**** BEGIN LOGGING AT Sat Jul 21 02:59:58 2007 Jul 21 07:15:20 hello I got a quick question about the OE manual Jul 21 07:16:09 for some reason I can't access the page "Installation Scripts: Running Scripts during package install and/or removal" which http://www.openembedded.org/user-manual&dpage=recipes_installation_scripts Jul 21 07:21:56 Philippe : Hi! Jul 21 07:34:51 good morning all Jul 21 08:07:06 good morning all Jul 21 08:07:42 morning koen Jul 21 08:08:10 how is the weather over there in overijssel Jul 21 08:10:45 very sunny actually Jul 21 08:27:29 ~ping Crofton Jul 21 08:27:30 pong Crofton Jul 21 09:07:47 how to find which .bb file include gdb into image? Jul 21 09:33:57 Jon Bradbury is now know as jonb Jul 21 09:34:30 koen : I have installed oe and am building task-base Jul 21 09:35:02 It's taking a long time, as it appears to be downloading loads of stuff. Jul 21 09:35:20 it will only download that stuff once Jul 21 09:35:38 Yes, I realise that. It's a bit like make in that regard Jul 21 09:36:01 task-base should give me a 2.6.17 kerkel image? Jul 21 09:37:01 I am using branch org.openembedded.dev Jul 21 09:37:21 (basically following the "Getting Started" instructions in the OE wiki. Jul 21 09:39:16 Now doing task 192 of 1146 (!) Jul 21 09:39:21 bitbake virtual/kernel Jul 21 09:39:40 You mean I don't need task-base? Jul 21 09:41:20 doesn't virtual/kernel give a Zaurus image? Or is it the same as a NBP image? Or does it know it's for the NBP because I specified it in MACHINE in the local.conf? Jul 21 09:45:07 jonb : virtual/kernel will build toolchain and kernel for the machine you have specified but not a complete filesystem Jul 21 09:46:00 03koen 07org.oe.dev * r8d1228f6... 10/ (1 packages/brutefir/brutefir_1.0i.bb): brutefir: depend on all fftw flavours Jul 21 09:46:07 jonb : to get a filesystem + kernel build in one go you need bitbake angstrom-console-image or angstrom-x11-image Jul 21 09:47:23 Neither of those targets are in org.openembedded.dev/packages/meta. Jul 21 09:55:08 check packages/angstrom and packages/images Jul 21 09:58:50 koen : They are there. I thought all the bbfiles lived in org.openembedded.dev/packages/meta Jul 21 09:59:01 Oh well, it's a learning curve... :) Jul 21 09:59:19 task 247 of 1146.... Jul 21 10:58:11 I need to talk to Phillipe or florian regarding BooSt on the NetBook Pro. Jul 21 11:27:22 * koen waits for girlfriend to finish the new harry potter book Jul 21 11:29:23 lol Jul 21 11:29:53 i cant seem to get my usb stick to work :( keyboard works nice but no usb Jul 21 11:30:03 using angstrom Jul 21 11:30:21 i see its been scanned successfully using dmesg Jul 21 11:31:12 but when i put it into storage mode i get g_file_storage gadget-lun0: unable to open backing file: /dev/hda Jul 21 11:31:18 any ideas? Jul 21 11:41:21 Hello Jul 21 11:41:39 Has anyone had any problems with bitbake 1.8 trying to build glibc-intermediate-2.3.2+cvs20040726 for no reason? Jul 21 11:42:03 I use 1.8 to build an image, but it keeps trying to build glibc-intermediate-2.3.2+cvs20040726 Jul 21 11:42:14 hm Jul 21 11:42:20 why this old libc? Jul 21 11:42:27 angstroem uses 2.5 Jul 21 11:42:28 I revert (or reverted, before those new sanity tests) to 1.6, and everything worked fine Jul 21 11:42:31 Yeah Jul 21 11:42:33 Me too Jul 21 11:42:38 and this works fine Jul 21 11:42:49 Another user reported the same problem Jul 21 11:42:58 beside i havent test it in the last weeks, because I am using uclibc Jul 21 11:43:00 I wouldn't come to you if I didn't know if it was just me Jul 21 11:43:03 Aha Jul 21 11:43:10 hm put in the bugtracker Jul 21 11:43:12 please Jul 21 11:43:17 if it isnt already there Jul 21 11:43:28 I don't know the reason why it does that, but is that OK? Jul 21 11:43:31 It's not Jul 21 11:43:33 now guadec is over and rp have some time for this Jul 21 11:43:43 when it is really a bitbake error Jul 21 11:44:03 Well, it doesn't happen in 1.6, so I'll guess it's bitbake Jul 21 11:44:04 1.8 Jul 21 11:44:17 no Jul 21 11:44:21 the recipe can be false Jul 21 11:44:47 Well, if the recipe was false, then it would happen under 1.6 too, wouldn't it? Jul 21 11:45:32 not in all cases Jul 21 11:45:45 can you pastebin the error? Jul 21 11:46:05 did you try glibc 2.5? Jul 21 11:47:25 glibc is already built at full Jul 21 11:47:34 My toolchain is already set up Jul 21 11:47:52 It just happens for no apparent reason Jul 21 11:48:59 Well, there is no error, other than bitbake trying to build the package for no reason (and the package failing to configure, because of gcc, and because it's too old, but that is of no issue). Jul 21 11:50:10 which modules do i need to get my usb host working with flash drive? Jul 21 11:50:20 at the moment i only have g_ether loaded Jul 21 11:50:40 have done usb-gadget storage Jul 21 11:51:06 !oebug 2671 Jul 21 11:51:08 * * Bug 2671, Status: NEW, Created: 2007-07-21 04:50 Jul 21 11:51:09 * * B_Linuz(AT)yahoo.com: Bitbake 1.8 tries to build glibc-intermediate-2.3.2+cvs20040726 for no apparent reason. Jul 21 11:51:10 * * http://bugs.openembedded.org/show_bug.cgi?id=2671 Jul 21 11:51:17 :) Jul 21 11:52:12 These damn new sanity tests do not let me build with bitbake 1.6 Jul 21 11:54:22 You can turn them off I believe Jul 21 11:54:55 Yeap Jul 21 11:55:16 So what happens when you do? Jul 21 11:55:53 they arent damned Jul 21 11:56:15 Well, I comment the BB_MIN_VERSION and everything is fine Jul 21 11:56:26 03todd 07org.oe.dev * r7d9e3c03... 10/ (1 conf/machine/hx4700.conf): hx4700: prefer version of xserver-kdrive that has w100 support (1.2.0) Jul 21 11:56:32 03todd 07org.oe.dev * r9312d5d8... 10/ (1 conf/machine/hx4700.conf): disapproval of revision '7d9e3c0371703df6cb39dcdb41d1d9131946eb1e' Jul 21 11:56:36 03todd 07org.oe.dev * ra5720325... 10/ (1 conf/machine/hx4700.conf): Jul 21 11:56:36 hx4700.conf: prefer version 1.2.0 of xserver-kdrive since that version Jul 21 11:56:36 supports w100. Jul 21 11:56:42 03todd 07org.oe.dev * r0c6870e8... 10/ (1 conf/machine/hx4700.conf): disapproval of revision 'a57203255798a5084799d04f08eba13dacae3f56' Jul 21 11:56:47 Well, actually they aren't Jul 21 11:56:48 03todd 07org.oe.dev * r32975812... 10/ (1 conf/machine/hx4700.conf): hx4700.conf: prefer version 1.2.0 of xserver-kdrive since it supports w100 Jul 21 11:58:19 I think I couldn't change the mouse button (either with xmodmap or with xmonobut) under 1.2.0 Jul 21 11:59:13 Maybe it's just on my machine, but test if you can change the mouse buttons under 1.2.0 Jul 21 11:59:17 It might be a bug Jul 21 12:14:25 hi all! Jul 21 12:14:59 hi jan_js Jul 21 12:15:30 hi woglinde Jul 21 12:20:22 does anybody know what's the status of h3900 in angstrom ? Jul 21 12:29:29 03koen 07org.oe.dev * re520909b... 10/ (1 packages/jack/jack_0.99.0.bb packages/jack/jack_0.103.0.bb): jack: update to 0.103.0, fix packaging, remove broken 0.99.0 version Jul 21 12:29:59 * * OE Bug 2672 has been created by jstoefer(AT)schuett-elektronik.de Jul 21 12:30:01 * * angstrom-x11-image does not boot on h3850 Jul 21 12:30:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2672 Jul 21 12:32:53 how to find which .bb file include gdb into image? Jul 21 12:32:59 how to find which .bb file include gdb into image? Jul 21 12:36:48 readon run bitbake with -g Jul 21 12:36:53 and look in depends.dot Jul 21 12:39:51 psokolovsky_: I have read on the angstrom site that you are machine mentor for h3900 - do you know wether an x11 image can be built already ? Jul 21 12:43:58 moin! Jul 21 12:44:15 moin zecke Jul 21 12:45:12 Is pidgin supposed to use nspr.h? Jul 21 12:45:25 r0 didn't, but r1 uses it Jul 21 12:45:29 For some reason Jul 21 12:46:16 Specifically, the ssl-nss plugin uses it Jul 21 12:46:31 But the version is still at 2.0.2 Jul 21 12:48:50 hey zezom Jul 21 12:48:52 ups Jul 21 12:48:57 ~lart tab completion Jul 21 12:48:57 * ibot puts on a hockey mask and jumps out at tab completion Jul 21 12:49:01 hey zecke Jul 21 12:52:51 hey mickeyl Jul 21 12:53:12 mickeyl: could you respond to my reply to todd about X guis? Jul 21 12:54:27 moin koen Jul 21 12:55:13 hey woglinde Jul 21 12:57:52 yo koen. sure, once i have read it, i'll do Jul 21 12:57:59 woglinde, it is very strange that i didn't find any dependency on gdb, but it is included in my images,why? Jul 21 12:58:12 readon???? Jul 21 12:58:19 did you grep the file? Jul 21 12:58:28 aor tasks-depend.dot? Jul 21 12:58:47 yes Jul 21 12:58:52 mickeyl: thanks Jul 21 12:59:03 Readon: grep -r is your friend Jul 21 12:59:07 Pidgin 2.0.2-r1 gives me a CROSS COMPILE Badness when it tries to use /usr/include/mozilla-1.7.12/nss (which doesn't even exist) for ssl-nss.c, does anyone else get that? Jul 21 12:59:26 It didn't do that in 2.0.2-r0, so I'm guessing something changed in the bb file Jul 21 12:59:41 woglinde, i do it, it's all about gdbm but gdb Jul 21 13:00:07 koen, do you mean grep -r gdb *, to find it by myself? Jul 21 13:00:20 Readon: yes Jul 21 13:00:28 it gave me the answer in 2 seconds Jul 21 13:01:12 koen,perhaps it contains the info about gdbm. that is huge number Jul 21 13:03:15 koen: you're back home? Jul 21 13:03:45 I see that the update only changed jabber to xmpp Jul 21 13:05:49 mickeyl: yes Jul 21 13:06:08 koen: had a chance to see the clutter foo talk? Jul 21 13:06:12 yes Jul 21 13:06:14 it rocked Jul 21 13:06:23 cool. was it taped? Jul 21 13:07:51 yes Jul 21 13:08:08 excellent. Jul 21 13:08:41 we need to work on the press release about GMAE using an OE powered buildsystem Jul 21 13:08:50 "eat that, scratchbox!" Jul 21 13:09:45 haha Jul 21 13:09:45 yeah Jul 21 13:09:58 mickeyl: it's clear now that both nokia and google are doing a linux phone Jul 21 13:10:17 hence all the badmouthing of openmoko from the nokia folks Jul 21 13:10:27 i think so, too Jul 21 13:10:39 i just hope it'll take them long enough for us to rock Jul 21 13:10:51 then again, they probably won't beat our degree of freedom Jul 21 13:11:01 that's for sure Jul 21 13:11:28 * koen curses libgsmd again for /var/tmp/1337.socket instead of dbus Jul 21 13:11:45 can someone introduce harald and werner to the 21st century? Jul 21 13:12:11 koen: probably not. however, I think they wouldn't mind a libgsmd-dbus Jul 21 13:12:14 21st century and posix is still around :) Jul 21 13:12:53 mickeyl: when libgsmd-dbus happens only 2 people will use the socket: harald and werner :) Jul 21 13:13:10 hehe Jul 21 13:13:14 yeah Jul 21 13:13:24 harald is a dependency mania Jul 21 13:13:25 c Jul 21 13:13:40 noone can't change that Jul 21 13:13:47 "I make the software fragile just to avoid dbus" type of maniac? Jul 21 13:14:06 no. more like "dbus introduces too many dependencies" Jul 21 13:14:08 koen: dbus needing a magic printf to not crash is a good argument against it :) Jul 21 13:14:47 you can remove that printf nowadays Jul 21 13:15:14 google teaming with Nokia for Linux phone? Jul 21 13:15:24 Crofton: no, 2 seperate things Jul 21 13:15:26 AFAIK Jul 21 13:15:30 interesting Jul 21 13:15:41 WebKit, WebKit, WebKit :) Jul 21 13:15:49 * zecke should start a WebKit company as well :} Jul 21 13:15:58 zecke: there was a big webkit buz at guadec Jul 21 13:16:05 * Crofton goes digging for cave Jul 21 13:16:32 Crofton: I suspect your ossie recipes need to depend on fftwf now Jul 21 13:17:14 since you were using libfftw3f instead of libfftw3 Jul 21 13:22:00 zecke|shopping: http://www.gnome.org/mobile/, bottom line Jul 21 13:23:51 koen : Am still building task-base! Jul 21 13:25:23 koen : But even if it completes in my lifetime and I do virtual/kernel, it will still produce the same zImage that is in the unstable images area. So I won't have achieved much, because those zImages wont decompress on the NBP. Jul 21 13:26:16 ken : The bootloader complains : "Load error, zlib inflate failed!" Jul 21 13:28:40 koen : It's got to be something about the way the nand filesystem is partitioned. I am using the following commands in the nbkpro.img boot script Jul 21 13:28:46 nand -ignore erase 18 7e8 Jul 21 13:29:06 nand partition e8 Jul 21 13:30:12 ..but since I have no idea what these are actually doing (or the other commands I've seen in some of the Psion images), I don't know how to debug it. Some one, somewhere, must have some documentation on BooSt! Jul 21 13:31:03 yes, florian and philippe Jul 21 13:51:30 mickeyl: do you like avahi? Jul 21 13:51:51 koen: what i've seen, it's amazing. no practical experience yet though Jul 21 13:52:00 koen : how can I get a message to them? Jul 21 13:52:07 mickeyl: btw, liam is going to talk to lennart about replacing alsa-states with pulseaudio Jul 21 13:52:25 jonb: florian.boor@kernelconcepts.de Jul 21 13:52:46 koen: oh that would be handy. so that we could control the alsa scenarios via pulse? Jul 21 13:52:54 yes Jul 21 13:52:55 * mickeyl hates calling alsactl via system() Jul 21 13:52:57 great! Jul 21 13:53:55 XorA explained to the X dudes that ts and screen don't have to match up Jul 21 13:54:04 so X might get saner in the future :) Jul 21 13:54:31 that would be nice Jul 21 13:57:07 mickeyl: the intel dude that offered hardware just sent a python question to oe-devel :) Jul 21 13:58:14 * XorA|gone examines his new mindstorms Jul 21 13:58:22 ah, hat's him. so i might try to answer? Jul 21 13:58:38 XorA|gone: bastard! Jul 21 13:59:13 koen: I'm surprised it isn't enabled anyway. let me rebuild the current kit and check Jul 21 14:00:06 bbiab, need to brew a Roiboos Vanilla Tea Jul 21 14:00:24 * XorA|gone likes roibos and pinapple Jul 21 14:00:59 rooibos presumably? Jul 21 14:01:23 (redwood) Jul 21 14:02:15 koen : eMail sent to florian. Hope it doesn't sound too desparate. Jul 21 14:03:28 koen : need to find out why the images fail on zlib inflate too - any ideas? Jul 21 14:03:42 no, sorry Jul 21 14:04:38 koen : have you got any other boot files for the NBK lying around that I could look at? Jul 21 14:05:08 no, I showed all the secret locations already :) Jul 21 14:05:27 how did you get e17 running then? Jul 21 14:05:45 make a chroot on microdrive and install openzaurus Jul 21 14:06:42 did you use the same kernel image as the (working) gpe install in ~anonymous? Jul 21 14:07:17 iirc, yes Jul 21 14:07:50 have you still got the files? I don't yet know how to do what you did. Jul 21 14:08:34 I put some files on cf and it started flashing by itself Jul 21 14:08:47 I followed the instructions florian gave me Jul 21 14:09:16 can you send me copies Jul 21 14:09:32 I don't have copies :( Jul 21 14:09:36 irc conversation Jul 21 14:09:58 Ah. so the cf disk is lost too? Jul 21 14:10:53 drop me an email at the dr_zee address. Jul 21 14:11:15 pretty please Jul 21 14:11:25 http://www.flickr.com/photos/koenkooi/864716684/in/photostream/ Jul 21 14:15:11 nice picture of some IT guys Jul 21 14:26:04 koen : Hey! The biuild (base-tasks) completed with no errors! Who-hoo! Jul 21 14:26:41 koen : now for virtual/kernel Jul 21 14:27:17 koen : YES! It's completed! Jul 21 14:34:35 koen : I have generated zImage-2.6.17-netbook-pro-20070721142628 Jul 21 14:39:03 ~lart ISP Jul 21 14:39:04 * ibot strangles ISP with a doohicky mouse cord Jul 21 14:40:20 still can't find which packages included gdb in.Does bitbake would rebuild the image, doesn't it? Jul 21 14:42:15 03mickeyl 07org.oe.dev * rbf4d3cbe... 10/ (32 files in 2 dirs): efl: next round of updates / cleanups Jul 21 14:42:23 03mickeyl 07org.oe.dev * r68b677cc... 10/ (1 classes/efl1.bbclass): efl.bbclass: update snapshot date Jul 21 14:42:28 03mickeyl 07org.oe.dev * r486c22df... 10/ (3 files in 3 dirs): python: add subpackage python-debugger including pdb.py* and bdb.py* Jul 21 14:42:53 koen: r486c22df should help Mark Gross. I responded via ml as well. Jul 21 14:43:27 koen : the new image won't inflate on the Pro. Can I make it non compressed? Jul 21 14:44:49 when i rebuild angstrom with DISTRO_TYPE='release' but gdb strace tslib-cal etc, are still included in images, is that a bug of angstrom or bitbake? Jul 21 14:47:26 How do you tell OE to create an image that's not zimage (ie, not compressed)? Jul 21 14:55:57 you mean kernel? Jul 21 14:56:46 Readon : yes. I have the option to not uncompress the image on the target machine - and the zimage OE is creating will not work with the machine Jul 21 14:57:58 Have tried to decompress by renaming zimage to zimage.gz and using gzip, but it doesn't recognise the file type Jul 21 15:00:09 i don't know, i think zImage file need not compressed yourself, it would uncompressed itself. Jul 21 15:03:54 mickeyl: Jul 21 15:03:56 - --enable-software-16-x11 \ Jul 21 15:03:56 + --disable-software-16-x11 \ Jul 21 15:04:03 mickeyl: is that intentional? Jul 21 15:05:16 Readon : the problem is it is failing to uncompress on the target machine, so I thought it might be an idea to try an umcompressed image instead (and tell the target not to decompress it) Jul 21 15:05:41 jonb: 'make Image' in the kernel source tree should do that Jul 21 15:05:59 do you think you have enough space for that? Jul 21 15:06:08 i met that before Jul 21 15:07:05 the code build zImage is very stable. Jul 21 15:12:07 koen : I don't know where the kernel source is in this BitBake environment Jul 21 15:12:48 jonb: the only source you are likely to find is patches against a tarball Jul 21 15:13:24 if it's compressed it must be possible to UNcompress it Jul 21 15:13:58 jonb: I just joined so I don't know the context Jul 21 15:14:15 jonb: but in the OE metadata (not BitBake environment) you have a .bb file that can build a linux kernel Jul 21 15:14:40 jonb: this .bb file says from where to get the source tarball, e.g. ftp.kernel.org and then can add a couple of patches to it Jul 21 15:15:10 zecke : what's the bitbake target name? Jul 21 15:17:18 jonb: virtual/kernel, will select the kernel for your MACHINR Jul 21 15:17:40 koen: yes, software-x16 is broken in the snapshot. the snapshots are not up to par anyways. I'm going to move to cvs builds again and add a respective set of known good SRCDATES. Jul 21 15:17:44 jonb: use -g to get a graph or -v (+ dry run) to see which PROVIER gets selected Jul 21 15:17:45 brb Jul 21 15:17:49 That's the one I just built. It gave me a zimage Jul 21 15:18:14 koen: there's a whole lot of new things coming in EFL. it makes no sense (right now) to use random snapshots until they're really releases. Jul 21 15:18:41 when i rebuild angstrom with DISTRO_TYPE='release' but gdb strace tslib-cal etc, are still included in images, is that a bug of angstrom or bitbake? Jul 21 15:25:05 Philippe : are you here? Jul 21 15:29:29 morning Jul 21 15:32:23 is it normal that log.image_cmd files contains a line like ./usr/bin/[[ Jul 21 15:32:46 i think [[ file is a symbol of bug. Jul 21 15:34:00 mickeyl: are their any usefull targets people can build and test to see if EFL is working? Jul 21 15:34:55 koen: yes. exhibit is a pretty useful picture viewer Jul 21 15:35:01 ok Jul 21 15:35:14 did you enable insane.bbclass on your buildmachine? Jul 21 15:35:22 i'm committing the new cvs stuff later tonite Jul 21 15:35:27 yes :) Jul 21 15:38:42 mickeyl: when is the new o-hand openmoko stuff going to get merged? Jul 21 15:38:48 it's "after guadec" now :) Jul 21 15:40:35 koen: hehe. always pushing :) no idea offhand, I'm going to call them early next week. Jul 21 15:40:51 personally I want it before the end of this month Jul 21 15:41:08 the o-hand gang told me to ask you for details :) Jul 21 15:41:09 so that those who get shipped a neo can already get the new stuff Jul 21 15:41:24 koen: hehe, ok. i'm going to write them tomorrow. Jul 21 15:41:32 I guess openmoko won't switch EABI now? Jul 21 15:41:33 we also need to write an announcement Jul 21 15:41:43 koen: why not? Jul 21 15:42:15 koen: I want to switch a) to the new stuff b) to org.openembedded.dev c) to include angstrom.inc -- all at the same time Jul 21 15:42:31 which is hopefully by the end of this month Jul 21 15:43:23 ah, cool Jul 21 15:43:29 shout when you need a hand Jul 21 15:43:44 only boring books to read the next 3 weeks Jul 21 15:45:37 yeah, definitely need some help. thanks! Jul 21 15:49:28 at last, i think i find a bug of bitbake Jul 21 15:49:44 where can i submit it? Jul 21 15:50:06 bugs.openembedded.org Jul 21 15:53:34 i cann't write one does any one can give me a hand? Jul 21 15:54:48 it seems that task-distro-base task was not recalculated after conf file was modified and the images was rebuilt. Jul 21 15:56:21 i wouldn't count that as a bug. you need to bump the PR of a task if you want it to be recomputed Jul 21 15:56:31 or do bitbake -c rebuild explicitly Jul 21 15:56:40 it's a minor annoyance, but not a bug (imho) Jul 21 15:57:02 it would recalculate itself Jul 21 16:02:47 im trying built openbox natively but i got this error Jul 21 16:02:49 /bin/sed: can't read /data/build/koen/OE/build/tmp/angstrom/work/armv5te-angstrom-linux-gnueabi/glib-2.0-2.12.12-r2/glib-2.12.12/glib/libglib-2.0.la: No such file or directory Jul 21 16:03:14 if my libglib-2.0 is in /usr/lib, why sed tr get it in this directory ? Jul 21 16:03:25 s/tr/try Jul 21 16:07:06 mickeyl, it is a bug, i think. i tried -c rebuild , it does not work. Jul 21 16:08:57 i must use bitbake -c clean task-distro-base;bitbake angstrom_xxx_image. Jul 21 16:09:14 so rebuild does work, but you are using it wrong Jul 21 16:09:51 koen, no,the correct is been tested before test rebuild Jul 21 16:10:04 no Jul 21 16:10:18 you can try by yourself Jul 21 16:10:23 I did Jul 21 16:10:44 and if you want to disagree with 2 core developers go ahead, but don't do it in this channel Jul 21 16:11:11 are you using 1.8.6 release? Jul 21 16:11:31 [17:10] koen: and if you want to disagree with 2 core developers go ahead, but don't do it in this channel Jul 21 16:11:44 yes, 1.8.6 here Jul 21 16:13:48 i just want it to be better. Jul 21 16:15:08 koen: can you give me a little help with my native compilation ? Jul 21 16:27:54 Jul 21 16:37:51 Readon: ast to recalculation. Would you be willing to take a +5 minute speed penalty to get this? Jul 21 16:38:50 but this would make a mistake when conf has been modified. Jul 21 16:39:30 Readon: could you please rephrase the sentence I don't understand what you try to say Jul 21 16:39:31 i meet this on different machine. Jul 21 16:41:26 zecke: i write a text file for everything need to be known. Jul 21 16:41:51 I don't understand what your problem/concern is... anyway have fun Jul 21 16:42:55 i mean i modified conf/distro/angstromxxx.conf file, but images not act as what i specific Jul 21 16:43:52 that might be caused by bitbake who did not recalculate depends of task-distro-base. Jul 21 16:44:16 Readon: BitBake could implement tracking every var. This would require +500MB of data and +5 minutes on each run to check differences. Are you willing to take such a penalty? Jul 21 16:45:17 Readon: it would be easily to implement with O(n^2) (n number of lines of each bb file), you might try to implement it :) Jul 21 16:45:36 zecke: don't wast your time, mickeyl told him it's not a bitbake problem, since bitbake doesn't know about ipkgs in staging, but mr readon knows better than us Jul 21 16:45:53 s/staging/deploy/ Jul 21 16:49:56 what we need to do is write "advanced faq" for bitbake, so people understand that the choices are no arbitrary, but essentially result of many compromises, i.e. the only realistic way to have stuff done Jul 21 16:56:39 here it is. http://bugs.openembedded.org/attachment.cgi?id=1749 Jul 21 16:58:26 I only see someone ignoring advice in that log Jul 21 16:59:35 in any case, lets clarify whether there is a problem or not through this bug entry Jul 21 16:59:45 it might as well be a feature request Jul 21 17:00:35 Noone yet told "that's by design", so let me to it. Jul 21 17:00:49 heh, k Jul 21 17:01:03 Also, "users should not modify anything except local.conf" Jul 21 17:01:31 modifying a distro confif w/o knowing how all that stuff works will lead to unexpected results Jul 21 17:05:16 "if you change the resulting package, bump PR or rebuild" Jul 21 17:05:40 bumping PR and rebuilding have been suggested multiple times Jul 21 17:09:14 the guy doesn't understand why that should be done, so such suggestions are empty words for him. that means that folks who really need to do that, should find explanation of the issues involved, and the rest kept away from doing at all ;-) Jul 21 17:10:30 I would technically be possible to do a SHASUM of env after *.conf is read, and get some .bb files eg tasks to auto rebuild when this SHASUM changes Jul 21 17:10:53 sorry if this is the wrong place: I've built a root filesystem for my asus730 and written it to an SD card, now what's the correct root= paramater to give the kernel? Jul 21 17:10:55 * koen thinks DATE changes Jul 21 17:11:29 lmm: root=mmcblk0p1 rootdelay=10 Jul 21 17:11:36 thanks Jul 21 17:11:51 lmm: not kernel 2.6.22 I hope Jul 21 17:12:40 XorA|gone: no, 2.6.21-hh12 Jul 21 17:13:52 ok i got it, that is my fault, how to remove the entry? Jul 21 17:14:10 bitbake -c rebuild task-base Jul 21 17:14:43 XorA|gone, koen: think distributed development, think package upgrade paths. adding any automation into that risky stuff. unless of course we go same way as distributed SCMs went: hashes as PRs Jul 21 17:15:05 but that doesn't solve upgrade path problem - there's no history, like in SCM Jul 21 17:16:07 psokolovsky_: Feed ideas into bug tracker, familly is round and I got to go socialise, this is an issue proper discussion is worth, as Readon is the 10,000 person to ask the same questions Jul 21 17:16:33 bye xora Jul 21 17:16:36 have fun Jul 21 17:16:57 psokolovsky_: actually the current situation is the wanted behaviour Jul 21 17:17:17 XorA|gone: imho, it will be wprth discussion in next year ;-). so far, we just coming to revamp OE to work properly at all ;-) Jul 21 17:17:23 psokolovsky_: the 'r0' package has the same contents as any other 'r0' package out there Jul 21 17:17:44 how do I stop wince reformatting my SD card when I stick it in, or don't I? Jul 21 17:17:46 people wanting different contents in the _same_ package should be taken out and shot Jul 21 17:17:50 koen: exactly, that's to what I came after long thinking too ;-). now, need to PR (another PR) that to the masses ;-) Jul 21 17:18:24 lmm: create a small fat partition on p1, move linux to p2, change root to mmcblk0p2 Jul 21 17:18:36 koen: thanks Jul 21 17:19:31 small, but fat :) Jul 21 17:37:47 hmm, it's hanging at init Jul 21 17:38:53 is this likely to have been caused by enabling the iwmmxt optimisations? Jul 21 17:39:59 mmm Jul 21 17:40:02 iwmmxt :) Jul 21 17:40:18 wish I could solve the problem with my clock getting canucked Jul 21 17:40:23 when I reset out of linux/oe Jul 21 17:40:32 on my palm t5 but I am sure a fix will come in the near future Jul 21 17:40:56 is there any test apps in opie or a standard way of running a fully charged t5 battery down Jul 21 17:41:03 to get a rough idea what my average battery lifespan is Jul 21 17:41:10 I can't tell if the battery is worn or what Jul 21 17:41:31 I know this, after the nightmare I had last time just resoldering the existing battery leads to the t5 mb Jul 21 17:41:41 when I change the battery I am putting a socket on the mb Jul 21 17:41:51 I just need to find a very low low profile socket Jul 21 18:48:31 NOTE: package gtk+-2.10.14-r0: task do_configure: started Jul 21 18:48:32 * koen taps foot Jul 21 18:49:08 hm how that? Jul 21 18:49:13 args Jul 21 18:49:16 food foot Jul 21 19:32:04 03koen 07org.oe.dev * rd7f98d5d... 10/ (22 files in 3 dirs): gtk+: add 2.10.14, the latest stable release Jul 21 19:32:19 hehe Jul 21 19:32:21 koen Jul 21 20:46:56 hmmm Jul 21 20:47:03 I wanted to ask mickeyl something Jul 21 20:47:09 but I forgot what Jul 21 20:48:29 koen: "Mickey, do you want to marry me?" Jul 21 20:49:16 could be :) Jul 21 20:49:31 sorry, promiscuity is not allowed where i live Jul 21 20:49:45 your wife forbids it? Jul 21 20:49:50 hahaha Jul 21 20:49:51 yeah Jul 21 20:49:51 :D Jul 21 21:05:42 mickeyl: koen: RP: are we going to invite tf again? Jul 21 21:05:51 zecke: absolutely Jul 21 21:05:52 yes Jul 21 21:06:00 (we already did) Jul 21 21:07:20 i think tf brings some fresh blood Jul 21 21:07:21 okay will get a http://www.rikscha-mobil.de/ so tf hasn't to walk :) Jul 21 21:07:25 hahaha Jul 21 21:07:26 yeah Jul 21 21:07:34 he still remembers that walk Jul 21 21:07:38 i asked him @ guadec Jul 21 21:07:41 * zecke is currently cycling like mad Jul 21 21:07:44 and he's looking forward to the next Jul 21 21:09:15 NOTE: the following files were installed but not shipped in any package: Jul 21 21:09:16 NOTE: /home/pkg/oe/fic-gta01/tmp/staging/arm-angstrom-linux-gnueabi/share/ewl/examples/ewl_pdf.c Jul 21 21:09:16 NOTE: /home/pkg/oe/fic-gta01/tmp/staging/arm-angstrom-linux-gnueabi/lib/ewl/tests/libewl_pdf.so.0 Jul 21 21:09:16 NOTE: /home/pkg/oe/fic-gta01/tmp/staging/arm-angstrom-linux-gnueabi/lib/ewl/tests/libewl_pdf.so Jul 21 21:09:20 hmmm Jul 21 21:09:24 installed into staging...? Jul 21 21:09:26 that doesn't look right Jul 21 21:10:28 * mickeyl dives into Jul 21 21:13:14 touchscreen does seem to work on n800 Jul 21 21:14:54 err.. does not seem, that is Jul 21 21:16:22 mickeyl: efl stuff used to give out really weird messages like that Jul 21 21:16:30 mickeyl: and I used to blame binconfig and bad code Jul 21 21:17:16 drat, pavelm just came online Jul 21 21:17:36 * koen prepares for stupid and unnecessary questions Jul 21 21:18:00 *g* Jul 21 21:18:03 ljp: but it boots? Jul 21 21:18:18 * Crofton is glad to be a low profile nerd Jul 21 21:19:01 it's not hard to figure out why suspend on x86 is broken since pavelm wrote it Jul 21 21:19:34 * koen ponders about packaging debootstrap for angstrom Jul 21 21:19:35 koen: found it. this time it's a problem in EPDF installing right into the directories it gets from pkg-config as opposed to only using a suffix Jul 21 21:19:38 or prefix Jul 21 21:20:14 mickeyl: tf promised to look into the poppler stuff this week Jul 21 21:20:21 koen: ok, cool Jul 21 21:29:26 ok, it's like that Jul 21 21:29:26 in configure.in they have Jul 21 21:29:26 EWL_TEST_LIBS_DIR="`$PKG_CONFIG --variable=testsdir ewl`" Jul 21 21:29:27 and then they use this dir to install additional stuff to Jul 21 21:29:39 thanks to our pkg-config sed'ing this dir points to the staging area Jul 21 21:29:43 how to fix this path to point to ${prefix}/foo/bar ? Jul 21 21:29:50 s/path/test/ Jul 21 21:30:37 * mickeyl looks around for autofoo masters Jul 21 21:30:52 * woglinde hides Jul 21 21:30:59 * * OE Bug 2674 has been created by lorn.potter(AT)gmail.com Jul 21 21:31:01 * * touchscreen on nokia 800 not working Jul 21 21:31:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2674 Jul 21 21:33:53 hi all Jul 21 21:34:27 * mickeyl hacks it out with a fixed path (boooooooooooooo) Jul 21 21:41:51 koen: the 20070711 image does boot. Jul 21 21:43:29 starting a build Jul 21 21:44:55 hi florian Jul 21 21:46:22 mickeyl: it is wrong to use pkg-config for getting this dir Jul 21 22:06:08 zecke: i dunno. me no expert in autofoo. i just know that it doesn't work together with the way we "correct" paths in foo.pc to point to staging Jul 21 22:06:24 perhaps we should "correct" not all paths unconditionally Jul 21 22:18:35 zecke, not necessarily. if it's asking pkg-config for "installed" paths, then its entirely valid for it to assume those are indeed correct paths for the target. arguably we'd use "uninstalled" .pc's for staging, but with the .libs/ crap hacked out? i dunno Jul 21 22:18:37 heh Jul 21 22:20:26 hi kergoth Jul 21 22:22:33 kergoth: a) the .pc are 'uninstalled' just like the .la files b) it trades --prefix of autoconf vs. some .pc value (pc does not mean political correct...) Jul 21 22:23:54 good evening evryone here Jul 21 22:24:02 hi Jul 21 22:24:02 hi jan_js Jul 21 22:24:36 you know I allways come with questions, don't you ? Jul 21 22:24:45 hihi Jul 21 22:24:47 ask Jul 21 22:25:20 i dont think you see what i'm saying. if we were to put foo.pc and foo-uninstalled.pc in both in staging, and only mangle uninstalled, and that PKG_CONFIG call were to use PKG_CONFIG_DISABLE_UNINSTALLED, then the paths would be correct.. i'm not saying thats the best solution, just throwing it out as a possibility.. i havent seen oe's pc stuff in a while. Jul 21 22:26:00 I still wander wether it is possible to build images for h3900, because there is no directory in the angstrom download area. Jul 21 22:26:14 ???? Jul 21 22:26:16 I have allready built a minimal image so far. Jul 21 22:26:34 kergoth: you are right on both points. But I believe it should honor --prefix of autoconf more than a random .pc file :) Jul 21 22:27:14 there are images for h3800 and even for h3600, but not for h3900 Jul 21 22:28:14 o.k., maybee not so important - I willl just try to build them. Jul 21 22:29:48 * kergoth nods Jul 21 22:48:11 some one can point a good place to read about native compilation in zaurus/agnstrom ? Jul 21 22:48:46 psycoman: just a wild guess, openzaurus.org? Jul 21 22:48:54 psycoman: the package names of the toolchain should not have cnaged Jul 21 22:48:57 changed even Jul 21 22:49:23 i think my problem is bigger than toolchain names Jul 21 22:49:36 i will past here my problem Jul 21 22:49:41 not here! Jul 21 22:49:43 ~pastebin Jul 21 22:49:46 [pastebin] a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Jul 21 22:49:47 im tring compile openbox natively Jul 21 22:49:54 yeah i now Jul 21 22:49:57 it is one line Jul 21 22:49:57 :) Jul 21 22:50:07 one line can be here rigth ? Jul 21 22:50:18 /bin/sed: can't read /data/build/koen/OE/build/tmp/angstrom/work/armv5te-angstrom-linux-gnueabi/glib-2.0-2.12.12-r2/glib-2.12.12/glib/libglib-2.0.la: No such file or directory Jul 21 22:50:26 my problem Jul 21 22:50:39 psycoman: I can clearly see it you != koen Jul 21 22:50:57 why ? Jul 21 22:51:06 psycoman: sorry, just a bad joke Jul 21 22:51:20 psycoman: what is the file pointing to the .la file? Jul 21 22:51:33 my qustion Jul 21 22:51:48 why the make is tring to get file in this directory Jul 21 22:51:55 if the file is in /usr/lib Jul 21 22:52:21 psycoman: well, I think libtool is doing that but you don#t have enough context in your one line Jul 21 22:52:22 sounds like koen sneaked into .ipk :) Jul 21 22:52:42 psycoman: there is one .la file pointing to /data/build/.... it would be nice to know the name of that file Jul 21 22:52:56 psycoman: and then you can look into packaging of the package that installs this file Jul 21 22:52:59 I'd guess it's glib? Jul 21 22:53:24 polyonymous: well, if glib is present why should it say it is in /data/build? Jul 21 22:53:35 but the glib in /usr/lib is not a symlink Jul 21 22:53:42 zecke, umm.. I think I'm wrong Jul 21 22:53:50 psycoman: stop talking, start greping? Jul 21 22:54:25 oka Jul 21 22:54:26 * polyonymous greps the workdir Jul 21 22:55:01 grep -rl /home/ ~/OE/angstrom/tmp/work/armv5te-angstrom-linux-gnueabi/*/install/*/usr/lib/*.la - gives heck of a lot of stuff :) Jul 21 22:55:27 mumbls something about disabled insanity checker regarding INSTALLED=yes in .la files Jul 21 22:55:41 one point Jul 21 22:55:49 i tring to compile openbox in zaurus Jul 21 22:55:58 not in my desktop with oe builder Jul 21 22:55:59 I was just about to ask about insanity... Jul 21 22:56:01 psycoman: stop taling, start greping on your zaurus :) Jul 21 22:56:07 xD Jul 21 22:56:38 polyonymous: it doesn't know which .la files got installed and greps everything (e.g. .la files installed by libtool in share) Jul 21 22:57:00 polyonymous: some one need to add some dirs to a black list and the reports should be meaningful Jul 21 22:57:19 zecke, what greps everything? Jul 21 22:57:24 hi everybody Jul 21 22:57:49 polyonymous: the insanity check for .la files is searching in the whole STAGING_LIBDIR Jul 21 22:58:03 zecke, well, we're not dealing with staging here... Jul 21 22:58:15 polyonymous: sure we do :) Jul 21 22:58:34 polyonymous: some one is linking against glib, and glib.la says it is INSTALLED Jul 21 22:58:37 no, we're well past staging, we're in ipks now Jul 21 22:59:03 no, the problem is that we have workdir (not even staging, heh) in .la that is packaged. Jul 21 22:59:05 polyonymous: and then the -dev packages points to the installed (in the staging area) glib.la Jul 21 22:59:32 polyonymous: ah right, but I think it is the same issue :) Jul 21 22:59:33 zecke, yup. but we need to QA check it in installed, not in staging. Jul 21 22:59:39 well, sure it's related. Jul 21 22:59:59 psycoman: did it find something yet? Jul 21 23:00:03 anyway, psycoman compile your openbox on desktop :) Jul 21 23:00:07 zecke: do you mean me with tomas in OEDEM mail? Jul 21 23:00:22 thesing: no, the tf_ of o-hand.com Jul 21 23:00:33 thesing: but you are welcome as well, same as ph5 Jul 21 23:00:37 you're not thomas enough! Jul 21 23:00:45 not yeat, i think i not know how use grep very well Jul 21 23:01:08 thesing: I try to be 200% careful to name tf_ tomas (without 'h') Jul 21 23:01:26 oka searching now Jul 21 23:01:46 polyonymous: i have some trobles in my ow builder too Jul 21 23:01:48 polyonymous: well I the real Thomas (with h ) ;) Jul 21 23:02:04 thesing, then you're too Thomas :) Jul 21 23:02:15 im follow the steps in oe.org but bitbake nano dont work, i will see it later Jul 21 23:03:28 zecke: I think I will come to OEDEM. Who knows where it will be the next time ;) Jul 21 23:04:33 thesing hehe Jul 21 23:16:26 zecke: sorry i cannot find any usefful information but can be my search syntax Jul 21 23:16:30 im tring Jul 21 23:17:59 psycoman: you will find it :) Jul 21 23:18:22 yeah.. Jul 21 23:18:56 i have many, many la point to this @@ Jul 21 23:19:11 pastbin ? Jul 21 23:20:15 ~pastebin Jul 21 23:20:17 rumour has it, pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well Jul 21 23:20:17 yes please Jul 21 23:22:31 dpast.com/hold/14977 Jul 21 23:22:58 dpast doesn't work here Jul 21 23:23:42 oka Jul 21 23:24:20 * mickeyl -> bed Jul 21 23:24:21 g'night Jul 21 23:24:25 mickeyl: night Jul 21 23:25:03 pastebin.ca/629082 Jul 21 23:25:49 psycoman: bad :) Jul 21 23:26:04 what =/ Jul 21 23:26:08 psycoman: could you go to bugs.openembedded.org and file a bug report and attach the output? Jul 21 23:26:17 yeap Jul 21 23:26:27 thnkx the support ^^ Jul 21 23:26:37 psycoman: the quick-fix/hack is to sed this /data/build/koen... with /usr/lib Jul 21 23:26:57 i put any special information in topic ? Jul 21 23:27:01 i will do zecke Jul 21 23:27:18 psycoman: "libtool breakage in -dev packages" Jul 21 23:27:24 thnx Jul 21 23:34:19 I'm searching for a *usable* audio player on a Terrier, any suggestion ? bmp won't work and xmms is painful... Jul 21 23:34:51 xmms use 40% of my processor @@ Jul 21 23:35:04 i agree it is painfull Jul 21 23:35:07 oh, it works Jul 21 23:35:19 but the UI is just unusable on a Z Jul 21 23:35:19 im using mplayer at all Jul 21 23:35:31 you can try mp3blaster for console Jul 21 23:36:05 I was thinking about trying eclair, but I'll need to make a recipe first Jul 21 23:36:18 it isn't in OE yet Jul 21 23:37:05 compile it Jul 21 23:37:34 yup. It will be a good way to learn ;) Jul 21 23:37:57 i can try help with you need Jul 21 23:45:46 * thesing -> sleep Jul 21 23:45:52 Night all. Jul 21 23:46:01 nigth Jul 21 23:51:34 03mickeyl 07org.oe.dev * r4f67bea2... 10/ (54 files in 7 dirs): Jul 21 23:51:34 EFL: move from building snapshots to cvs again. Reasons: Jul 21 23:51:34 1.) EFL gained a lot of momentum leading to constant changes Jul 21 23:51:34 2.) Snapshots were just random cvs dates again. Might as well use SRCDATE to pin them down Jul 21 23:51:34 3.) Jul 21 23:51:46 it is possible to inherit a class and then rewrite vars in it, right ? Jul 21 23:52:10 mhmm... I'd better pull again then :) Jul 22 00:08:00 * * OE Bug 2675 has been created by waldecir(AT)gmail.com Jul 22 00:08:02 * * libtool breakage in -dev package Jul 22 00:08:04 * * http://bugs.openembedded.org/show_bug.cgi?id=2675 Jul 22 02:25:34 Hello **** ENDING LOGGING AT Sun Jul 22 02:59:58 2007