**** BEGIN LOGGING AT Tue Jan 17 02:59:58 2012 Jan 17 09:13:52 morning all Jan 17 09:16:22 good morning Jan 17 09:32:43 good morning Jan 17 09:34:22 good morning everyone Jan 17 10:39:02 hello i m getting this error Jan 17 10:39:04 sed: can't read /media/eeecc9e6-2a1f-42f8-9ecd-f7ef700f894c/sanket/oe/build/tmp/tmp/work/i686-linux/autoconf213-native-2.13-r14.0/autoconf-2.13/bin/*.in: No such file or directory Jan 17 10:39:05 | ERROR: Function do_configure failed Jan 17 10:52:15 !help Jan 17 10:58:53 JaMa: ping Jan 17 11:54:04 reisei: pong Jan 17 12:05:49 JaMa: got an error "symbol lookup error: /usr/lib/egl/egl_dri2.so: undefined symbol: _glapi_get_proc_address" after installing mesa :-( can you advice me how to fix it? Jan 17 12:36:09 JaMa: no suggestions? Jan 17 12:38:12 cno Jan 17 12:38:14 sorry Jan 17 12:45:11 that's sad Jan 17 12:49:22 hi Jan 17 13:55:33 How can I specify the directory into which the archive specified by SRC_URI is extracted? Jan 17 14:00:30 How can I specify the directory into which the archive specified by SRC_URI is extracted? By default the files are extracted into the WORKDIR of the recipy and not into ${WORKDIR}/${PN}-${PV} Jan 17 14:03:21 freds: typically you would set S = "${WORKDIR}/whatever-path-gets-extracted" Jan 17 14:05:27 the archive doesn't have a root folder (and thus extracts everything to WORKDIR). Jan 17 14:06:15 I was sure there was an option for this but I can't find it at the moment Jan 17 14:07:17 ok, thanks Jan 17 14:08:00 freds: aha, I think you can add ;subdir=dirname to the SRC_URI line Jan 17 14:09:05 I guess it doesn't get used very often since most people tar up their sources in the traditional manner with a subdirectory inside Jan 17 14:09:43 bluelightning: hi there. Iirc we were using that mechanism for $LOGO_SIZE in oe-classic. Jan 17 14:09:57 I think we originally added it for things like glibc addons, which need to be unpacked inside the main source tree, but it should work equally well for this :) Jan 17 14:10:01 * kergoth yawns Jan 17 14:10:07 hi ant_work, kergoth Jan 17 14:10:12 hey Jan 17 14:10:29 ant_work: hmm, yes, I remember we talked about this... does it not work now? Jan 17 14:10:41 or does it just work differently? Jan 17 14:11:22 * ant_work checking old linux.inc Jan 17 14:11:42 I added ;subdir=${PN}-${PV}. It works! Thanka. Jan 17 14:11:59 ok, we gave SRC_URI = "file://${LOGO_SIZE}/logo_linux_clut224.ppm.bz2" Jan 17 14:12:03 and expected if [ -e ${WORKDIR}/${LOGO_SIZE}/logo_linux_clut224.ppm ]; then Jan 17 14:13:45 bluelightning: afaik that fetching in a separate dir did not work offhand Jan 17 14:14:00 with bitbake used in oe-core Jan 17 14:14:22 should probably open a bug if that's the case Jan 17 14:15:48 hm http://cgit.openembedded.org/openembedded/log/?qt=grep&q=subdir Jan 17 14:16:04 ah, yes, it was Tartarus Jan 17 14:26:30 kergoth: indeed I'll recheck. If my memory doesn't fail the issue was the file living in a subdir was fetched in $WORKDIR Jan 17 14:29:39 ant_work: well, the subdir option does work, I just tested it Jan 17 14:30:47 ant_work: was the issue we had earlier that with the old fetcher there was an implicit subdir if the file in SRC_URI was in a named subdir? Jan 17 14:30:58 yes, maybe in oe-classic the path was automatically repeated Jan 17 14:31:16 from src to dest Jan 17 14:32:56 hmm, not sure what to do about that... if we were to restore the old behaviour we'd have to assess what impact it would have on current recipes, if any Jan 17 14:33:08 probably was the fetc module v1, not fetcher2 Jan 17 14:33:12 yeah Jan 17 14:36:12 yeah, the behavior of the in-metadata unpacker was always to retain the subdir of the file URI in the destination. It was really a special case, though. I could see an argument for not restoring it, and instead telling people to use subdir= if its intended Jan 17 14:38:40 agreed Jan 17 14:38:58 sounds good to me also Jan 17 14:39:30 btw the list I posted above contains mostly obsolete recipes Jan 17 14:45:40 bluelightning: curiosly tzdata from meta-oe is using subdir while the corresponding oe-core recipe doesn't Jan 17 14:46:13 sigh Jan 17 14:46:19 why do we still have tzdata in meta-oe? Jan 17 14:46:27 ant_work: the recipe is doomed..I have a patch on the list Jan 17 14:46:33 bluelightning: ^^ Jan 17 14:46:53 ant_work: ah right, good, thanks for taking care of it :) Jan 17 14:47:11 I've almost finished cleaning the outstanding doubles ;) Jan 17 14:47:37 still hoping udev will disappear soon from meta-oe Jan 17 14:48:38 what replaced it, the bits included with the c library? Jan 17 14:48:56 no, the oe-core recipe, which was judged inferior Jan 17 14:49:21 ahh, right Jan 17 14:50:07 kergoth: talk me more about those bits... Jan 17 14:51:29 the tz- recipes seem to know '# libc is removing zoneinfo files Jan 17 14:51:29 from package' Jan 17 14:52:17 still actual? Jan 17 17:35:19 is there an IMAGE_CMD present to generate a directory instead of a tar.gz… I don't really see anything Jan 17 17:36:07 i suspect you would have to do something as root for all the files... Jan 17 17:36:20 trying to directly target something like NFS w/o the extra steps Jan 17 17:36:38 yep, you'd have to sudo which would be a problem due to not having access to the user tty to prompt with Jan 17 17:36:43 unless its passwordless, of course Jan 17 17:37:36 i suspected as much, thought i'd ask if anyone attempted it Jan 17 17:38:00 not that I know of. could definitely see it being useful, if the user is willing to trust it, of course :) Jan 17 17:38:26 maybe better to just have it make a tarball and have a separate script that does the rest of hte footwork from there Jan 17 17:39:34 kergoth: i was thinking about just adding to IMAGE_CMD_tar Jan 17 17:39:39 and just extracting it with a sudo Jan 17 17:39:56 there is some wasted time with the tarring though Jan 17 17:42:51 * kergoth nods Jan 17 19:06:14 * Tartarus starts using snapshot.debian.net way more to fix some checksum issues in 2011.03-maintenace Jan 17 19:23:02 So, silly question Jan 17 19:23:16 Anyone out there, that's not fetching from a mirror first, that's fetched icon-naming-utils-native in oe-core recently? Jan 17 19:41:36 before I import the package, has anyone alreayd done libcap-ng for oe-core? Jan 17 19:41:45 (or even oe classic?) Jan 17 19:51:28 libcap is the capability lib or the pcap lib? Jan 17 20:14:36 aloha Jan 17 20:17:51 pcap is pcap, cap isn't pcap :) Jan 17 20:17:53 afaik, anyway Jan 17 20:18:29 libcap (and libcap-ng) are posix capability libraries Jan 17 20:45:55 re Jan 17 20:47:58 ohayZ flo Jan 17 21:10:59 ok Jan 17 21:11:06 so it was the capability one Jan 17 22:02:48 florian: iwmmx indeed speeds things up Jan 17 22:03:07 cs_nbp *g* Jan 17 22:03:57 cs_nbp: very good... I would not have expected much of a difference. Jan 17 22:04:21 florian: well it's not like putting a dual core in it but its definitely a difference Jan 17 22:04:27 eg menu opens faster Jan 17 22:04:30 terminal too Jan 17 22:06:32 cs_nbp, does images with iwmmx work now? Jan 17 22:06:47 there were problems like with mplayer ending up in the wrong feed Jan 17 22:07:19 well mplayer is actually a little way to go Jan 17 22:07:23 so can't test Jan 17 22:07:27 ok Jan 17 22:07:32 but the rest seems to work Jan 17 22:07:33 I've no armv5 stuff Jan 17 22:07:39 pcmcia network, sd cards Jan 17 22:07:50 but it was a report by an user trying to bitbake shr images on ezx Jan 17 22:08:13 ah k, nice to know because i wanted to steal some recipes from shr Jan 17 22:08:21 the name was angelox or something like that he's in #openmoko-cdevel too Jan 17 22:08:32 but it was at oe-classic times Jan 17 22:08:44 nah got oe-core and meta-oe Jan 17 22:08:57 then he bought an n900 and it broke(hardware defect) Jan 17 22:09:08 omg Jan 17 22:09:10 indeed Jan 17 22:09:20 and n900 aren't cheap in south america Jan 17 22:09:27 it's like doubble of the price Jan 17 22:09:58 so maybe he'll get back to ezx, who knows Jan 17 22:10:27 :) ah yes now i remember the device Jan 17 22:10:58 slide kbd Jan 17 22:11:10 on n900 yes Jan 17 22:11:12 I have it too Jan 17 22:11:30 i never had any slide device, too afraid ill break it Jan 17 22:11:37 im great at breakin stuff Jan 17 22:11:47 htcdream has it too but I don't advise to buy it Jan 17 22:11:51 it's very unsuported Jan 17 22:12:16 yes i know, ill stay w my gta02 anyways Jan 17 22:13:05 ezx Jan 17 22:13:08 sounds familiar Jan 17 22:13:21 whats ezx? cant find useful links :D Jan 17 22:13:45 motorola linux phone platform Jan 17 22:13:47 thought so Jan 17 22:13:54 cs_nbp, yes gta02 is one of the best phones Jan 17 22:14:01 I had an E680 once upon a time :) Jan 17 22:14:02 ah ok, only find server landlords Jan 17 22:14:03 ultra-well supported Jan 17 22:14:06 t'was a fun phone Jan 17 22:14:23 GNUtoo: think so too, at the beginning it was a bit tough, but now im glad i stayed through it Jan 17 22:14:34 indeed lol Jan 17 22:14:37 wifi problems Jan 17 22:14:39 glamo problems Jan 17 22:14:41 etc... Jan 17 22:14:46 now most of it are solved Jan 17 22:14:58 glamo sources and specs leaked Jan 17 22:15:08 I know Jan 17 22:15:11 lookin forward to whats coming out of that Jan 17 22:15:17 now they are even officially released Jan 17 22:15:18 legally Jan 17 22:15:22 yeehaw Jan 17 22:15:27 *sshoot*shoot Jan 17 22:15:49 http://wiki.openmoko.org/wiki/Community_Updates/2011-12-01 Jan 17 22:15:56 anyways i dont understand the top secret polocy of some corps Jan 17 22:16:13 indeed neither do I Jan 17 22:16:26 for glamo they want to hide that it's a bad hardware maybe Jan 17 22:16:31 but everybody already knows it Jan 17 22:16:36 *g* Jan 17 22:16:42 so what? mygawd stand by ur product Jan 17 22:16:50 bunch of wussies Jan 17 22:17:03 the only option would have to get the doc out to get the driver improved early Jan 17 22:17:21 so there would be less users complaining Jan 17 22:17:26 and less bad publicity Jan 17 22:17:27 yeah imagine 3 years ago Jan 17 22:18:14 but at least they now opened it, there is some who take their secrets to the grave Jan 17 22:18:52 maybe the opened it because it was leaked? Jan 17 22:18:59 woglinde: btw whuts *g* mean ? Jan 17 22:19:16 cs_nbp *grin* Jan 17 22:19:17 bah was half reverse engineered anyways Jan 17 22:19:20 aaah Jan 17 22:19:23 :D Jan 17 22:20:48 I tough it meant something like gee Jan 17 22:21:11 like OMG Jan 17 22:21:20 yeah everybody learnin here Jan 17 22:21:23 :D Jan 17 22:21:47 I often makes mistakes with english expressions like heh, g etc... Jan 17 22:22:05 before I tough that heh means something like hey Jan 17 22:22:06 not a native speaker, too, eh? Jan 17 22:22:07 *meant Jan 17 22:22:11 indeed Jan 17 22:22:39 only 2300 tasks to go-.- Jan 17 22:23:02 btw theres kind of an emacs recipe floating around Jan 17 22:23:16 but currently only the fix for it is in openpandore repo Jan 17 22:23:30 cs_nbp, yes Jan 17 22:23:30 smone has more info? Jan 17 22:23:37 yes I was involved in it Jan 17 22:23:52 at oe-classic times Jan 17 22:23:58 emacs and emacs-x11 Jan 17 22:24:05 i only need nox Jan 17 22:24:25 where can i find it? Jan 17 22:24:36 in oe-classic Jan 17 22:24:37 need emacs and wanderlust quickly Jan 17 22:24:39 ah ok Jan 17 22:24:45 what's this openpandora emacs? Jan 17 22:24:51 can i import that as a meta-layer?^^ Jan 17 22:24:55 no Jan 17 22:24:58 just import the recipe Jan 17 22:25:07 in meta-oe Jan 17 22:25:10 and send a patch Jan 17 22:25:15 I'll be happy to try it Jan 17 22:25:20 k Jan 17 22:25:41 lisp extensions should install flawlessly, or should there be a recipe, too Jan 17 22:25:48 like eg wanderlust Jan 17 22:25:55 cs_nbp, note that now that there are per device sysroot it should be easier to point qemu to that Jan 17 22:26:00 (nice japanese-origin mail client) Jan 17 22:26:04 instead of assembling a sysroot from multiple parts Jan 17 22:26:17 hmm Jan 17 22:26:36 I think if it's nor part of emacs(the lisp extension) you should do a separate recipe Jan 17 22:26:46 but first get emacs running Jan 17 22:27:01 thats clear :) Jan 17 22:27:16 heres pandora http://204.124.182.13/p/pandora-oe/source/tree/08932f053357287239a6f0dc17d74981c107637c/packages/emacs Jan 17 22:27:27 looks different from yesterday -.- Jan 17 22:28:02 hmmm Jan 17 22:28:06 that's for oe-classic Jan 17 22:28:13 it's thir fork of oe-classic Jan 17 22:28:16 there is a meta-pandora now Jan 17 22:28:27 y and they only have an emacs patch Jan 17 22:28:31 so use the oe-classic one Jan 17 22:28:40 from github? Jan 17 22:28:40 because that one says: Jan 17 22:28:41 3 years 1 month Jan 17 22:28:51 y i know Jan 17 22:28:57 and they speak of errors Jan 17 22:29:00 from git.openembedded.org Jan 17 22:29:00 and breakage Jan 17 22:31:11 that one i think http://git.openembedded.org/openembedded/tree/recipes/emacs, and i just git clone that into eg recipes-apps of my distro? Jan 17 22:31:12 http://git.openembedded.org/openembedded/tree/recipes/emacs Jan 17 22:31:18 :D Jan 17 22:31:37 up high, down low, too slow Jan 17 22:32:06 no Jan 17 22:32:11 what you should do is that Jan 17 22:32:14 1)clone http://git.openembedded.org/openembedded/tree/recipes/emacs Jan 17 22:32:19 oops Jan 17 22:32:39 git clone git://git.openembedded.org/openembedded Jan 17 22:32:46 cd recipes/emacs Jan 17 22:32:56 copy with cp the emacs recipes you want Jan 17 22:33:04 into a layer Jan 17 22:33:10 like meta-openembedded Jan 17 22:33:17 at the right place Jan 17 22:33:19 and bitbake it Jan 17 22:33:21 it'll fails Jan 17 22:33:26 so fix the errors Jan 17 22:33:33 yeeh like usual ;) Jan 17 22:33:42 possibly update the recipe to an updated version of emacs while you're doing that Jan 17 22:33:49 and send a patch Jan 17 22:34:10 yeh ill try that Jan 17 22:34:17 right now Jan 17 22:34:22 ok nice Jan 17 22:34:27 ah no 200 tasks to go Jan 17 22:34:31 *2000 Jan 17 22:34:36 then ill try Jan 17 22:34:51 try to change the TREEDIR stuff Jan 17 22:34:54 in emacs.inc Jan 17 22:34:59 to point to sysroot Jan 17 22:35:16 more precisely the sysroot of your machine Jan 17 22:35:16 so i need emacs.inc in any case? wasn't sure and havn't finished reading the manual Jan 17 22:35:32 copy the whole emacs directory Jan 17 22:35:38 it'll be easier Jan 17 22:35:41 and remove what's not used Jan 17 22:35:45 or do the countrary Jan 17 22:35:49 copy emacs recipe Jan 17 22:35:54 and add what's needed Jan 17 22:35:59 it'll fails and tell you Jan 17 22:36:36 ur french? Jan 17 22:36:45 contraire gave u away :D Jan 17 22:36:59 im half french Jan 17 22:38:01 yes I'm french Jan 17 22:38:09 allez les bleues Jan 17 22:38:56 lol I'm not very sports-on-tv Jan 17 22:39:12 yeh me neither, starting with not having a tv Jan 17 22:39:35 my wife watches biathlon on the net, that's our tv experience Jan 17 22:39:38 lol Jan 17 22:39:39 ok Jan 17 22:41:07 damn forgot to bitbake virtual/libgl for webkit again Jan 17 22:41:30 suppose i could add a patch for that, too? Jan 17 22:47:16 yeah, tworaz has the switch to disable that mentioned in the jlime thread Jan 17 22:48:13 huhwhut? webkit? or all packages that fail? Jan 17 22:48:21 webkit Jan 17 22:48:54 i have the resolution in my fail.log Jan 17 22:49:08 but suppose it could be automatized Jan 17 22:51:05 so TREEDIR = "${WORKDIR}/qemu-treedir" should be /rightpath/oe/build/tmp-eglibc/work/sysroot/qemu-treedir ? Jan 17 22:53:34 ah i see, its in sysroots and then the right subdir :) Jan 17 23:32:15 smone knows how to get rid of google autocompletion quickly? Jan 17 23:32:52 cs_nbp: use yahoo Jan 17 23:33:00 :D Jan 17 23:33:11 or binggg Jan 17 23:33:40 y cause autocompletion draws 100% CPU on my pxa255 Jan 17 23:33:49 for 20 secs Jan 17 23:33:59 wow Jan 17 23:34:08 which browser Jan 17 23:34:13 midori Jan 17 23:34:23 128mb ram + 256 mb swap Jan 17 23:34:25 yeah well Jan 17 23:34:37 did you optimize for pxa Jan 17 23:34:52 since it has an extra coprocessor to do media Jan 17 23:34:57 i hope the repo maintainer did Jan 17 23:35:01 and I think gcc can use it if you ask it for Jan 17 23:35:39 or you mean the build? yes i build for machine netbookpro, that is pxa255 Jan 17 23:35:57 or you mean a third thing? :D Jan 17 23:36:21 well you must be using a sane gcc and then using right set of options Jan 17 23:36:49 I'm not that deep in it yet, at present only using premade recipes Jan 17 23:37:05 at least the recipe is for a distro on the right machine Jan 17 23:37:37 add mcpu=iwmmxt to your build and see if that helps Jan 17 23:37:44 it can simply bork too Jan 17 23:38:01 already have that in kernel config, speeds things up considerably Jan 17 23:38:02 ok Jan 17 23:38:33 I know mplayer uses it Jan 17 23:38:36 beuhargl, bing is from microsoft Jan 17 23:40:32 without javascript google is way faster Jan 17 23:41:15 bing autocompletion is 100 times faster than google o_O Jan 17 23:41:28 my build just failed... it was missing ofono during do_rootfs Jan 17 23:41:48 * florian wonders how this could happen Jan 17 23:41:50 hmm Jan 17 23:41:56 you need some stuff prebuilt Jan 17 23:42:02 like virtual/libsdl Jan 17 23:42:14 dont know if its that Jan 17 23:42:20 but i dont get ofono failure Jan 17 23:42:21 nope Jan 17 23:42:36 but I rmember I had to add it somewhere as a dependency Jan 17 23:42:54 when did u last pull? Jan 17 23:43:00 yesterday Jan 17 23:43:17 k me 2 hours ago Jan 17 23:43:35 my built went through with 2 errors "returning a non-zero exit code" Jan 17 23:44:45 ah no it failed, too -.- Jan 17 23:45:42 connman-gnome is what requires ofono Jan 17 23:46:00 ah k i have meta-gnome in my layers Jan 17 23:46:40 because i didnt change anything and for me it breaks at * extract_archive: Cannot create symlink from ./var/run to 'volatile/run': File exists. Jan 17 23:48:44 but wait Jan 17 23:48:47 * opkg_install_cmd: Cannot install package connman. Jan 17 23:55:49 ERROR: emacs: Recipe file does not have license file information (LIC_FILES_CHKSUM) Jan 17 23:56:09 how can I obtain such a license checksum? Jan 17 23:58:24 find which file in S has license information and use md5sum Jan 18 00:02:37 does LICENSE = "GPLv2" count as license information? and then the md5 goes into the recipe i assume Jan 18 00:07:50 no, that's in recipe I guess not file in $S, look for examples.. Jan 18 00:10:36 cs_nbp: NOTE: Tasks Summary: Attempted 4334 tasks of which 4332 didn't need to be rerun and 0 failed. Jan 18 00:10:56 florian: what, no var/run stuff? Jan 18 00:11:05 ok as a little cheat I removed *mplayer* Jan 18 00:11:12 cs_nbp: nope, nothing Jan 18 00:12:50 florian: apart old git checkout :D Jan 18 00:12:55 'old' Jan 18 00:13:20 had mplayer already out, now kicked xnoise Jan 18 00:13:26 extremely old ;) Jan 18 00:13:49 it isn't guilty Jan 18 00:13:55 I left it in... Jan 18 00:14:09 yeh it's connman Jan 18 00:14:13 prolly Jan 18 00:14:19 since it complained bout that Jan 18 00:14:23 just build ofono manually Jan 18 00:14:33 well ofono ran through Jan 18 00:14:40 should be easy to fix Jan 18 00:14:44 just not configuration of connman Jan 18 00:14:47 k Jan 18 00:15:00 Running task 5055 of 5059 Jan 18 00:15:05 o_O Jan 18 00:17:35 btw there seems to be only russian locale complete so far Jan 18 00:20:53 funny Jan 18 00:21:00 time to get a little rest Jan 18 00:21:04 Good night Jan 18 00:21:07 gn8z **** ENDING LOGGING AT Wed Jan 18 02:59:56 2012