**** BEGIN LOGGING AT Fri Nov 19 02:59:58 2010 Nov 19 04:07:04 huh, that was easy Nov 19 04:07:12 * kergoth plays around with a nicer bitbake parsing progress bar Nov 19 04:14:05 * kergoth rather likes that this shows an ETA Nov 19 04:15:14 * mrj10 is psyched for multithreaded parsing, it dominates my workflow these days with fixing and rebuilding little packages all over the place Nov 19 04:19:47 kergoth: heh may be you should think about playing video of the day from youtube. I am sure not many will complain about parsing time then :) Nov 19 04:20:31 heh Nov 19 04:22:26 khem: check out the current parallel-parsing branch, added experimental usage of the 'progressbar' python module for a nicer progress bar Nov 19 04:24:29 is there a way to "unstage" something? Nov 19 04:24:33 non-cached case is smoking fast again, btw Nov 19 04:24:36 grg: -c clean Nov 19 04:25:34 kergoth, nah, that leaves headers and stuff in my sysroot Nov 19 04:25:47 well, -c clean uninstalls the pstage package from the sysroot Nov 19 04:25:50 so don't know what to tell you Nov 19 04:26:16 i suppose i'm going to have to use pstage one day Nov 19 04:27:28 its default unless you're on old OE Nov 19 04:31:35 i have found that -c clean doesn't remove all the headers/libs from build/tmp-/sysroots either Nov 19 04:31:46 time to remove PSTAGING_DISABLED = "1" from my local.conf Nov 19 04:31:47 err, i have also found that -c clean does not remove Nov 19 04:32:43 this was one of the biggest reasons its default now :P Nov 19 04:32:45 kergoth: cool. Lemme try it on my core2duo Nov 19 04:33:08 i don't have PSTAGING_DISABLED, i'm using the default local.conf from angstrom-setup-scripts Nov 19 04:33:24 i'm probably just missing something Nov 19 04:34:09 kergoth: ah this is somewhre on github ? Nov 19 04:34:23 why didnt u create a user branch on bitbake repo Nov 19 04:34:27 khem: git clone git://github.com/kergoth/bitbake; cd bitbake; git checkout parallel-parsing Nov 19 04:34:34 because git.oe doesn't let you remove your own branches Nov 19 04:34:37 or rebase them, or.. Nov 19 04:34:40 just a giant pain in the ass Nov 19 04:34:49 so i never put personal branches there now Nov 19 04:34:58 (though bitbake may not have the same hooks as oe, i don't recall) Nov 19 04:35:50 strange you sould be able to remove own branches at least is that disabled ? Nov 19 04:36:03 well sort of makes sense Nov 19 04:36:15 one could delete a remote branch Nov 19 04:37:03 http://pastebin.com/0n5KakG7 Nov 19 04:37:08 this is really nasty Nov 19 04:37:15 you can delete any branch, we'd need access control in the hooks within the repo, which we don't have right now Nov 19 04:37:22 s/can delete/could delete/ Nov 19 04:37:36 at this point my build machine is hosed Nov 19 04:37:53 yikes Nov 19 04:39:16 yeah Nov 19 04:39:28 somethine spawns a bunch of processes Nov 19 04:39:29 Crofton: which pakcage is that Nov 19 04:39:48 is it regenerating the libtool ? Nov 19 04:39:56 or using an existing one Nov 19 04:39:59 http://pastebin.com/MMfjKUDp Nov 19 04:40:56 try to disable parallel make for this recipe Nov 19 04:41:45 hmm there is pastedroid Nov 19 04:42:21 which one? Nov 19 04:42:59 Crofton: which recipe is causing this Nov 19 04:43:38 It is hard to tell Nov 19 04:45:17 runnign bitbake into a tee Nov 19 04:46:15 hmmm it should say at the end which recipe failed Nov 19 04:46:20 or ps -ef Nov 19 04:46:25 will tell you too Nov 19 04:46:29 prolly Nov 19 04:51:42 I don't think the recipe fails Nov 19 04:51:55 guile and xmms are in configure when it blows up Nov 19 04:52:03 checking the changelogs/ Nov 19 04:54:27 nothing recent Nov 19 05:07:39 khem: try it out? Nov 19 05:07:45 hmmmmmmm Nov 19 05:08:09 * kergoth wonders how to resurrect chunking of the tasks without making the code uglier Nov 19 05:08:58 ok Nov 19 05:09:05 I will ahve to come back to this tomorrow Nov 19 05:09:24 hopefully my build build for work does not catch this problem Nov 19 05:09:41 and hopefully someone tries a clean build on the release branch Nov 19 05:31:44 hmm Nov 19 05:46:31 handhelds.org is down Nov 19 05:46:50 there are some recipes that fetch from cvs server on it Nov 19 05:50:42 kergoth: nice progress bar and ETA Nov 19 05:51:18 kergoth: definitely better Nov 19 05:51:35 I see 3 python processes instead of 1 Nov 19 05:52:16 kergoth_: time to parsing is almost half now Nov 19 06:00:21 khem, I hear that a lot...that hh.org is down. Nov 19 06:01:55 03Graham Gower  07master * rd7fbdeb376 10openembedded.git/recipes/freesmartphone/mioctl_git.bb: (log message trimmed) Nov 19 06:01:55 mioctl_git.bb: Add glib-2.0 to DEPENDS. Nov 19 06:01:55 | make[3]: Entering directory `/mnt/oe/tmp/work/mipsel-oe-linux/mioctl-1_0.1.0+gitr0+43fae6cf5e3aa57f5d7fed467896d2d4d0f69679-r0/git/tools/mioctl/src' Nov 19 06:01:55 | mipsel-oe-linux-gcc -march=mips32 -DHAVE_CONFIG_H -I. -I.. -pthread Nov 19 06:01:55 -I/mnt/oe/tmp/sysroots/mipsel-oe-linux/usr/include/glib-2.0 Nov 19 06:01:55 -I/mnt/oe/tmp/sysroots/mipsel-oe-linux/usr/lib/glib-2.0/include -include Nov 19 06:01:56 ../config.h -isystem/mnt/oe/tmp/sysroots/mipsel-oe-linux/usr/include Nov 19 06:04:25 ka6sox: I wonder what we should do with recipes which fetch from those servers Nov 19 06:04:44 ka6sox: if its not maintained may be time to get rid of those recipes Nov 19 06:05:07 khem, either that *or* make a repo of the recipies locally on our servers. Nov 19 06:05:35 along with the sources Nov 19 06:05:46 I think that Zecke used to do some of that. Nov 19 06:06:11 who is admin on hh.org I wonder Nov 19 06:06:40 if we have someone here on oe community it will be nice to move over thoe repos to somewhere live Nov 19 06:07:37 yep...I think @ 1 time there was some cross pollination...dunno if that is still the case Nov 19 06:07:55 maybe our talking here will refresh someones minds. Nov 19 06:08:17 you mean some repos are mirrored on osuosl Nov 19 06:09:23 Iknow that they had *some* resources there...but I don't think that they are hosted tehre. Nov 19 06:09:36 hmm I see Nov 19 06:10:12 they had an ipaq cluster there. Nov 19 06:20:59 ka6sox: I see that handhelds.org is down Nov 19 06:21:18 yup...let me see who their ISP is. Nov 19 06:22:11 http://osuosl.org/services/hosting/communities/handheldsorg Nov 19 06:23:34 I guess I'll have to ping someone. Nov 19 06:24:34 okay I have pinged on the osuosl folks...someone *may* be around. Nov 19 06:32:41 ka6sox: cool thx Nov 19 06:33:06 gpe is hosted at linuxtogo which is accessible Nov 19 06:33:12 but opie no luck Nov 19 06:34:47 I guess we could create mirrors and freshen them up periodically. Nov 19 06:34:51 as backups. Nov 19 06:37:23 http://gitorious.org/opie Nov 19 06:40:21 I'm able to browse the repo (for opie) let me see if I can clone Nov 19 06:42:52 well git is not the problem Nov 19 06:42:59 issue is for the older tar releases Nov 19 06:43:15 I guess at one point we just have to give up on older versions Nov 19 06:43:19 oh... Nov 19 06:43:22 nm then... Nov 19 06:43:38 I just changed the preferred versions for opie Nov 19 06:43:42 we have git recipes Nov 19 06:43:52 let me see how far it takes me Nov 19 06:44:05 this is for 2010.12? Nov 19 06:44:37 yes Nov 19 06:44:52 actually minimal distro Nov 19 06:45:20 qt-gui for minimal...interesting. Nov 19 06:46:40 khem, shr/nokia900/shr-lite-image (and vala-terminal will be broken) Nov 19 06:48:23 JaMa: how come ? Nov 19 06:48:38 JaMa: any blockers that need attention Nov 19 06:50:28 khem: are you talking about that older gcc linaro patch issue? 01:10:57 < khem> JaMa: I need your help to identify the cairo issue Nov 19 06:51:16 JaMa: yes Nov 19 06:51:33 the older one where we identified the patch to back out Nov 19 06:52:07 but in order to fix the problem we need to reproduce the issue and kind of get to bottom in cairo Nov 19 06:52:21 and thn create a testcase for gcc so it can be fixed properly Nov 19 06:52:30 linaro guys want to fix it Nov 19 06:52:36 we can help with inputs Nov 19 06:57:06 * JaMa already removed those gcc+cairo builds :/ Nov 19 06:57:15 and now gtg to work Nov 19 07:05:41 gm Nov 19 07:13:42 gm Nov 19 07:14:39 khem, did you have an comment/reply/idea on the #if 0 for str* and mem* functions I had to use for uclibc++ in order to get it compile Nov 19 07:58:20 good morning Nov 19 08:54:43 morning Nov 19 08:58:18 anyone an idea what is wrong with handhelds.org or where it is or who is responsible for it? I get: Nov 19 08:58:19 NOTE: Fetch cvs://anoncvs:anoncvs@anoncvs.handhelds.org/cvs;module=apps/update-rc.d;tag=r0_7 cvs [checkout aborted]: connect to anoncvs.handhelds.org(128.31.0.25):2401 failed: Connection timed out Nov 19 08:58:34 seems the thing is down or gone Nov 19 09:04:04 eFfeM_work: once you used your testing script http://wiki.openembedded.net/index.php/TestingScript how do you add the result in the wiki page? Nov 19 09:04:11 eFfeM_work: by hand? Nov 19 09:04:50 yes, I run this under hudson then visually inspect the console output Nov 19 09:05:06 eFfeM_work: ok thanks Nov 19 09:05:26 of course the last line ( 123 tasks run, 0 did not need to be rerun 0 failed is the most important one) Nov 19 09:05:45 but normally I glance over the file to see if there are any oddities (and or grep on error or note) Nov 19 09:05:58 sometimes wish the tools were a little less verbose Nov 19 09:15:03 hm.. I can't update the Main_Page.. someone should add a link to 'testing' in the Developers frame Nov 19 09:15:19 this link: http://wiki.openembedded.net/index.php/Testing Nov 19 09:23:54 or even under Distributions and Devices Nov 19 09:29:48 Hi , Nov 19 09:30:26 In oe build i am getting error while building gst-plugins-bad Nov 19 09:30:43 please help me how i can resolve it Nov 19 09:30:55 here is the log Nov 19 09:31:25 make[3]: Leaving directory `/home/mistral/stuff/openembedded/tmp/work/armv7a-angstrom-linux-gnueabi/gst-plugins-bad-0.10.20-r11.0/gst-plugins-bad-0.10.20/ext/jack' | /home/mistral/stuff/openembedded/tmp/sysroots/armv7a-angstrom-linux-gnueabi/usr/lib/libgstreamer-0.10.so: undefined reference to `register_printf_specifier@GLIBC_2.10' | collect2: ld returned 1 exit status | make[4]: *** [generate_tables] Error 1 | make[4]: *** Waiting for u Nov 19 09:51:47 eFfeM_work: read channel logs about handhelds.org Nov 19 09:52:10 Jay7: any hint on date or period Nov 19 09:52:20 about 5 hrs ago Nov 19 09:52:22 this day Nov 19 09:52:28 ah ok, thanks Nov 19 09:52:56 ka6sox should pings anyone who may know status of hh.org Nov 19 09:53:57 they have some resources on osuosl.org Nov 19 09:55:05 Jay7: read it, thanks, actually did not get to reading the logs yet today, so missed it Nov 19 09:55:24 yw Nov 19 09:57:54 hm.. good, all images for angstrom-2008.1 was built successfully this night Nov 19 09:58:24 even for efikamx Nov 19 09:58:29 and even opie-image Nov 19 10:01:34 atm gpe stuff seems a bit broken Nov 19 10:14:08 ant_work: which gpe bits are broken? Nov 19 10:14:13 good morning bt.w Nov 19 10:14:35 hello Nov 19 10:14:45 x11-gpe-image breaks because of 'checking for CONTACTSDB... configure: error: Package requirements (sqlite libgpepimc) were not met:' Nov 19 10:15:44 gpe-image was breaking because of gpe-package manager iirc (still looking for ipk) Nov 19 10:16:07 first one seems a .pc issue Nov 19 10:16:17 second one need heavier patch Nov 19 10:17:58 yes the first might be a pc issue or just a missing dependency. Nov 19 10:18:04 03Martin Jansa  07master * r3548f2b2cb 10openembedded.git/recipes/opensync/wbxml2_0.10.8.bb: Nov 19 10:18:04 wbxml2: add recipe for 0.10.8 Nov 19 10:18:04 Signed-off-by: Martin Jansa Nov 19 10:18:14 03Marco Trevisan (Treviño)  07master * rde7821a440 10openembedded.git/recipes/gnome/ (2 files in 2 dirs): Nov 19 10:18:14 libsoup-2.4: added TLS 1.2 support patch Nov 19 10:18:14 * Added patch to libsoup-2.4 to make it work with secure connections Nov 19 10:18:14 when using gnutls v2.10 and higher. Nov 19 10:18:14 * Now webkit based browsers works with HTTPS too. Nov 19 10:18:14 Signed-off-by: Martin Jansa Nov 19 10:18:18 03Marco Trevisan (Treviño)  07master * r14e7159af3 10openembedded.git/recipes/e17/ (3 files in 2 dirs): Nov 19 10:18:18 eve: theme updated for small screen devices Nov 19 10:18:18 Signed-off-by: Martin Jansa Nov 19 10:18:22 03Martin Jansa  07master * ra941898ca5 10openembedded.git/recipes/libxml/ (4 files in 2 dirs): Nov 19 10:18:22 libxml2: add recipe for 2.7.8 version Nov 19 10:18:22 Signed-off-by: Martin Jansa Nov 19 10:18:24 03Martin Jansa  07master * r0d09102f17 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Nov 19 10:18:24 EFL: bump SRCREV for fixed keyboard resize Nov 19 10:18:24 Signed-off-by: Martin Jansa Nov 19 10:19:13 florian: I think the issue has been brought out by 'pkgconfig.bbclass: don't blindly install every . pc file' commit Nov 19 10:19:26 can't check right now Nov 19 10:19:42 (image was building until nov. 18.) Nov 19 10:20:21 probably this commit exposed some flaws in the packages Nov 19 10:21:16 ant_work: obviously... I guess libcontactsdb does the right things. This might be worth a discussion on the mailinglist. Nov 19 10:25:16 ant_work: A workaround for the second one might be to remove gpe-package from the task for now. this is one of the things I wanted to get fixed lately but I just lack the time to do so. Nov 19 10:25:35 hm..I found something perhaps Nov 19 10:26:49 florian: pls see http://cgit.openembedded.org/cgit.cgi/openembedded/log/recipes/gpe-package Nov 19 10:27:36 "gpe-package really depends on libipkg, not IPKG_VARIANT. It will not build against a recent opkg" Nov 19 10:27:40 ? Nov 19 10:28:42 Quite likely yes... Nov 19 10:29:40 that's not exactly my bread...I'm led to believe to you and grg ;) Nov 19 10:32:04 morning Nov 19 10:36:47 HI ..while building oe for am3517-evm i am getting error while building building Gstreamer plugins Nov 19 10:36:56 Please help mw .. Nov 19 10:55:18 <_julian_> hi Nov 19 10:55:31 <_julian_> should bitbake work with python 2.7? Nov 19 11:05:12 yes Nov 19 11:13:48 <_julian_> XorA: I get this error: http://pastebin.com/zneTgm9s Nov 19 11:15:16 * XorA doesnt understand python :-( Nov 19 11:18:35 _julian_: which bb version you use? Nov 19 11:18:51 <_julian_> Jay7: 1.10.0 Nov 19 11:19:20 _julian_: try to upgrade to bb master Nov 19 11:19:35 and report it again when kergoth will be alive ;) Nov 19 11:19:51 <_julian_> ok (c: Nov 19 11:21:23 seems I should learn python.. Nov 19 11:21:44 I've reached some limits w/o python knowledge Nov 19 11:33:50 03Martin Jansa  07master * r48df5f0743 10openembedded.git/recipes/shr/ (5 files): Nov 19 11:33:50 shr: bump SRCREVs to versions compatible with latest EFL Nov 19 11:33:50 Signed-off-by: Martin Jansa Nov 19 11:36:12 I guess I can cherry-pick this one to release-2010.12 right? because those recipes are buildable only with shr-autorev (which is included only in shr.conf) Nov 19 11:37:12 but I should also remove shr-autorev.conf from shr.conf (only for release :/) Nov 19 11:38:07 <_julian_> Jay7: should I use trunk or some of the branches? Nov 19 11:38:20 bb is in git now Nov 19 11:38:31 <_julian_> oh, which is the git repository? Nov 19 11:38:46 git://git.openembedded.org/bitbake Nov 19 11:39:15 if you wish you can join to parallel-parsing testing :) Nov 19 11:39:52 <_julian_> Jay7: actually first I want to get something to work at all (c: it's my first test with openembedded Nov 19 11:40:19 ok, then try git master from that repo Nov 19 11:44:42 03Martin Jansa  07master * rc666d01d72 10openembedded.git/conf/distro/shr.conf: Nov 19 11:44:42 shr.conf: remove include for shr-autorev.inc Nov 19 11:44:42 * I'll include it in local.conf generated with SHR Makefile instead Nov 19 11:44:42 Signed-off-by: Martin Jansa Nov 19 11:50:51 _julian_: just checked my Fedora 14 is python 2.7 and running bitbake git fin Nov 19 11:50:58 fine Nov 19 11:51:14 <_julian_> XorA: yes, git seems to be ok. jusst need to fiddle around with all this configuration stuff now Nov 19 11:51:52 * XorA is sitting on the cutting edge again :-) Nov 19 12:12:57 after installing bb and oe i got the following error message when trying to do a 'bitbake nano' Nov 19 12:13:00 http://tiny.cc/bberror Nov 19 12:13:04 insanely long Nov 19 12:13:34 but the recurring element is this: ERROR: Exception: Message:EOL while scanning string literal (, line 1) Nov 19 12:13:41 i wonder what that is Nov 19 12:13:45 any idea? Nov 19 12:13:54 where did i go wrong? Nov 19 12:15:07 alkopop79: no DISTRO defined probably Nov 19 12:15:08 alkopop79: use pastebin please Nov 19 12:15:17 ok Nov 19 12:15:22 asorry Nov 19 12:15:34 google docs is cool but I'm lazy to login Nov 19 12:15:53 then can have no google login at all Nov 19 12:16:05 sorry Nov 19 12:16:06 again Nov 19 12:16:08 http://pastebin.com/c8h6uXpS Nov 19 12:16:19 the distro is ubuntu Nov 19 12:16:37 and i want a binary image for angstrom Nov 19 12:18:47 alkopop79: check DISTRO spelling in your local.conf Nov 19 12:18:47 alkopop79: as XorA said, check your local.conf Nov 19 12:18:55 he Nov 19 12:18:59 ant_work: :) Nov 19 12:19:17 well, afk for some hours Nov 19 12:21:14 kergoth: I've done some runs of parallel bb w/o hangs Nov 19 12:21:44 but should do again on idle system Nov 19 12:24:11 Jay7: finally, which kernel scheduler do you suggest? Nov 19 12:25:10 ok Nov 19 12:25:16 will do! thx! Nov 19 12:26:46 obi: hello Nov 19 12:27:33 hi ant_work Nov 19 12:27:35 obi: the pkgconfig patch didn't break anything, just exposed a flaw Nov 19 12:28:15 did removing do_install work? Nov 19 12:28:27 will test later on Nov 19 12:28:59 ok Nov 19 12:29:01 03Martin Jansa  07master * r50ed401cda 10openembedded.git/recipes/libxml/ (2 files in 2 dirs): Nov 19 12:29:01 libxml2-2.7.8: shared library versionning was broken in 2.7.8 release, add another patch from git to fix it (instead of rebuilding all against non-versioned symbols Nov 19 12:29:01 Signed-off-by: Martin Jansa Nov 19 12:29:02 03Martin Jansa  07master * r4373bc4d0c 10openembedded.git/recipes/libxslt/libxslt.inc: Nov 19 12:29:02 libxslt: bump INC_PR to rebuild against newer libxml2 in case someone was fast enough to rebuild it against libxml2 without versioned symbols Nov 19 12:29:02 Signed-off-by: Martin Jansa Nov 19 12:29:04 03Martin Jansa  07master * rd549667bc0 10openembedded.git/recipes/openmoko-3rdparty/mokosuite2_git.bb: Nov 19 12:29:04 mokosuite2: bump SRCREV to pick patches from Marco Trevisan Nov 19 12:29:04 Signed-off-by: Martin Jansa Nov 19 12:29:05 03Martin Jansa  07master * r7e673ab79c 10openembedded.git/recipes/gnome/libsoup-2.4_2.29.91.bb: Nov 19 12:29:05 libsoup-2.4_2.29.91: bump PR in case someone was fast enough to rebuild it against libxml2 without versioned symbols Nov 19 12:29:05 Signed-off-by: Martin Jansa Nov 19 12:29:34 bad libxml2 bad! spank it Nov 19 12:29:56 i suspect that there are some more packages which have the same problem, but fixing them is easy and overall recipe quality will be improved a little ;) Nov 19 12:31:39 library version is a very subtle topic in GNU, it's very complex Nov 19 12:37:39 jay7, i checked the distro spelling and it seems all right Nov 19 12:37:49 on the other hand it was commented Nov 19 12:38:28 uncommented and got a different error  Nov 19 12:39:29 unable to parse conf/bitbake.conf (/stuff/build/conf/local.conf:36 unparsed line: ' " ' Nov 19 12:39:39 what is that abvout? Nov 19 12:39:42 about Nov 19 12:52:19 argh, spelling Nov 19 12:52:26 ericben|away, sorry I forgott to activate the serial ports with cpld_uart_active(4); Nov 19 12:52:28 working on that Nov 19 13:06:21 ericben|away, hmmm still doesn't work, I'll ask in #kernelnewbies(not on freenode) Nov 19 13:37:09 03Koen Kooi  07org.openembedded.dev * rd990744d83 10openembedded.git/recipes/linux/ (11 files in 4 dirs): Nov 19 13:37:09 linux-omap 2.6.36rc: move to 2.6.37rc Nov 19 13:37:09 Signed-off-by: Koen Kooi Nov 19 13:50:15 hrm Nov 19 13:54:19 * Tartarus confirms updating openssl.inc for linux-uclibcspe now Nov 19 14:08:42 03Koen Kooi  07org.openembedded.dev * rb9b8fdd4c5 10openembedded.git/recipes/linux/linux-omap-2.6.37rc/beagleboard/defconfig: Nov 19 14:08:42 linux-omap 2.6.37rc: tweak beagleboard defconfig Nov 19 14:08:42 Signed-off-by: Koen Kooi Nov 19 14:21:26 * kergoth debugs issue with backgrounding the cache write to disk Nov 19 15:18:57 is patchwork down? I get The server at patchwork.openembedded.org is taking too long to respond. Nov 19 15:20:07 it's taking its time here too Nov 19 15:20:12 * kergoth_ waits and sees if it loads Nov 19 15:20:15 nope Nov 19 15:20:17 down here too Nov 19 15:21:04 any more volunteers to test parallel parsing? Nov 19 15:21:16 would like to fully pound the thing before it goes anywhere near master Nov 19 15:21:45 kergoth_: I've been playing with it a bit here as time permits; no problems thus far. Nov 19 15:21:51 cool Nov 19 15:22:26 03Tom Rini  07org.openembedded.dev * r9aa0531ed3 10openembedded.git/recipes/lmbench/ (lmbench-2.0.4/gnu-os-rundir.patch lmbench_2.0.4.bb): Nov 19 15:22:26 lmbench 2.0.4: Add a patch to fix a problem in 'gnu-os' Nov 19 15:22:26 Signed-off-by: Tom Rini Nov 19 15:22:30 03Tom Rini  07org.openembedded.dev * r9692544e16 10openembedded.git/recipes/openssl/openssl.inc: Nov 19 15:22:30 openssl.inc: Update powerpc logic Nov 19 15:22:30 We 'spe' needs similar tests to 'eabi'. Nov 19 15:22:30 Signed-off-by: Tom Rini Nov 19 15:22:31 foerster: thanks Nov 19 15:22:47 kergoth_: np Nov 19 15:23:08 hmm Nov 19 15:23:08 I've just recently started diving head first into OE and bitbake to develop a product, so I'm having fun learning all of it. Nov 19 15:23:19 nice, welcome Nov 19 15:24:12 foerster: welcome aboard! Nov 19 15:24:30 eFfeM_work: thanks Nov 19 15:25:14 I've been thoroughly pleased with OE. Compared to the homegrown build systems and a previous stint using ltib, OE is great! Nov 19 15:25:28 glad you like it Nov 19 15:29:49 still struggling with oe, the latest error message is http://pastebin.com/w6PH5xP0 Nov 19 15:30:10 strange since both nano and help2man seems working fine Nov 19 15:30:17 dunno about perl Nov 19 15:30:22 huh Nov 19 15:30:27 poke Tartarus :) Nov 19 15:32:00 thx! Nov 19 15:32:14 * Tartarus opens Nov 19 15:32:26 the hell? Nov 19 15:33:16 alkopop79: What distro/machine? Nov 19 15:33:23 and what version of bitbake Nov 19 15:33:25 i should think about suppressing log messages during the parsing progress display Nov 19 15:33:33 and showing them after the parsing is done, or upon exception Nov 19 15:33:53 also uniq()ing them, so we see each angstrom message *once* only Nov 19 15:34:09 hmmm Nov 19 15:35:17 oe on ubuntu Nov 19 15:35:26 for angstrom (beagleboard) Nov 19 15:37:42 dunno how to check the bitbake version Nov 19 15:38:00 weekend ... Nov 19 15:38:36 alkopop79: bitbake --version Nov 19 15:39:01 thx! bitake 1.8.18 Nov 19 15:42:02 i'm not quite sure what to do now Nov 19 15:42:36 I'm debugging now :) Nov 19 15:50:27 i hope RP is going to do the bitbake sync with poky, there's an awful lot of commits here :) Nov 19 15:50:34 * kergoth_ is reviewing poky commit logs for bitbake changes Nov 19 15:52:40 I wonder what's become of RP. Haven't heard from him for some time. Nov 19 15:52:49 probably busy with yocto Nov 19 15:53:07 So... Nov 19 15:53:17 bitbake 1.8.18 is supposed to work, yes? Nov 19 15:53:32 I grabbed the tag from bitbake git rather than tarball Nov 19 15:53:45 and even after fixing my help2man introduced problem, it's dying in a loop here Nov 19 15:54:17 alkopop79: Can you remove the perl-native-runtime line from recipes/help2man/help2man_1.37.1.bb and try? Nov 19 15:54:23 should, as far as i know anyway Nov 19 15:54:27 i haven't tried it in a while though Nov 19 15:54:44 wil try, thx! Nov 19 15:58:15 Tartarus: I feel safer with bitbake 1.10.1 Nov 19 15:59:16 tried Nov 19 15:59:22 nothing changes Nov 19 15:59:25 mckoan: I do too, but if release-2010.12 is going to support it, it should work Nov 19 15:59:26 still the same error Nov 19 15:59:35 alkopop79: Same exact? Or just an error still? Nov 19 15:59:41 the very same Nov 19 15:59:48 * kergoth_ thinks its been months since he's used 1.8 Nov 19 15:59:58 and you removed DEPENDS_virtclass-native = "perl-native-runtime" Nov 19 15:59:59 ? Nov 19 16:00:20 * kergoth_ creates a 1.8 test setup Nov 19 16:00:50 oh, wait, i created one already, i just haven'jt used it :) Nov 19 16:00:55 it might be the wrong file Nov 19 16:01:02 since it's different Nov 19 16:01:24 f*kin virtual machine does not allow me to copypaste Nov 19 16:01:30 just a sec Nov 19 16:01:50 the file is help2man_1.36.4.bb Nov 19 16:02:09 1.37.1 Nov 19 16:02:13 is the one with the bad line Nov 19 16:02:26 DEPENDS = "gettext-native perl-native liblocale-gettext-perl-native" Nov 19 16:02:33 oke Nov 19 16:02:36 check it Nov 19 16:03:40 no 1.37 Nov 19 16:03:48 only 1.36.4 Nov 19 16:05:23 what branch are you using? Nov 19 16:05:29 do a git branch inside the openembedded dir Nov 19 16:08:26 so i did a git branch in that directory Nov 19 16:08:51 * kergoth_ tests angstrom-2010.1/qemux86/micro-image w/ 1.8 Nov 19 16:08:53 and got "* org.openembedded.dev" Nov 19 16:09:12 and yep, there's alkopop79's perl-native-runtime-native error, that was quick Nov 19 16:09:37 Yeah Nov 19 16:09:40 And that's an easy fix Nov 19 16:09:47 is it? Nov 19 16:09:54 wtf is up with perl-native-runtime anyway Nov 19 16:10:03 its in ASSUME_PROVIDED by default, yet RDEPENDS_automake-native includes it Nov 19 16:10:04 kergoth_: it's a not real package Nov 19 16:10:13 how exactly is that package going to be useful without its dependency being built? Nov 19 16:10:13 Just as a way to say "yeah, circular dep hell" Nov 19 16:10:24 ah Nov 19 16:10:25 I shouldn't have bothered to list it Nov 19 16:10:28 But.. Nov 19 16:10:34 03Graeme Gregory  07org.openembedded.dev * re5df86c282 10openembedded.git/conf/machine/ (omapzoom2.conf omapzoom36x.conf): omapzoom2/3.conf : got serial tty wrong, fixed this time Nov 19 16:10:35 Remove the offending line, try again please kergoth Nov 19 16:10:42 it's only mentioned in 3 places in oe Nov 19 16:10:44 03Graeme Gregory  07org.openembedded.dev * r5fa94e3018 10openembedded.git/recipes/gnome/at-spi_1.32.0.bb: at-spi_1.32.0.bb : add new version that doesnt error with odd errors in Makefile Nov 19 16:10:46 03Graeme Gregory  07org.openembedded.dev * r32dcd4f48e 10openembedded.git/recipes/sysvinit/sysvinit_2.86.bb: sysvinit_2.86.bb : upped PR to catch zoom2/3 changes Nov 19 16:10:47 help2man, automake, and bitbake.conf Nov 19 16:10:47 03Graeme Gregory  07org.openembedded.dev * r7205d3b5e9 10openembedded.git/recipes/u-boot/ (4 files in 3 dirs): u-boot_git.bb : update to latest for zoom2/zoom3 machines Nov 19 16:11:15 bit confused Nov 19 16:13:30 alkopop79: You must have a recipes/help2man/help2man_1.37.1.bb file Nov 19 16:14:40 i only have 1.36.4.bb Nov 19 16:15:52 i have it opened Nov 19 16:18:14 that's the file's content: http://pastebin.com/Vjdphv3r Nov 19 16:19:18 Tartarus: is there an explanation of perl-native-runtime somwhere? Nov 19 16:19:41 kergoth_ no, not that I know of Nov 19 16:19:54 gah, trying to do too much at once, losing track of it Nov 19 16:20:03 kergoth_: So if you drop that line do you see a diff circular dep? Nov 19 16:20:06 Or am I insane/ Nov 19 16:21:48 linux rabbithole Nov 19 16:23:45 * kergoth_ checks Nov 19 16:24:27 Tartarus: yikes Nov 19 16:24:39 Tartarus: its not just a circular one, its one that *dies* trying and failing to find teh chain Nov 19 16:24:45 find_chains(revdep, deepcopy(prev_chain)) Nov 19 16:24:45 File "/home/kergoth/Code/oe/projects/1.8-test/bitbake/lib/bb/runqueue.py", line 248, in find_chains Nov 19 16:24:48 for dep in explored_deps[revdep]: Nov 19 16:24:50 KeyError: 474 Nov 19 16:25:00 * kergoth_ checks master for fixes to this area Nov 19 16:25:26 oh hell Nov 19 16:25:38 i forgot just how much crap went into 1.10 from 1.8 ;) Nov 19 16:26:30 So I'm not crazy, great Nov 19 16:26:35 kergoth_, well, wait Nov 19 16:26:38 Lets post on the ML Nov 19 16:26:41 no i am Nov 19 16:26:53 The whole "lets require 1.8.x still" thing seems kinda silly Nov 19 16:27:09 Lemme do a quick bisect and see how long it's been broken Nov 19 16:27:29 well, i was part of that tsc discussion :) some of us felt maintaining support for both shouldn't be terribly difficult, and we've basically been doing so anyway for months Nov 19 16:27:54 Have we? Nov 19 16:28:03 I wonder how long 1.8 hasn't been able to parse the metadata Nov 19 16:28:19 any suggestions how to fix this? should I downgrade bitbake? Nov 19 16:28:42 Tartarus: we thought we were, but *shrug* Nov 19 16:28:43 alkopop79: You can upgrade bitbake to 1.10.x Nov 19 16:28:49 i think a bisect of the metadata is in order Nov 19 16:28:49 kergoth_, exactly :) Nov 19 16:29:06 yey! is that cd to directory and than git pull? Nov 19 16:29:28 alkopop79: Did you get bitbake from git? Nov 19 16:29:30 i love you, git Nov 19 16:29:38 it is a separate project Nov 19 16:29:56 ok Nov 19 16:29:59 so... Nov 19 16:30:29 http://download.berlios.de/bitbake/bitbake-1.10.1.tar.gz Nov 19 16:30:48 rafa: do you think the weirdsharkness is an upstream problem ? i.e., OE vs. jlime ? Nov 19 16:31:04 03Bernhard Reutner-Fischer  07master * r823a02185e 10bitbake.git/lib/bb/fetch/ (__init__.py bzr.py hg.py osc.py perforce.py svn.py): Nov 19 16:31:04 fetch: add common helper _strip_leading_slashes() Nov 19 16:31:04 Several fetcher need a way to strip leading slashes off a local path. Nov 19 16:31:04 This helper-function consolidates all such occurances. Nov 19 16:31:04 Signed-off-by: Bernhard Reutner-Fischer Nov 19 16:31:07 damnit, whats the syntax for that again.. Nov 19 16:31:08 hmm Nov 19 16:31:14 03Bernhard Reutner-Fischer  07master * re88834fb7c 10bitbake.git/lib/bb/fetch/ (9 files): Nov 19 16:31:15 fetch: be more pythonic Nov 19 16:31:15 no functional changes Nov 19 16:31:15 Signed-off-by: Bernhard Reutner-Fischer Nov 19 16:31:16 03Bernhard Reutner-Fischer  07master * rc360b01df1 10bitbake.git/lib/bb/fetch/__init__.py: Nov 19 16:31:16 fetch: use os.path.join Nov 19 16:31:16 Signed-off-by: Bernhard Reutner-Fischer Nov 19 16:31:17 03Bernhard Reutner-Fischer  07master * rb8bb4433de 10bitbake.git/lib/bb/fetch/ (__init__.py bzr.py git.py hg.py repo.py svn.py): Nov 19 16:31:17 fetch: rename suppports_srcrev to supports_srcrev Nov 19 16:31:17 osc had it already spelled correctly?! Nov 19 16:31:18 Signed-off-by: Bernhard Reutner-Fischer Nov 19 16:31:18 03Bernhard Reutner-Fischer  07master * r8da9744fcd 10bitbake.git/lib/bb/fetch/svn.py: Nov 19 16:31:19 svn fetcher: warn people to switch to SRCREV Nov 19 16:31:19 as noted by rp in ac00ca89a4e43cd4f38ba86455079d31be78e644 Nov 19 16:31:20 kergoth_, ok it is recent Nov 19 16:31:23 Signed-off-by: Bernhard Reutner-Fischer Nov 19 16:32:32 alkopop79: if you got bitbake from git, just cd bitbake; git checkout v1.10.1, or the 1.10 branch. otherwise the tarball tom directed you to Nov 19 16:32:39 bitbake is, as he said, a separate project Nov 19 16:32:52 Tartarus: is it? i was busy trying to remember the syntax to tell checkout to direclty check out the tree as of a month ago Nov 19 16:33:07 i know you can tell it to check out the *ref* as of a month ago, but that uses the reflog, not the regular commit log Nov 19 16:33:09 i tried but did not work Nov 19 16:33:11 kergoth_: is your 'parallel parsing' branch rebased on master? Nov 19 16:33:18 kergoth_, if i find the time i'll look into moving oe.path to bb during the weekend Nov 19 16:33:18 ant_work: yes, its based on master Nov 19 16:33:33 ok, then is missing the latest commits Nov 19 16:33:36 yep Nov 19 16:33:40 * Tartarus tries git bisect run aain Nov 19 16:33:51 kergoth_, thank you for reviewing these nitpick bb patchlets! Nov 19 16:33:53 kergoth_, I just paged down the log :( Nov 19 16:33:53 * blindvt` & Nov 19 16:33:57 blindvt`: np, thanks for working on it Nov 19 16:34:10 wpwrak: no idea.. maybe I could check the jlime part.. but because i am not an oe expert i could not say about upstream much Nov 19 16:34:11 Tartarus: hehe. i just did git log --since="1 month ago" and hit Nov 19 16:34:25 guess i could have done git rev-list --since="1 month ago" --reverse|head -n 1 or something Nov 19 16:34:27 wpwrak: some people here could to know I guess Nov 19 16:36:11 kergoth_, heh, I shoulda started this somewhere other than the laptop :) Nov 19 16:36:18 noob question Nov 19 16:36:25 hah, thats what i was just thinking, i just started a bisect too for the hell of it Nov 19 16:36:25 once the tarball is extracted Nov 19 16:36:44 how should I configure it? when i do ./configure Nov 19 16:36:50 you don't need to. Nov 19 16:36:55 you can run btibake right out of its source tree Nov 19 16:36:57 blindvt, until khem wakes up, do you have seen the uclibc issue wrt log2f? Nov 19 16:36:58 alright ... in the replaces/conflicts of wireshark of OE-derived jlime, i found things like "Conflicts: tshark, wireshark, (<1.0.5)". is that a valid entry ? Nov 19 16:36:58 it says no such file directory Nov 19 16:37:03 PATH=$PWD/bitbake-1.10.1/bin:$PATH Nov 19 16:37:28 wpwrak: that's not valid, something split the RCONFLICTS wrong Nov 19 16:37:31 how can i run it from the source tree? Nov 19 16:37:36 would be my guess Nov 19 16:37:38 alkopop79: i just told you Nov 19 16:37:44 add it to your path the way i just said, and run bitbake Nov 19 16:38:06 or adjust whatever script you source that does it for you to add that instead Nov 19 16:38:32 kergoth_: thanks ! what to do about it ? (i have a work-around as far as what i'm doing is concerned) Nov 19 16:38:38 kergoth_: was playing around with your parallel-parsing branch. This is trivial, but shaves ~1/2 second (on my setup) during parsing: https://github.com/foerster/bitbake/commit/0a6429d7588c6fe7d1491e1a3497629de227124b Nov 19 16:39:02 kergot_: minimal impact, but speed improvement nonetheless... Nov 19 16:39:35 blindvt`: any low hanging patch for the uclibc-git / log2f issue? Nov 19 16:39:49 thx kergoth, updated to 1.10 Nov 19 16:40:01 foerster: haha, i have a similar one locally :) Nov 19 16:40:13 i tested mod 5, 10, /100, /200, a few other variations Nov 19 16:40:54 the only downside, before i switched to the new progressbar, was it made it "feel" slower Nov 19 16:40:55 gotcha. This still gives nice smooth updates, without sending 8000 events ;) Nov 19 16:41:05 due to the less granular updating of the parsed recipe count Nov 19 16:41:11 but with the new progress bar, it probably isn't noticable Nov 19 16:41:14 funny, my old friend "Ansgtrom does not support bluez-lib becasue bluex 3.x has been replaced by bluez4" Nov 19 16:41:25 my i add, i installed bluez4 Nov 19 16:41:33 this is insane :D Nov 19 16:41:42 installed? Nov 19 16:41:46 foerster: i'll go ahead and try adding yours and test it locally real quick, thanks for looking at it Nov 19 16:42:08 ahh, that's just a note, not an error Nov 19 16:42:08 foerster: hell, thanks for knowing enough to implement that change, between me and RP, and now blindvt`, we're the only ones that touch bitbake really Nov 19 16:42:10 another question: what does a 1: or 2: at the beginning of a package version mean ? i see that such versions are sometimes referred to with the full version, sometimes with just the part after the n: Nov 19 16:42:11 hehe Nov 19 16:42:33 kergoth_: cool, np. I haven't done any python in about 4 years, so it's fun looking around. Nov 19 16:42:38 would appreciate any comments on the changes to the code, if there's anything you think could be done better Nov 19 16:42:42 * kergoth_ nods, fair enough Nov 19 16:43:32 foerster: hey, try sending me a pull request for that. i've never used that in github, i want to try it. hehe Nov 19 16:43:42 i think it's fine, no errors so far...still woking though Nov 19 16:43:45 thx guys Nov 19 16:43:47 wpwrak: the package EPOCH Nov 19 16:43:51 kergoth_: me either, let me see how ;) Nov 19 16:45:09 d'oh Nov 19 16:45:09 ooh, https://github.com/kergoth/bitbake/network is fancy Nov 19 16:45:13 * kergoth_ hugs github Nov 19 16:45:33 damnit, sent pull reguest on master Nov 19 16:45:49 oh, interesting Nov 19 16:46:03 it sends a group of commits, i see Nov 19 16:46:03 XorA: oh .. so what's the rule for comparing versions with epochs ? if both version strings have one, include it in the comparison, else ignore it and hope for the best ? Nov 19 16:46:09 never used github, today was first day ;) Nov 19 16:46:17 oh, it went against my master instead of my parallel-parsing, right Nov 19 16:46:25 still, thats pretty cool Nov 19 16:46:32 yea, i have no clue how to use it yet ;) Nov 19 16:46:34 i like that i can comment on it right there in my pull request view Nov 19 16:46:44 wpwrak: ?: always wins over without ?: and the higher ? always wins Nov 19 16:46:56 closed that one :) Nov 19 16:47:03 fixed Nov 19 16:47:09 wpwrak: its a way to reset messed up versioning, line when people call kernels 2.6.XX-rc1 Nov 19 16:47:20 like Nov 19 16:47:41 wpwrak: if no epoch, treat it like "0:" Nov 19 16:47:41 ah, nice, in the Files Changed tab of the pull request - thats the full diff Nov 19 16:48:12 very nice Nov 19 16:49:48 hanrod Nov 19 16:49:58 hanrotd Nov 19 16:50:22 XorA, pb_: thanks ! i see a libxfixes3, version 1:4.0.4-r0. many things depend on libxfixes3 (>= 4.0.4). so that's broken ? (e.g., matchbox-wm is one of them with such a dependency) Nov 19 16:50:55 wpwrak: no, 1:4.0.4-r0 >> 0:4.0.4 Nov 19 16:51:00 so the dependency is ok Nov 19 16:51:10 kergoth_, ok I can see where the bisect is going, some other part of the change is killing it Nov 19 16:51:52 pb_: aah yes, that works. thanks ! Nov 19 16:55:46 Jay7, foerster: do me a favor and check from time to time to see if we have any rogue bitbake processes or zombies after your build completes or you ^C it -- i've seen it happen, but i'm not sure if its something that was already fixed -- thought it was. still, keep an eye out in case, particularly if you happen to notice it parsing slower than it should Nov 19 16:55:51 heh Nov 19 16:56:59 bitbake is running for a while Nov 19 16:57:08 thank you very much guys! Nov 19 16:57:12 nice Nov 19 16:57:22 brilliant! thx! Nov 19 16:57:57 foerster: fyi, github tip, something i didn't spot for a while. switch to master, then on the source tab go to the branch list, next to switch tags, and click the compare button on a branch, and it'll show a nice list of commits between the two branches, with a diff Nov 19 16:58:08 quite nice for asking for general comments on a set of changes Nov 19 16:58:33 * kergoth_ thinks its quite silly that its buried under Source, not under Commits, but whatever.. :) Nov 19 16:59:03 kergoth_: thanks. This is the first day I've actually used github for anything (other than browsing other people's stuff). Nov 19 16:59:10 * kergoth_ nods Nov 19 16:59:43 i mainly use it for my personal branches on oe/bitbake (as personal branches on our git server don't allow deletion/rebase), and for my dotfiles and vim config Nov 19 17:09:42 Tartarus: get the commit isolated? Nov 19 17:09:54 Yeah, see above :( Nov 19 17:10:00 I think I have an idea at least Nov 19 17:10:09 Given the error message from the other part Nov 19 17:10:47 some other part of which change? :) Nov 19 17:13:12 my help2man one Nov 19 17:13:30 ahh Nov 19 17:13:31 weird Nov 19 17:15:31 hi valhalla Nov 19 17:16:26 It's the autotools change Nov 19 17:16:59 * kergoth_ ponders Nov 19 17:38:09 does anyone else think this is really silly? -b apparently takes a regular expression which it tries to match against BBFILES Nov 19 17:38:28 -b builds a *file* you specify, skipping cross recipe deps, why should it give a crap about bbfiles? Nov 19 17:38:37 also, it loads the cache, even though it doesn't need it, or use it Nov 19 17:42:58 hmmm Nov 19 17:47:29 damn, something is fork bombing clean builds of an image that uses gnome Nov 19 17:51:45 kergoth_, ok, i have no idea why this is blowing up 1.8, help :) Nov 19 17:51:50 it's the autotools.bbclass part Nov 19 17:52:04 in addition to the odd way perl-native-runtime is evaluated Nov 19 17:52:10 (or rather, mangled) Nov 19 18:09:50 why does buildFile claim to not be using the cache, then initializes the cache, loading the cache data from disk? Nov 19 18:09:54 what the hell Nov 19 18:12:24 Tartarus: if you comment out DEPENDS_virtclass-native = "perl-native-runtime", it falls back to the full DEPENDS the target recipe was using, giant loop, boom. needs DEPENDS_virtclass-native = "", i'm sure Nov 19 18:12:30 * kergoth_ hasn't tested that yet, but.. Nov 19 18:12:46 also, RDEPENDS_pn-help2man should be RDEPENDS_${PN}_pn-help2man, otherwise that applies to -dev/-dbg too Nov 19 18:17:11 as far as i can tell, bitbake -b is initializing the cache, loading it from disk *twice* when it claims to not use the cache at all. -e with -b does the same Nov 19 18:20:01 hrm Nov 19 18:23:25 if i change this, it would lose the application of bbappends to the -b file -- this isn't necessarily surprising, but i'll see if i can fix these other issues without breaking that Nov 19 18:23:28 hmm Nov 19 18:31:44 gm all Nov 19 18:32:11 * khem is miffed with fuss over release Nov 19 18:32:30 * kergoth_ doesn't see what the big deal is Nov 19 18:32:44 now koen is not happy with branch Nov 19 18:33:05 he was not happy when I asked to lift foot of pedal for master commits Nov 19 18:33:26 I wonder if people under estimate the efforts to make a release Nov 19 18:34:09 don't worry Nov 19 18:34:14 no plan ever survives Nov 19 18:34:23 a flaky release will do no good to us. It will be a wasteful thing to do and users will lose trust on very onset Nov 19 18:34:45 but, if the release is in a branch, people will piss and moan about it Nov 19 18:34:50 Crofton: heh Nov 19 18:35:04 see comments that frequently appear here about stable Nov 19 18:35:21 Crofton: frankly if there are enough users I will volunteer to maintain it till next release Nov 19 18:35:37 I stress on expanding userbase of OE Nov 19 18:35:40 sure, but not wearing your OE hat :) Nov 19 18:36:02 yes ofcourse Nov 19 18:36:22 just remove the branch after we tag the release, users won't have anything unusual to piss and moan about Nov 19 18:36:26 completely different from stable Nov 19 18:36:29 sure Nov 19 18:36:32 6months release cycle is not that bad Nov 19 18:36:51 the problem is .us west coasts emails and goes to bed Nov 19 18:37:05 and .eu wakes up and does understand the intent Nov 19 18:37:14 kergoth_: How would we relate that release to git Nov 19 18:37:30 can we create a tag on a branch and delete the branch itself Nov 19 18:37:48 of course. Nov 19 18:37:49 git doesn't care Nov 19 18:38:04 Crofton: well pb is in .eu Nov 19 18:38:09 the hostory will survive, right? Nov 19 18:38:15 if you compare master to the tag, you'll see that the tag has commits that aren't on master, but who cares? Nov 19 18:38:22 you can always just merge tag to master at release, too Nov 19 18:38:29 in which case git will show no differences Nov 19 18:38:43 that is what I thought would happen Nov 19 18:38:56 * kergoth_ nods Nov 19 18:39:07 create branch, cherry pick to branch, tag branch, merge branch to master, delete branch Nov 19 18:39:11 pretty straightforward Nov 19 18:39:46 kergoth_: tag will move to master on merge ? Nov 19 18:39:56 no Nov 19 18:40:08 nor should it, because master will contain stuff that isn't in the release Nov 19 18:40:16 kergoth_: so can people clone and checkout the tag and they will get the release Nov 19 18:40:21 right Nov 19 18:40:22 yes Nov 19 18:40:30 that's the case whether you merge to master or not Nov 19 18:40:48 kergoth_: ok so it will be tag on anonymous branch Nov 19 18:40:50 the merge just makes it so git log will show nothing in the release that isn't also in master Nov 19 18:40:54 no Nov 19 18:40:56 if we delete the branch Nov 19 18:40:58 tags aren't of branches, period Nov 19 18:41:02 tags mark *commits* Nov 19 18:41:11 when you tag a branch, you're tagging the comimt the branch points to right now Nov 19 18:41:19 kergoth_: I am trying to imaging tree :) Nov 19 18:41:54 I hate sound Nov 19 18:42:02 kergoth_: ok so its in the myriad Nov 19 18:42:34 kergoth_: ok. sounds that I get it now Nov 19 18:42:57 so we will bake it on this branch, then merge the branch back into master Nov 19 18:43:02 at release Nov 19 18:43:17 and create a release tag before merging Nov 19 18:43:19 yes, but *tag* the branch, not master after merging the branch Nov 19 18:43:20 right Nov 19 18:43:35 i think you've got it :) Nov 19 18:44:06 right branch out -> chery-pick -> tag releage on this branch -> merge into master -> announce the release Nov 19 18:44:26 * khem feels better Nov 19 18:44:57 Crofton: are you working over weekend Nov 19 18:45:10 did u get over that libtool issue Nov 19 18:45:19 most likely working some Nov 19 18:45:39 kergoth: thanks for hte fix, pushing soon Nov 19 18:45:46 khem, will bug about 2 c-p's for the release Nov 19 18:45:48 in 5 min Nov 19 18:46:06 Tartarus: as of now we need quite a few things Nov 19 18:46:23 so this week meaning upto sunday we could get stuff in easily Nov 19 18:46:30 next week we should be more careful Nov 19 18:46:30 Tartarus: np Nov 19 18:46:47 kergoth_: so how is git commits organised internally Nov 19 18:46:53 sakoman_, ping Nov 19 18:46:57 khem: read Pro Git Nov 19 18:46:58 ;) Nov 19 18:46:58 is is a sort of unsorted tree Nov 19 18:47:09 Crofton: pong Nov 19 18:47:17 a commit is just an object that holds commit message and other metadata, plus a pointer to a git tree object, plus pointer(s) to its parent commits Nov 19 18:47:21 that it Nov 19 18:47:23 its a digraph Nov 19 18:47:27 do you have local commits to make hte gnome sound control work? Nov 19 18:48:48 Crofton: the GNOME volume control has been deprecated in 2.30 IIRC Nov 19 18:49:36 kergoth_: ah Nov 19 18:49:48 kergoth_: its like compiler;s IR Nov 19 18:49:54 hmm Nov 19 18:50:06 what do you use to control the mixer? Nov 19 18:50:35 now, a non-annotated tag *is* a ref, identical to a branch. a branch is just a ref, a simple pointer to a commit (or, in the case of a symbolic ref, a pointer to another ref) Nov 19 18:50:46 Crofton: lately I've just been using the gnome-mplayer volume control Nov 19 18:50:51 an annotated tag is just that, an object that points the same way a ref does, + metadata Nov 19 18:50:55 kergoth_: and branch is a moving tag ? Nov 19 18:51:04 I am trying to check the input hardware Nov 19 18:51:04 so these are just pointers to points in time in the digraph Nov 19 18:51:25 Crofton: you can run alsamixer from a terminal window Nov 19 18:51:30 technically tehre's no difference between an unannotated tag and a branch, other than its path in the repository Nov 19 18:51:39 yeah Nov 19 18:51:39 you can *technically* rewrite a tag, but you're not supposed to ;) Nov 19 18:51:43 doing that Nov 19 18:51:47 kergoth_: right Nov 19 18:51:59 note: refs/tags/*, refs/heads/* -- both refs Nov 19 18:51:59 of course the stupid kb confused wrt to the function keys Nov 19 18:52:27 to be even more precise, when you have an annotated tag, what you have is a refs/tag/ ref pointing at the annotated tag object Nov 19 18:52:32 rather than pointing at a commit object Nov 19 18:52:35 its just a slight indirection Nov 19 18:52:42 khem, I am running a build with the pkdconfig.class commit reverted Nov 19 18:52:53 to see if that causes me to run out of processes Nov 19 18:53:01 Crofton: hmm ok Nov 19 18:53:26 Crofton: could be but I doubt. but all bets are open Nov 19 18:53:31 yeah Nov 19 18:53:56 we need more people doing clean builds Nov 19 18:54:24 03Tom Rini  07org.openembedded.dev * r5cfa19be82 10openembedded.git/recipes/help2man/help2man_1.37.1.bb: Nov 19 18:54:24 help2man: Fix DEPENDS on virtclass-native Nov 19 18:54:24 Signed-off-by: Tom Rini Nov 19 18:54:35 khem: 5cfa19be826f67b153481bba74cec8f5c4537287 && 9692544e1683a96555e5fcea4bbf2b9052c57a68 for the release branch? Nov 19 18:54:37 Crofton: I have done 3 clean build before comming that Nov 19 18:54:38 sakoman_, got any suggestions for anything that is a quick sound input test? Nov 19 18:54:53 but I might have built images whch dont matter to u Nov 19 18:55:13 Crofton: play a wave file :) Nov 19 18:55:15 Crofton: heh, I just connect my iPod and capture the output Nov 19 18:55:20 input Nov 19 18:55:29 yeah, that is where I am headed again Nov 19 18:55:40 it is the simplest way for me Nov 19 18:56:36 Tartarus: help2man Nov 19 18:56:39 is fine Nov 19 18:56:44 let me look at the other one Nov 19 18:56:46 k Nov 19 18:56:53 I tested on both uclibc and eglibc Nov 19 18:57:30 there we go, now bitbake -e -b takes ~1sec, thats an improvement Nov 19 18:57:39 now to make sure i didn't break regular -e Nov 19 19:01:07 kergoth_: btw. I have been using parallel-parsing branch without any events since yesterday Nov 19 19:01:12 it doubled parsing speed Nov 19 19:01:17 on my poor laptop Nov 19 19:01:24 nice Nov 19 19:01:26 which has core2duo Nov 19 19:01:34 its spoiling me, i don't want to go back Nov 19 19:01:36 and I build 6 images from scratch for different machines Nov 19 19:01:40 all went fine Nov 19 19:01:42 sweet Nov 19 19:01:44 glad to hear it Nov 19 19:01:46 wel some of them failed Nov 19 19:01:50 but that was not bitbake Nov 19 19:01:53 hehe Nov 19 19:02:07 i want to do a bit more poking at this, then will have to see if we can get it into master sometime soon Nov 19 19:02:26 * kergoth_ pushes -b speed improvement Nov 19 19:03:19 cool. Nov 19 19:03:44 kergoth_: I will look at parsing branch today Nov 19 19:12:15 arg, damnit Nov 19 19:12:16 hrm Nov 19 19:13:33 sakoman_, alsamixer is giving me no level sliders for inputs Nov 19 19:14:03 khem, your builds include gnome Nov 19 19:15:16 * Crofton cries over the state of linux audio Nov 19 19:18:49 Crofton: it may be off the right side of the screen Nov 19 19:19:38 I just have TX1 and TX2 Nov 19 19:19:45 which I can change, but no sliders Nov 19 19:20:28 kergoth_: 50 bitbake -p w/o hangs Nov 19 19:20:50 Crofton: no, I think its x11-image at best Nov 19 19:20:59 Crofton: have you used tab to select Capture devices? Nov 19 19:21:33 yeah Nov 19 19:21:40 I type alsamixer, then when it launches hit tab Nov 19 19:21:43 khem, I strongly suspect it is in gnome Nov 19 19:21:51 sakoman_, yep Nov 19 19:22:04 pulseaudio? Nov 19 19:22:15 no pulse at this run level Nov 19 19:23:04 parsing for angstrom is looking better Nov 19 19:23:10 but ETA on right side Nov 19 19:23:38 Crofton: I see what you mean (just tried it) That sucks! Nov 19 19:23:45 rofl Nov 19 19:23:48 Jay7: nice Nov 19 19:23:51 ok, I am not crazy :) Nov 19 19:24:39 Crofton: you should fix that issue :-) Nov 19 19:24:42 Jay7: i can't seem to break it either, other than temporarily due to local changes, anyway :) Nov 19 19:24:47 heh Nov 19 19:24:56 I can't figure out what the problem is :) Nov 19 19:24:59 * sakoman_ goes back to phone call Nov 19 19:25:02 ok Nov 19 19:25:08 thanks for confirming there is a problem Nov 19 19:25:19 Jay7: what i'd like to do next is capture log messages sent during the parsing and show them afterward, to stop disrupting it due to the angstrom messages Nov 19 19:25:25 heh Nov 19 19:25:52 er, i think i need to fix the ncurses ui's handling of the new parsing events, too, though that ui is broken for a number of other reasons.. Nov 19 19:26:37 kergoth_: cool Nov 19 19:27:39 khem: opie-images was built ok Nov 19 19:28:03 so, please cherry-pick that fix of opie-notes to release Nov 19 19:29:54 Jay7: done Nov 19 19:30:02 already Nov 19 19:30:09 ok, good Nov 19 19:30:11 sakoman_, any idea the last time you saw alsamixer wrk for capture Nov 19 19:30:30 Jay7: it did not build for me. It could not fetch stuff from hh.org Nov 19 19:30:37 whats the deal with hh.org Nov 19 19:30:51 the site seems to be down in general Nov 19 19:30:52 ah.. seems I've fetched it before.. Nov 19 19:31:00 I can mirror this somewhere Nov 19 19:31:36 hmmm we can put then on sources.openembedded.org Nov 19 19:32:06 Tartarus: both patches seems fine for release Nov 19 19:32:29 oe@oedev:~/sources$ ls -d *handheld* | wc -l Nov 19 19:32:30 160 Nov 19 19:32:39 well.. half is .md5 files Nov 19 19:32:41 so only 80 Nov 19 19:32:59 are these cvs source tars ? Nov 19 19:33:12 opie.core.applets.suspendapplet_anoncvs.handhelds.org_v1_2_4_.tar.gz Nov 19 19:33:16 something like this Nov 19 19:33:30 yeah those are the ones Nov 19 19:33:34 apps.update-rc.d_anoncvs.handhelds.org_r0_7_.tar.gz Nov 19 19:33:45 I've seen complains about this Nov 19 19:33:51 can you upload them somewhere Nov 19 19:34:01 I can share on http Nov 19 19:34:15 I was trying to get opie git images goinf Nov 19 19:34:21 but better to fetch it from there and place on mirror Nov 19 19:34:49 I can put it on mirror but I need to access them Nov 19 19:35:27 * kergoth_ tries to make bitbake -n less horrifically slow Nov 19 19:37:05 http://jay-home.ath.cx/handhelds.org/ Nov 19 19:37:41 wget -np -r -l2 should be ok :) Nov 19 19:37:50 or even -l1 Nov 19 19:42:11 Jay7: ok Nov 19 19:43:08 http://pastebin.com/91h6hAV3 Nov 19 19:43:22 Jay7: also the md5 please Nov 19 19:43:26 strange messages from gcc-cross_4.5 do_package Nov 19 19:43:46 khem: there are *.md5 Nov 19 19:44:10 Jay7: cool Nov 19 19:44:17 Jay7: that gcc message is normal Nov 19 19:44:23 ignore them for now Nov 19 19:45:12 khem: I dislike /usr/lib prefix Nov 19 19:50:54 hrm, this -n speedup cluttered things up a bit Nov 19 19:50:59 runqueue really needs some work Nov 19 19:51:00 hmmm Nov 19 19:51:26 re Nov 19 19:51:56 florian: do u admin handhelds.org ? Nov 19 19:53:11 http://pastebin.com/qiaZ6BjU Nov 19 19:53:19 other thing from minimal + ben-nanonote Nov 19 19:53:25 looks like race again Nov 19 19:54:05 all following images for minimal + ben-nanonote was built ok Nov 19 19:56:01 well.. all other combinations was successfull Nov 19 19:56:10 I should update testing page Nov 19 19:57:02 khem: *cough* Nov 19 19:57:32 khem: no Nov 19 19:57:37 Jay7: copied stuff you can delete it Nov 19 19:57:43 florian: hehe ok Nov 19 19:58:16 khem: what stuff exactly? Nov 19 19:58:40 Jay7: opie Nov 19 19:58:53 ah, ok Nov 19 19:59:16 jay fish can remember max 8 seconds Nov 19 19:59:21 :) Nov 19 19:59:40 ERROR: BBCooker instance has no attribute 'bb_cache' Nov 19 20:00:04 kergoth_: thats with parallel parsing on another older p4 box running ubuntu 10.04 Nov 19 20:01:20 ~seen bluelightening Nov 19 20:01:28 khem: i haven't seen 'bluelightening' Nov 19 20:02:28 ~seen bluelightning Nov 19 20:02:29 bluelightning <~bluelight@pdpc/supporter/professional/bluelightning> was last seen on IRC in channel #opie, 6d 3h 25m 4s ago, saying: 'ah ok'. Nov 19 20:03:30 * khem also owns onoffon nick Nov 19 20:03:45 khem: i guess the only one who would be able to help with hh.org would be france Nov 19 20:03:56 he might be banned here ;) Nov 19 20:04:01 France you mean the country Nov 19 20:04:10 :) Nov 19 20:04:18 hehe Nov 19 20:04:28 I was reading about gpe conflict Nov 19 20:04:37 hh.org and ltg.org Nov 19 20:04:39 do we have a ban on french ips? ;) Nov 19 20:05:14 quite some projects suffered... including opie and ipkg Nov 19 20:05:44 but yes it might be easy to find some backround info using google :) Nov 19 20:06:15 there was that fight over gpe channel on freenode Nov 19 20:07:25 khem: btw. did you see that we hav some content here now? http://wiki.openembedded.org/index.php/Donations Nov 19 20:08:00 florian: awesome Nov 19 20:08:34 add my CPU time there ;) Nov 19 20:08:37 florian: the corps I sent this earlier promised to donate but did not fullfil Nov 19 20:08:56 florian: we should update news section more often Nov 19 20:09:04 yes indeed Nov 19 20:09:10 like when we add new recipes or new stuff Nov 19 20:09:16 it does not have to be very grand Nov 19 20:09:30 translate commit messages to twitter :) Nov 19 20:09:40 I meant the news section Nov 19 20:09:42 we need more people with a little bit of free time Nov 19 20:10:04 khem: still thinking about the 2nd commit? Nov 19 20:10:19 things like latest qt recipes or whatever should raise some interest Nov 19 20:10:19 florian:true. I so many times thought about adding stuff but some critical parts are not open to all editors Nov 19 20:10:30 Tartarus: no I looked at it Nov 19 20:10:32 said its fine Nov 19 20:10:50 florian: yes, like we have a testing page Nov 19 20:10:58 and we are doing release soon Nov 19 20:10:59 mediawiki isn't better choice for access-control.. Nov 19 20:11:15 someone was working on wiki facelift Nov 19 20:11:16 khem: hmm... for the news section it would be good if everyone would bea able to edit it Nov 19 20:11:34 florian: it just keeps devs/users upbeat Nov 19 20:11:39 spammers will be happy :) Nov 19 20:11:42 i'm not a friend of using wikis for websites anyway Nov 19 20:11:50 if we have more stuff in news Nov 19 20:12:08 well. make registration tough Nov 19 20:12:18 and let only registered users modify it Nov 19 20:12:23 right Nov 19 20:12:27 should be good Nov 19 20:14:43 khem: thanks, missed that :) Nov 19 20:16:45 Tartarus: np Nov 19 20:17:07 kergoth_: I am seeing some issues with parallel processing branch on i686/ubuntu10.04 Nov 19 20:17:16 what issues? Nov 19 20:17:28 i haven't done much testing on different build machines yet Nov 19 20:17:36 ERROR: BBCooker instance has no attribute 'bb_cache' Nov 19 20:17:45 kergoth_: runqueue.py references cooker.bb_cache which you just removed Nov 19 20:17:57 ERROR: BBCooker instance has no attribute 'bb_cache' Nov 19 20:18:09 oh, i didn't push, sorry, i have that fixed here Nov 19 20:18:10 I've not seen this Nov 19 20:18:12 give me a sec Nov 19 20:18:17 Jay7: if you haven't pulled recently, you won't :) Nov 19 20:18:21 i've been hacking on it more today Nov 19 20:18:26 hehe Nov 19 20:19:21 foerster, khem: git pull Nov 19 20:19:44 I'll pull also then :) Nov 19 20:19:48 :) Nov 19 20:20:03 today i've mostly been fixing -b, improving -b's performacne, and working on -n Nov 19 20:20:16 -i :) Nov 19 20:20:41 khem, the pkgconfig.class commit seems to be causing troubles Nov 19 20:20:54 dry-run is now only a few seconds more than the parse, whereas it took around 2.75 minutes more before Nov 19 20:20:57 should be more useful for scripting now Nov 19 20:21:09 though i still advise using -g and parsing the .dot for most things that use that info Nov 19 20:21:14 kergoth_: SRCREV and SRCDATE are same for cvs based recipes Nov 19 20:21:25 Crofton: wow Nov 19 20:21:34 kergoth_: seems fixed, thanks Nov 19 20:21:40 np Nov 19 20:21:49 Crofton: can you send and email to mailing list with your summary and request a backout Nov 19 20:21:57 will do Nov 19 20:22:07 let me make sure build really runs to completion first Nov 19 20:22:53 anyone know how alsamixer works? Nov 19 20:24:33 Crofton: ok Nov 19 20:24:38 Crofton: hmmm Nov 19 20:24:40 Crofton: not much - just installed it and got sound working on my system the other day (I have an x96 board) Nov 19 20:24:50 x86* Nov 19 20:24:54 its just a curses frontend to alsa drivers isnt it Nov 19 20:25:31 my parsing is down to 5min from 12 mins on this machine Nov 19 20:26:18 mrj10: are you at UIUC ? Nov 19 20:27:02 one of my relative went there too. He said its fullform was University of Indians and University of Chinese :) Nov 19 20:28:50 khem, ping? Nov 19 20:29:56 ka6sox: hi Nov 19 20:30:09 ka6sox: I put some more tar files on our mirror Nov 19 20:30:13 for opie Nov 19 20:30:32 hi, I am chasing down handhelds.org today and wanted to know if we needed both the wiki and cvs? Nov 19 20:31:02 ka6sox: I think we dont need it anymore Nov 19 20:31:16 opie is now on gitorious for future Nov 19 20:31:26 and we have cached the older tars on our mirror Nov 19 20:31:33 okay good..thanks Nov 19 20:31:57 someone should convert the ipkg-utils repo into a git repo somewhere Nov 19 20:31:59 if they haven't already Nov 19 20:32:30 kergoth_: those were @ hh.org too? Nov 19 20:32:41 ka6sox: yes I think Nov 19 20:32:47 yep Nov 19 20:32:52 I have put up the tar for that on mirror Nov 19 20:32:57 * kergoth_ nods Nov 19 20:33:00 but if we need to do developement Nov 19 20:33:04 still would be good to rescue it from their clutches Nov 19 20:33:10 i dont think its been touched in years anyway Nov 19 20:33:15 then we could host it on code.google.com Nov 19 20:33:20 or github somewhere Nov 19 20:33:21 yeah Nov 19 20:33:30 I've worked on re-implementing ipkgfind. Nov 19 20:33:30 or even our git server Nov 19 20:34:10 its 90% done. Nov 19 20:34:15 nice Nov 19 20:34:22 kergoth_: our git fetcher has some issues with repos on gitorious Nov 19 20:34:29 hmm, k Nov 19 20:34:43 kergoth_: same issue I was facing with efikamx kernel I am seeing it on opie too Nov 19 20:36:16 reparse is 18 secs wow Nov 19 20:37:03 ka6sox: imho, better to backup wiki and cvs somewhere anyway Nov 19 20:37:37 Jay7: ya, I'll see if I can get it. Nov 19 20:37:44 there was some software for ipaqs afaik Nov 19 20:37:55 so non-OE people can be interested too Nov 19 20:38:02 I need to remember the contact I had a HH.org that I made a while ago. Nov 19 20:41:42 hi Nov 19 20:42:00 kergoth_: is there a var to tell bitbake continue looking for cvs stash tarballs on mirrors Nov 19 20:42:06 ericben: hey Nov 19 20:42:08 kergoth_: is http://pastebin.com/pnN1zPFr correct for amend.inc's MULTIMACH_ARCH fix ? Nov 19 20:42:11 hi khem Nov 19 20:42:37 khem: for testing release, is there an interest to test against machine in an overlay ? Nov 19 20:43:11 ericben: not quite, line 57 is setting it and returning nothing instead of returning it Nov 19 20:43:15 ericben: otherwise looks right Nov 19 20:43:15 all the machines I can test on are in external overlay (armv4, armv5, armv6, armv7, x86) Nov 19 20:43:36 ericben: we wont be able to include that in release notes as it wont build out of box Nov 19 20:43:48 but that does not stop you Nov 19 20:44:15 ftp://ftp.gwdg.de/pub/linux/handhelds/ Nov 19 20:44:22 we can think about mentioning it in another section of external overlay testing or somesuch Nov 19 20:45:13 kergoth_: ok, in that case can I directly set the variables in the function instead of returning the value ? (this is maybe a stupid question as I don't know much of python) Nov 19 20:46:07 http://web.archive.org/web/20070708083652/www.handhelds.org/downloads.html Nov 19 20:46:15 here are some HH mirrors Nov 19 20:49:05 Jay7: thanks Nov 19 20:51:01 btw, are kernels-hh fetchable? Nov 19 20:51:19 does anyone build for ipaqs? Nov 19 20:52:06 hh.org used to. Nov 19 20:52:14 but I don't know anymore Nov 19 20:53:11 hm.. I'll try to do clean build for h3600 e.g. Nov 19 20:53:56 mad_soft is building for h3600 and hx4700 Nov 19 20:55:09 was building Nov 19 20:55:26 ppp question. When I try to launch pppd, I get a message back saying that I need to mknod a link for ppp in /dev. I am looking through the recipes and it looks like there is something to accomplish this in the initscripts recipe Nov 19 20:55:34 how do I get this to apply to my build? Nov 19 20:55:46 kergoth_: this seems to work : http://pastebin.com/9dVDV8As Nov 19 20:57:43 kergoth_: https://github.com/foerster/bitbake/compare/parallel-parsing...optimize-use-xrange Nov 19 20:57:56 kergoth_: again, minor speedup - about 1s improvement. Just picking off low hanging fruit ;) Nov 19 21:18:03 martinmeba: is initscript included in your image Nov 19 21:20:24 khem: in my main build task, i have "initscripts" included in RDEPENDS Nov 19 21:21:09 martinmeba: can you see that its included in image using opkg Nov 19 21:21:23 you can also write some postinstall for ppp too Nov 19 21:21:53 khem: opkg reports modutils-initscripts Nov 19 21:22:20 and yeah - I was thinking of doing a postinstall Nov 19 21:22:31 i just figured that it must have already been handled somehwere Nov 19 21:24:58 not necessarily Nov 19 21:38:36 03Chris Larson  07master * r16a9d689e6 10bitbake.git/lib/bb/cache.py: Nov 19 21:38:36 cache: make loadDataFull a classmethod Nov 19 21:38:36 Signed-off-by: Chris Larson Nov 19 21:38:47 03Chris Larson  07master * r6dac30d225 10bitbake.git/lib/bb/runqueue.py: Nov 19 21:38:47 runqueue: speed up bitbake -n Nov 19 21:38:47 From scratch, bitbake -n task-boot times: Nov 19 21:38:47 Before: 4:39.73 Nov 19 21:38:47 After: 1:52.58 Nov 19 21:38:47 Signed-off-by: Chris Larson Nov 19 21:38:49 03Chris Larson  07master * r60b9b18eaf 10bitbake.git/lib/bb/ (fetch/__init__.py persist_data.py): Nov 19 21:38:49 Revert "persist_data: cache connection and use cursor" Nov 19 21:38:49 Caching the database connection can cause serious issues if it results in Nov 19 21:38:49 multiple processes (e.g. multiple tasks) simultaneously using the same Nov 19 21:38:50 connection. Nov 19 21:38:50 This reverts commit 8a6876752b90efd81d92f0947bfc9527d8260969. Nov 19 21:38:51 Signed-off-by: Chris Larson Nov 19 21:52:19 blindvt: well done on the srcdate patch, that's so damn annoying i'm tempted to go fix them all myself :P Nov 19 21:52:25 the warnings, that is Nov 19 21:57:18 has anyone seen this message: *** %n in writable segment detected *** Nov 19 21:57:47 nothing seems broken and I keep forgetting to debug it... :) Nov 19 22:01:11 kergoth: don't know if you saw this earlier https://github.com/foerster/bitbake/compare/parallel-parsing...optimize-use-xrange Nov 19 22:01:43 kergoth: minor speed increase, shaves off another second or so. Looking through, picking off some low hanging fruit. Nov 19 22:01:59 nice, thats a good obvious one, will apply Nov 19 22:02:15 khem, I am getting more confidence the pkgconfig.class needs reverting Nov 19 22:03:19 kergoth: I have the ncurses UI at least semi-functional right now. OE base.bbclass prints directly though which mucks up the ncurses display. Nov 19 22:04:20 yeah i've noticed that one too, we need to stop all direct output to stdout/stderr in favor of communications with the UI Nov 19 22:05:05 can the server capture stdout and send everything to the UI? Nov 19 22:05:19 or, does it currently, and just print it? Nov 19 22:07:20 we capture all output from tasks, but not output during parsing, as afar as i know -- and i'm not sure about the stdout of python tasks offhand Nov 19 22:07:30 it would be much easier if we were using the xmlrpc server Nov 19 22:07:36 with the 'none' server, the server and ui share thes ame process Nov 19 22:07:46 if we neuter stdout/stderr, it'll hose the UI too Nov 19 22:07:52 (entirely, i mean) Nov 19 22:08:58 yea, I saw that. Was scratching my head b/c it called BitBakeServerFork, but in none it doesn't fork :) Nov 19 22:09:18 hehe Nov 19 22:10:00 iirc there were performance problems with the xmlrpc one, but we need to fix it if that's the case Nov 19 22:11:10 xml/rpc will open up a number of options, including making it easier for non-python UIs to be implemented (e.g. eclipse build control) Nov 19 22:11:31 cool. We should look into getting it functional then. Nov 19 22:13:53 Crofton: ok thats a good thing Nov 19 22:14:12 Crofton: although it will be intresting to see which recipe is causing the problem Nov 19 22:14:27 may be we just are needing to fix the recipe and it will work Nov 19 22:15:25 yeah Nov 19 22:15:30 * kergoth greps the repo for 'print' calls in classes Nov 19 22:15:52 I will ask for a revert due to a mysterious build failure Nov 19 22:15:59 sorry I can't get any closer then that Nov 19 22:16:45 Crofton: thats ok Nov 19 22:17:24 * Crofton is juggling several critical things Nov 19 22:19:28 what to build tonight? :) Nov 19 22:20:07 Jay7: build any gnome based images out of release branch Nov 19 22:20:25 i.e. from master e.g.? Nov 19 22:20:29 Jay7: and meta-toolchains Nov 19 22:20:44 no from release-2010.12 branch Nov 19 22:20:57 ah.. I've misunderstood 'out of' :) Nov 19 22:21:05 lack of english knowledge.. Nov 19 22:21:14 ok Nov 19 22:22:00 same here Nov 19 22:24:27 Crofton: what issues wrt pkgconfig changes? Nov 19 22:25:50 well.. now I'll use parallel-bb for build :) Nov 19 22:27:37 basically it fork bombs my build machine Nov 19 22:30:22 I see, I found probably a single package failing after that commit Nov 19 22:30:31 nothing serious, though Nov 19 22:30:33 btw Nov 19 22:30:37 obi: ping Nov 19 22:30:51 ant__: Crofton seems to have bigger problems Nov 19 22:30:57 he Nov 19 22:31:06 khem: I have one bigger Nov 19 22:31:12 Crofton: what image should we try to reproduce it Nov 19 22:31:21 trying to build dietlibc -> gcc segfault Nov 19 22:34:31 khem: did you mean under 'meta-toolchains' meta/meta-toolchain-*.bb? Nov 19 22:35:13 khem: yea, x11-gpe-image builds too having solved the issue of missing .pc Nov 19 22:35:52 add task-angstrom-gnome.bb to an image Nov 19 22:36:00 maybe just try buidling that Nov 19 22:36:37 ok, still studying that libgpepimc recipe... Nov 19 22:37:08 hm... Nov 19 22:37:46 are meta-toolchain* recipes distro-dependent? Nov 19 22:38:29 i.e. should I build it for angstrom and minimal? Nov 19 22:41:40 and I've found only one real gnome-based image - angstrom-gnome-image :) Nov 19 22:42:15 udumpty-image have gnome parts too Nov 19 22:45:11 khem: may I push that commit to master? Nov 19 22:48:00 git checkout -b release-2010.12 origin/release-2010.12 Nov 19 22:48:00 fatal: git checkout: updating paths is incompatible with switching branches. Nov 19 22:48:03 ? Nov 19 22:48:42 well.. I've started toolchains and angstrom-gnome-image build for set of machines Nov 19 22:48:58 * Jay7 -> sleep Nov 19 23:12:27 ant__: yes sure and also cherry pick that to release branch Nov 19 23:12:46 tharvey: hmm strange Nov 19 23:12:47 ok, I'm now distracted by another finding... Nov 19 23:13:16 cp: cannot stat `/oe/build/tmp/work/i686-linux/stagemanager-native-0.0.1-r15/sysroot-destdir//oe/build/tmp/*': No such file or directory Nov 19 23:13:43 that comes from run.do_populate_sysroot Nov 19 23:14:06 yes but that destdir is not correct Nov 19 23:14:12 interestingly there are two calls to 'cp' and the first one is commented...? Nov 19 23:14:21 tharvey: did you do a git pull --rebase Nov 19 23:14:25 before branching Nov 19 23:14:47 sysrootdest = bb.data.expand('${SYSROOT_DESTDIR}${STAGING_DIR_TARGET}', d) Nov 19 23:14:49 hmm Nov 19 23:14:55 khem, doh... sorry I did not Nov 19 23:16:04 kkhem: that seems the culprit os.system(bb.data.expand('cp -pPR ${SYSROOT_DESTDIR}${TMPDIR}/* ${TMPDIR}/', d)) Nov 19 23:19:49 that should be switched to use python methods rather than system Nov 19 23:38:22 khem, did my revert email get to the list? Nov 19 23:42:12 hrm Nov 19 23:48:46 ant__: pong Nov 19 23:50:25 hi, I was locally testing the libgpepimc.pc issue and seems ok Nov 19 23:50:45 but now a rebuild from packaged-staging is failing... Nov 19 23:51:09 other reason it seems Nov 19 23:52:23 ok Nov 19 23:52:45 he he Nov 19 23:52:46 libxml2-2.7.8: shared library versionning was broken Nov 19 23:52:54 now is broken again ;) Nov 19 23:55:30 JaMa|Off: patched it 12 hours ago..let's retry from scratch Nov 19 23:56:36 obi: about libgpepimc patch, may I add your Acked-by? Nov 19 23:57:47 ha..another one Nov 19 23:57:50 NOTE: package git-native-1.7.0.2-r6: task packagestage_scenefunc: Failed Nov 19 23:57:50 ERROR: Function 'packagestage_scenefunc' failed Nov 19 23:58:03 i haven't seen it. what was your solution? Nov 19 23:58:17 just removing do_install Nov 19 23:58:27 ok, then feel free to add my ack Nov 19 23:58:36 k, thx Nov 20 00:07:36 anyone familar with python here? Nov 20 00:08:25 Crofton|road: the problem with reverting the pkgconfig patch is, that it would just hide broken install rules (either in .bb files or in the package itself) and break other (combinations of) packages instead. i'm building angstrom-task-gnome right now. if a problem is caused by the pkgconfig patch, then it's easy to fix Nov 20 00:08:44 balister_: I think that on #python have sure! Nov 20 00:08:56 s/caused/uncovered/ Nov 20 00:09:09 this is an OE packaging issue Nov 20 00:09:22 can you pastebin ? Nov 20 00:09:31 looks like we do not install platform.py Nov 20 00:10:26 not much to pastebin Nov 20 00:11:18 if I INHERIT += "rm_work" is there a simple global config I can use to disable that for a specific recipe? Nov 20 00:12:31 03Andrea Adami  07org.openembedded.dev * r7953bc1414 10openembedded.git/recipes/libgpepimc/ (libgpepimc.inc libgpepimc_0.9.bb libgpepimc_svn.bb): Nov 20 00:12:31 libgpepimc: correctly install libgpepimc.pc file Nov 20 00:12:31 * by removing now superfluous custom do_install task Nov 20 00:12:31 * fix build of libtododb and libcontactsdb Nov 20 00:12:31 * bump PR Nov 20 00:12:32 Acked-by: Andreas Oberritter Nov 20 00:12:32 Signed-off-by: Andrea Adami Nov 20 00:12:51 tharvey: yes, create an empty task in the .bb Nov 20 00:13:03 with same name :) Nov 20 00:14:31 ya that makes sense... was hoping to come up with something in local.conf Nov 20 00:14:59 well, you inherit it in local.conf...I don't see the point Nov 20 00:16:29 I use to do it for the single packages I'm working on Nov 20 00:20:09 ya likely makes more sense to touch those bb's Nov 20 00:24:20 or use bbappend or amend.inc to do it Nov 20 00:34:29 I'm using the latest srctree class and a bit confused. If I bitbake mypackage.bb then I see in the log 'mypackge-ver: execute blah." And it gets packaged nicely into an ipk file. However I think at the package to the RDEPENDS of a task which itself is in the IMAGE_INSTALL of an image recipe. Now whenever I build the image, even though mypackage is staged, it runs through the packages do_ functions ... Nov 20 00:34:47 bah... -c clean and git-native and libxml2 build now, on pstage rebuild... Nov 20 00:38:27 he..strange things happen here... Nov 20 00:38:31 NOTE: package gtk+-2.20.1-r10.4: task do_configure: Failed Nov 20 00:38:31 ERROR: Couldn't install the staging package to a temp directory Nov 20 00:38:40 wtf? Nov 20 00:39:36 oh..ERROR: copying inserted file: No space left on device Nov 20 00:40:00 no way../dev/sdb5 73G 11G 59G 16% /oe Nov 20 00:41:42 Crofton: my host is haunted too Nov 20 00:43:55 sigh.. again Nov 20 00:43:57 NOTE: Failure installing prestage package Nov 20 00:44:20 * ant__ is about rebooting Nov 20 00:54:34 khem, how are you managing copies from dev to the release branch Nov 20 01:06:27 is there something I can put in a machine file to mark a module for autoload? Nov 20 01:07:10 module_autoload_g_ether = "g_ether" Nov 20 01:07:14 this looks good .... Nov 20 01:07:29 Crofton, module_autoload_ar6000 = "ar6000" Nov 20 01:07:32 right Nov 20 01:08:21 Crofton: push the patch to master Nov 20 01:08:30 then cherry pick into branch Nov 20 01:08:35 ok Nov 20 01:08:42 who is doing the cherry picking? Nov 20 01:09:01 anyone as long as the patch is discussed on IRC or ml Nov 20 01:09:17 if u prefer u can ask it to be merged too Nov 20 01:09:18 I am concerned we will miss stuff Nov 20 01:09:38 I am observing master too Nov 20 01:09:41 ok Nov 20 01:10:07 for this to work, we need someone watching for premature divergence :) Nov 20 01:10:12 but we focus on getting branch into stable state Nov 20 01:10:24 it wont happen Nov 20 01:10:35 and 2 weeks is short time Nov 20 01:10:54 unless people push radical stuff without review Nov 20 01:12:05 I am happy with progress so far Nov 20 01:12:19 we are really fixing issues that need to be fixed for a release Nov 20 01:12:56 bother Nov 20 01:13:04 module_autoload_usrp_e = "usrp_e" Nov 20 01:13:15 this gitorious pulls failing is a concern Nov 20 01:13:19 what to do about the underscore in the moudle name Nov 20 01:13:24 module anme Nov 20 01:13:26 arg Nov 20 01:13:29 bad typing Nov 20 01:13:47 hmmm _ is a override char Nov 20 01:14:06 we should have chosen something else Nov 20 01:14:22 I should have not used _ in the module name Nov 20 01:14:31 yeah jsut use - Nov 20 01:14:58 looks ok Nov 20 01:15:02 g_ether is the same Nov 20 01:16:04 I wonder if we should address this issue in future bb2 language **** ENDING LOGGING AT Sat Nov 20 02:59:57 2010