**** BEGIN LOGGING AT Thu Oct 07 02:59:56 2010 Oct 07 03:19:08 need guidance on how to trouble shoot a recipe parsing error. Oct 07 03:19:33 need more details :) Oct 07 03:19:39 ideally, pastebin the exact error Oct 07 03:19:47 and the recipe, if possible Oct 07 03:20:10 exception being thrown from bitbake/lib/bb/cooker.py line 994 Oct 07 03:20:19 It happens parsing any recipe Oct 07 03:21:40 again, need more details Oct 07 03:21:46 "exception being thrown" tells me nothing Oct 07 03:22:26 http://pastebin.com/BaZTyAbS Oct 07 03:22:38 exactly, it tells me nothing also :) Oct 07 03:23:18 crank up parsing debugging Oct 07 03:23:21 -l Parsing -l Parsing -l Parsing Oct 07 03:23:31 alternatively, you can use -D, but no point cranking up debugging on everything Oct 07 03:24:38 bitbake -l Parsing -l Parsing -l Parsing pciutils Oct 07 03:24:38 ? Oct 07 03:24:44 correct? Oct 07 03:28:36 yeah Oct 07 03:28:42 also, what bitbake version? Oct 07 03:30:06 master from about a week ago Oct 07 03:31:56 commit bed8e09971dc577f5443ad3d89aa14634c54eb16 Oct 07 03:31:56 Author: Chris Larson Oct 07 03:31:56 Date: Tue Sep 28 08:52:32 2010 -0700 Oct 07 03:32:49 should be fine Oct 07 03:43:49 http://pastebin.com/a5uvWD6z Oct 07 03:44:16 I am not seeing anything useful from said -l Parsing -l Parsing -l Parsing command line. Oct 07 03:44:30 that's not including all the parsing messages Oct 07 03:44:31 odd Oct 07 03:44:32 This was obtained by redirecting stdout stderr to a file. Oct 07 03:44:58 bitbake -l Parsing -l Parsing pciutils > /tmp/dick 2>&1 Oct 07 03:45:21 yeah, thats only including the level 1 messages, must be a bug in the debug levels.. grr Oct 07 03:45:24 try -DDD Oct 07 03:47:00 you can also try 1.10, see if the behavior is different, if so, we know the bug lies between 1.10 and master Oct 07 03:47:10 thanks kergoth! http://pastebin.com/guAaAeuM Oct 07 04:04:52 kergoth: found a problem in one of my recipes: initscripts Oct 07 04:05:24 thanks for your help. But bitbake was of no help in pointing me to that. Seems parsing errors should give line no. and recipe file. Oct 07 04:06:26 Do you think anyone is interested in a new ppp-2.4.5 recipe, switched over to use CMake? Oct 07 04:09:29 dickelbeck submit Oct 07 04:10:09 a semantic analyser for bitbake ? is that what you want Oct 07 04:13:24 khem: how do I submit? you mean send a patch to the devel mailing list? Oct 07 04:14:45 khem: regarding parsing, the bitbake exception said it was a parsing error. A good thing for a parser to do is capture the line number and filename of the problem text. Oct 07 04:38:09 dickelbeck: it should be showing that info, and does most of hte time, not sure how to reproduce that behavior, but will attempt it Oct 07 04:44:23 dickelbeck: i'd be curious as to whether 1.10 errors in a more useful way Oct 07 04:48:11 here is the bad recipe file, to give you an idea http://pastebin.ubuntu.com/507758/ Oct 07 04:48:53 it was a conflicted merge using subversion. Oct 07 04:49:39 Can you give me the exact git commands to swap over to 1.10 in the same directory? Oct 07 04:50:02 or do you have what you need with the bad recipe? Oct 07 04:53:58 recipes/python/python-pygobject_2.20.0.bb do_configure failed With error "configure: error: could not find Python headers" How to resolve this error ? Oct 07 04:54:15 checking for headers required to compile python extensions... not found Oct 07 06:36:58 gm Oct 07 06:37:28 eFfeM_work, gm Oct 07 07:00:31 Should the mounting of removable ext3 media default to writeback instead of ordered? Everywhere on the web it says ordered is the default, but on my Beagle it's mounted as writeback Oct 07 07:04:54 is it the busybox mount that decides that writeback should be default? the mount.sh script used by udev seems just to use the defaults. Oct 07 07:11:03 Crofton|work: in ${TMPDIR} a file called legacy-staging.log is generated which lists the recipes that used legacy staging when building your image Oct 07 07:11:54 python-pygobject_2.20.0.bb failed to build Oct 07 07:12:03 error "configure: error: could not find Python headers" How to resolve this error Oct 07 07:12:08 Ananth_: it means it did not find python-dev most probably Oct 07 07:12:30 Ananth_: try to dig into config.log and find out which header it was looking for Oct 07 07:12:57 then we can trace which package should have provided that and add that to DEPENDS Oct 07 07:16:02 khem: while trying to build recipes/python/python-pygobject_2.20.0.bb recepe error was obtaind. And configure log shows that checking for headers required to compile python extensions... not found Oct 07 07:16:12 checking for python... /media/500_p4/open_embedded/overo-oe/tmp/sysroots/i686-linux/usr/bin/python Oct 07 07:16:49 03Martin Jansa  07master * r5c5e270e7b 10openembedded.git/recipes/linux/ (4 files in 3 dirs): Oct 07 07:16:50 linux(-kexecboot)_git: bump SRCREV for 2.6.36-rc7 Oct 07 07:16:50 Signed-off-by: Martin Jansa Oct 07 07:16:59 03Martin Jansa  07master * r40901c6b04 10openembedded.git/conf/machine/om-gta02.conf: Oct 07 07:16:59 om-gta02: upgrade preferred virtual/kernel from 2.6.32 to 2.6.34 Oct 07 07:16:59 Signed-off-by: Martin Jansa Oct 07 07:17:54 morning Oct 07 07:31:40 khem: In log it was looking on python extensions and the result was not found Oct 07 07:31:51 Where to add python-dev Oct 07 07:39:27 anyone any idea how to easily mask off some bits from a numerical value in bash? expr does not seem to have bitwise operators Oct 07 07:55:34 03Roger Monk  07org.openembedded.dev * r02750772e0 10openembedded.git/recipes/ti/ (3 files): Oct 07 07:55:34 ti-edma3lld: Add Latest EDMA3LLD v1 release + fixup SRC_URI + HOMEPAGE Oct 07 07:55:34 Signed-off-by: Roger Monk Oct 07 07:55:34 Signed-off-by: Koen Kooi Oct 07 07:55:41 03Roger Monk  07org.openembedded.dev * rc71ed1f946 10openembedded.git/recipes/ti/ti-paths.inc: Oct 07 07:55:41 ti-sysliink: Add initial syslink support Oct 07 07:55:41 Signed-off-by: Roger Monk Oct 07 07:55:41 Signed-off-by: Koen Kooi Oct 07 07:55:43 03Roger Monk  07org.openembedded.dev * r32336b9a0e 10openembedded.git/recipes/ti/ (ti-ipc.inc ti-paths.inc): Oct 07 07:55:43 ti-ipc: Add Initial Support for TI IPC (Inter Process Communication) Support Software Oct 07 07:55:43 * This contains IPC support software for uni- and multi- processor communications tools Oct 07 07:55:43 Signed-off-by: Roger Monk Oct 07 07:55:44 Signed-off-by: Koen Kooi Oct 07 07:55:56 03Roger Monk  07org.openembedded.dev * r859432d157 10openembedded.git/recipes/ti/ (3 files): (log message trimmed) Oct 07 07:55:56 ti-edma3lld-sysbios: Add new EDMA3LLD (v2.x) Oct 07 07:55:56 * Add support for EDMA3 Low Level Driver (v2.10.03.04) Oct 07 07:55:56 * This version is based on SysBios (aka BIOS6) and hence needs a different package name Oct 07 07:55:56 * Some targets supports both EDMA3LLDv1 and EDMA3LLDv2 Oct 07 07:55:56 * Allow side-by-side install (alongside v1) using different install paths Oct 07 07:55:57 Signed-off-by: Roger Monk Oct 07 07:55:57 03Roger Monk  07org.openembedded.dev * r2a7b6e70e9 10openembedded.git/recipes/ti/ti-xdctools_3.20.04.68.bb: Oct 07 07:55:58 ti-xdctools: Add XDCTools v3.20.04.68 Oct 07 07:55:58 * Add Latest Release Oct 07 07:55:59 Signed-off-by: Roger Monk Oct 07 07:55:59 Signed-off-by: Koen Kooi Oct 07 08:29:07 khem: I've found strange change in eglibc locale packages locale-base-en-us now RPROVIDES glibc-binary-localedata-en_US, before it was Provides: virtual-locale-en-us, virtual-locale-en, shouldn't it provide both? Oct 07 08:31:18 khem: prepared RDEPENDS are overwritten in output_locale_binary_rdepends() in eglibc-package.bbclass Oct 07 08:31:30 s/RDEPENDS/RPROVIDES Oct 07 08:34:17 What do I have to add to my iamge to get the qt4 powervr-plugin? I tried "libgles-omap3 devmem2 qt4-embedded-plugin-gfxdriver-gfxpvregl", but don't get any ipk for the plugin. Oct 07 08:38:37 khem: I'll try something like http://paste.pocoo.org/show/272111/ Oct 07 08:55:51 hey guys i just noticed a very misleading error message when i accidentally put "file://http://download-url" with the extra http:// in it! it would fail in task staging_helper /dream/oe1.6/dm800/build/tmp/work/mipsel-oe-linux/gcc-cross-4.4.4-r0.1/temp/run.staging_helper.17345: 796: Syntax error: "else" unexpected (expecting "then") Oct 07 08:56:08 shouldn't it rather complain about a malformed uri or something? Oct 07 08:59:50 fraxinas: on git head ? Oct 07 09:00:37 good morning Oct 07 09:00:43 hi florian Oct 07 09:02:26 fraxinas: actually i meant bitbake git head, there was a patch last night that does something with errors if fetch fails: Oct 07 09:02:27 http://cgit.openembedded.org/cgit.cgi/bitbake/commit/?id=12066bf508a1459b1971a7ce666799e1f0124e36 Oct 07 09:04:53 bitabke 1.8 Oct 07 09:05:04 bake even Oct 07 09:19:24 Is there a useful way to allow bitbake to include a file that is in a di Oct 07 09:19:29 fferent overlay? Oct 07 09:20:17 There was a mechanism to do this altering FILESPATHBASE basing on collections but this doesn't seem to work any more using layers Oct 07 09:53:58 florian: you can always use a hard coded path to the file in the other layer (I know, not too nice) Oct 07 09:54:39 or probably use the amend.inc or .bbappend mechanism Oct 07 09:54:47 eFfeM_work: that's what i'd like to avoid :) Oct 07 09:54:53 guessed that Oct 07 09:55:01 * eFfeM_work also dislikes hard coded paths Oct 07 09:55:42 just gave it a try... it seems like it extends FILESPATHBASE already Oct 07 10:18:52 03Martin Jansa  07master * r7b18d478d1 10openembedded.git/recipes/tasks/task-shr-feed.bb: Oct 07 10:18:52 task-shr-feed: add mokojeweled Oct 07 10:18:52 Signed-off-by: Martin Jansa Oct 07 10:18:59 03Martin Jansa  07master * r20e431e8ba 10openembedded.git/recipes/syslog-ng/ (syslog-ng_1.6.11.bb syslog-ng_1.6.8.bb): Oct 07 10:18:59 syslog-ng: remove old 1.6 versions Oct 07 10:18:59 Signed-off-by: Martin Jansa Oct 07 10:19:00 03Martin Jansa  07master * r0b4933c5b1 10openembedded.git/recipes/openmoko-3rdparty/mokojeweled_git.bb: Oct 07 10:19:01 mokojeweled: add recipe for new game Oct 07 10:19:01 Signed-off-by: Martin Jansa Oct 07 10:19:05 03Martin Jansa  07master * rb3bf82db70 10openembedded.git/recipes/syslog-ng/ (syslog-ng.inc syslog-ng_2.0.5.bb): Oct 07 10:19:05 syslog-ng: add .inc with common declarations and use u-a for initscript as rsyslog/busybos/sysklogd does Oct 07 10:19:05 Signed-off-by: Martin Jansa Oct 07 10:49:24 03Koen Kooi  07org.openembedded.dev * rd2939bad15 10openembedded.git/recipes/powervr-drivers/ (libgles-omap3.inc libgles-omap3/cputype libgles-omap3/rc.pvr): Oct 07 10:49:24 libgles-omap3: fixup ti816x Oct 07 10:49:24 * needs a better fix in cputype, but this will do for now Oct 07 10:49:24 Signed-off-by: Koen Kooi Oct 07 11:16:29 someone having time to peek at this and if ok give me a 2nd ack ? http://article.gmane.org/gmane.comp.handhelds.openembedded/37835 Oct 07 11:21:01 03Martin Jansa  07master * re27d80de88 10openembedded.git/ (2 files in 2 dirs): libeflvala: bump SRCREV for version compatible with latest vala Oct 07 11:24:02 eFfeM_work: to be really correct you should bump PR in those recipes using those bbclasses Oct 07 11:24:36 eFfeM_work: building image with those here, if it goes well (not sure about ${PN} usage in image.bbclass) I'll ack Oct 07 11:32:38 JaMa: great, didn't test very much as it was yesterday late when I discovered this Oct 07 11:32:48 and yeah bumping PR would be good Oct 07 11:35:31 75 recipes inherit image, cpan is inherited by 159 recipes, distutils by 199 and ruby by (surprise) 0 Oct 07 11:36:43 is it really worth the effort to bump all those PRs ? Oct 07 11:37:52 JaMa: btw was playing with recipe-sanity when I discovered this, got errors on files that seemingly looked good Oct 07 11:38:02 planning a recipe-sanity run on all recipes Oct 07 11:39:09 well when I've relpaced all those RDEPENDS/RRECOMMENDS/RSUGGESTS/RPROVIDES/RCONFLICTS/RREPLACES I was asked for PR bump all Oct 07 11:39:43 eFfeM_work: maybe not needed for images as they are probably called as bitbake targets so rebuilt ondemand Oct 07 11:39:52 mickey|office: is rubyextension.bbclass still of any use, saw you created it in 2005 but there still does not seem to be a user Oct 07 11:40:28 JaMa maybe there is a perl inc that will rebuild all, not sure about that Oct 07 11:41:03 but can bump if desired, will cost some time though Oct 07 11:42:16 I'll ack even without PR bump.. (planing to bump DISTRO_PR in near future) so it's up to you Oct 07 11:42:53 i'll ask on the list Oct 07 11:56:12 eFfeM_work: wah, that's old. I don't think anyone cares about ruby in OE these days Oct 07 11:56:26 there is still a ruby recipe Oct 07 11:57:13 koen updated it on sep 18, so I guess he feels it is useful Oct 07 11:57:40 but no rubyextension users Oct 07 11:57:48 yeah, noone is building extensions Oct 07 11:58:22 should we keep the class? if not can you cleanup ? Oct 07 11:58:40 (and if yes, I'll fix the RDEPENDS) Oct 07 12:10:56 03Koen Kooi  07org.openembedded.dev * r4797af3e03 10openembedded.git/recipes/powervr-drivers/ (libgles-omap3.inc libgles-omap3/rc.pvr): libgles-omap3: 10205 is still ES5, not ES6 Oct 07 12:13:02 eFfeM_work: see git log ttf-fonts/ttf.inc for explanation why i added _${PN} and then had to remove it Oct 07 12:14:09 eFfeM_work: the rest probably wen't unnoticed in my last cleanup because of missing space before = (but strange I guess I grepped it with and without space) Oct 07 12:14:26 i used grep -w Oct 07 12:14:49 butw your commit of un 2 also mentions those classes Oct 07 12:15:07 module and gtk-icon-cahce aare apparently fixed Oct 07 12:15:51 if it is without _${PN} what is the meaning? Does it apply to all generated packages ? Oct 07 12:15:57 anyway recipe-sanity does not like it Oct 07 12:16:47 eFfeM_work: ttf.inc has problem that packages using this include are creating many sub-packages Oct 07 12:16:57 and all subpackages should have that RDEPENDS Oct 07 12:17:02 yeah, did log -p and saw the comment Oct 07 12:17:21 without _${PN} it's applied to all created packages (like -dbg -static etc too) Oct 07 12:18:23 ok, got that (actually that tells me a PR bump is not that important as without it old packages will just RDEPEND etc on too much Oct 07 12:18:30 but no functional breakage Oct 07 12:19:26 yes, RDEPEND is kind of safe, RPROVIDES etc are a bit dangerous (in meaning that rebuilding those can fix things which were already broken) Oct 07 12:19:54 true Oct 07 12:21:19 actually i think for most of the list adding _${PN} would be ok, not too sure on fakeroot-native and nylon-feed Oct 07 13:43:14 eFfeM_work: hello Oct 07 13:43:21 I saw you mentioned nylon Oct 07 13:43:23 ant_work: hi Oct 07 13:43:34 So, does anyone know why libtiff 4.0.0beta5 is called 3.9.2+4.0.0beta5, to us? Oct 07 13:43:37 WHen it's not upstream Oct 07 13:44:02 (There's both 3.9.4 and 4.0.0beta6 that we should package up, but what we have today make that harder..) Oct 07 13:44:03 ant_work: yeah, see also the mail on the list (on RDEPENDS) Oct 07 13:44:07 well, yesterday Iwas grepping for MACHINE_KERNEL_VERSION = "2.4" Oct 07 13:44:19 and neon.conf has it... Oct 07 13:44:31 u..nylon.conf, sorry Oct 07 13:44:52 candidate for obsolescence? Oct 07 13:45:24 If I want to override the feeds for the Angstrom distro, what is the best way to do that? I want to provide my own feeds for my custom Angstrom build. Oct 07 13:45:35 eFfeM_work: SRCDATE := "20050527" Oct 07 13:46:39 ant_work: there is no maintainer in MAINTAINERS, but git blame lists Martin Dietze as having done a lot of things on that file on 2009-05-19 Oct 07 13:48:43 svolpe: that seems more of a question for the #angstrom channel, but you can do somethign like bitbake package-index and put your deploy/ipk dir on a webserver and put the feed url on it Oct 07 13:50:14 eFfeM_work: I already have all that set up I'm just trying to figure out how to automate inserting the correct feeds into /etc/opkg. I will check on the Angstrom channel. I'm working of of OE so that is why I checked this channel but I agree the Angstrom channel is probably the correct choice. Oct 07 13:50:51 svolpe: don't know the exact procedure for that, but I added a feed once manually Oct 07 13:50:54 ant_work: http://br1.einfach.org/wiki/NyLon Oct 07 13:51:42 eFfeM_work: thats what I have done in the past, I made a package to override the /etc/opkg stuff but it seems to me there should be a cleaner way than that since my approach is just a bad hack ;-) Oct 07 13:52:00 eFfeM_work: thanks for the info Oct 07 13:52:17 yw Oct 07 13:52:27 ant_work: http://herbert.the-little-red-haired-girl.org/en/nylon/index.html Oct 07 13:52:31 maybe drop him a mail Oct 07 13:52:46 yes, I'll try to be fair Oct 07 13:52:49 ;) Oct 07 14:01:35 I know this is not a machine support channel .. but if I got openembedded working on a new type of machine how do I get it added ... do I just start a thread on the mailing list Oct 07 14:01:37 seems not too bitrotten Oct 07 14:04:25 cargoudel: yes, send the patches on the ml using git send-mail Oct 07 14:05:57 okay back. http://pastebin.com/yNcJSdYu So I have that and when it runs the fetch it grabs a tar.gz and places it in the download. Something about how it does the checkout in bitbake causes it to not download main.cpp. As explained earlier the fetch command for SVN is in fetch/svn.py. Did someone here write that because I'm curious if it has like an exception list. Even if I delete the tar.gz file it fetches and run it again it fetches one that's Oct 07 14:07:31 and documentation for a machine should go on the wiki .... like if there are special instructions for the machine .... for example there is a custom kernel config and a custom kernel .... this is for the ts-7500 btw Oct 07 14:14:27 kergoth: Hello. Maybe you know this: http://pastebin.com/ipS7cdew Oct 07 14:15:23 all it does is runs svn co Oct 07 14:15:30 if you crank up the debugging, you can see the exact command it runs. Oct 07 14:15:37 there's no magic here Oct 07 14:15:45 How is it creating the tar.gz? Oct 07 14:15:55 by running tar? Oct 07 14:16:15 So it checks it out then tars it then puts it in downloads? Oct 07 14:16:33 yes Oct 07 14:16:39 line 162 of svn.py runs tar Oct 07 14:16:51 the only exclude is .svn, in current bitbake master, previous had no excludes, it tars everything Oct 07 14:17:32 How do I crank up the debugging? Oct 07 14:18:03 we have documentation for a reason, both bitbake user manual and oe user manual Oct 07 14:18:09 or you could type bitbake -h and see it right there.. Oct 07 14:19:30 dear OE developers! where can I find a problem "task do_populate_staging: Failed"? Oct 07 14:20:30 i have logfile, it tells nothing Oct 07 14:21:02 I know the package name where error signalled, but It have no problems... Oct 07 14:21:22 syntax problems, i mean Oct 07 14:21:51 the failed message is below the actual messages from the task about what happened, and the portion of the log Oct 07 14:22:03 also, do_populate_staging is old, are you on a stable branch or something? Oct 07 14:23:18 hmmm... stable_2009 branch, was downloaded at aug 27 Oct 07 14:23:35 k Oct 07 14:24:29 all logs that I have: Oct 07 14:24:31 ERROR: Function do_stage failed Oct 07 14:25:26 well, when bitbake runs that task, it runs the run. script that lives by the log. files Oct 07 14:25:33 so run that yourself, perhaps with sh -x Oct 07 14:25:39 and see exactly what its doing when it fails Oct 07 14:25:48 or use bitbake -D, which does that for you, iirc Oct 07 14:28:25 [dv@dvpc temp]$ ls -1 Oct 07 14:28:27 log.do_stage.25966 Oct 07 14:28:28 log.populate_staging_preamble.25966 Oct 07 14:28:30 run.do_populate_staging.25966 Oct 07 14:28:31 run.do_stage.25966* Oct 07 14:28:33 run.populate_staging_preamble.25966* Oct 07 14:28:39 yes, what about it? Oct 07 14:29:12 you just said that do_stage is failing, so i'd think that would be a good one to run.. Oct 07 14:29:24 running _preamble says: ... Updating, ... Updating. And stops. No errors Oct 07 14:29:59 ? Oct 07 14:30:06 i'm not sure what makes you think running the preamble would be useful Oct 07 14:30:13 do_stage says no errors. It doing something and then stops with return code =0 Oct 07 14:30:13 conisdering that wasn't mentioend in any of your errors Oct 07 14:30:31 well, like i said, those are the script bitbake runs Oct 07 14:30:54 both of them returns 0 Oct 07 14:31:02 yes, you said that already. Oct 07 14:31:29 what else can I do to diagnose? Oct 07 14:32:05 nothing that I know of. logs and run scripts are all there is, other than adding bb.note()s to the task or diagnosing by inspecting the tasks and functions being run Oct 07 14:32:21 and again, cranking up bitbake debugging, but i don't know that it will be much help Oct 07 14:33:02 I know the reciept that failed. But I don't see "populate_staging" definition in it Oct 07 14:33:22 recipes don't define everything. Oct 07 14:33:33 all recipes inherit classes, which define common functionality Oct 07 14:33:51 beyond the 'inherit' lines in recipes, all classes listed in the INHERIT variable, as well as base.bbclass, get loaded Oct 07 14:46:17 okay Oct 07 14:46:27 us letting bitbake run nonexistant tasks as though they were no-ops Oct 07 14:46:31 not good. Oct 07 14:46:53 typo the task name in its definition and now it happily runs and does nothing.. Oct 07 14:47:46 sorry for my bad english, what you mean? I'm running non-existant task? Oct 07 14:47:54 no, i'm not talking to you Oct 07 14:48:00 this channel is for more than just support Oct 07 14:56:49 nah with -v -DD it doesn't output the SVN command used. Oct 07 14:57:17 is svn.py running runfetchcmd, or its own system() calls? Oct 07 14:57:26 runfetchcmd knows to show the command if debugging, iirc Oct 07 14:57:45 I haven't studied the source code. I'm just trying to use it mostly. Oct 07 14:59:34 kergoth: I added a warning in bitbake in poky for that Oct 07 14:59:55 kergoth: Need to sync bitbake up again sometime soon given I have things mostly working again now Oct 07 15:00:08 kergoth: I pushed the fetcher bits of sstate so its all there now Oct 07 15:01:16 cool Oct 07 15:02:13 pb_: if you would, or someone else in the channel, i need an unban again :) Oct 07 15:02:48 kergoth: done Oct 07 15:04:35 kergoth_: OT: did you ever play with awesome WM? Oct 07 15:04:57 yeah, i'm using it on my work laptop now. haven't spent any time configuring it yet, still running stock Oct 07 15:04:59 but i'm loving it Oct 07 15:05:07 i love the vim-like navigation - hehe Oct 07 15:05:30 kergoth_: I am using ion3 but xrandr support is bad... awesome manages that very nicely Oct 07 15:05:39 ah Oct 07 15:05:47 kergoth_: but... can one configure different tile layouts? Oct 07 15:05:58 kergoth_: e.g. in ion3 i manually move tiles around, split things... Oct 07 15:06:27 you can move tiles around, but they always obey the layout scheme for the current workspace - which you can change at any time by clicking on the icon in the conrer of the screen Oct 07 15:06:39 from floating to various tiled layouts Oct 07 15:06:40 mickey|sports: ping? Oct 07 15:06:45 mickey|office: ping? Oct 07 15:06:51 not sure how much manual control you can have over it, think it always obeys the scheme Oct 07 15:07:07 kergoth_: yes, i clicked on them... did you ever try to create a new scheme? Oct 07 15:07:26 one thing that's kind of cool, though i haven't used it much, is that workspaces are just tags, and a window can have multiple tags, so show up in multipel specific workspaces where you want it Oct 07 15:07:28 not yet Oct 07 15:07:36 haven't sat down to really tweak it to my liking yet Oct 07 15:07:38 need to Oct 07 15:09:10 03Martin Jansa  07master * rd19b69a60a 10openembedded.git/recipes/qi/ (files/0001-Glamo-242-timings.patch qi_git.bb): Oct 07 15:09:10 qi: add patch for better glamo timings Oct 07 15:09:10 Signed-off-by: Martin Jansa Oct 07 15:09:23 i am at the cross road... fix ion3... try awesome... the xrandr support is nice... the tags look good to... one can name them Oct 07 15:10:42 the philosophical opposition to xft by the ion3 author always kind of grated on me, but other than that i loved it -- don't need xrandr at this time though Oct 07 15:11:08 isnt ion3 dead now Oct 07 15:19:02 khem: well, the author decided to torture himself and move to windows Oct 07 15:50:51 kergoth: So my svn is on revision 5 and I have PR = "r5" but it looks like from what I can tell after removing and cleaning everything it's grabbing revision 1 only. Revision 1 only had trunk,tags,branches as the folders and its the only thing checked out everytime. Oct 07 15:51:04 set SRCREV in your recipe Oct 07 15:51:08 PR sets the revision of the *recipe* Oct 07 15:51:14 not the revision of the *source* Oct 07 15:51:15 makes sense thanks Oct 07 15:51:25 PR needs to be bumped for all changes to the recipe that affect output, e.g. changes to packaging,e tc.. Oct 07 15:51:28 np Oct 07 15:51:40 good job isolating that one :) Oct 07 15:59:00 I have added python-native to DEPENDS in my recipe. But bitbake does not pick it up when i build it. Oct 07 15:59:12 The recipe works only when I do a >bitbake -b python-native and then do a >bitbake -b myrecipe. Oct 07 15:59:19 Any idea how i may specify bitbake to build python-native first and then "myrecipe" later? Oct 07 16:00:02 Hello everyone Oct 07 16:00:24 Questions, maybe someone tried putting nginx (httpd) under OE? Oct 07 16:00:42 I am having hard time looking for bitbake recipie or instructions Oct 07 16:26:04 gm Oct 07 16:27:35 hi all Oct 07 16:39:37 * kergoth ponders Oct 07 16:42:28 Does OE override the include directories? "error: X11/Xlib.h: No such file or directory" but I can plainly see /usr/include/X11/Xlib.h when I do locate Xlib.h Oct 07 16:43:37 anyone using overlays to append to COMPATIBLE_MACHINE? seems that amend.inc doesn't work for that Oct 07 16:44:37 foobaz, yes, it overrides includes so as to include target based libs, not host Oct 07 16:47:25 my isuse with augmenting COMPATIBLE_MACHINE using amend.bbclass must be that amend parses after anonymous functions and base.bbclass weeds out recipes based on COMPATIBLE_MACHINES in an anon function correct? Oct 07 16:47:54 yeah, that's correct Oct 07 16:48:18 so sounds like another case where bbappend may be the better choice for me Oct 07 16:48:23 tharvey: so if I want to use xlib would I need to download or copy those to say my repo or does OE have a standard way for using say xlib because "CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include" it knows something I don't. Oct 07 16:48:29 *if* anon python functions work in an amend.inc, and i'm not sure if that's the case, you could implement it yourself by raising the exception Oct 07 16:48:38 foobaz: xlib is in oe.. Oct 07 16:48:43 go look, its one of the 7,000+ recipes Oct 07 16:48:48 or grep Oct 07 16:49:06 CROSS COMPILE Badness means the makefiles are explicitly pulling in /usr paths, you need to fix that Oct 07 16:49:31 ah. You mean look in one of the recipes to see how they're using it in the .bb files? Oct 07 16:49:40 foobaz: your recipe that needs xlib would simply have an DEPENDS = "xlib" in it to make sure xlib is built and exists prior. Your recipe that is looking for xlib needs to use OE to look in STAGING vs host paths Oct 07 16:49:42 or do you mean it's an actual recipe? Oct 07 16:50:16 foobaz what is it thats looking for xlib in / of host? are you having an issue with an OE recipe or trying to cross-compile outside of OE? Oct 07 16:50:31 oh no that's was me being silly Oct 07 16:51:05 foobaz: ls openembedded/recipes/xorg-lib/ Oct 07 16:51:09 ta da Oct 07 16:51:30 kergoth, not sure what you meant by raising the exception? Oct 07 16:51:42 all COMPATIBLE_MACHINE is is a convenience Oct 07 16:51:51 under the hood it raises an exception that tells bitbake "this recipe isn't buildable" Oct 07 16:52:07 yes, I see where thats done in base.bbclass Oct 07 16:52:28 i'm saying if anonymous python functions work in an amend.inc, you could raise it yourself. but i'm not sure if thats the case Oct 07 16:52:34 and yes, this is where bbappend works better Oct 07 16:52:49 its properly integrated into the parsing, so it happens after the recipe but before finalizing (overrides, anonymous functions, etc) Oct 07 16:53:05 whereas amend.inc is parsed *from* an anonymous python function in amend.bbclass Oct 07 16:53:16 there's no telling which other anonymous python functions will run before vs after it Oct 07 16:53:48 right... in my case I'm trying to add a machine to COMPATIBLE_MACHINE as well so not raise the exception but ignore it Oct 07 16:54:13 ahh, right, you're hosed ;) Oct 07 16:54:30 if I recall, bbappend has a shortcomming of not looking at PV or something like that? Oct 07 16:54:42 bbappend goes by filename, it has to be an exact match Oct 07 16:54:58 ah, right... Oct 07 16:55:07 lets say you append to foo_1.0.bb with foo_1.0.bbappend, and upstream adds 1.1, and no version preferences are specified Oct 07 16:55:13 now it builds 1.1, without your appended bits Oct 07 16:55:22 right, so you need to lockdown your versions Oct 07 16:55:25 yup Oct 07 16:55:35 which can often be a good idea anyway Oct 07 16:55:41 agreed Oct 07 16:56:00 ideally we'd have a better interface for updating your locked down versions, a browseable interface Oct 07 16:56:07 "lets see, this has 2 newer versions *click*" Oct 07 16:56:08 heh Oct 07 16:56:12 Oct 07 16:57:07 anyone know if I should be worried about bitbake git head? bbappend is not in 1.10.0 right? Oct 07 16:58:07 master *should* be working fine, and more people using it means we can fix those issues sooner :) Oct 07 16:58:20 right, it was added after 1.10 released, and is a new feature, doesn't belong in the 1.10.x bugfix series Oct 07 16:58:58 someone mentioned syntax errors in recipes don't show useful errors from the parser at the moment for some reason, haven't reproduced that yet Oct 07 16:59:29 so your using bitbake git yourselef? Oct 07 16:59:34 yep Oct 07 16:59:44 if it breaks, come yell at me Oct 07 16:59:49 ok I feel better thx :) Oct 07 16:59:51 :) Oct 07 17:01:07 bitbake is in svn, not git? Oct 07 17:01:41 homepage goes straight to manual which doesn't cover where to get it :) Oct 07 17:02:05 svn is deprecated Oct 07 17:02:10 its on the oe git server now Oct 07 17:02:16 git://git.openembedded.org/bibake Oct 07 17:02:20 typoed Oct 07 17:02:22 heh Oct 07 17:02:31 thx Oct 07 17:02:50 http://developer.berlios.de/projects/bitbake/ is out of date Oct 07 17:07:14 the only thing in the svn is a file saying to get it from git :) Oct 07 17:24:29 03Steffen Sledz  07org.openembedded.dev * r65bd13c8c9 10openembedded.git/recipes/dbus/ (4 files in 2 dirs): Oct 07 17:24:29 dbus: backport SO_REUSEADDR handling for tcp dbusses from dbus git repo Oct 07 17:24:29 Signed-off-by: Steffen Sledz Oct 07 17:34:07 Is it possible to put an image in DEPENDS? Oct 07 17:34:44 What's your end goal? Oct 07 17:34:55 To put an initramfs image in my other image Oct 07 17:35:00 I just want to make sure it gets built first. Oct 07 17:35:12 See linux-kexecboot and linux-preboot for examples of how to have OE handle initramfs Oct 07 17:35:31 Those build the initramfs inside the kernel; I would like it separate Oct 07 17:36:24 Then it's an initrd not an initramfs, no? Oct 07 17:37:21 Tartarus: Well it depends on how you define it. It is still a cpio.gz archive that gets loaded at boot. Oct 07 17:37:27 It's just not statically compiled into the kernel. Oct 07 17:37:55 right and you want it to be used to figure out stuff before the real init goes? Oct 07 17:38:17 Yeah. Oct 07 17:38:28 k :) Oct 07 17:38:35 Yes, you should be able to do that Oct 07 17:38:43 Just see how it's done for normal initramfs Oct 07 17:38:46 just with DEPENDS = initramfs-image-foo ? Oct 07 17:38:49 inside my image? Oct 07 17:38:50 no Oct 07 17:38:58 lemme find the magic Oct 07 17:39:35 Oh yeah, is OE typically not used for x86_64 archs? Oct 07 17:39:45 I wouldn't mind a kexecboot for my main desktop computer Oct 07 17:42:14 So, in your image Oct 07 17:42:32 do_rootfs[depends] += "your-initramfs-image:do_roots" Oct 07 17:42:44 er, do_rootfs Oct 07 17:43:57 Tartarus: I will try that Oct 07 17:44:00 Thanks :) Oct 07 17:45:09 03Steffen Sledz  07org.openembedded.dev * rd1008e24d3 10openembedded.git/recipes/dbus/ (dbus/anon.patch dbus_1.2.24.bb dbus_1.3.1.bb): Oct 07 17:45:10 dbus: allow anon dbus connections at hipox machine Oct 07 17:45:10 Signed-off-by: Steffen Sledz Oct 07 17:51:23 03Frans Meulenbroeks  07org.openembedded.dev * r590c96185e 10openembedded.git/recipes/binutils/ (9 files in 5 dirs): Oct 07 17:51:23 binutils : moved unused files to obsolete dir Oct 07 17:51:23 Signed-off-by: Frans Meulenbroeks Oct 07 17:51:32 03Frans Meulenbroeks  07org.openembedded.dev * r7a00a76925 10openembedded.git/recipes/powervr-drivers/omap3-sgx-modules/texture-width.patch: Oct 07 17:51:33 powervr-drivers : moved unused files to obsolete dir Oct 07 17:51:33 Signed-off-by: Frans Meulenbroeks Oct 07 17:51:33 03Frans Meulenbroeks  07org.openembedded.dev * r83d23b9881 10openembedded.git/recipes/busybox/ (28 files in 7 dirs): Oct 07 17:51:33 busybox : moved unused files to obsolete dir Oct 07 17:51:33 Signed-off-by: Frans Meulenbroeks Oct 07 17:51:34 03Frans Meulenbroeks  07org.openembedded.dev * r87bef56c09 10openembedded.git/recipes/qpe-gaim/files/ (fix-compilation.patch libopie2.patch): Oct 07 17:51:35 qpe-gaim : moved unused files to obsolete dir Oct 07 17:51:35 Signed-off-by: Frans Meulenbroeks Oct 07 17:51:42 03Frans Meulenbroeks  07org.openembedded.dev * r723862ee99 10openembedded.git/recipes/visiscript/files/compilefix.patch: Oct 07 17:51:42 visiscript : moved unused files to obsolete dir Oct 07 17:51:42 Signed-off-by: Frans Meulenbroeks Oct 07 17:52:05 03Frans Meulenbroeks  07org.openembedded.dev * r07d4842135 10openembedded.git/recipes/qte/qte-2.3.10/deprecated/ (7 files): Oct 07 17:52:05 qte : moved unused files to obsolete dir Oct 07 17:52:05 Signed-off-by: Frans Meulenbroeks Oct 07 17:52:05 03Frans Meulenbroeks  07org.openembedded.dev * r02d509a55b 10openembedded.git/recipes/qt4/qt-4.6.0/ (2 files): Oct 07 17:52:06 qt4 : moved unused files to obsolete dir Oct 07 17:52:06 Signed-off-by: Frans Meulenbroeks Oct 07 17:52:07 03Frans Meulenbroeks  07org.openembedded.dev * rb300be99e3 10openembedded.git/recipes/v4l2apps/v4l2apps-20020317/flags.patch: Oct 07 17:52:07 v4l2apps : moved unused files to obsolete dir Oct 07 17:52:08 Signed-off-by: Frans Meulenbroeks Oct 07 17:52:57 ok 4 letters to do; about 40 dirs, guess this is done in a week or so Oct 07 17:56:35 eFfeM: sheesh. thanks again for doing this :) Oct 07 17:57:08 yw Oct 07 17:58:18 i've decided when i make recipe changes to seriously consider taking maintenance if the recipe is unmaintained, unless its a mass sed change Oct 07 17:58:26 s/maintenance/ownership/ Oct 07 17:58:43 heh Oct 07 17:58:56 you saw me punt any claim to xerces-c today :) Oct 07 17:59:29 03Stefan Schmidt  07master * r74e1d02aab 10openembedded.git/: Merge branch 'master' of git.openembedded.org:openembedded Oct 07 17:59:32 03Stefan Schmidt  07master * r9c39314133 10openembedded.git/recipes/cacao/cacao.inc: Oct 07 17:59:32 cacao.inc: Fix path to cacaoh staging Oct 07 17:59:32 Problem found and fixed by dfoley Oct 07 17:59:32 Signed-off-by: dfoley Oct 07 17:59:33 Signed-off-by: Stefan Schmidt Oct 07 17:59:35 yeah, I do something like that too, but some of my changes are more cosmetical Oct 07 18:00:20 btw still could use some opinions/acks on the RDEPENDS patch (including whether or not to bump PR for recipes inheriting those classes Oct 07 18:00:44 * kergoth has no opinion, expects the distros will have one eventually Oct 07 18:01:17 khem, you around? Oct 07 18:02:24 kergoth btw saw the discussion of COMPATIBLE_MACHINE a short while ago. if a recipe does not build because of C_M not matching I now get Noting PROVIDES followed by a traceback; is it possible to avoid the traceback? (and if I do a bitbake -b I only get the traceback but no other message) Oct 07 18:02:41 try to peek in the code for a solution but didn't have success Oct 07 18:09:25 that would be a bug Oct 07 18:29:58 eFfeM: can you remind me what was decided for a removal of old versions policy? Oct 07 18:30:02 * kergoth 's brain is failing Oct 07 18:30:22 kergoth, hatever you can get away with Oct 07 18:30:27 :-) Oct 07 18:30:30 haha Oct 07 18:30:32 so true Oct 07 18:30:34 be respectful of potential out of tree users Oct 07 18:30:52 understand the recipes you working with Oct 07 18:30:55 well, as long as compatibility is retained, probably fine Oct 07 18:31:03 in this case, just going to bump gzip, i doubt anyone cares Oct 07 18:31:09 i'm sure zcat works the same ;) Oct 07 18:31:12 heh Oct 07 18:31:19 There's always git checkout ..., cp recipes/... my-overlay/ Oct 07 18:31:20 seriously Oct 07 18:31:22 and, afk Oct 07 18:31:23 do the easy stuff first Oct 07 18:32:46 kergoth this was in the march tsc minutes, but linuxtogo seems nonresponsive Oct 07 18:33:05 think it belongs in with the policy pages Oct 07 18:34:39 here is what the TSC said: Oct 07 18:34:40 http://pastebin.ca/1957053 Oct 07 18:35:03 sad that i can't remember Oct 07 18:35:06 considering i was in the discussion Oct 07 18:35:09 * kergoth rolls eyes at self Oct 07 18:35:16 you are getting old Oct 07 18:35:38 clearly Oct 07 18:37:09 hmm Oct 07 18:39:21 kergoth, posted an email on C_M with log Oct 07 18:39:24 now afk for a while Oct 07 18:39:28 cool, thanks Oct 07 18:39:30 will take a look Oct 07 18:40:42 03Chris Larson  07master * r4a26ee3d7e 10openembedded.git/ (4 files in 3 dirs): (log message trimmed) Oct 07 18:40:43 gzip: update to 1.4, add native Oct 07 18:40:43 - CVE-2010-0001 affects gzip < 1.4, so updated to 1.4. Also removed prior Oct 07 18:40:43 version, as no distributions appeared to pin it. Oct 07 18:40:43 - Added BBCLASSEXTEND="native" Oct 07 18:40:43 - Added myself as the maintainer in MAINTAINERS Oct 07 18:40:43 - Move all binaries into /bin, not just the 3 we use with update-alternatives. Oct 07 18:47:39 khem: started rebuild from scratch with your libtool patches and quite a few recipes are failing (not sure about real reason..) http://paste.pocoo.org/show/272395/ Oct 07 18:48:38 khem: mostly seems to trying to use wrong libpython.. /usr/lib/libpython2.6.so: file not recognized: File format not recognized Oct 07 18:52:06 Hi.. I recently installed gnuradio on beagleboard.. when I do opkg list-installed, it shows me gnuradio, but I am not sure which directory it is in.. Can anyone give any idea? Oct 07 18:58:42 sam2, this is probably better asked on the gnuradio channel :) Oct 07 18:59:04 but look in /usr/sharegnuradio/examples Oct 07 19:06:51 thanks Oct 07 19:07:02 @Crofton ..thanks Oct 07 19:09:51 re Oct 07 19:10:14 hi florian Oct 07 19:10:44 can you make a spreadsheet of finances for the eV GA? Oct 07 19:11:32 Crofton|work: should be easy :) Oct 07 19:12:21 yeah Oct 07 19:12:43 * florian needs to book a flight! Oct 07 19:15:50 khem, ping Oct 07 19:21:12 JaMa: cool Oct 07 19:21:18 JaMa: I have fixed few locally Oct 07 19:21:26 Crofton|work: hi Oct 07 19:21:30 wats up Oct 07 19:21:45 I still have the shamt issue Oct 07 19:22:28 whcih shmat.c file is used by eglibc for armv7? Oct 07 19:22:32 JaMa: that eglibc change looked ok too. Oct 07 19:22:36 I see three Oct 07 19:22:50 ./libc/ports/sysdeps/unix/sysv/aix/shmat.c Oct 07 19:22:54 ./libc/sysvipc/shmat.c Oct 07 19:22:56 not that one Oct 07 19:23:00 ./libc/sysdeps/unix/sysv/linux/shmat.c Oct 07 19:23:07 the last one Oct 07 19:23:13 I gueesed the aix one was bad :) Oct 07 19:23:43 hmmm was that being used ? Oct 07 19:23:53 dunno Oct 07 19:24:39 ./build-arm-angstrom-linux-gnueabi/sysvipc/shmat.o Oct 07 19:24:41 suggetss Oct 07 19:24:46 ./libc/sysvipc/shmat.c Oct 07 19:24:50 this one though .... Oct 07 19:24:57 hmmm Oct 07 19:25:30 Crofton|work: thats when you dont have the syscall Oct 07 19:25:46 yeah Oct 07 19:25:58 that is the .o O have Oct 07 19:26:01 I ahve Oct 07 19:26:12 * Crofton|work is not good at libc building and config Oct 07 19:26:29 Crofton|work: what kernel headers do u use Oct 07 19:26:41 whatever angstrom 2010 chose Oct 07 19:26:53 I suspect it is not makeingthe syscall Oct 07 19:27:15 I did an strace and I never saw the shmat Oct 07 19:27:17 call Oct 07 19:27:46 where would the headers get installed Oct 07 19:27:54 hmm Oct 07 19:27:56 hmm... noone tried to build on a machine running AIX yet, right? ;) Oct 07 19:28:06 is there en eglibc config log I can look at Oct 07 19:28:10 heh Oct 07 19:30:15 Crofton|work: do_compile.log Oct 07 19:30:22 khem: so should I push it with INC_PR bump? Oct 07 19:30:47 khem: I didn't understand why it's in rdepends function Oct 07 19:32:23 khem, not sure what I am looking at/for Oct 07 19:34:11 Crofton|work: that sysvipc looks ok Oct 07 19:34:45 JaMa: look into classes/glibc-package.bbclass Oct 07 19:36:09 Crofton|work: as a raw test copy ./libc/sysdeps/unix/sysv/linux/shmat.c to ./libc/sysvipc/shmat.c Oct 07 19:36:16 and recompile Oct 07 19:36:27 copy libc.so.6 to target Oct 07 19:36:36 and see if it makes things better Oct 07 19:36:40 ok Oct 07 19:36:51 we want to use the linux one Oct 07 19:37:09 not sure but it could be Oct 07 19:37:28 what is weird is that this used to work Oct 07 19:37:34 Crofton|work: if you did not have this problem before then it could be something else ending up here Oct 07 19:38:06 Crofton|work: ok you can strip gcc off all linaro patches Oct 07 19:38:07 the error message I get ties the failure to a return code from the shmat Oct 07 19:38:13 and see if it works again Oct 07 19:38:33 shmat may fail for several reasons Oct 07 19:38:41 if you have old working build Oct 07 19:38:48 then I would suggest to swap kernels Oct 07 19:39:01 and use one from a good build Oct 07 19:40:21 I've got a card here I can try in a bit Oct 07 19:40:34 the shm stuff in the kernel hasn't changed Oct 07 19:40:59 I'm confused why it appears eglibc is not using the linux syscall Oct 07 19:41:19 03Khem Raj  07master * rdba9226506 10openembedded.git/recipes/webm/libvpx.inc: Oct 07 19:41:19 libvpx.inc: Do not inherit autotools Oct 07 19:41:19 * Define own versions of do_compile and do_configure Oct 07 19:41:19 this package has configure but its not autconf one. Oct 07 19:41:19 Signed-off-by: Khem Raj Oct 07 19:41:30 03Zhai Edwin  07master * r071ed13f7f 10openembedded.git/recipes/lighttpd/ (4 files in 2 dirs): Oct 07 19:41:30 lighttpd: Add 1.4.28 as new recipe. Oct 07 19:41:30 Signed-off-by: Zhai Edwin Oct 07 19:41:30 Signed-off-by: Khem Raj Oct 07 19:41:55 Crofton|work: its using syscall see the build log it gets it on the fly during compilation Oct 07 19:42:47 ok Oct 07 19:42:50 weird Oct 07 19:43:09 Is OE a good way of making pxe bootable images? Or would I be better off looking at other more mainstream linux distro options? Oct 07 19:43:29 I wonder why I do not see it in the strace Oct 07 19:44:17 03Khem Raj  07master * r8f6eadc641 10openembedded.git/recipes/dhcp/dhcp4.inc: Oct 07 19:44:18 dhcp4.inc: We dont have LIC_FILES_CHKSUM Oct 07 19:44:18 Signed-off-by: Khem Raj Oct 07 19:44:18 kelvie_: with OE you can do almost anything Oct 07 19:46:56 03Khem Raj  07master * r8e1eef82a0 10openembedded.git/recipes/gcc/ (gcc-4.5.inc gcc-4.5/thumb-asm-out-of-range-pr43961.patch): Oct 07 19:46:56 gcc-4.5: Fix asm out of range errors seen on thumb when using -O1 Oct 07 19:46:56 Signed-off-by: Khem Raj Oct 07 19:52:59 are x86_64 machines supported well in OE? Oct 07 19:53:07 I don't see any machines that use them Oct 07 20:01:06 kelvie_: they are but they are just not upstream Oct 07 20:01:28 kelvie_: biggest opposition is the _ in arch name which conflicts with our override mechanism Oct 07 20:01:52 although we did not pay much attention to getting it included really Oct 07 20:01:59 I had patches posted last year Oct 07 20:03:50 x86_64 is not really a cpu used in embedded systems Oct 07 20:05:13 eFfeM: thats not true Oct 07 20:07:15 gettext native fails... is this a known issue? Oct 07 20:07:19 configure.ac:23: option `silent-rules' not recognized Oct 07 20:07:30 florian: new to me Oct 07 20:08:17 ok sounds like i will have to find out why this happens Oct 07 20:08:50 florian: whats your host make version Oct 07 20:08:58 * khem thinks florian is using 3.82 Oct 07 20:09:46 khem, ofc there are devices that use it; issue as far as I see it is that power consumption is too much for the typical embedded system Oct 07 20:09:49 close :-) Oct 07 20:10:01 khem: its 3.81 Oct 07 20:10:10 florian: that should be ok Oct 07 20:10:41 eFfeM: core routers running IP they are embedded systems too Oct 07 20:10:50 * eFfeM is used to developing for consumer devices Oct 07 20:10:51 eFfeM: all ATCA chassis Oct 07 20:11:00 khem ah ok, didn't think of those Oct 07 20:11:10 eFfeM: embedded is a wide area not limited to consumer electronis only Oct 07 20:11:35 all carrier grade stuff Oct 07 20:11:41 as i said I am used to consumer devices; nowadays I do industrial stuff but as it is used in mobile environments power usage is important too Oct 07 20:11:47 uses x86_64 quite a bit Oct 07 20:11:59 hi. Oct 07 20:12:07 does someone knows how to use MTD_OOPS ? Oct 07 20:12:10 true, didn't think of those (and nut sure if they are heavy OE users) Oct 07 20:12:26 alsa-utils is depending on build machine ncurses Oct 07 20:12:35 not good Oct 07 20:12:53 eFfeM: they are mind you MontaVista is using OE Oct 07 20:13:55 CURSESLIB=`ncursesw5-config --libs` Oct 07 20:13:56 heh Oct 07 20:14:12 that will never be correct on cross compiling env Oct 07 20:15:03 khem, I think I am going insane Oct 07 20:15:24 crfoton not good Oct 07 20:15:31 no Oct 07 20:15:34 wait Oct 07 20:15:35 Crofton: have patience system issues are hard nuts to crack :) Oct 07 20:15:47 too many usd card Oct 07 20:19:41 weird Oct 07 20:19:43 working Oct 07 20:19:46 arg Oct 07 20:20:01 Crofton: old kernel + new root file sysrem ? Oct 07 20:20:12 no, I can't find the failing card Oct 07 20:20:19 hrmmm Oct 07 20:23:20 * Crofton rocks back and forth in the corner Oct 07 20:24:31 Crofton, I use typex to number the cards and I have a file where I put what's in what card number Oct 07 20:24:41 the idea came from buglabs Oct 07 20:24:51 yeah Oct 07 20:24:56 I need to do something like that Oct 07 20:25:02 because I tried to do it before but nothing stayed long enough Oct 07 20:25:08 typex is fine Oct 07 20:25:38 * khem is hungry Oct 07 20:25:40 btw do you know how to debug early boot issues without serial? Oct 07 20:25:46 SIGFOOD bbiab Oct 07 20:25:53 lol ok Oct 07 20:47:38 03Michael 'Mickey' Lauer  07org.openembedded.dev * r72c2ac9400 10openembedded.git/recipes/libgee/libgee_0.6.0.bb: libgee 0.6.0: all versions depend on gobject-introspection now Oct 07 20:51:00 03Frans Meulenbroeks  07org.openembedded.dev * rc04571329f 10openembedded.git/recipes/barebox/barebox_git.bb: Oct 07 20:51:00 barebox_git.bb: added missing quote Oct 07 20:51:00 Signed-off-by: Frans Meulenbroeks Oct 07 20:51:21 03Frans Meulenbroeks  07org.openembedded.dev * r7484e1d174 10openembedded.git/classes/cpan-base.bbclass: Oct 07 20:51:21 RDEPENDS -> RDEPENDS_${PN} Oct 07 20:51:21 Signed-off-by: Frans Meulenbroeks Oct 07 20:51:21 03Frans Meulenbroeks  07org.openembedded.dev * r01d946dc27 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Oct 07 21:03:51 khem: can you push your other libtool-2.4 patches somewhere? Oct 07 21:04:23 khem: seems like even binutils is failing here now with: Oct 07 21:04:35 ERROR: QA Issue with staging: libopcodes.la failed sanity test (workdir) in path /OE/tmpdir-shr/sysroots/armv7a-oe-linux-gnueabi/usr/lib Oct 07 21:04:38 ERROR: QA Issue with staging: libbfd.la failed sanity test (workdir) in path /OE/tmpdir-shr/sysroots/armv7a-oe-linux-gnueabi/usr/lib Oct 07 21:22:01 JaMa: yes I am fixing that binutils stuff atm :) Oct 07 21:23:45 JaMa: thats the only one which is creating corrupt .la files in my console-image Oct 07 21:24:06 and thats because binutils uses its own version of libtool macros Oct 07 21:24:09 which needs some love Oct 07 22:00:13 03Michael 'Mickey' Lauer  07org.openembedded.dev * ra95ddbbc16 10openembedded.git/recipes/vala/ (vala_0.11.0.1.bb vala_0.11.0.bb): Oct 07 22:00:13 vala: 0.11.0 -> 0.11.0.1 Oct 07 22:00:13 NOTE: And *boom*, we are back to FSO-releases. Oct 07 22:05:58 03Khem Raj  07master * r6719909330 10openembedded.git/recipes/alsa/alsa-utils_1.0.20.bb: Oct 07 22:05:58 alsa-utils_1.0.20.bb: Use --with-alsa-prefix and --with-alsa-inc-prefix Oct 07 22:05:58 Signed-off-by: Khem Raj Oct 07 22:07:32 These commit messages look like mickey|sports had good vacations :-) Oct 07 22:12:31 hehe :) Oct 07 22:12:39 ya, feeling pretty relaxed Oct 07 22:15:48 :-) Oct 07 22:16:26 i should try the same :) Oct 07 22:47:03 khem, i've done a bitbake world with your libtool2.4 patch set (v2). Here's a list of tasks that failed, which previously succeeded: http://pastebin.com/iCb3jJw8 Oct 07 22:47:30 ah Oct 07 22:47:40 libtool2.4 patch set ? Oct 07 22:47:58 i need to write some python code to automate this and make it more robust (that was just manually awking and may have missed some). Oct 07 22:48:07 dj-death, see mailing list Oct 07 22:51:39 khem: there ? Oct 07 22:54:32 dj-death: yes whats up Oct 07 22:55:41 grg: cool can you post that as reply to my mail to ml ? Oct 07 22:55:48 that way we will keep the list Oct 07 22:55:50 sure Oct 07 22:55:52 in archive Oct 07 22:56:42 khem: I'm very interested by your resolve-sysroot.patch Oct 07 22:57:10 grg: I am getting binutils to behave once I have that then I have a console-image Oct 07 22:57:15 and I will try to boot it Oct 07 22:57:29 dj-death: its still WIP Oct 07 22:57:46 khem: I've seen a problem when using libtool 2.4 Oct 07 22:57:50 although I am not expecting changes to libtool itself Oct 07 22:57:51 khem: http://lists.gnu.org/archive/html/bug-libtool/2010-09/msg00064.html Oct 07 22:58:15 khem: did you make that patch to fix the same issue ? Oct 07 22:58:45 basicly compilation is failing with something like this : Oct 07 22:58:47 libtool: link: `=/usr/lib/libiconv.la' is not a valid libtool archive Oct 07 22:59:06 yeah there are two problems Oct 07 22:59:12 where '=' should be sysroot Oct 07 22:59:22 1. you have to pass --with-sysroot to every package that uses it Oct 07 22:59:33 and then you have to pass the same to libtool too Oct 07 22:59:38 when it is building itself Oct 07 22:59:42 I noticed the --with-sysroot option Oct 07 22:59:45 2. you need my patch Oct 07 22:59:58 there might be more coming I dont know yet Oct 07 23:00:37 1. is not really a problem ;) Oct 07 23:02:14 try with my patch Oct 07 23:02:18 you will be happier Oct 07 23:14:02 is there a known issue regarding failure to build gdb-cross-sdk? can't build meta-toolchain Oct 07 23:14:49 grg: hmmm btw. you ran it using old libtool right Oct 07 23:15:23 grg: or did you use new libtool 2.4 ? Oct 07 23:15:42 khem, i had LIBTOOL_VER = "2.4", LIBTOOL_HAS_SYSROOT = "yes" in my local.conf Oct 07 23:15:48 oh ok Oct 07 23:15:51 thats fine Oct 07 23:16:02 need to do one without i suppose Oct 07 23:16:38 yeah both are useful Oct 07 23:16:43 without is more important Oct 07 23:16:50 as we dont want to break existing stuff Oct 07 23:17:03 if you can make sure that without runs are same Oct 07 23:17:11 we can very well merge it into master Oct 07 23:17:25 and then fix 2.4 issues one by one Oct 07 23:17:34 * khem drives off Oct 07 23:50:48 trying to wrap my head around tinderbox to see if others have successfully built gdb-cross-sdk recently... If I goto http://tinderbox.openembedded.net/packages/?name=gcc-cross-sdk - I see version info that doesn't make sense Oct 07 23:52:04 ah wrong recipe... according to tinderbox there has 'not' been much success with 'gdb-cross-sdk' recently Oct 08 00:03:27 recipe or installing it? Oct 08 00:03:39 I updated some of the stuff that gets nuked recently so it installs Oct 08 00:30:05 Tartarus, looks like my git may be accidently out of date my script that creates it wasn't checking out the org.openembedded.dev branch correctly Oct 08 00:30:26 although I thought that the master branch was org.openembedded.dev? Oct 08 00:32:54 it is, they're the same Oct 08 00:33:11 one's a symbolic ref to the other Oct 08 00:36:26 ah, ok - well I was definately a bit out of sync, pulled now and will kick off a build thx **** ENDING LOGGING AT Fri Oct 08 02:59:57 2010