**** BEGIN LOGGING AT Wed Nov 25 03:35:38 2009
Nov 25 05:53:05 <cdbot2> * * OE Bug 4179 has been RESOLVED (FIXED) by bernhard.kaindl(AT)gmx.net
Nov 25 05:53:07 <cdbot2> * * external toolchain generated by  meta-toolchain depends on the original build location,
Nov 25 05:53:09 <cdbot2> * * http://bugs.openembedded.net/show_bug.cgi?id=4179
Nov 25 06:01:04 <cdbot2> * * OE Bug 2393 has been RESOLVED (FIXED) by bernhard.kaindl(AT)gmx.net
Nov 25 06:01:06 <cdbot2> * * meta-toolchain/sdk does not build.
Nov 25 06:01:08 <cdbot2> * * http://bugs.openembedded.net/show_bug.cgi?id=2393
Nov 25 07:22:05 <cdbot2> * * OE Bug 1904 has been RESOLVED (FIXED) by bernhard.kaindl(AT)gmx.net
Nov 25 07:22:07 <cdbot2> * * glibc-intermediate build failure (icecc compile farm)
Nov 25 07:22:09 <cdbot2> * * http://bugs.openembedded.net/show_bug.cgi?id=1904
Nov 25 07:29:57 <tsjsieb> good morning
Nov 25 08:17:41 <XorA> morning
Nov 25 08:52:53 <mickey|office> morning
Nov 25 08:54:32 <XorA> hey mickey|office please send ack for new members vote otherwise I cant start the voting tomorrow
Nov 25 08:55:00 <mickey|office> righto
Nov 25 08:55:02 <XorA> mickey|office: acording to policy vote proposal must be supported by two members
Nov 25 08:55:44 <mickey|office> underway
Nov 25 08:56:12 <XorA> mickey|office: thanks
Nov 25 08:56:29 <XorA> mickey|office: at least 5 people need to vote to make vote valid, Im sure we can make that in 2 days
Nov 25 08:56:30 <mickey|office> np
Nov 25 08:56:38 <mickey|office> yeah, sounds reasonable
Nov 25 08:56:56 <mickey|office> we can stay in the planned schedule for the TSC voting then
Nov 25 08:57:12 <XorA> mickey|office: yes
Nov 25 08:57:17 <mickey|office> btw., do folks listed on the Wiki TSC page need to reapply or are they already in the pool?
Nov 25 08:57:27 <XorA> mickey|office: they need to re-apply
Nov 25 08:57:49 <XorA> mickey|office: my original email stated this
Nov 25 08:58:03 <mickey|office> ok, didn't read closely enough
Nov 25 08:58:07 <mickey|office> thanks, will mail then
Nov 25 08:58:11 <XorA> mickey|office: kick people to do so
Nov 25 08:58:15 <mickey|office> will do
Nov 25 08:58:28 <XorA> for TSC Im trying to follow rules to letter
Nov 25 08:58:33 <mickey|office> yep
Nov 25 08:58:33 <XorA> as its an important task
Nov 25 08:58:42 <mickey|office> do you maintain a list of applicants so far?
Nov 25 08:58:49 <mickey|office> in the open somewhere i mean
Nov 25 08:58:51 <CIA-80> 03Sebastian Spaeth <Sebastian@SSpaeth.de> 07org.openembedded.dev * r1f0f2940bf 10openembedded.git/recipes/navit/ (files/navit.launcher navit.inc): 
Nov 25 08:58:51 <CIA-80> navit: Do not set /proc/sys/vm/overcommit in an app launcher, especially not without ever changing it back.
Nov 25 08:58:51 <CIA-80> * This reverts parts of 742fb4ae71e6a7e655385086ba9edebbf00b6ebd
Nov 25 08:58:51 <CIA-80> * If someone wants to do use this, it should be done on a more finegrained distro basis
Nov 25 08:58:53 <CIA-80> Signed-off-by: Sebastian Spaeth <Sebastian@SSpaeth.de>
Nov 25 08:58:55 <XorA> mickey|office: its on the list
Nov 25 08:59:01 <mickey|office> ok
Nov 25 08:59:16 <XorA> mickey|office: candindacy requests must be public on members list
Nov 25 08:59:24 <mickey|office> ok, then everything is clear
Nov 25 08:59:26 <mickey|office> thanks
Nov 25 08:59:41 <XorA> the policy was designed for transparency I think
Nov 25 09:00:04 <mickey|office> yeah, makes sense
Nov 25 09:01:12 <XorA> I shall look into the software for counting at the weekend and see if I can figure it out
Nov 25 09:01:24 <XorA> otherwise I looked up how to count them manually :-D
Nov 25 09:01:28 <mickey|office> hehe
Nov 25 09:01:38 <mickey|office> thankfully we're not KDE e.V. yet
Nov 25 09:01:53 <mickey|office> so even if you need to resort to manual counting, it should be somewhat handleable
Nov 25 09:02:21 <XorA> mickey|office: yeah, I have done hand counts for about 100 votes before
Nov 25 09:02:28 <XorA> using same system
Nov 25 09:03:55 <XorA> mickey|office: we just have to get -people trained into proper voting system now :-D
Nov 25 09:04:15 <mickey|office> hehe, true
Nov 25 09:04:39 <hrw> morning
Nov 25 09:04:42 <mickey|office> hey hrw 
Nov 25 09:04:45 <mickey|office> morning
Nov 25 09:07:38 <hrw> XorA: I forgot that member proposal has to be acked by two members - would ack them right away
Nov 25 09:09:34 <XorA> hrw: is done now, I shall continue to kick people into new voting procedures are common to people
Nov 25 09:09:55 <hrw> I know that it is done nw
Nov 25 09:10:44 <XorA> this stuff is new to everyone so part of my board duties is to make sure it goes smoothly
Nov 25 09:10:52 <hrw> ;)
Nov 25 09:12:29 <hrw> 160GB fullHD cameras produce too big files
Nov 25 09:13:54 <mickey|office> hehe
Nov 25 09:13:57 * mickey|office still on DV
Nov 25 09:14:49 <RP> morning all
Nov 25 09:14:53 <hrw> mickey|office: it was my brother-in-law camera. 5 minutes per GB
Nov 25 09:14:56 <hrw> RP: hi
Nov 25 09:15:15 <RP> XorA: Thanks for getting this stuff rolling :)
Nov 25 09:24:34 <mickey|office> hi RP
Nov 25 09:28:38 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r46a3e348d5 10openembedded.git/conf/distro/include/angstrom-glibc.inc: angstrom-glibc: switch back to -O2
Nov 25 09:28:49 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r24c8c7b88f 10openembedded.git/ (8 files in 4 dirs): linux-omap 2.6.29: add smsc patches and update overo defconfig
Nov 25 09:33:13 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r379c751a9a 10openembedded.git/contrib/angstrom/build-feeds.sh: angstrom-feed-builder: add more stuff
Nov 25 09:33:24 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r53ee8aa889 10openembedded.git/recipes/tasks/angstrom-task-gnome.bb: angstrom task gnome: clean up package list a bit
Nov 25 09:33:34 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * re2eceea9c9 10openembedded.git/recipes/angstrom/angstrom-feed-configs.bb: angstrom feed configs: rrecommend opkg-nogpg-nocurl
Nov 25 09:33:35 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r810296df41 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: add more boards
Nov 25 10:05:36 <CIA-80> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * r4293b84e5b 10openembedded.git/recipes/linux/linux.inc: linux.inc: create DEPLOY_DIR_IMAGE in do_devicetree_image
Nov 25 10:48:42 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r035b4d579c 10openembedded.git/recipes/pam/ (libpam-base-files.bb libpam-base-files/pam.d/polkit): libpam-base-files: polkit is provided by policykit itself
Nov 25 10:52:07 <hrw> XorA: so we vote by answering to mail or by sending you separate mail?
Nov 25 10:52:33 <XorA> hrw: I will send you a poll
Nov 25 10:52:50 <XorA> hrw: same as debian
Nov 25 10:52:55 <XorA> hrw: in theory
Nov 25 10:53:22 <hrw> ok
Nov 25 10:53:38 <hrw> shit
Nov 25 10:53:58 <hrw> desktop-(samba)->sheevaplug-(usb)->hdd is 2MB/s
Nov 25 10:54:24 <hrw> desktop-1gbEswitch-sheevaplug
Nov 25 10:55:20 <sgh> Hi
Nov 25 10:56:28 <sgh> it seems that i686-generic.conf allways override PROFERRED_PROVIDER_virtual/xserver. Where do I set it if I want to make sure that I have preferrence over i686-generic.conf ?
Nov 25 11:00:12 <sgh> writing PROFERRED_PROVIDER_virtual/xserver="xserver-xorg" in local.conf does not actually set it.
Nov 25 11:00:44 <XorA> sgh: i686-generic setting that variable sounds like a bug
Nov 25 11:01:06 <sgh> XorA: it is not even set with a ?=
Nov 25 11:01:29 <XorA> sgh: it shoudk be ?= but I think maybe it shouldnt be set at all
Nov 25 11:02:14 <sgh> XorA: doing a "grep xserver *" in machines is not particullary nice reading in that case :/
Nov 25 11:02:21 <pb_> I dunno, I would have thought that if you didn't want the definitions from i686-generic.conf then you should just be selecting a different MACHINE.
Nov 25 11:03:00 <pb_> if you want to override it locally (i.e. in local.conf) then you can set PREFERRED_PROVIDER_virtual/xserver_local = "..."
Nov 25 11:03:12 <sgh> pb_: or overrride the machine .... I'll try that instead.
Nov 25 11:03:55 <pb_> but there is so little in the i686-generic machine definition that, if it doesn't match your needs, I would recommend that you simply don't use it.  it's not as if it is a big complicated piece of mechanism that would be difficult to replicate.
Nov 25 11:04:32 <pb_> mickey|office: good morning
Nov 25 11:04:43 <sgh> pb_: thanks .....
Nov 25 11:05:50 <woglinde> jo
Nov 25 11:06:37 <hrw> oj
Nov 25 11:07:23 <woglinde> time to update libogg and libvorbis
Nov 25 11:07:28 <mickey|office> morning pb_
Nov 25 11:07:35 <mickey|office> which inotify library are people using these days?
Nov 25 11:07:44 <woglinde> hi mickeyl
Nov 25 11:08:03 <woglinde> why you need it?
Nov 25 11:08:10 <woglinde> I never saw any sin
Nov 25 11:08:51 <mickey|office> in using a library over the glibc interface or in inotify itself?
Nov 25 11:14:13 <pb_> hi woglinde 
Nov 25 11:14:48 <pb_> mickey|office: I don't think I've ever used anything other than the "raw" api that glibc provides
Nov 25 11:15:59 <mickey|office> pb_: i see, I recall someone having added some glib-style API on top of it, but i don't remember who or where :)
Nov 25 11:16:24 <mickey|office> i guess the glibc API will do
Nov 25 11:16:31 * mickey|office goes and writes a .vapi for it
Nov 25 11:16:47 <pb_> oh yeah, you're right, I do remember that now as well
Nov 25 11:16:49 <pb_> I never used it though
Nov 25 11:17:10 <pb_> I guess you already checked that it isn't in glib itself, right?
Nov 25 11:17:35 <pb_> hm.  http://www.kernel.org/pub/linux/kernel/people/rml/inotify/glib/README
Nov 25 11:17:56 <mickey|office> ah, that could b eit
Nov 25 11:18:10 <mickey|office> looks rather old though ;)
Nov 25 11:18:15 <mickey|office> ancient actually
Nov 25 11:18:17 <pb_> yeah
Nov 25 11:18:29 <pb_> the inotify api is "mature" though so I don't suppose it would have needed to change
Nov 25 11:18:38 <mickey|office> true
Nov 25 11:19:12 <pb_> plenty of code in glibc and glib themselves dates from the 1990s, I wouldn't necessarily dismiss that stuff because it is a few years old :-}
Nov 25 11:19:24 <mickey|office> hehe, of course not
Nov 25 11:20:06 <mickey|office> ah, and desrt wrote http://blogs.gnome.org/desrt/2005/09/04/introducing-libinotify/
Nov 25 11:20:18 <pb_> it is a bit odd that glib itself doesn't contain inotify support by now, actually
Nov 25 11:20:28 <pb_> I would have thought that would be a useful enough function to have been included.
Nov 25 11:20:39 <pb_> maybe it's too hard to emulate on non-linux
Nov 25 11:21:09 <mickey|office> yes, that might be the culprit
Nov 25 11:27:04 <mickey|office> ok, desrt's lib is still in gnome git
Nov 25 11:27:09 <mickey|office> build and compiles
Nov 25 11:27:16 <pb_> very good
Nov 25 11:27:22 <mickey|office> i guess i'll try to use that now
Nov 25 11:27:47 <mickey|office> saves me from some lines of code doing the mainloop integration
Nov 25 11:27:57 <pb_> yeah
Nov 25 11:35:33 <rvfo> May Anyone help me? I need to compile the task-base without libx11, gtk+, gconf, because I only need a simple linux image with the minimal and QT Libraries,  I remove the lines  PREFERRED_PROVIDER_(avahi, gtk, x11..), but doesn't work to compile this minimal linux, what is the way to do that?
Nov 25 11:44:08 <rvfo> I'm compiling the minimal distro fro a x86 machine, How is the usually final size for this machine?
Nov 25 11:44:30 <koobe> rvfo: try bitbake -g to see the dependencies and understand why there are more packages built than you would expect
Nov 25 11:44:54 <rvfo> koobe: thanks
Nov 25 11:46:38 <CIA-80> 03Thomas Zimmermann <ml@vdm-design.de> 07org.openembedded.dev * r5a38714e50 10openembedded.git/recipes/tasks/task-shr-feed.bb: task-shr-feed: add transmission to the feed, fixes SHR bug #745
Nov 25 11:52:27 <hrw> btw - did someone tried to make build entirely in tmpfs?
Nov 25 11:53:03 <woglinde> nope
Nov 25 11:53:46 <hrw> I am waiting for new build machine at BugLabs for it
Nov 25 11:53:54 <hrw> 12GB ram should make it fit
Nov 25 11:55:39 <DJWillis> hrw: I considered doing that, 8GB of space was an epic fail ;-). Let us know how it goes as I tihnk it's a great idea.
Nov 25 11:56:16 <mario-goulart> hrw: yesterday I was talking to otavio about something similar.  I thought about storing ccache data in a ram filesystem.
Nov 25 11:57:14 <hrw> DJWillis: I did build with 6GB of disk space in past. this require rm_work and split into few builds
Nov 25 11:58:12 <pb_> DJWillis: if you've got enough ram then I don't imagine you will notice much difference from building in tmpfs compared to any other filesystem.  oe builds are not, generally, filesystem-intensive enough for it that to have much effect on performance.
Nov 25 12:03:19 <CIA-80> 03Thomas Zimmermann <ml@vdm-design.de> 07org.openembedded.dev * r84024549aa 10openembedded.git/recipes/tasks/task-shr-feed.bb: task-shr-feed: add xf86-input-tslib to the feed, with this TS can be calibrated again
Nov 25 12:06:10 <XorA> hmm debian-vote doesnt handle group elections
Nov 25 12:06:32 <broonie> debian-vote?
Nov 25 12:06:40 <broonie> Do you mean devotee?
Nov 25 12:07:02 <broonie> (and what do you mean by a group election?)
Nov 25 12:07:03 <XorA> broonie: no, the other implementation
Nov 25 12:07:19 <broonie> I didn't even realise there was another one...
Nov 25 12:07:23 <XorA> broonie: having issues getting hold of devotee as the instructions dont work with my tla
Nov 25 12:07:58 <broonie> I don't think Manoj is using tla any more, the repository is probably broken.
Nov 25 12:08:03 <XorA> broonie: group election, you have Y candidates and X roles to fill
Nov 25 12:08:08 <broonie> it'll be in git IIRC.
Nov 25 12:08:20 <XorA> broonie: be cool if you had a link
Nov 25 12:08:26 <broonie> XorA: SPI uses almost the same voting system as Debian and has some software to do that.
Nov 25 12:08:42 <broonie> XorA: No idea where the source is but it really ought to be public.
Nov 25 12:11:03 <broonie> Can't find devotee easily but asking secratary@debian.org where the current stuff is will probably help.
Nov 25 12:12:57 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r2c8134baa5 10openembedded.git/classes/pkgconfig.bbclass: 
Nov 25 12:12:57 <CIA-80> pkgconfig bbclass: also sed out bogus includes
Nov 25 12:12:57 <CIA-80> RP says: In theory the recipes themselves should be patched to stop this nonesense getting in there in the first place
Nov 25 12:12:57 <CIA-80> but that is appropriate for pkgconfig.bbclass
Nov 25 12:13:10 <broonie> XorA: lunch, anyway?
Nov 25 12:13:33 <XorA> yes
Nov 25 12:25:19 <woglinde> hm so
Nov 25 12:25:25 <woglinde> libvorbis is simplified
Nov 25 12:29:27 <woglinde> hm seems libogg can be simplified as well
Nov 25 12:47:10 <sgh> gcc-cross can't compile. The log is here : http://pastebin.com/d2b69e961
Nov 25 12:47:21 <sgh> It does not seem like it actually fail ....
Nov 25 12:48:36 <sgh> ah ... it fails on line 243 
Nov 25 12:51:39 <woglinde> hm objc
Nov 25 12:52:05 <woglinde> angstroem?
Nov 25 13:05:15 <hrw> ~hail nfs
Nov 25 13:05:17 * ibot bows down to nfs and chants, "I'M NOT WORTHY!!"
Nov 25 13:06:00 <hrw> desktop->sheeva->usbhdd over sshfs: 2MB/s, over samba: 2MB/s, over nfs: 10-20MB/s
Nov 25 13:13:23 <CIA-80> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * rfbb10c0340 10openembedded.git/recipes/freesmartphone/fsogsmd_git.bb: fsogsmd: no longer requires libxml2 nor broadband info
Nov 25 13:13:23 <CIA-80> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * re9383dc132 10openembedded.git/recipes/freesmartphone/libfsoframework_git.bb: libfsoframework: add missing dependency on alsa-lib
Nov 25 13:13:24 <CIA-80> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * r6c58a537d4 10openembedded.git/ (3 files in 2 dirs): vala: add 0.7.8.2 release
Nov 25 13:13:24 <CIA-80> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * rc8da8a939d 10openembedded.git/conf/distro/include/sane-toolchain.inc: sane-toolchain.inc: bump linux-userspace-headers to match with bumped eglibc version
Nov 25 13:14:42 <mckoan> good morning
Nov 25 13:14:55 <hrw> hi Marco
Nov 25 13:15:31 * hrw -> coffe
Nov 25 13:15:32 <hrw> e
Nov 25 13:18:55 <XorA> hmmm, ubuntu in tentacle porn shocker -D
Nov 25 13:28:59 <broonie> hrm?
Nov 25 13:29:04 <hrw> que?
Nov 25 13:29:26 <Crofton|work> broonie, you should know better than to ask
Nov 25 13:29:58 <woglinde> hrw you could try phomelfs in the future
Nov 25 13:30:14 <woglinde> hrw which nfs version 3 or 4?
Nov 25 13:30:15 <XorA> broonie: read /.
Nov 25 13:31:20 <hrw> -t nfs so rather 3
Nov 25 13:32:17 <mike_cw> can anyone tell me what would cause a -c clean to fail on missing tmp/staging/staging.lock, and how I might fix it?
Nov 25 13:38:03 <XorA> ~blame ibot
Nov 25 13:38:07 * ibot blames ibot (and Canada) for all the evil in the world
Nov 25 14:06:25 <kergoth> pb_: doener: kergoth: git for-each-ref --format='%(upstream)' $(git rev-parse --symbolic-full-name HEAD)
Nov 25 14:06:27 <kergoth> pb_: fyi :)
Nov 25 14:06:51 * kergoth adds that to his text file of random git tips
Nov 25 14:11:30 <kergoth> meh, still not completely happy with these scripts for keeping track of cherry picks.. it functions, yes, but there are corner cases... for example, take a commit FOO which becomes commits BAR, BAZ, and MEH upstream, but there's a part of foo that isn't encompassed by BAR+BAZ+MEH.. how the hell do you reflect that?  maybe i just wont, and instead just diff the two trees after all the commits have been dealt with to catch any remnant changes
Nov 25 14:16:26 <hrw> kergoth: what does that git magic do?
Nov 25 14:16:54 <kergoth> when you create a local branch, you can supply a remote tracking branch its based on, which it merges from when you pull
Nov 25 14:17:04 <kergoth> that command gets that tracking branch given the current branch
Nov 25 14:17:11 <kergoth> useful if you want to compare the current branch and its "upstream"
Nov 25 14:17:27 <hrw> ok
Nov 25 14:22:35 * kergoth is trying, trying to make it easier to keep track of what's gone back and forth between a private overlay and upstream oe
Nov 25 14:28:12 * RP proposed adding pkgconfig-native to the default dependencies and dropping pkgconfig.bbclass
Nov 25 14:28:29 <RP> s/proposed/proposes/
Nov 25 14:39:50 <kergoth> as far as I'm concerned, anyone not using pkg-config should start using it, so that sounds okay, but I'm always hesitant about touching the default deps..  another case where itd be nice to capture and store information from previous builds, so we could store the fact that a given SRC_URI results in a source tree that uses pkg-config (has .pc.in/.pc files) :)
Nov 25 14:40:40 <RP> kergoth: You need pkg-config if it has any pkgconfig dependencies too
Nov 25 14:41:19 <kergoth> true.  that could be automatically determined for autoconf based buildsystems by macro usage, but not for custom ones
Nov 25 14:44:27 <kergoth> i wish i could do rev-list-like sorting on commits going into git-diff-tree, without using rev-list
Nov 25 14:46:13 <rvfo> May you help me, please? I was trying to compile the  libxcb_1.0.bb recipes and get the error:  unable to parse xproto.xml. All Error debug is here: http://pastebin.com/m2b35bf78
Nov 25 14:52:35 <CIA-80> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r3ab82dfb59 10openembedded.git/ (conf/checksums.ini recipes/inkscape/inkscape_0.47.bb): inkscape: add 0.47
Nov 25 14:58:27 <kergoth> there we are, have a little function that takes a list of arbitrary commits and reorders them by author date first, then by commit date, oldest to newest, and uses that to determine which of the two commits is really the origin..
Nov 25 14:58:27 <kergoth> hmm
Nov 25 15:03:45 <pb_> kergoth: oh, very good :-)
Nov 25 15:04:06 <kergoth> ♥ #git
Nov 25 15:04:14 <kergoth> learned quite a bit via osmosis
Nov 25 15:04:18 <kergoth> heh
Nov 25 15:04:24 <pb_> heh
Nov 25 15:04:37 <pb_> yeah.  I still find myself wincing at the name, but I'm slowly getting used to that.
Nov 25 15:05:09 <kergoth> heh
Nov 25 15:11:27 <dies_felices> Hello
Nov 25 15:13:09 <dies_felices> How do I change the prefix path safely?
Nov 25 15:14:51 <kergoth> safely?
Nov 25 15:15:09 <kergoth> I'd suggest looking at the micro distro, it changes prefix to flatten the filesystem
Nov 25 15:15:46 <dies_felices> Thanks.
Nov 25 15:16:23 * kergoth wonders if an oe tree is too large to go on github
Nov 25 15:16:59 <broonie> yes
Nov 25 15:17:16 <broonie> limit is (was?) 100M
Nov 25 15:21:51 <kergoth> hrmph, figured :)
Nov 25 15:21:55 <kergoth> thanks for the info
Nov 25 15:22:27 * kergoth thinks about paying for an online VM to host stuff like that, personal repositories
Nov 25 15:22:50 <broonie> You could ask them nicely.
Nov 25 15:23:02 <kergoth> heh :)
Nov 25 15:23:29 <broonie> They may be willing to maintain an OE clone - they seem to only count new objects in your quota (otherwise Linux wouldn't work either).;
Nov 25 15:23:52 <kergoth> ah, thats an idea.. so then I'd just maintain a fork of it
Nov 25 15:23:56 <kergoth> good thought
Nov 25 15:25:38 * kergoth likes how github shows readme files automatically, handy feature
Nov 25 15:25:41 <Crofton|work> at OEDEM people seemed to think that we could add repos for devs on request to our existing server
Nov 25 15:25:52 <Crofton|work> I'd need to double check the minutes though
Nov 25 15:27:39 <kergoth> that would be nice.
Nov 25 15:27:48 <kergoth> are we using gitosis there? would make it easy
Nov 25 15:29:11 <th1_> that commit bot CIA-80, can I get the source for it anywhere?
Nov 25 15:29:15 <hrw> we use gitosis
Nov 25 15:29:19 <hrw> th1_: git pull
Nov 25 15:29:32 <th1_> from where?
Nov 25 15:30:06 <pb_> th1_: http://code.google.com/p/cia-vc/source/checkout
Nov 25 15:30:25 <th1_> cool thanks
Nov 25 15:31:36 <kergoth> who's administering the git server right now?
Nov 25 15:31:56 <pb_> kergoth: cbrake administers it in general, though I don't know if he is actively administering it at this instant
Nov 25 15:32:02 <kergoth> ah
Nov 25 15:32:31 <RP> kergoth: We are using gitosis
Nov 25 15:32:37 <RP> kergoth: I have admint access, what do you need?
Nov 25 15:33:26 <kergoth> we were just talking about the whole personal repositories thing, to alleviate the need to request administrative assistance for branch removal / non-fast-forward / etc.. do you recall the results of that discussion from oedem, by chance?
Nov 25 15:33:32 * kergoth tries to figure out an odd packging issue
Nov 25 15:34:02 <RP> kergoth: I think I said I'd setup a contrib repo for general access by people
Nov 25 15:34:21 <RP> kergoth: I think a repo per person wasn't felt to be a great idea
Nov 25 15:35:06 <broonie> I remember a post to the list to that effect, dunno if it was you that volunteered at the time but well done :)
Nov 25 15:35:18 <kergoth> i strongly disagree, contrib will become an unmaintainable mess
Nov 25 15:35:24 <kergoth> but if that's the consensus
Nov 25 15:36:00 <RP> kergoth: well, its better than having the main repo be a total mess
Nov 25 15:36:37 <kergoth> the main repo is already a mess, because we can't delete our own branches when we're done with them :)
Nov 25 15:36:43 <kergoth> i agree its an improvement
Nov 25 15:37:13 <RP> kergoth: What would be really cool is if gitosis actually exported the user to an environmental variable
Nov 25 15:37:20 <RP> Then the commit hooks could do stuff with that
Nov 25 15:37:57 <kergoth> ah, to auth access to per user branches?
Nov 25 15:38:03 <RP> kergoth: right
Nov 25 15:38:09 <RP> kergoth: or to allow branch deletion
Nov 25 15:38:35 <RP> gitosis doesn't have this functionality of course :/
Nov 25 15:39:40 * Crofton|work can't remember what we thought, but I sort of like the repo per person still :)
Nov 25 15:39:49 <kergoth> in the past, I've set things up manually to distinguish behavior between connections by their authenticating private keys, even if the user is the same
Nov 25 15:40:03 <kergoth> so i know its possible, just a matter of setting up the authorized keys in a certain way
Nov 25 15:40:19 <broonie> kergoth: gitosis already does that.
Nov 25 15:40:37 <broonie> kergoth: eg, to give read functionality separately to write.
Nov 25 15:41:02 <kergoth> nod.  well, RP wants that info passed along in the environment to the child processes, i guess
Nov 25 15:41:04 <kergoth> hmm
Nov 25 15:41:12 * Crofton|work thinks most people disagreed with him on this :)
Nov 25 15:41:35 <RP> kergoth: I want to access it in the commit hooks themselves
Nov 25 15:42:10 <RP> Crofton: anyone who wants ro run per user repositories is welcome to. I'm not becomming an admin though and will not do that
Nov 25 15:43:00 <Crofton|work> yeah, I think the final conclusion was exactly that. Those that do the admin make the decisions
Nov 25 15:48:54 <kergoth> RP: from the looks of it, exporting the user as an env var would be a oneliner in gitosis/serve.py
Nov 25 15:49:02 <kergoth> hmmm
Nov 25 15:49:06 <kgilmer|zzz> mornin
Nov 25 15:49:31 * kergoth checks to see if git-shell filters the env
Nov 25 15:49:50 <RP> kergoth: yes. There is a policy for the OE server of only using released packages though
Nov 25 15:49:58 <RP> kergoth: So this would have to be upstream first
Nov 25 15:50:22 * RP has sent an email to the author of gitosis asking questions
Nov 25 15:50:41 <kergoth> it must be a common request, access control by branch
Nov 25 15:51:01 <RP> and the hooks are the perfect place to do it
Nov 25 15:51:05 <kergoth> yep
Nov 25 15:51:42 <kgilmer|zzz> RP svn.o-hand.com is down?
Nov 25 15:55:39 <kergoth> RP: http://gist.github.com/242252 - http://kergoth.pastey.net/129429
Nov 25 16:00:39 <jconnolly> anyone know of any smartcard related recipes/packages in OE?
Nov 25 16:00:45 <hrw> re mickeyl 
Nov 25 16:01:11 <mickeyl> hi 
Nov 25 16:01:14 <kergoth> RP: though, http://kergoth.pastey.net/129431 is probably a better example, since by default log-origins shows diffs in the output
Nov 25 16:13:17 <JaMa|Gone> woglinde: hi
Nov 25 16:15:06 <chouimat> morning
Nov 25 16:24:28 <kergoth> hrm, the only real issue i see with this is the case i mentioned earlier, where one commit consists of multiple upstream commits, or vice versa, and it marks the thing as already in both sides, because it can't know if parts are missing, or were just removed to work around conflicts
Nov 25 16:24:40 <kergoth> ah well, guess the final diff to sanity check is needed indeed
Nov 25 16:24:42 * kergoth talks to himself
Nov 25 16:27:04 <CIA-80> 03Klaus Kurzmann <mok@fluxnetz.de> 07org.openembedded.dev * r932a9ec2a8 10openembedded.git/conf/distro/include/sane-srcrevs-fso.inc: 
Nov 25 16:27:04 <CIA-80> sane-srcrevs-fso.inc: bump fsodeviced and others to get new alsa stuff
Nov 25 16:27:04 <CIA-80> Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>
Nov 25 16:28:20 <mickeyl> mrmoku: thanks
Nov 25 16:30:23 <rvfo_> Help, I tried to see the dependency tree using "bitbake -g task-base" and "dot2gxl task-dependency.dot", but when i do this command i get a segmentation fault, Anyone knows why? tanks
Nov 25 16:33:20 <hrw> rvfo_: I use grep and less to view that file
Nov 25 16:33:38 <tobi_> hi, are there any qt/embedded cracks around here who are familiar with the gumstix overo series? :)
Nov 25 16:33:44 <rvfo_> hrw: thanks
Nov 25 16:35:44 <rvfo_> hrw: Do you know where can I get recipes and examples to compile a task-base without x11 and gtk++, I remove the lines PREFFERED_PROVIDERD form minimal distro. but doesn't work
Nov 25 16:35:51 <kergoth> i wish git was slightly smarter about copies/renames :\
Nov 25 16:35:59 <kergoth> particularly wrt rev-list, though i see why it isnt
Nov 25 16:36:09 * kergoth has to do git log-origins HEAD oe/org.openembedded.dev -- recipes/ncurses packages/ncurses
Nov 25 16:37:11 <hrw> rvfo_: those libs will not end in your rootfs
Nov 25 16:37:28 <hrw> something depends on them - I suspect bluez4 (as usual)
Nov 25 16:37:47 <hrw> rvfo_: remove bluetooth from MACHINE_FEATURES of your target device config and then do build
Nov 25 16:39:33 <rvfo_> hrw: I'm trying to remove because I get a buil fail all time in the libxcb_1.0.bb recipe, an unable parser
Nov 25 16:40:07 <mrmoku> mickeyl: np :)
Nov 25 16:40:07 <rvfo_> hrw: This is the error  make[1]: *** [xproto.c] Error 6
Nov 25 16:43:44 <rvfo_> hrw: i don't know what to do more, because my MACHINE_FEATURES only have: "kernel26 ext2 vfat keyboard pci serial usbgadget usbhost uboot"
Nov 25 16:43:55 <rvfo_> hrw; may you help me, please?
Nov 25 16:46:51 <hrw> rvfo_: no idea what is wrong
Nov 25 16:47:13 <Goddchen> hey everybody
Nov 25 16:47:35 <Goddchen> i do have a question on a kernel module
Nov 25 16:47:55 <rvfo_> hrw: thanks, i will try again and search to solve, thanks a lot
Nov 25 16:48:14 <Goddchen> i want to include openvpn to my image. openvpn recipe compiles just fine, but when oe wants to add it to the rootfs (do_rootfs) i argues about a missing kernel-module-tun
Nov 25 16:48:30 <Goddchen> but i'm completely lost on how to solve that dependency :(
Nov 25 16:50:28 <Goddchen> i tried "bitbake kernel-module-tun" but there is no such thing, then i just enabled tun as module in my kernel config via menuconfig and recompiled the kernel, but kernel-module-tun is not generated at all :(
Nov 25 16:51:59 <rsalveti> Goddchen: hm, generally the kernel modules are generated by the linux recipe that you build
Nov 25 16:52:06 <rvfo_> hrw: SOLVED the problem with libxcb_1.0.bb - SOLUTION: I forgot to install the xcb-proto in the host machine
Nov 25 16:52:27 <rsalveti> Goddchen: check the .config at the workdir to see if the module was actually compiled
Nov 25 16:52:36 <Goddchen> ok, so where exactly do i specify what modules should be built? i use an arm machine, kernel version 2.6.28
Nov 25 16:52:59 <hrw> rvfo_: thats a bug then
Nov 25 16:53:00 <rsalveti> Goddchen: defining it at the defconfig of the kernel/machine you have
Nov 25 16:53:05 <Goddchen> 2(rsalveti2): you mean the one in tmp/work/taskit-stamp9g20-angstrom-linux-gnueabi/linux-2.6.28-r10/linux-2.6.28$, right?
Nov 25 16:53:22 <rsalveti> Goddchen: yep
Nov 25 16:53:51 <Goddchen> i activated it in recipes/linux/linux-2.6.28/at91sam9g20ek/defconfig
Nov 25 16:54:19 <Goddchen> let me try a virtual/kernel -c clean
Nov 25 16:54:22 <rsalveti> Goddchen: yep, sounds right
Nov 25 16:54:42 <rsalveti> Goddchen: try building and checking the work dir, check if the final .config is the same one you edit and if the module was compiled
Nov 25 16:54:52 * rsalveti going to get the food
Nov 25 16:55:04 <Goddchen> ok, let me see
Nov 25 16:58:42 <Goddchen> when i do "bitbake virtual/kernel" it should build the correct kernel that is configured in my config file, right?
Nov 25 17:00:32 <rsalveti> Goddchen: yep, and the provider would be just linux
Nov 25 17:00:35 <rsalveti> as it seems at your work dir
Nov 25 17:01:03 <Goddchen> ok i'm doing "bitbake virtual/linux" atm
Nov 25 17:09:23 <mwester> re
Nov 25 17:09:35 <woglinde_> jo  mwester
Nov 25 17:10:12 <mwester> urgh.  oe ml rejected my email.  I am persona non grata! :D
Nov 25 17:11:12 <mike_cw> Goddchen: I think openvpn has an RDEPENDS on kernel-module-tun that should actually be an RRECOMMENDS
Nov 25 17:11:45 <mike_cw> i had a similar problem a while ago. i probably should have posted a patch
Nov 25 17:11:50 <woglinde_> mwester graylist again?
Nov 25 17:11:59 <Goddchen> yes it has a RDEPENDS
Nov 25 17:12:08 <Goddchen> but it indeed needs the tun module, doesn't it?
Nov 25 17:12:24 <mike_cw> yes but it might be compiled into the kernel
Nov 25 17:12:37 <mwester> woglinde_,  thank goodness no -- it is something stupid that *I* did this time -- I used the wrong email account to send the email.  So all is well.
Nov 25 17:12:46 <mike_cw> there is a note in the OE docs somewhere that says any recipe that RDEPENDS on a kernel module is a bug
Nov 25 17:13:23 <hrw> ~hail qt
Nov 25 17:13:31 * ibot bows down to qt and chants, "I'M NOT WORTHY!!"
Nov 25 17:14:20 * chouimat hates November
Nov 25 17:14:22 <kergoth> anyone looked at adding the bits to build kernel modules to the kernel-dev package?
Nov 25 17:14:23 <mwester> Right - the problem is that RDEPENDS causes massive problems if (for example) the tun driver is built into the kernel -- in that case there is no kernel module, so the dependency is unresolvable, even though it should work just fine.  RRECOMMENDS handles this case more elegantly.
Nov 25 17:14:27 <kergoth> headers + kbuild bits?
Nov 25 17:15:14 <pb_> still not very elegantly, though, since you get loads of "unsatisfied recommendation for..." diagnostics.  it'd be nicer if the kernel-image Provide:d kernel-module-tun in that kind of situation.
Nov 25 17:16:18 <Goddchen> yes you're right, i can be compiled into the kernel... but anyway, i just want to satisfy that dependency now to get my rootfs updated with openvpn
Nov 25 17:17:44 <mike_cw> if your kernel config builds tun as a module then you will be in business i think, but if not then changing the openvpn recipe to RRECOMMENDS fixed it for me
Nov 25 17:18:16 <mwester> pb_: that's a great idea.
Nov 25 17:20:54 <Goddchen> hm, seems like it's not the defconfig i edited that is copied to the work dir :(
Nov 25 17:21:16 <Goddchen> it sais something like:
Nov 25 17:21:17 <Goddchen> NOTE: Running task 307 of 545 (ID: 3, /home/swk/osgi/openembedded/stuff/openembedded/recipes/linux/linux_2.6.28.bb, do_unpack)
Nov 25 17:21:17 <Goddchen> NOTE: Unpacking ../sources/linux-2.6.28.tar.bz2 to tmp/work/taskit-stamp9g20-angstrom-linux-gnueabi/linux-2.6.28-r10/
Nov 25 17:21:17 <Goddchen> NOTE: Unpacking ../openembedded/recipes/linux/linux/defconfig to tmp/work/taskit-stamp9g20-angstrom-linux-gnueabi/linux-2.6.28-r10/
Nov 25 17:21:29 <Goddchen> but why does it take linux/linux/defconfig
Nov 25 17:21:43 <Goddchen> which btw is completely wrong, wrong architecture, wrong cpu...
Nov 25 17:39:07 <hrw> because your machine does not provide one
Nov 25 17:43:17 <Goddchen> why not? what is with the one under 1recipes/linux/linux-2.6.28/at91sam9g20ek/defconfig?
Nov 25 17:44:47 <Goddchen> and a more general question: if i manage to get the correct .config with tun module enabled to compile, will that resolve the "kernel-module-tun" dependency?
Nov 25 17:45:41 <DJWillis> Is there any easy way (other than -DDD and a lot of time) to work out why a bb has suddenly decided to fail with a dependency loop when it has gone unchanged for some time.
Nov 25 17:46:33 <DJWillis> Goddchen: the module will get packaged like that and it will be added to the kernel-modules meta package rdepends.
Nov 25 17:46:35 <pb_> not a general way, but if you post the .bb file here then someone might have a suggestion.
Nov 25 17:50:57 <DJWillis> pb_: Its a Linux-OMAP based kernel recipe, http://git.openpandora.org/cgi-bin/gitweb.cgi?p=openpandora.oe.git;a=blob;f=recipes/linux/omap3-pandora-kernel_2.6.27-pandora.bb;h=6a6a8248aefdb45684289d60e00001790c20c4de;hb=HEAD (it's in my overlay at the moment). Been building fine for months
Nov 25 17:51:49 <pb_> oh, right, so just a regular kernel recipe?  do other kernels have the same problem?
Nov 25 17:53:21 <XorA|gone> khem: you have returned :-)
Nov 25 17:53:24 <JaMa> DJWillis: is it the only recipe with devicetree defined inside as well as in linux.inc include?
Nov 25 17:53:48 <khem> XorA|gone: sort of. I was not feeling well for past couple of weeks
Nov 25 17:54:10 <XorA|gone> khem: sorry to here that, BTW there is a membership election going on for you currently
Nov 25 17:54:38 <khem> +1 from me :)
Nov 25 17:54:46 <JaMa> DJWillis: I know why dependeny loop exist in ./linux_2.6.23+2.6.24-rc6.bb
Nov 25 17:55:02 <XorA|gone> khem: goodo, can you send your wish to by a member of the TSC to openembedded-members
Nov 25 17:55:12 <XorA|gone> khem: just a short note will do
Nov 25 17:55:23 <XorA|gone> khem: the TSC election is also running
Nov 25 17:56:09 <Goddchen> hm, still NOTE: Unpacking ../openembedded/recipes/linux/linux/defconfig to tmp/work/taskit-stamp9g20-angstrom-linux-gnueabi/linux-2.6.28-r10/
Nov 25 17:56:37 <Goddchen> i have a local overlay and i created linux/linux-2.6.28/taskit-stamp9g20/defconfig there
Nov 25 17:56:41 <XorA|gone> khem: we will add you to the mailing list when you are voted in on saturday
Nov 25 17:56:42 <Goddchen> was that wrong?
Nov 25 17:56:47 <XorA|gone> khem: members list that is
Nov 25 17:56:55 <khem> XorA|gone: cool thanks
Nov 25 17:56:57 <DJWillis> JaMa: there is a local linux.inc in there but its a copy of the one from oe.dev
Nov 25 17:58:36 <rvfo_> The openssl source site, dosen't work, how can i change the website to a mirrors and all works fine with md5?
Nov 25 18:00:18 <Goddchen> why is oe getting the config from recipes/linux/linux/defconfig
Nov 25 18:00:19 <Goddchen> ?
Nov 25 18:00:32 <JaMa> DJWillis: check last few patches to kernel.bbclass, but I don't see anything wrong in your recipe..
Nov 25 18:01:39 <Goddchen> is there any way for me to see where oe looks for config files?
Nov 25 18:16:38 <Goddchen> no ideas?
Nov 25 18:16:59 <Goddchen> i currently try to run with -DDD but it doesn't say where it looks after the defconfig file :(
Nov 25 18:24:42 <Goddchen> even when i create recipes/linux/linux/defconfig in my local overlay it uses still the one in the oe directory :(
Nov 25 18:32:41 <hrw> bye
Nov 25 18:44:44 <kergoth> Goddchen: config files are usually pulled in via file:// urls in SRC_URI, which are found via FILESPATH, which is constructed with FILESPATHBASE, FILESPATHPKG, and OVERRIDES.
Nov 25 18:44:53 <kergoth> Goddchen: see the FILESPATH vars in conf/bitbake.conf
Nov 25 18:45:24 <Goddchen> k
Nov 25 18:46:17 <Goddchen> k i see
Nov 25 18:50:31 <rvfo_> Anyone could help me, please? I don't know what to do. I've tried to compile a simple OE image using task-base.bb. I don't know why bitbake tried to compile all xorg libraries, 2500 recipes. I've just needed a simple linux with a console. My distro and machine's conf doesn't have anything like the minimal. Please, how may I escape the xorg libraries?
Nov 25 19:02:00 <likewise> rvfo: do not use task-base.bb. I'm sorry, but it sucks when it comes to build dependencies.
Nov 25 19:12:24 <rvfo_> likewise: I tried to compile gconf-dbus_svn.bb and get a fetch error svn://developer.imendio.com/svn/gconf-dbus what may I do to solve this?
Nov 25 19:14:21 <CIA-80> 03Martin Jansa <Martin.Jansa@gmail.com> 07org.openembedded.dev * rbba98a56fb 10openembedded.git/conf/distro/ (include/preferred-shr-versions.inc shr.conf): 
Nov 25 19:14:21 <CIA-80> shr: use newer classpath and jamvm
Nov 25 19:14:21 <CIA-80> Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Nov 25 19:16:09 <Goddchen> bye
Nov 25 19:19:00 <likewise> rvfo: don't know
Nov 25 19:27:28 <woglinde_> jo florian
Nov 25 19:27:32 <woglinde_> and rschuster
Nov 25 19:27:33 <florian> re
Nov 25 19:27:40 <rschuster> moin
Nov 25 19:31:54 <florian> hi rsalveti 
Nov 25 19:31:57 <florian> eh rschuster 
Nov 25 19:32:13 <rsalveti> florian: hi
Nov 25 19:32:21 <florian> :-)
Nov 25 19:35:26 <kgilmer> RP my new build box is here and it is FAST.
Nov 25 19:35:41 <kgilmer> i should have 'build from scratch' numbers shortly
Nov 25 19:37:48 <khem> bitbake -D may be 
Nov 25 19:38:21 * khem was replying to old message from Goddchen
Nov 25 19:38:22 <pb___> kgilmer: very good
Nov 25 19:38:30 <khem> Hi pb___ 
Nov 25 19:38:32 <woglinde_> he khem
Nov 25 19:38:38 <khem> woglinde_: hello
Nov 25 19:38:38 <woglinde_> khem where were you?
Nov 25 19:38:43 <khem> sick 
Nov 25 19:38:49 <khem> still recovering
Nov 25 19:39:02 <woglinde_> oh :(
Nov 25 19:39:15 <woglinde_> khem I fixed a parallel make install bug in uClibc ;)
Nov 25 19:39:25 <khem> woglinde_: very cool
Nov 25 19:39:35 <woglinde_> now we have libc.so in place
Nov 25 19:39:53 <kgilmer> pb_, hey i've been working on eclipse docs on the wiki
Nov 25 19:39:54 <khem> nice I am behind on reading my mails
Nov 25 19:39:59 <kgilmer> still a WIP
Nov 25 19:40:22 <kgilmer> i have the "install OE" section done
Nov 25 19:43:12 <kgilmer> watching this machine run, it seems like untaring source packages is a serial bottleneck
Nov 25 19:43:33 <kgilmer> all cpus go idle while something is being uncompressed
Nov 25 19:43:37 <kgilmer> except for one
Nov 25 19:48:03 <khem> woglinde_: RUNTIME_PREFIX points to <install_root>/usr 
Nov 25 19:48:20 <woglinde_> khem uh
Nov 25 19:48:24 <woglinde_> let mee see
Nov 25 19:48:44 <khem> now ldd is installed into $(PREFIX)$(RUNTIME_PREFIX)bin/ldd
Nov 25 19:49:15 <woglinde_> yeah
Nov 25 19:49:16 <khem> I thought runtime prefix pointed to / of root file system
Nov 25 19:49:20 <khem> but I could be wrong
Nov 25 19:49:34 <woglinde_> hm thats okay
Nov 25 19:49:38 <woglinde_> in our case
Nov 25 19:49:49 <woglinde_> so I gave it not so much toughts
Nov 25 19:50:28 <khem> hmm we invoke make with RUNTIME_PREFIX=/
Nov 25 19:50:44 <khem> now
Nov 25 19:51:20 <khem> Does ldd get installed into /usr/bin or just /bin now ?
Nov 25 19:53:23 <woglinde_> bin for micro
Nov 25 19:53:29 <woglinde_> /usr/bin for angstroem
Nov 25 19:53:32 <woglinde_> hm
Nov 25 19:53:35 <woglinde_> I hope
Nov 25 19:53:40 <woglinde_> didnt test it for sure
Nov 25 19:54:08 <woglinde_> but you can try it
Nov 25 19:54:10 <woglinde_> *g*
Nov 25 19:54:46 <khem> oh thats because of $(PREFIX)
Nov 25 19:54:50 <khem> got it
Nov 25 19:55:28 <khem> so I think PREFIX=${D}
Nov 25 19:55:39 <mickeyl> who is maintaining linux-libc-headers?
Nov 25 19:55:54 <khem> mickeyl: I do sometimes
Nov 25 19:56:18 <mickeyl> khem: who do i need to bribe to get linux/rfkill.h in there?
Nov 25 19:56:22 <mickeyl> welcome back, btw.
Nov 25 19:57:08 <mickeyl> hmm, i'm afraid it's an oversight upstream
Nov 25 19:57:16 <signal11> hi, I own a computer
Nov 25 19:57:25 <signal11> if I find that a recipe does not build, what are the first steps I should take?
Nov 25 19:58:24 <woglinde_> signal11 fix it
Nov 25 19:58:33 <signal11> woglinde_: thanks, that's helpful (:
Nov 25 19:58:41 <khem> mickeyl: thx. we use install headers now a days
Nov 25 19:59:01 <woglinde_> signal11 hm second step find someone that fixes it for you
Nov 25 19:59:06 <signal11> woglinde_: it seems like the source is broken, but others have compiled it for other archs sucessfully
Nov 25 19:59:06 <khem> mickeyl: Did you make a case on lkml ?
Nov 25 19:59:25 <mickeyl> no, i wasn't even sure whether it's an oversight on our side or not
Nov 25 19:59:28 <signal11> libdvb is the recipe I'm having problems with
Nov 25 19:59:36 <woglinde_> signal11 *g* just say which recipe breaks for which arch
Nov 25 19:59:38 <woglinde_> ah
Nov 25 19:59:45 <signal11> for beagleboard
Nov 25 19:59:56 <woglinde_> I guess we will have a really old libdvb
Nov 25 20:00:01 <woglinde_> but let me see
Nov 25 20:00:11 <khem> mickeyl: we dont babysit the headers that much with newer kernels we use the make target thats maintained by kernel guys
Nov 25 20:00:13 <signal11> http://tinderbox.openembedded.net/packages/libdvb/
Nov 25 20:00:24 <signal11> it's succeeded on other archs at least, which is puzzling
Nov 25 20:00:30 <signal11> I'll run it again and get the error
Nov 25 20:00:31 <woglinde_> hm okay latest available
Nov 25 20:00:40 <mickeyl> khem: ok, that means for future kernels i need to bother upstream, for now i will patch linux-libc-headers in OE
Nov 25 20:00:57 <woglinde_> signal11 hm I see no fail
Nov 25 20:01:00 <signal11> yeah
Nov 25 20:01:09 <signal11> that's what's puzzling
Nov 25 20:01:27 <khem> mickeyl: yeah its good otherwise we may export headers which should not and become incompatible
Nov 25 20:01:29 <woglinde_> the pastebin the error from the log file
Nov 25 20:01:46 <signal11> I will in a bit once I rerun it
Nov 25 20:01:51 <mickeyl> khem: well, rfkill.h says pretty much it's for userspace only
Nov 25 20:01:52 <khem> mickeyl: which kernel ver is it ?
Nov 25 20:02:21 <mickeyl> i think it's missing since a couple of versions
Nov 25 20:02:23 <khem> mickeyl: OK seems that it could be an oversight upstream then
Nov 25 20:02:28 <mickeyl> but at least 2.6.30 and 2.6.31 should have it
Nov 25 20:03:58 <khem> http://www.mjmwired.net/kernel/Documentation/rfkill.txt convinces me that it should be exported
Nov 25 20:04:56 <woglinde> hms
Nov 25 20:05:04 <woglinde> re
Nov 25 20:06:41 <mickeyl> khem: yep
Nov 25 20:06:49 <mickeyl> rfkill.h even more so
Nov 25 20:07:11 <mickeyl> | /* and that's all userspace gets */
Nov 25 20:07:16 <mickeyl> at the bottom
Nov 25 20:07:26 <woglinde> re signal
Nov 25 20:07:31 <khem> yeah they even relicenced it non GPL for same
Nov 25 20:07:35 <signal11> I think I ran out of HDD space again! (not related to the error I had earlier with this)
Nov 25 20:07:48 <mickeyl> khem: do you know how to best patch linux-libc-headers to include it?
Nov 25 20:08:06 <mickeyl> i can't see how the "good" headers are marked
Nov 25 20:08:08 <likewise> hi khem, long time no see
Nov 25 20:08:19 <signal11_> woglinde: yeah, I missed whatever you said
Nov 25 20:08:41 <khem> likewise: was sick and took foot of the pedal :)
Nov 25 20:09:56 <likewise> How can I convince my binutils-cross to not be fortified.Adding "-U_FORTIFY_SOURCE" to CFLAGS_FOR_BUILD in the binutils-cross should be enough, right?
Nov 25 20:10:14 <woglinde> signal I was out too
Nov 25 20:10:33 <signal11> oh
Nov 25 20:10:36 <likewise> khem: welcome back, good to hear it's past you now
Nov 25 20:10:51 <woglinde> khem swine flue?
Nov 25 20:11:08 <khem> mickeyl: add unifdef-y += rfkill.h in include/linux/Kbuild
Nov 25 20:11:30 <khem> woglinde: may be they do not test for it in hospital anymore
Nov 25 20:11:40 <mickeyl> khem: awesome, thanks
Nov 25 20:11:46 <khem> but I had more problems than that
Nov 25 20:11:53 <woglinde> args
Nov 25 20:12:04 <woglinde> libtool installs stuff under /usr/include
Nov 25 20:12:12 <woglinde> for micro
Nov 25 20:12:29 <khem> woglinde: and aclocal too for micro
Nov 25 20:12:53 <khem> woglinde: right now when you change layout that is different than build systems there are few culprits 
Nov 25 20:13:09 <mickeyl> hmm
Nov 25 20:16:32 <woglinde> khem I meant aclocal stuff
Nov 25 20:16:58 <khem> woglinde: yes
Nov 25 20:17:10 <khem> thats what I got stuck at when I was trying big things with micro
Nov 25 20:17:20 <khem> like build native sdk
Nov 25 20:20:21 <woglinde> hm libtool isnt autotools bases self?
Nov 25 20:20:53 <khem> likewise: -U_FORTIFY_SOURCE in CFLAGS should be enough yes
Nov 25 20:20:54 <nik0n> is there any reason why stable an dminimal distribution still uses tinylogin by default?
Nov 25 20:21:01 <nik0n> i thought tinylogin is dead
Nov 25 20:21:35 <woglinde> nik0n discussion was about
Nov 25 20:21:45 <woglinde> that you have to set busybos suid
Nov 25 20:21:55 <woglinde> you dont want this in all cases
Nov 25 20:23:03 <nik0n> i must lie but i think my busybox is not suid root
Nov 25 20:23:18 <nik0n> i have to verify that tomorrow
Nov 25 20:23:41 <woglinde> nik0n ask pb_ about it
Nov 25 20:23:57 <likewise> khem: thanks, I added them to binutils.bb, not binutils-cross.bb it seems. rebuilding now.
Nov 25 20:23:59 <khem> likewise: you could also add to EXTRA_OEMAKE += "CFLAGS=-U_FORTIFY_SOURCE"
Nov 25 20:24:21 <khem> likewise: what is you build system OS ?
Nov 25 20:24:36 <khem> seems like old ubuntu
Nov 25 20:24:37 <likewise> khem: I'm adding the nios2 toolchain.
Nov 25 20:24:40 <nik0n> woglinde kthx
Nov 25 20:24:45 <khem> likewise: I see
Nov 25 20:24:59 <khem> likewise: are you using latest versions of gcc and binutils
Nov 25 20:25:15 <likewise> khem: nios2 toolchains are old (even the "new" one by Wind River)
Nov 25 20:25:25 <khem> likewise: hmmm ok
Nov 25 20:25:47 <khem> likewise: binutils and gcc were fixed for FORTIFYing
Nov 25 20:25:53 <khem> may be the versions you have are older than that
Nov 25 20:26:05 <signal11> woglinde: http://pastebin.com/m7554ef29
Nov 25 20:26:37 <likewise> khem: binutils-cross for nios2 is 2.15 and gcc-cross for nios2 is 3.4.6
Nov 25 20:26:57 <likewise> khem: I could indeed look at the fixes.
Nov 25 20:27:01 <khem> yuck ok
Nov 25 20:27:25 <khem> likewise: you might want to use old gcc on your build box 
Nov 25 20:27:27 <dies_felices> re
Nov 25 20:27:46 <likewise> khem: there is a newer (Wind River) toolchain but they packaged up the sources their own way.
Nov 25 20:28:05 <likewise> khem: I'ld rather disable FORTIFY, then newer toolchains should cope.
Nov 25 20:28:14 <khem> thats right
Nov 25 20:28:24 <likewise> khem: ever worked with nios2?
Nov 25 20:28:40 <khem> likewise: no not seriously
Nov 25 20:28:48 <khem> but I read about it just for fun
Nov 25 20:28:58 <khem> and some uclibc reviews
Nov 25 20:29:04 <likewise> khem: ok
Nov 25 20:29:38 <khem> likewise: are you porting the nios patches to one of existing versions of gcc and binutils ?
Nov 25 20:29:41 <khem> or adding all new
Nov 25 20:29:48 <likewise> khem: nios2/oe will be extra pain in the butt, because even linux-libc-header staging will depend on the soft-programmable SoC architecture.
Nov 25 20:30:07 <khem> hmm
Nov 25 20:30:20 <likewise> khem: I'm just plugging the old (mmu-less) toolchain from nioswiki GIT into OE.
Nov 25 20:30:22 <khem> oh yeah extensa has similar issues
Nov 25 20:30:49 <likewise> khem: I wish I could dissect the Wind River toolchain to: upstream toolchain + additional patches.
Nov 25 20:30:54 <likewise> khem: and then add those patches to OE.
Nov 25 20:31:13 <likewise> khem: that's what I did with the PowerPC e200/e300 toolchain I added to OE.
Nov 25 20:31:30 <khem> signal11: hmm it seems string.h is not included
Nov 25 20:31:45 <signal11> odd
Nov 25 20:31:53 <khem> likewise: that would be ideal
Nov 25 20:31:58 <signal11> I mean, I just extracted the tarball and compiled natively, you know, normally
Nov 25 20:31:59 <signal11> and it worked
Nov 25 20:32:06 <likewise> khem: and, keep your foot near the brakes - don't go sick on us this winter! 
Nov 25 20:32:11 <likewise> :-)
Nov 25 20:32:19 <signal11> I mean, not using bitbake, openembedded, etc... like normal world compiling
Nov 25 20:32:22 <khem> signal11: error: ‘memset’ was not declared in this scope
Nov 25 20:32:23 <likewise> khem: the world needs you
Nov 25 20:32:31 <signal11> hmm
Nov 25 20:32:53 <signal11> I see
Nov 25 20:33:20 <khem> likewise: yeah I am better now. last week was horrible one of my teammates had a stroke and passed away
Nov 25 20:33:46 <woglinde> ah
Nov 25 20:33:50 <woglinde> thats easy to fix
Nov 25 20:33:59 <woglinde> #include <cstdio>
Nov 25 20:34:09 <woglinde> or #include <cstring>
Nov 25 20:34:16 <signal11> yeah, but it shouldn'tbe broken
Nov 25 20:34:17 <khem> for c++ and #include <string.h>
Nov 25 20:34:20 <khem> for c
Nov 25 20:34:23 <signal11> I'm trying to see if the included patch is breaking it
Nov 25 20:34:27 <signal11> since it compiles fine natively
Nov 25 20:34:31 <signal11> from the tarball
Nov 25 20:34:54 <likewise> khem: my condolences
Nov 25 20:35:20 <signal11> woglinde: what's bothering me is that it's somehow building for other archs/distros
Nov 25 20:35:28 <khem> likewise: thx it was a shock because he was hale and hearty he was visiting us from australia
Nov 25 20:36:29 <khem> good guy was not connected to a power grid dwelled upon 120 acres of land and ran house on solar power
Nov 25 20:36:52 <khem> bought a distrilling machine to produce ethenol to run his car 
Nov 25 20:37:01 <khem> 0 carbon footprint 
Nov 25 20:37:01 <signal11> is that the only thing he used it for?
Nov 25 20:37:18 <khem> I dont know
Nov 25 20:37:23 <signal11> :P
Nov 25 20:37:28 <likewise> khem: wow. he left a family?
Nov 25 20:37:55 <khem> likewise: yes he is survived by his wife and two children 
Nov 25 20:38:01 <khem> kids in college
Nov 25 20:38:33 <signal11> it works with the patch
Nov 25 20:38:42 <signal11> so why is it breaking in bitbake, but not otherwise
Nov 25 20:38:45 <signal11> and not for some other archs?
Nov 25 20:38:51 <signal11> gcc version?
Nov 25 20:39:25 <likewise> khem: must be real bad to hear your husband/dad died far away on a biz trip, this way.... Makes you think.   Could it be diagnosed to trombosis or something?
Nov 25 20:40:37 <khem> yes I bet its hard on the family
Nov 25 20:45:18 <dies_felices> I thought I give devshell a try but the C compiler can not compile executables.
Nov 25 20:45:31 <dies_felices> When I run ./configure
Nov 25 20:46:05 <signal11> http://bugs.gentoo.org/show_bug.cgi?id=220165
Nov 25 20:46:05 <signal11> ahhhhhhhhhh
Nov 25 20:47:03 <khem> signal11: someone already ran into it eh ?
Nov 25 20:47:11 <signal11> khem: apparently, on gentoo, over a year ago
Nov 25 20:47:14 <khem> is it fixed ?
Nov 25 20:47:20 <signal11> I was searching for the wrong stuff earlier
Nov 25 20:47:23 <signal11> yes, with that patch
Nov 25 20:47:34 <signal11> I guess that needs to be added to the recipe tree thing?
Nov 25 20:47:43 <khem> yes sure post a patch
Nov 25 20:48:36 <woglinde> yeah
Nov 25 20:48:46 <woglinde> signall I will check it in later
Nov 25 20:48:51 <woglinde> thanks for spotting
Nov 25 20:49:36 <signal11> where would I post a patch?
Nov 25 20:49:51 <woglinde> signal11 as I said I will do it
Nov 25 20:49:55 <signal11> oh, thanks
Nov 25 20:49:56 <khem> openembedded-devel
Nov 25 20:50:07 <signal11> oh, mailing list?
Nov 25 20:50:15 <signal11> I guess the other sucessfull builds used a different gcc
Nov 25 20:50:18 <signal11> version
Nov 25 20:50:25 <signal11> I hate gcc version differences!
Nov 25 20:50:35 <khem> yeah
Nov 25 20:51:18 <woglinde> hm why?
Nov 25 20:51:24 <woglinde> forgetting the header
Nov 25 20:51:30 <woglinde> isnt a compiler fault
Nov 25 20:51:41 <signal11> that's true
Nov 25 20:51:41 <woglinde> the compiler fault was not complaining about it
Nov 25 20:51:43 <signal11> but, how did it work on other versions of the compiler
Nov 25 20:51:44 <signal11> yes
Nov 25 20:51:51 <woglinde> in earlier versions
Nov 25 20:51:55 <signal11> it should've always complained :P
Nov 25 20:52:01 <woglinde> dodnt know
Nov 25 20:54:13 <khem> older compilers includes string.h indirectly
Nov 25 20:54:37 <signal11> hmm
Nov 25 21:00:36 <dies_felices> Am going bye.
Nov 25 21:00:48 <woglinde> bye dies_felices
Nov 25 21:17:36 <woglinde> so
Nov 25 21:17:49 <woglinde> updates for ogg vorbis and theora done
Nov 25 21:18:01 <woglinde> now I have to test it live
Nov 25 21:30:49 <DJWillis> woglinde: Theora 1.1?
Nov 25 21:33:16 <woglinde> hm swype
Nov 25 21:33:21 <woglinde> djwillis jupp
Nov 25 21:34:23 <DJWillis> woglinde: Danke
Nov 25 21:34:35 <woglinde> but I have to test it
Nov 25 21:34:40 <woglinde> he where do you know german
Nov 25 21:34:57 <woglinde> hm swype is intressting
Nov 25 21:35:04 <woglinde> seems the stole a bit from raster
Nov 25 21:35:15 <woglinde> guess I should talk with him
Nov 25 21:39:08 <DJWillis> woglinde: my German sucks, it really does. Spent time near Stuttgart as a child but please, thanks and yes are about my limits ;-).
Nov 25 21:39:27 <woglinde> hehe
Nov 25 21:39:35 <woglinde> really wonder about this channel
Nov 25 21:39:41 <woglinde> khem was in ulm
Nov 25 21:39:43 <woglinde> or near
Nov 25 21:40:32 <khem> woglinde: I was in ulm, am Donau
Nov 25 21:40:39 <DJWillis> I think an OE map of devs would be interesting to look at.
Nov 25 21:41:14 * khem has docs appt in an hour
Nov 25 21:42:27 <woglinde> hms we are 5 versions behind with libpng
Nov 25 21:42:29 <woglinde> *sigh*
Nov 25 21:43:52 <DJWillis> woglinde: do you want to pick that up, it was on my TODO as we want a newer build for the Pandora, same with a chuck of the other media libs/SDL etc.
Nov 25 21:45:30 <woglinde> hm which sdllibs are behind?
Nov 25 21:45:34 <woglinde> wasnt this much
Nov 25 21:46:08 <DJWillis> woglinde: I can't recall, have it in my notepad somewhere (i.e. not with me and my laptop ;-))
Nov 25 21:47:33 <woglinde> png should be easy
Nov 25 21:48:24 <woglinde> hm
Nov 25 21:48:48 <woglinde> newer libiconv was on my list to
Nov 25 21:50:10 <DJWillis> woglinde: what are you doing all this cleanup work for? just as it needs to be done or do you have some motivation for something cool.
Nov 25 21:51:58 <woglinde> hm
Nov 25 21:52:03 <woglinde> dont know
Nov 25 21:52:11 <woglinde> I saw this and I thought
Nov 25 21:52:12 <woglinde> I can fix this
Nov 25 21:52:21 <woglinde> so I did
Nov 25 21:53:24 <DJWillis> woglinde: I seem to find a lot of that, that's why the PAM and Xfce stuff got done (not that it's all there yet ;-))
Nov 25 21:56:26 <woglinde> *g*
Nov 25 21:56:34 <woglinde> send me a pandora
Nov 25 21:56:34 <woglinde> *g*
Nov 25 22:06:36 <RP> kgilmer: Pleased its doing the job :)
Nov 25 22:06:49 <RP> kergoth: Sorry, I wandered off earlier. Those scripts look nice 
Nov 25 22:07:36 <RP> khem: Welcome back :)
Nov 25 22:10:28 <DJWillis> woglinde: if I had anything that was not a devboard I would ;-).
Nov 25 22:10:37 <woglinde> *g*
Nov 25 22:10:38 <woglinde> okay
Nov 25 22:21:58 <likewise> Guys, where is UCLIBC_STAGE_PREFIX defined?     grep -rn -e 'UCLIBC_STAGE_PREFIX.*=.*' openembedded.git   does give me zip.
Nov 25 22:22:15 <RP> likewise: It used to be in the uclibc recipes
Nov 25 22:22:25 * likewise looks for a typo
Nov 25 22:22:49 <RP> likewise: It was recently removed
Nov 25 22:23:12 <likewise> RP: ah, but it's being used all over the place and breaks my current build
Nov 25 22:23:25 <RP> likewise: Where is it used?
Nov 25 22:23:49 <RP> likewise: Hmm, I suspect its used in older uclibc versions that have not been converted :/
Nov 25 22:24:31 <likewise> RP: indeed :-)
Nov 25 22:25:15 <RP> likewise: Do you really need the old uclibc version? If yes, I'd suggest trying the changes against it
Nov 25 22:26:00 <RP> likewise: which version are you using?
Nov 25 22:26:29 <likewise> RP: I'm dragging in a 2.15 / 3.4.6 / 0.9.27 -ish toolchain for the new arch "nios2" I'm bringing to .dev.
Nov 25 22:27:15 <likewise> RP: np, I will manage. Just wondered if the variable was supposed to go.
Nov 25 22:28:04 <RP> likewise: Well, its bad we have these dangling variables. I'll see if I can clean that up
Nov 25 22:28:08 <RP> woglinde: ^^^ :/
Nov 25 22:29:12 <woglinde> hms
Nov 25 22:29:32 <woglinde> yeah I had bad feeling saying likewise working wth 0.9.27
Nov 25 22:29:34 <woglinde> args seeing
Nov 25 22:30:22 <RP> woglinde: I can backport the changes to those versions, not sure if they'll work though?
Nov 25 22:30:44 <woglinde> hm sorry I have no motivation to support this old version
Nov 25 22:31:39 <RP> woglinde: right, I understand
Nov 25 22:31:45 <woglinde> hm
Nov 25 22:32:01 <woglinde> even git recipe has it
Nov 25 22:32:13 <woglinde> maybee a task for khem to clean it up
Nov 25 22:32:37 <likewise> woglinde: but a heads-up on the ml would have been good, as it is rather core (or I just missed it). How's the changing commit called?
Nov 25 22:33:29 <woglinde> likewise hm
Nov 25 22:33:50 <woglinde> likewise sorry didnt believe there is someone who uses this old stuff
Nov 25 22:33:51 <CIA-80> 03Klaus Kurzmann <mok@fluxnetz.de> 07org.openembedded.dev * r5720be48bd 10openembedded.git/recipes/freesmartphone/ (fsodeviced/fsodeviced fsodeviced_git.bb): 
Nov 25 22:33:51 <CIA-80> fsodeviced: adjust path in initscript - fsodeviced moved to /usr/sbin
Nov 25 22:33:51 <CIA-80> Signed-off-by: Klaus Kurzmann <mok@fluxnetz.de>
Nov 25 22:34:22 <woglinde> f56a544c49284c258c5c46520752b7be0c0f2379
Nov 25 22:34:29 <woglinde> qgit rockz
Nov 25 22:34:30 <likewise> woglinde: thanks
Nov 25 22:35:15 <likewise> woglinde: if we don't use it we should remove it. However, it's hard to tell what gets used.
Nov 25 22:35:30 <woglinde> yeah
Nov 25 22:36:00 <woglinde> hm seems any a task to unify initial
Nov 25 22:36:25 <likewise> woglinde: you call this old version? recipes/uclibc/uclibc-initial_git.bb :-)
Nov 25 22:36:27 <woglinde> wonder now why khem didnt do it
Nov 25 22:36:36 <woglinde> yeah sorry
Nov 25 22:36:43 <likewise> he was sick, he has an excuse :-)
Nov 25 22:36:46 <RP> I'll push an untested fix
Nov 25 22:36:56 <woglinde> I thought it breaks your stuff
Nov 25 22:38:42 <RP> woglinde: by fix, I mean converting it all to use do_install
Nov 25 22:39:00 <woglinde> yeah
Nov 25 22:39:03 <woglinde> push it
Nov 25 22:39:29 <woglinde>  args
Nov 25 22:39:35 <woglinde> this preffered xorg stuff
Nov 25 22:39:35 <CIA-80> 03Richard Purdie <rpurdie@linux.intel.com> 07org.openembedded.dev * r77ec121b1b 10openembedded.git/recipes/uclibc/ (6 files): 
Nov 25 22:39:35 <CIA-80> uclibc-initial: Convert to us do_install for staging. 0.9.30.1 is tested and works, anyone using older versions will need to verify this
Nov 25 22:39:35 <CIA-80> Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
Nov 25 22:44:14 <likewise> woglinde: btw, what targets do you build uclibc against (out of interest)?
Nov 25 22:47:06 <woglinde> simpad and beagleboard now
Nov 25 23:07:19 <likewise> the nice thing about working in a branch is that my screw-ups do not appear on this IRC channel.
Nov 25 23:07:32 <likewise> hi rwhitby 
Nov 25 23:09:33 <woglinde> likewise?
Nov 25 23:09:37 <woglinde> hm is cia fixed now?
Nov 25 23:09:46 <woglinde> mailinglist seems not
Nov 25 23:10:10 <Crofton|work> likewise,  :)
Nov 25 23:10:14 <likewise> don't know, did cia report branches?
Nov 25 23:10:23 <Crofton|work> it has in the past
Nov 25 23:10:32 <likewise> ah, then the black-out is my luck :-)
Nov 25 23:10:46 <kergoth> its been talked about, limiting it to master.. don't know if it happened or not
Nov 25 23:11:24 <woglinde> hi crofton
Nov 25 23:11:27 <woglinde> re kergoth
Nov 25 23:11:37 <likewise> I actually managed to screw up some gittery, when I picked up working on likewise/nios2 again (after rm -rf) I accidently committed to nios2 (creating another branch)... 
Nov 25 23:11:58 <kergoth> hey
Nov 25 23:12:13 * kergoth hates that we can't push fast forward commits to our private branches
Nov 25 23:12:32 <likewise> kergoth: ah, so it's not me. I never understood why.
Nov 25 23:12:40 <Crofton> hey woglinde 
Nov 25 23:12:59 <kergoth> likewise: once the hooks run, they all run under the git user, no longer know who was who, so it can't auth based on branch.
Nov 25 23:13:07 <kergoth> likewise: its all or nothing, and right now its nothing ;)
Nov 25 23:13:10 <likewise> ah ok
Nov 25 23:13:19 <kergoth> we were talking about it earlier, would be nice to fix
Nov 25 23:13:29 <kergoth> RP emailed the gitosis folks
Nov 25 23:15:17 <likewise> what would be the best way to split up my branch work as a set of patches? git format-patch will make one patch per commit, I want one patch per set of files...
Nov 25 23:15:51 <kergoth> "set of files"?
Nov 25 23:16:34 <kergoth> i think itd probably be easiest for you to just use git diff.  you can generate a diff of master vs your branch, and use filterdiff from patchutils to filter the diff by path
Nov 25 23:17:33 <kergoth> alternatively, do a squash merge of the branch to master, git reset HEAD~1, and use git add / add -p to add specific files / dirs to be committed as new commits..
Nov 25 23:29:05 <likewise> kergoth: that last options *sounds* nice, but when merging my branch against .dev I got in a conflict circle I never got out, I'm too afraid to screw-up the branch or .dev...
Nov 25 23:29:57 <likewise> kegoth: With "set of files" I rather meant "one patch per set of related diffs".
Nov 25 23:30:03 <likewise> +r
Nov 25 23:31:20 <kergoth> oh
Nov 25 23:31:27 <kergoth> you might just want to rebase -i on top of master
Nov 25 23:31:35 <kergoth> with that you can rearrange commits, merge them, skip them, ..
Nov 25 23:35:32 <ant__> morning all
Nov 25 23:37:35 <woglinde> jo ant
Nov 25 23:38:00 <ant__> hey woglinde !
Nov 25 23:38:13 <ant__> I'll have to come to Berlin soon
Nov 25 23:38:33 <ant__> JAnuary probably
Nov 25 23:40:55 <ant__> kergoth: I 'fixed' with the hammer two recipes having QA issues (.so sin non -dev)
Nov 25 23:41:14 <ant__> rewriting PACKAGES = .. is probably not the most elegant way...
Nov 25 23:41:26 <ant__> http://fr.pastebin.ca/1687454   http://fr.pastebin.ca/1687455
Nov 25 23:41:50 <JaMa> ant__: hi ant
Nov 25 23:41:53 <ant__> if these happen to just look a bit sane I'll post to the ML. 
Nov 25 23:42:15 <JaMa> ant__: I guess you would be interested http://cgit.openembedded.net/cgit.cgi/openembedded/commit/?h=shr/merge&id=25826f99372c073a5c81e07735dd62dc27b9b015
Nov 25 23:42:21 <ant__> JaMa: hay, took a breathe after ?
Nov 25 23:42:23 <ant__> ;)
Nov 25 23:42:37 <ant__> ah.. very nice
Nov 25 23:42:57 <ant__> yesterday and today I had only the Nokia Maps with me...no zaurus...
Nov 25 23:43:10 <ant__> small screen of my E51 suxx a bit!
Nov 25 23:44:43 <kergoth> and you're showing these to me why exactly?  if you want review, post them to the channel, not a random individual you pick out of the crowd.  addresing one person specifically is idiotic unless you know they're the right *specific* person to be talking to, which is not the case here
Nov 25 23:45:04 <ant__> kergoth: you happen to be one who knows
Nov 25 23:45:10 <kergoth> yes, one of many
Nov 25 23:45:19 <ant__> all the internals, almost
Nov 25 23:45:37 <kergoth> and that means you get to poke me with every random little problem you hit?
Nov 25 23:45:52 <ant__> ok ok 
Nov 25 23:46:04 <ant__> wifey angry with you today?
Nov 25 23:46:07 <ant__> raining ther?
Nov 25 23:46:10 <kergoth> i'll take a look, but its extremely presumptuous to think you can waste someone's time like that
Nov 25 23:46:32 <ant__> see, mobody cared for years about QA
Nov 25 23:46:36 <kergoth> there are many people that know how this stuff works, ask and wait for an answer
Nov 25 23:46:46 <kergoth> and again, that means you get to pester me incessently?
Nov 25 23:46:54 <ant__> I happened to read some irc chats btw you and pb
Nov 25 23:47:01 <ant__> that was enlightning
Nov 25 23:47:18 <ant__> sorry if I bother you incessently
Nov 25 23:47:33 <ant__> or if you feel like
Nov 25 23:47:45 <ant__> anyway, just forget
Nov 25 23:47:48 <ant__> thx 
Nov 25 23:48:36 <kergoth> i appreciate that you're trying to fix problems, but just ask the channel, I'm not the only one in here who can answer questions
Nov 25 23:49:12 <ant__> during EU night...almost the only one :)
Nov 25 23:49:33 <kergoth> sadly
Nov 25 23:50:21 <kergoth> the libpcre one looks wrong.  -dev includes the .so in libdir automatically, and that FILES_${PN} explicitly excludes it..
Nov 25 23:50:52 <kergoth> the other one looks like its in the right direction, but i expect you could avoid the PACKAGES change and fix it via FILES changes, but I'd really have to build the thing to check into that..
Nov 25 23:51:00 <ant__> ok, so just changing the order is enough?
Nov 25 23:51:02 * kergoth doesn't have an OE build env handy, been doing mvl6 work recently
Nov 25 23:51:26 <ant__> ah, ok, I'll try to work on that
Nov 25 23:51:46 <kergoth> changing PACKAGES shouldn't be necessary in pcre either, ${PN}'s fILES glob doesn't include ${libdir}/libpcre.so, only ${libdir}/libpcre.so.*
Nov 25 23:51:50 <kergoth> unless I'm missing something, which is possible
Nov 25 23:52:05 <ant__> you see, I was right to be ashamed to post them straight to the ML :)
Nov 25 23:52:31 <ant__> they 'happen' to work, yes, with the hammer
Nov 25 23:52:38 <ant__> no QA
Nov 25 23:53:14 <kergoth> yeah, I'm sure it fixes it, i just think we can probably avoid changing so many variables.  which file(s) are getting QA errors in libpcre?
Nov 25 23:53:29 <ant__> iirc  libpcre.so
Nov 25 23:53:41 <ant__> 1 swc
Nov 25 23:53:58 <kergoth> k
Nov 25 23:54:26 <ant__> non dev contains .so, libpcre, /work/armv5te-angstrom-linux-gnueabi/libpcre-7.6-r4/packages-split/libpcre/usr/lib/libpcre.so
Nov 25 23:55:51 <ant__> and btw there is this 'custom' do_split_packages
Nov 25 23:56:14 <ant__> first time I found it
Nov 25 23:58:50 <kergoth> interesting
Nov 25 23:58:50 <kergoth> hmm
Nov 26 00:00:06 <ant__> he..you see..that's not a trivial recipe....
Nov 26 00:01:47 <kergoth> I'd suggest checking the pkgdata, see what the FILES vars end up at the end, after the do_split_packages calls
Nov 26 00:02:05 <kergoth> cat tmp/pkgdata/runtime/libpcre, cat tmp/pkgdata/runtime/libpcre-dev
Nov 26 00:02:26 <kergoth> i expect those calls are overwriting the stock values for the 'pcre' lib, since they apply to all libs in the libdir..
Nov 26 00:03:06 * kergoth kicks off a build, will take a while though
Nov 26 00:03:41 <ant__> kergoth: thx for your time... these are the 2 LAST QA I get
Nov 26 00:03:48 <kergoth> nice
Nov 26 00:03:52 <kergoth> that's good news
Nov 26 00:03:57 <ant__> building console- opie- x11-image
Nov 26 00:04:17 <ant__> there are some about desktop files, perhaps fake
Nov 26 00:04:28 <ant__> I mean, not real QA
Nov 26 00:05:07 <ant__> the two a.m are the last two about packaging
Nov 26 00:05:25 <ant__> and this is indeeed very good news
Nov 26 00:05:28 <Crofton|work> ant__, it is a .us holiday tomorrow and Friday, so things will be really slow at this time
Nov 26 00:05:53 <ant__> np, I got a 2000km trip in 20 hours...
Nov 26 00:06:04 <ant__> I'll rest for 2 days :)
Nov 26 00:06:18 <Crofton|work> cool
Nov 26 00:06:23 <Crofton|work> where are you going?
Nov 26 00:06:52 <ant__> I went from Verona-IT to Strassbourg then to PAris
Nov 26 00:07:06 <ant__> today 12 hours ago, back...
Nov 26 00:07:13 <ant__> I'm at home since 2 beers
Nov 26 00:07:20 <ant__> er..since 30 mins...
Nov 26 00:08:16 <ant__> you see, two customers decided both we had to have year meeting today
Nov 26 00:08:28 <ant__> :/
Nov 26 00:08:51 <Crofton|work> sounds fun
Nov 26 00:09:35 <ant__> unfortunately in Europe now average is 100km/hour
Nov 26 00:09:42 <ant__> average
Nov 26 00:10:44 <ant__> he...when I saw the new installed speed traps, tutors and cameras I sold the old Alfa Romeo
Nov 26 00:11:32 <woglinde> hehe good old germany
Nov 26 00:11:46 <ant__> well, not on West side..
Nov 26 00:11:57 <ant__> Basel- Frankfurt almost all 110 
Nov 26 00:12:17 <ant__> too many Franch and Dutch ruling over there :)
Nov 26 00:12:44 <ant__> Germans don't like
Nov 26 00:12:53 <ant__> so they put speed limits :]
Nov 26 00:13:25 <ant__> (..people say...)
Nov 26 00:55:28 <ant__> yahwn....zZZz  good night
Nov 26 01:15:36 <B_Lizzard_> Should work now
Nov 26 01:20:12 <B_Lizzard_> Upload the config.php file for the forum
Nov 26 01:23:01 <B_Lizzard_> I'm going to bed, will work on the rest tomorow
Nov 26 01:23:25 <B_Lizzard_> Do you like any part of this?
Nov 26 01:25:11 <B_Lizzard_> Ooops, wrong window.
Nov 26 01:25:17 <B_Lizzard_> It *is* too late...
**** ENDING LOGGING AT Thu Nov 26 02:59:56 2009