**** BEGIN LOGGING AT Sun Nov 21 02:59:58 2010 Nov 21 03:20:55 * sigh * Nov 21 03:20:57 dunno what I did Nov 21 03:21:01 but now my lifecam isn't detected Nov 21 03:21:03 * sigh * Nov 21 03:21:05 goddamnit Nov 21 03:29:59 obi: ping Nov 21 03:30:13 xmms configure fails Nov 21 04:02:45 ERROR: procps-3.2.8: http://procps.sourceforge.net/procps-3.2.8.tar.gz has no entry in conf/checksums.ini, not checking URI Nov 21 04:02:57 I try to compile the stablw2009 Nov 21 04:03:09 anyone have an idea why? Nov 21 05:46:36 Crofton: how did u nail it to xmms Nov 21 07:03:56 03Khem Raj  07master * ref2d81873b 10openembedded.git/recipes/openssl/ (6 files in 3 dirs): Nov 21 07:03:56 openssl: Add mips64{el}-linux target awareness Nov 21 07:03:56 Signed-off-by: Khem Raj Nov 21 07:32:48 does anyone here know if oe-dev kernel is setup to support NEON? Nov 21 07:45:39 lostincake: yes if you choose a machine with NEON it should be enabled Nov 21 07:47:57 Hey zecke Nov 21 07:52:38 khem: sweet Nov 21 07:52:56 haven't tried anything neon yet, but using the neon flags let the compiler compile and executable run Nov 21 08:38:58 gm Nov 21 08:39:36 ge Nov 21 09:00:44 khem: I've suspected xmms when building angstrom-gnome-image Nov 21 09:01:12 then just tried bitbake xmms with same effect Nov 21 09:01:27 morning Nov 21 10:00:31 morning Nov 21 10:02:18 morning Nov 21 10:19:30 ERROR: procps-3.2.8: http://procps.sourceforge.net/procps-3.2.8.tar.gz has no entry in conf/checksums.ini, not checking URI Nov 21 10:19:33 I try to compile the stablw2009 Nov 21 10:19:36 anyone have an idea why? Nov 21 11:14:05 concerning handhelds's cvs Nov 21 11:14:30 mirroring the sources on sources.openembedded.org is good but it's not used as a default for cvs Nov 21 11:14:54 in order to use it we need to add a PREMIRRORS() {cvs://.*/.* http://sources.openembedded.org/} Nov 21 11:15:12 or to used own-mirrors 's class with SOURCE_MIRROR_URL set to http://sources.openembedded.org/ Nov 21 11:15:46 I've launched a build from scratch using jay's testbuild script and it always fail on update-rc Nov 21 11:16:01 http://tinderbox.openembedded.net/builders/eukrea/ Nov 21 11:20:55 btw.. Nov 21 11:21:12 I should place updated version into contrib Nov 21 11:21:36 Jay7: i've downloaded the 0.1 from your site Nov 21 11:22:00 one problem : not possible to set IMAGES_minimal-uclibc="blabl" in the .conf Nov 21 11:22:28 hm.. you can define build set "minimal-uclibc" Nov 21 11:22:38 at least bash complains of this I don't know if that affects the behaviour or the builder Nov 21 11:23:19 what is bash message? Nov 21 11:25:00 Jay7: seems I can't reproduce it now Nov 21 11:25:03 strange Nov 21 11:25:21 ericben: feel free to bother me about it :) Nov 21 11:26:25 bash: IMAGES_minimal-uclibc=console-image meta-toolchain: command not found Nov 21 11:26:40 I have minimal-uclibc in DISTROS Nov 21 11:26:54 and IMAGES_minimal-uclibc="console-image meta-toolchain" Nov 21 11:27:49 Jay7: this is the error log (3 lines up) Nov 21 11:28:15 ericben: pastebin your .conf please Nov 21 11:28:17 anyway : my test build are all failing because of update-rc Nov 21 11:29:20 Jay7: http://pastebin.com/bap2Xa0J Nov 21 11:30:29 angstrom-2010 will not autoconverted to angstrom_2010 :) Nov 21 11:30:48 DISTRO_NAME_angstrom_2010="angstrom-2010.x" Nov 21 11:30:56 ok Nov 21 11:31:43 and images are expanding per build set.. Nov 21 11:32:54 not in you ebuild Nov 21 11:32:58 oeps Nov 21 11:33:27 so you can define build set "uclibc-minimal", include there all machines you needed and then use IMAGES_uclibc-minimal for defining images to build for that build set.. Nov 21 11:33:27 patchwork still down? Nov 21 11:33:47 Jay7: ok . This buildset concept is in OE or in your build script ? Nov 21 11:33:51 cazze: seems so Nov 21 11:33:53 cazze, down? Nov 21 11:33:59 ericben: in my script only Nov 21 11:34:15 Jay7: ok, I should study your script more in details before using it then ;-) Nov 21 11:34:22 ka6sox-work: i can't go to the web page to get some patches :-) Nov 21 11:34:41 hmmm...let me log in and see what is up. Nov 21 11:34:54 ericben: build-sets are expanded by distros, machines and images are expanded by build-sets Nov 21 11:35:40 that was sufficient for me, but if you have other opinion, just say me :) Nov 21 11:36:06 Jay7: no this will be fine I think, I only need to write the right config file :-D Nov 21 11:36:17 yeah.. Nov 21 11:37:51 side effect of buildset is building all included machines in one TMPDIR Nov 21 11:39:04 Jay7: TOOLCHAIN_TARGET_TASK_minimal="task-sdk-base" Nov 21 11:39:12 will not be used if I nuderstand well ? Nov 21 11:39:14 ericben: this will not work at all Nov 21 11:39:19 yes :( Nov 21 11:39:22 ok I understand well ;-) Nov 21 11:39:35 you can use it as local per-config variable only :) Nov 21 11:39:59 so, e.g. IMAGES="console-image $TOOLCHAIN_TARGET_TASK" :) Nov 21 11:40:26 config is just shell script too :) Nov 21 11:40:52 Jay7: yes I've now seen that Nov 21 11:40:55 btw, you can whitelist this variable in testbuilder Nov 21 11:41:06 then it will be passed to bitbake Nov 21 11:41:09 Jay7: yes, I'm playing with this now Nov 21 11:41:51 btw, 'group' == 'build-set' in comments :) Nov 21 11:43:40 hehe.. Nov 21 11:43:53 ericben: '-' can't be used in shell variable name Nov 21 11:44:02 Jay7: seems so Nov 21 11:44:25 that' why you have angstrom_2008 and not angstrom-2008 ;) Nov 21 11:44:48 yes, and that why I need DISTRO_NAME :) Nov 21 11:46:13 obi, Crofton: xmms build was failed on do_configure Nov 21 11:46:32 http://tinderbox.openembedded.net/packages/1014817/ Nov 21 11:56:35 patchwork seems up now...needed kicking Nov 21 11:58:04 ka6sox-work: nice, thx Nov 21 12:12:54 ERROR: procps-3.2.8: http://procps.sourceforge.net/procps-3.2.8.tar.gz has no entry in conf/checksums.ini, not checking URI Nov 21 12:12:58 I try to compile the stable2009 Nov 21 12:13:00 anyone have an idea why? Nov 21 12:13:08 and how to fix it? Nov 21 12:25:28 j24: create checksum and add it to conf/checksums.ini Nov 21 12:25:51 I can't remember what format was in that file Nov 21 12:26:14 current oe.dev have no that file Nov 21 12:26:32 stable2009 is dead Nov 21 12:28:45 hmm, are recursive package (ipkg) install dependencies considered okay ? i can think of two possible interpretations: if package A needs B to run and B needs A to run, all is fine, as long a we install A and B together. however, if A needs B to run and B needs A to be installed, that wouldn't be so good. as far as i understand things, this distinction isn't being made, correct ? Nov 21 12:53:53 ericben: backreading the log, I have submitted a patch for update-rc.d to the list Nov 21 12:54:06 davidlt: any progress on php? Nov 21 12:54:25 Zero, 18th + latest u-boot compiled Nov 21 12:54:48 Haven't tested yet if it works, but flashed the chip and gonna do that quite soon Nov 21 12:54:48 good morning Nov 21 12:54:58 hi ant Nov 21 12:55:10 hey eFfeM Nov 21 12:55:17 noticed the new breakage? Nov 21 12:55:27 davidlt: was it php or php-native you had problems with Nov 21 12:55:30 4 images on 4 failed tonight :/ Nov 21 12:55:48 ant_: no, why do they break ? Nov 21 12:56:18 eFfeM: php-5.2.13 Nov 21 12:56:23 haven't bee too much online this morning Nov 21 12:56:48 davidlt: and php-native does build for you ? Nov 21 12:56:58 ls Nov 21 12:57:18 see e.g. http://www.pastebin.ca/1997864 Nov 21 12:57:36 2 days ago issue was not present Nov 21 12:58:52 eFfeM: right, it works just fine. php is not capable of finding any lib at all. Check last e-mail and mailing list, switching off flags you always end up with something else missing. Nov 21 12:59:52 eFfeM: something is broken, but what I don't know. Maybe paths, maybe some other package. Nov 21 13:00:40 eFfeM: I gave a quick look at run.do_compile and etc. It looked normal, all paths were given. Nov 21 13:00:42 davidlt: can you upload your config.log, config.status, config.guess and configure files these are in your work dir in php-5.2.13-r7.1/php-5.2.13 Nov 21 13:00:50 davidlt: e.g use filebin.ca Nov 21 13:01:13 eFfeM, the only related change is sysvinit I guess Nov 21 13:02:11 ant_: strange, I do not really see a patch that affects this (sysvinit could be but is I doubt it) Nov 21 13:02:19 you could try bisecting Nov 21 13:03:31 * eFfeM does not like packages with workaround in the name, sounds too much like "there is a better solution" Nov 21 13:03:56 eFfeM: darn, I had to delete all files because I run out of space. I could launch new build of php, but probably gonna take 2-4 hours to finish. Nov 21 13:04:34 eFfeM, I can't really find suspicious commits Nov 21 13:04:46 though, I was playing with bitbake master.. Nov 21 13:06:45 ant_: from task-base.bb: # bluez-dtl1-workaround above is workaround for bitbake not handling DEPENDS on it in Nov 21 13:06:45 # kernel.bbclass. It should be there as long as kernel-module-dtl1-cs is, but not longer. Nov 21 13:08:35 but this comment is there already for quite a while Nov 21 13:09:51 eFfeM, how is the performance of git pulls these days? Nov 21 13:10:16 ka6sox-work: no idea have Nov 21 13:10:23 haven Nov 21 13:10:30 havebn't tried it for quite a while Nov 21 13:10:35 ouch, that was hard :-) Nov 21 13:10:42 okay, thanks Nov 21 13:11:33 ant_: do you still have recipes/bluez/bluez-dtl1-workaround_1.0.bb ? Nov 21 13:21:36 eFfeM: I restarting php compilation and sending you links to requested files via e-mail. Nov 21 13:34:17 ok Nov 21 13:36:19 eFfeM, yes, bluez-dtl1-workaround is there Nov 21 13:37:07 ant_: strange because that it what it complains about Nov 21 13:37:30 ant_: can you build it ? Nov 21 13:38:18 Jay7: please try whether "inherit lib_package" in gtk+-1.2*.bb fixes that. or else: FILES_${PN}-dev += "/usr/bin/gtk-config" Nov 21 13:38:29 I'm trying to go back with bitbake master Nov 21 13:40:36 03Frans Meulenbroeks  07org.openembedded.dev * rc3b5b68541 10openembedded.git/recipes/flex/flex.inc: Nov 21 13:40:36 flex: added perl to dependencies Nov 21 13:40:36 Without it (rebuilding from packaged staging) I got: Nov 21 13:40:36 | /home/frans/oe/tmp_minimal/work/i686-linux/flex-native-2.5.35-r5.1/temp/run.do_configure.12240: Nov 21 13:40:36 /home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/bin/automake: Nov 21 13:40:37 /home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/bin/perl: bad interpreter: No such file or directory Nov 21 13:40:38 Signed-off-by: Frans Meulenbroeks Nov 21 13:42:24 eFfeM, the package itself builds Nov 21 13:43:55 afterwards the images build Nov 21 13:44:00 strange Nov 21 13:44:04 bbl Nov 21 13:50:48 ant_: guess there is a missing dep one way or another Nov 21 14:01:57 Jay7: ok with a clean config file your script worked fine until the server crashed ... more news tomorrow as I don't want to go to the office now ... -bash: /bin/ls: No such file or directory :-D Nov 21 14:02:23 hi all Nov 21 14:03:19 Jay7: sorry... forgot to answer last night. i have been in a little telco at this time but fixed these issues while i was talking :) Nov 21 14:05:16 nooo Nov 21 14:05:37 perl-native is a bad thing to add as a dep :( Nov 21 14:06:34 heh Nov 21 14:06:35 eFfeM: We want to blacklist perl-native from pstaging rather than add it as a dep to everything Nov 21 14:07:11 Unless we want to stop using host perl all together, things go right to hell Nov 21 14:35:41 So, um Nov 21 14:36:00 Can anyone explain why RDEPENDS makes sense in the context of meta-toolchain* and its deps? Nov 21 14:36:40 Jay7: what do u mean by "create checksum and add it to conf/checksums.ini" Nov 21 14:37:08 checksums go in the recipe now Nov 21 14:39:53 Crofton|road: btu I've to use a stable_2009 Nov 21 14:40:02 provide by the vendor Nov 21 14:40:07 so no choice Nov 21 14:40:15 ok Nov 21 14:40:37 look at conf/checksums.ini Nov 21 14:41:01 I believe what you need to add is in the messages from bitbake Nov 21 14:41:37 Tartarus, ah, you find the issue wrt perl threads? Nov 21 14:41:54 Crofton|road: ok Nov 21 14:42:20 ant_: I know there's a rats nest :) Nov 21 14:42:34 let me dig the logs for you.. Nov 21 14:42:48 Well, no point Nov 21 14:42:53 We really need to rethink perl Nov 21 14:43:26 And either (1) move towards using host perl and a local cpan config for perl modules we want, along with kicking making target perl be the only thing to use the perl-native we build Nov 21 14:43:36 (2) move perl-native up to the perl that _Everything_ uses Nov 21 14:43:49 fact is I cannot build on Gentoo Nov 21 14:44:15 build what? Nov 21 14:44:23 "This Perl not built to support threads" Nov 21 14:44:28 Oh Nov 21 14:44:32 and bitbake errors out Nov 21 14:44:42 No, I'm referring to the problems we see when perl-native is brought in from pstaging way way early Nov 21 14:44:57 * Tartarus is fixing the scripts that hard code a too long path to perl atm, hitting that anyhow now :( Nov 21 14:45:00 well, in my case seems not pulled Nov 21 14:45:09 is Gentoo perl the problem Nov 21 14:45:12 Crofton|road: I'll take a look later to rebase against the current one but I need to finish to write the kernel and barebox first Nov 21 14:45:37 sure Nov 21 14:46:36 Have problems with php, everything is in 'image' directory, but only few files are staged. I am missing php, php-cgi, etc. on my flashed chip. Nov 21 14:46:48 then again, I had issues with lzma-native Nov 21 14:49:38 Tartarus: then please tell me how to get flex building from packaged staging ? Nov 21 14:53:35 I feel if a package needs another package it should be listed in DEPENDS, for me that is the purpose of DEPENDS, not? Nov 21 14:54:03 eFfeM: blacklist perl-native Nov 21 14:54:22 perl is very special, that's why we have perl-runtime-native in conf/bitbake.conf listed under ASSUME_PROVIDED Nov 21 14:54:41 All sorts of stuff depends on host perl but we didn't want to deal with it Nov 21 14:55:26 Tartarus: sorry I don't know what you mean with blacklist perl-native Nov 21 14:55:33 See my email Nov 21 14:56:27 * Tartarus hates branches Nov 21 14:57:22 03Tom Rini  07org.openembedded.dev * r2a548400c6 10openembedded.git/recipes/perl/ (perl-native_5.10.1.bb perl-native_5.8.8.bb): Nov 21 14:57:22 perl-native: Mangle scripts to use /usr/bin/env perl Nov 21 14:57:22 Signed-off-by: Tom Rini Nov 21 14:57:39 why the hatred for branches Nov 21 14:57:53 I never remember to check which one I'm in when I commit :) Nov 21 14:58:01 heh Nov 21 14:58:02 ok Nov 21 14:58:45 Tartarus: why is the +PSTAGING_DISABLED = "1 not in the main recipe (or, why is your patch from jul 27 not in) ? Nov 21 14:58:46 Crofton|road, so, unrelated Q. In the context of meta-toolchain, does RDEPENDS on the stuff that we install into the toolchain make any sense to you? Nov 21 14:59:08 * Tartarus can't figure out why we did it that way, nor why suddenly folks see task-sdk-host (and other RDEPENDS bits) not being built Nov 21 14:59:20 well, kergoth_ has me all confused about DEPENDS/RDEPENDS again, so I am the wrong guy to ask :) Nov 21 14:59:28 heh Nov 21 14:59:29 true Nov 21 14:59:45 He's probably sleeping in or something :) Nov 21 15:05:11 eFfeM: say again? Nov 21 15:05:30 Why wasn't it in? see the whole thread Nov 21 15:05:51 It was an RFC and we didn't quite get agreement Nov 21 15:07:17 not sure on the agreement thing but definitely it is not concluded Nov 21 15:07:29 but left or right we have an issue Nov 21 15:08:01 btw i think i never saw issues with the length of the path to perl but this could be my setup Nov 21 15:11:13 should we add a limit on the path name where you may build ? or should we indeed have something back like cross-perl ?? Nov 21 15:11:42 Well Nov 21 15:11:55 I'd prefer to fix things to use /usr/bin/env Nov 21 15:12:22 Since we (Mentor) have to fix them to build things in our auto-builder Nov 21 15:12:34 (Hudson creates very long paths) Nov 21 15:16:04 Tartarus: seems a good plan to me (although i didn't bump into long path issues with hudson yet, maybe it also depends on the way it is installed) Nov 21 15:17:17 It's a little bit installed and a lot about how many axis you have Nov 21 15:17:31 In what I posted about on oe-members there's MACHINE DISTRO and BBPKGS axis Nov 21 15:17:38 and then there's an axis for the host type Nov 21 15:18:15 So it's building in /var/hudson/workspace/sb_openembedded_autobuilder/BBPKGS/meta-toolchain/DISTRO/minimal/MACHINE/qemumipsel/label/sb-u804/ for example Nov 21 15:18:33 toss in sysroot and a bit more and you're past the #! limit Nov 21 15:21:09 Tartarus: understood. I don't do matrix builds, I've just linearized the 6 builds I wanted to do Nov 21 15:21:15 but the env solution is fine with me Nov 21 15:26:47 How many build hosts do you have available? Nov 21 15:26:50 I bet that's part of it :) Nov 21 15:29:12 yes, we're still setting this up, I think currently we have two hosts, but not 100% sure about that, our IT guys are managing that Nov 21 15:30:29 basically the currrent setup meets the company's needs. I just asked if it was ok to use the free cycles during nights and weekend to compile some things for testing-next Nov 21 15:30:56 where I picked machines and recipes in which work at least has some interest Nov 21 15:31:23 Yeah, same deal here Nov 21 15:31:28 Except we've got a lotta resources :) Nov 21 15:31:42 (in part since this is a pool shared w/ other products) Nov 21 15:32:21 * eFfeM thinks Axon is a little bit smaller than Mentor ;-) Nov 21 15:32:59 :) Nov 21 15:33:16 just a little bit Nov 21 15:33:53 guess we have about 5 devs using OE Nov 21 15:59:44 03Frans Meulenbroeks  07org.openembedded.dev * re8f1549409 10openembedded.git/recipes/tasks/task-internetserver.bb: Nov 21 15:59:44 task-internetserver: added php-cgi and php-cli Nov 21 15:59:44 Signed-off-by: Frans Meulenbroeks Nov 21 16:44:36 * khem yawns Nov 21 17:11:42 03Khem Raj  07master * r0bf3d4989b 10openembedded.git/recipes/gcc/gcc-4.5.inc: Nov 21 17:11:42 gcc-4.5.inc: Add a space after mips64-nomultilib patch. Nov 21 17:11:42 Signed-off-by: Khem Raj Nov 21 17:17:15 hey khem Nov 21 17:17:25 sorry for putting the perl-native patch into 2010.12 first Nov 21 17:17:32 too early & forgot to check where i was Nov 21 17:18:29 plus side, just mmaybe a few more of these builds i've got going will finish and then i can start to turn on tinderbox :) Nov 21 17:19:11 Tartarus: hmm Nov 21 17:19:26 Tartarus: dont you use some sort of git gui Nov 21 17:19:29 that can help you Nov 21 17:19:56 and use the command I posted in my original email that will only push the ref to a given branch Nov 21 17:20:13 even if you have pending pushed for other branches Nov 21 17:20:16 they will stay local Nov 21 17:20:55 mips64 native-sdk-image booted on qemu and the compiler on target also seem to be able to compile a hello world Nov 21 17:21:01 yeah, i need to be more careful, heh Nov 21 17:21:23 visual tools are very helpful in such cases Nov 21 17:21:40 its easy to miss git branch :) Nov 21 17:21:46 also doing a 'git branch' before git commit :) Nov 21 17:22:13 well, you seem to miss that often :) Nov 21 17:22:16 * Tartarus waits and sees if a local build of mipsel+uclibc+nas-server-image fails in pixman Nov 21 17:22:21 I could imagine a wailing infant Nov 21 17:22:48 I wanted to build a media center kind of image Nov 21 17:22:59 which has a video player and audio player Nov 21 17:23:09 and can play almost all formats Nov 21 17:23:31 and prolly a browser for netflix Nov 21 17:23:38 do we have something Nov 21 17:23:45 already Nov 21 17:24:38 Tartarus: how many build can your server do / Nov 21 17:24:53 I mean how fast it can do stuff Nov 21 17:25:05 khem: seen http://pastebin.com/d0sALfab ? Nov 21 17:25:42 Tartarus: ok known problem :_ Nov 21 17:25:44 khem: So, I time'd qemuppc+minimal-uclibc+minimal-image at ~50min Nov 21 17:25:52 is that samba Nov 21 17:25:56 yeah, samba Nov 21 17:26:11 ok I have seen this problem and also reduced it I guess Nov 21 17:26:15 but no solutions Nov 21 17:26:29 so lemme see what I had as workaround Nov 21 17:26:30 Tartarus: put the git branch into your prompt Nov 21 17:26:34 easy enough, and resolves that issue Nov 21 17:26:48 kergoth: fantastic idea Nov 21 17:26:58 I have machine/distro in my prompt Nov 21 17:27:04 willl add this too Nov 21 17:29:37 kergoth: So, question about DEPENDS vs RDEPENDS. Is it really right in meta-toolchain to use RDEPENDS_${PN} for stuff to be installed? Nov 21 17:29:47 'cuz bitbake -g meta-toolchain then doesn't show task-sdk-host Nov 21 17:29:55 And then we fail Nov 21 17:29:59 https://github.com/kergoth/homefiles/blob/master/.zsh/conf.d/S55_git is what i use, but that's with zsh Nov 21 17:30:02 (also using the parallel branch, btw) Nov 21 17:30:30 you checked all 3 .dot files? :) Nov 21 17:30:35 runtime doesn't go in the same .dot as the rest.. Nov 21 17:30:46 yeah Nov 21 17:30:51 And the task actually fails :) Nov 21 17:31:20 https://github.com/kergoth/homefiles/blob/master/bin/scm-prompt-output is what i used to use with bash Nov 21 17:31:41 less functional than the zsh setup though, that doesn't show if you're in the middle of a rebase/am/etc Nov 21 17:32:06 yeah, zsh has been on my switch-to list for feels like nearly a decade now, heh Nov 21 17:32:20 RDEPENDS_${PN} could work fine, it depends on the tasks being executed Nov 21 17:32:25 Tartarus: pass -O0 to samba Nov 21 17:32:37 and also try ARM_INSTRUCTION_SET = "arm" Nov 21 17:32:55 it was one of these two workarounds or maybe both cant remember Nov 21 17:33:09 khem, ok Nov 21 17:34:46 will do that after confirming the pixman __tls stuff Nov 21 17:43:55 khem, hmm, qemush4 DISTRO=minimal minimal-image builds, minimal-uclibc gcc doesn't and complains about some stuff being the wrong endian Nov 21 17:44:42 known or pastebin? Nov 21 17:45:03 (i'll turn on tinderbox once i'm pretty sure we're done w/ #!/too/long/a/path bugs) Nov 21 17:45:17 ans we just may be :) Nov 21 17:48:19 03Chris Larson  07org.openembedded.dev * rab1cf46448 10openembedded.git/classes/utils.bbclass: Nov 21 17:48:19 utils.bbclass: add dirs/pushd/popd Nov 21 17:48:19 These can be useful, and even necessary in some tasks, yet are bash, so let's Nov 21 17:48:19 ensure they're available so we can avoid the /bin/sh is bash requirement in Nov 21 17:48:19 the future. Nov 21 17:48:19 Signed-off-by: Chris Larson Nov 21 18:09:10 * Tartarus peeks at v86d Nov 21 18:09:53 Er, hmm Nov 21 18:10:01 NOTE: the following files were installed but not shipped in any package: Nov 21 18:10:07 NOTE: /var/hudson/workspace/sb_openembedded_autobuilder/BBPKGS/nas-server-image/DISTRO/minimal/MACHINE/qemux86/label/sb-u804/tmp/sysroots/i686-linux/usr/i686/bin/i686-oe-linux-klcc Nov 21 18:11:35 * Tartarus peeks at -e Nov 21 18:15:55 03Philip Balister  07org.openembedded.dev * r8d91e8a1b6 10openembedded.git/contrib/python/generate-manifest-2.6.py: Nov 21 18:15:55 generate-manifest-2.6.py : Added platform module to the generator script. Nov 21 18:15:55 This is needed so that http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=c59afd9fd456766c473c4fb7f7cc4db3e8623c39 Nov 21 18:15:55 does not overwrite the manifest file. Nov 21 18:17:19 did anyone manage to get a backup of the actual cvs repositories from hh.o before it died, or do we just not have history for things like ipkg-utils now? Nov 21 18:17:31 Ah, A-derp Nov 21 18:17:37 shouldn't have ${D} in the FILES line Nov 21 18:18:18 ant_: ping Nov 21 18:19:09 * Tartarus checks, fixes branch, gets back to fixing Nov 21 18:20:29 Tartarus: point to tinderbox ? Nov 21 18:21:35 ericben, thanks for pointing out my mistake Nov 21 18:21:43 khem: Haven't enabled until I was sure there weren't local issues in the build, which is finally this build. So next one I'll turn it on Nov 21 18:21:43 sec.. Nov 21 18:22:23 http://pastebin.com/GhAcWZ8N Nov 21 18:22:31 Tartarus: ok Nov 21 18:24:37 khem: OK to pull d4c4033f9af05651e106cd8e9f8577862df20427 into the release? Nov 21 18:24:37 03Tom Rini  07org.openembedded.dev * rd4c4033f9a 10openembedded.git/recipes/klibc/klcc-cross_1.5.20.bb: Nov 21 18:24:37 klcc-cross: Fix FILES_klcc-cross line. Nov 21 18:24:37 ${D} shouldn't be used here. Nov 21 18:24:37 Signed-off-by: Tom Rini Nov 21 18:24:42 ^^^ :) Nov 21 18:24:55 Tartarus: you are using 0.9.31 ? Nov 21 18:25:07 khem: I'm using whatever sane-toolchain sets as the default, yes. Nov 21 18:25:21 I believe that's 0.9.31 Nov 21 18:25:24 Tartarus: ok I think we need to use git version Nov 21 18:25:35 You update and I'll pull stuff in and re-kick these Nov 21 18:25:45 03Bob Foerster  07master * r73b40f0644 10bitbake.git/lib/bb/ (cooker.py runqueue.py taskdata.py utils.py): (log message trimmed) Nov 21 18:25:45 Prefer xrange over range for small performance gain. Nov 21 18:25:45 range() allocates an actual list when called. xrange() is just an iterator Nov 21 18:25:45 and creates the next range item on demand. This provides a slight Nov 21 18:25:45 performance increase. Nov 21 18:25:45 In python 3, range will do what xrange does currently, but the upgrade will Nov 21 18:25:46 be handled by the 2to3 tool. Nov 21 18:25:47 Tartarus: ok Nov 21 18:28:16 Tartarus: d4c4033f9af05651e106cd8e9f8577862df20427 looks ok for branch Nov 21 18:29:16 k, thanks Nov 21 18:29:47 ok, stupid git question Nov 21 18:29:55 How can I fast-forward other than a git pull? Nov 21 18:30:58 * Tartarus googles Nov 21 18:33:16 pull isn't magic Nov 21 18:33:20 pull is git fetch + git merge Nov 21 18:33:21 that's it Nov 21 18:33:29 (unless pull --rebase, in which case its fetch + git rebase) Nov 21 18:33:38 so just git merge Nov 21 18:36:36 03Khem Raj  07master * ra69c8a4da6 10openembedded.git/conf/distro/include/sane-toolchain.inc: Nov 21 18:36:36 sane-toolchain.inc: Use git as preferred version for uclibc Nov 21 18:36:36 * Also define PREFERRED_EGLIBC and PREFERRED_GLIBC Nov 21 18:36:36 Signed-off-by: Khem Raj Nov 21 18:37:54 Tartarus: pull and rebuild Nov 21 18:39:52 k, kicking things off Nov 21 19:01:24 damn, it would be nice to try to incorporate the actual loading of the cache file from disk into the progress bar somehow Nov 21 19:01:49 since when you're fully cached, it sits there seemingly stalled while it loads it, *then* you get a fast bar as it processes that Nov 21 19:02:02 ideally itd show the progress bar immediately, so you know its working Nov 21 19:02:23 that much i have working, but the ETA is all off due to the amount of time from start to first update Nov 21 19:02:32 hmm Nov 21 19:04:33 interesting Nov 21 19:11:26 hi Nov 21 19:11:47 I got this error pixman_0.21.2.bb do_compile failed Nov 21 19:11:52 hum Nov 21 19:11:54 http://pastebin.com/hi2SVSuF Nov 21 19:11:58 i compiled yesterday successfully Nov 21 19:12:00 i'll check for you Nov 21 19:12:09 I would be grateful if anybody could give me a hint about how could it be caused and how to solve it. Nov 21 19:12:18 distro=kaeilos-2010 Nov 21 19:12:20 What machine? Nov 21 19:12:27 obi, ping Nov 21 19:12:42 angelox_123: ronetix-pm9263 or akita Nov 21 19:12:55 pong Nov 21 19:13:15 has anything gone into dev that might fix the gtk problem? Nov 21 19:13:35 mckoan: Do you tried bitbake -c clean /home/koan/devel/openembedded/recipes/xorg-lib/pixman_0.21.2.bb ? Nov 21 19:13:41 and after Nov 21 19:13:50 mckoan: bitbake -c build /home/koan/devel/openembedded/recipes/xorg-lib/pixman_0.21.2.bb ? Nov 21 19:14:27 angelox_123: doing now Nov 21 19:14:38 Crofton: no, i'm still running bitbake xmms... currently at xextproto-native. it's a slow machine, so it won't finish until tomorrow i guess Nov 21 19:14:42 angelox_123: NoProvider: /home/koan/devel/openembedded/recipes/xorg-lib/pixman_0.21.2.bb Nov 21 19:15:11 what does your current diff look like? Nov 21 19:15:30 sorry Nov 21 19:15:33 i made a mistake Nov 21 19:15:36 Am I supposed to be building perl-native? Nov 21 19:15:37 * mckoan is puzzled Nov 21 19:15:43 bitbake -c clean -b /home/koan/devel/openembedded/recipes/xorg-lib/pixman_0.21.2.bb Nov 21 19:15:44 try now Nov 21 19:15:48 and after Nov 21 19:15:53 bitbake -c build -b /home/koan/devel/openembedded/recipes/xorg-lib/pixman_0.21.2.bb Nov 21 19:17:24 angelox_123: ok, same error Nov 21 19:17:37 Crofton: http://pastebin.com/5kcdk8d9 Nov 21 19:18:02 angelox_123: http://pastebin.com/ZH6b08uD Nov 21 19:18:48 mckoan: Can you try compile another version,only to test? Nov 21 19:18:57 let me try to dupe that by hand and try another build Nov 21 19:19:55 Crofton: you can get the raw patch from pastebin, too: http://pastebin.com/raw.php?i=5kcdk8d9 Nov 21 19:20:20 angelox_123: another pixman version? Nov 21 19:20:32 mckoan: yes Nov 21 19:21:13 Crofton: i'll be afk from now for some hours, but i'll read whatever you'll write later ;) Nov 21 19:23:13 ok Nov 21 19:23:19 have a good day Nov 21 19:23:25 thanks for working on this Nov 21 19:23:26 pixman_0.20.0.bb failed Nov 21 19:23:50 try Nov 21 19:24:18 pixman-0.18.4.bb Nov 21 19:24:45 fun, Intel specify i7's DDR3 controler's voltage at 1.5V typical 1.575 max and dealers sell 1.65V memories for these CPU ... that should explain why my build machine is crashing under load .... Nov 21 19:26:14 angelox_123: worse sh: ipkg-build: command not found Nov 21 19:26:24 strange Nov 21 19:26:40 should you ask to kaeilos mainteiners Nov 21 19:26:41 here Nov 21 19:27:15 angelox_123: that's me :-D Nov 21 19:27:18 :) Nov 21 19:27:24 hehehe Nov 21 19:27:25 :) Nov 21 19:27:36 ask for master of here,like khem,khergoth... Nov 21 19:27:39 I don't know pixman Nov 21 19:27:46 me too Nov 21 19:27:54 angelox_123: I'll try with ML or tomorrow Nov 21 19:27:56 but i think is a problem of deppendencies Nov 21 19:28:03 ok,should be better Nov 21 19:28:12 angelox_123: thanks Nov 21 19:28:25 np Nov 21 19:43:46 I'll push on org.openembedded.dev soon Nov 21 19:45:00 angelox_123, hi! did you have time to look into the r-project + fortran compiler problem? Nov 21 19:45:17 any opinions on my update-rc.d patch? or are we going to solve this in another way for the release Nov 21 19:45:31 (btw could also use a 2nd ack for the php url fix) Nov 21 19:45:48 re Nov 21 19:45:57 eFfeM: what is the issue with update.rc Nov 21 19:46:06 stale URI ? Nov 21 19:47:00 ericben: would you propose the patch to make OE search oe sourcemirror Nov 21 19:47:18 so we can get the tars for dysfunction hh.org Nov 21 19:47:42 whats issue with pixman? Nov 21 19:49:40 khem: once I've tested it :) Nov 21 19:51:00 is it ok to push? Nov 21 19:51:17 because I've read that the release will be rebased on top of master Nov 21 19:51:25 I want to push to master Nov 21 19:51:32 ericben, hi Nov 21 19:53:13 ericben: nice thx Nov 21 19:53:35 GNUtoo|laptop: not everything from master will be in release Nov 21 19:53:48 hi GNUtoo|laptop dinner time Nov 21 19:53:50 GNUtoo|laptop: but if a patch helps the release then it will be cherry picked Nov 21 19:54:05 I want only to push to org.openembedded.dev Nov 21 19:54:13 * khem off to shower Nov 21 19:54:25 can I do it or do I have to wait or something like that? Nov 21 19:54:27 GNUtoo|laptop: which patch is that Nov 21 19:54:44 compilations fixes, and a linux-n900 recipe Nov 21 19:54:46 you can push the stuff to master like usual Nov 21 19:54:53 ok thanks a lot Nov 21 19:55:00 that's what I wanted to know Nov 21 19:55:03 thanks Nov 21 19:55:12 make sure that if some of them are needed for release then ask for cherry picking them into release Nov 21 19:55:15 too Nov 21 19:55:28 ok Nov 21 19:55:41 I am looking at master every now and then but I may miss stuff and may not know if it fixes something for release Nov 21 19:56:36 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * ra3aa7c162c 10openembedded.git/recipes/xorg-driver/xf86-input-synaptics_1.3.0.bb: (log message trimmed) Nov 21 19:56:37 xf86-input-synaptics 1.3.0: fix CROSS COMPILE Badness Nov 21 19:56:37 It was failing to compile that way: Nov 21 19:56:37 make[2]: Entering directory `.../work/i686-angstrom-linux/xf86-input-synaptics-1_1.3.0-r14.0/xf86-input-synaptics-1.3.0/tools' Nov 21 19:56:37 CC synclient.o Nov 21 19:56:37 CC syndaemon-syndaemon.o Nov 21 19:56:38 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/xorg Nov 21 19:56:44 I don't know if my fixes are really critical Nov 21 19:56:47 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * ra54020e17c 10openembedded.git/recipes/wesnoth/wesnoth.inc: (log message trimmed) Nov 21 19:56:47 wesnoth.inc: fix missing lua5.1 dependency and bump PR Nov 21 19:56:47 ERROR: Logfile of failure stored in: .../work/i686-angstrom-linux/wesnoth-wvga-1.8.4-r0.1/temp/log.do_configure.28289 Nov 21 19:56:47 Log data follows: Nov 21 19:56:47 | CMake Error at .../sysroots/x86_64-linux/usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:70 (MESSAGE): Nov 21 19:56:48 | Could NOT find Lua51 (missing: LUA_LIBRARIES LUA_INCLUDE_DIR) Nov 21 19:56:48 | Call Stack (most recent call first): Nov 21 19:56:49 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r30a23e1429 10openembedded.git/recipes/libzrtpcpp/libzrtpcpp_1.4.6.bb: (log message trimmed) Nov 21 19:56:50 libzrtpcpp 1.4.6: fix compilation error and bump PR Nov 21 19:56:50 Without that fix we have: Nov 21 19:56:51 | arm-oe-linux-gnueabi-g++: .../sysroots/armv7a-oe-linux-gnueabi/lib/libgcrypt.so: No such file or directory Nov 21 19:56:51 | arm-oe-linux-gnueabi-g++: .../sysroots/armv7a-oe-linux-gnueabi/lib/libgpg-error.so: No such file or directory Nov 21 19:56:52 | make[3]: *** [libzrtpcpp.la] Error 1 Nov 21 19:56:52 | make[3]: Leaving directory `.../work/armv7a-oe-linux-gnueabi/libzrtpcpp-1.4.6-r0/libzrtpcpp-1.4.6/src' Nov 21 19:56:53 03Klaus Kurzmann  07org.openembedded.dev * rd9b47fc890 10openembedded.git/recipes/linux/ (linux-nokia900/defconfig linux-nokia900_git.bb): Nov 21 19:56:53 Add new recipe linux-nokia900_git.bb Nov 21 19:59:10 khem, was not watching the chat, update-rc.d fetched from hh.org, as there was only one file fetched, I extracted it from a tarball I still had, and inlined the file Nov 21 20:05:10 * Tartarus kicks the not reporting failed builds thing Nov 21 20:06:02 eFfeM: I see no bad to have update-rc.d in srctree Nov 21 20:07:21 but I'm not enough skilled to give Ack's :) Nov 21 20:11:41 jay7 if you want to you can test it (remove the update-rc.d tarball and checksum from your downloads dir first to be sure Nov 21 20:11:54 and then if it builds nicely mail a Tested-by line Nov 21 20:16:55 obi: hey Nov 21 20:17:18 obi: I can run test and get results in 2 hours Nov 21 20:17:37 obi: what way to test better? inherit or FILES_* Nov 21 20:33:29 hi, libgnashplugin.so is installed into usr/lib64/mozilla/plugins in the image folder Nov 21 20:33:53 libgnashplugin.so is the mozilla plugin for gnash(the free adobe flash replacement) Nov 21 20:34:00 I wonder how to workarround Nov 21 20:34:07 because lib64 is clearly wrong Nov 21 20:34:28 basically I run a 64bit OS on my build system and my arm target is only 32bit Nov 21 20:34:30 so... Nov 21 20:34:33 should I: Nov 21 20:34:41 *fix the autotools stuff Nov 21 20:34:43 or: Nov 21 20:34:54 *move the installed libs? Nov 21 20:35:12 gnash 0.8.7 doesn't compile anymore Nov 21 20:35:21 so I created a gnash 0.8.8 recipe Nov 21 20:35:24 GNUtoo|laptop: is there any configure args to specify lib destination dir? Nov 21 20:35:27 which compiles but has that issue Nov 21 20:35:31 I'll look good idea Nov 21 20:36:43 I'll look better but I think it can Nov 21 20:38:06 how does 64bit mips stuff work? Nov 21 20:38:12 they install to lib64? Nov 21 20:38:44 but then you should provide good path.. I don't know where to get arch instructions size Nov 21 20:38:55 i.e. 32 vs 64 bits Nov 21 20:39:03 GNUtoo|laptop: I don't have any Nov 21 20:39:20 any 64bit non x86 :) Nov 21 20:39:29 ok Nov 21 20:40:04 for instance loongson CPU are 64bit (the ones in the lemotes ) Nov 21 20:40:42 I have only ben-nanonote which is 32bit Nov 21 20:41:04 ok Nov 21 20:41:10 btw thanks for yesterday Nov 21 20:41:15 yw :) Nov 21 20:41:30 I'll explain: Nov 21 20:41:32 so have you control over your n900 now? :) Nov 21 20:41:36 yes Nov 21 20:41:39 I'll explain Nov 21 20:41:43 good :) Nov 21 20:41:45 basically my computer froze Nov 21 20:41:50 eFfeM: I think if ericben proposes the patch then this will not be needed. So wait for sometime Nov 21 20:41:50 that interupted a build Nov 21 20:42:04 which made deployed kernel size 0 Nov 21 20:42:15 wich I flashed Nov 21 20:42:19 *which Nov 21 20:42:21 hehe :) Nov 21 20:42:28 btw Nov 21 20:42:31 so the previous kernel remained on the device Nov 21 20:42:33 and so.... Nov 21 20:42:40 there was a mismatch between the dirs Nov 21 20:42:50 we have kernel size check routine in OE but it checks upper limit Nov 21 20:43:06 can we enable some QA check for size 0? Nov 21 20:43:28 (unsure that this is good idea anyway) Nov 21 20:43:32 it would have no effect Nov 21 20:43:36 as the build was interupted Nov 21 20:43:41 when it should not have been Nov 21 20:43:56 well.. then you should fix installer to not flash 0-size kernels :) Nov 21 20:43:59 and that a bitbake shr-image thought that it was ok Nov 21 20:44:05 yes Nov 21 20:44:08 GNUtoo|laptop: a target package installing into lib64 because build box is 64bit is a problem in anycase Nov 21 20:44:23 GNUtoo|laptop: find out the relevant parts of configure where this is determined Nov 21 20:44:24 khem, yes I know, that's why I asked how to resolve it Nov 21 20:44:27 ok Nov 21 20:44:30 I'll do it now Nov 21 20:44:41 because OE passes all libdir and other knobs that are common Nov 21 20:44:45 so it should not happen Nov 21 20:45:00 unless gnash configure is doing something whacky Nov 21 20:46:28 gnash isn't great for cross-compiling Nov 21 20:46:37 for instance it disables pkg-config calls Nov 21 20:46:38 etc... Nov 21 20:46:45 basically it's meant to be compiled by hand Nov 21 20:46:47 hmm Nov 21 20:46:52 so it doesn't work well with oe Nov 21 20:47:00 too bad Nov 21 20:47:05 beat it to behave Nov 21 20:47:16 that's what I'll do Nov 21 20:47:17 email the author Nov 21 20:47:19 with some -I Nov 21 20:47:22 he promised to fix it Nov 21 20:47:27 ah? Nov 21 20:47:30 several years back at FOSDEM Nov 21 20:47:34 ok Nov 21 20:47:39 I'll ask him Nov 21 20:47:42 he's on irc Nov 21 20:47:44 Crofton: btw. did that issue got resolved for you now Nov 21 20:47:49 with the patch Nov 21 20:48:09 exploding pkgconfig? Nov 21 20:48:11 testing now Nov 21 20:48:12 limitless spawning of threads with libtool Nov 21 20:48:13 yes Nov 21 20:48:15 ok Nov 21 20:48:16 according to #gnash it's --with-npapi-plugindir=DIR Nov 21 20:48:17 cool Nov 21 20:48:40 so what should I pass to it? Nov 21 20:48:54 ${libdir}/mozilla/plugins Nov 21 20:48:55 ? Nov 21 20:49:09 libdir can be lib64 in case of 64bit mips? Nov 21 20:49:25 GNUtoo|laptop: yes thats fine Nov 21 20:50:19 ok thanks a lot will try now Nov 21 20:53:03 I hope I will be able to push it this afternoon/night Nov 21 20:53:07 (gnash) Nov 21 21:00:21 wow Nov 21 21:00:37 * Jay7 have looked at tinderbox's waterfall Nov 21 21:01:04 mentor-graphics-autobuilder is powerfull :) Nov 21 21:01:56 Tartarus: looking amazing :) what is behind? Nov 21 21:03:21 hmmmm Nov 21 21:03:23 can't test: Nov 21 21:03:49 | .../work/x86_64-linux/xcb-proto-native-1.6-r3.0/temp/run.do_configure.30415: .../sysroots/x86_64-linux/usr/bin/automake: .../sysroots/x86_64-linux/usr/bin/perl: bad interpreter: No such file or directory Nov 21 21:04:01 I'll look Nov 21 21:06:00 ok race Nov 21 21:16:38 Hi all Nov 21 21:16:42 again :) hehe Nov 21 21:16:48 http://angelox_123.pastebin.com/m7msWHEw Nov 21 21:16:55 it cant install xfce-utils Nov 21 21:17:07 i think is a duplicate entry on code Nov 21 21:17:08 :( Nov 21 21:21:14 any help? Nov 21 21:25:40 another version works fine Nov 21 21:26:14 I'll push soon Nov 21 21:27:02 angelox_123, I already saw that Nov 21 21:27:08 fix the install scripts Nov 21 21:27:10 maybe Nov 21 21:27:22 on source isn't ? Nov 21 21:27:45 s/scripts/Makefile.am Nov 21 21:29:42 ok Nov 21 21:38:36 khem, i've pushed the fixes for uClibc++ now. Please verify (or ask Frans to double-check, not sure, is he eFfeM ?) Nov 21 21:46:05 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r36fa802068 10openembedded.git/recipes/gnash/gnash_0.8.8.bb: (log message trimmed) Nov 21 21:46:05 gnash : add 0.8.8 version Nov 21 21:46:05 The 0.8.8 version fixes youtube. Nov 21 21:46:05 Note that I had to add one more include: Nov 21 21:46:05 -I${STAGING_INCDIR}/cairo Nov 21 21:46:05 Because gnash didn't find cairo: Nov 21 21:46:05 | ../sysroots/armv7a-oe-linux-gnueabi/usr/include/gtk-2.0/gdk/gdkscreen.h:31:19: fatal error: cairo.h: No such file or directory Nov 21 21:46:15 blindvt, yes he is Nov 21 21:46:39 I'll go Nov 21 21:46:45 obi, xmms build fails Nov 21 21:46:49 http://pastebin.com/vy0JFCjj Nov 21 21:47:01 bye GNUtoo|laptop Nov 21 21:47:55 XorA: oh, literki fails to build ? I have to check - it did build when I submitted it :) Nov 21 21:48:35 but the chain of events looks like an outside bug... Nov 21 21:52:04 Crofton: which solution you test this time? Nov 21 21:54:26 hehe.. clean quit.. Nov 21 21:56:18 one that obi pastebined earlier Nov 21 21:56:42 gtk-config is installed in bindir and packaged in the -dev package Nov 21 21:57:21 [16:31] Jay7: please try whether "inherit lib_package" in gtk+-1.2*.bb fixes that. or else: FILES_${PN}-dev += "/usr/bin/gtk-config" Nov 21 21:57:32 one of this? Nov 21 22:02:15 the fiels is in there Nov 21 22:02:28 +FILES_${PN}-dev += "${datadir}/gtk-1.2/include ${libdir}/gtk-1.2/include ${bindir}/gtk-config" Nov 21 22:03:02 did you really mean "/usr/bin/gtk-config"? Nov 21 22:03:38 yes. does it get packaged into the ipk? Nov 21 22:03:51 gtk+-1.2*-dev.ipk Nov 21 22:04:38 trying the inherit one Nov 21 22:04:48 that shouldn't make a difference Nov 21 22:05:13 the FILES one confuses me Nov 21 22:07:43 why? Nov 21 22:08:05 does the get gtk-config from the machien you are buidling on? Nov 21 22:08:45 no, it's the paths in FILES are relative to the work dir Nov 21 22:09:00 s/it's// Nov 21 22:09:11 khem, thats for the pointers on icecc...I'll talk to likewise next time I see him. Nov 21 22:22:25 03Martin Jansa  07master * r46c8038ae2 10openembedded.git/recipes/linux/ (linux-2.6.34/shr.patch linux-openmoko_2.6.34.bb): Nov 21 22:22:25 linux(-openmoko)_2.6.34: add missing WSOD fix from 2.6.32 to shr.patch Nov 21 22:22:25 Signed-off-by: Martin Jansa Nov 21 22:23:12 Crofton: i've made a wrong assumption. inherit lib_package will probably fix it, but might change the packaging. inherit binconfig should be sufficient. lib_package inherits binconfig, too Nov 21 22:24:45 Crofton: the modification of FILES_${PN}-dev will also be done by binconfig, so there's no need to do it in the recipe. Nov 21 22:25:38 obi: so what to change in recipe? Nov 21 22:27:50 Jay7: http://pastebin.com/raw.php?i=w7V9S7vg Nov 21 22:28:28 and bump PR, please Nov 21 22:28:32 ok, so inherit binconfig is the only change? Nov 21 22:29:08 basically, yes Nov 21 22:29:40 I'll just clean the recipe, and not bump PR :) Nov 21 22:36:34 hm.. Nov 21 22:36:40 xmms built ok Nov 21 22:37:07 Crofton: commit it please :) Nov 21 22:37:38 obi: or you Nov 21 22:38:41 my login doesn't work Nov 21 22:38:55 but i'd just send it to the list Nov 21 22:43:02 can i add your signed-off? Nov 21 22:46:07 or ... your ack or the like Nov 21 22:46:58 assuming it also works for Crofton, of course Nov 21 22:49:34 I'll let my build run to completion Nov 21 23:11:20 is this still the prefered way to send patches? http://wiki.openembedded.org/index.php/How_to_submit_a_patch_to_OpenEmbedded Nov 21 23:15:35 probably Nov 21 23:24:50 cazze: "This page was last modified on 17 November 2010, at 00:25." -- did you really think the patch submission process had changed in the past 4 days? :P Nov 21 23:26:16 obi: I'll prefer Tested-by :) Nov 21 23:26:40 kergoth: héhé, didn't see the date. but i found it strange i didn't see the patches posted to the mailinglist comming up on the patchwork website Nov 21 23:27:20 sounds like an IT problem, something up with patchwork or the mailing list, not the process :) Nov 21 23:27:37 swfdec is failing http://pastebin.com/VGVmFzph Nov 21 23:27:45 kergoth: ok :-) Nov 21 23:27:55 there where some list issues Nov 21 23:28:06 and we may have a pw issue, need to check Nov 21 23:41:30 03Chris Larson  07master * r44289f3b9f 10openembedded.git/recipes/mktemp/mktemp_1.6.bb: Nov 21 23:41:30 mktemp: fix for bindir vs base_bindir Nov 21 23:41:30 Signed-off-by: Chris Larson Nov 21 23:45:22 how to get in a device internet? i can ssh into it Nov 21 23:46:45 angelox_123: it depends Nov 21 23:47:00 how? Nov 21 23:47:00 angelox_123: how are you connected to device? Nov 21 23:47:06 via usb Nov 21 23:47:15 usbnet? Nov 21 23:47:17 no Nov 21 23:47:20 i mean Nov 21 23:47:23 ops Nov 21 23:47:27 i don't know Nov 21 23:47:37 i followed the tutorial USB networking from openmok Nov 21 23:47:41 i can ssh into it Nov 21 23:47:44 usbnet then :) Nov 21 23:47:50 sorry,newbie :) Nov 21 23:48:31 what is your inetrnet connection? Nov 21 23:48:42 how is it done? Nov 21 23:48:51 i have a router Nov 21 23:49:05 GNUtoo helped me once,but he aren't online :( Nov 21 23:49:36 you have at least two ways - route ip address of device to router and nat it there Nov 21 23:49:42 or nat it on your host Nov 21 23:50:17 which way is more easy/fast ? Nov 21 23:50:18 :) Nov 21 23:50:30 if you have no experience with iptables - routing can be easier :) Nov 21 23:50:59 but I'm unsure about your router abilities Nov 21 23:51:37 do you have any tutorial ? Nov 21 23:51:52 I haven't Nov 21 23:52:03 I just former ISP sysadmin :) Nov 21 23:52:15 *I'm Nov 21 23:52:27 :) Nov 21 23:53:28 Crofton: i have a fix for that Nov 21 23:53:48 ok Nov 21 23:53:59 I commented them out and let the build run Nov 21 23:54:20 angelox_123, I use this in PC to get internet access in openmoko: Nov 21 23:54:21 sudo ifconfig eth1 192.168.2.200 Nov 21 23:54:21 sudo iptables -A POSTROUTING -t nat -j MASQUERADE -s 192.168.2.192/26 Nov 21 23:54:21 sudo sysctl -w net.ipv4.ip_forward=1 Nov 21 23:55:52 pespin|netbook: what is your resolv.conf on openmoko? Nov 21 23:56:11 the problem is,i have a network manager on Fedora :( that makes mistakes Nov 21 23:57:05 i'll try remove my nm Nov 21 23:58:00 angelox_123, nameserver 208.67.222.222 Nov 21 23:58:00 nameserver 208.67.220.220 Nov 21 23:58:05 opendsn ones I think Nov 21 23:58:38 *opends Nov 21 23:58:51 *opendns Nov 21 23:59:02 omg, can't write, time to sleep hehe Nov 21 23:59:27 angelox_123, you told me you'd try to look at fortran not compiling in OE. Did you have time to look at it? Nov 22 00:03:52 Crofton: i'm about to send the patches to the list. you can add your ack or nack there, if you like, or just commit them Nov 22 00:04:47 ok Nov 22 00:05:05 is patchwaork really not working? Nov 22 00:06:07 on the site the last date of a patch is 19 nov Nov 22 00:06:17 we are 22 nov now Nov 22 00:06:22 seems strange Nov 22 00:17:27 anybody, ping (test connection) Nov 22 00:19:14 use /ctcp angelox_123 ping Nov 22 00:19:26 ok Nov 22 00:19:27 thank you Nov 22 00:19:28 :) Nov 22 00:19:30 worked Nov 22 00:19:32 :) Nov 22 00:19:35 now Nov 22 00:19:46 i need /etc/opkg/**** files of any angstrom distribution Nov 22 00:19:55 * Jay7 is trying to setup zoneminder on freebsd.. Nov 22 01:06:56 hi all Nov 22 01:07:10 alx_: Hi Nov 22 01:08:19 Bye all Nov 22 01:08:22 good night Nov 22 01:08:49 i'm stuck with my wish to customize psplash (as many, I noted on forum) unluckily I didn't understood how to do "my way": I wanted to do a bb recipe that provides psplash inspired to what I've read, but I didn't understood override system (i wrote a custom image based on angstrom) Nov 22 01:14:36 So what's your recipe look like? Nov 22 01:15:08 And where do your artwork files live, exactly Nov 22 01:50:21 does anyone why I have very small fonts on qt embedded ? Nov 22 02:03:27 03Khem Raj  07master * r2138038303 10openembedded.git/recipes/libiconv/ (libiconv-1.13.1/autoconf.patch libiconv_1.13.1.bb): Nov 22 02:03:27 libiconv_1.13.1.bb: Make it libtool version independent Nov 22 02:03:27 * The patch hardcodes macros from libtools 2.2.6 which Nov 22 02:03:27 cause problem with 2.4 so here we let autotools class copy the macros Nov 22 02:03:27 from staging. Nov 22 02:03:28 Signed-off-by: Khem Raj **** ENDING LOGGING AT Mon Nov 22 02:59:58 2010