**** BEGIN LOGGING AT Fri Nov 07 02:59:57 2008 Nov 07 05:56:05 * * OE Bug 4798 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Nov 07 05:56:07 * * glibc fails with -march=i686 or pentiumpro or pentium3 Nov 07 05:56:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4798 Nov 07 05:56:17 03Denys Dmytriyenko  07org.openembedded.dev * r0f1aa27588 10openembedded.git/packages/glibc/ (files/march-i686.patch glibc_2.6.1.bb): glibc: Add support for compiling glibc with march=i686 Nov 07 07:49:05 morning Nov 07 08:28:52 morning Nov 07 09:13:48 hi in the package glib-2.0, there were a commit that splite the libgobject from libglib. Nov 07 09:14:07 This cause some problem while upgrading Nov 07 09:14:48 because in libgobject it's depended on libglib in some version Nov 07 09:15:01 for example 2.16.1 Nov 07 09:15:18 This change comes from r4 and bump to r5 Nov 07 09:16:13 but in opk file libgobject only know it's depends on 2.16.1 Nov 07 09:16:20 not 2.16.1-r5 Nov 07 09:17:07 therefore if we install libgobject first, it will find something conflict in libglib Nov 07 09:17:39 does anyone have any suggestion about this? Nov 07 09:20:51 pb_, do you have any suggestion? Nov 07 09:50:31 03Michael 'Mickey' Lauer  07org.openembedded.dev * rf9c6ab0d65 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc: bump freesmartphone.org and openmoko recipes to prepare for FSO ms4 Nov 07 10:02:18 ticktock: I guess you need to add an appropriate Replaces: line to libgobject. Nov 07 10:02:24 hi blindman Nov 07 10:05:10 hi pb Nov 07 10:07:27 03Michael 'Mickey' Lauer  07org.openembedded.dev * reca5576b3d 10openembedded.git/ (3 files in 2 dirs): Nov 07 10:07:27 u-boot-openmoko: bump stable version, remove unstable version since u-boot-openmoko has been deprecated. Nov 07 10:07:27 Further development will went into the Qi bootloader. Nov 07 10:07:35 pb_, hmmm but libglib still exists therefore it's seems not good to set RREPLACES to libgobjects Nov 07 10:09:29 pb_, maybe I should upgrade from 2.16.1 to 2.18.1 :P Then I can sure that libglib will be upgraded before libgobject Nov 07 10:13:31 ticktock: right, it shouldn't matter that glib still exists. Nov 07 10:13:50 upgrading from 2.16.4 to 2.18.1 is probably a fine idea anyway but I don't think it will directly solve your problem Nov 07 10:18:08 I see, thanks :) Nov 07 12:19:49 stefan_schmidt: bist du wach? ich haette einen patch fuer dich Nov 07 12:19:57 i have a problem building libtool-native_1.5.10.bb Nov 07 12:20:26 oups, sry, wrong language.... stefan_schmidt i have another patch which may be of interest - do you have time for it? Nov 07 12:20:28 http://rafb.net/p/XIsD3I88.html Nov 07 12:20:51 http://rafb.net/p/IxqKAy40.html and that's the config.log Nov 07 12:23:45 mickeyl|away: ping Nov 07 12:25:12 morning all Nov 07 14:11:46 I don't suppose anyone has built an image for an asus wl-500g router recently? Nov 07 14:12:33 there's a machine config, and an asusoe distro config, but bitbake chokes right away, and all my googling returns stuff from 2005 Nov 07 14:16:12 i'm trying to use bitbake -g to create a dependency graph, but looking at the resulting depends.dot file shows a bunch of nodes that are each only a single character, with a collection of nodes spelling out different package names. Am I making a common mistake somewhere? Nov 07 14:33:24 slothlove: its broke :-( Nov 07 14:44:17 cbrake: the dependency graph generation in general? Nov 07 14:44:29 it looks like it's trying to parse through "IMAGE_INSTALL" and choking Nov 07 14:45:03 slothlove: yes, -g has not been working correctly for a long time Nov 07 14:45:51 cbrake: blast. well, thanks for the info. Nov 07 14:48:51 hi Nov 07 14:58:30 hey Nov 07 14:58:35 hi kergoth Nov 07 15:08:03 "The rule matching engine got converted from a rule list to a token Nov 07 15:08:04 array which reduced the in-memory rules representation of a full Nov 07 15:08:04 featured distro with thousands of udev rules from 1.2MB to 0.12 MB." Nov 07 15:08:12 nice. (udev 131 announcement) Nov 07 15:17:14 kergoth yeah Nov 07 15:17:21 good news Nov 07 15:17:25 for this beast Nov 07 15:21:26 * kergoth chuckles at yesterday's xkcd Nov 07 15:27:44 good morning all! Nov 07 15:32:11 hi hubar Nov 07 15:34:06 hey hubar Nov 07 15:34:08 hi wog! Nov 07 15:34:23 and kergoth! Nov 07 15:36:43 hmm facebook is interesting... Nov 07 15:36:55 my first time using it :) Nov 07 15:42:17 woglinde: any luck on apache2? :) Nov 07 15:53:40 * kergoth wonders if git ever got partial checkout support Nov 07 16:18:32 hubar re Nov 07 16:18:51 hubar the sucking apache developer assuming you cant crosscompile apache2 Nov 07 16:18:56 I will prove them wrong Nov 07 16:19:25 woglinde: :) Nov 07 16:19:26 woglinde, i added an apache2 recipie a while back - it wasn't perfect, but it worked on 2.2. Nov 07 16:19:51 does it need some love? i'd be happy to help out. Nov 07 16:20:12 HopsNBarley: Where do I get that recipe? Nov 07 16:20:41 openembedded/packages/apache2/.... Nov 07 16:20:43 HopsNBarley we need to set one variable Nov 07 16:20:55 for configure Nov 07 16:21:02 otherwise no crosscompiling Nov 07 16:22:17 apache built years and years ago when i was in oe, but it was a pain even then, probably always will be Nov 07 16:22:24 heh Nov 07 16:22:34 g'day kergoth Nov 07 16:22:38 kergoth that was apache 1.3 Nov 07 16:22:39 hi woglinde Nov 07 16:22:42 he pb Nov 07 16:22:49 i spent a couple of days working that recipie. it works on 2.2.3. Nov 07 16:23:04 I will check in 2.2.10 anyway Nov 07 16:23:10 hey pb__ Nov 07 16:23:39 HopsNBarley: Is that the apache2_2.2.3.bb? Nov 07 16:24:11 hubar, yeah. it used to work - i'm running apache2/modphp on slug and ppc based oe. Nov 07 16:24:51 HopsNBarley: That thing breaks on chartables.c now. Nov 07 16:25:10 HopsNBarley how that? set you the void compare to long variable in configure in sites file? Nov 07 16:25:33 okay, that's probably a simple patch update - i'm build now against a reasonably current head. Nov 07 16:26:05 woglinde, the recipie required some site file updates. don't recall exactly what they were. Nov 07 16:26:16 otherwise it breaks at this Nov 07 16:31:40 woglinde, it build for me on arm5 target. Nov 07 16:32:13 what target are you using? Nov 07 16:33:42 I am having difficulty determining how the kernel-module-* packages work. The only hooks I see are in the image/task recipes and the PACKAGE_DYNAMIC tag in kernel.bbclass. My end goal is to build and add custom modules to the rootfs without adding them to the kernel source tree using patches. Any suggestions? Nov 07 16:34:32 pretty sure we have a module.bbclass specifically for that sort of thing and multiple .bb files for external modules in the repository.. at least, we used to Nov 07 16:35:48 I will check that out, thanks! Nov 07 16:35:56 hubar, what target are you building for? Nov 07 16:36:05 HopsNBarley: x86 Nov 07 16:37:41 hubar, never tried x86 OE. Nov 07 16:39:36 HopsNBarley: the build gave me these errors Nov 07 16:39:37 | In file included from pcre.c:470: Nov 07 16:39:39 | chartables.c:21:27: error: pcre_internal.h: No such file or directory Nov 07 16:40:10 chartables.c seems to be from dftables patch. Nov 07 16:40:45 I tried to whack the patch, but then a lot of other consequences flooded in and knocked me out cold. :( Nov 07 16:40:57 this is for 2.2.10? Nov 07 16:42:18 pb__: greetings Nov 07 16:42:36 I see the module.bbclass file and found recipes that use this class to specify external modules. I am still confused, however, on how the kernel-modules-* packages. They do not seem to have a .bb file or be provided anywhere. I'm guessing my confusion stems from my lack of understanding regarding PACKAGES_DYNAMIC. Nov 07 16:42:46 HopsNBarley: no, for 2.2.3. Nov 07 16:42:56 HopsNBarley: the one that is in git right now. Nov 07 16:43:53 ah - get it. lemme try an x86 build - won't be till later or tomorrow though.... Nov 07 16:44:28 Actually this error: http://tinderbox.openembedded.net/public/logs/1884443.txt Nov 07 16:46:17 hubar - thanks. Nov 07 16:48:57 HopsNBarley: sure, I have been banging my head against wall for apache2 build for a couple of days already. :) Nov 07 16:50:43 it took me days to get that recipie in the average state its in. Nov 07 16:53:28 so Nov 07 16:53:33 now time for apache3 Nov 07 16:53:35 ups Nov 07 16:53:35 2 Nov 07 16:54:11 woglinde: How do I check it out? Nov 07 16:54:40 ???? Nov 07 16:54:48 I will test here and commit Nov 07 16:55:38 Do I need to reclone everything (cvs checkout)? Nov 07 16:57:53 cvs? Nov 07 16:57:58 cvs is old and dead Nov 07 16:59:25 hi plague Nov 07 17:00:51 HopsNBarley it isnt smart to patch configure when it will regenerate by autotools Nov 07 17:03:30 hm and do not understand the -native package Nov 07 17:05:08 but thanks for the work sofar Nov 07 17:07:19 woglinde: have you commit the changes yet? :) Nov 07 17:08:33 the -native package is to make the generate tables program, which generates some tables. i verified that these were the same on little/big endian machines, so i populate staging with the programs, and patch the build to call them rather than die on a cross-compiled version. Nov 07 17:12:43 hubar be patiend Nov 07 17:12:57 hops hm I put the pkgconfig back to dev Nov 07 17:13:04 woglinde: :) Sorry. Nov 07 17:13:10 the made no sense in the normal pn Nov 07 17:13:22 so now lets try compile it Nov 07 17:13:59 jo florian Nov 07 17:31:59 hms this suckz Nov 07 17:32:02 checking for /dev/zero... configure: error: cannot check for file existence when cross compiling Nov 07 17:34:36 supply that info via the site info, for a Linux TARGET_OS you know that exists Nov 07 17:34:47 if it isnt pulling from the cache in the configure.ac, then make it do so. see the autoconf manual Nov 07 17:36:50 kergoth I know how to fixit Nov 07 17:36:59 ah, good Nov 07 17:37:01 but the sucking apache devs arent Nov 07 17:37:03 still annoying, yes Nov 07 17:37:25 comes from my alltime favorite Nov 07 17:37:29 apr!!!!!!!!!!!!!!! Nov 07 17:37:32 this is why we need people to focus on certain packages imo. to really try to talk to the upstream devs about fixing the issues there and stuff Nov 07 17:37:37 ~lart apache for inventing apr Nov 07 17:37:37 * ibot beats apache into protomatter with the andromeda galaxy for inventing apr Nov 07 17:40:57 woglinde: I could build apr, but not apr-util. :) Nov 07 17:41:20 well, the apr in packages/ anyway. Nov 07 17:41:22 * mwester hasn't seen that lart before. Nov 07 17:41:36 hubar just wait a couple of minutes Nov 07 17:41:41 then I have fixed it Nov 07 17:41:50 woglinde: :) I am! Nov 07 17:42:18 args!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Nov 07 17:42:21 what the fuck Nov 07 17:42:28 its already in site-conf Nov 07 17:42:40 but the include apr seems not to load it Nov 07 17:44:33 I hate apache more and more Nov 07 17:48:14 i want a patch from here to be applied into my build "/home/a/oe/openembedded/packages/xorg-xserver/xserver-kdrive-xomap" . this is my local conf : BBFILES = "/home/adas/oe/openembedded/packages/*/*.bb" . I do not see the patch reflected in /home/a/oe/tmp/work/beagleboard-angstrom-linux-gnueabi/xserver-xorg-1_1.4.2-r9/xorg-server-1.4.2/hw/kdrive/src . am i relating it correctly? Nov 07 17:52:08 hm why the hack the CONFIG_SITE variable isnt provided to the internal configure run of apr Nov 07 18:00:17 hubar sorry Nov 07 18:00:29 it will take a little bit longer Nov 07 18:00:41 tomorrow i will finish it Nov 07 18:00:55 okay! :) Nov 07 18:01:07 Thanks a lot for the effort. Nov 07 18:01:24 yes Nov 07 18:01:39 apr is the root of all evil Nov 07 18:02:08 hi, I'm trying to build libtool-native, but configure task fails with this error: Nov 07 18:02:08 /usr/bin/autoreconf: unrecognized option `--exclude=libtoolize' Nov 07 18:02:08 someone can help me? Nov 07 18:03:49 shouldn't it be something more like --exclude-libtoolize? Nov 07 18:04:31 I found this in wiki.openembedded.net FAQ: Nov 07 18:04:31 For some reason the base system's autoreconf is being used instead of the OE version (from autoconf-native) Nov 07 18:04:31 but how to tell him to use autoreconf from autoconf-native package? Nov 07 18:04:37 valhalla: no Nov 07 18:04:48 guess85: it shouldnt be using /usr/bin/autoreconf, it should be using -our- autoreconf, as built by autoconf-native Nov 07 18:04:58 err, maybe not Nov 07 18:05:09 ignore me, i forget the interdependencies, probably talking out my ass Nov 07 18:05:13 :P Nov 07 18:05:17 I have this: Nov 07 18:05:17 | NOTE: Executing autoreconf --verbose --install --force --exclude=libtoolize -I /home/claudiu/oe/oetmp/work/i686-linux/libtool-native-1.5.10-r9/libtool-1.5.10/ -I/home/claudiu/oe/oetmp/staging/i686-linux/share/aclocal-1.10 -I /home/claudiu/oe/oetmp/staging/i686-linux/share/aclocal Nov 07 18:05:17 | /usr/bin/autoreconf: unrecognized option `--exclude=libtoolize' Nov 07 18:05:17 | Try `/usr/bin/autoreconf --help' for more information. Nov 07 18:07:25 guess85: that option is supported in newer versions of autoconf. try building autoconf-native manually and then try libtool-native, see if that helps Nov 07 18:09:21 uops, autoreconf Nov 07 18:12:17 bitbake has built autoconf-native-2.61 as a dependency, and in oetmp/work/i686-linux/autoconf-native-2.61-r1/autoconf-2.61/bin/ I have autoreconf Nov 07 18:12:17 but how can I tell OE to use this autoreconf and not /usr/bin/autoreconf Nov 07 18:12:17 ? Nov 07 18:17:46 if do this: Nov 07 18:17:47 - go in libtool-native dir (oetmp/work/i686-linux/libtool-native-1.5.10-r9/libtool-1.5.10/) Nov 07 18:17:47 - and type this: Nov 07 18:17:47 oetmp/work/i686-linux/autoconf-native-2.61-r1/autoconf-2.61/bin/autoreconf --verbose --install --force --exclude=libtoolize -I /home/claudiu/oe/oetmp/work/i686-linux/libtool-native-1.5.10-r9/libtool-1.5.10/ -I/home/claudiu/oe/oetmp/staging/i686-linux/share/aclocal-1.10 -I /home/claudiu/oe/oetmp/staging/i686-linux/share/aclocal Nov 07 18:17:47 it works, so I only have to use autoreconf located in the built autoconf-native dir Nov 07 18:18:33 OpenEmbedded already adds staging stuff to its paths, i have no idea why itd be using that one specifically :\ Nov 07 18:18:40 s/that one/the one in /usr/bin// Nov 07 18:21:15 i found something here: http://wiki.openembedded.net/index.php/Category_talk:FAQ#2._Error:_autoreconf2.50:_unrecognized_option_--exclude.3Dlibtoolize Nov 07 18:21:16 but I didn't understand what I have exactly to do Nov 07 18:21:55 FOSDEM is Feb 7-8 Nov 07 18:32:18 morning Nov 07 18:33:44 gm Nov 07 18:39:11 I am not having any luck determining what the PACKAGES_DYNAMIC field does. No mention in the bitbake manual as far as I can tell. Anyone care to assist? Nov 07 18:39:15 my baby decided to not sleep yet... Nov 07 18:39:32 ah Nov 07 18:39:43 so you must stay up with her Nov 07 18:39:43 Crofton|irssi: they decided finally about fosdem date? Nov 07 18:39:48 yeah Nov 07 18:39:57 it is on the website Nov 07 18:40:05 cool - time to book tickets until they are cheap Nov 07 18:40:20 * Crofton|irssi is hoping the prices drop some ..... Nov 07 19:00:30 re Nov 07 19:25:28 re Nov 07 20:14:04 Crofton|irssi: 53 EUR for me Nov 07 20:14:42 both ways, but I have to add 150 pln (~40 eur) for bus szczecin<>berlin Nov 07 20:24:16 anyone got a test rootfs for an eten g500? Nov 07 20:24:29 * hrw -> sleep Nov 07 20:24:39 have a nice weekend Nov 07 20:31:57 hrw|gone: urg, more like 400-500 euro's for me Nov 07 20:36:03 Crofton|irssi: i would think atleast 600-700 euros Nov 07 20:39:04 not quite that bad .... Nov 07 21:54:14 if anyone else is interest in working/testing with the M8050, give me a shout Nov 07 21:54:36 free hardware in exchange for code Nov 07 21:55:52 prpplague: M8050? Nov 07 21:56:15 http://www.elinux.org/M8050 , a S3C2440 based industrial PDA Nov 07 21:57:10 prpplague: IC... Nov 07 21:57:39 prpplague: too many handhelds around here already... Nov 07 21:59:56 FOM: yea, i'm sure there are a number of handhelds, but not too many have a 5800mAh battery, with barcode scanner, and magnetic card reader Nov 07 22:00:47 prpplague: watch out for Symbol Technologies, or has Moto finished running them (hopefully) into the ground? Nov 07 22:01:29 FOM: we have license agreements with moto/symbol Nov 07 22:01:45 prpplague: oic, well then have at it ;-) Nov 07 22:01:46 Big Nov 07 22:02:02 prpplague seem good for warehouse ... Nov 07 22:02:12 no good for medical device ... :S Nov 07 22:02:29 gremlin[it]: yea, its more intended for retail areas, like for grocery stores and restaurants Nov 07 22:02:38 handheld credit card processing, for special events, perhaps? Nov 07 22:02:47 gremlin[it]: yea not intended for medical Nov 07 22:02:54 mwester: exactly Nov 07 22:03:29 we have a second unit pre-production for more warehouse stuff, i.e. has a pistol grip Nov 07 22:03:42 mwester: handheld credit card processing for criminals :) Nov 07 22:04:04 Crofton: shhhh Nov 07 22:04:42 Crofton|irssi: ;) Criminals and vendors at sporting events are indistinguishable, in my experience. Just consider the price of a hot dog and a beer at a baseball game. Nov 07 22:04:50 hehe Nov 07 22:13:30 mwester: yea the M8050 is fairly big, for two reasons, 1) customers don't like handhelds that their employees might steal 2) needs big enough battery to last 2 shifts of 8 hours without recharging Nov 07 22:51:07 * * OE Bug 4814 has been created by cliff.brake(AT)gmail.com Nov 07 22:51:09 * * qemu fails to build on ubuntu 8.10 Nov 07 22:51:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4814 Nov 07 23:01:43 re Nov 07 23:02:03 hallo Nov 07 23:02:15 jo ant whats up? Nov 07 23:02:51 too many miles in two days..ich spinne... Nov 07 23:09:53 woglinde: are you using >glib-2.18.1 ? Nov 07 23:09:59 hm Nov 07 23:10:07 I only compiled for some weeks Nov 07 23:10:22 it's about atomic-thumb... Nov 07 23:10:23 http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2008-November/002717.html Nov 07 23:10:58 whats the problem? Nov 07 23:11:10 just take the patch from 2.6.14.1 Nov 07 23:11:27 why is glib-2.18.1 not yet updated? Nov 07 23:11:45 aks pb Nov 07 23:11:55 ok Nov 07 23:12:02 are yo building with thumb code? Nov 07 23:12:06 I'll retry Nov 07 23:12:22 it was failing on glib ^^ Nov 07 23:12:34 yeah Nov 07 23:12:41 then just put the patch theier Nov 07 23:12:49 yea Nov 07 23:28:42 h bluelightning Nov 07 23:29:39 hi woglinde Nov 07 23:34:31 mikkov_: http://pastebin.ca/1248268 sorry its in german but its basically due to a missing dependency. if I do apt-get install -f Nov 08 00:34:23 03Stefan Schmidt  07org.openembedded.dev * r88e8d54252 10openembedded.git/packages/images/fso-image.bb: fso-image: Use the correct name for the fso illume theme Nov 08 01:46:04 well, monotone to git trigger me to learn OE today Nov 08 01:46:40 after installing all packages, git clone the whole OE, i don't know how to try the 'bitbake world' Nov 08 01:47:06 noticed there are a build/ and conf/*.conf, how can I use them to try one of it? Nov 08 01:47:30 the wiki site etc is not updated for git fully **** ENDING LOGGING AT Sat Nov 08 03:00:02 2008