**** BEGIN LOGGING AT Tue Mar 23 02:59:57 2010 Mar 23 04:28:08 khem, ping Mar 23 05:09:57 I am getting error while building "bitbake console-image" Mar 23 05:10:06 It passed @ 1700 tasks Mar 23 05:10:31 It is unable to fetch from "http://www.angstrom-distribution.org/unstable/sources/git_opensource.wolfsonmicro.com.linux-2.6-asoc.tar.gz" Mar 23 05:10:45 Anyone has any idea about how to fix it Mar 23 05:10:47 SagS: that is no problem Mar 23 05:11:01 I'm building it for "smdk-2440" Mar 23 05:11:17 SagS: before hitting SCMs we try to download a tarball from a static source, if it is not there it will fallback to use git to get it. Mar 23 05:11:57 but my build stops after this error Mar 23 05:12:16 It gives an error "fatal: Unable to look up opensource.wolfsonmicro.com (port 9418) (Name or service not known)" Mar 23 05:13:35 SagS: well, you are behind firewall. Mar 23 05:13:40 Yep Mar 23 05:13:48 SagS: then ask your admins for real internet. Mar 23 05:13:57 I cant access FTP sites Mar 23 05:14:42 SagS: no kidding. Get real internet. OpenEmbedded will sometime attempt to use svn,git,mercurial,cvs to get the sourcecode Mar 23 05:14:42 zecke : OKey.. So u think it shall resolve my issues if my firewall is disabled...? Mar 23 05:15:08 SagS: if your company does not allow you to use svn, git, mercurial, cvs and others then you have a problem. Mar 23 05:15:31 zecke : thanks for the help.. I shall try it now only... Mar 23 05:16:30 SagS: which recipe where you building? Mar 23 05:19:14 I'm building it for "smdk-2440" Mar 23 05:19:22 SagS: you can also go to, repo.or.cz, then clone the wolfon git repository, change linux-smdk2440_...git to use this repository with http protocol instead Mar 23 05:19:57 SagS: well, the error message does tell you the name of the "recipe" (the bb file), it is easy to find. In your case it was recipes/linux/linux-smdk2440_2.6.20+git.bb Mar 23 05:22:46 Thanks zecke for the help Mar 23 05:22:58 i need to restart now to configure my real internet Mar 23 05:23:01 logging off... Mar 23 05:23:13 good luck **** BEGIN LOGGING AT Tue Mar 23 11:58:15 2010 Mar 23 12:12:02 MWelchUK_work: please rememeber about bumping PR each time Mar 23 12:12:12 MWelchUK_work: http://patchwork.openembedded.org/patch/1724/ for example Mar 23 12:13:02 hrw, I assume that only needs to be done if the patch is actually applied. Mar 23 12:13:16 MWelchUK_work: if package contents change Mar 23 12:13:41 MWelchUK_work: when you send patches to oeml then assume that they will get applied - makes life easier Mar 23 12:13:42 hrw, yup. Mar 23 12:14:54 I added most of your work into my tree for checking Mar 23 12:16:26 hrw, Ok, I wasn't sure it was ready yet and I'm sure I didn't sign'em Mar 23 12:18:09 MWelchUK_work: signing is easy: git commit -s Mar 23 12:19:47 I'm trying create an sdk with python-dev and libusb1-dev. Is following/modifying http://docs.openembedded.org/usermanual/html/commonuse_qte_sdk.html the path to follow? Mar 23 12:20:13 can someone help me to debug posible thumb issue? Mar 23 12:20:18 http://pastebin.ca/1850197 Mar 23 12:20:49 JaMa|Wrk: pb__ is your man Mar 23 12:20:57 tasslehoff: well, it depends. if it is only for your use... it might be easier to extend the task-meta-toolchain.bb Mar 23 12:22:00 zecke: it's only for me, so I'll try that. Mar 23 12:28:08 ~curse lighttpd Mar 23 12:28:09 May you be reincarnated as a Windows XP administrator, lighttpd ! Mar 23 12:30:57 hrw if you think lighttpd is a curse, try to get up apache + php + mysql with oe Mar 23 12:31:40 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=564556 gave me light Mar 23 12:33:36 ant_work: I know he is, but I don't want to bother him each time :) Mar 23 12:35:08 zecke: you meant modifying task-sdk-host and task-sdk-bare? Mar 23 12:35:30 tasslehoff: yeah, sounds right. :) Mar 23 12:35:35 :) Mar 23 12:37:16 hrw: ah ok, i'm not using v6 (at least not afaik) Mar 23 12:38:46 * XorA thinks we will be on ipv8 before we see mass ipv6 rollouts Mar 23 12:39:15 does someone know a tool which will mass download videos from youtube? Mar 23 12:39:28 but not by links but by search phrases Mar 23 12:41:18 totem Mar 23 12:43:25 actually currently ipv6 deployment picks up in speed Mar 23 12:44:58 XorA if you go to the far east you can see mass ipv6 rollout today Mar 23 12:45:02 rkirti: ping Mar 23 12:45:44 eFfeM: I think I went too far or not far enough. :) Mar 23 12:46:19 * XorA would love ipv6 and make all his machines routable Mar 23 12:47:21 zecke: actually thought tw is one of the countries where there is a reasonable rollout Mar 23 12:48:46 zecke then again cn used ipv6 in the 2008 olympics, so maybe that is a reason for tw to stay on v4 Mar 23 12:49:03 eFfeM: hehe, I have seen ipv6 in japan Mar 23 12:49:17 hrw: youtube-dl Mar 23 12:49:39 03Sebastian Spaeth  07org.openembedded.dev * r6013125501 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh+git://git@git.openembedded.net/openembedded into org.openembedded.dev Mar 23 12:49:40 03Sebastian Spaeth  07org.openembedded.dev * rd5977a2bfe 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh+git://git@git.openembedded.net/openembedded into org.openembedded.dev Mar 23 12:49:40 03Sebastian Spaeth  07org.openembedded.dev * rd454200a43 10openembedded.git/recipes/abiword/abiword-2.5.inc: Mar 23 12:49:41 abiword-2.5.inc: Resemble DESCRIPTION more to valid English grammar Mar 23 12:49:41 I really, really think the abiword recipes need cleanup we have old versions back to 2.4.* in here. Mar 23 12:49:41 Signed-off-by: Sebastian Spaeth Mar 23 12:49:43 03Sebastian Spaeth  07org.openembedded.dev * ra501f95529 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh+git://git@git.openembedded.net/openembedded into org.openembedded.dev Mar 23 12:49:44 broonie: it fetches one Mar 23 12:50:16 hrw: for x in list ; do ... Mar 23 12:50:47 broonie: sure, but I hoped for tool which will extract list too; :D Mar 23 13:03:42 ~curse make Mar 23 13:03:43 make: INTERNAL: Exiting with 5 jobserver tokens available; should be 4! Mar 23 13:03:43 May you be reincarnated as a Windows XP administrator, make ! Mar 23 13:08:48 in linux-omap-psp_2.6.32.bb, SRC_URI has 'file://defconfig'. I don't see how do_unpack() knows to copy this out of recipes/linux/linux-omap-psp-2.6.32/omap3evm/ like it does. Mar 23 13:09:35 according to the bitbake manual, the path searched is defined by FILESPATH and FILESDIR, neither of which show up with 'printenv' in devshell Mar 23 13:14:42 see FILESPATHPKG etc. in bitbake.conf Mar 23 13:19:02 I tried adding libusb1-dev to RDEPENDS in task-sdl-bare.bb and compiled meta-toolchain, but I have no libusb1-stuff in my sdk. Where did I go wrong? Mar 23 13:21:12 ant_work, thanks. Mar 23 13:33:25 hm. problem was that it didn't rebuild everything because I didn't bump PR Mar 23 14:01:38 hey -- does anyone know if there is now a way to specify md5sums/shasums for files in .bb files instead of checksums.ini? Mar 23 14:02:15 mems, there is on the dev branch Mar 23 14:02:22 of bitbake? Mar 23 14:02:34 no OE Mar 23 14:02:45 It's OE specific at the moment Mar 23 14:02:58 mems: lots of examples are already there Mar 23 14:03:09 JaMa|Wrk: can you point me at one? Mar 23 14:03:50 http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=c853d5b877257a4b2addc762ea2fea918968c7e4 Mar 23 14:04:48 JaMa|Wrk: thanks :-D Mar 23 14:04:49 mems, recipes/cscope/cscope_15.7a.bb Mar 23 14:04:51 hmm not so good as example because it already expects ;name=archive in SRC_URI from .inc Mar 23 14:06:07 JaMa|Wrk, mems, recipes/cscope/cscope_15.7a.bb is all in one file, there's also very little else to it. Mar 23 14:06:09 so this won't work with the stable/2009 branch at the moment? Mar 23 14:06:19 mems, AFAIK no Mar 23 14:06:22 ok Mar 23 14:06:43 MWelchUK_work: yeah, I agree much better example than mine Mar 23 14:07:06 Just took me longer to find it :-) Mar 23 14:07:26 and I would pick one with name=archive :) Mar 23 14:07:26 cheers. any ideas of when this behaviour might find it's way into stable? Mar 23 14:07:48 Probably won't be merged with the current stable tree. Mar 23 14:07:57 At a guess. Mar 23 14:08:19 I think that is a good guess Mar 23 14:08:25 since it needs bitbake changes Mar 23 14:08:26 sounds like you might be implying that there's a new stable tree coming... Mar 23 14:08:38 when the users demand it :) Mar 23 14:08:41 heh Mar 23 14:08:56 Again, only a guess given the branch is called stable/2009 Mar 23 14:09:10 MWelchUK_work, good guess Mar 23 14:09:37 where should I send a patch for the dev branch? Mar 23 14:09:38 the idea is to provide a slowly moving target for people, not keep up with .dev :) Mar 23 14:09:44 oe dev ml Mar 23 14:09:47 absolutely, and it's great Mar 23 14:10:00 It's a shame we don't do things more like debian, with unstable, testing and stable, though this will probably result in lots of work... Mar 23 14:10:11 it's just that there's a broken URI in it for ipython... Mar 23 14:10:52 mems, you can always ask for a patch to be included. Mar 23 14:10:57 into stable? Mar 23 14:11:04 To fix the URI - yeah. Mar 23 14:11:07 ok Mar 23 14:11:38 Stable is only any use if it will build. As long as your not up-reving it at the same time. Mar 23 14:13:08 thanks. Mar 23 14:13:12 np Mar 23 14:13:57 Maybe you guys can help me understand some of the history of bitbake. I've seen some references to "cbitbake", but it doesn't appear to actually exist, is it vaporware? Mar 23 14:14:34 mems: mostly yes, IIRC the directory existed in trunk/cbitbake on the svn of berlios Mar 23 14:14:56 ok, not much interest in developing it then? Mar 23 14:15:08 mems: the slowness on parsing is coming from the fact we have to evaluate python scripts Mar 23 14:15:19 khem: isn't DEPENDS = "ruby-native" in ruby_1.8.7-p248.bb now with BBCLASSEXTEND = "native", causing bitbake tasks dependency loops? Mar 23 14:16:01 zecke: so are then any "plans" to change stuff so it's faster? Mar 23 14:16:37 khem: it's says that Identifying dependency loops will take a while (forever).. so I'm just guessing from last commits which could be pulled to task-shr-feed (ie ruby) Mar 23 14:16:39 mems, some are trying to trim out older versions of stuff. MAking the tree smaller will make it faster... Mar 23 14:16:47 mems: now we come to varpoware but I have to call it a day Mar 23 14:17:00 mems: 1.) change OE to use less anonymous functions that are evaluated after parsing Mar 23 14:17:23 mems: 2.) we do have a working lexer/parser in C/Cython. It is based on work of Marc Singer Mar 23 14:18:37 khem: yeah, for shorter runqueue it finished soon: http://pastebin.ca/1850285 Mar 23 14:19:20 zecke: ok, thanks for shedding some light Mar 23 14:20:07 khem: I'll push fix soon Mar 23 14:20:20 mems: e.g. even with the lexer in a test environment for parsing... we spend most of the time in python evaluation. So we do have like 0.6s to lex all 6000 files from then, and 30 seconds to run anonymous python functions Mar 23 14:20:31 bye Mar 23 14:22:40 e2fsprogs-libs will soon vanish Mar 23 14:24:47 hello, anybody to help me with this error : http://effik.pastebin.com/Ah6L9DLT ? Mar 23 14:33:53 03Martin Jansa  07org.openembedded.dev * r471981651b 10openembedded.git/recipes/ruby/ (ruby.inc ruby_1.8.5.bb ruby_1.8.7-p248.bb): Mar 23 14:33:53 ruby: resolve dependency loop virtual:native:ruby -> ruby-native Mar 23 14:33:53 Signed-off-by: Martin Jansa Mar 23 14:37:58 has anyone built cmake lately? Mar 23 14:45:06 What do I need to depend on in my sdk to get the linux headers? I tried linux-libc-headers-dev, but can't find headers like compiler.h Mar 23 14:50:59 morning Mar 23 14:52:32 Which branch is the most recent stable branch? Mar 23 14:55:06 kergoth: I have delved into all the ramifications but I do like the idea of BBVERSIONS Mar 23 14:55:11 havent Mar 23 14:56:59 hello, anybody to help me with this error : http://effik.pastebin.com/Ah6L9DLT ? (compile angstrom) Mar 23 14:59:27 XorA: I'm thinking about getting it into bitbake master so its available when 1.10 comes out, and then see how people make use of it when we bump the minimum required bitbake version, but once its in, its in, and removing it later would be a pain... hence the discussion on the frontend :) I'm rather torn on the whole thing Mar 23 15:00:06 * kergoth shrugs, hmm Mar 23 15:01:03 sakoman: your pulseaudio patch worked for me too, can you push it? Mar 23 15:01:50 JaMa|Wrk: sorry, someone with commit privileges will have to do that Mar 23 15:02:31 is it in pw? Mar 23 15:02:31 sakoman: and can you send it as git patch? Mar 23 15:02:51 sakoman: otherwise I have to find out your real name and e-mail somehow to give you credits :) Mar 23 15:05:18 is there any way to pull a library into an image without having a program that depends on it first? Mar 23 15:05:51 adding libgeos-c1 to IMAGE_INSTALL does not work Mar 23 15:05:58 JaMa|Wrk: the info you need should be here: http://pastebin.com/krTf2rQk Mar 23 15:06:46 sakoman: thanks Mar 23 15:07:47 JaMa|Wrk: are you going to push the very similar tslib patch? Mar 23 15:08:50 sakoman: haven't seen tslib patch (and haven't build tslib here) Mar 23 15:09:14 JaMa|Wrk: someone posted the patch here yesterday Mar 23 15:09:33 if you try to build tslib you'll notice a similar issue Mar 23 15:09:46 cbrake: adding the library package to IMAGE_INSTALL should work fine, but I'm honestly not sure if you need to use the pre-shlib-renaming name or the post-shlib-renaming name :) Mar 23 15:09:50 cbrake: maybe try the other Mar 23 15:10:18 kergoth: what would be the opposite of libgeos-c1? Mar 23 15:11:13 dunno, check the recipe. if foo emits libfoo.so.1, the foo package will become libfoo1 Mar 23 15:11:25 pre-rename = the entries in PACKAGES Mar 23 15:12:01 hello, anybody to help me with this error : http://effik.pastebin.com/Ah6L9DLT ? (compile angstrom) Mar 23 15:13:19 Hello! Mar 23 15:13:44 Is there a jtag facility in one device, the target devices: avr32, arm ? Mar 23 15:13:59 We'd like to use one programmer/debugger for these devices. Mar 23 15:14:52 from Atmel Vendor e.g. Mar 23 15:17:07 pls anybody help me, i neet it , http://effik.pastebin.com/Ah6L9DLT Mar 23 15:17:35 * MWelchUK_work thinks that wget may be failing with PARALLEL_MAKE = "-j 2" Mar 23 15:21:45 03Martin Jansa  07org.openembedded.dev * r30332a51b3 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Mar 23 15:21:45 shr-splash: bump SRCREV for hide VT cursor patch Mar 23 15:21:45 Signed-off-by: Martin Jansa Mar 23 15:21:56 03Steve Sakoman  07org.openembedded.dev * r9cf826839c 10openembedded.git/recipes/pulseaudio/pulseaudio.inc: Mar 23 15:21:56 pulseaudio: patch configure.ac to work around install error Mar 23 15:21:56 Signed-off-by: Martin Jansa Mar 23 15:23:14 03Martin Jansa  07org.openembedded.dev * r575091d50e 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Mar 23 15:23:14 pyphonelog: bump SRCREV for dbus name fix Mar 23 15:23:14 Signed-off-by: Martin Jansa Mar 23 15:28:21 XorA: still experimenting, but moved it from gist to github so its a bit more comprehensible - http://github.com/kergoth/Nano-OE/tree/master/recipes/nano/ Mar 23 15:28:24 heh Mar 23 15:30:32 03Steve Sakoman  07org.openembedded.dev * r5b6b153cde 10openembedded.git/recipes/pulseaudio/files/configure-ac.patch: Mar 23 15:30:32 pulseaudio: patch configure.ac to work around install error Mar 23 15:30:32 Signed-off-by: Martin Jansa Mar 23 15:46:54 | /sdr/build/tmp/staging/i686-linux/usr/bin/install: cannot create regular file `../../../bin/omniNames': File exists Mar 23 15:47:04 does this sort of error ring a bell with anyone? Mar 23 15:47:47 hmm, could be its trying to install a file into a dir that exists as a file, or something Mar 23 15:47:51 odd one Mar 23 15:48:26 it is from someone else, but I am thinnking this is not the first appearance of such an error recently, trying to check on my machine now Mar 23 15:48:55 Crofton: Fix the reference counting on enable/disable in the MMC codel. Mar 23 15:49:10 hmm Mar 23 15:49:16 broonie, you assume I am smart :) Mar 23 15:49:34 Crofton: Oh, FFS. Actually, try turning off REGULATOR_USE_DUMMY. Mar 23 15:49:39 it seems like I can ignore the problem for now :) Mar 23 15:49:46 ok, I wasn't sure about that option Mar 23 15:50:03 I just wanted to make people aware the message could occur Mar 23 15:50:29 03Koen Kooi  07org.openembedded.dev * rc8df97b149 10openembedded.git/recipes/qt4/ (qt-4.6.2.inc qt4-embedded.inc qt4-x11-free.inc qt4.inc): Mar 23 15:50:29 qt4: sync package between -native, -sdk and target packages Mar 23 15:50:29 * also sed out path to ${WORKDIR} in target package Mar 23 15:57:29 hmm. Mar 23 16:00:23 03Martin Jansa  07org.openembedded.dev * r3e3bc649e8 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Mar 23 16:00:23 pyphonelog: fix SRCREV, sorry Mar 23 16:00:23 Signed-off-by: Martin Jansa Mar 23 16:09:55 * kergoth mutters Mar 23 16:12:13 03Koen Kooi  07org.openembedded.dev * r8cee206a83 10openembedded.git/recipes/x-load/ (x-load-git/beagleboard/name.patch x-load_git.bb): x-load git: move beagleboard to 1.44 Mar 23 16:20:58 03Martin Jansa  07org.openembedded.dev * r0ac4a54f9a 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Mar 23 16:20:58 libphone-ui: fix weak assignment in sane-srcrevs.inc Mar 23 16:20:58 Signed-off-by: Martin Jansa Mar 23 16:21:10 me hopes that ^ is last of series of mine stupid commits Mar 23 16:21:24 hehe Mar 23 16:21:46 JaMa|Wrk: you just want your name more in weekly reports :-D Mar 23 16:22:13 hehe Mar 23 16:22:43 then I should start using more sofisticated commit messages :) Mar 23 16:23:19 because with -m like this I would look stupid even with BIG number :) Mar 23 16:23:59 * kergoth realizes he doesn't seem to do recipe fixes much anymore, mainly invasive class stuff and the occasional cleanup Mar 23 16:24:23 and no more commits from work as this system doesn't know that if I select something with mouse I mean to copy it to clipboard Mar 23 16:24:39 hi, I get this message from oestat-client: "NOTE: oestats: error starting build, disabling stats" printing vars from oestats_send I get: vars {'build_arch': 'x86_64', 'metadata_branch': None, 'builder': 'ao2', 'metadata_revision': None, 'machine': 'a780', 'distro': 'minimal'} any clue of what's wrong? Mar 23 16:24:58 ao2: bitbake master? Mar 23 16:25:27 ao2: then you're probably missing Depends as Event.data is a bit shorter now :/ Mar 23 16:25:57 so it just doesn't work? I am using bitbake 1.8.18 Mar 23 16:26:33 that's not master, than maybe you should check why you have None for metadata_revision and metadata_branch Mar 23 16:26:46 Could someone look over this patch and apply it please? http://patchwork.openembedded.org/patch/1724/ Mar 23 16:26:57 maybe you need alias git-branch="git branch" Mar 23 16:26:57 alias git-log="git log" Mar 23 16:27:29 as at least before bitbake was using those names like git-branch which is not available on all systems Mar 23 16:28:03 This one as well: http://patchwork.openembedded.org/patch/1743/ Mar 23 16:28:04 JaMa|Wrk, let me try, it could very well be git-1.7 which broke stuff Mar 23 16:34:49 hi guys Mar 23 16:35:20 i'm from dream multimedia and i work with opendreambox 1.6 Mar 23 16:35:52 but i just can't believe that bitbake sits at "NOTE: Resolving any missing task queue dependencies" for about 5 minutes for the easiest tasks Mar 23 16:36:05 does anybody have an idea how i could accelerate that? Mar 23 16:36:26 you must be on an extremely slow box, and using bitbake 1.8 Mar 23 16:36:54 if you want to stay on bitbake 1.8, try cherry picking commit d715da0ab6fecb85d1bd4722dbc315a6ec079599 from master Mar 23 16:37:00 that speeds up runqueue generation a fair bit Mar 23 16:37:25 hopefully enough :) Mar 23 16:37:40 yeah 1.8 Mar 23 16:37:45 is that like outdated already again? Mar 23 16:38:03 no, just master has things it doesn't, as you'd expect for a development branch Mar 23 16:38:16 okay hm Mar 23 16:38:46 so my box is an amd 64 3800+ with 6gb ram Mar 23 16:39:14 but i'll try that cherry picking with bitbake ... mom Mar 23 16:39:18 mine is a core 2 duo with 4gb ram and it takes no more than 30 seconds Mar 23 16:39:20 http://git.openembedded.org/cgit.cgi/bitbake/patch/?id=d715da0ab6fecb85d1bd4722dbc315a6ec079599 Mar 23 16:39:29 thats the patch file you'll want, if you don't want to use git Mar 23 16:39:39 git seems fine Mar 23 16:39:46 JaMa|Wrk, it's not it. OE uses lower level commands than those and doesn't seem to be using the git- syntax either. I'll take a closer look I just thought it was a known issue. Mar 23 16:40:50 kergoth since it's a script do i not have to do anything to make it take into effect? Mar 23 16:40:55 like flush a cache or anything? Mar 23 16:41:33 okay id re-handles bitbake files Mar 23 16:41:36 it Mar 23 16:44:43 the bitbake metadata cache is tmp/cache/ by default, but i don't think you need to force the matter Mar 23 16:45:09 yeah i guess it's already using the updated script Mar 23 16:45:15 since it rebuilds the cache right now Mar 23 16:45:26 (that too takes forever compared to bb 1.6. btw ;) Mar 23 16:45:31 but enough with the bitching :) Mar 23 16:46:03 patches welcome. Mar 23 16:46:27 i am so busy making patches USING the toolchain... :) Mar 23 16:47:31 hmm Mar 23 16:47:40 NOTE: Resolving any missing task queue dependencies for over a minute already Mar 23 16:49:00 python's running at full cpu usage Mar 23 16:49:12 okay preparing runqueue Mar 23 16:50:04 not sure if it actually sped up anything... i didn't do an exact benchmark Mar 23 16:54:17 RP: thoughts on 1.10 release timeframe? i expect there are still some kinks to be worked out first, like the eventhandlers that rely on the old code, but beyond that.. Mar 23 16:57:27 kergoth: a bitbake -c clean of a single packet still takes.... real 6m20.694s Mar 23 16:58:44 out of which <= NOTE: Resolving any missing task queue dependencies took like 6m19 Mar 23 16:59:03 I dunno, takes nowhere near that here. I'd suggest opening a bug in the BTS Mar 23 16:59:24 where is that? Mar 23 16:59:30 read the topic of the channel. Mar 23 16:59:34 oh Mar 23 16:59:44 fair plan Mar 23 16:59:51 thanks :) Mar 23 17:00:28 Obviously, we need to do a lot more profiling of bitbake to find and fix the performance issues, but its a substantial time investment, I haven't had the time personally, I suspect the same is true for others that hack on the bitbake code Mar 23 17:00:41 and of course, there are really only like 3 of us that touch bitbake itself Mar 23 17:01:25 speaking of title, kergoth, dont you own this channel? Mar 23 17:01:28 most people just use it and OE, and never think to touch its code or the code in the OE classes. Mar 23 17:01:47 I think I have admin rights with chanserv for it, yeah, though I'm not th eonly one Mar 23 17:01:49 * kergoth checks Mar 23 17:02:02 kergoth: either spread the pleasure or fix the topic :-D Mar 23 17:02:46 kergoth, even people that do not hack on the code could profile to get an idea of hot spots Mar 23 17:03:31 yeah, true Mar 23 17:03:56 yay Mar 23 17:04:14 hmm, looks like only mickeyl and i have the permissions right now Mar 23 17:04:25 * kergoth hasn't touched chanserv in ages Mar 23 17:06:20 be an idea to spread that to the new TSC I guess Mar 23 17:06:39 thatd work, or the folks in leadership roles of the e.V. Mar 23 17:06:54 for now, you should have access, try /msg chanserv op #oe XorA Mar 23 17:07:28 * kergoth is rather disappointed with the small number of TSC nominees Mar 23 17:07:48 kergoth: cool, thanks Mar 23 17:07:53 we should run an ev membership drive again Mar 23 17:07:56 increase the pool Mar 23 17:08:04 * kergoth is rather disappointed with the small number of people that feel comfortable with the intricacies of the system beyond just building and fixing a recipe, too Mar 23 17:08:15 that's a good idea Mar 23 17:09:43 I do think OE needs more "community" Mar 23 17:09:49 rather than users Mar 23 17:10:42 I think thats more common a problem then you think Mar 23 17:10:52 and ubuntu is muscling in on us, and they have more marketing $$$ Mar 23 17:12:48 * kergoth isn't sure how he feels about that Mar 23 17:13:24 kergoth: filed @ http://bugs.openembedded.org/show_bug.cgi?id=5421 Mar 23 17:14:35 re Mar 23 17:15:31 Depends: update-rc.d, libwrap0, libc6, libgcc1, libblkid1, libuuid1 Mar 23 17:15:43 someone remember what removes package versions? Mar 23 17:16:20 there's a utility function in bb for breaking up rdeps with the version stuff, explode_deps Mar 23 17:16:24 iirc Mar 23 17:30:13 pls anybody help me, i neet it , http://effik.pastebin.com/Ah6L9DLT Mar 23 17:34:47 shit. all my packages got wrong Mar 23 17:34:59 anyway it shown build deps so Mar 23 17:35:14 ~die e2fsprogs-libs, die! Mar 23 17:35:15 * ibot murders e2fsprogs-libs, die!, takes two shots to the head then crumples to the ground, lifeless... Mar 23 17:36:31 we use e2fsprogs-libs to get libblkid and libuuid and now they are in util-linux-ng Mar 23 17:36:53 arg, i hate checksum mismatches Mar 23 17:39:15 hmm, would be nice if we had a way to utilize a .asc file on the server to verify with gpg rather than checksums, as an option Mar 23 17:39:17 03Stefan Schmidt  07org.openembedded.dev * r0e960229a3 10openembedded.git/recipes/lowpan-tools/lowpan-tools_git.bb: (log message trimmed) Mar 23 17:39:17 lowpan-tools_git.bb: Fix SRC_URI and update SRCREV. Mar 23 17:39:17 The old rev did not even build and to checkout a newer one we need to follow the Mar 23 17:39:17 URI changes rom upstream: Mar 23 17:39:17 | cc1: warnings being treated as errors Mar 23 17:39:18 | izconfig.c: In function 'do_set_short': Mar 23 17:39:19 | izconfig.c:186: error: dereferencing pointer 'sa' does break strict-aliasing rules Mar 23 17:39:21 * kergoth ponders Mar 23 17:42:39 shit. my commit was wrong Mar 23 17:44:09 * kergoth thinks about an option like bitbake -s but which shows all available versions Mar 23 17:44:54 kergoth: -ss? Mar 23 17:45:00 hehe Mar 23 17:45:26 which reminds me, i still want a commandline way to run bbshell commands, to have a command-based interface like svn/git instead of all the convoluted args Mar 23 17:45:34 bbshell build foo, or whatever Mar 23 17:46:58 CIA-2: wakeup Mar 23 17:46:59 03Marcin Juszkiewicz  07org.openembedded.dev * r7316e1c380 10openembedded.git/recipes/xorg-xserver/ (xserver-xorg-conf/bug20/xorg.conf xserver-xorg-conf_0.1.bb): Mar 23 17:47:00 xserver-xorg-conf: really add BUG 2.0 Mar 23 17:47:00 Signed-off-by: Marcin Juszkiewicz Mar 23 17:47:35 03Roman I Khimov  07org.openembedded.dev * r58b2dab3bd 10openembedded.git/recipes/ulogd/ (4 files in 2 dirs): Mar 23 17:47:35 ulogd: new recipe Mar 23 17:47:35 Signed-off-by: Roman I Khimov Mar 23 17:47:36 03Roman I Khimov  07org.openembedded.dev * r4ca8c88071 10openembedded.git/recipes/fileschanged/ (3 files in 2 dirs): Mar 23 17:47:36 fileschanged: new recipe Mar 23 17:47:37 Signed-off-by: Roman I Khimov Mar 23 17:47:38 03Roman I Khimov  07org.openembedded.dev * rcff666e4f8 10openembedded.git/recipes/gamin/ (gamin_0.1.10.bb gamin_0.1.8.bb): Mar 23 17:47:38 gamin: add missing dependency on glib-2.0 Mar 23 17:47:38 Signed-off-by: Roman I Khimov Mar 23 17:47:38 03Roman I Khimov  07org.openembedded.dev * r7e85836d47 10openembedded.git/recipes/fprobe/ (files/default files/init.d fprobe-ulog_1.1.bb): Mar 23 17:47:39 fprobe-ulog: new recipe Mar 23 17:47:39 Signed-off-by: Roman I Khimov Mar 23 17:47:40 03Roman I Khimov  07org.openembedded.dev * r2ae318d327 10openembedded.git/recipes/htb/htb-init_0.8.5.bb: Mar 23 17:47:40 htb-init: new recipe Mar 23 17:47:41 Signed-off-by: Roman I Khimov Mar 23 17:48:15 does Roman sits on irc? Mar 23 17:49:08 * kergoth wonders why some of those patches hit the list at all, does roman not have commit access yet? Mar 23 17:49:27 Crofton: I sent one patch for inclusion in stable/2009 - can you ack? Mar 23 17:49:37 kergoth: I think that we should vote for his r/w Mar 23 17:50:00 will write mail Mar 23 17:50:29 definitely, i think he's proven himself to be competent, if he doesn't have r/w yet he should Mar 23 17:51:18 He does, that's why there's rik/ branches Mar 23 17:51:56 he does? Mar 23 17:52:07 argh... I lost track Mar 23 17:52:18 * hrw -> brown bag Mar 23 17:52:20 Roman has r/w access I'd swear Mar 23 17:52:25 'cuz I know rik/ branches are his Mar 23 17:52:29 Or are we talking about someone else? Mar 23 17:54:43 03Roman I Khimov  07org.openembedded.dev * r55e25fb988 10openembedded.git/recipes/htb/htb-init_0.8.5.bb: Mar 23 17:54:43 Revert "htb-init: new recipe" Mar 23 17:54:43 Sorry, overlooked somehow that it's already in OE.dev. Mar 23 17:54:43 Signed-off-by: Roman I Khimov Mar 23 17:54:50 nevermind Mar 23 17:57:01 ah, then why does he seem to want approval on everything? i thought you only need acks on big changes to important recipes or classes Mar 23 17:58:10 hrm Mar 23 18:01:43 said kergoth_ to kergoth Mar 23 18:01:48 remind him to push the "easy" stuff Mar 23 18:01:49 hehe Mar 23 18:02:00 what about mwelch? Mar 23 18:02:54 +1 from me Mar 23 18:03:12 Crofton: ack my stable patch please: http://patchwork.openembedded.org/patch/1824/ Mar 23 18:03:24 maybe we should make the commit access process be we suggest people to the TSC? Mar 23 18:03:37 let me read it first :) Mar 23 18:03:43 mkey Mar 23 18:03:54 only breaks bug right :) Mar 23 18:03:59 * hrw -> daughter Mar 23 18:04:07 Crofton: it fixes bug20 not breaks Mar 23 18:04:09 ;) Mar 23 18:04:18 bb in ~1h Mar 23 18:04:36 :) Mar 23 18:04:37 done Mar 23 18:28:47 Different kernel names my CF-card differently. I'm sure I have seen something about configuring this but I don't remember where it was. Could anyone give me a hint? Mar 23 18:30:04 anyone know why opkg-utils isn't in OE build stats? Mar 23 18:31:21 I tried building it and QA picked up that it built for my host machine, not the target Mar 23 18:31:31 I'm trying to create an SDK, but can't figure out what package to depend on to include the linux headers (like omapfb.h and errorhandler.h). Sorry for repeating, but had to disconnect earlier before I could get any answer :) Mar 23 18:32:47 I tried linux-libc-headers-dev, but that was apparently not the one Mar 23 18:34:07 Some very board specific stuff might not be easily SDK'able today Mar 23 18:41:28 Hmm. Any other solutions to speed up development? I'd like to be able to compile a rootfs and create something sdk-ish that the other devs can use to create applications. Then I can bundle their apps and my rootfs in a package. Mar 23 18:42:57 Well, make linux-libc-headers use the same sources as your kernel does Mar 23 18:43:01 Rather than just vanilla upstream Mar 23 18:46:29 Tartarus: ah. I'll take a look. thanks Mar 23 19:06:29 hi Mar 23 19:19:06 re Mar 23 19:28:30 03Roman I Khimov  07org.openembedded.dev * rd8ce5c488c 10openembedded.git/recipes/htb-init/htb-init_0.8.5.bb: Mar 23 19:28:30 htb-init: fix ip location, use init script, update package_arch Mar 23 19:28:30 OE has alternatives setup for ip and the link is installed in /bin/ip, rather Mar 23 19:28:30 than /sbin/ip, so fix that. Also, setup rc links by default and set Mar 23 19:28:30 package_arch to all since it's just an init script. Mar 23 19:28:31 Update RDEPENDS since versioning doesn't work here. Mar 23 19:28:32 Signed-off-by: Roman I Khimov Mar 23 19:40:08 re Mar 23 19:40:13 no more Toshiba laptops Mar 23 19:41:39 my wife has Toshiba L30. worked fine under xp home but catched something and is not usable anymore. so I installed Debian on it Mar 23 19:42:58 disaster... rtl8139 ethernet works slowly only with irqpoll (which suxx), wifi (atheros) disconnects more often then connects so I plugged in ralink usb wifi... usbport got disconnected after few MB of data Mar 23 19:56:03 03Koen Kooi  07org.openembedded.dev * rf04722e241 10openembedded.git/recipes/mozilla/ (16 files in 2 dirs): firefox 3.6.2: add new version Mar 23 19:56:13 03Koen Kooi  07org.openembedded.dev * r771026d7d5 10openembedded.git/recipes/xbmc/libcdio_0.81.bb: libcdio: be more precise about license Mar 23 20:21:42 jovox, ping Mar 23 20:22:27 zenlinuxPDX, pong Mar 23 20:22:46 jovox, are you still having issues moving a CF card between devices breaking your networking? Mar 23 20:23:31 yes, just compiled an earlier kernel to see if I can get it working better. Testing as we speak Mar 23 20:23:57 I just thought I'd mention that I've encountered that problem before - and it had to do with the udev rule generation Mar 23 20:24:31 on bootup when you move the card to another system udev will see the different MAC address of your NIC and try to create a new ethernet interface Mar 23 20:24:46 and if your networking scripts assume that eth0 is your main interface, breakage can result Mar 23 20:25:04 oh.. Mar 23 20:25:22 you should have some .rules files in /etc/udev somewhere Mar 23 20:25:39 look for one with a name like "persistent-net.rules" or thereabouts Mar 23 20:25:39 So, udev stores this information somewhere.? Mar 23 20:25:41 and look inside it Mar 23 20:26:36 this may or may not be your problem, but I thought I'd suggest it Mar 23 20:27:06 ok, I will, right now my machine is a bit broken.. It doesnt boot beacuse I'm having trouble with the cf card has been renamed from sda to hdc.. as soon as I get it running I will have a look Mar 23 20:27:22 Thank you very much Mar 23 20:27:53 jovox, good luck Mar 23 20:33:39 zenlinuxPDX, there is one 70-persistent-net.rules Mar 23 20:33:54 are there multiple eth interfaces defined in it? Mar 23 20:34:18 10 Mar 23 20:34:22 bingo Mar 23 20:34:41 really? Mar 23 20:34:43 I'm trying to remember how I resolved this for a project at work Mar 23 20:35:05 I think I had a script which deleted the rules file - and that would force it to get re-generated on the next bootup Mar 23 20:35:21 so if you moved a card to a new hardware device, it would delete the file and reboot the system automatically Mar 23 20:35:27 total hack, but I recall it worked Mar 23 20:35:55 it was only during development that we'd move our CF cards into different hardware - it never happened in production Mar 23 20:36:16 oneshel, do you remember this? Mar 23 20:36:21 ok I see Mar 23 20:36:36 I will try to remove the file and move the cfcard to another machine Mar 23 20:36:46 zenlinuxPDX: the debian network issue when moving cards? yup, vaguely Mar 23 20:36:50 I just rm the file? Mar 23 20:37:02 jovox, I believe so Mar 23 20:37:10 oki, give me a minute Mar 23 20:37:11 might as well move it someplace else just to be safe so you can restore it Mar 23 20:37:35 yupp Mar 23 20:41:36 it works! Mar 23 20:41:57 zenlinuxPDX, thank you soo much! You saved my lot of time! Mar 23 20:42:12 it worked? eggselent! Mar 23 21:48:09 jovox: good news Mar 23 21:48:34 jovox: but what happened when you did ifconfig on the various machines then? Mar 23 21:50:41 it said ~"no such device" Mar 23 21:51:37 jovox: strange, because I faced to the same problem when I move HD amid different PCs Mar 23 21:52:19 jovox: but each time I do ifconfig (with no parameters) I see one ethN more (N++) Mar 23 21:53:02 kergoth: hi Mar 23 21:53:12 kergoth: One bitbake dev has taken time to profile bitbake ;-) Mar 23 21:53:37 But is also frequently overly overloaded? :) Mar 23 21:53:45 kergoth: Its actually pretty optimised now, the key future improvement is do less junk in the metadata ;-) Mar 23 21:53:51 Tartarus: well, yes Mar 23 21:54:06 mckoan, correct, but ifconfig eth0 said "no such device" but I think there were divices with higher numbers Mar 23 21:54:23 Tartarus: but I have done it and the results are still pretty much valid as those parts of the codebase haven't changed much until mayeb recently Mar 23 21:54:32 Tartarus: I'm all in favour of more people looking though :) Mar 23 21:54:37 when I unloaded and loaded the modules for the nic I saw that Mar 23 21:54:43 kergoth: As for 1.10, I think soon would be nice Mar 23 21:55:02 kergoth: I have ideas about stamp handling pending but I think those can wait for a 1.12 Mar 23 22:05:30 * kergoth_ nods, hmmm Mar 23 22:06:15 imagemagick-native doesn't compile anymore with autotools 2.65? Mar 23 22:06:25 I've profiled bitbake in the past, by the way, and worked on it in that way, but it was years and years ago :) Mar 23 22:09:29 kergoth: Right. Mine was more recent but before the recent parser changes. I don't think they've changed much though Mar 23 22:11:30 I think the better use of our time than that may be to focus on the metadata as you suggest, and maybe do more cleanup and refactoring in bitbake, make it easier for us to grasp and maintain, and pursue further performance stuff at that point Mar 23 22:11:34 * kergoth shrugs Mar 23 22:13:44 kergoth: I agree Mar 23 22:14:09 kergoth: but running the performance list now and again doesn't hurt. I think I even added an option for it :) Mar 23 22:14:23 yeah you did, though it interacts a bit oddly with other options Mar 23 22:14:31 * kergoth should run that on occasion Mar 23 22:14:46 kergoth: I think thats the client/server stuff breaking it :/ Mar 23 22:15:30 hmm Mar 23 22:15:51 kergoth: it shouldn't be hard to fix Mar 23 22:16:19 figured it wouldn't be, just one of those things that wasn't sufficiently annoying to dive into it, yet :) Mar 23 22:17:05 03Antonio Ospite  07org.openembedded.dev * r8ad0e50c77 10openembedded.git/classes/base.bbclass: (log message trimmed) Mar 23 22:17:05 base.bbclass: add missing import of metadata_scm Mar 23 22:17:05 In commit 89b7e433719f43f1c36c76cb8856d559014e99bc metadata_scm.bbclass Mar 23 22:17:05 functions were factored out from base.bbclass, but the new file was not Mar 23 22:17:05 included. Mar 23 22:17:06 This prevented METADATA_SCM, METADATA_REVISION, METADATA_BRANCH to be Mar 23 22:17:07 set; for instance this caused oestats-client stop working. Mar 23 22:20:24 kergoth: I know the feeling... Mar 23 22:21:07 03Khem Raj  07org.openembedded.dev * ref6e7b308e 10openembedded.git/recipes/geos/geos_3.2.0.bb: (log message trimmed) Mar 23 22:21:08 geos_3.2.0.bb: Refine PACKAGES and contents. Mar 23 22:21:08 * Create geos-c1 for C libraries Mar 23 22:21:08 * Forcibly package libgeos into geos package Mar 23 22:21:08 * Enhance description. Mar 23 22:21:08 * Add --disable-inline, its not needed if gcc 4.4 or newer Mar 23 22:21:08 is in use. but many distros still use older gcc 4.x Mar 23 22:25:01 ah rats, bbclassextend breaks packaged-staging Mar 23 22:25:16 check_stamp_fn is bust with it Mar 23 22:31:01 kergoth: I think bitbake could do with sharing a bit more info with tasks it runs Mar 23 22:31:25 kergoth: Any opinion on some carefully controlled set variables being added? Mar 23 22:31:51 kergoth: packaged-staging breaks as FILE != fn in the case of virtual classes :/ Mar 23 22:32:11 runqueue.check_stamp_fn doesn't like this Mar 23 22:35:23 kergoth: http://git.pokylinux.org/cgit.cgi/poky/commit/?id=589df80ca1b825afd138531bb796aa8617ed0991 Mar 23 22:35:37 I realise this isn't going to fly ;-) Mar 23 22:36:32 03Khem Raj  07org.openembedded.dev * rca018970a8 10openembedded.git/recipes/lzma/ (lzma-native_4.65.bb lzma.inc): Mar 23 22:36:32 lzma: Use BBCLASSEXTEND to get native recipe. Mar 23 22:36:32 Signed-off-by: Khem Raj Mar 23 22:40:52 can somone comment on that patch? http://patchwork.openembedded.org/patch/1827/ Mar 23 23:31:42 RP: Any idea why bitbake -n is so slow? Mar 24 00:22:46 Laibsch: sorry, 8:00 is one 1am of your time Mar 24 02:15:57 zecke: howdy, are you in Japan Mar 24 02:34:07 khem: no, Taiwan **** ENDING LOGGING AT Wed Mar 24 02:59:57 2010