**** BEGIN LOGGING AT Tue Aug 03 02:59:57 2010 Aug 03 06:04:18 03sledz  07org.openembedded.dev * r4946a4b375 10openembedded.git/recipes/rsyslog/ (files/rsyslog-resume.patch rsyslog_5.5.4.bb): Aug 03 06:04:18 rsyslog: break potential infinite loop in actionDoRetry Aug 03 06:04:18 If a module always returns RS_RET_OK (like ompipe does) the actionDoRetry Aug 03 06:04:18 loop may not have leaved faked ACT_STATE_SUSP state in case iResumeOKinRow Aug 03 06:04:18 had ever reached a count of 1000. Aug 03 06:04:18 Signed-off-by: Steffen Sledz Aug 03 07:04:11 03Steffen Sledz  07org.openembedded.dev * rb4a82e59cb 10openembedded.git/recipes/rsyslog/rsyslog_5.5.6.bb: Aug 03 07:04:11 rsyslog: new version 5.5.6 Aug 03 07:04:11 Signed-off-by: Steffen Sledz Aug 03 07:06:20 alo Aug 03 07:12:12 hi hrw, all Aug 03 07:25:16 03Martin Jansa  07org.openembedded.dev * r2c0037cd2a 10openembedded.git/ (20 files in 5 dirs): Aug 03 07:25:16 xorg: bump latest versions from 2010-08-02 Aug 03 07:25:16 Signed-off-by: Martin Jansa Aug 03 07:25:19 03Martin Jansa  07org.openembedded.dev * r9e5b50834f 10openembedded.git/recipes/eglibc/eglibc_2.12.bb: Aug 03 07:25:19 eglibc-2.12: bump SRCREV for alignment traps fix Aug 03 07:25:19 * but still lots of alignment traps on armv4t :/ Aug 03 07:25:19 Signed-off-by: Martin Jansa Aug 03 07:25:20 03Sebastian Krzyszkowiak  07org.openembedded.dev * rab327b9ce1 10openembedded.git/recipes/shr/shr-theme-niebiee.bb: Aug 03 07:25:20 shr-theme-niebiee: add phoneui-shr-theme-niebiee to RDEPENDS Aug 03 07:25:20 Signed-off-by: Martin Jansa Aug 03 07:25:21 03Martin Jansa  07org.openembedded.dev * r5fe65f7cd7 10openembedded.git/recipes/xorg-xserver/xorg-xserver-common.inc: Aug 03 07:25:21 xorg-xserver: new style staging Aug 03 07:25:21 Signed-off-by: Martin Jansa Aug 03 07:25:22 03Martin Jansa  07org.openembedded.dev * r0effb58081 10openembedded.git/recipes/gcc/ (gcc-4.5.inc gcc-4.5/gcc-pr43698-arm-rev-instr.patch): Aug 03 07:25:22 gcc-4.5: bump SRCREV after 4.5.1 release Aug 03 07:25:23 Signed-off-by: Martin Jansa Aug 03 07:25:27 03Martin Jansa  07org.openembedded.dev * r0a1635529a 10openembedded.git/recipes/kexecboot/ (5 files in 4 dirs): Aug 03 07:25:27 linux-kexecboot: add 2.6.35 Aug 03 07:25:27 Signed-off-by: Martin Jansa Aug 03 07:25:27 03Martin Jansa  07org.openembedded.dev * ra3ce3df6c3 10openembedded.git/recipes/tasks/task-shr-feed.bb: Aug 03 07:25:39 03Martin Jansa  07org.openembedded.dev * rc0f6de52cb 10openembedded.git/recipes/linux/ (3 files in 2 dirs): Aug 03 07:25:39 linux-openmoko-2.6.34: add patch for proper gsm sysfs node path Aug 03 07:25:39 Signed-off-by: Martin Jansa Aug 03 07:25:44 03Martin Jansa  07org.openembedded.dev * r445a10176f 10openembedded.git/recipes/maxima/maxima_5.21.1.bb: Aug 03 07:25:44 maxima: oe-stylize Aug 03 07:25:44 Signed-off-by: Martin Jansa Aug 03 07:43:26 good morning Aug 03 07:45:06 03Chase Maupin  07org.openembedded.dev * r5e75fa8c1b 10openembedded.git/classes/distribute_license.bbclass: (log message trimmed) Aug 03 07:45:06 distribute_license: added class to copy license files Aug 03 07:45:06 * This class can be used to copy license files from package Aug 03 07:45:06 sources into the deploy directory so they can be distributed. Aug 03 07:45:06 * By default it searches the top-level source directory for Aug 03 07:45:06 files matching COPYING* or LICENSE* Aug 03 07:45:07 * Search depth and pattern can be modified using the Aug 03 07:45:13 03Koen Kooi  07org.openembedded.dev * r49ff3926ef 10openembedded.git/recipes/fftw/fftw.inc: fftw: remove legacy staging Aug 03 08:04:45 I see, JaMa is here :) Aug 03 08:05:43 :) Aug 03 08:09:29 g'morning Aug 03 08:28:02 anyone got a good tip on a guide for network booting using Uboot? Aug 03 08:31:15 quick question, if anyone knows the answer. Why is there no tincantools-zippy.diff for the 2.6.32 kernel? Aug 03 08:32:26 Proxyles: http://www.lmgtfy.com/?q=u-boot+tftp Aug 03 08:33:29 eFfeM_work, well, question was if you knew of a good one ;P Aug 03 08:36:53 Proxyles: all you need is just tftpd and then uboot/tftp command Aug 03 08:37:00 Proxyles: nearly nobrain operation Aug 03 08:37:56 hrw, ok thx Aug 03 08:38:44 huhu Aug 03 08:41:05 Proxyles: and quite well explained in the first link if you peform the query I gave above Aug 03 08:41:55 eFfeM_work, reading it now, thank you all for the kind help Aug 03 08:44:09 Proxyles: there was one gotcha I run into: if you need to fetch the kernel from other ip than what the default gateway is, you can't use dhcp, bootp Aug 03 08:44:19 as dhcp overrides serverip Aug 03 08:44:28 Proxyles: yw Aug 03 08:44:33 (or something like that, just looked it up in my notes) Aug 03 08:45:11 zub, a thanks for the heads up Aug 03 08:45:37 n/p Aug 03 08:46:58 cd .. Aug 03 08:47:02 zub, good point, I always use tftp Aug 03 08:47:11 whoops Aug 03 10:12:05 does RREPLACES work at all? I am trying to get rid of a paticular package on my rootfs when another new package, providing the same functionality, is installed Aug 03 10:12:13 however it does not seem to work out Aug 03 10:12:23 what exactly is happening? Aug 03 10:12:28 I tried RREPLACES = "packagename" and RREPLACES_${PN} = "packagename" Aug 03 10:12:31 well nothing Aug 03 10:12:39 the new package is installed, the old remains Aug 03 10:13:46 it happens, that the binaries do not clash because one puts its stuff in sbin and the other one in bin Aug 03 10:13:58 that's the expected behaviour. Aug 03 10:14:06 but the old package does not get removed Aug 03 10:14:08 I think RREPLACES just does not do quite what you think it does. Aug 03 10:14:24 if you want the old package to be removed, you need to mention it in RCONFLICTS and RPROVIDES as well as RREPLACES. Aug 03 10:14:44 RREPLACES on its own just means that the new package is allowed to overwrite files from the old one, which is not going to do anything for you if the files do not actually overlap. Aug 03 10:14:52 I went by this: Used to specify that the current package replaces an older package with a different name. During package installation the package that is being replaced will be removed since it is no longer needed when this package is installed. Aug 03 10:14:56 which is from the OE manual Aug 03 10:15:01 ah, that is incorrect Aug 03 10:15:24 pb_ hey Aug 03 10:15:28 that explains it then :) Aug 03 10:15:37 may I join the lesson? Aug 03 10:15:58 iirc there is another undocumented thing Aug 03 10:16:04 see Aug 03 10:16:07 -PACKAGE_ARCH_${PN} = "${MACHINE_ARCH}" Aug 03 10:16:09 +PACKAGE_ARCH_pn-${PN} = "${MACHINE_ARCH}" Aug 03 10:16:28 ant_work: that change is just bogus Aug 03 10:16:40 (assuming it's in a recipe, anyway) Aug 03 10:16:48 ah, which are the right cases to use _pn Aug 03 10:17:03 (was a proposed patch btw) Aug 03 10:17:27 Jin^eLD: see http://www.debian.org/doc/debian-policy/ch-relationships.html, specifically section 7.6 Aug 03 10:17:30 ah, no, is committed Aug 03 10:17:42 pb_: btw do I need RCONFLICHTS = "bla" or RCONFLICTS_${PN}="blah"? Aug 03 10:17:47 pb_: thx, reading up Aug 03 10:17:57 Jin^eLD: the latter, in general Aug 03 10:18:16 ant_work: _pn is only appropriate in global configurations (e.g. distro config) Aug 03 10:18:30 ant_work: there is no point using _pn in an individual recipe since it will (in general) always match. Aug 03 10:18:59 there are a few corner cases where it is appropriate in an .inc file, or a .bb file which is included from others, but I don't think that was the case being proposed here. Aug 03 10:19:07 pb_: [oe-commits] Khem Raj : recipes: Replace ${PN} override with pn-${PN} for PACKAGE_ARCH Aug 03 10:20:01 pb_: _pn override can be used in local.conf, isn't? Aug 03 10:20:08 ant_work: see http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-July/022114.html and the replies to that post Aug 03 10:20:25 it is a bit of a shame that khem seems to just have committed his patch anyway without explaining why he thinks it is a good idea. Aug 03 10:20:33 ant_work: yes, local.conf is fine Aug 03 10:21:08 pb_: well, manual needs updating... Aug 03 10:21:27 * Jay7 is trying to remember who will donate some zauruses in US.. Aug 03 10:21:42 I've found another local tester here :) Aug 03 10:22:06 so I'll take out my tosa Aug 03 10:22:17 pb_ once I feel I've retained it, I'll update the manual Aug 03 10:22:20 thx Aug 03 10:22:22 ant_work: good idea Aug 03 10:23:21 this overrides stuff is not an easy cake... Aug 03 10:28:53 pb_: thanks, that worked now Aug 03 10:29:00 very good Aug 03 10:36:37 03Martin Jansa  07org.openembedded.dev * r0d3da75671 10openembedded.git/recipes/libgee/libgee_0.5.2.bb: Aug 03 10:36:37 libgee-0.5.2: fix SRC_URI for -native Aug 03 10:36:37 Signed-off-by: Martin Jansa Aug 03 10:36:48 03Martin Jansa  07org.openembedded.dev * rb196b4ea4b 10openembedded.git/ (4 files in 4 dirs): Aug 03 10:36:48 linux-openmoko-2.6.32: upgrade stable patch to 2.6.32.17, move more stuff to modules, update om-gta02 to load them Aug 03 10:36:48 Signed-off-by: Martin Jansa Aug 03 11:49:39 03Dirk Opfer  07org.openembedded.dev * r2b61601cc8 10openembedded.git/recipes/linknx/ (4 files in 2 dirs): Aug 03 11:49:40 linknx: 0.0.1.26 -> 0.0.1.27 Aug 03 11:49:40 bump version number of cvs recipe Aug 03 11:49:50 03Dirk Opfer  07org.openembedded.dev * rcc58e6fe45 10openembedded.git/recipes/owfs/owfs_2.7p28.bb: Aug 03 11:49:50 owfs: 2.7p28 -> 2.8p0 Aug 03 11:49:50 remove legacy style staging Aug 03 11:49:55 03Dirk Opfer  07org.openembedded.dev * r305940c1b6 10openembedded.git/recipes/pthsem/ (pthsem_2.0.7.bb pthsem_2.0.8.bb): Aug 03 11:49:55 pthsem: 2.0.7 -> 2.0.8 Aug 03 11:49:55 Get rid of legacy staging Aug 03 11:49:56 03Dirk Opfer  07org.openembedded.dev * r2415f71c42 10openembedded.git/recipes/argp-standalone/argp-standalone_1.3.bb: argp-standalone: Remove staging null ops Aug 03 11:58:33 hm, lots of svn recipes seem to fail for me during do_fetch. E.g. navit_svn.bb Aug 03 11:58:44 anyone an idea on this? does it build for you? Aug 03 11:58:51 see http://www.pastie.org/1072628 for my log Aug 03 12:03:14 eFfeM_work: try rm -f /home/frans/workspace/tst/tmp/downloads/navit_navit.svn.sourceforge.net_.svnroot.navit.trunk_3358_.tar.gz*, works for me Aug 03 12:03:46 JaMa: will try, no idea why i got this Aug 03 12:06:32 JaMa: no success same error as before, could this be due to the https in the proto field? btw no idea why someone would want to use https for this Aug 03 12:07:58 eFfeM_work: yeah.. try "svn co" by hand and if it asks for certificate, reply with 'p' for permanent Aug 03 12:09:29 JaMa: i'm not really good at svn, but apart from that it should work without user intervention. I have at least 62 recipes that seem to have this issue Aug 03 12:10:00 (actually bitbake -k world gave 169 recipes that have a failing do_fetch) Aug 03 12:12:20 eFfeM_work: you can try with "svn --trust-server-cert --non-interactive" Aug 03 12:12:54 in bitbake.conf FETCHCMD_svn Aug 03 12:14:21 JaMa, re: a646269c2ada7691d8a7f7455ba4528c7cca3483 this is obviously a kernel bug. It should of course test 'write && !capable' Aug 03 12:14:51 JaMa, care to fix this there? TIA Aug 03 12:15:18 ehm, that was not really the point I was trying to make. The observation is that we have lots of recipes that do not build out of the box (actually I expected it to fetch things from a mirror, but apparently it isn't). Aug 03 12:15:43 03Frans Meulenbroeks  07org.openembedded.dev * rf2893b80cf 10openembedded.git/recipes/ (502 files in 206 dirs): (log message trimmed) Aug 03 12:15:43 500+ recipes: remove do_stage Aug 03 12:15:43 This patch removes all occurrences of Aug 03 12:15:43 do_stage() { Aug 03 12:15:43 autotools_stage_all Aug 03 12:15:44 } Aug 03 12:15:44 including all kind of variants w.r.t whitespace Aug 03 12:15:53 there are also recipes that point to files that are gone and that are not mirrorred (try cdtool 2.1.8) Aug 03 12:16:14 whoever committed my patch should have added the ack's Aug 03 12:21:16 ah the ack's are in it, but they do not show up in the cia log Aug 03 12:21:48 03Koen Kooi  07org.openembedded.dev * r99f6333dd2 10openembedded.git/recipes/libidl/ (libidl-native_0.8.3.bb libidl_0.8.2.bb libidl_0.8.3.bb): libidl: remove really old versions Aug 03 12:23:53 blindvt`: the problem is that you can build on some host with older kernel where it's not fixed yet.. Aug 03 12:24:56 eFfeM_work: there is many recipes failing to fetch.. se my report when I was downloading really everything (all version and all sources not only for on MACHINE if there was override).. Aug 03 12:27:30 JaMa, nm. i just see that this was fixed already, sorry Aug 03 12:27:57 blindvt`: and there is also workarround in OE qemu Aug 03 12:28:54 JaMa: only vaguely recall the report, but if the sources are gone the recipe is useless; either archeology is needed to recover them, or the recipes should be killed Aug 03 12:29:17 for cdtool that one is still in debian Aug 03 12:32:54 eFfeM_work: I've fixed some of them if I was able to find credible mirror.. for many others I was able to download archives manually from OE distros source stashes (so I was able to build them after manual intervention) Aug 03 12:33:19 eFfeM_work: but redirecting SRC_URI ie to slugos source stash seems wrong so I kept recipe as it was Aug 03 12:33:47 we could redirect the URI to the oe mirror dir Aug 03 12:33:53 and stash it there Aug 03 12:34:06 less Aug 03 12:35:08 btw is there a better way than a patch to change the default env for u-boot; ? Aug 03 12:35:10 03Enrico Scholz  07org.openembedded.dev * rc8ed7ab954 10openembedded.git/recipes/neon/neon_0.26.0.bb: Aug 03 12:35:10 neon: updated 0.26.0 to 0.29.3 Aug 03 12:35:10 Recipe was moved because old (and new) recipe have a negative default Aug 03 12:35:10 preference and only 0.25.5 is referenced by distributions. Aug 03 12:35:10 Signed-off-by: Enrico Scholz Aug 03 12:36:08 Hello, I'm getting | install: cannot stat `/stuff/tmp/work/beagleboard-angstrom-linux-gnueabi/angstrom-version-1_2010.7-test-20100803-r4/lsb_release': No such file or directory Aug 03 12:36:10 any clues? Aug 03 12:36:33 yes, the file or directory does not exist Aug 03 12:36:57 check path first to see what is wrong and whether it is file or dir before asking Aug 03 12:37:18 and it also helps to provide some context (e.g. what you did to get that) Aug 03 12:37:47 marcompile: http://catb.org/esr/faqs/smart-questions.html Aug 03 12:41:46 its on angstrom-version.bb, do_install Aug 03 12:42:44 as you can see, I am building angstrom for beagleboard. I git pulled the oe repo yesterday, cleared my tmp folder and started everything again Aug 03 12:43:51 maybe ask in #angstrom; suggest you read the link I gave you first :-) Aug 03 12:45:19 maybe the recipe is broken Aug 03 12:45:22 i have this file at /stuff/openembedded/recipes/angstrom/angstrom-version Aug 03 12:48:46 eFfeM_work, thanks anyway. i'll get this going Aug 03 12:49:58 marcompile: i don't think I've been too much of a help, the only point I wanted to make is taht you really haven't given enough info for anyone to diagnose the problem, explain what you did etc Aug 03 12:50:16 and I doubt the recipe is broken, people build angstrom for beagleboard all the time Aug 03 12:50:41 eFfeM_work, yeah I noticed.. I just tought it would be a common issue freshly talked about here Aug 03 12:50:55 doesn Aug 03 12:51:06 yes I noticed Aug 03 12:51:11 doesn't ring a bell for me but gl chasing it down Aug 03 12:57:19 eFfeM_work: just run into the svn issue; have you made any progress with it? Aug 03 12:57:47 zub, nope, didn't really try this was a spinoff of my bitbake world experiment Aug 03 12:57:49 I wonder what's the cause... I updated system (possibly svn), OE, bitbake. So there are too many degrees of freedom. Aug 03 12:58:04 zub have you installed svn on the system? Aug 03 12:58:10 zub, what recipe? Aug 03 12:58:31 I mean svn form my distro; OE builds its own? Aug 03 12:58:58 I got this going. It was a problem with PARALLEL_MAKE, (it was "-j2"). I disabled and it is running fine now Aug 03 12:59:09 Maybe I should report this somewhere? Aug 03 12:59:49 marcompile: cool, yes please report at the angstrom mailing list (you can also disable parallel make in a recipe, guess the angstrom people like it if you provide a patch) Aug 03 13:00:06 zub, no it hink it uses svn from the host Aug 03 13:00:22 thought you had a new system so wanted to make sure you had svn on it Aug 03 13:00:49 nope, just updates and I have svn Aug 03 13:01:06 zub, what recipe ? Aug 03 13:01:12 first recipe that fails for me was matchbox-panel-2 Aug 03 13:01:23 trying svn by hand, e.g. svn ls svn://svn.o-hand.com/repos/matchbox/trunk/matchbox-panel-2 Aug 03 13:01:26 times out Aug 03 13:01:35 so maybe o-hand svn issue? Aug 03 13:01:46 but http://svn.o-hand.com/view/matchbox/trunk/matchbox-panel-2/ seems ok Aug 03 13:02:40 hm, this recipe was also on the list of files I had problems with Aug 03 13:03:40 would that ^ svn ls ... work for you? Aug 03 13:04:50 zub that is awfully silent Aug 03 13:05:05 svn: Can't connect to host 'svn.o-hand.com': Connection timed out Aug 03 13:05:10 eveok, same here Aug 03 13:05:15 s/eve//; Aug 03 13:05:51 can't ping them either, but your http url works fine for me Aug 03 13:05:59 * eFfeM_work is wondering if this is a firewall issue Aug 03 13:06:14 I tried from two different boxes and both time out Aug 03 13:07:23 substituting svn: with http: makes it pass Aug 03 13:07:49 o-hand went paranoid? :) Aug 03 13:08:14 hm, here too, i'm not an svn wiz. anyone an idea? Aug 03 13:08:39 indeed they don't ping... wtf? Aug 03 13:08:58 svn.o-hand.com is there Aug 03 13:09:05 * RP is the admin for the server... Aug 03 13:09:17 wow, didn't expect that :) Aug 03 13:09:30 RP svn ls svn://.... does not work but svn ls http:// does Aug 03 13:09:41 any idea why, the recipe uses svn: Aug 03 13:09:47 matchbox-panel-2 Aug 03 13:10:32 eFfeM_work: It should be using http I expect Aug 03 13:10:52 RP, so the recipe is broken (guess it used to work) Aug 03 13:11:05 eFfeM_work: The server hasn't changed in a long time Aug 03 13:11:33 then something on oe (or in bitbake) changed. Aug 03 13:11:46 RP does svn ls svn://... work for you ? Aug 03 13:12:48 eFfeM_work: no Aug 03 13:13:15 eFfeM_work: The poky recipe says proto=http Aug 03 13:13:44 oh I see: Aug 03 13:13:45 SRC_URI = "svn://svn.o-hand.com/repos/matchbox/trunk;module=${PN};proto=http \ Aug 03 13:13:54 svn:/, but proto=http Aug 03 13:14:06 maybe handling of this changed in bitbake? Aug 03 13:15:27 but svn log neither on the recipe in question, not on lib/bb/fetch/svn.py indicates any recent change Aug 03 13:15:31 I mean git log Aug 03 13:21:06 RP the oe recipe also says proto=http, but for me it does not work Aug 03 13:21:14 i'm on bitbake 1.10 Aug 03 13:21:19 zub what bb version are you using? Aug 03 13:21:45 1.10 (tried master too) Aug 03 13:23:12 zub bitbake -b matchbox-panel-2_svn.bb works for me Aug 03 13:23:13 ???? Aug 03 13:23:54 zub: Are you behind a firewall and need a proxy? Aug 03 13:24:32 I'm behind a NAT, and some sort of FW too, but svn has always worked for me Aug 03 13:25:13 eFfeM_work ???? ? Aug 03 13:25:34 zub, I tried with -b and it works Aug 03 13:25:38 hm Aug 03 13:25:58 $ bitbake -b /mnt/data/SHR/shr-unstable/openembedded/recipes/matchbox2/matchbox-panel-2_svn.bb -c fetch Aug 03 13:26:10 for me fails Aug 03 13:26:32 trying with -D, the fetcher spis what it does, and it seems weird Aug 03 13:26:48 maybe it's semi-downloaded somehwere and fetcher just tried to update Aug 03 13:27:08 althought I rm-ed what I thought relevant in download dir Aug 03 13:27:40 but I see: svn: Failed to add directory 'applets': an unversioned directory of the same name already exists Aug 03 13:28:27 no idea what was wrong now it works for me Aug 03 13:29:01 ok, resolved after rm -rf .../downloads/svn/svn.o-hand.com Aug 03 13:29:17 the checked out tree got corrupted somehow and this kept the fetcher failing Aug 03 13:29:27 sorry for the noise Aug 03 13:29:51 strange thing is that I had issues with the same file Aug 03 13:32:56 for you it resolved itself? Aug 03 13:33:59 it seems Aug 03 13:34:13 testing the other svn's i had an issue with Aug 03 13:36:46 zub: that because 2 different recipes are requesting newer/older SRCREV of in same checkout dir Aug 03 13:37:30 zub: first svn upgrade to newer creates applets dir, then downgrade marks applets ad unversioned and another upgrade fails because applets dir already exists Aug 03 13:37:50 ok, now it makes sense Aug 03 13:39:59 ah ok Aug 03 13:42:43 zub: you can try matchbox2/matchbox-panel-2-icon-themes_0.0.1.bb SRCREV = "2098" instead of SRCREV = "1907" if it still works Aug 03 13:43:36 or just merge those 2 recipes (matchbox2/matchbox-panel-2_svn.bb) and create icon-themes as subpackages Aug 03 13:44:51 ok, I'll c Aug 03 13:46:41 03Koen Kooi  07org.openembedded.dev * r58527a1f82 10openembedded.git/recipes/angstrom/angstrom-uboot-scripts.bb: angstrom-uboot-scripts: bump PR Aug 03 13:46:51 03jkridner@beagleboard.org  07org.openembedded.dev * r6f9c2666d2 10openembedded.git/recipes/angstrom/angstrom-uboot-scripts/ (2 files): Aug 03 13:46:51 angstrom-uboot-scripts: added default boot scripts Aug 03 13:46:51 Added default boot scripts for the BeagleBoard validation image. Aug 03 13:46:51 Signed-off-by: Koen Kooi Aug 03 13:47:13 thanks koen Aug 03 13:50:05 ensc|w, hello Aug 03 13:50:44 ensc|w, I used your ncurses 5.7 recipe, and I'm trying to build asterisk now, and it is failing with error: Aug 03 13:50:54 | configure: *** The TERMCAP installation appears to be missing or broken. Aug 03 13:51:30 maybe its a bug with asterisk or anything else, but maybe its good to just let you know that this happens Aug 03 13:51:36 if you want I can send you the full log Aug 03 14:04:24 JaMa: just bumping the SRCREV for matchbox-panel-2-icon-themes doesn't work - the dirs have changes and do_fatch fails :-/ Aug 03 14:04:35 can't investigate more ATM Aug 03 14:08:18 morning Aug 03 14:08:47 g'day kergoth Aug 03 14:10:29 hi kergoth Aug 03 14:10:56 7am meetings are just wrong. Aug 03 14:11:08 * kergoth yawns Aug 03 14:11:10 JaMa: Making them the same recipe does not make sense Aug 03 14:11:14 marcompile: the asterisk recipe seems to be very old and do referred SRCURI does not exist anymore Aug 03 14:11:29 using the same svn revison might be bettre, or fix the fetcher so it actually works :/ Aug 03 14:11:31 kergoth: yeah, that sucks. I find it hard enough to get up by 8am in time for the builders. Aug 03 14:11:47 pb_: Your builders start late ;-) Aug 03 14:11:54 pb_: I've been having this problem too... Aug 03 14:12:06 RP: heh. well, any earlier would be unacceptable :-} Aug 03 14:12:15 * RP now has two rooms nearly finished Aug 03 14:12:21 ah, two more than me Aug 03 14:12:31 * pb_ has a nice set of trenches though Aug 03 14:12:45 pb_: I did have a 12 month head start so it doesn't say much ;-) Aug 03 14:13:00 ah yes :-} Aug 03 14:13:23 * RP does have planning permission for the work on the garage Aug 03 14:13:30 and the neighbour no longer speaks to me Aug 03 14:13:30 we rather hope to have a habitable building by Christmas. fingers crossed. Aug 03 14:13:39 oh dear. what did you do to him? Aug 03 14:13:55 pb_: Applied to extend the garage despite his objection Aug 03 14:15:00 ah, right. why doesn't he like it? Aug 03 14:15:24 pb_: It will add shade to some of his garden at certain times of the day Aug 03 14:15:46 marcompile: termcap itself is dead and to be replaced by tinfo Aug 03 14:16:24 doh. oh well, I guess he will get over it. Aug 03 14:16:38 pb_: Fundamentally if I want to change anything its automatically disliked Aug 03 14:17:07 that's a nuisance. it's always a pain when that happens. Aug 03 14:17:22 luckily our neighbours are quite relaxed about that sort of thing. Aug 03 14:17:35 pb_: The ones on the other side are fine Aug 03 14:17:50 I think we might have some slightly tricky negotiations coming up when my wife wants to build her studio, but they have been fine about the house itself. Aug 03 14:18:20 I haven't been agressive and threatening towards them, invaded their privacy or done any other manner of things I'm alleged to have done by the other one. Aug 03 14:18:56 good grief Aug 03 14:19:11 hi Aug 03 14:19:15 yo zecke Aug 03 14:19:16 rofl Aug 03 14:19:28 hi zecke Aug 03 14:19:33 that was a weird conversation to come into the middle off :) Aug 03 14:19:51 :) Aug 03 14:20:12 my first thought is you were talking about bitbake or soemthing Aug 03 14:20:17 not your houses Aug 03 14:20:31 bitbake has been agressive and threatening? :) Aug 03 14:20:46 heh Aug 03 14:20:55 the user base, er neighbbors Aug 03 14:20:56 It's certainly invading the privacy of all those recipies with its constant scanning. Aug 03 14:21:26 marcompile: asterisk-1.6.0.28 seems to be able to use -ltinfo so it might be worth to update to this version (at least) Aug 03 14:22:05 listening to a guy talk about security who is really excited Aug 03 14:24:27 RP: poor boy, but I am happy you have a house. Aug 03 14:25:51 i'm building asterisk-1.6.2.10 Aug 03 14:26:35 i get the termcap broken with the original 1.4.23.1 recipe and also my new 1.6.2.10 Aug 03 14:27:05 I just disabled (with ./configure --without-termcap Aug 03 14:28:36 hi, I've an issue with libsdl-mixer: http://pastebin.com/rNQmqZTv I bet I'll have to dig in autotools Aug 03 14:28:55 I'll look configure options Aug 03 14:31:30 hmmm Aug 03 14:32:24 http://pastebin.com/GU7n8wPn Aug 03 14:34:24 hi khem Aug 03 14:34:39 with your old patches I've that: Aug 03 14:35:22 arm-oe-linux-gnueabi/bin/ld: warning: libc.so, needed by .../arm-oe-linux-gnueabi/lib/libgcc_s.so.1, not found (try using -rpath or -rpath-link) Aug 03 14:36:31 broonie, indeed, moreover it's stressing to be under constant evaluation Aug 03 14:36:46 specially when autorev is there Aug 03 14:37:36 hi btw Aug 03 14:40:33 hmmm Aug 03 14:40:42 floor seem to come from libc Aug 03 14:41:50 I thought it was in libm, but I guess libc might be a possible candidate too. Aug 03 14:42:35 It's libm here... Aug 03 14:44:29 ok Aug 03 14:44:34 but it says: Aug 03 14:44:58 warning: libc.so, needed by ... libgcc_s.so.1, not found (try using -rpath or -rpath-link) Aug 03 14:45:01 but anyay Aug 03 14:45:03 I'll look Aug 03 14:45:07 with the binutils Aug 03 14:45:12 er, what does that have to do with floor()? Aug 03 14:45:50 hmmm maybe it doesn't find libc and so has a floor issue Aug 03 14:45:59 but I may be ultra-wrong Aug 03 14:46:16 we use eglibc btw Aug 03 14:48:22 GNUtoo|laptop: hello, thats an existing warning Aug 03 14:48:35 hi Aug 03 14:50:36 floor is in libm Aug 03 14:50:49 sorry I'm a bit tired today Aug 03 14:51:07 000284f8 W floor Aug 03 14:51:19 (arm-oe-linux-gnueabi-nm libm.so.6| grep floor) Aug 03 14:51:35 from sysroots/armv6-novfp-oe-linux-gnueabi/lib Aug 03 14:51:45 hmmm Aug 03 14:52:09 it lacks -lm then Aug 03 14:52:14 no? Aug 03 14:54:14 GNUtoo|laptop: yes its in libm Aug 03 14:55:18 GNUtoo|laptop: whats real problem ? Aug 03 14:55:30 is some app not finding floor ? Aug 03 14:55:38 not exactly Aug 03 14:56:20 I've no music in wesnoth, that's because libsdl-mixer didn't find libvorbisdec but libvorbisdec is staged....so I looked at the configure logs and found that: Aug 03 14:56:32 http://pastebin.com/GU7n8wPn Aug 03 14:56:46 I can pastebin the test source code too if you want Aug 03 14:57:38 add -lm to LDFLAGS Aug 03 14:58:29 libmikmod.so could have asked for it too thru DT_NEEDED what does readelf -d libmikmod.so show you Aug 03 14:59:06 I'll look (I didn't know DT_NEEDED) Aug 03 14:59:33 yeah, it looks like libmikmod was wrongly linked without -lm Aug 03 15:00:24 I grep NEEDED and paste? Aug 03 15:00:32 *pastebin Aug 03 15:01:09 http://pastebin.com/c0sr0HRv Aug 03 15:01:17 no libm in NEEDED Aug 03 15:02:05 btw will there be some replacement for ldd planed in the future? Aug 03 15:02:11 (not related) Aug 03 15:02:26 not that I know of. why does ldd need replacing? Aug 03 15:02:35 security issues Aug 03 15:02:52 basically I heard that you could craft a binary which execute code when you ldd it Aug 03 15:03:00 that's because of the way ldd works Aug 03 15:03:24 if I understand well it's a script which uses a variable to put libc in a special mode Aug 03 15:03:31 but my explanations may be inaccurate Aug 03 15:04:13 I'm not shure how other OS like FreeBSD handle that Aug 03 15:04:44 oh, right. this is the thing where you provide your own PT_INTERP. Aug 03 15:05:40 if you're worried about that then you can just patch ldd to not use that mode of operation. Aug 03 15:05:58 ok Aug 03 15:07:05 btw how should I fix my issue? modify the .m4 macros that check for libvorbis? Aug 03 15:07:17 no, fix the linking of libmikmod Aug 03 15:07:48 how? Aug 03 15:08:02 do you have more precision? Aug 03 15:08:10 dunno, you'd need to inspect what it's actually doing during final link. clearly it doesn't have -lm in there and should do. Aug 03 15:09:06 ok Aug 03 15:15:33 GNUtoo|laptop: yeah, ldd is just a shell script and it does run the target binary. See `which ldd`, man ld.so Aug 03 15:16:00 so you could create a binary that would abuse this and ldd on it would execute arbitrary code Aug 03 15:18:47 yes Aug 03 15:18:49 ok Aug 03 15:20:17 but you can list part of what ldd lists with readelf -d Aug 03 15:26:17 ok it uses libtool Aug 03 15:26:49 zub, I know you can do it with some binutils. I didn't remember the exact command tough Aug 03 15:26:53 thanks Aug 03 15:27:17 libmikimod uses libtool Aug 03 15:29:32 and no traces of -lm in: Aug 03 15:29:37 http://pastebin.com/6F6MV7Ee Aug 03 15:29:45 I'll try to fix that Aug 03 15:34:31 hmmm Aug 03 15:34:42 kergoth_, hi. what about warning about redundant DEPENDS_native (redundant as being implied by BBCLASSEXTEND=native DEPENDS)? something like http://paste.debian.net/82137/ Aug 03 15:34:48 kergoth_, just a thought.. Aug 03 15:34:53 I don't find the correct Makefile.am Aug 03 15:35:00 it's nested Aug 03 15:35:09 and there is no Makefile.am in: Aug 03 15:35:34 GNUtoo|laptop: on first glance, EXTRA_OECONF='-lm' might fix it Aug 03 15:35:38 armv6-novfp-oe-linux-gnueabi/libmikmod-3.2.0-beta2-r1/libmikmod-3.2.0-beta2/libmikmod Aug 03 15:35:51 ensc|w, isn't that risky? but I can try that Aug 03 15:36:04 you mean OEMAKE? Aug 03 15:36:12 wrong... EXTRA_OECONF='LIBRARY_LIB=-lm' Aug 03 15:36:13 because that will do ./configure ... -lm Aug 03 15:36:19 ah ok thanks a lot Aug 03 15:38:00 ./configure says LIBS Aug 03 15:38:03 I'll try that Aug 03 15:38:05 thanks a lot Aug 03 15:38:43 blindvt`, hmm, the concept is a good one, but it's almost something better relegated to recipe_sanity, since its a cleanup suggestion, not a problem Aug 03 15:40:25 kergoth_, the reasoning would be that DEPENDS_virtclass-native can be deduced from DEPENDS. See e.g. autoconf.inc or gnome-doc-utils.inc (?!) Aug 03 15:40:37 yeah, I know Aug 03 15:41:31 oh sweet. I wasn't aware of recipe_sanity Aug 03 15:42:13 many of the things it reports are things like this, vars you just don't need to bother setting Aug 03 15:42:25 not sure if itd be easy to put there due to the native logic already being in native, but something to consider Aug 03 15:43:46 http://pastebin.com/SwvWDNj9 still same issue Aug 03 15:43:58 I'll try library_lib Aug 03 15:44:01 i'd estimate that about 3% of all current DEPENDS_{virtclass-,}native are really needed and justified Aug 03 15:44:49 yikes :) Aug 03 15:44:51 but anyway. it was just a thought, maybe someone wants to come up with a real suggestion :) Aug 03 15:45:12 yeah like I said, its a good idea Aug 03 15:51:10 ensc|w, thanks a lot!!! it worked Aug 03 15:51:22 I'll try to rebuild libsdl-mixer now Aug 03 15:51:32 0x00000001 (NEEDED) Shared library: [libm.so.6] Aug 03 16:00:16 03Frans Meulenbroeks  07org.openembedded.dev * r636d7431fd 10openembedded.git/recipes/sane/ (sane-backends_1.0.19.bb sane-backends_1.0.20.bb): Aug 03 16:00:17 sane-backends: removed legacy staging Aug 03 16:00:17 Signed-off-by: Frans Meulenbroeks Aug 03 16:03:12 thanks a lot!!! I've some music now Aug 03 16:03:18 03Frans Meulenbroeks  07org.openembedded.dev * rfece966c35 10openembedded.git/recipes/freetype/ (4 files): Aug 03 16:03:18 freetype: removed 2.1 version Aug 03 16:03:18 this one had an empty do_install so probably the removal Aug 03 16:03:18 of do_stage broke it; as it is quite old removed it Aug 03 16:03:18 Signed-off-by: Frans Meulenbroeks Aug 03 16:03:28 I'll push as soon as I go to to the wifi cafee Aug 03 16:13:44 03Scott Garman  07org.openembedded.dev * r9ed124062c 10openembedded.git/recipes/tar/ (tar_1.22.bb tar_1.23.bb): Aug 03 16:13:44 tar: Upgrade to v1.23 Aug 03 16:13:44 * Addresses CVE-2010-0624 Aug 03 16:13:44 Signed-off-by: Scott Garman Aug 03 16:13:44 Acked-by: Frans Meulenbroeks Aug 03 16:14:34 03Frans Meulenbroeks  07org.openembedded.dev * rbc14d2764a 10openembedded.git/recipes/squashfs-tools/ (squashfs-tools-native_4.0.bb squashfs-tools.inc): Aug 03 16:14:34 squashfs-tools: added dependency to attr/attr-native Aug 03 16:14:34 Signed-off-by: Frans Meulenbroeks Aug 03 16:14:45 03Frans Meulenbroeks  07org.openembedded.dev * rf70303411d 10openembedded.git/recipes/udev/attr_2.4.44.bb: Aug 03 16:14:46 attr: added BBCLASSEXTEND = "native" Aug 03 16:14:46 added BBCLASSEXTEND = "native" Aug 03 16:14:46 this is needed for squashfs-tools, otherwise it is not possible to build Aug 03 16:14:46 mksquashfs and hence no squashfs creation on the build system Aug 03 16:14:46 Signed-off-by: Frans Meulenbroeks Aug 03 16:19:09 does anyone want to make psplash work with an 8 bpp framebuffer and send me a patch? Aug 03 16:23:56 gchiii: what an offer :) Aug 03 16:25:32 gchiii: and to mee too Aug 03 16:25:49 we in kexecboot still have no support for 8bpp modes Aug 03 16:27:02 zub: in my meager defense, i did look at it and make an attempt, then got a big headache. i could offer my grattitude and possibly a starbucks as reward (if i knew a good way to give starbucks over tcp/ip) Aug 03 16:27:32 :) Aug 03 16:28:05 hi mickey|office Aug 03 16:28:08 gchiii: what hardware you using? Aug 03 16:34:28 Jay7: an at91rm9200, with a slightly broken epson s1d13706 display chip, and some other display that is frustratingly close to an s1d15605 driver Aug 03 16:41:24 gchiii: one time I'll look on 8bpp.. Aug 03 16:41:32 but not this time Aug 03 17:09:33 hmm, has anyone tried the cron daemon from busybox? it seems to completely ignore /etc/crontab Aug 03 17:10:06 I wonder if I am doing anything wrong Aug 03 18:15:55 hm, there is really no recipe for literki in oe yet ? Aug 03 18:20:18 there are one or two packages we do not have recipes for yet Aug 03 18:24:45 kergoth: is there a way to pass answer to a process invoked from python script Aug 03 18:25:17 kergoth: I am calling ssh and if I am logging into that machine for first time it asks for yes and no Aug 03 18:25:51 to add it Aug 03 18:26:07 use popen, you can access stdin, stdout, and stderr Aug 03 18:26:18 that's assuming it asks on stdin, and not directly via /dev/tty or something, of course Aug 03 18:26:32 ah I am using popen Aug 03 18:26:47 Crofton|road: I know :) I'm just puzzled that its authors ships an ipk but there is no .bb anywhere Aug 03 18:26:51 writing one Aug 03 18:27:04 if you use subprocess, the popen object has stdin, stderr, and stdout members Aug 03 18:27:10 p.stdin.write() Aug 03 18:27:19 alternatively, you could look at Pexpect Aug 03 18:31:21 * kergoth_ 's working on image_features + package_dbg again, need to get that solid and merged so he can move on to the next pending item Aug 03 18:31:45 content, err = p.communicate("Yes") Aug 03 18:31:52 kergoth_: will that do it ? Aug 03 18:32:12 kergoth: yeah image_features is nice think Aug 03 18:32:15 ah, right, I forgot communicate can take stdin. it might, it depends on what it does with the input before it sends the prompt. most likely will work Aug 03 18:32:17 I have your patches Aug 03 18:32:22 in my local tree Aug 03 18:32:31 wanna a copy :_ Aug 03 18:32:32 otherwise you could read stdout until you get the prompt, then stdin.write(), then communicate() to complete it Aug 03 18:32:50 right now I'm testing a fix for kernel-modules dep'ing on -dbgs Aug 03 18:32:55 k Aug 03 18:32:56 think perl still needs a fix for that too Aug 03 18:33:40 hmmm this prompt is for some machines in a loop and some not Aug 03 18:34:03 there's 1) add PACKAGE_INSTALL_ATTEMPTONLY to rootfs_ipk, 2) add image_features, 3) fix recipes that rdep on all packages to exclude them, then it should be good Aug 03 18:34:10 hmm Aug 03 18:34:35 well, if you get tired of messing with it, and don't mind another dep, Pexpect would work well Aug 03 18:34:43 expect was designed for that sort of scripting Aug 03 18:34:49 wait X seconds for this string, send that, etc Aug 03 18:35:01 * kergoth_ hasn't messed with pexpect himself though Aug 03 18:35:16 if all you have to deal with is the single prompt, probably more than you'd need Aug 03 18:36:49 Use communicate() rather than .stdin.write, .stdout.read or .stderr.read to avoid deadlocks due to any of the other OS pipe buffers filling up and blocking the child process. Aug 03 18:37:11 the buffering can be a problem when you do that sort of scripting of interactive apps Aug 03 18:37:43 NOTE: linux-omap-psp-2.6.32-r84+gitra6bad4464f985fdd3bed72e1b82dcbfc004d7869: do_patch: Elapsed time: 8.05 seconds Aug 03 18:37:43 NOTE: linux-omap-psp-2.6.32-r84+gitra6bad4464f985fdd3bed72e1b82dcbfc004d7869: do_patch: Cpu usage: 25.3% Aug 03 18:37:50 heh, i really underestimated how much cpu these other tasks take Aug 03 18:43:31 hm, trying to build literki fails, seemingly because freetype-config outputs "-isystem//tmp/sysroots/armv4t-oe-linux-gnueabi/usr/include" - anyone knows about that ? Aug 03 18:48:34 fully, since "./tmp/sysroots/x86_64-linux/usr/bin/armv4t-oe-linux-gnueabi/freetype-config --cflags" run manually gives correct output Aug 03 18:48:40 s/fully/funny/ Aug 03 18:49:29 (wrong statements, forget it) Aug 03 18:59:56 anybody having problems building libxcb? Aug 03 19:01:23 "No module named xcbgen.state" Aug 03 19:03:54 Attempting "apt-get install python-xcbgen" to fix. Aug 03 19:07:36 still fails. :( Aug 03 19:39:47 03Roman I Khimov  07org.openembedded.dev * rd4eb4f3ea8 10openembedded.git/recipes/openldap/ (openldap-2.4.21/openldap-m4-pthread.patch openldap_2.4.21.bb): (log message trimmed) Aug 03 19:39:47 openldap: update 2.4.21 to 2.4.23 Aug 03 19:39:47 * bugfixes mostly, considered as safe upgrade Aug 03 19:39:47 * fixes CVE-2010-0211 and CVE-2010-0212 Aug 03 19:39:47 * switch to PACKAGES_DYNAMIC for backends, allows to kill some recipe bloat, Aug 03 19:39:47 might be nice to do additional OECONF cleanup later (probably adding more Aug 03 19:39:48 modules as they're nicely packaged) Aug 03 19:44:16 03Denys Dmytriyenko  07org.openembedded.dev * r69d694e393 10openembedded.git/recipes/strace/strace_4.5.20.bb: Aug 03 19:44:16 strace: add latest 4.5.20 version Aug 03 19:44:16 Earlier versions may have build issues associated with linux header changes: Aug 03 19:44:16 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=560516 Aug 03 19:44:16 Thanks to Pierluigi Passaro Aug 03 19:44:16 Signed-off-by: Denys Dmytriyenko Aug 03 19:46:36 03Khem Raj  07org.openembedded.dev * r4f96a91b1b 10openembedded.git/ (4 files in 2 dirs): Aug 03 19:46:37 bitbake.conf,cmake.bbclass,cross.bbclass,icecc.bbclass: Point the toolchain paths to new multi-machine friendly toolchain install location. Aug 03 19:46:37 Signed-off-by: Khem Raj Aug 03 19:46:37 Acked-by: Koen Kooi Aug 03 19:46:37 Acked-by: Martin Jansa Aug 03 19:46:37 03Khem Raj  07org.openembedded.dev * r3c5585b873 10openembedded.git/recipes/mesa/mesa_7.2.bb: Aug 03 19:46:38 mesa_7.2.bb: Tinker DEBUG_OPTIMIZATION to avoid gcc ICE Aug 03 19:46:38 Signed-off-by: Khem Raj Aug 03 19:46:39 03Khem Raj  07org.openembedded.dev * r57be258994 10openembedded.git/recipes/uclibc/uclibc_git.bb: Aug 03 19:46:39 uclibc_git.bb: move to latest git tip. Aug 03 19:46:40 Signed-off-by: Khem Raj Aug 03 19:46:41 03Khem Raj  07org.openembedded.dev * rc52193ee03 10openembedded.git/recipes/binutils/binutils-cross.inc: Aug 03 19:46:41 binutils-cross.inc: Dont install duplicate copies of tools create relative symlinks instead. Aug 03 19:46:42 * Dont generate shared libraries and bfd. Aug 03 19:46:42 Signed-off-by: Khem Raj Aug 03 19:46:43 Acked-by: Koen Kooi Aug 03 19:46:55 mpfr_3.0.0.bb: Override needs to be thumb instead of armv4t Aug 03 19:46:55 * This covers the case when compiling with thumb mode Aug 03 19:46:55 Signed-off-by: Khem Raj Aug 03 19:46:55 03Khem Raj  07org.openembedded.dev * r07733978e2 10openembedded.git/recipes/ (10 files in 2 dirs): Aug 03 19:46:55 gcc,eglibc: Use the paths for mutli-machine safe toolchain install. Aug 03 19:46:55 Signed-off-by: Khem Raj Aug 03 19:46:56 Acked-by: Koen Kooi Aug 03 19:46:56 Acked-by: Martin Jansa Aug 03 20:01:43 03Roman I Khimov  07org.openembedded.dev * r060285f9fd 10openembedded.git/recipes/byacc/ (byacc-native_20081225.bb byacc.inc): Aug 03 20:01:43 byacc: kill -native recipe, replace with BBCLASSEXTEND Aug 03 20:01:43 Signed-off-by: Roman I Khimov Aug 03 20:02:51 03Graham Gower  07org.openembedded.dev * r11d08db810 10openembedded.git/ (27 files in 10 dirs): (log message trimmed) Aug 03 20:02:51 opkg recipe overhaul - removal of opkg-nogpg and opkg-nogpg-nocurl. Aug 03 20:02:51 None of the distros in OE appear to want/need gpg or curl support in opkg; Aug 03 20:02:51 so have opkg.inc remove them by default. This diff makes things more consistent Aug 03 20:02:51 across all distros/tasks/images. Aug 03 20:02:51 The slugos recipe is kept intact and further divided from the more modern Aug 03 20:02:52 recipes. The opkg-native recipe is the sane-srcrev version even when building Aug 03 20:07:11 03Koen Kooi  07org.openembedded.dev * r6332fb06d8 10openembedded.git/recipes/opkg/ (opkg.inc opkg_svn.bb): opkg: first stab at restoring upgrade paths Aug 03 20:15:48 anyone able to build libxcb? Aug 03 20:26:18 jkridner: lemme try Aug 03 20:26:38 jkridner: I've built it ok libxcb-1.6-r3.1.do_rm_work Aug 03 20:26:58 jkridner: for armv4t and armv7 Aug 03 20:28:41 I'm on an Ubuntu 10.04 machine with python-xcbgen installed, but still getting xcbgen.state module missing error. Aug 03 20:29:09 koen suggested 'bitbake -c clean python-native', but that made no difference. Aug 03 20:29:21 python-native seems to be being built for my task. Aug 03 20:30:20 does anyone know what provides python's threading? Aug 03 20:32:07 python-native gets built when I run 'bitbake libxcb'. Aug 03 20:33:01 I don't know if there is some way to skip that dependency. Aug 03 20:33:18 jkridner: what does bitbake -g say Aug 03 20:33:22 who pulls in python Aug 03 20:33:51 jkridner: my box is also ubuntu 10.04 Aug 03 20:33:57 hey does anyone know what should I include in my image to have python-threading? Aug 03 20:34:52 marcompile: doesnt python come with it already Aug 03 20:36:11 khem, I tought so, but I included python on my image and its not there. I can install python-threading in angstrom (with opkg), but I don't see the recipe for bitbake Aug 03 20:36:38 maybe that ipk is built with python? Aug 03 20:42:37 libxcb -> libxslt-native -> libxml2-native -> python-native Aug 03 20:49:59 marcompile: hmm thats a package emitted from python recipe Aug 03 20:50:10 marcompile: a recipe can emit more than 1 packages Aug 03 20:50:31 I see now Aug 03 20:51:03 but my recipe shouldn't DEPEND on python-threading for example, it should just IMAGE_INSTALL that, correct? Aug 03 20:51:22 see recipes/python/python-2.6-manifest.inc Aug 03 20:51:49 it should depend on python Aug 03 20:51:54 well Aug 03 20:52:07 RDEPENDS_${PN} = "python-threading" Aug 03 20:52:13 should work too Aug 03 20:52:45 jkridner: ah native python Aug 03 20:52:52 jkridner: you need to build it Aug 03 20:52:56 unfortunately Aug 03 20:53:10 btw. my build is going well thus far for libxcb Aug 03 20:53:34 well, it is built, but it seems to fail. not sure why it would work for you. Aug 03 20:54:55 I will know soon Aug 03 20:55:00 if it builds or not Aug 03 20:55:07 its still doing other stuff Aug 03 21:00:19 jkridner: could you paste yout build.log somewhere for configure stage Aug 03 21:00:42 it could be that XCBPROTO_XCBPYTHONDIR is not set properly for some reason Aug 03 21:02:21 khem: do you just mean my log.do_compile? Aug 03 21:02:33 log.do_compile and log.do_configure Aug 03 21:05:52 khem: http://www.beagleboard.org/~arago/libxcb-debug/ Aug 03 21:06:27 I don't see that property being set. Aug 03 21:06:50 did it build for you? Aug 03 21:10:14 jkridner: seeing the same issue here, debugging it Aug 03 21:10:33 I believe it is due to global do_stage removal, though... Aug 03 21:12:15 did we add 2.6.35 recipes already :) Aug 03 21:12:21 its 2 days old now Aug 03 21:13:06 btw, libxcb error - http://pastebin.com/3SUCzK7x Aug 03 21:14:40 * jkridner hates breakages on release-candidate day. :( Aug 03 21:16:07 denix: yeah Aug 03 21:16:09 jkridner: why do you update to the latest dev branch then? :) Aug 03 21:16:10 # Aug 03 21:17:00 checking for python script directory... ${prefix}/lib/python2.6/site-packages Aug 03 21:17:00 checking for python extension module directory... ${exec_prefix}/lib/python2.6/site-packages Aug 03 21:17:15 I wonder why ${prefix} did not expand Aug 03 21:17:27 denix: :) because I introduced patches and haven't actually had a working version yet (also not a good thing on release candidate day). Aug 03 21:18:31 hi khem Aug 03 21:18:51 ant___: hello Aug 03 21:19:06 ant___: I have gfortan thing yet in my sandbox Aug 03 21:19:16 want to fix it more elegantly Aug 03 21:19:27 great Aug 03 21:19:36 ant___: btw did that do_stage removal work for klibc Aug 03 21:19:43 I'm looking at klibc again, now Aug 03 21:19:57 but install -m 755 klcc/klcc ${TOOLCHAIN_PATH}/bin/${TARGET_PREFIX}klcc Aug 03 21:20:12 works but then the compiler is not found... Aug 03 21:20:27 | checking for arm-angstrom-linux-gnueabi-gcc... arm-angstrom-linux-gnueabi-klcc Aug 03 21:20:27 | checking whether the C compiler works... no Aug 03 21:20:53 jkridner: checking XCBPROTO_XCBPYTHONDIR... /home/ubuntu/angstrom-setup-scripts/build/tmp-angstrom_2008_1/sysroots/armv7a-angstrom-linux-gnueabi/usr/lib/python2.6/site-packages Aug 03 21:22:28 my box is still churning toolchain should not have updated gcc :) Aug 03 21:23:43 jkridner: does xcbgen.state Aug 03 21:23:53 exist in /home/ubuntu/angstrom-setup-scripts/build/tmp-angstrom_2008_1/sysroots/armv7a-angstrom-linux-gnueabi/usr/lib/python2.6/site-packages Aug 03 21:27:18 hmm, I don't even have /usr/lib/python2.6 in my sysroots... Aug 03 21:28:32 but I do have it in i686-linux, instead of armv7a-*-linux Aug 03 21:30:10 denix: thats ok Aug 03 21:30:22 denix: it should have pick from native one I guess Aug 03 21:30:41 khem: another strange path? http://pastebin.com/FMtYU447 Aug 03 21:31:38 khem: but still, there is no xcb* in python2.6 or site-packages Aug 03 21:33:42 ant___: thats ok Aug 03 21:33:49 we could remove it though Aug 03 21:33:57 denix: yeah Aug 03 21:34:06 * khem tried to bake libxcb git Aug 03 21:35:30 hmm same error Aug 03 21:35:34 khem: xcb-proto is supposed to stage xcbgen - it even has it in WORKDIR/image, but... Aug 03 21:39:20 grr Aug 03 21:39:22 hmm Aug 03 21:39:51 khem: here's xcb-proto do_stage http://pastebin.com/gHhxEUsw and it clearly supposed to stage xcbgen, but it's not in sysroots/armv7a-*-linux Aug 03 21:43:19 denix: thats what I concluded thus far :) Aug 03 21:43:30 lemme see Aug 03 21:50:20 jkridner: denix : alright Aug 03 21:50:23 here is a patch Aug 03 21:51:57 http://pastebin.com/TUdJd5QK Aug 03 21:52:16 after applying Aug 03 21:52:48 bitbake -c clean xcb-proto Aug 03 21:52:58 bitbake libxcb Aug 03 21:53:30 are you sure about the patch? doesn't seem right Aug 03 21:53:36 yes Aug 03 21:54:21 unless you meant just the xcb-proto.inc Aug 03 21:54:26 WAIT !!!! Aug 03 21:54:27 wrong one Aug 03 21:55:18 http://pastebin.com/38bc2FhZ Aug 03 21:55:24 yes :) Aug 03 21:55:30 git diff Aug 03 21:55:36 git diff . Aug 03 21:55:40 http://pastebin.com/XnnR6ix9 Aug 03 21:55:40 make difference Aug 03 21:55:58 yep Aug 03 21:56:00 right, I already extracted the needed piece :) Aug 03 21:56:12 QA tested before pastebininit :) Aug 03 21:56:20 smart boy Aug 03 21:57:07 khem: I'm trying a rebuild from pstage for armv4 Aug 03 21:57:09 fails Aug 03 21:57:23 nite all Aug 03 21:57:27 cya tomorrow Aug 03 21:57:39 i really hate opkg sometimes Aug 03 21:57:45 ant___: hm Aug 03 21:57:47 gnu-config: sh: quilt: command not found Aug 03 21:57:54 kergoth_: heh Aug 03 21:57:57 this should be working. it says there's no package, but the package exists Aug 03 21:58:11 ERROR: Task 1148 (virtual:native:/oe/openembedded/recipes/gnu-config/gnu-config_git.bb, do_patch) failed with 256 Aug 03 21:58:13 same thing happened to me Aug 03 21:58:30 that was for kergoth Aug 03 22:00:34 argh, xcb is not INC_PR-ed :) Aug 03 22:01:29 I am doing that here Aug 03 22:01:29 wtf Aug 03 22:02:04 khem: ah, if you are about to push the fix, I'll wait then :) Aug 03 22:03:22 denix: http://pastebin.com/NT9M3i4D Aug 03 22:03:32 hmm Aug 03 22:03:37 i can install this package outside of do_rootfs Aug 03 22:03:42 but it fails inside do_rootfs Aug 03 22:03:43 * kergoth_ is confused Aug 03 22:04:07 denix: I will push it Aug 03 22:04:32 khem: thanks Aug 03 22:06:37 ok builds, ship it :) Aug 03 22:06:50 denix: test it though for runtime I havent done that Aug 03 22:07:39 03Khem Raj  07org.openembedded.dev * rb0061845a6 10openembedded.git/recipes/xcb/ (8 files): Aug 03 22:07:39 xcb-proto: Get rid of legacy staging. Aug 03 22:07:39 * Convert to using INC_PR Aug 03 22:07:39 Signed-off-by: Khem Raj Aug 03 22:08:37 ant___: why is quilt missing Aug 03 22:09:04 thanks khem. Aug 03 22:09:24 khem: fwiw is not on my Ubuntu buildhost Aug 03 22:09:34 jkridner: np Aug 03 22:09:42 I suppose is a missing dep Aug 03 22:09:51 ant___: then it should install quilt-native Aug 03 22:10:01 heh Aug 03 22:10:10 fails too :/ Aug 03 22:10:29 * khem doesnt have packaged_staging here Aug 03 22:10:37 I have to clean out some things before... Aug 03 22:10:41 ERROR: Task 1148 (virtual:native:/oe/openembedded/recipes/gnu-config/gnu-config_git.bb, do_patch) failed with 256 Aug 03 22:10:51 now Aug 03 22:10:52 ERROR: Task 74 (virtual:native:/oe/openembedded/recipes/linux-libc-headers/linux-libc-headers_2.6.31.bb, do_patch) failed with 256 Aug 03 22:11:44 ant___: figure out why quilt-native was ignored by packaged_staging Aug 03 22:12:15 seems those two are first failing packages Aug 03 22:16:46 kergoth: I use bitbake -g -u depexp to find deplist but when I dont have gui whats the other way ? Aug 03 22:17:06 khem: I prefer to pull and rebuild anew Aug 03 22:17:33 so I can test the commits Aug 03 22:18:16 ant___: ok Aug 03 22:18:25 btw did you adandon gentoo ? Aug 03 22:18:32 perhaps I'm missing some bit... Aug 03 22:18:35 abandon Aug 03 22:18:50 no, is still on first partition ;) Aug 03 22:19:07 just that binutils infamous issue Aug 03 22:19:09 heh btw now the binutils problem should not happen Aug 03 22:19:21 ok, I'll be glad to verify :) Aug 03 22:19:22 I have averted it Aug 03 22:19:53 I encourage devs to use various distros Aug 03 22:19:57 this ubuntu is a bit unstable imho Aug 03 22:20:03 heh Aug 03 22:20:06 which one Aug 03 22:20:12 10.4 + updates Aug 03 22:20:19 hmmm works ok here Aug 03 22:20:29 never had issues since upgrade Aug 03 22:20:33 I have to say it survived to kernel and grub2 upgrades ..;) Aug 03 22:20:40 I have to reboot only when kernel update happens Aug 03 22:20:47 otherwise keeps cranking Aug 03 22:21:05 bah..Firefox freezes system now and then... Aug 03 22:21:23 at least, I have Firefox and XChat Aug 03 22:21:31 latter seems more tested ;) Aug 03 22:21:39 less error-prone :p Aug 03 22:21:59 I use reconq Aug 03 22:22:00 oh, yes, some terminals are open... Aug 03 22:22:05 less firefox Aug 03 22:22:08 and no xcaht Aug 03 22:22:10 I hope it can bear that Aug 03 22:22:17 so I prolly am safe :) Aug 03 22:22:20 heh Aug 03 22:22:30 XChat or Firefox? Aug 03 22:22:37 I dont use either Aug 03 22:22:39 I could bet Aug 03 22:23:27 btw since some days the applet for shutdown is broken Aug 03 22:23:47 does logout it seems Aug 03 22:24:07 ok, let me reboot in gentoo then Aug 03 22:25:19 heh someone just asked how to get setup with gentoo here at work :) Aug 03 22:25:22 ironical Aug 03 22:26:06 ant__: booted into gentoo :) moo Aug 03 22:26:32 omg..is like driving Ferrari Aug 03 22:26:47 gnome is on steroids (no compiz) Aug 03 22:27:00 is 10x snappy :p Aug 03 22:28:10 ant__: when compared to what ? Aug 03 22:28:28 to Ub*** Aug 03 22:28:43 hmmm ok cool Aug 03 22:28:50 I have not tried gentoo in years Aug 03 22:29:02 with today's machine is fun Aug 03 22:29:04 but I should try it when I buy a fast machine Aug 03 22:29:41 emerge system is about 1,5 hours Aug 03 22:29:51 with Xeon class Aug 03 22:29:57 quad Aug 03 22:30:08 can't imagine with SSD Aug 03 22:48:36 good night Aug 03 23:03:00 grrr Aug 03 23:03:06 this is proving to be much more trouble than it should be. Aug 03 23:03:14 * kergoth_ thinks about revamping rootfs*.bbclass Aug 03 23:08:55 are there some best practice wrt software that should build a helper tool as native but does not ? Something we could tell automake maybe ? Aug 03 23:12:02 automake doesn't have built in support for that, it has no knowledge of or support for crosscompilation directly. you'll have to add your own rules for the build machine tools to the .am Aug 03 23:13:17 do you know a package that does that already for an example ? Aug 03 23:14:48 grep around for BUILD_CC or CC_FOR_BUILD Aug 03 23:15:00 ok Aug 03 23:15:23 also, do we need AC_CANONICAL_BUILD in configure.in ? Aug 03 23:16:17 if you want to support crosscompilation, you likely want all of those, yes, but it depends on what you're needing to do with your buildsystem Aug 03 23:16:27 goddamnit opkg, this makes no sense Aug 03 23:17:08 it's a tool built to compile patterns (mkpat in gnugo) Aug 03 23:26:50 guh Aug 03 23:26:55 we really need a replacement for opkg Aug 03 23:27:56 or improve it Aug 03 23:28:09 a *lot* Aug 03 23:28:13 this is terrible Aug 03 23:28:18 * kergoth_ grumbles Aug 03 23:28:44 summarise your experiences Aug 03 23:28:54 in a rant/blog Aug 03 23:29:02 good idea Aug 03 23:29:10 get this Aug 03 23:29:15 if i add this package to PACKAGE_INSTALL, it installs fine Aug 03 23:29:29 if i run a second opkg-cl install after that one to install the exact same package, it says the package doesn't exist Aug 03 23:29:38 with the exact same commandline arguments Aug 03 23:29:52 * kergoth_ scratches head Aug 03 23:30:29 the only idea that comes to mind is that opkg's status/info is changed in that install in a way that prevents subsequent installs, i.e. losing the opkg update info Aug 03 23:31:39 hmm Aug 03 23:32:42 too bad AC_CANONICAL_BUILD itself does not give us a BUILD_CC for free... Aug 03 23:39:13 yann, autoconf/automake support for crosscompilation is cursory at best, and libtool's is nonexistant Aug 03 23:54:33 aha Aug 03 23:54:39 i can't believe this hasn't been spotted before Aug 03 23:59:23 kergoth_: any easy way to build 2 versions (native+cross) of a given c file ? I remember seeing something long ago, bug it may be just for cmake :) Aug 03 23:59:48 ugh, that gets ugly fast. i don't recall the best way to do that offhand, without writing the rules manually Aug 04 00:02:14 nothing like libfoo_bar_CC vs prog_bar_CC ? Aug 04 00:02:25 can't find that in the doc... Aug 04 00:04:18 ah, google helps, that exists Aug 04 00:05:10 not sure that works for CC, though, maybe just CFLAGS and such Aug 04 00:05:13 let's test Aug 04 00:07:28 right, that does not :( Aug 04 00:07:56 that's frustrating, full support is *almost* there Aug 04 00:21:45 kergoth_: I haven't put all together yet, but that will finally work at low cost Aug 04 00:22:16 hm well Aug 04 00:44:39 03Chris Larson  07org.openembedded.dev * racc720fa80 10openembedded.git/classes/rootfs_ipk.bbclass: Aug 04 00:44:39 rootfs_ipk.bbclass: install the package manager in a separate pass from the other packages Aug 04 00:44:39 It seems that installing opkg/opkg-collateral blows away the configuration we Aug 04 00:44:39 need to install more packages, so lets make that explicit by installing those Aug 04 00:44:39 things last. Aug 04 00:44:39 Signed-off-by: Chris Larson Aug 04 00:44:43 03Chris Larson  07org.openembedded.dev * rd67c4e5ba4 10openembedded.git/ (3 files in 2 dirs): Aug 04 00:44:44 Recipe fixups for package_dbg usage Aug 04 00:44:44 Signed-off-by: Chris Larson Aug 04 00:45:25 03Chris Larson  07org.openembedded.dev * r2de66c5504 10openembedded.git/classes/rootfs_ipk.bbclass: Aug 04 00:45:25 rootfs_ipk.bbclass: add support for PACKAGE_INSTALL_ATTEMPTONLY Aug 04 00:45:26 Signed-off-by: Chris Larson Aug 04 00:47:05 khem, progress, these commits + the image_features patch get you something functional Aug 04 00:54:32 gah Aug 04 00:54:34 hmm Aug 04 00:56:13 * kergoth_ tests something Aug 04 00:57:48 * kergoth_ mutters Aug 04 01:12:12 damned, there are too many intrications between gnugo libs and progs, I give up on that method Aug 04 01:12:34 I'll have to build a native package first, and use the binaries from there Aug 04 01:16:10 using BBCLASSEXTEND=native, how can I specify that gnugo DEPENDS on gnugo-native ? Aug 04 01:16:44 DEPENDS_virtclass-native ? Aug 04 01:18:38 kergoth, what's PACKAGE_INSTALL_ATTEMPTONLY about? Aug 04 01:19:06 ah no, that does not work Aug 04 01:19:35 kergoth: any hint ? or should I stick to "inherits native" ? Aug 04 01:23:45 grg: it ignores a failure to install them. tries, doesn't fail Aug 04 01:23:52 grg: rootfs_deb already had it, just copied it over Aug 04 01:24:13 kergoth, sure. But when is that useful? Aug 04 01:24:22 grg: need it for my IMAGE_FEATURES patch, that + package_dbg and you can pull in all debugging files for all packages in your image in one go Aug 04 01:24:58 IMAGE_FEATURES += "dbg dev" will pull in all -dbg and -dev packages for the packages going into the image Aug 04 01:27:16 kergoth, hmm that sounds pretty good actually Aug 04 01:27:19 needed to use attemptonly to ensure it doesn't explode if a -dev/-dbg/-doc package doesn't exist Aug 04 01:27:21 http://github.com/kergoth/OpenEmbedded/commit/d09ccec62ec6f92acbbbdd24ac4f4c192de4e476 Aug 04 01:27:49 INHERIT += package_dbg will make it work a bit better, since it makes the -dbg packages per binary package, rather than per recipe Aug 04 01:32:51 kergoth, the only problem with having all dbg packages installed is gdb runs out of memory when trying to debug certain things with large dependencies, e.g. a gtk+ app :( Aug 04 01:33:10 on small memory machines anyway Aug 04 01:33:23 heh, hard to avoid that, probably best off using gdbserver in that case Aug 04 01:56:35 * kergoth thinks about renaming IMAGE_FEATURES to something relating to 'package groups' instead, more generic Aug 04 02:47:07 grg: heh there is gdbserver for a reason Aug 04 02:47:37 I have a massive patch to not inherit autotools_stage Aug 04 02:51:26 kergoth: can you do some braindump about srctree Aug 04 02:51:44 anyone here familiar with the zippy board, how it is detected and modules loaded? Aug 04 02:51:45 19:51 < khem> kergoth: can you do some braindump about srctree **** ENDING LOGGING AT Wed Aug 04 02:59:57 2010