**** BEGIN LOGGING AT Wed Apr 15 02:59:57 2009 Apr 15 04:16:02 03Carsten Haitzler (Rasterman  07org.openembedded.dev * rada816f513 10openembedded.git/recipes/sysstat/sysstat_8.1.6.bb: sysstat: fix build on omap3 to actually work (gnu_hash problem like ltrace) Apr 15 04:23:58 03Koen Kooi  07fso/milestone5.5 * rfa029f4bf5 10openembedded.git/ (conf/checksums.ini recipes/sqlite/sqlite3_3.6.5.bb): sqlite3: update to 3.6.5 Apr 15 04:24:12 03Angus Ainslie  07fso/milestone5.5 * rbb7d29d1b8 10openembedded.git/recipes/tasks/task-openmoko-feed.bb: task-openmoko-feed : add fennec to the feed Apr 15 06:07:43 03Carsten Haitzler (Rasterman)  07org.openembedded.dev * r6f4ba50ad3 10openembedded.git/recipes/xserver-kdrive-common/ (xserver-kdrive-common/Xserver xserver-kdrive-common_0.1.bb): xserver-kdrive-common: fix Xserver script to detect overo and run xorg Apr 15 06:37:48 03Koen Kooi  07stable/2009 * rbbf4c102a8 10openembedded.git/recipes/gdbm/gdbm_1.8.3.bb: Apr 15 06:37:48 gdbm: use LDFLAGS (fixes QA error) Apr 15 06:37:48 Signed-off-by: Koen Kooi Apr 15 06:37:48 Acked-by: Denys Dmytriyenko Apr 15 06:37:58 03Carsten Haitzler (Rasterman)  07stable/2009 * rf8bdf2dd32 10openembedded.git/recipes/ltrace/ltrace_0.4.bb: Apr 15 06:37:58 ltrace: fix build (at least for omap3) Apr 15 06:37:58 Signed-off-by: Koen Kooi Apr 15 06:37:58 Acked-by: Denys Dmytriyenko Apr 15 07:10:14 good morning Apr 15 07:13:58 hmm, my email to openembedded-devel was rejected Apr 15 07:16:22 akheron: ML si not moderated, so why you say that? Apr 15 07:16:30 s/si/is Apr 15 07:17:15 I got a reply: "You are not allowed to post to this mailing list, and your message has been automatically rejected." etc Apr 15 07:17:48 akheron: did you subscribe before? Apr 15 07:17:51 does it check the From: address against all subscribers? Apr 15 07:18:03 03Koen Kooi  07org.openembedded.dev * r272ae318b0 10openembedded.git/ (7 files in 3 dirs): udev: add 141, needs more work since it only creates static libs and the OE-only kernel match patch needs to get evaluated Apr 15 07:18:07 yes, but my subscriber address is of the form user+openembedded-devel@domain.com Apr 15 07:18:23 to automatically move the messages to openembedded-devel box Apr 15 07:19:09 akheron: I think that's not important Apr 15 07:19:33 akheron: did you receive a subscription confirmation? Apr 15 07:19:55 yes, and I can log in to mailman Apr 15 07:20:26 akheron: did you acknowledge the subscription confirmation? Apr 15 07:20:30 actually, I've subscribed already in Oct 2008 Apr 15 07:20:41 and have been receiving mail from the list since that Apr 15 07:20:59 so there should be no problem with the subscription Apr 15 07:21:28 akheron: so you was already subscribed with the same address, wasn't you? Apr 15 07:24:32 hmm? Apr 15 07:25:13 i'm subscribed with user+openembedded-devel@domain.com and sent mail From: user@domain.com Apr 15 07:25:34 To: openembedded-devel@lists.openembedded.org Apr 15 07:26:56 ... and was rejected Apr 15 07:37:16 akheron: of course it was rejected, you must use the same address you used to subscribe. Apr 15 07:46:46 ok, so the + part matters after all Apr 15 08:39:01 mckoan, console-image build and seem work of Kaleios distro for qemux86 target ... now building x11-image Apr 15 08:49:17 gremlin[it]: thanks Apr 15 08:49:45 gremlin[it]: I successfully built x11-image yesterday night with stable branch Apr 15 09:03:50 morning Apr 15 09:39:47 03Dirk Opfer  07org.openembedded.dev * rf374f6e0e6 10openembedded.git/recipes/uclibc/uclibc.inc: uclibc: use LDFLAGS for ldd and ldconfig (fixes QA error) Apr 15 09:49:28 mckoan, x11-image with KaleiOS work better than Angstrom for me !!! Apr 15 09:49:54 hrw|gone, If you wish to have a look, I have a couple of patches for directfb (a 1.3.0 recipe) and cairo-directfb (a 1.8.0 recipe) into git://repo.or.cz/openembedded/mini2440.git head "for-upstream"... these were against stable/2009 but they go into -dev of course. I separated the 'checksum' patch because it might not apply easy on -dev Apr 15 09:50:02 - progressbar on splash don't increase Apr 15 09:50:08 + keyboard work Apr 15 09:50:18 mouse don't work (as angstrom) Apr 15 10:04:00 gremlin[it]: ok, this afternoon I'll test it on a vortex86sx Apr 15 10:04:52 mckoan, probably adding modules for ps/2 mouse on the kernel and/or final image will make also mouse work Apr 15 10:11:14 gremlin[it]: may be Apr 15 10:13:29 03Florian Boor  07org.openembedded.dev * rdf32f8c84d 10openembedded.git/recipes/images/initramfs-image.bb: initramfs-image: Make sure static busybox gets into the initfs image. Apr 15 10:13:30 03Florian Boor  07org.openembedded.dev * r99fc9b7e92 10openembedded.git/conf/machine/micro2440.conf: micro2440.conf: New device: FriendlyARM micro2440 Apr 15 10:13:30 03Florian Boor  07org.openembedded.dev * r0328f2d784 10openembedded.git/recipes/linux/ (14 files in 2 dirs): linux: initial micro2440 support for 2.6.29 using mini2440 kernel patches. Apr 15 10:14:57 florian, didn;t knew you have write access. care to pull the rest of my mini patches ? Apr 15 10:16:29 BusError: sounds like a good idea. let me take a look Apr 15 10:16:36 * florian always likes patches Apr 15 10:17:48 florian: good morning Apr 15 10:17:56 03Koen Kooi  07org.openembedded.dev * r1705a068b0 10openembedded.git/ (conf/checksums.ini recipes/haserl/haserl_0.9.25.bb): haserl: update to 0.9.25 Apr 15 10:17:57 03Koen Kooi  07org.openembedded.dev * r3b164e7cc7 10openembedded.git/recipes/packagekit/packagekit_0.4.6.bb: packagekit: fix RDEPENDS Apr 15 10:17:57 03Koen Kooi  07org.openembedded.dev * re91620e695 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 15 10:17:58 03Koen Kooi  07org.openembedded.dev * r878f0abaac 10openembedded.git/ (conf/checksums.ini recipes/gnome/gnome-bluetooth_2.27.4.bb): gnome-bluetooth: add 2.27.4 Apr 15 10:18:15 pb_: hillct Apr 15 10:18:16 florian, check the dev-mini-stable branch on my OE tree. Apr 15 10:18:22 hi pb_ :) Apr 15 10:18:26 heh Apr 15 10:18:37 gnaa Apr 15 10:18:57 BusError: ok, i'll take look in a few minutes Apr 15 10:19:17 as soon as I found out why this stupd device here fails to start! Apr 15 10:35:58 hello Apr 15 10:36:04 could someone help me ? Apr 15 10:36:20 flameman: just ask Apr 15 10:36:31 i want to use the bitbake to check out kernel tree and patches for zaurus/akita Apr 15 10:37:01 i launched -c fetch virtual/kernel Apr 15 10:37:18 an the bitbake fetches Apr 15 10:37:47 then i launched -c prepare but it issues errors that i can't understand Apr 15 10:37:55 is this procedure right ? Apr 15 10:39:01 '-c prepare' doesn't sound right. where did you get that from? Apr 15 10:39:14 gremlin suggestion Apr 15 10:39:16 '-c configure' might be better Apr 15 10:39:25 if you just want to patch, try '-c patch' Apr 15 10:39:30 or yes, '-c configure' Apr 15 10:39:47 i want to obtain a kernel tree for akita + patch, or ... better a patched kernel :P Apr 15 10:49:58 ok i launched -c configure :P Apr 15 10:59:27 well done Apr 15 11:27:51 03Koen Kooi  07org.openembedded.dev * r29b717d285 10openembedded.git/conf/checksums.ini: checksums: add avr32 gdb patch Apr 15 11:34:54 damn Apr 15 11:35:05 -c configure is building stuff Apr 15 11:35:10 i do not want it ! Apr 15 11:35:20 i only want to fetch the sources Apr 15 11:44:08 -c fetch Apr 15 11:44:24 i did -c fetch Apr 15 11:44:48 possibly bb is building the -native stuff to manage a kernel Apr 15 11:45:07 :P Apr 15 11:45:45 -c unpack or -c patch ? Apr 15 11:46:01 -c fetch download things from the web, umm may be i should combine fetch + patch Apr 15 11:46:12 yes, fetch, unpack, patch Apr 15 11:46:22 let me try out :) Apr 15 11:46:34 bitbake -c patch runs do_fetch and do_unpack too Apr 15 11:46:45 you don't have to do it manually Apr 15 11:47:31 ah, ok, so -c patch only Apr 15 11:47:49 flameman: You can list all tasks for a given recipe with 'bitbake -c listtasks '. Apr 15 11:48:19 good ! Apr 15 11:49:23 and_work the zaurus has been working since our last pm Apr 15 11:49:36 it seems nothing has been damaged Apr 15 11:50:04 I've found out that only C-7x0 models could have the fuse(s) blown Apr 15 11:50:09 (also the recharge led is orange ) Apr 15 11:50:28 hi all ... i remember a thread on the ML about how to clean all except native and cross tools on OE ... someone remember how to do (if it's just a couple of commend) ? Apr 15 11:50:29 mmm no, i am quite sure the c1000 has an half diode bridge Apr 15 11:50:47 in order to protect the power circuit Apr 15 11:50:56 a different beast Apr 15 11:51:16 - ok, i launched -c patch, crossing my fingers :P Apr 15 11:58:34 ok i get the 2.6.23 source tree :P Apr 15 11:58:36 yahoooooooooooooo Apr 15 11:58:57 thanks to everybody :) Apr 15 11:59:07 keep fingers crossed ... Apr 15 12:03:50 03Koen Kooi  07org.openembedded.dev * r883da3d954 10openembedded.git/recipes/python/ (2 files in 2 dirs): Apr 15 12:03:50 python 2.6.1: fix build for architectures lacking libffi support Apr 15 12:03:50 * if your patch does: +res = os.system(cmd) ; print "ret %d" % res, DO NOT change the second 'res' to 'ret'! Apr 15 12:03:50 * this fixes a regression since python 2.5.x Apr 15 12:27:43 gremlin[it]: I'm booting x11-image on vortex86sx just now Apr 15 12:29:24 mckoan, good ... now i try KaleiOS with rpm package ... Apr 15 12:36:05 hmm, building helloword-image fails on fetching http://tinylogin.busybox.net/downloads/tinylogin-1.4.tar.bz2 since its not available any more (tinylogin merged into busybox). Apr 15 12:39:25 morning Apr 15 12:39:35 m-g-l yes tinylogin no more exist (about) Apr 15 12:39:38 hi hrw ! Apr 15 12:39:50 * hrw got EDB9301 devboard today Apr 15 12:40:19 hrw, cirruslogic ? Apr 15 12:40:23 yes Apr 15 12:40:54 hrw, ... bleahhhhh i got a couple of custom EP9307 boards ... worst documentation i never saw :( Apr 15 12:41:19 gremlin[it]: you use it? Apr 15 12:41:57 hrw, yes ... i did a Linux BSP for such board ... before start use OE Apr 15 12:42:38 mainly little modification on eth, serial and i2c and custom LCD Apr 15 12:43:34 http://www.gremlin.it/download/gremlin-ep9307a.jpg Apr 15 12:43:37 http://www.gremlin.it/download/gremlin-ep9307b.jpg Apr 15 12:44:39 gremlin[it]: big Apr 15 12:44:50 edb9301 is just a bit bigger then ngw100 Apr 15 12:45:14 hrw, yes the display is a 7" ... Apr 15 12:46:12 hrw, it have a bit power supply on board ... 3xUSB 4 or 5 SERIAL ... a RF module ... some othe stuff i don't remember Apr 15 12:47:20 working on org.openembedded.dev branch I can see that IMAGE_LOGIN_MANAGER defaults to tinylogin in image.bbclass - I assume this needs to be fixed to "busybox" right? Apr 15 12:49:11 as tinylogin is now merged into busybox... Apr 15 12:53:32 I hate redboot Apr 15 12:55:23 hi, I need help on strategy: Apr 15 12:55:27 hrw, i agree !!! Apr 15 12:56:07 I've a 2.4.19-rmk7-omap1 kernel Apr 15 12:56:12 RedBoot> load -m TFTP zImage-ep9301 Apr 15 12:56:12 Unrecognized image type: 0xe1a00000 Apr 15 12:56:15 etc Apr 15 12:56:28 Gnutoo: flush it in toilet and use 2.6 Apr 15 12:56:37 * hrw -> off for some Apr 15 12:56:44 hrw, lol...I can't there are proprietary modules Apr 15 12:57:02 hrw|gone, by the way serial works with reedboot Apr 15 12:57:12 (xyzmodem transfers) Apr 15 12:58:47 so I was saying...that I've a 2.4 kernel =>no nptl Apr 15 12:59:04 so I need to chose a libc...which one should I choose? Apr 15 12:59:17 and I think I'll use 2.4 headers Apr 15 12:59:29 uclibc would work without ntpl I think Apr 15 12:59:37 but will eglibc work? Apr 15 12:59:45 recent glibc won't Apr 15 13:01:07 03Steffen Sledz  07org.openembedded.dev * r617ed6a71a 10openembedded.git/ (5 files in 5 dirs): hipox: HIPOX machine uses u-boot-2009.03 for NOR/NAND combi boot Apr 15 13:09:58 03Steffen Sledz  07org.openembedded.dev * r2ec3ed3b79 10openembedded.git/recipes/u-boot/u-boot-2009.03/hipox/00-hipox.patch: u-boot: tiny printout for HIPOX machine fixed Apr 15 13:14:45 Gnutoo: headers 2.6 + glibc 2.3.5 Apr 15 13:15:04 Gnutoo: openzaurus setting you should try to Apr 15 13:15:31 ok thanks...I tried header 2.6+uclibc but I couldn't compile modutils Apr 15 13:17:08 but will this old glibc be compatible with most of the applications? Apr 15 13:17:24 else I could find patches for glibc 2.6 Apr 15 13:17:28 s/2/6/2.5 Apr 15 13:30:49 anyone know why I am getting "AttributeError: 'NoneType' object has no attribute 'split'" from BitBake set up as per the wiki? Apr 15 13:30:58 wrong python, crappy linux distro? Apr 15 13:33:33 also is /stuff/bitbake/lib/bb/COW.py:29: DeprecationWarning: the sets module is deprecated a problem? Apr 15 13:53:48 tried with bitbake 1.8.12 and from svn and the same effect.. sigh >.< Apr 15 13:54:07 morning Apr 15 13:54:45 DeprecationWarning are harmless Apr 15 13:56:49 fair enough Apr 15 13:56:56 the attributeerror bit is annoying me though Apr 15 13:57:16 I have a developer who says he's running suse 11.0 and python 2.5 and it works.. I've got 11.1 and 2.6 and it doesn't.. Apr 15 13:57:38 bitbake shouldn't know the difference though right? the ubuntu package of bitbake didn't change when they moved to python 2.6 Apr 15 14:11:58 has anyone of you ever had this error while trying to compile a kernel (in this case in OE, using devshell) Apr 15 14:12:00 arm-angstrom-linux-gnueabi-ld: unrecognized option '-Wl,-rpath-link,/home/hakkesteegt/oe/stable_2009/build/tmp/staging/armv5te-angstrom-linux-gnueabi/usr/lib' Apr 15 14:17:48 ~seen rp Apr 15 14:17:50 rp is currently on #maemo (20h 6m 27s) #handhelds (20h 6m 27s) #oe (20h 6m 27s) #openmoko (20h 6m 27s), last said: 'hi XorA '. Apr 15 14:17:53 hi Apr 15 14:18:06 woglinde: bah you woke me :-) Apr 15 14:18:19 gm xora Apr 15 14:36:18 hi booxter Apr 15 14:36:45 hi Apr 15 14:47:38 uf... Apr 15 14:47:43 [42949376.830000] Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(2,0) Apr 15 14:47:46 progress Apr 15 14:47:47 hi hrw Apr 15 14:48:49 ~curse edb9301 for using 57600 for redboot Apr 15 14:48:50 May you be reincarnated as a Windows XP administrator, edb9301 for using 57600 for redboot ! Apr 15 14:49:22 uh Apr 15 14:56:50 fsking redboot Apr 15 14:57:09 load -v -r -b 0x01000000 /zImage-ep9301;exec -c"console=ttyAM0,57600n8 root=/dev/nfs ip=dhcp nfsroot=192.168.1.112:/srv/nfs/at91" Apr 15 15:13:38 morning Apr 15 15:14:22 hi kergoth Apr 15 15:16:37 03Carsten Haitzler (Rasterman  07org.openembedded.dev * r79e58369ba 10openembedded.git/recipes/libfribidi/libfribidi-native_0.10.4.bb: libfribidi: add native version Apr 15 15:16:38 03Carsten Haitzler (Rasterman  07org.openembedded.dev * rf7c65b66ca 10openembedded.git/recipes/efl1/ (evas-native_svn.bb evas.inc): evas: use fribidi-native package Apr 15 15:16:38 03Carsten Haitzler (Rasterman  07org.openembedded.dev * r9e28cecc90 10openembedded.git/conf/distro/include/sane-srcrevs.inc: efl: update to srcrev with fribidi build fixes and e win management fixes Apr 15 15:28:33 hi jay7 Apr 15 15:28:47 mickeyl, jay7: hi Apr 15 15:29:07 uf.. edb9301 boots Apr 15 15:29:22 2.6.29+patch Apr 15 15:30:06 Been facing following problem with gumstix oe: http://pastebin.com/m7d8ed2b4 Apr 15 15:32:46 woglinde, hrw: morning :) Apr 15 15:33:37 Any one seen this? Apr 15 15:34:45 hm seems curl is missing some deps Apr 15 15:35:13 Strange, any clues how I can check out which ones? Apr 15 15:37:13 hm mkdir is in coreutils Apr 15 15:38:08 how do I check that coreutils are ok? Apr 15 15:38:20 bitbake coreutils-native will probably help Apr 15 15:38:41 I've been building with oe to gumstix for few months, but it's still striking me some times ;) Apr 15 15:46:34 huh, going to have to read through these amd64 patches Apr 15 15:47:39 o... I am surprised Apr 15 15:47:48 edb9301 ethernet do 22Mbps Apr 15 15:49:34 hrw hm? Apr 15 15:51:01 woglinde: edb9301 has onboard ethernet which do 22Mbps (according to iperf). nice value compared to other devices Apr 15 15:55:15 okay Apr 15 16:07:06 http://marcin.juszkiewicz.com.pl/2009/04/15/edb9301-hacking/ Apr 15 16:12:11 http://www.flickr.com/photos/hrwandil/3444358903/ Apr 15 16:13:56 that has to wait until i'm home :) Apr 15 16:14:03 3k/sec no worky Apr 15 16:16:43 mickey|ICE875: Fix PPP/3G on the HTC :-) Apr 15 16:17:04 ~shoot alphaone Apr 15 16:17:05 * ibot shoots alphaone in the head with a quantum singularity weapon! Apr 15 16:17:24 :D Apr 15 16:17:59 nah, I'll rather continue to whine about it on #htc-linux ;) Apr 15 16:18:07 mickeyl: time for closed phone with 3g? Apr 15 16:18:09 mickey|ICE875: That's the spirit Apr 15 16:18:18 mickey|ICE875: So how long until they ban you? Apr 15 16:18:19 :-) Apr 15 16:18:26 ~lart llvm for using IO_putc Apr 15 16:18:26 * ibot executes killall -HUP llvm for using IO_putc Apr 15 16:19:25 hrw: looks like it. as no vendors are putting out open 3g phones, we're back to anti-vendor-ports again... welcome to 2002 *cough* Apr 15 16:19:56 2002 is better then om-gta01/02 anyway ;D Apr 15 16:20:14 hrw *g* Apr 15 16:20:16 nah Apr 15 16:20:21 that's wrong Apr 15 16:20:39 a buzz-fixed gta02 is pretty good Apr 15 16:21:10 mickey|ICE875: but has gsm modem from <2002 Apr 15 16:21:15 what is the significance of GUI_MACHINE_CLASS in machine.conf file? Apr 15 16:21:15 still I'd love to have something faster Apr 15 16:21:28 vsr: forget about it basically Apr 15 16:21:29 vsr: basically none. it's a legacy Apr 15 16:21:44 is there any way we need to specify about resolution of the lcd somewhere Apr 15 16:22:01 or this is adjusted dynamically by XServer Apr 15 16:22:17 vst hm with newer xserver and edid you shouldnt have to do this Apr 15 16:22:27 using info provided by fb driver Apr 15 16:22:28 okay Apr 15 16:22:52 there's no need to, but feel free to provide these information in the machine.conf Apr 15 16:22:58 some utilities could use it Apr 15 16:23:05 03Marcin Juszkiewicz  07org.openembedded.dev * r3bbbc7dabc 10openembedded.git/recipes/linux/ (3 files in 2 dirs): Apr 15 16:23:05 linux: make 2.6.29 default for ep93xx (tested on edb9301) Apr 15 16:23:05 - added EDB9301 support (from LAKML) Apr 15 16:23:05 - hacked machine ID of EDB9301 (board here has other then official one) Apr 15 16:23:06 like, for instance, prerendering icons during build time Apr 15 16:23:36 with xserver-xorg i have not been able to use xcalibrate. i enabled it that is Xextention init was not being called in the function which calls inits of all extention. but it is not working at all Apr 15 16:23:38 second devboard in this week took to 2.6.29 ;D Apr 15 16:24:03 hrw remove the old cruft? Apr 15 16:24:04 is there any way to get xcalibrate working with xserver-org or xserver-org used xinput extention only Apr 15 16:24:17 http://www.flickr.com/photos/hrwandil/3445158164/ and http://www.flickr.com/photos/hrwandil/3423727862/ Apr 15 16:24:33 ops.. forgot to update defconfig in OE Apr 15 16:24:46 hm so now lets see if qemu can handle llvm Apr 15 16:25:27 woglinde, hi I'm re-trying uclibc with kernel headers from 2.4.19(before it couldn't compile modutils because the kernel headers were 2.6.x) Apr 15 16:26:00 gnutoo ah right Apr 15 16:26:11 03Marcin Juszkiewicz  07org.openembedded.dev * r24f749b45d 10openembedded.git/recipes/linux/ (linux/ep93xx/defconfig linux_2.6.29.bb): linux: updated defconfig for ep93xx Apr 15 16:26:14 hm this is maybe not a uclibc problem only Apr 15 16:26:33 http://www.run.montefiore.ulg.ac.be/~martin/enp-compiling.html is nice page for RedBoot users Apr 15 16:26:39 args Apr 15 16:26:43 I forgot Apr 15 16:29:17 anyone tried to use xtscalibrate routine with xorg here? Apr 15 16:29:42 vsr sorry nope Apr 15 16:30:52 okay, i try hard to get that working since ppl are saying xserver-kdriver should not be used since no one supports that Apr 15 16:33:00 who are these people? Apr 15 16:33:07 kdrive still works fine on many platforms Apr 15 16:33:11 where as xorg may not yet Apr 15 16:33:32 xorg may be the future, but afaict it's not quite there yet Apr 15 16:34:10 yeah, kdrive is still the right answer for many situations. Apr 15 16:35:27 xorg is still quite a lot bigger, for one thing, or at least it was last time I checked. Apr 15 16:35:58 and, although it's nice to have everything be a plugin with xorg, loading all those modules is slow if you only have a simple, fixed set of hardware. Apr 15 16:36:04 okay i was scared of using kdrive since i thought it gone obselete soon Apr 15 16:36:10 yep Apr 15 16:36:28 why xorg does not support xcalibrate? Apr 15 16:36:30 pb_: what's your take on the movement of ts calibration and filtering into the kernel? Apr 15 16:36:43 kergoth: would also like to hear your opinion on that Apr 15 16:37:16 hm some qemu expert here? Apr 15 16:37:38 mickey|ICE875: I haven't been paying much attention to it, but putting it in the kernel seems like a fairly silly idea at first glance. Apr 15 16:37:58 vsr: probably just because nobody has bothered to merge it. you're welcome to do that if you want. Apr 15 16:38:52 how do to tell qemu to look up ld-uClibc.so.0 under oetmp2/staging/arm-oabi-angstrom-linux-uclibc/usr/lib/ instead of /lib/ Apr 15 16:39:17 yeah, i've been wondering that myself. they're essentially duplicating the whole tslib infrastructure in kernelspace Apr 15 16:39:17 only to bypass one userland module Apr 15 16:39:18 what do you mean by merging it, there is already a xcalibrate extention, i was in xorg-devel. i heard that it is not supported since it is supposed to be part of xinput and xinput does not have support yet since xorg is designed for desktop machines Apr 15 16:39:19 s/only/just/ Apr 15 16:39:55 mickeyl: right, it's hard to see what the benefit would be. Apr 15 16:40:24 pb_: i will look into it further and try to do that. unfortunately it does not work right now with only minor modifications Apr 15 16:40:50 I guess you could argue that it's in some way "neater" for userspace programs to be delivered the touchscreen events ready-cooked, so that the coordinates they see correspond to screen pixels, but I can't really see that it makes a lot of difference in the real world. Apr 15 16:40:57 and i don't buy the "minimize external dependencies" thing... Apr 15 16:41:14 it's not like the kernel is standalone anyways Apr 15 16:41:28 well, quite. Apr 15 16:42:00 anyway, I would have thought that touchscreen filtering was as clear an example of "policy" as one might expect to see. don't the kernel dudes have a fixation on no policy in the kernel, or did they change their mind on that? Apr 15 16:42:01 heh Apr 15 16:42:26 pb_: I completely agree -- it's totally inconsistent. Apr 15 16:43:18 I don't mind it in-kernel, but if they accept that in-kernel then there are a number of other patches that I would like to see re-submitted for acceptance into the kernel. Apr 15 16:44:14 personally I would much rather see less stuff in the kernel, not more. I have felt for a long time that things like the virtual console (with accompanying fonts, keymaps and the like) belong more in user space than in the kernel. Apr 15 16:44:30 and touchscreen support is of course much the same as that stuff. Apr 15 16:45:26 make sense, yeah Apr 15 16:48:46 quite apart from anything else, putting this sort of unnecessary thing into the kernel seems like it might be contributing to the general all-over bloat that seems to have been a feature of all recent 2.6 releases. it would be nice to do something to reverse that trend. Apr 15 16:49:11 s/2.6/2.x/ Apr 15 16:50:20 I wish only for consistency. If the argument is that is is no longer reasonable to expect that a bare kernel can boot an embedded device, and that one must provide an initrd to do so (things like setting the mac addr, loading firmware to the network hardware), then I just cannot see the reason for touchscreen handling in-kernel. Apr 15 16:50:21 have a nice rest of day Apr 15 16:50:47 mmm....| ./include/asm/param.h:14:52: error: asm/proc/page.h: No such file or directory Apr 15 16:50:56 I should look at that Apr 15 16:51:05 I bet I will have lot of similar errors Apr 15 16:51:42 mwester: yes, that's a reasonable point of view. Apr 15 16:53:21 The argument for touchscreen filtering was that it was able to cut out context thrashes bouncing through user space. Apr 15 16:53:43 broonie: that's a spurious argument though, since calling through a shared library doesn't involve a context switch Apr 15 16:53:43 So it provided a performance win that couldn't be emulated with a pure userspace solution. Apr 15 16:54:08 It wasn't the shared library, it was the I/O overhead for pulling in the points. Apr 15 16:54:34 are there any recipes that create (CF/SD/USB/etc) disk images from items in deploy, ie partition and extract, etc - I use an external script and am wondering if it would be easy to do as a recipe - would need root access to mount partitions etc Apr 15 16:54:47 * broonie notes that if this is rubbish you should push back on the patch set, I've not actually evaluated what they're doing here at all. Apr 15 16:56:32 broonie: that does sound a bit bogus to me. the overhead of a syscall is, what, a couple of hundred nanoseonds? Even if you had 1000 points per second to process, which is implausibly large for current touchscreens, you'd save less than a millisecond of system call overhead. Apr 15 16:56:50 It wasn't the syscalls themselves AIUI. Apr 15 16:56:50 so that's a lot of kernel infrastructure for something below 0.1% speedup Apr 15 16:58:09 ah. what was it, exactly? I'm struggling to think what else "context thrashes" could be referring to. Apr 15 16:58:34 if it's not context switches, and not syscalls, that doesn't seem to leave much. Apr 15 16:58:53 Scheduling the user space process. Though looking at their page it appears that it was actually just crap filtering they wanted to improve Apr 15 16:58:56 http://wiki.openmoko.org/wiki/Touchscreen_Filters Apr 15 16:59:38 Context switch in this case == thread/process switch Apr 15 17:00:45 right, I see. Apr 15 17:01:23 that rationale does seem pretty flaky to me. the only case where you would get a significant win in performance would be if you are throwing away a large percentage of the raw touchscreen events in the filtering. Apr 15 17:01:35 They are doing so, their touchscreen is crap. Apr 15 17:01:58 i.e. if you either have a filter algorithm which takes many input events to produce a single output, or if you're just discarding one in N (for some moderate N) as being complete garbage. Apr 15 17:02:14 and, in either of those cases, it sounds like you're basically onto a loser anyway. Apr 15 17:02:21 oh, right, fair enough. Apr 15 17:02:48 yeah, if your hardware is that awful then I can see it being attractive at least on paper. Apr 15 17:02:56 Yeah, if you're hardware's good you're fine. Apr 15 17:03:18 I'm still not very convinced that it would translate to a measurable real-world speedup even in the lousy hardware case. Touchscreen events just don't arrive often enough for it to be a very big deal. Apr 15 17:03:46 and, if you _do_ have a high speed stream of touchscreen events, the scheduling quantum will take over and you will end up processing multiple events per context switch anyway. Apr 15 17:04:00 You'd hope. Apr 15 17:04:08 I would. Apr 15 17:05:46 there seems to be a bit of hand waiving there Apr 15 17:05:52 "and our tests show that it's the right thing to do" Apr 15 17:05:56 and where are those test results? Apr 15 17:05:59 kergoth: yeah Apr 15 17:06:14 I was just trying to understand what they meant by "sending the event takes 8.45ms" Apr 15 17:06:20 * kergoth nods Apr 15 17:06:54 anyway, gotta go buy some curtain fabric for wifey now. Apr 15 17:06:56 later all Apr 15 17:07:08 I agree; I also suspect that the real problem is tslib filters. Apr 15 17:08:14 heh, patches welcome ;) Apr 15 17:09:32 Well, the underlying problem is the ts hardware and the hardware I have to support generally seems to deliver good quality points :) Apr 15 17:15:45 bye Apr 15 17:38:27 mmm...what are proc-armo proc-armv in the kernel headers? Apr 15 17:45:21 (what's the difference) Apr 15 17:52:25 found it Apr 15 17:54:34 armo = 26bit and armv = 32bit Apr 15 18:24:10 03Koen Kooi  07org.openembedded.dev * rabf92f42db 10openembedded.git/recipes/xorg-xserver/xserver-xorg_1.5.3.bb: xserver-xorg: make avr32 build go a bit further Apr 15 18:43:38 ~seen timtim Apr 15 18:43:42 timtim was last seen on IRC in channel #oe, 176d 21h 1m 53s ago, saying: 'my tuppence... :)'. Apr 15 18:44:00 hm Apr 15 18:45:17 anyone seen Tim Ellis lately, 2 weeks ago he said I could test the zoneminder recipe but it's not committed yet to git Apr 15 19:25:56 re Apr 15 19:26:56 pb___: wb Apr 15 19:38:45 bye Apr 15 19:54:22 gremlin[it]: any news? Apr 15 20:00:26 mckoan, no about sheevaplug ... this evening i talk with EBV guys Apr 15 20:03:25 gremlin[it]: so we'll update tomorrow Apr 15 20:05:26 mckoan, yes ... nad cross fingers when you sleep ! Apr 15 20:06:10 i do that everynight i start a build Apr 15 20:07:31 LOL Apr 15 20:08:04 usually my nightly builds are successful Apr 15 20:08:35 everything i toucch fails Apr 15 20:08:58 ntfn: keep off :-D Apr 15 20:09:03 like that skittles comercial only instead of skittles its fail Apr 15 20:09:39 but hey i just setup (included) oestats in my build env Apr 15 20:09:48 gn Apr 15 20:10:10 adios Apr 15 20:16:59 mmm...busybox fails because a syscall is not avaliable ( http://rafb.net/p/8c618C75.html ) ...what should I do? Apr 15 20:17:13 upgrade kernel? Apr 15 20:17:28 XorA, I can't Apr 15 20:17:41 I've proprietary drivers such as the framebuffer Apr 15 20:17:41 code around it then Apr 15 20:17:48 older busybox, I would expect Apr 15 20:17:49 ok but how? Apr 15 20:17:51 ok Apr 15 20:17:53 use an older busybox ? Apr 15 20:17:58 ok I'll try Apr 15 20:18:26 evrything(except the kernel installation ) was successfull until now... Apr 15 20:18:53 (there were problems but I fixed them) Apr 15 20:18:57 prefered package ver Apr 15 20:20:21 i'll try CONFIG_MONOTONIC_SYSCALL Apr 15 20:22:34 mmm I already have it Apr 15 20:23:10 ah disabling not enabling...I'll try Apr 15 20:33:42 I think that worked Apr 15 20:33:45 thanks Apr 15 20:45:40 wow....it...works!!! Apr 15 20:46:28 let's bitbake something else than the minimal image then Apr 15 21:02:35 hmm, bitbake really shouldn't dump the log with BBINCLUDELOGS when the reason for the task interruptions was a ^C Apr 15 21:19:52 03Carsten Haitzler (Rasterman  07fso/milestone5.5 * r1f792293a0 10openembedded.git/recipes/efl1/ (evas-native_svn.bb evas.inc): evas: use fribidi-native package Apr 15 21:19:52 03Carsten Haitzler (Rasterman  07fso/milestone5.5 * r510a789101 10openembedded.git/recipes/libfribidi/libfribidi-native_0.10.4.bb: libfribidi: add native version Apr 15 21:42:38 hey, is there a wiki page or website somewhere detailing the proper configuration needed to bitbake canadian-sdk? Apr 15 21:42:52 (for host=mingw32 and target=arm-linuxgnueabi) Apr 15 21:43:14 I got tripped up because I needed to define SDK_EXEEXT and I'm wondering if there is other magic Apr 15 21:46:10 bitbake task-sdk-canadian-host Apr 15 21:46:32 ah I've been bitbaking canadian-sdk Apr 15 21:46:37 I think that is the same thing, just packaged... Apr 15 21:48:02 there are a bunch of variables in canadian-sdk.bbclass, I was mostly wondering which ones I should be setting and where Apr 15 21:55:12 dhd: may be look for SDK_* Apr 15 22:00:39 03Khem Raj  07org.openembedded.dev * r8993791a6b 10openembedded.git/ (10 files in 3 dirs): Apr 15 22:00:39 quilt-0.48: Add new version 0.48 of quilt. Apr 15 22:00:39 checksums.ini: Add checksum for quilt 0.48 Apr 15 22:00:39 * Adds the latest release of quilt recipes. Apr 15 22:00:39 * Move PR out of quilt.inc into respective bb files. Apr 15 22:00:41 03Khem Raj  07org.openembedded.dev * r9396869c27 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 15 22:00:46 03Khem Raj  07org.openembedded.dev * r3189734f8d 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 15 22:00:49 03Khem Raj  07org.openembedded.dev * ra91ef39c62 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 15 22:06:22 florian, hi, first day of LE was quite fun, but really exhausting Apr 15 22:06:36 florian, I think the bootpack for TX might be ready soon Apr 15 22:06:43 btw. what about the patches ? Apr 15 22:08:47 marex: opps right, i'll take care of patches tomorrow. have some more to check in Apr 15 22:10:20 florian, any word on LinuxTag :) Apr 15 22:11:00 http://www.linuxtag.org/2009/en/about/newsarchive/entry/article/linuxtag-call-for-projects-freie-projekte-nach-berlin.html Apr 15 22:11:06 this is confusing ... Apr 15 22:11:17 Crofton|work: no sorry Apr 15 22:11:42 I suppose we have announced out participation Apr 15 22:12:57 Crofton|work: in vcc yes but not in public... and their multiple deadlines are even more confusing Apr 15 22:13:14 agreed Apr 15 22:14:02 * florian remembers he wrote the submission at the last day of the original deadline Apr 15 22:14:18 but they had multiple deadlines last year too Apr 15 22:18:10 florian, no problem :) Apr 15 22:19:27 * Jay7 -> sleep Apr 15 22:52:28 whatcha makin florian Apr 15 22:53:37 ntfn: kernel hacking, why? Apr 15 22:57:48 just wondering Apr 15 22:58:09 ntfn: or because of the deadlines? Apr 15 22:58:20 deadlines Apr 15 22:58:31 suck Apr 15 22:58:55 ntfn: ah ok, we wnat to show oe at LinuxTag Apr 15 23:00:11 time to get some sleep Apr 15 23:00:14 good night Apr 15 23:01:00 adios Apr 16 01:03:25 03Mike Westerhof  07org.openembedded.dev * rfb7eba513b 10openembedded.git/ (5 files in 3 dirs): Apr 16 01:03:25 Openmoko: linux-openmoko-2.6.29 - commit new recipe tied to upstream kernel Apr 16 01:03:25 git repo, which is currently at 2.6.29-rc3 + Openmoko patches. Note that Apr 16 01:03:25 the preferred kernel remains unchanged; select this kernel via local.conf. Apr 16 01:33:35 03Angus Ainslie  07fso/milestone5.5 * r934816eb0a 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs : bump efl Apr 16 01:40:20 03Angus Ainslie  07fso/milestone5.5 * r57f7a1a311 10openembedded.git/recipes/python/python-ecore_svn.bb: python-ecore : remove patch ecore_x_window_del Apr 16 01:40:31 03Angus Ainslie  07fso/milestone5.5 * r7692bafc07 10openembedded.git/recipes/images/ (fso-image.inc fso-paroli-image.bb): Apr 16 01:40:31 fso-paroli-image : fix fso-image.inc Apr 16 01:40:31 fso-image : remove paroli install **** ENDING LOGGING AT Thu Apr 16 02:59:57 2009