**** BEGIN LOGGING AT Fri Apr 17 02:59:57 2009
Apr 17 04:32:27 <CIA-2> 03Angus Ainslie <nytowl@openmoko.org> 07fso/milestone5.5 * ra3ce3ea2c7 10openembedded.git/recipes/linux/ (2 files in 2 dirs): 
Apr 17 04:32:27 <CIA-2> linux-openmoko-devel : re-eanble fix-install patch as the
Apr 17 04:32:27 <CIA-2> build still fails on some machines.
Apr 17 05:08:12 <adub> i keep getting a pango modules error when trying to install packages from the x11 repository
Apr 17 05:49:10 <CIA-2> 03Carsten Haitzler (Rasterman <raster@rasterman.com> 07org.openembedded.dev * rd92391844e 10openembedded.git/conf/distro/include/sane-srcrevs.inc: efl: newer svnrev with exquisite/ecore-fb fixes and more.
Apr 17 05:49:20 <CIA-2> 03Carsten Haitzler (Rasterman <raster@rasterman.com> 07org.openembedded.dev * r212e36417a 10openembedded.git/ (36 files in 16 dirs): (log message trimmed)
Apr 17 05:49:20 <CIA-2> angstrom, exquisite, psplash, opkg, angsrtom-images, sysvinit: abstract splash
Apr 17 05:49:20 <CIA-2> this abstracts psplash to be generic. now as long as something provides a
Apr 17 05:49:20 <CIA-2> splashfuncs file that sysvinit (and other scripts) can source, and all the
Apr 17 05:49:22 <CIA-2> approproate init hooks to start the splash etc. you can use psplash,
Apr 17 05:49:24 <CIA-2> exquisitie, usplash or anything that tickles your fancy. this moves splash
Apr 17 05:49:26 <CIA-2> toa ${SPLASH} variable to include in your image (or override). the default in
Apr 17 06:32:26 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r5eb66dfe4f 10openembedded.git/recipes/angstrom/ (e-wm-config-angstrom.bb e-wm-config-angstrom/e.src): e-wm-config-angstrom: connman module is disabled in recent builds, so don't try to load it
Apr 17 06:41:13 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r93f9c2b0df 10openembedded.git/recipes/images/ (12 files): various image: weakly assign $SPLASH so it isn't unset on expansion
Apr 17 06:50:04 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * rd8947237a0 10openembedded.git/recipes/psplash/psplash-zap_1.0.bb: psplash-zap: recommend psplash
Apr 17 06:55:54 <mpoullet|work> good morning
Apr 17 07:08:04 <mckoan> good morning
Apr 17 07:11:46 <CIA-2> 03Steffen Sledz <sledz@dresearch.de> 07org.openembedded.dev * re0499fda0d 10openembedded.git/recipes/linux/ (4 files in 3 dirs): 
Apr 17 07:11:46 <CIA-2> linux-2.6.24, linux-rt-2.6.24: squashfs-lzma-2.6.24.patch imported to oe repo
Apr 17 07:11:46 <CIA-2> (former URL no longer available)
Apr 17 07:58:19 <ana> i have made some changes to defconfig and when i build a image it does not load the kernel modules. it says disagrees about the symbol struct module
Apr 17 07:58:21 <ana> why is it
Apr 17 07:58:41 <ana> should i recompile all external packages which are built as external modules? 
Apr 17 08:03:04 <woglinde> ~seen rp
Apr 17 08:03:20 <ibot> rp is currently on #maemo (5h 9m 13s) #handhelds (5h 9m 13s) #oe (5h 9m 13s) #openmoko (5h 9m 13s), last said: 'ant_work: It was part of the sharp 2.4 junk'.
Apr 17 08:04:11 <ana> hello woglinde: you have any idea on this subject? do i need to recompile external packages which produce modules if i change the defconfig?
Apr 17 08:04:29 <woglinde> ana depends
Apr 17 08:04:45 <ana> what it depends on
Apr 17 08:04:50 <woglinde> but in newer kernels you have this module-magic-hash
Apr 17 08:05:08 <woglinde> and this can change when you rework defconfig
Apr 17 08:05:24 <ana> do i need to enable that option?
Apr 17 08:05:27 <woglinde> and than the kernel may refuse to load your compiled modules
Apr 17 08:05:32 <woglinde> nope
Apr 17 08:05:41 <woglinde> and thats why I said it depends
Apr 17 08:05:45 <ana> so what should i do?
Apr 17 08:06:00 <woglinde> hm???
Apr 17 08:06:06 <ana> if i want the earlier compiled modules to work with new kernel with changed defconfig
Apr 17 08:06:16 <woglinde> try
Apr 17 08:06:20 <woglinde> and when it dont work
Apr 17 08:06:28 <woglinde> you have to recompile the modules
Apr 17 08:07:17 <ana> i dont find this option in the kernel
Apr 17 08:08:17 <ana> woglinde: i use 2628 and i dont find this option in the kernel
Apr 17 08:08:27 <woglinde> which option?
Apr 17 08:08:38 <ana> module-magic-has
Apr 17 08:08:46 <ana> module-magic-hash
Apr 17 08:09:21 <woglinde> its
Apr 17 08:09:22 <woglinde> CONFIG_MODULE_SRCVERSION_ALL=y
Apr 17 08:09:22 <woglinde> CONFIG_MODULE_UNLOAD=y
Apr 17 08:09:22 <woglinde> CONFIG_MODVERSIONS=y
Apr 17 08:09:39 <woglinde> hm okay I see
Apr 17 08:09:45 <woglinde> on ubuntu CONFIG_MODULE_FORCE_LOAD
Apr 17 08:09:47 <woglinde> is disabled
Apr 17 08:09:53 <woglinde> you enable it
Apr 17 08:10:07 <ana> so these options need to disabled to make it work?
Apr 17 08:10:28 <woglinde> and then the out of tree modules might be able to work on kernel with newer defconfig
Apr 17 08:11:19 <ana> i have them enabled. do i need to disbaled it
Apr 17 08:11:31 <ana> http://pastebin.ca/1394744
Apr 17 08:11:50 <woglinde> *sigh*
Apr 17 08:12:03 <woglinde> please read the Documentation on this option
Apr 17 08:12:13 <ana> i have them enabled but it does not work
Apr 17 08:12:17 <ana> so i asked you
Apr 17 08:12:31 <woglinde> did you try insmod -f ?
Apr 17 08:14:20 <gremlin[it]> ping hrw|gone 
Apr 17 08:14:36 <woglinde> hi gremlin
Apr 17 08:18:23 <BusError> woglinde, FYI I fixed the directfb problem I had on arm on stable/2009 by upgrading to 1.3 upstream. If you want the patch, I got it on a branch
Apr 17 08:18:49 <ana> woglinde: i tried giving a insmod -f i get a message invalid module format
Apr 17 08:25:09 <gremlin[it]> hi woglinde 
Apr 17 08:25:13 <woglinde> args
Apr 17 08:25:36 <woglinde> my compiled libva dontworks with iegd
Apr 17 08:26:17 <mckoan> hi gremlin[it] 
Apr 17 08:28:48 <woglinde> hms
Apr 17 08:28:57 <woglinde> I should read the error message carefully
Apr 17 08:30:08 <gremlin[it]> i'm trying to build console-image with rpm package, it stop cause populating rootfs isn't able to find the pgp keys for rpm ... someone have clue about it 
Apr 17 08:30:46 <gremlin[it]> at now a solution could be siply to not check signature of rpm 
Apr 17 08:31:23 <woglinde> gremlin hm I think should be an option to disable the crypto stuff
Apr 17 08:35:02 <gremlin[it]> woglinde, any suggestion about where to check ? some bbclass file i suppose 
Apr 17 08:35:53 <woglinde> hi florian
Apr 17 08:41:51 <gremlin[it]> hi mckoan 
Apr 17 08:52:41 <Longfield> Hi ... I'm on the stable branch, and since my last git pull yesterday, I am told that all packages have no entry in checksum.ini
Apr 17 08:53:34 <Longfield> I have seen a checksum discussion lately on the mailing-list, so there definitely must be some work around this these days ... but what's the "correct" way to handle this now ?
Apr 17 08:54:28 <florian> hi all
Apr 17 09:17:07 <pb_> florian: good morning
Apr 17 09:17:28 <florian> hey pb_ 
Apr 17 09:17:42 <woglinde> hi pb
Apr 17 09:17:44 <hrw> morning
Apr 17 09:17:47 <florian> hi hrw 
Apr 17 09:17:48 <woglinde> ho hrw
Apr 17 09:18:40 <hrw> I got usb-ethernet card today which appears as usbX instead of ethX
Apr 17 09:18:51 <woglinde> hrw why not
Apr 17 09:18:55 <hrw> [591970.653307] dm9601: No valid MAC address in EEPROM, using d6:db:0b:8f:b2:60
Apr 17 09:18:55 <hrw> [591970.657102] usb0: register 'dm9601' at usb-0000:00:1a.7-5.4.1, Davicom DM9601 USB Ethernet, d6:db:0b:8f:b2:60
Apr 17 09:19:04 <hrw> woglinde: second dm9601 card appears as eth0
Apr 17 09:19:08 <woglinde> let udev rename it
Apr 17 09:19:16 <hrw> [591993.938390] eth1: register 'dm9601' at usb-0000:00:1a.7-5.3, Davicom DM9601 USB Ethernet, 00:60:6e:7f:0c:20
Apr 17 09:21:37 <woglinde> hrw hm oh strange
Apr 17 09:26:39 <hrw> woglinde: and udev rule is not so good solution as I move those cards between many devices
Apr 17 09:29:13 <woglinde> hm
Apr 17 09:32:23 <gremlin[it]> hrw, i have trouble with rpm packaging ... "You have enabled checking of packages via GPG keys." but on rootfs's yum.conf there is "gpgcheck = 0" ... i don'r understand ... and so do_rootfs fail 
Apr 17 09:35:38 <hrw> gremlin[it]: I worked on rpm rootfs long time ago (few months)
Apr 17 09:35:56 <hrw> when I was merging it to OE I got rootfs working so did not checked how it was done
Apr 17 09:36:52 <hrw> and my nightly build shown that my queue of updates for stable/2009 require work from scratch again
Apr 17 09:37:08 <gremlin[it]> mhh ok i'll try to check, can you answer a question koen ask on mailing list in reply to a my patch on rpm-native package, i really don't know the answer (why a -native recipe generate a package) ... 
Apr 17 09:38:13 <woglinde> gremlin you install native packages on your target device?
Apr 17 09:39:01 <hrw> gremlin[it]: add PACKAGES = "" into rpm-native recipe after 'require rpm_${PV}.inc'
Apr 17 09:39:28 <hrw> gremlin[it]: or exchange require and inherit lines
Apr 17 09:39:35 <hrw> hi Graeme
Apr 17 09:39:47 <woglinde> hi xora
Apr 17 09:45:13 <florian> hi thebohemian 
Apr 17 09:45:20 <thebohemian> hi florian
Apr 17 09:45:37 <hrw> ok, time to add serial ports to machine
Apr 17 09:45:54 <hrw> from 4 -> 10 or just 8..
Apr 17 09:45:55 <hrw> ;D
Apr 17 09:45:58 <hrw> bye
Apr 17 09:54:27 <gremlin[it]> ok thanks hrw ! 
Apr 17 10:03:26 <hrw> re
Apr 17 10:04:10 <woglinde> re hrw
Apr 17 10:07:30 <hrw> heh.. I have 7 normal serial ports now and have to reconfigure kernel ;(
Apr 17 10:11:54 <hrw> CONFIG_SERIAL_8250_NR_UARTS should be CONFIG_SERIAL_NR_UARTS rather
Apr 17 10:12:47 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * r3e13bc01b2 10openembedded.git/recipes/iphone/ (4 files): 
Apr 17 10:12:47 <CIA-2> iphone: start adding toolchain elements for iphone
Apr 17 10:12:47 <CIA-2> * iphone-sources: script to download and extract rootfs / SDKs
Apr 17 10:12:47 <CIA-2> * iphone-sources.manifest: config file describing a firmware version
Apr 17 10:12:47 <CIA-2> * iphone-rootfs.bb: stage libraries from firmware image
Apr 17 10:12:48 <CIA-2> * iphone-sdks.bb: stage headers from SDKS
Apr 17 10:14:29 <hrw> ~curse serial maintainer
Apr 17 10:14:30 <ibot> May you be reincarnated as a Windows XP administrator, serial maintainer !
Apr 17 10:15:26 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * r15b556e91d 10openembedded.git/recipes/iphone/apple-csu_0.30.bb: apple-csu: add core libraries for iPhone
Apr 17 10:15:55 <broonie> hrw: IIRC there is no serial maintainer ATM.
Apr 17 10:16:24 <hrw> broonie: but once was.. and made those ugly NR_UARTS things
Apr 17 10:16:25 <XorA> broonie: to be fair, there has been no real serial work on kernel since 2.0
Apr 17 10:16:33 <XorA> broonie: its all been backwards progress
Apr 17 10:16:42 <broonie> Indeed.
Apr 17 10:16:55 <broonie> It's picked up a bit recently, though-  Alan Cox has been doing a lot of cleanups in the TTY layer.
Apr 17 10:17:04 * XorA remebers the days he could plug in 2x 16c954 baords and they just worked
Apr 17 10:17:05 <hrw> 'Maximum number of 8250/16550 serial ports' in reality is 'Maximum number of serial ports'
Apr 17 10:17:25 <XorA> at 2M without patches
Apr 17 10:17:25 <broonie> IIRC that's actually overridable by a kernel command line parameter now.
Apr 17 10:17:46 <hrw> and why mos9845/9835 rely on 8250..
Apr 17 10:18:04 <XorA> hrw: compatible register sets?
Apr 17 10:18:22 <XorA> hrw: most uarts Ive used are in some way 8250 capable just to make them easy to bootstrap
Apr 17 10:19:34 <hrw> looks like I will have ttyS0, then dead S1-S3, then S4-S9
Apr 17 10:19:56 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * re7f0c7a787 10openembedded.git/conf/machine/iphone.conf: iphone.conf: add basic machine description for iPhone
Apr 17 10:20:55 <hrw> and prefer to not think how would situation look if I would plug mos9835 (2xrs232, 1xlpt) and moxa (8xrs232)
Apr 17 10:25:24 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * r5d90e1e748 10openembedded.git/recipes/iphone/gcc-iphone-cross_git.bb: gcc-iphone-cross: add rough cross-compiler for iPhone
Apr 17 10:25:26 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * r319b8006cf 10openembedded.git/conf/distro/iphone-compat.conf: iphone-compat.conf: add mini distro with toolchain for iPhone
Apr 17 10:31:19 <gremlin[it]> hrw, your host machine is a redhat like ? 
Apr 17 10:37:45 <hrw> gremlin[it]: no, Debian
Apr 17 10:38:50 <gremlin[it]> mhhh seem native-yum use some info from my host system 
Apr 17 10:39:05 <gremlin[it]> " Error: Cannot retrieve repository metadata (repomd.xml) for repository: fedora. Please verify its path and try again"
Apr 17 10:39:33 <gremlin[it]> hrw, obviously there isn't "fedora" repository in OE's rootfs 
Apr 17 10:40:08 <gremlin[it]> food time for me, i'll read later 
Apr 17 10:42:28 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r989f05cdca 10openembedded.git/recipes/tasks/task-beagleboard-demo.bb: task-beagleboard-demo: remove psplash hardcode
Apr 17 10:42:35 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r4debb8b0f8 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev
Apr 17 10:42:38 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r8821a1efe4 10openembedded.git/conf/distro/iphone-compat.conf: iphone-compat: require angstrom, since minimal is just a copy of angstrom with added bugs, which we don't need
Apr 17 11:00:24 <CIA-2> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * r22b8d2b81d 10openembedded.git/conf/distro/ (include/fso-autorev.inc iphone-compat.conf): 
Apr 17 11:00:24 <CIA-2> iphone.conf: revert 8821a1efe4bf6a5cb88792b5d5f250ef480604cb, since apart from the polemics, it's the call of
Apr 17 11:00:24 <CIA-2> the maintainer, not yours.
Apr 17 11:01:19 <pb_> good work mickey
Apr 17 11:01:28 <mickeyl> morning pb_ 
Apr 17 11:01:36 <pb_> mickeyl: good morning :-)
Apr 17 11:02:23 <mickeyl> i guess things were too calm for a while ;)
Apr 17 11:02:32 <pb_> heh, it seems
Apr 17 11:07:40 <hrw> hi mickeyl 
Apr 17 11:07:47 <mickeyl> morning hrw 
Apr 17 11:08:00 <mickeyl> troubles with the serial layer again?
Apr 17 11:08:35 <hrw> mickeyl: finally got it to working like I wanted
Apr 17 11:08:48 <hrw> ttyS0 is mainboard serial, S1-S6 are mos9845 ones
Apr 17 11:12:08 <hrw> mickeyl: your commit touched sane-srcrevs!
Apr 17 11:12:47 <mickeyl> luckily not
Apr 17 11:12:52 <mickeyl> just fso-autorev
Apr 17 11:13:07 <mickeyl> which was a mistake as well, but a bearable one
Apr 17 11:13:11 <mickeyl> imo ;)
Apr 17 11:18:27 <hrw> oik ;D
Apr 17 11:18:41 * hrw wait for 'built-test2.sh base-image' end
Apr 17 11:19:08 <hrw> script is nightmare of resource use
Apr 17 11:19:27 <hrw> do builds for glibc and uclibc for 4-5 archs
Apr 17 11:28:43 * hrw -> food
Apr 17 11:34:49 <mickeyl> ~bon appetit
Apr 17 11:34:50 <ibot> bon appetit is, like, smacznego.  Guten Appetit.  Eet Smakelijk.  God Appetitt.  Buon Appetito.  Buen apetito  Bom Apetite. buen apetito  Smaklig måltid!.  Hyvää ruokahalua.  Bo Proveito
Apr 17 11:34:55 <Ainulindale> hey mickeyl 
Apr 17 11:35:18 <Ainulindale> mickeyl: did you see the log I posted about the race in oeventsd generating CPU requests?
Apr 17 11:36:58 <mickeyl> i don#t think so
Apr 17 11:41:07 <carlosguerreiro> hello
Apr 17 11:45:42 <ant_work> RP: ping
Apr 17 11:46:04 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * rbd3f015607 10openembedded.git/conf/distro/iphone-compat.conf: 
Apr 17 11:46:04 <CIA-2> iphone-compat: require angstrom, since minimal is just a copy of angstrom with added bugs, which we don't need
Apr 17 11:46:04 <CIA-2> From the my jabber logs with Jeremy, the maintainer of iphone-compat:
Apr 17 11:46:05 <CIA-2> 12:37 Koen: better inherit angstrom, since minimal is just angstrom with added bugs
Apr 17 11:46:07 <CIA-2> 12:38 Sharky: ah ok
Apr 17 11:46:09 <CIA-2> 12:38 Sharky: feel free to change it
Apr 17 11:46:11 <CIA-2> Mickey: next time check things before reverting (and violating the revert policy)
Apr 17 11:47:15 <mickeyl> *sigh*
Apr 17 11:48:30 * mickeyl puts on his oh-well-i-dont-care-hat
Apr 17 11:50:53 <Crofton> mickeyl, I hope that someone could create a distro in OE, with fewer lines in it than minimal has
Apr 17 11:51:21 <Crofton> it really looks like minimal grw beyond the definition of the minimum amount of stuff needed to build simple images
Apr 17 11:52:04 <hrw> and name it 'really-minimal'?
Apr 17 11:52:48 <Ainulindale> hrw: "über minimal"
Apr 17 11:52:51 <Ainulindale> We did that for SHR.
Apr 17 11:53:06 <Ainulindale> http://trac.shr-project.org/trac/wiki/Get%20Started#SHR-ImageULTRA%C3%9CBERLITEContent <=
Apr 17 11:54:32 <Crofton> then it will pick up features and we will need really-minimal-this-time :)
Apr 17 11:55:16 <hrw> ~curse task-base
Apr 17 11:55:17 <ibot> May you be reincarnated as a Windows XP administrator, task-base !
Apr 17 11:55:18 <mickeyl> if it bothers you, i can rename it
Apr 17 11:55:33 <hrw> without changing task-base we can not say about minimal...
Apr 17 11:55:47 <mickeyl> all my work on making any alternatives to angstrom seems to void these days anyways
Apr 17 11:56:09 <mickeyl> if the majority of folks don't care about mixing OE and angstrom, the why should I
Apr 17 11:56:11 <mickeyl> *shrug*
Apr 17 11:56:15 <hrw> 'if your machine has pci or usbhost or pcmcia' then 'build x11, gstreamer because you may plug wifi or bluetooth'
Apr 17 11:57:04 <hrw> and I lack idea how to move task-base-bt to separate and keep it working
Apr 17 12:00:33 <ant_work> hrw: MACHINE_FEATURES ?
Apr 17 12:00:44 <ant_work> like USE flag
Apr 17 12:01:15 <hrw> ant_work: I was one of inventors of MACHINE/DISTRO/COMBINED_FEATURES
Apr 17 12:01:40 <ant_work> what's the blocker?
Apr 17 12:02:07 <hrw> time for tests
Apr 17 12:02:24 <hrw> my machine do 5 archs, 2 libcs build now
Apr 17 12:02:42 <hrw> and is at 2447/2908 of first run
Apr 17 12:02:57 <Crofton> mickeyl, so you can now create iphone apps with OE?
Apr 17 12:03:08 <hrw> Crofton: not quite
Apr 17 12:03:26 <mickeyl> Crofton: no idea, ask Jeremy
Apr 17 12:03:29 <Crofton> ah
Apr 17 12:03:29 <hrw> Crofton: you can run misc stuff on iphone/ipod-touch 
Apr 17 12:03:30 <Crofton> ok
Apr 17 12:03:35 <mickeyl> i still create iPhone apps with Xcode
Apr 17 12:03:38 <hrw> Crofton: clutter based apps for example
Apr 17 12:03:46 <mickeyl> since that's the only blessed way to get in the AppStore
Apr 17 12:03:51 <hrw> RP did that ~year ago
Apr 17 12:03:54 * hrw -> coffee
Apr 17 12:04:11 <Crofton> ah
Apr 17 12:04:25 <Crofton> I'm wondering what the distro does
Apr 17 12:36:52 <CIA-2> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * r0f5bd8682d 10openembedded.git/ (2 files in 2 dirs): libeflvala: bump to 0.0.1.0 and fix Q/A packaging error
Apr 17 12:37:00 <mickeyl> Ainulindale: that should be it
Apr 17 12:37:03 <CIA-2> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * ra9bb0bd17e 10openembedded.git/ (2 files in 2 dirs): libgsm0710: bump version to 1.1.1 release; add author and homepage
Apr 17 12:38:48 <Ainulindale> mickeyl: thanks a lot :-)
Apr 17 12:38:52 <Ainulindale> mickeyl: want a hug?
Apr 17 12:38:54 <Ainulindale> :-)
Apr 17 12:38:57 <Ainulindale> by the way mickeyl 
Apr 17 12:39:05 <Ainulindale> mickeyl: in our overlay, recipes/shr are mostly clean
Apr 17 12:39:13 <Ainulindale> I'd like to know if you could review them
Apr 17 12:39:25 <mickeyl> sure
Apr 17 12:39:27 <Ainulindale> If you do I commit today these recipes
Apr 17 12:39:40 <mickeyl> oh, today...
Apr 17 12:39:49 <Ainulindale> Spent all yesterday working on that :-)
Apr 17 12:40:31 <mickeyl> ok, let me have a look after lunch
Apr 17 12:40:38 <mickeyl> hungry now
Apr 17 12:53:56 <gremlin[it]> no way .. yum still continue to look for "fedora" repository ... that is present on my host but no on rootfs ... hrw .. how i can debug ??? 
Apr 17 13:03:05 <hrw> gremlin[it]: strace?
Apr 17 13:03:17 <hrw> gremlin[it]: my knowledge about yum is very limited
Apr 17 13:04:11 <gremlin[it]> mhh ok ... the strange thing is that also adding '-v -d 10' (verbose and debug messages) no additional output happen 
Apr 17 13:04:48 <gremlin[it]> btw, yum is in python ... strace could be useful ? 
Apr 17 13:05:41 <hrw> why not
Apr 17 13:06:56 <BusError> is there a way to clean all the non-bases packages (aka keep toolchain, libc etc) for rebuild ?
Apr 17 13:14:22 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r9284c24637 10openembedded.git/recipes/pulseaudio/ (pulseaudio.inc pulseaudio_0.9.10.bb pulseaudio_0.9.11.bb): pulseaudio: fix volatiles and ship policykit file
Apr 17 13:14:22 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r068ae7f600 10openembedded.git/ (3 files in 2 dirs): 
Apr 17 13:14:22 <CIA-2> packagekit: fix service and policy files, update git version
Apr 17 13:14:22 <CIA-2> * patches against 0.4.6 have been applied upstream
Apr 17 13:14:24 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * ra610a5d172 10openembedded.git/recipes/policykit/policykit_0.9.bb: policykit: fix permissions and volatiles
Apr 17 13:14:51 <mckoan> gremlin[it]: strace is always useful, sometime too much :-D
Apr 17 13:19:40 <Ainulindale> mickey|lunch: you didn't cherry pick to ms5.5
Apr 17 13:27:44 <CIA-2> 03Julien 'Ainulindalë' Cassignol <ainulindale@gmail.com> 07fso/milestone5.5 * rb394c8ca89 10openembedded.git/ (2 files in 2 dirs): 
Apr 17 13:27:44 <CIA-2> libeflvala: bump to 0.0.1.0 and fix Q/A packaging error
Apr 17 13:27:44 <CIA-2> Conflicts:
Apr 17 13:27:44 <CIA-2> conf/distro/include/sane-srcrevs.inc
Apr 17 13:27:53 <gremlin[it]> hrw, I just notice : work/tmp/deploy/glibc/rpm contain only repodata files but didn't exist any .rpm file ... and so all repodata info report there isn't packages 
Apr 17 13:28:32 <hrw> srange
Apr 17 13:28:50 <hrw> uf.. 3 archs passed 'base-image' in stable/2009/glibc build
Apr 17 13:41:31 <woglinde> hrw and uclibc?
Apr 17 13:45:43 <gremlin[it]> hrw, gotcha !!!! movefile: Stating source file failed... [Errno 2] No such file or directory: '/mnt/spare/oe/PSAc/tmp/work/i586-kaeilos-linux/busybox-1.13.2-r17/rpm/RPMS/i586/busybox-linuxrc-1.13.2-r17.i586.rpm
Apr 17 13:48:37 <gremlin[it]> the file is /mnt/spare/oe/PSAc/tmp/work/i586-kaeilos-linux/busybox-1.13.2-r17/rpm/RPMS/i586/busybox-linuxrc-1.13.2-r17.1.i586.rpm
Apr 17 13:51:18 <hrw> woglinde: uclibc will be next
Apr 17 13:51:58 <hrw> woglinde: http://pastebin.ca/1394955
Apr 17 13:55:26 <hrw> woglinde: thats my build test script
Apr 17 14:07:23 <CIA-2> 03Michael 'Mickey' Lauer <mickey@vanille-media.de> 07org.openembedded.dev * r9c42b0b88c 10openembedded.git/ (3 files in 2 dirs): 
Apr 17 14:07:23 <CIA-2> libgsm0710mux: bump to 0.3.3 release
Apr 17 14:07:23 <CIA-2> fso-abyss: bump to 0.3.3 release
Apr 17 14:19:23 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * re0b9b37479 10openembedded.git/classes/insane.bbclass: insane.bbclass: add an entry for darwin9/arm
Apr 17 14:32:09 <hrw> ok, uclibc now
Apr 17 15:22:17 <CIA-2> 03Steffen Sledz <sledz@dresearch.de> 07org.openembedded.dev * rc838626c2b 10openembedded.git/recipes/linux/ (3 files in 2 dirs): linux-2.6.24: nand and ubifs patches for HIPOX machine
Apr 17 15:47:33 <woglinde> hm oh LiMo phones from NEC and panasonic
Apr 17 16:18:06 <CIA-2> 03kergoth 07bitbake-1.8 * r1159 10/lib/bb/fetch/__init__.py: 
Apr 17 16:18:06 <CIA-2> bb.fetch: make try_mirror go through the dispatcher rather than using wget directly.
Apr 17 16:18:06 <CIA-2> Signed-off-by: Chris Larson <clarson@mvista.com>
Apr 17 16:18:31 <CIA-2> 03kergoth * r1160 10bitbake/lib/bb/fetch/__init__.py: 
Apr 17 16:18:31 <CIA-2> bb.fetch: make try_mirror go through the dispatcher rather than using wget directly.
Apr 17 16:18:31 <CIA-2> Signed-off-by: Chris Larson <clarson@mvista.com>
Apr 17 16:19:01 <kergoth> (there, now you could set SRC_TARBALL_STASH to a file:// url or something)
Apr 17 16:27:44 <woglinde> till later
Apr 17 16:43:23 <hrw> bye
Apr 17 16:45:02 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r012465dda9 10openembedded.git/recipes/opkg/ (5 files in 2 dirs): opkg: only ping splash when splashfuncs are present
Apr 17 16:49:29 <CIA-2> 03Koen Kooi <koen@openembedded.org> 07org.openembedded.dev * r54c2f02920 10openembedded.git/recipes/sysvinit/ (sysvinit/angstrom/rc sysvinit_2.86.bb): sysvinit: only ping splash when splashfuncs are present
Apr 17 17:57:24 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * r3b76c92985 10openembedded.git/conf/bitbake.conf: 
Apr 17 17:57:24 <CIA-2> bitbake.conf: fix package population for darwin
Apr 17 17:57:24 <CIA-2> * define SOLIBS and SOLIBSDEV for darwin9 (identical to darwin8)
Apr 17 17:57:24 <CIA-2> * make use of SOLIBS and SOLIBSDEV to populate FILE_* instead of hardcoded .so.* and .so
Apr 17 17:58:04 <CIA-2> 03Jeremy Lainé <jeremy.laine@m4x.org> 07org.openembedded.dev * ra4f79016c5 10openembedded.git/classes/lib_package.bbclass: 
Apr 17 17:58:04 <CIA-2> lib_package.bbclass: fix package population on darwin
Apr 17 17:58:04 <CIA-2> * make use of SOLIBS and SOLIBSDEV to populate FILE_* instead of hardcoded .so.* and .so
Apr 17 18:52:28 <jeremy_laine> khem: ping?
Apr 17 19:09:23 <Tartarus> hmm, missing dep problem in canadian-sdk it seems
Apr 17 19:09:33 <Tartarus> had a build fail overnight, a few -k's and it worked
Apr 17 19:09:39 * Tartarus hopes to recreate
Apr 17 19:43:10 <Gnutoo> mmmthere was a talk long time ago about hal dependencies of xserver...I'll look at it because kdrive depend on hal and hal depend on recent 2.6 kernel header
Apr 17 19:46:49 <Gnutoo> ah...it was a bugreport...
Apr 17 19:46:50 <Gnutoo> http://bugs.openembedded.net/show_bug.cgi?id=3757
Apr 17 19:48:34 <pb___> I think there was also some discussion on the mailing list.
Apr 17 20:01:26 <Tartarus> Bingo, found it, minor goof-up in canadian-sdk.bbclass wrt DEPENDS_prepend
Apr 17 20:02:30 <Gnutoo> it must be long time ago then...
Apr 17 20:04:57 <Tartarus> dhd, so, I think your 1st patch is fine and I'll push that shortly
Apr 17 20:05:03 <Tartarus> NAK'ing the other with a reason in a moment
Apr 17 20:05:38 <Gnutoo> mmm...when I moved some files(local recipes...not much)...it wanted to recompile them...strange
Apr 17 20:07:18 <Gnutoo> (it is recompiling uclibc-initial)
Apr 17 20:11:22 <Gnutoo> ah ok I understand...
Apr 17 20:11:34 <Gnutoo> I'll read more on collections...
Apr 17 20:11:47 <woglinde> hi
Apr 17 20:11:50 <Gnutoo> hi
Apr 17 20:11:58 <Gnutoo> I hope we can have 3 repos
Apr 17 20:12:06 <Gnutoo> s/repos/collection of bb files
Apr 17 20:12:37 <kergoth> you can have as many as you want
Apr 17 20:13:26 <Gnutoo> ok so it must be a keyword problem...I'll change the pma430 keyword to local...thanks a lot
Apr 17 20:17:10 <Gnutoo> mmm...I don't understand how to put as many collections as I want....
Apr 17 20:18:00 <kergoth> neither COLLECTIONS nor BBFILE_COLLECTIONS have any limitations on the number of entries
Apr 17 20:18:20 <Gnutoo> ok 
Apr 17 20:18:30 <Gnutoo> I'll try puting priorities < 10
Apr 17 20:18:40 <Gnutoo> else if it doesn't work I'll pastebin my setup
Apr 17 20:18:54 <kergoth> sounds good
Apr 17 20:22:04 <Gnutoo> also I'll pull 
Apr 17 20:31:11 <Gnutoo> http://rafb.net/p/fjl4Hv61.html
Apr 17 20:31:16 <Gnutoo> here it is
Apr 17 20:31:22 <Gnutoo> but it's realy strange...
Apr 17 20:32:22 <kergoth> what's the behavior?
Apr 17 20:32:44 <kergoth> pma430 isn't in BBFILE_COLLECTIONS, is tha tintentional?
Apr 17 20:32:47 <kergoth> blah, cant' tyep
Apr 17 20:32:48 <kergoth> lkfashg
Apr 17 20:33:24 <Gnutoo> ok i'll add it thanks sorry
Apr 17 20:33:31 <kergoth> :)
Apr 17 20:33:39 <kergoth> what problem are you hitting, though? i didn't see it
Apr 17 20:34:25 <Gnutoo> (must be that pma430 isn't in bbfiles collection...I've checked the paths and a lot of things but not that...sorry)
Apr 17 20:34:42 <Gnutoo> that it didn't took my pma430 recipes and wanted to rebuild uclibc-initial
Apr 17 20:35:06 <kergoth> ah
Apr 17 20:35:14 <kergoth> it should have still parsed those recipes since they're in BBFILES, but they wouldnt be preferred
Apr 17 20:35:47 <Gnutoo> yes as didn't took I meant didn't prefer sorry
Apr 17 20:36:00 <kergoth> ahh, right
Apr 17 20:36:01 <kergoth> okay then :)
Apr 17 20:40:35 <Gnutoo> mmm....still have the problem...strange...NOTE: Running task 1 of 1 (ID: 0, /home/embedded/oe/org.openembedded.dev/recipes/uclibc/uclibc-initial_0.9.30.1.bb, do_clean)
Apr 17 20:40:47 <Gnutoo> I'll re-double check my options
Apr 17 20:42:08 <Gnutoo> ah strange
Apr 17 20:42:43 <Gnutoo>  peek uclibc-initial PR => r2
Apr 17 20:42:55 <Gnutoo> NOTE: package uclibc-initial-0.9.30.1-r1: task do_setscene: completed
Apr 17 20:43:47 <kergoth> i think itll prefer the latest rev, and the latest ver too if it isnt locked down.. so if v5 is in upstream and v4 is in your repo, itd use v5 even though your overlay has higher priority, unless you lock it down to v4
Apr 17 20:43:49 <kergoth> if that makes sense
Apr 17 20:43:53 * kergoth isn't quite sure htough
Apr 17 20:44:12 <woglinde> ~seen pb
Apr 17 20:44:13 <woglinde> args
Apr 17 20:44:15 <Gnutoo> peek is from my repo
Apr 17 20:44:16 <woglinde> ~seen rb
Apr 17 20:44:16 <ibot> pb <n=pb@166.Red-88-18-114.staticIP.rima-tde.net> was last seen on IRC in channel #tomcat, 176d 2h 51s ago, saying: 'arghh'.
Apr 17 20:44:18 <ibot> rb <n=rb@86.81.163.228> was last seen on IRC in channel #kde, 749d 12h 15m 39s ago, saying: 'Like, set it so one can manually arrange the icons.'.
Apr 17 20:44:28 <Gnutoo> r2 is in my repo and r1 in org.openembedded.dev
Apr 17 20:44:29 <woglinde> ~seen rp
Apr 17 20:44:30 <ibot> rp is currently on #maemo (17h 50m 23s) #handhelds (17h 50m 23s) #oe (17h 50m 23s) #openmoko (17h 50m 23s), last said: 'ant_work: It was part of the sharp 2.4 junk'.
Apr 17 20:44:51 <woglinde> hi ant
Apr 17 20:45:17 <kergoth> Gnutoo: your OEDIR is ${HOME}/oe?  you use OEDIR in the regexes, but not in the bbfiles.  just making sure :)
Apr 17 20:45:23 * kergoth shrugs
Apr 17 20:45:29 <Gnutoo> yes it is
Apr 17 20:47:43 <Gnutoo> I'll put the same evrywhere
Apr 17 20:48:28 <Gnutoo> mmm was oedir
Apr 17 20:48:34 <Gnutoo> it is: None
Apr 17 20:48:42 <Gnutoo> strange...it was in the manual
Apr 17 20:49:01 <Gnutoo> must be a holder and not a real variable then
Apr 17 20:49:02 <kergoth> you'd have to 1) set OEDIR in your env, and 2) add it to the bb env whitelist, or 3) set it explicitly in your .conf
Apr 17 20:49:14 <kergoth> some env setup scripts use it so you dont have to hardcode paths in the .conf
Apr 17 20:49:34 <kergoth> also, dont forget about bitbake -e
Apr 17 20:49:44 <kergoth> can look through it to see the actual value of BBFILES, etc
Apr 17 20:49:52 <kergoth> helps w/ diagnosis
Apr 17 20:51:36 <Gnutoo> i'm using bitbake -i
Apr 17 20:51:57 <ant__> hi woglinde 
Apr 17 20:52:06 <ant__> hello kergoth 
Apr 17 20:52:16 <ant__> Gnutoo: instancabile...
Apr 17 20:52:51 <Gnutoo> what do you mean by instancabile...that I often do something?
Apr 17 20:53:00 <Gnutoo> or that I keep on until it works?
Apr 17 20:53:30 <ant__> no, just that you're busy since 2-3 days with that
Apr 17 20:53:41 <ant__> nobreath-taking task...
Apr 17 20:54:31 <Gnutoo> ah ok
Apr 17 20:54:44 <Gnutoo> there are times where I'm very busy with oe and time where I'm not
Apr 17 20:55:02 <ant__> I hope for ya ;-)
Apr 17 20:57:08 <Gnutoo> and I'm a bit lazy...for instance I made emacs cross compile but I didn't commit it because it needed to test too much patches for qemu: I would have needed to wrote a custom software for doing it
Apr 17 20:57:15 <Gnutoo> but it's on my todo list
Apr 17 20:57:38 <Gnutoo> (I needed to find the one that made the cross-compilation possible)
Apr 17 20:59:24 <Gnutoo> ah ok I understand...bitbake -i 's parsing was too fast
Apr 17 20:59:31 <Gnutoo> it didn't reparse evrything
Apr 17 21:03:07 <kergoth> everyone ever seen bitbake choke when trying to find loops in the dependencies?  very deep callstack in find_chains, getting a KeyError on the "revdep" in explored_deps
Apr 17 21:03:20 <kergoth> i know i have a dep loop, but i can't isolate it, because bitbake is crashinga nd burning :)
Apr 17 21:06:18 <Gnutoo> thanks a lot it works now
Apr 17 21:20:46 <lyakh> hi, any idea about superh support (sh4al-dsp fpu-less)?
Apr 17 21:22:13 <Gnutoo> ouch...configure: error: C compiler cannot create executables ...I think I must recompile my toolchain
Apr 17 21:22:25 <Gnutoo> (I did the stupid thing of cleaning uclibc-initial)
Apr 17 21:23:42 <Gnutoo> mmm...(.text+0x28): undefined reference to `__uClibc_main'
Apr 17 21:30:28 <spaetz> is there a reason, why the hg (mercurial) support is so sparse for oe/bitbake?
Apr 17 21:31:07 <spaetz> the only bug I could find is that there is no equivalent to git -ls -r (for remotely identifying the current rev)
Apr 17 21:31:11 <woglinde> spaetz hm we havent much projects now from what which we need the tip version
Apr 17 21:31:21 <woglinde> so nobdy cared
Apr 17 21:31:25 <kergoth> I'd guess someone hasn't taken the time to improve it.  patches welcome
Apr 17 21:31:26 <spaetz> however "hg identify http://mips.complang.tuwien.ac.at/hg/cacao" shows the current tip
Apr 17 21:31:38 <spaetz> at least with my current hg
Apr 17 21:31:56 <woglinde> spaetz you need the latest cacao version?
Apr 17 21:32:04 <spaetz> nahh, just an example
Apr 17 21:32:22 <spaetz> I have small project for the openmoko and I did it with hg
Apr 17 21:32:48 <spaetz> but switched to git, as I never managed to get a .bb file that fetched the sources with hg
Apr 17 21:33:42 <Tartarus> chicken and egg problem :)
Apr 17 21:33:50 <spaetz> with openmoko/FSO/SHR we often need the tip of programs
Apr 17 21:33:55 <spaetz> living on the edge...
Apr 17 21:34:00 <spaetz> Tartarus: seems like it
Apr 17 21:34:10 <spaetz> :)
Apr 17 21:34:29 <kergoth> as with most open source projects, most work is done by those who need that functionality.. so.. have at it :)
Apr 17 21:35:05 <spaetz> kergoth: OE/bitbake is a bit daunting, so I hoped for an easier solution
Apr 17 21:35:17 <spaetz> I was afraid of getting your answer :) but fair enough
Apr 17 21:35:46 <Tartarus> heh, its just python
Apr 17 21:35:55 <Tartarus> lib/bb/fetcher/hg.py, or so, needs hacking on :)
Apr 17 21:37:14 <Tartarus> kergoth, do you have access to bitbake?
Apr 17 21:37:21 <kergoth> yep
Apr 17 21:37:26 <Tartarus> !oebug 6073
Apr 17 21:37:27 <cdbot2> * * Bug 6073, Status: HTTP transfer failed :(
Apr 17 21:37:30 <Tartarus> !oebug 5073
Apr 17 21:37:31 <cdbot2> * * Bug 5073, Status: UNCONFIRMED, Created: 2009-03-30 18:46
Apr 17 21:37:32 <cdbot2> * * trini(AT)embeddedalley.com: Local git (git:///path/to/it) URIs don&apos;t work
Apr 17 21:37:33 <cdbot2> * * http://bugs.openembedded.net/show_bug.cgi?id=5073
Apr 17 21:37:42 <Tartarus> There's a patch there too, if you have time :)
Apr 17 21:37:48 <kergoth> i'll add it to the todo list
Apr 17 21:37:57 <kergoth> i still have like 12 patches of my own i need to get pushed to various places
Apr 17 21:38:02 <Tartarus> heh, k, thanks
Apr 17 21:38:10 <kergoth> had to take the time to get things cleaned up and ready and all, ,you know
Apr 17 21:38:13 <kergoth> err, hard
Apr 17 21:38:14 <kergoth> not had
Apr 17 21:38:17 <ant__> !oebug 5002
Apr 17 21:38:18 <cdbot2> * * Bug 5002, Status: UNCONFIRMED, Created: 2009-01-29 12:28
Apr 17 21:38:19 <cdbot2> * * roman(AT)khimov.ru: Graphviz deps tree generation broken
Apr 17 21:38:20 <cdbot2> * * http://bugs.openembedded.net/show_bug.cgi?id=5002
Apr 17 21:38:31 <kergoth> thanks for the reminder, ant__ 
Apr 17 21:38:33 <kergoth> :)
Apr 17 21:38:33 <Tartarus> yeah, understood
Apr 17 21:38:37 <ant__> kergoth: I tried to catch RP for ack..no trace of him
Apr 17 21:38:43 <ant__> ..today
Apr 17 21:38:50 <kergoth> i'll see if i can handle both of those tonight or tomorrow
Apr 17 21:38:53 <kergoth> weekend > *
Apr 17 21:39:15 <ant__> zecke is still in Far East, isn't?
Apr 17 21:41:20 <kergoth> Tartarus: that patch looks good, i'll test with them both applied just to make sure nothing explodes horribly, and get them committed
Apr 17 21:41:43 <kergoth> I've been manually sending patches to richard via jabber for review for bitbake stuff, he seems too busy to peruse the MLs lately
Apr 17 21:41:44 <kergoth> heh
Apr 17 21:43:11 <Tartarus> OK, time to start on a series of canadian stuff for stable
Apr 17 21:43:27 <CIA-2> 03David Huggins-Daines <dhuggins@cs.cmu.edu> 07org.openembedded.dev * r71cc8fba29 10openembedded.git/recipes/meta/canadian-sdk.bb: 
Apr 17 21:43:27 <CIA-2> canadian-sdk: Synchronize (to some extent) canadian-sdk.bb with meta-toolchain.bb.
Apr 17 21:43:27 <CIA-2> Unlike meta-toolchain, it does not build an -extras package.
Apr 17 21:43:33 <CIA-2> 03Tom Rini <trini@embeddedalley.com> 07org.openembedded.dev * r669c8f6503 10openembedded.git/ (3 files in 3 dirs): 
Apr 17 21:43:33 <CIA-2> canadian-sdk: We DO want DEPENDS_prepend to always bring in host virtual gcc/binutils
Apr 17 21:43:33 <CIA-2> Without this gdb-canadian-sdk would try and build without a host compiler as
Apr 17 21:43:34 <CIA-2> it wasn't fixing up this missing dep itself. The comment in the bbclass was
Apr 17 21:43:36 <CIA-2> wrong (checked -g on all recipes using this bbclass).
Apr 17 21:43:40 <CIA-2> 03Tom Rini <trini@embeddedalley.com> 07org.openembedded.dev * rad507618ca 10openembedded.git/recipes/ (3 files in 2 dirs): 
Apr 17 21:43:42 <CIA-2> mingw-gcc, mingw-runtime, mingw-w32api: Fix DEFAULT_PREFERENCES to be correct
Apr 17 21:43:45 <CIA-2> We want a default of -1 (not 0) and 1 for "mingw32" so that the default list
Apr 17 21:43:46 <CIA-2> of overrides will catch and use them.
Apr 17 21:44:14 <woglinde> ant seems so
Apr 17 21:44:39 <woglinde> richard is busy with poulsbo
Apr 17 21:45:32 <woglinde> ant but the summer term started this week
Apr 17 21:45:40 <woglinde> so he will be here sooner or later
Apr 17 21:47:16 <kergoth> shoot, that reminds me
Apr 17 21:47:23 <kergoth> i keep forfgetting to make the bitbake git repo on git.openembedded.net
Apr 17 21:47:26 <kergoth> grr
Apr 17 21:50:20 <ant__> ach..packaged_staging build break again...
Apr 17 21:50:33 <kergoth> Tartarus: tarball stash url amuses me when using git:///
Apr 17 21:50:46 <kergoth> like, yeah, good luck fetching /home/clarson/blahblah.tar.gz on angstrom..
Apr 17 21:50:50 <kergoth> :)
Apr 17 21:51:18 <Tartarus> yes, that is funny :)
Apr 17 21:53:49 <Tartarus> yay for git cherry-pick
Apr 17 21:53:59 <Tartarus> Now to figure out a good for loop to compile test all of this stuff for stable
Apr 17 21:59:05 * Tartarus bg's 22 bitbake -g's and waits
Apr 17 21:59:42 <Tartarus> mwester, I assume SlugOS in stable/2009 is something you're taking care of, yes?
Apr 17 22:00:32 <kergoth> Tartarus: I'm getting a failure with that patch.  it's trying to run rsync to get refs, ideas?  perhaps due to the set of ud.proto in localpath()?
Apr 17 22:01:19 <kergoth> Tartarus: it works if i copy the ud.host/ud.proto=file block into localpath also
Apr 17 22:01:20 <kergoth> hmm
Apr 17 22:01:47 <Tartarus> hmm, ok :)
Apr 17 22:09:39 <kergoth> wow, thats a new one
Apr 17 22:09:54 <kergoth> "Task failed: Unknown fetch Error: dictionary changed size during iteration"
Apr 17 22:09:58 <kergoth> lovely
Apr 17 22:10:20 <ant__> what are you doing so dirty?
Apr 17 22:10:51 <kergoth> just testing a local git repo per Tartarus's patch
Apr 17 22:10:54 <kergoth> odd
Apr 17 22:11:06 * kergoth 's headache isg etting worse
Apr 17 22:14:29 <Tartarus> Whee, this should keep this room warm:
Apr 17 22:14:34 <Tartarus> for MACH in at91sam9263ek qemuarm omap591
Apr 17 22:14:35 <Tartarus> 2osk beagleboard vortex86sx qemux86 mpc8313e-rdb ; do for GCC in 4.1.2 4.2.4 4.3.3; do for META in meta-toolchain meta-toolchain-fso meta-toolchain-gpe meta-toolchain-gpe-sbox meta-toolchain-neuros meta-toolchain-opie meta-toolchain-sbox canadian-sdk; do MACHINE=$MACH ANGSTROM_GCC_VERSION=$GCC bitbake $META; done; done; done
Apr 17 22:14:58 <Tartarus> (and TMPDIR will change on gcc version at least)
Apr 17 22:15:18 <kergoth> hah
Apr 17 22:15:19 <kergoth> fun 
Apr 17 22:15:22 <ant__> Tartarus: cold fusion there?
Apr 17 22:15:27 <Tartarus> As an aside, this is why I bitch when people don't touch all toolchain versions
Apr 17 22:15:40 <Tartarus> that also adds in 4.3.2 and 4.3.1
Apr 17 22:16:14 <Tartarus> (and, sadly all of the 4.3.x's won't do canadian-sdk, due to mingw/gcc 4.3.x unhappiness i haven't had time to track and fix
Apr 17 22:16:27 * kergoth thinks Tartarus should be the official toolchain maintainer according to MAINTAINERS, if he isnt already
Apr 17 22:16:38 <Tartarus> heh, that's khem :)
Apr 17 22:16:48 <Tartarus> I just make sure the SDK stuff keeps working
Apr 17 22:16:48 <kergoth> ah :)
Apr 17 22:26:06 <kergoth> ah ha! somehow my urldata is changing inside the urldata loop in bb.fetch.go
Apr 17 22:26:08 * kergoth debugs
Apr 17 22:29:44 <kergoth> arg, i think its my try_mirrors change, somehow
Apr 17 22:29:46 <kergoth> bite me!
Apr 17 22:29:50 * kergoth grumbles
Apr 17 22:30:31 <The_Niz> lies
Apr 17 22:41:30 <denix> Tartarus: wanted to clarify what /proc/atags is, as I never used or even seen it before...
Apr 17 22:49:27 <Tartarus> Got an arm board?  Should have it
Apr 17 22:49:46 <Tartarus> Know what atags are or no?
Apr 17 22:49:58 <denix> you mean ARM-branded board? don't have any :)
Apr 17 22:50:33 <Tartarus> heh
Apr 17 22:50:34 <denix> don't know, that's what I'm asking you about :) I guess I should have googled first... :)
Apr 17 22:50:38 <Tartarus> any arm board should do
Apr 17 22:51:06 * Tartarus tries to find a good link
Apr 17 22:51:16 <Tartarus> But, atags are how arm passes info from firmware to the kernel
Apr 17 22:51:23 <Tartarus> ATAG_FOO, even
Apr 17 22:51:35 <denix> are you sure any arm-powered board should do?
Apr 17 22:51:36 <Tartarus> I assume, but havne't looked in a while, that /proc/atags is just that info exported
Apr 17 22:52:11 <Tartarus> digging out my n810
Apr 17 22:52:55 <kergoth> son of a bitch
Apr 17 22:53:14 <kergoth> bb.fetch must die
Apr 17 22:53:44 <ant__> he..survives ^C sometimes
Apr 17 22:53:55 <denix> Tartarus: http://pastebin.com/m64c3f98e
Apr 17 22:54:00 <kergoth> i want try_mirror to not call wget directly, but instead use bb.fetch's dispatcher, right
Apr 17 22:54:03 <kergoth> which makes sense
Apr 17 22:54:35 <kergoth> i don't want it to set up the urldata for those paths with bb.fetch.init, because if it does so, those urls are in the urldata for the file, and would be fetched by bb.fetch.go()
Apr 17 22:55:11 <Tartarus> denix, yeah
Apr 17 22:55:15 <kergoth> but i can't do it locally either, because wget's mirrors handling uses bb.fetch.localpath, which in turn calls bb.fetch.init, which manipulates urldata, hosing the current iteration of urldata
Apr 17 22:55:16 <Tartarus> I wonder if it's a 2.4 thing or what
Apr 17 22:55:22 <Tartarus> Koen brought it up, bug him
Apr 17 22:55:27 <kergoth> I'm going to have to revert this for now, and just fucking rewrite bb.fetch
Apr 17 22:55:36 * kergoth is getting very annoyed
Apr 17 22:55:39 <Tartarus> I just assumed it was something that exists, since I know arm uses ATAG_FOO to pass stuff rom firmware to kernel
Apr 17 22:55:40 <Tartarus> *from
Apr 17 22:56:22 <denix> Tartarus: ok, thanks
Apr 17 22:57:07 <khem> denix: kexec stuff ?
Apr 17 22:57:35 <denix> khem: not sure, maybe...
Apr 17 22:57:52 <ant__> kergoth: can you imagine race-conditions during rebuild from pstage? it seems PARALLEL_MAKE = "" would solve an odd issue here...
Apr 17 22:57:57 <Tartarus> yes, thats it
Apr 17 22:58:02 <Tartarus> 1st hit in google is kexec :)
Apr 17 22:58:49 <mwester> /proc/atags is related to kexec
Apr 17 22:58:56 <Tartarus> hey mwester 
Apr 17 22:59:04 <Tartarus> see my Q re slugos && stable/2009?
Apr 17 22:59:10 <kergoth> ant__: nothing comes to mind offhand
Apr 17 22:59:15 <mwester> if you don't have kexec enabled in your defconfig, you won't even see the option for /proc/atags.
Apr 17 22:59:40 <mwester> I have not been building SlugOS on stable/2009 -- thanks for the reminder; I need to set up the autobuilder to do so.
Apr 17 22:59:41 <mwester> :)
Apr 17 22:59:46 <mwester> I'll work on that this weekend.
Apr 17 23:00:02 <ant__> kergoth: I'll commit it..is just zaurus-updater
Apr 17 23:00:17 * kergoth is so annoyed at bb.fetch right now, he can barely think straight
Apr 17 23:00:46 <ant__> however now console-image fails on sysvinit
Apr 17 23:01:12 <ant__> ah..commit of today...
Apr 17 23:01:40 <ant__> error is: http://tinderbox.openembedded.net/builds/139984/
Apr 17 23:02:02 <ant__> culprit? is http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&id=54c2f02920fde00de5f9d260173b6bbe1de8426a
Apr 17 23:03:06 <ant__> hm  sounds ccache error...
Apr 17 23:03:23 <Tartarus> mwester, k
Apr 17 23:03:34 <Tartarus> Do you want to pull in your SDK fix or should I put it with my SDK updates?
Apr 17 23:04:03 <mwester> Go ahead and put it with your updates; might as well get it all in one neat tidy commit.
Apr 17 23:04:28 <Tartarus> k
Apr 17 23:04:37 <Tartarus> time to interrupt the big loop and add a task :)
Apr 17 23:05:27 <kergoth> whew, think i have a workaround for this fucking thing
Apr 17 23:08:35 <kergoth> whew, that was a pain in the ass
Apr 17 23:08:46 * kergoth 's glad he's getting back up to speed quickly
Apr 17 23:12:57 <dhd> /a/way
Apr 17 23:13:00 <dhd> oops!
Apr 17 23:13:18 <dhd> Tartarus: thanks (in advance) for pushing the Canadian stuff to stable
Apr 17 23:13:21 <CIA-2> 03kergoth 07bitbake-1.8 * r1161 10/lib/bb/fetch/git.py: 
Apr 17 23:13:21 <CIA-2> bb.fetch.git: make empty host pass a file uri to git.
Apr 17 23:13:21 <CIA-2> Signed-off-by: Tom Rini <trini@embeddedalley.com>
Apr 17 23:13:21 <CIA-2> Signed-off-by: Chris Larson <clarson@mvista.com>
Apr 17 23:13:27 <CIA-2> 03kergoth 07bitbake-1.8 * r1162 10/lib/bb/fetch/__init__.py: 
Apr 17 23:13:27 <CIA-2> bb.fetch.go: iterate over SRC_URI, not urldata.
Apr 17 23:13:27 <CIA-2> urldata can contain urls other than just what's in SRC_URI. The urls of
Apr 17 23:13:28 <CIA-2> mirrors, premirrors, stashes, etc can end up in there as well, and we don't
Apr 17 23:13:29 <CIA-2> want bb.fetch.go to be fetching those things directly.
Apr 17 23:13:31 <CIA-2> Signed-off-by: Chris Larson <clarson@mvista.com>
Apr 17 23:13:59 <CIA-2> 03kergoth * r1163 10bitbake/lib/bb/fetch/git.py: 
Apr 17 23:13:59 <CIA-2> bb.fetch.git: make empty host pass a file uri to git.
Apr 17 23:13:59 <CIA-2> Signed-off-by: Tom Rini <trini@embeddedalley.com>
Apr 17 23:13:59 <CIA-2> Signed-off-by: Chris Larson <clarson@mvista.com>
Apr 17 23:14:11 <CIA-2> 03kergoth * r1164 10bitbake/lib/bb/fetch/__init__.py: 
Apr 17 23:14:11 <CIA-2> bb.fetch.go: iterate over SRC_URI, not urldata.
Apr 17 23:14:11 <CIA-2> urldata can contain urls other than just what's in SRC_URI. The urls of
Apr 17 23:14:11 <CIA-2> mirrors, premirrors, stashes, etc can end up in there as well, and we don't
Apr 17 23:14:11 <CIA-2> want bb.fetch.go to be fetching those things directly.
Apr 17 23:14:13 <CIA-2> Signed-off-by: Chris Larson <clarson@mvista.com>
Apr 17 23:14:15 <kergoth> whee
Apr 17 23:15:51 <Tartarus> Yay :)
Apr 17 23:16:37 <kergoth> that urldata thing was really, really annoying.  surprised it hasn't caused issues before now
Apr 17 23:17:13 * kergoth thinks he's almost back to 100% when it comes to OpenEmbedded
Apr 17 23:19:04 <ant__> 'nite folks
Apr 17 23:21:11 <CIA-2> 03kergoth 07bitbake-1.8 * r1165 10/lib/bb/cooker.py: 
Apr 17 23:21:11 <CIA-2> Fix rdepends in graphviz graph generation.
Apr 17 23:21:11 <CIA-2> Signed-off-by: Roman Khimov <roman@khimov.ru>
Apr 17 23:21:11 <CIA-2> Signed-off-by: Chris Larson <clarson@mvista.com>
Apr 17 23:21:59 <kergoth> hmm, mayb ei should revert this
Apr 17 23:22:11 <kergoth> i didn't realize, bu tit looks like its fixed already, and differently, in trunk
Apr 17 23:23:10 <kergoth> eh, the trunk commit is a bit more invasive, probably best off not pulling it into 1.8
Apr 17 23:23:15 <kergoth> at least the bug is worked around now
Apr 18 00:26:22 * kergoth thinks about trying to rewrite bb.fetch
Apr 18 00:37:48 * kergoth wonders what itd take to add git/cvs/etc as urlopeners to urllib2
Apr 18 01:41:38 <svolpe> which is better to use in Angstrom eglibc or uclibc, I have used uclibc in the past but I have never used eglibc.
Apr 18 01:57:46 * odonell waves
Apr 18 01:58:32 <odonell> I have some OE terminology questions. Recently I've seen the packages directory change to recipes, is that correct? Technically the same thing, just renamed the directory?
Apr 18 01:58:37 <CIA-2> 03Carsten Haitzler (Rasterman <raster@rasterman.com> 07org.openembedded.dev * r523f0e2302 10openembedded.git/recipes/libelf/libelf_0.8.6.bb: libelf: fix another GNU_HASH issues while building libelf
Apr 18 01:59:20 <raster> svolpe: eglibc?
Apr 18 01:59:55 <raster> i've never touched it. i use regular glibc and it does the job well
Apr 18 02:00:07 <raster> odonell:  yes.
Apr 18 02:00:49 <odonell> raster, eglibc is source compatible with glibc, as long as you leave all of the options turned on. It's almost glibc, but with some additional patches maintained by CodeSourcery.
Apr 18 02:01:07 <odonell> raster, CodeSourcery's official producst use EGLIBC.
Apr 18 02:01:35 <raster> then.. why use it instead of glibc?
Apr 18 02:01:45 <raster> (that at least as best i can tell is working just fine)
Apr 18 02:02:22 <odonell> You can tune it for embedded systems, tear out various pieces you don't need.
Apr 18 02:02:37 <raster> i'd be wary of tearing out pieces of glibc
Apr 18 02:02:50 <raster> u'd be surprised how much userspace depends on it
Apr 18 02:03:02 <raster> unless of course u are runing on some woefully tiny embedded system
Apr 18 02:03:05 <odonell> In an embedded linux system you can control taht.
Apr 18 02:03:32 <raster> isnt the point of linux to be able to re-use so much of the good software already there... without having to port?
Apr 18 02:03:35 <raster> or re-do it?
Apr 18 02:04:14 <odonell> And we align the option groups (things you tear out), with various POSIX option groups.
Apr 18 02:04:33 <raster> ie to improve the long-term maintainability by being able to simply inherit work from previous products directly and easily as well as ggraft in any new oss code and improve time to market by buildingon existing oss stuff
Apr 18 02:04:52 <odonell> eglibc merges from glibc regularly.
Apr 18 02:05:00 <odonell> It just has different goals.
Apr 18 02:05:03 <raster> sure
Apr 18 02:05:06 <raster> but what does it buy you?
Apr 18 02:05:24 <raster> if all you are doing is chopping bits out of glibc (effectively)?
Apr 18 02:05:26 <odonell> better support for cross-compiling and cross-testing.
Apr 18 02:05:31 <raster> ok - i'll use less flash for my libc
Apr 18 02:05:33 <odonell> Is probably the top of the list.
Apr 18 02:05:35 <raster> thats about it
Apr 18 02:06:04 <odonell> It's more than that. uClibc and newlib aren't quite API compatible with applications that expect glibc.
Apr 18 02:06:04 <raster> hmm to date - glibc has behaved well for me under OE for cross-compiling
Apr 18 02:06:06 <kergoth> hey raster, how goes it?
Apr 18 02:06:14 <raster> that has been one of things not giving me problems
Apr 18 02:06:17 <raster> kergoth: yo!
Apr 18 02:06:38 <raster> odonell:  oh yes. i know. i dont really recommend ucliubc as a result.
Apr 18 02:06:55 <odonell> I have other terminology questions regarding OE. I don't quite understand where the meta-toolchain package fits into the whole picture.
Apr 18 02:06:59 <raster> if apps go "ooh i'm on linux" they tend to go "ooh and so i have glibc!"
Apr 18 02:07:11 <odonell> Yes. That's true linux apps assume glibc.
Apr 18 02:07:37 <raster> kergoth: been busy fixing up angstrom. it's time it got some love. especially for some targets peolpe dont seem to care about
Apr 18 02:07:38 <odonell> uClinux-dist tries to solve some of those problems by providing a set of patches sources.
Apr 18 02:08:18 <odonell> Is an external toolchain very different from a meta-toolchain?
Apr 18 02:08:23 <raster> odonell:  dont know about the meta-toolchain stuff... never has it wafted around under my nose.
Apr 18 02:08:25 <kergoth> heh, thats always true, things only get fixed that individuals care about.  such is open source
Apr 18 02:08:42 <raster> all i know is oe builds itself its own toolchain for your target
Apr 18 02:08:50 <raster> and that seems to never be a problem
Apr 18 02:09:14 <raster> kergoth:  yeah. for a while mickey scared me by saying "you must coordinate etc. etc."
Apr 18 02:09:20 * kergoth chuckles
Apr 18 02:09:21 <odonell> Right, except that for example, I have a toolchain already built, which has been verified by running tests on hardware, and I'd like to use that to build OE :-)
Apr 18 02:09:23 <raster> so i basically never pushed my changes
Apr 18 02:09:43 <raster> but after a while i realised.. at least oe.dev git.. is possibly more broken than e svn...
Apr 18 02:09:48 <raster> people push willy-nilly
Apr 18 02:10:06 <raster> push things that are broken (missing files/dependencies etc.)
Apr 18 02:10:17 <raster> so i cant make the git repo any worse by my pushing :)
Apr 18 02:10:52 <raster>  odonell in theory u can replace the oe built toolchain with an external one
Apr 18 02:10:55 <raster> somehow
Apr 18 02:10:59 <raster> i dont know how - iv'e never done it
Apr 18 02:11:20 <kergoth> some people bitch and whine if dev breaks, but i think it not breaking is a pipe dream.  it's the active line of development, shit happens.  it's much more difficult than for a normal oss project, since it's very difficult ot test every combination of target and build setup
Apr 18 02:11:26 <raster> i am very happy letting oe build my toolchain for me as to date it's done a good job of it for everything from armv4 up to cortex-a8
Apr 18 02:11:34 <odonell> Does OE run the toolcahin testsuite on your target to verify the package it just built is working?
Apr 18 02:11:36 <kergoth> shit happens, we try to keep it stable, and fix it fast when it breaks, anything else is a fantasy
Apr 18 02:12:05 <raster> odonell:  you cant as oer is not dependant on there being a target
Apr 18 02:12:20 <raster> it runs entirely 100% detached from the target.
Apr 18 02:12:40 <raster> kergoth:  thats what i noticed. it was very much different to what iwas fed when getting oe commit access
Apr 18 02:12:43 <odonell> OK, so nothing is valid.
Apr 18 02:12:49 <odonell> Sorry, "validated"
Apr 18 02:13:02 <kergoth> there's a stable branch now, and i wish them luck
Apr 18 02:13:13 <raster> odonell: i do actually reflash/run my images i build on my targets. they do work. 
Apr 18 02:13:20 <raster> oe has earnt my trust.
Apr 18 02:13:29 * kergoth still thinks more individual responsibility and splitting up actively individually maintained recipes from the recipes maintained by the entire team would do a lot more for stability than trying to keep the whole world stable
Apr 18 02:13:42 <raster> kergoth: i know. unfortunately the stable branch is even relatively slow in picking up changes
Apr 18 02:13:45 <raster> it pickes some up fast
Apr 18 02:13:48 * kergoth nods
Apr 18 02:13:51 <raster> but others.. i doubt it will ever merge
Apr 18 02:14:24 <raster> right now - i just focus on anything that comes from my ugly code factory
Apr 18 02:14:26 <raster> ie e, and efl
Apr 18 02:14:39 <raster> and my own illume image thats really just my personalyl test-bed for it all
Apr 18 02:14:58 <odonell> Where can one talk about packages that don't seem to buidl?
Apr 18 02:15:18 <raster> and i build that for several targets (om-gta02, palmt650, nokia800, rokre6, overo)
Apr 18 02:15:19 <odonell> Or, recipes.
Apr 18 02:15:27 <kergoth> if a package isnt' building, open a bug, or check MAINTAINERS and contact its maintainer, or both
Apr 18 02:15:34 <raster> odonell:  #oe, oe mailing lists... 
Apr 18 02:15:43 <kergoth> or fix it yourself, which would be optimal, as itd teach you more about OpenEmbedded and solve the problem
Apr 18 02:15:50 <kergoth> :)
Apr 18 02:16:09 <odonell> I find it useful to coordinate somewhere so everyone isn't fixing the same thing.
Apr 18 02:16:39 <kergoth> that's highly unlikely, unless it takes you weeks to fix.  again, open a bug, contact the maintainer, or as rasters says, use the mailing list or talk about it here
Apr 18 02:18:21 <odonell> For example, has anyone noticed breakage in libtool-cross-2.2.4? It configures with --host=$target, and tries to run target executables on the host.
Apr 18 02:19:13 <raster> hmm
Apr 18 02:19:16 <raster> havent noticed it
Apr 18 02:20:30 <kergoth> huh?
Apr 18 02:20:36 <raster> it built find ro me and worked - without complaint
Apr 18 02:20:48 <kergoth> libtool doesn't run target executables.  it doesn't run executables at all.  it calls the linker and compiler with some arguments
Apr 18 02:21:04 <kergoth> you're misinterpreting whatever problem you're seeing.  show us logs and describe the exact behavior you're seeing
Apr 18 02:21:14 <raster> yeha
Apr 18 02:21:20 <raster> libtool is shell-hell
Apr 18 02:21:24 <raster> its alll shellscripts
Apr 18 02:21:34 <raster> (*and boy does it make compiling go slooooooooooooooooooooooow)
Apr 18 02:22:00 <odonell> NOTE: package libtool-cross-2.2.4-r23: task do_configure: started
Apr 18 02:22:28 <kergoth> if you have a log, put it in a pastey
Apr 18 02:22:40 <kergoth> raster: tried dolt?  seems interesting
Apr 18 02:23:02 <odonell> Ah.
Apr 18 02:23:30 <odonell> OK, checking config.log, it seems like upstream OE has changed the way it constructs the cross-compiler name.
Apr 18 02:23:52 <odonell> TARGET_VENDOR no longer seems to be honoured.
Apr 18 02:24:16 <raster> kergoth: dolt.. is it a libtool droping?
Apr 18 02:24:19 <raster> der
Apr 18 02:24:21 <raster> eerrr
Apr 18 02:24:24 <raster> drop-in?
Apr 18 02:24:32 <kergoth> raster: http://lists.debian.org/debian-devel/2008/04/msg00286.html
Apr 18 02:24:41 <raster> or another jam/cmake/scons/whatever  ball of fun?
Apr 18 02:24:44 <kergoth> the poky libtool package has a patch to enable it and stuff
Apr 18 02:25:08 <kergoth> no, it just generates a script to compile the bits at configure time rather than at runtime via libtool
Apr 18 02:25:08 <raster> hmm
Apr 18 02:25:12 <kergoth> only actually calls libtool for linking
Apr 18 02:25:12 <raster> drop-in eh...
Apr 18 02:25:16 <kergoth> its a start, speeds up some builds
Apr 18 02:25:48 <raster> hmm
Apr 18 02:25:55 <raster> even linking - libtool could be ditched
Apr 18 02:26:06 <raster> linux doesnt need any of libtool's magic
Apr 18 02:26:09 <raster> well - h don't
Apr 18 02:26:12 <raster> err i don't
Apr 18 02:26:27 <raster> gcc and glibc happily do all the things i need
Apr 18 02:26:36 <kergoth> the .la crap needs to die, painfully
Apr 18 02:26:37 <kergoth> but quickly
Apr 18 02:26:39 <raster> libtool is just a big fat blob of sh to get in the way
Apr 18 02:26:41 <raster> oh hell yeah
Apr 18 02:26:42 <raster> .la's
Apr 18 02:26:47 <raster> i cant turn it OFF in libtool
Apr 18 02:27:06 <raster> there isnt a --dont-bloody-produce-la-files-you-moron
Apr 18 02:27:08 <raster> option
Apr 18 02:27:09 <odonell> My build/conf/local.conf contains TARGET_VENDOR="-none", is this still used?
Apr 18 02:27:27 <raster> why does anyone need these things on a modern os?
Apr 18 02:27:39 <mib_4xh10a> what options do i need to set to enable udev in kernel
Apr 18 02:27:41 <odonell> I need the target to read arm-none-linux-gnueabi, what might override this? The distro settings?
Apr 18 02:27:49 <raster> on i've never set the vendor myself
Apr 18 02:28:06 <odonell> OK. Yes.
Apr 18 02:28:23 <odonell>  openembedded/conf/distro/include/angstrom.inc overrides my TARGET_VENDOR settings with "-angstrom"
Apr 18 02:28:31 <odonell> Hrm.
Apr 18 02:30:48 <odonell> I would still like to use the angstrom distro, but specify a binary toolchain to build it.
Apr 18 02:32:05 <odonell> The distro settings should probably be loaded first, and allow the user to override them.
Apr 18 02:32:22 <raster> kergoth:  bugger. no dolt on ubuntu yet
Apr 18 02:32:26 <kergoth> how is that going to happen, when its usually the local.conf that defines DISTRO so it knows what to load, odonell
Apr 18 02:32:52 <kergoth> try overriding TARGET_PREFIX directly, I'd say
Apr 18 02:33:02 <odonell> kergoth, Great hint, thanks.
Apr 18 02:33:05 <kergoth> np
Apr 18 02:33:20 <kergoth> and if you're curious about how TARGET_VENDOR is used, and how TARGET_PREFIX is set, read bitbake.conf
Apr 18 02:33:25 <odonell> kergoth, You split the conf file in two, or parse it twice.
Apr 18 02:33:54 <kergoth> that's needless complexity, particularly given 90% of the cases that you're looking to solve can be worked around by having the distro .conf use ?=
Apr 18 02:34:16 <kergoth> and parsing it twice could actually cause breakage, due to things like +=, =+, .=, =., _prepend, _append, etc
Apr 18 02:37:01 <odonell> I'm assuming all of this is shell? Parse the conf in a subshell, which returns the distro, and leaves the current shell vars untouched, then source it into the current shell and continue processing.
Apr 18 02:37:35 <odonell> For safety you can use subshells to do things that might otherwise effect current shell vars.
Apr 18 02:38:06 <kergoth> no, it isnt.
Apr 18 02:38:10 <kergoth> bitbake metadata is directly parsed
Apr 18 02:38:17 <kergoth> if you want to rewrite the parser, send a patch
Apr 18 02:38:35 <odonell> OK, if it's not shell, then my suggestion is not at all useful.
Apr 18 02:39:02 <kergoth> no, please, send a patch. the parser sucks, it needs to be rewritten :D
Apr 18 02:39:57 <odonell> At this point I expect raster to comment about "to be able to re-use so much of the good software already there" ... why not use a shell to do all of this?
Apr 18 02:40:30 <odonell> Any POSIX shell supports =, +=, arrays, functions, variables, etc.
Apr 18 02:41:01 <odonell> kergoth, Thanks. Overriding TARGET_PREFIX allows my build to successfully continue.
Apr 18 02:41:07 <kergoth> glad to hear it
Apr 18 02:41:15 <kergoth> we specifically chose not to use shell
Apr 18 02:41:19 <kergoth> portage used shell, we scrapped it
Apr 18 02:41:29 <odonell> May I ask why?
Apr 18 02:41:30 <kergoth> directly parsed gives us more control, and lets us convert to another format in the future
Apr 18 02:41:36 <kergoth> you aren't locked in
Apr 18 02:41:49 <kergoth> I'm pretty sure all of the original design discussions are still buried in a wiki somewhere
Apr 18 02:42:18 <kergoth> we're also able to write tasks in python, inline python scripts, and embed python into variable references.
Apr 18 02:42:45 <odonell> Yes, I've used zope, and the ability to use scraps of python is quite useful.
Apr 18 02:42:57 <odonell> Why not write it all in python then?
Apr 18 02:43:04 <odonell> I digress.
Apr 18 02:43:17 <odonell> I don't mean to have a design discussion like this ;_
Apr 18 02:43:33 <kergoth> fewer people know python.  this way the hard work can be done in python tasks in classes, but the user can still write do_configure, do_compile, etc in shell
Apr 18 02:43:37 <kergoth> it's been working well thus far
Apr 18 02:43:45 <kergoth> that aspect, anyway :)
Apr 18 02:44:01 <raster> its relatively good
Apr 18 02:44:12 <raster> it means peolpe familiar with build systems
Apr 18 02:44:20 <raster> and even their own src are able to re-use their normal shell
Apr 18 02:44:21 <kergoth> believe me, i have *long* list of things that bug me about OpenEmbedded, but that isn't one of them
Apr 18 02:44:36 <raster> (someo-ne who doesnt know shell doign this kind of stuff should be shot)
Apr 18 02:44:41 <raster> and it makes it comfortable
Apr 18 02:45:00 <raster> but for those willing to go the extra mile - they can do some python goop
Apr 18 02:45:01 <raster> :)
Apr 18 02:45:07 <raster> i've never needed to myself.
Apr 18 02:45:58 <odonell> So far nothing has really bugged me about OE. It works with some minimal tweaking. My only complaint is the massive amount of non-standard terminlogy and concepts that I don't understand.
Apr 18 02:46:11 <raster> odonell:  how long have u been using oe now?
Apr 18 02:46:18 <odonell> recipes, meta-toolchains, sdk's (not really sdks).
Apr 18 02:46:32 <odonell> I've only been using OE for a couple of months.
Apr 18 02:46:49 <odonell> I've probably just missed the "Terms" wiki page.
Apr 18 02:48:17 <kergoth> https://docs.google.com/View?docID=dzb388z_67cb57qmgq
Apr 18 02:48:49 * kergoth adds new stuff to his list weekly
Apr 18 02:48:51 <raster> odonell: hmm and no complaints? wow!nothing bugs you?
Apr 18 02:49:01 <odonell> Nothing yet.
Apr 18 02:49:05 <odonell> I'm hard to irk.
Apr 18 02:49:13 <raster> wow
Apr 18 02:49:14 <odonell> I'm a glibc ports developer...
Apr 18 02:49:17 <raster> i must be easy to irk
Apr 18 02:49:31 <odonell> I learned to develop alongside the core glibc maintainers which were always a pain.
Apr 18 02:49:38 <raster> i have come to respect oe - but i still have my disagreements with it
Apr 18 02:49:43 <kergoth> odonell: i dont envy you
Apr 18 02:49:59 <raster> even just getting oe set up and working is a small mountain of work
Apr 18 02:50:28 <odonell> I recently added external binary toolchain support to uClinux-dist (for the Sourcery G++ tools) and it was a hassle. OE was a breeze.
Apr 18 02:50:56 <odonell> uClinux-dist runs the compiler through a frontend program which munges your command line.
Apr 18 02:50:58 <kergoth> set up should be slightly easier now, though the docs haven't been updated to reflect it.  can just set COLLECTIONS to avoid setting BBFILES, BBFILE_{PATTERN,PRIORITY}_*, BBFILE_COLLECTIONS.  or can let it default COLLECTIONS based on BBPATH
Apr 18 02:51:07 <kergoth> i need to commit some bits to speed up the collections bits though
Apr 18 02:51:30 <odonell> I still don't understand what a meta-toolchain is, why I would want one, and how to use it.
Apr 18 02:52:02 <odonell> I have a known good binary toolchain, and I want to build all my recipes with it, and then later install the appropriate sysroot to the target image.
Apr 18 02:52:42 <kergoth> neither do i, I've never used that stuff.  i expect its for producing an external toolchain/sdk of some sort.  *shrug*
Apr 18 02:52:44 <odonell> The later is probably going to be a manual step, unless meta-toolchain solves that e.g. installing headers and the correct sysroot to the target (before image compression or creation time).
Apr 18 02:53:58 <odonell> Well you've all been a big help.
Apr 18 02:54:03 <odonell> Thanks!
Apr 18 02:54:09 <kergoth> there is an external-toolchain recipe that can use an external toolchain and generate target packages from it
Apr 18 02:54:12 <kergoth> if it was built by OpenEmbedded
Apr 18 02:54:21 <kergoth> I'm not sure if thats the best way to go, but i remember seeing it
Apr 18 02:54:23 <odonell> That sounds useful.
Apr 18 02:54:59 <kergoth> at any rate, if you come across anything you think would help improve bitbake or oe, let us know
Apr 18 02:55:06 * kergoth is trying to slowly introduce some changes
Apr 18 02:55:26 <odonell> The external-toolchain recipe would have to know 1) My CFLAGS, 2) Copy the correct sysroot based on CFLAGS, 3) Copy the headers.
Apr 18 02:55:35 * odonell looks
Apr 18 02:56:19 <odonell> How do I use external-toolchain?
Apr 18 02:56:51 <kergoth> no idea.  maybe its in the OpenEmbedded manual, otherwise read the recipe :)
Apr 18 02:57:03 <kergoth> I'm sure someone knows, but the channel is usually dead around this time
**** ENDING LOGGING AT Sat Apr 18 02:59:57 2009