipes/ti shows latest version is 1_61_03 ? Jan 21 23:06:49 its just like a nervous tick or something. just keep doing it for no reason Jan 21 23:06:49 tharvey where in oe or at ti? Jan 21 23:07:29 kergoth: heh, maybe you should leave those channels Jan 21 23:07:32 woglinde, don't understand your question? are there two sets of recipes? Jan 21 23:07:32 tharvey: that version is not publicly available yet Jan 21 23:07:51 denix ;) Jan 21 23:07:56 pb__: perhaps.. but once in a while there's a useful discussion. i need a noise filter :) Jan 21 23:07:58 I'm testing it now, once finished it will be pushed to OE Jan 21 23:08:19 denix what is changed? Jan 21 23:08:35 woglinde, what are you building from? I see you've built some flavors of it as well Jan 21 23:08:38 *sigh* I dont have the time to test the dsp stuff Jan 21 23:08:54 denix, is there a more up to date tree with this stuff? Jan 21 23:08:58 tharvey I downloaded stuff from ti Jan 21 23:09:05 and wrote some recipes here Jan 21 23:09:13 wandern, you mean arago? Jan 21 23:09:34 tharvey yes denix makes arago Jan 21 23:10:00 tharvey: I am currently testing the next branch of arago... Jan 21 23:10:00 not sure what git tree I should be using... oe-dev, gumstix-oe, or arago - confused :) Jan 21 23:10:17 trying to use gstreamer-ti on overo Jan 21 23:10:33 woglinde, the vendor will have a working recipe on the day the sw is widely available :) Jan 21 23:11:23 Laibsch: in compatibility-providers.conf: PREFERRED_PROVIDER_virtual/psplash ?= "psplash-angstrom" Jan 21 23:11:32 crofton?= Jan 21 23:11:33 tharvey: if you can wait little bit, those new recipes will be pushed to arago master and oe.dev as well Jan 21 23:11:57 woglinde, wrt to dsplink Jan 21 23:11:57 denix, do you think 'a little bit' is a matter of days, or longer? Jan 21 23:12:08 crofton there were some newer versions on the ti site Jan 21 23:12:17 but I only tried to compile it Jan 21 23:12:26 tharvey: but if you need cutting edge, you are welcome to use our wip branches Jan 21 23:12:50 denix, happy to test anything that builds Jan 21 23:12:53 the problem with that is some of the recipes are for the unreleased components Jan 21 23:13:24 so, we have to wait for those become available from the download site, before pushing recipes for them Jan 21 23:13:29 crofton but anyway the whole dsp stuff needs cleanup at ti side Jan 21 23:13:36 woglinde: hi Jan 21 23:13:50 ant__: Are you sure that file is included in minimal? Jan 21 23:13:56 denix, what about building with slightly dated components? Jan 21 23:14:00 rp hm how can I compile gettext-native-0.14.1 now with BBCLASSEXTEND? Jan 21 23:14:13 ant__: you said you wanted to discuss the Z updater.sh thing? Or are we done? Jan 21 23:14:20 woglinde: bitbake gettext-native? Jan 21 23:14:20 hm..right..but I get that NOTE byuilding minimal Jan 21 23:14:23 tharvey: matter of days, maybe couple weeks at most, hopefully Jan 21 23:14:29 ant__: I know Jan 21 23:14:31 woglinde: I appreciate there is a small problem with the -b option :/ Jan 21 23:14:32 rp that would build 0.17 Jan 21 23:14:35 Laibsch: we have RP :) Jan 21 23:15:01 bitbake -c build gettext-native-0.14.1 says is not avail Jan 21 23:15:05 ant__: as I said a couple of times now. I don't consider that such a major thing. And I certainly wouldn't want to waste Richard's time for that Jan 21 23:15:15 where bitbake -c build gettext-0.14.1 works Jan 21 23:15:24 woglinde: I wonder if bitbake gettext-0.14.1-native works ? Jan 21 23:15:39 rp uh Jan 21 23:15:42 let me try Jan 21 23:15:43 tharvey: what platform? Jan 21 23:16:01 denix, is there any value in me creating a local recipe using dsplink 1.63? not clear if it will copoerate with the other parts in recipes/ti Jan 21 23:16:06 Laibsch: RP is probably the only one knowing about the thing in depth Jan 21 23:16:11 woglinde: or gettext-native-0.14.1 should really work Jan 21 23:16:17 ERROR: Nothing PROVIDES 'gettext-0.14.1-native' Jan 21 23:16:21 denix, overo - have a v4l2 camera driver I'm itching to get working with dsp encode Jan 21 23:16:29 woglinde: gettext-native-0.14.1-native ? Jan 21 23:16:31 ERROR: Nothing PROVIDES 'gettext-native-0.14.1' Jan 21 23:16:37 denix, omap35x Jan 21 23:16:37 rp lol Jan 21 23:16:52 woglinde: There should be something there even if its wrong ;-) Jan 21 23:17:01 khem: are you going to push Guo's request? Jan 21 23:17:05 ERROR: Nothing PROVIDES 'gettext-native-0.14.1-native' Jan 21 23:17:13 okay I will force the version for build Jan 21 23:17:36 ant__: this is no black magic. I can find out, I think, if I dug into. for the X time, this is VERY minor. Jan 21 23:17:42 rp I know its not a normal use case Jan 21 23:17:48 Laibsch: right Jan 21 23:17:55 but I it helps quicktest if package builds without reparsing Jan 21 23:18:12 woglinde: We do need a way to handle this problem Jan 21 23:18:12 from what I'm reading people have gotten ti-dsplink-module to build in the past - just not clear when things broke and what srcrev I need to go back to Jan 21 23:18:25 tharvey: omap3 based socs are supported by thjose recipes. try them, you may want to lock to sligtly older versions, whichever is publicly available Jan 21 23:18:28 woglinde: Interestingly the PROVIDES for gettext-native are empty :/ Jan 21 23:18:47 khem: http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=9f7e313d9136d588e399422d556f458ba1772912 could have seen a better commit message, IMHO Jan 21 23:19:06 In another note, there is a little discussion on using the Amazon cloud stuff to do OE builds on the gumstix list Jan 21 23:19:09 rp at poky too? Jan 21 23:19:17 woglinde: yes Jan 21 23:19:22 okay Jan 21 23:19:29 woglinde: look at the recipe - it sets them Jan 21 23:19:38 woglinde: (to empty) Jan 21 23:19:52 hm ah Jan 21 23:20:05 woglinde: then the PROVIDES_append in bitbake.conf is overridden Jan 21 23:20:11 why? Jan 21 23:20:25 denix, what I don't quite understand is that there are newer and older versions of dsplink available from ti but there is only a ti-dsplink-module_1.61.3 recipe - not clear why there wouldn't be a newer recipe Jan 21 23:20:25 you set it empty? Jan 21 23:20:36 Laibsch: yes I thought afterwards I could have interactively rebased it but then I honored the author Jan 21 23:20:50 woglinde: PROVIDES_append appends to PROVIDES, but then PROVIDES is replaced by PROVIDES_virtclass-native Jan 21 23:21:02 khem: you can rebase and still honour the author ;-) Jan 21 23:21:23 tharvey: testing Jan 21 23:21:48 tharvey: sometimes specific combinations of component versions don't work with each other Jan 21 23:22:24 so, testing is being done to some combinations, not others Jan 21 23:22:38 denix, I just wonder what the people that I've seen post recently on lists which say they have built ti-dsplink-module built with - perhaps it was not for OMAP35 or for a diff kernel Jan 21 23:23:13 Laibsch: btw. I could boot 2.6.31 on qemumips and qemux86 Jan 21 23:23:14 Laibsch: http://handhelds.org/hypermail/oe/1/0126.html Jan 21 23:23:15 ok, well I'll take a shot at building a different dsplink source and see how that flys, and I'll keep an eye out for newer recipes from you - thanks Jan 21 23:23:23 Laibsch: not on arm though Jan 21 23:23:34 really? Jan 21 23:23:37 I need to test Jan 21 23:23:38 tharvey: e.g. dsplink-1.64, bios-5.41.1.9, but there is 5.41.2.x bios, which is not yet tested with everything else for all the platforms Jan 21 23:23:39 Laibsch: If I find sometime I will try to debug it Jan 21 23:23:50 I don't remember having any problem with the kernel itself Jan 21 23:23:53 for qemuarm Jan 21 23:23:59 Only in X Jan 21 23:24:00 Laibsch: try with MACHINE=qemumips DISTRO=minimal Jan 21 23:24:11 woglinde: Basically the override trumps the append/prepend Jan 21 23:24:23 khem: were you able to start X, you say? Jan 21 23:24:27 Laibsch: hmm I did not try X though Jan 21 23:24:32 OK Jan 21 23:24:37 Laibsch: I am console user Jan 21 23:24:41 as I said, I had no problems with the console Jan 21 23:24:42 at all Jan 21 23:24:51 qemuarm worked fine for me Jan 21 23:24:52 IIRC Jan 21 23:25:16 once I got opie-image running on angstrom about year back but thats all Jan 21 23:25:32 ant__: you should talk to pb_, then Jan 21 23:25:34 it had all gui fired up Jan 21 23:25:36 tharvey: also, from tinderbox you can see I currently build/test for davinci platforms Jan 21 23:25:41 ant__: He seems to know about that stuff Jan 21 23:25:55 nah, he just seems to have committed Jan 21 23:26:15 btw 2004 ! Jan 21 23:26:18 Laibsch: Guo's request to up glibc to 2.9 ? Jan 21 23:26:26 khem: yes Jan 21 23:26:32 Laibsch: I wud rather push it upto 2.10.1 Jan 21 23:26:33 ant__: I noticed Jan 21 23:26:39 khem: then do that Jan 21 23:27:09 khem: I don't think there is anyone more knowledgeable than you in OE when it comes to toolchain stuff Jan 21 23:27:12 ideally all should use eglibc instead :) Jan 21 23:27:16 hms which distro was it that still used gettext-0.14.1 Jan 21 23:27:21 khem: hehe Jan 21 23:27:27 seems to be your favourite, he? Jan 21 23:27:30 why? Jan 21 23:27:47 I was one of the promoters for it Jan 21 23:28:06 and its better suited for people who want 'embedded glibc' Jan 21 23:29:11 denix, strange that when listing by package tinderbox doesn't show you machine Jan 21 23:29:38 woglinde, your building for beagle - where does your ti-dsplink-1613 come from? Jan 21 23:30:33 hm I have recipes/ti/ti-dspbios-native_5.33.06.bb Jan 21 23:30:59 and recipes/ti/ti-cgt6x-native_6.1.11.bb Jan 21 23:30:59 khem: you saw the mail from Dallas? Jan 21 23:31:12 ti-dsplink-module_1.61.3.bb Jan 21 23:31:13 is oe Jan 21 23:31:45 hm I will push gettext now Jan 21 23:31:51 Laibsch: yes I replied too Jan 21 23:31:58 woglinde, why does tinderbox show version of your ti-dsplink-module as 1613? Jan 21 23:32:00 hm I will see if will be reverted Jan 21 23:32:24 tharvey do yo have the date at hand? Jan 21 23:32:29 woglinde: Why would it get reverted? Jan 21 23:32:42 rp maybee it breaks something Jan 21 23:33:01 dont know yet Jan 21 23:33:01 03Khem Raj  07org.openembedded.dev * r8ec8982add 10openembedded.git/conf/distro/include/sane-toolchain.inc: Jan 21 23:33:01 sane-toolchain.inc: Bump glibc PREFERRED_VERSION to 2.10.1 Jan 21 23:33:01 Signed-off-by: Khem Raj Jan 21 23:33:06 testing is only minimal Jan 21 23:33:07 probably woglinde builds from stable Jan 21 23:33:13 no Jan 21 23:33:16 woglinde, 11-26-2009? Jan 21 23:33:22 I never checkout stable Jan 21 23:33:27 http://tinderbox.openembedded.org/packages/ti-dsplink-module/ Jan 21 23:33:31 then old checkout Jan 21 23:33:33 * XorA gets himself confused inside BCLASSEXTENDED now Jan 21 23:33:43 xora *g* Jan 21 23:33:44 cause PV has changed from 1613 to 1_61_3 Jan 21 23:33:45 ask rp Jan 21 23:33:53 right Jan 21 23:33:53 need to apply a patch to the not native only Jan 21 23:34:05 there was problems with the version stuffs Jan 21 23:34:36 or maybe a do_compile_prepend to the not native only Jan 21 23:34:40 for instance 2.34.1 which was 2341 and 2.24 wich was 224 Jan 21 23:34:47 you can guess what was picked Jan 21 23:34:47 XorA: SRC_URI and SRC_URI_virtclass-native ? Jan 21 23:34:58 XorA: or see examples in poky Jan 21 23:35:21 args Jan 21 23:35:23 RP: I assume there isnt a _!native overide Jan 21 23:35:26 damn example Jan 21 23:35:30 XorA: no :/ Jan 21 23:35:39 make it 2.42 and 242 Jan 21 23:35:48 RP: bloomin not safe cross compile stuff :-) Jan 21 23:36:01 XorA: which recipe? Jan 21 23:36:09 RP: eggdbus Jan 21 23:36:12 XorA: I gave up on libtool in the end :/ Jan 21 23:36:36 XorA: Well, shelved it under "sometime when I'm feeling brave" Jan 21 23:37:08 * kergoth chuckles Jan 21 23:37:56 what is the best way to prevent auto**** making these progams, is there a simple patch to Makefile.am, just removing them from PROGRAMS var work? Jan 21 23:38:48 woglinde: have you committed all vdr stuff Jan 21 23:38:59 XorA: The latter I'd have thought but I'm not 100% sure what you're doing Jan 21 23:39:24 RP: two programs called during compile stage that I need to replace from ones built by a native version Jan 21 23:39:56 XorA: Hack the Makefile.am and remove from PROGRAMS, yes Jan 21 23:40:05 RP: awesome Jan 21 23:40:37 XorA: Usually I patch them to call the prorams without a path so it searches PATH, then you can just install them into staging Jan 21 23:41:07 RP: hmm, that might be a better hack, I shall check Jan 21 23:43:05 khem sure Jan 21 23:44:58 khem: does that change need review? Jan 21 23:45:09 laibsch not really Jan 21 23:45:22 laibsch was only split-up Jan 21 23:45:28 for .inc file Jan 21 23:45:36 i have probs compiling libx11 with errors like ....... ximcp/../../../libtool: No such file or directory Jan 21 23:45:42 and paul is fairly the only user of vdr so far Jan 21 23:45:52 woglinde: I meant the jump to glibc 2.10 Jan 21 23:45:53 pwgen hm Jan 21 23:45:58 laibsch oh Jan 21 23:46:12 woglinde: no I just wanted to mark the patchwork entries Jan 21 23:46:28 khem uh Jan 21 23:46:30 sorry Jan 21 23:46:35 I read it diffrent Jan 21 23:46:38 I tried the quemux86 MACHINE and angstrom Jan 21 23:46:58 woglinde, so I'm curious why your ti-dsplink-module-1613 is building for beagle and mine is failing for overo - same cpu arch, same dspbios/link 1.61.3 Jan 21 23:46:58 pwgen might be problem with our libtool renaming Jan 21 23:47:14 tharvey hm whats failing? Jan 21 23:47:19 exctalty Jan 21 23:47:24 do you have the log at hand? Jan 21 23:47:57 yesterday setting it to MACHINE=akita and angstrom it builds well ? Jan 21 23:49:36 http://pastebin.com/m1df929ea Jan 21 23:50:10 pwgen: are you on the most recent .dev tree? Jan 21 23:50:12 pwgen we rename the libtool inside builds Jan 21 23:50:37 maybee its broken for qemu stuff and the newer xlibs Jan 21 23:51:16 b6f9a84..8ec8982 org.openembedded.dev -> origin/org.openembedded.dev Jan 21 23:51:20 khem so to answer your question the latest patch from paul the splitupt into .inc files I didnt pushed yet Jan 21 23:51:33 ok Jan 21 23:51:43 his patches appear weird on formatting Jan 21 23:51:49 ah dolt again Jan 21 23:51:52 although it looks like they are git formatted Jan 21 23:52:04 pwgen mom Jan 21 23:52:15 woglinde, ti-dsplink-1.61.3 failure - http://pastebin.ca/1760926 Jan 21 23:52:28 ./gpp/src/arch/OMAP3530/shmem/Linux/omap3530_phy_shmem.c:131: error: implicit declaration of function 'IO_ADDRESS' Jan 21 23:52:39 pwgen look at 29c63b640c4e4c852cf3084232ef67c8fd9d3214 for fix Jan 21 23:53:29 woglinde, perhaps an issue building against linux-omap3-2.6.32-r51 - what kernel recipe are you using? Jan 21 23:53:46 tharvey he yes might be Jan 21 23:53:48 was older Jan 21 23:53:54 .29 I think Jan 21 23:53:57 in fact most of the ti support is stuck at 2.6.29 isnt' it? Jan 21 23:53:59 ya Jan 21 23:54:04 not sure Jan 21 23:54:12 I dont sill get what psp is Jan 21 23:54:13 waiting for that newer PSP that bumps support up to 2.6.32 (supposed to be end of Feb) Jan 21 23:54:17 and over stuff Jan 21 23:54:30 yeah maybee its overo related Jan 21 23:54:42 you can look the beagleboard stuff koen checked in Jan 21 23:54:51 maybee you can find a fix for overo too Jan 21 23:55:02 overo and beagle kernels mostly the same - but big diffs in 2.6.29 and later kernels due to omission of many of the TI patches Jan 21 23:55:22 tharvey you could maybee quikc test setting machine beagleboard Jan 21 23:55:27 and see if it builds Jan 21 23:55:40 ya, was going to kick off that build tonight... likely will take a while Jan 21 23:56:27 hm why? Jan 21 23:56:33 when it is the same arch? Jan 21 23:56:47 all what you have build for overo dont needs to rebuild Jan 21 23:56:49 only kernel Jan 21 23:56:50 ya, perhaps only the kernel will rebuild Jan 22 00:04:03 adding -DOLT to confiure.ac doesnt help Jan 22 00:04:26 pwgen aeh? Jan 22 00:04:39 you have to remove the whole dolt lines Jan 22 00:04:48 hm let me see what marco dit Jan 22 00:05:18 ah Jan 22 00:05:21 its a patch Jan 22 00:05:28 and the -DOLT means Jan 22 00:05:35 please remove this line Jan 22 00:05:35 the patch says something about adding AC_PRG_LIBTOOL and -DOLT ... Jan 22 00:06:05 mompl Jan 22 00:06:16 it says Jan 22 00:06:20 disabel staitc Jan 22 00:06:31 ac_porg_libtool Jan 22 00:06:40 hm you will get it Jan 22 00:07:01 *G* yeupp worked .. Jan 22 00:07:21 *g* Jan 22 00:07:37 RP: you are a fscking genius :-D Jan 22 00:07:54 xora what? Jan 22 00:07:59 ... pwgen.. go back and learn reading patches again ...... Jan 22 00:08:08 woglinde: is that in debate? Jan 22 00:08:25 *g* Jan 22 00:08:39 xora now you are fond with BBCLASSEXTEND ? Jan 22 00:09:15 woglinde: I love it and was the first in OE :-) Jan 22 00:09:27 ????? Jan 22 00:09:30 sure? Jan 22 00:09:31 XorA, when you are bored, you need to explain that remark Jan 22 00:09:39 woglinde: yes Jan 22 00:09:45 show me Jan 22 00:09:51 *g* Jan 22 00:10:12 Crofton: RPs advice about removing the patch works a dream and saves buggering around Jan 22 00:10:17 path Jan 22 00:13:14 'nite all Jan 22 00:17:11 woglinde, the missing macro IO_ADDRESS used in dsplink 1.63.03 is defined in kernel - not sure why its not defined, I suppose the files are not pulling in kernel headers or something Jan 22 00:17:32 I thought dsplink was patched into kernel these days - not clear why its built as an external module Jan 22 00:24:44 03Khem Raj  07org.openembedded.dev * r310b2cd891 10openembedded.git/recipes/whois/ (whois.inc whois_4.5.25.bb whois_5.0.0.bb): Jan 22 00:24:44 whois: Add recipe for version 5.0.0. Jan 22 00:24:44 * Move to using INC_PR Jan 22 00:24:44 * Add md5 and sha checksums to recipe. Jan 22 00:24:44 * Move more common parts like SRC_URI to whois.inc. Jan 22 00:24:46 * Thanks Paul Menzel for reporting the pending upgrade. Jan 22 00:24:50 Signed-off-by: Khem Raj Jan 22 00:24:52 03Jan Paesmans  07org.openembedded.dev * r5cc04010e7 10openembedded.git/recipes/gsoap/ (4 files in 2 dirs): (log message trimmed) Jan 22 00:24:55 gSOAP Jan 22 00:24:57 Hello, Jan 22 00:24:59 In the gSOAP package I noticed the following patch was applied to the Jan 22 00:25:01 native build: use-just-built-binary.patch Jan 22 00:25:03 This patches the makefiles of gSOAP so that the correct executable is used. Jan 22 00:25:07 However the problem originates from the fact that the makefile Jan 22 00:26:25 er, the commit message on that last one is awful Jan 22 00:26:38 please do not push stuff like that Jan 22 00:26:47 was very cordial :P Jan 22 00:26:49 http://cgit.openembedded.org/cgit.cgi/openembedded/log/ Jan 22 00:26:54 I though CIA achieved sentience Jan 22 00:26:59 heh Jan 22 00:27:08 XorA: Glad it worked :) Jan 22 00:27:27 read the log to see why that message sucks, if you are skimming the log to see what people are doing Jan 22 00:27:28 RP: I wrote a turorial on it as well :-) Jan 22 00:27:35 you can't make heads or tails of it Jan 22 00:28:07 XorA, I don't suppose you could write a tutorial on creating and submitting patches Jan 22 00:28:18 Crofton_|work: not tonight :-) Jan 22 00:28:26 Crofton_|work: and I still owe TSC some work as well Jan 22 00:28:28 XorA: :) Jan 22 00:28:28 I know Jan 22 00:28:48 I just think that would be a really great thing in light of todays list discussion Jan 22 00:28:56 I would do it, but I suck at that sort of thing Jan 22 00:29:11 I am reading you r latest article atm though Jan 22 00:29:47 just getting patches on the klist with good log messages thhat the script applied would go a long way to solving the patch backlog Jan 22 00:30:10 the new one covers making native :-) Jan 22 00:30:16 and autotolls Jan 22 00:30:27 so should be better as a reference than the eyeos one Jan 22 00:31:05 article? Jan 22 00:31:16 http://www.xora.org.uk/2010/01/22/openembeddedangstom-new-package-workflow-eggdbus/ Jan 22 00:31:29 ah Jan 22 00:31:37 kergoth: Ive been trying to document what I do to create recipes Jan 22 00:31:46 kergoth: trying to lower the barrier for people Jan 22 00:31:48 * kergoth has a terrible backlog on his google reader and instapaper and read it later Jan 22 00:31:52 nice, that's great to hear Jan 22 00:32:06 extend that article to include generating a pacth that git-am can apply :) Jan 22 00:32:13 i'll have to link this to the guys within mv making recipes, so i can stop denying their reviews with a crapload of comments ;) Jan 22 00:32:31 Crofton_|work: Ill do a small one, how to add a recipe to OE Jan 22 00:32:47 * XorA is tired now, its 12:30am Jan 22 00:33:26 xora hm Jan 22 00:33:38 you could have use the new checksum in recipe feature Jan 22 00:33:40 *g* Jan 22 00:33:51 woglinde: that seems to have gained zero traction Jan 22 00:33:52 no worries Jan 22 00:34:21 the problem is bitbake gives you the text to add to checksums.ini Jan 22 00:34:27 and the sorter makes it easy Jan 22 00:34:30 to add Jan 22 00:35:07 in recipe is about 100 times more complex as well Jan 22 00:35:32 SRC_URI = "${SOURCEFORGE_MIRROR}/wesnoth/${PN}-${PV}.tar.bz2;name=tarball" Jan 22 00:35:33 SRC_URI[tarball.md5sum] = "493826bbd9ba355930765a7e8fe3749a" Jan 22 00:35:33 SRC_URI[tarball.sha256sum] = "7ef047ae364278a5bf9bdc69228f77d825f793f1c4d9adae8b47f0882e7f30d7" Jan 22 00:35:40 thats hard? Jan 22 00:35:49 how do you examine the control file in the ipk to see the depends? Jan 22 00:35:55 when you start to include .inc and multiple sources Jan 22 00:36:19 Crofton: ar -x blah.ipk; tar zxvf control.tar.gz; less control Jan 22 00:36:27 dpkg-deb -i blah.ipk Jan 22 00:36:32 depending on your distro Jan 22 00:36:34 you shoudl mention that in the article :) Jan 22 00:36:57 mkdir foo Jan 22 00:37:01 mv deb foo Jan 22 00:37:09 cd foo Jan 22 00:37:19 dpkg -X deb . Jan 22 00:37:24 hm Jan 22 00:37:25 oh Jan 22 00:37:28 nevermind Jan 22 00:37:34 just pointig out that piece of the article is not clear Jan 22 00:38:19 Crofton_|work: its way faster to parse Jan 22 00:38:31 make changes to checksum.ini and you are stuck for another 10 mins Jan 22 00:38:38 untill it finished parsing Jan 22 00:38:43 sure Jan 22 00:39:12 ipkgfiles () { Jan 22 00:39:12 ar p $1 data.tar.gz | tar -tvz Jan 22 00:39:12 return $? Jan 22 00:39:12 } Jan 22 00:39:13 ipkgcontrol () { Jan 22 00:39:13 ar p $1 control.tar.gz | tar -zxO ./control Jan 22 00:39:14 return $? Jan 22 00:39:16 } Jan 22 00:39:18 courtesy ~/.sh/util Jan 22 00:39:21 but, people have worked out how to deal with checksums.ini Jan 22 00:39:26 03Graeme Gregory  07org.openembedded.dev * ra4cebe9f1d 10openembedded.git/recipes/eggdbus/ (eggdbus_0.6.bb files/gtk-doc.patch files/marshal.patch): Jan 22 00:39:26 eggdbus_0.6.bb : add this new recipe which is gobject bindings for dbus Jan 22 00:39:26 This is required by the newer policykit packages which I am working on Jan 22 00:39:26 pulling into OE. Jan 22 00:39:27 :) Jan 22 00:39:54 xora hm ah the _append stuff didnt work right? Jan 22 00:40:26 woglinde: the patch only applies to the not virtclass Jan 22 00:42:00 hm ah yes Jan 22 00:42:24 we havent persuaded kergoth to write a ! operator yet Jan 22 00:42:36 I suppose I should mention there are places like github that let you create git repos we can look at also Jan 22 00:43:13 03Stanislav Brabec  07org.openembedded.dev * r1922d0e881 10openembedded.git/ (4 files in 3 dirs): Jan 22 00:43:13 pidgin: Update to version 2.6.5 (security: CVE-2010-0013). Jan 22 00:43:13 * Fixes, security update for MSN plugin. Jan 22 00:43:13 * Removed unreferred recipes for vulnerable versions. Jan 22 00:43:13 anyway I am tired, its been a long day of coding/IT/coding/OE so Im off to sleep Jan 22 00:43:14 03Stanislav Brabec  07org.openembedded.dev * re2cc28b207 10openembedded.git/conf/checksums.ini: checksums.ini: Added missing checksum for balsa-2.4.2. Jan 22 00:43:18 03Stanislav Brabec  07org.openembedded.dev * r615c15be35 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Jan 22 00:44:22 nite xora Jan 22 00:47:42 jo raster Jan 22 00:48:01 what should out of tree kernel modules have access to in OE? only staging/.../kernel or are those only userspace kernel headers? Jan 22 00:48:45 trying to understand of dsplink is wrong in using the IO_ADDRESS macro, or if kernel headers are wrong in not including the file that defines it Jan 22 00:49:21 tharvey dont know Jan 22 00:49:27 maybe denix know it Jan 22 00:49:39 tharvey you could ask int #gst_ti too Jan 22 00:50:00 thx - didn't know there was a #gst_ti :) Jan 22 00:55:40 hmmm my git ammend for that gsoap did not get pushed instead the old message went thru crap Jan 22 00:55:54 args Jan 22 00:56:28 I forgot to git commit after git ammend :( seems Jan 22 00:56:40 I need a break Jan 22 00:56:53 woglinde, still not able to figure out why you were able to build dsplink-1.61.03 - should have encountered the same issue - do you have any IO_ADDRESS defined in your tmp/staging/? Jan 22 00:57:14 tharvey hm its 3 month Jan 22 00:57:20 so I didnt remember Jan 22 01:01:27 huh, is something up with xora's rss feed? doesn't include that post, yet anyway Jan 22 01:02:38 heh Jan 22 01:03:12 khem I know what you mean, earlier, I added the linux-omap kernel as a remote to gnuradio Jan 22 01:03:55 khem, just send something to the list about the gsoap commit :) Jan 22 01:04:28 encourage people to send proper commit messages so you do not need to use the amend thing Jan 22 01:05:18 keroght hm its planet now Jan 22 01:05:28 + on Jan 22 01:05:57 kergoth, I saw the article on xora's twitter Jan 22 01:06:21 ah Jan 22 01:11:09 woglinde, denix ti-dsplink-module builds fine against 2.6.29 - something in 2.6.32's kernel headers does not provide the macro for IO_ADDRESS Jan 22 01:11:53 khem: 5cc04010e70378484efe549b30b411f840e825f5 is even more sucky when it comes to commit messages Jan 22 01:17:28 although once you look beyond the shortlog, there is info Jan 22 01:17:42 but the shortlog info is key for people trying to dave time Jan 22 02:07:41 okay Jan 22 02:07:45 I will push now Jan 22 02:07:46 gettext Jan 22 02:07:49 seems to work Jan 22 02:09:24 03Henning Heinold  07org.openembedded.dev * rcb3a5b9cad 10openembedded.git/recipes/gettext/ (4 files): Jan 22 02:09:24 gettext: move to BBCLASSEXTEND, taken from popky Jan 22 02:09:24 * delete all native recipes now handled by BBCLASSEXTEND Jan 22 02:09:24 * bump PR's Jan 22 02:09:36 03Henning Heinold  07org.openembedded.dev * r0f29546c89 10openembedded.git/recipes/gettext/gettext_0.17.bb: Jan 22 02:09:36 gettext: add extra package for libintl Jan 22 02:09:36 * add extra package for libintl because it is mostly Jan 22 02:09:36 enough for running the stuff on targets Jan 22 02:09:37 * let runtime conflict it with proxy-libintl Jan 22 02:09:39 * bump PR **** ENDING LOGGING AT Fri Jan 22 02:59:57 2010