**** BEGIN LOGGING AT Mon Aug 06 02:59:57 2007 Aug 06 06:08:15 good morning all Aug 06 06:23:42 morning Aug 06 07:02:19 morning all Aug 06 07:04:07 hi Stelios Aug 06 07:05:13 hrw : hi ! Has the board arrived ? Aug 06 07:05:40 not yet Aug 06 07:06:53 hrw : hmmm let me know if does not arrive in a few days. Its send with register mail, so probably will be able to track it Aug 06 07:07:19 hrw : btw any news with the magicbox people ? Aug 06 07:08:44 morning all Aug 06 07:08:54 steliosk: will give you info when it arrive or at thursday if it will not. Aug 06 07:09:05 steliosk: end no news from magicbox ;( Aug 06 07:14:20 RP: hi Aug 06 07:14:28 RP: officially Bromley is part of London? Aug 06 07:17:56 hrw: Its part of something like greater London. I'd ask someone who works there for the exact technical meaning ;-) Aug 06 07:19:05 RP: http://blog.haerwu.biz/2007/08/06/my-openembedded-related-trips/ Aug 06 07:19:58 hrw: :) Aug 06 07:19:59 reload Aug 06 07:22:07 hrw: ated to OZ to try to get my Zau SL-C3000 actually useful. Never happened. Nor with the stock Sharp ROM. Aug 06 07:22:07 I have some needs from the kit, I wonder if ther Aug 06 07:22:12 gah, sorry Aug 06 07:23:05 ;D Aug 06 07:23:15 hrw: http://en.wikipedia.org/wiki/London_Borough_of_Bromley Aug 06 07:23:37 tx Aug 06 07:25:38 RP: where there any objections against your install kit patch? Aug 06 07:28:34 s/where/were/ Aug 06 07:29:27 RP: I will be back in ~1-1.5 hour Aug 06 07:47:09 koen: I wasn't aware of any Aug 06 07:51:33 koen: Do you want me to commit it? Aug 06 07:56:21 koen : http://rafb.net/p/7FNUqr82.html Aug 06 08:11:37 koen : its libx11 Aug 06 08:17:24 RP: yes please Aug 06 08:18:22 steliosk: the error is from a native build (makekeys needs to get run) Aug 06 08:19:17 koen : hmmm. i a building on a i386 machine using icecc Aug 06 08:30:17 i want to distribute some data files and copy them to target machine. when i use pkgdata_DATA = xxx pkgdatadir = /usr/local/xxx, but make told me that "*** No rule to make target `xxx.bmp', needed by `all-am'. Stop." then oe stopped, how to work around that? Aug 06 08:44:44 03koen 07org.oe.dev * r6e396056... 10/ (1 packages/openmoko-apps/openmoko-terminal_svn.bb): openmoko-terminal: fix PV Aug 06 08:58:48 koen : did you change libx11 during your recent commits ? can't find anything Aug 06 09:00:18 re Aug 06 09:06:43 koen : the prev version of libx11 had a do_compile() to handle the makekeys. In the new libx11 it has changed Aug 06 09:07:04 koen : and probably that's what causing the problem Aug 06 09:07:27 koen : i reverted to the old do_compile code and trying to build it now Aug 06 09:07:41 could you commit that? Aug 06 09:09:45 koen : did not work. same problem -> http://rafb.net/p/hBGbFq95.html Aug 06 09:11:04 that's not the same problem Aug 06 09:11:19 the first problem was a failed compile of a host binary Aug 06 09:11:31 the second problem is that it is trying to run a target binary Aug 06 09:13:27 koen : it still tries and builds makekeys using the target toolchain Aug 06 09:14:09 koen : in the first case it failed to link it. now it links ok but as expected can't run on the hosts Aug 06 09:36:31 hmmm Aug 06 09:36:37 I can't reproduce the error Aug 06 09:36:49 (for armv4t) Aug 06 09:41:11 i think i fixed it... Aug 06 09:42:08 NOTE: package libx11-1.1.2: completed Aug 06 09:43:19 good morning Aug 06 09:43:53 hi florian Aug 06 09:45:01 koen : http://rafb.net/p/DTFzdC87.html Aug 06 09:45:11 koen : try this Aug 06 09:51:09 koen: SINCE when OE use pkg-config 0.22?? Aug 06 10:17:27 hrw: since I accidentally removed the DEFAULT_PREF Aug 06 10:22:04 koen : did you try the patch ? Aug 06 10:22:40 no, but it looks ok Aug 06 10:23:57 ok. will let the build finish and commit after Aug 06 10:49:46 koen: ping? Aug 06 10:54:23 03koen 07org.oe.dev * r0f0d473a... 10/ (1 packages/pkgconfig/pkgconfig-native_0.22.bb): pkgconfig-native: reinstate default pref Aug 06 10:55:12 zecke: pong Aug 06 10:55:29 koen: could you try this blind fix? Aug 06 10:55:49 yes Aug 06 10:55:55 if you show me the fix :) Aug 06 10:55:57 comitted Aug 06 10:56:26 I hope you didn't mind me teasing you Aug 06 10:56:55 koen: it should compile :) Aug 06 10:57:15 removing -Werror would probably make it compile as well :) Aug 06 10:57:51 -Werror should get default in gcc Aug 06 10:58:37 zecke_ : hi ! Aug 06 10:58:41 ho Aug 06 10:58:58 zecke: fix works, thanks! Aug 06 11:00:12 koen: you still can only configure by changing the source code, and some how moko-finger-scroll can not draw the scroll indicator on the WebKitGtkPage (but you can scroll) Aug 06 11:01:12 koen: make it crash and send feedback. I'm currently on GRPS so it will get some more love to work on GPRS phones Aug 06 11:01:22 :) Aug 06 11:03:14 I think I made it crash :} Aug 06 11:09:45 bye Aug 06 11:18:42 koen : weird -> NOTE: package libx11-1.1.1: started Aug 06 11:19:09 libx11 was build allready 1.1.2 Aug 06 11:19:45 steliosk: It could be the RPROVIDES - koen, was the removed from the 1.1.1 verison? Aug 06 11:27:34 RP: It does not seem to effect the build though. building fs at the moment Aug 06 11:28:33 steliosk: Its a sublte effect which will end up causing pain. The 1.1.1 version does still have the RPROVIDES Aug 06 11:28:52 03rwhitby 07org.oe.dev * r51bbb663... 10/ (3 files in 3 dirs): ixp4xx-kernel: built the ixp4xx-npe module into the kernel, so it can accept a microcode image found in an MTD partition Aug 06 11:28:59 03rwhitby 07org.oe.dev * r8416ad85... 10/ (3 files in 3 dirs): slugos: Made sure the IXP modules are in the rootfs, and bumped the slugos revision in preparation for testing before a binary release Aug 06 11:30:22 stelios@farm-master:/storage/angstrom/org.openembedded.dev$ egrep -inr rprovides ./* | grep libx11 Aug 06 11:30:22 ./packages/obsolete/freedesktop-premodular-xlibs/libx11_6.2.1.bb:6:RPROVIDES = "virtual/libx11" Aug 06 11:30:22 ./packages/xorg-lib/libx11_git.bb:8:RPROVIDES = "virtual/libx11" Aug 06 11:30:22 ./packages/xorg-lib/libx11_1.1.1.bb:10:RPROVIDES = "virtual/libx11" Aug 06 11:30:22 stelios@farm-master:/storage/angstrom/org.openembedded.dev$ Aug 06 11:31:12 RP: It looks like you are correct about the RPROVIDES in 1.1.1 Aug 06 11:32:38 RP: should that be removed ? Aug 06 11:32:54 steliosk: Yes, I think so. That will then show up the real problem Aug 06 11:35:13 steliosk, tried any Efika builds? Aug 06 11:36:32 ~change 4700 kcs to pln Aug 06 11:36:38 i got some error trying that Aug 06 11:48:31 Crofton|home : yes. a minor glitch with libx11 Aug 06 11:49:00 hmm, no issue building libusb? Aug 06 11:49:47 Crofton|home : but both bootstrap-image and angstrom-x11-image build ok . not sure if they run thought Aug 06 11:49:56 hmm Aug 06 11:49:59 * * OE Bug 2690 has been RESOLVED (FIXED) by rpurdie(AT)rpsys.net Aug 06 11:49:59 time to try again Aug 06 11:50:01 * *  Script to make zaurus images usable Aug 06 11:50:03 * * http://bugs.openembedded.org/show_bug.cgi?id=2690 Aug 06 11:50:11 I wonder if I am having a host system interaction issue? Aug 06 11:50:33 Crofton|home : libusb builds fine Aug 06 11:51:12 RP: I only killed the virtual/xserver RPROVIDES Aug 06 11:51:13 Crofton|home : i did the build on i386 and used icecc, did not test amd64 Aug 06 11:51:29 the virtual/libx11 ones should go as well, though Aug 06 11:51:38 pulling and rebuilding Aug 06 11:51:51 it is good to have data points Aug 06 11:52:06 Crofton|home : hold on a sec to send the libx11 stuff Aug 06 11:52:24 I'm only working on bootstrap for now Aug 06 11:52:32 Crofton|home : btw managed to get Xilinx Platform cable usb jtag to work on ubuntu amd64 Aug 06 11:52:37 real men do not have displays on their embedded system s:) Aug 06 11:52:38 Crofton|home: libusb and libtool 1.5.24? Aug 06 11:52:48 hrw, most likely Aug 06 11:52:53 Crofton|home : hehe Aug 06 11:52:54 it worked on OSK ... Aug 06 11:54:14 can I do ANGSTROM_EXTRA_INSTALL_omap5912osk Aug 06 11:54:27 and ANGSTROM_EXTRA_INSTALL_efika ? Aug 06 11:55:24 ok building, and pulled all version overrides Aug 06 12:00:58 Crofton|home: you can, but it's ugly Aug 06 12:01:20 I know :) Aug 06 12:01:41 I am trying to minimize the number of changes to go from building Efika to OSK :) Aug 06 12:03:48 03ifaistos 07org.oe.dev * r112f00d0... 10/ (1 packages/xorg-lib/libx11_1.1.2.bb): Aug 06 12:03:48 packages/xorg-lib/libx11_1.1.2.bb : Fix an issue when building for Efika, where it used Aug 06 12:03:48 target toolchain to build makekeys which is need to run on the host Aug 06 12:03:54 03rpurdie 07org.oe.dev * r5a83d568... 10/ (11 files in 3 dirs): zaurus: Rename included conf files to .inc Aug 06 12:04:03 03rpurdie 07org.oe.dev * rbd74f0d9... 10/ (1 classes/image.bbclass conf/machine/include/zaurus-2.6.inc): zaurus-2.6: Create an install kit when generating images (from bug #2690 with modifications) Aug 06 12:04:44 koen: I removed the libx11 RPROVIDES in poky which is probably why its missing in OE Aug 06 12:05:12 hey mickeyl Aug 06 12:05:18 hi mickeyl Aug 06 12:05:34 hellow Aug 06 12:06:16 steliosk: could you put that fix in all 3 libx11 packages? Aug 06 12:07:11 mickeyl: can you check does SRCDATE="now" really works? Aug 06 12:07:41 koen : ok Aug 06 12:07:58 mickeyl: NOTE: fetch http://downloads.openmoko.org/sources/gsm_svn.openmoko.org_.trunk.src.target__now.tar.gz Aug 06 12:08:13 mickeyl: no information how old that 'now' is.. Aug 06 12:08:39 hrw: I would guess this is a stopgap until we get SRCREV support Aug 06 12:09:12 hrw: that's why SRCDATE=now should only be used by developers in private who know what consequences it has Aug 06 12:09:50 I complained about 'now' ending up in SRCDATE almost two years ago :) Aug 06 12:09:58 koen: hard to not agree Aug 06 12:10:21 Does cvs "now" support actually work? Aug 06 12:10:37 If not, I'd be tempted to rip that out of bitbake... Aug 06 12:11:01 cvs and svk are the only fetchers that use "now" after the recent changes Aug 06 12:11:10 and svk is probably just broken Aug 06 12:11:45 * koen thinks people can just do 'cvs up -dP' after do_unpack Aug 06 12:13:53 that's actually a neat idea Aug 06 12:14:40 mickeyl: openmoko.bbclass: do_unpack2() { svn up ${S} } addtask unpack2 after unpack Aug 06 12:15:07 that pollutes package space Aug 06 12:15:15 ya, something like that. Aug 06 12:16:56 mickeyl: only gtkterm2 doesn't build (fails on do_install) with upcoming autotools updates Aug 06 12:17:23 hmm, might scratch that and go for mrxvt anyway Aug 06 12:17:45 bitbake openmoko-devel-image with DISTRO=angstrom-2007.1 completes now (without gtkterm2) Aug 06 12:17:54 ah.. that reminds me that I need to boot 2.6.21 on c7x0 Aug 06 12:18:07 koen: ok, cool. Aug 06 12:19:17 mickeyl: and parsing got 10% faster for everyone :) Aug 06 12:19:19 What is the current thinking on a "stable branch"? Aug 06 12:19:36 Crofton|home: find someone to pay 2 fulltime OE devs :) Aug 06 12:19:45 heh Aug 06 12:20:07 and a few dozen autobuilders Aug 06 12:20:45 Crofton: it will come. Aug 06 12:20:56 first get atleast one autobuilder Aug 06 12:20:57 Crofton: roughly at the end of this year Aug 06 12:21:09 then add few more so amount of configs can be extended Aug 06 12:21:25 then people will more often commit fixes for just broken Aug 06 12:21:27 stuff Aug 06 12:22:33 as opposed to me chaning from hand coded kernel construction to make uImage :) Aug 06 12:37:19 03koen 07org.oe.dev * r75c7809d... 10/ (3 files in 2 dirs): libx11: remove RPROVIDES Aug 06 13:12:27 03ifaistos 07org.oe.dev * r20003f50... 10/ (3 files in 2 dirs): packages/xorg-lib/libx11 : Fix makekeys generation on all libx11 bb packages Aug 06 13:26:56 ~convert 6 lbs to kg Aug 06 13:26:58 6 lbs is approximately 2.72155 kg Aug 06 13:42:05 koen RP : This is from a clean build (tmp removed) -> http://rafb.net/p/qLmHTJ32.html Aug 06 13:44:34 steliosk: That exposes the problem properly. I'm not sure packages should be RDEPENDing on a virtual like that... Aug 06 13:48:10 RP : i am kind of lost with the recent changes to Angstrom. Haven't followed closely the updates/commits to see what changed Aug 06 13:50:06 steliosk: Poky X11 cleanup of recipes was included Aug 06 13:50:36 do we have a list of features / tested packages we're shooting for to make a first stable release? Aug 06 13:50:47 steliosk: We need to talk to koen about that issue. for now you can workaround by s#virtual/libx11#libx11# or diet-x11 Aug 06 13:52:07 dreamer : Not sure if others have discussed about it, but i think its one of the things to discuss at the OEDEM Aug 06 13:52:45 steliosk: thanks Aug 06 13:53:16 RP: well i was "lucky" enough to keep a copy of a working x11-image ;) Aug 06 13:53:38 is that going to be on skype/vent also, or just all the semi-locals Aug 06 13:53:44 we need a way to tag distros, machines, and recipes as "stable" Aug 06 13:55:16 Crofton|home : true. But we need to come up with a plan what a stable release is Aug 06 13:55:23 yeah Aug 06 13:56:08 good topics for OEDEM Aug 06 13:56:13 Following something the odd/even version scheme of the kernel probably Aug 06 13:57:27 later all Aug 06 13:57:32 bye Aug 06 14:23:57 hi all Aug 06 14:24:29 I need a suggestion. I'm trying to get OE to install the libc header files under /usr/include. When glibc builds, it creates the file: libc6-dev_2.3.5+cvs20050627-r13_arm.ipk, but I cannot seem to describe to OE that this is to be part of the rootfs image. Aug 06 14:25:07 I've said this in the distro.conf: DISTRO_EXTRA_RRECOMMENDS += "libc6-dev" Aug 06 14:25:15 T0mW: you need own image to create Aug 06 14:25:27 and : PREFERRED_VERSION_libc6-dev = "2.3.5+cvs20050627" Aug 06 14:25:34 we should BAN people which use DISTRO_EXTRA_* vars Aug 06 14:25:45 hrw, all the stuf is there in the ipk Aug 06 14:26:15 T0mW: then create own-image.bb recipe which will have all packages which you want to install listed Aug 06 14:26:39 T0mW: and because of debian renaming which you use you need to use 'glibc-dev' in recipe Aug 06 14:26:45 um, what about the task? Aug 06 14:26:52 oh Aug 06 14:27:25 why would I know that the renaming would occur? Is that in the packages.filelist or something? Aug 06 14:27:36 you can go any way - but end with OWN image NOT distro_extra_* var use Aug 06 14:27:53 T0mW: in recipes use 'other_recipe-dev' names Aug 06 14:29:13 What is wrong with DISTRO_EXTRA_RRECOMMENDS? That is in the task-base. Aug 06 14:29:27 crap, libusb builds with stock version of libtool/pkconfig etc Aug 06 14:29:46 T0mW: they ARE for DISTROS Aug 06 14:29:54 not for adding stuff by anyone to image Aug 06 14:30:12 T0mW: if you want image with extra content then proper way is creating own image recipe Aug 06 14:30:18 Well, I thought that I was making a Distro, a distro of linux for my target Aug 06 14:30:52 but stuff like libc6-dev should be done by kiind of tomw-devel-image.bb rather Aug 06 14:30:54 DISTRO is only relevant to the build host? Aug 06 14:30:57 no Aug 06 14:31:25 DISTRO_* is what distro maintainer(s) thinks that should be shipped in any image (except minimal one) Aug 06 14:31:45 T0mW: you really plan to ship libc6-dev with x11/console/xfce/e17/gpe/openmoko images? Aug 06 14:32:01 I don't use any GUI Aug 06 14:32:20 320x240 LCD and no pointing devices on the unit Aug 06 14:32:29 monochrome Aug 06 14:32:45 looks like it is time to add something like sdk-image into OE (if it is not in yet) Aug 06 14:32:48 so all that opie, poky, qt stuff is useless Aug 06 14:33:33 T0mW: look at openmoko-sdk-image.bb Aug 06 14:34:02 T0mW: you probably want kind of 'console-image.bb with task-native-sdk installed in it' Aug 06 14:34:11 In any event, during the build of glibc*.bb, it creates the libc6-dev*ipk, but doesn't explode the contents of it into the final rootfs Aug 06 14:34:41 did you rebuild task-base after playing with local.conf? Aug 06 14:35:05 hrw: problem is that this is a 2.4.21 kernel, apparently OE is broken with regards to building something so 'ancient'. Aug 06 14:35:06 yes Aug 06 14:35:15 bitbake -c rebuild task-base Aug 06 14:35:37 T0mW: in which way it is broken? I can help with it Aug 06 14:36:48 this out of order executioin of pieces of bb files is driving me insane Aug 06 14:36:57 headers for 2.4.21, I cannot seem to get headers setup for that kernel, so I have to use 2.6.11.1 headers to build with Aug 06 14:37:18 Crofton: heh, yeah libiconv builds AFTER glibc Aug 06 14:37:42 in my case, I want to see libusb compile, but it keeps teasing me Aug 06 14:37:47 configure is done now Aug 06 14:37:55 T0mW: basically kernel headers does not have to match kernel version Aug 06 14:38:07 Crofton|home: tried to remvoe aclocal.m4 before do_configure? Aug 06 14:38:41 hrw: correct. However, it seems to enforce pthreads if I use the 2.6 header (modified for correct 2.4 syscalls) Aug 06 14:38:44 hrw, lets wait until I have failure again :) Aug 06 14:38:59 T0mW: and 2.6.11.1 were ok for 2.4.18-crapix used by openzaurus Aug 06 14:40:22 T0mW: check how openzaurus was done in.oz354x branch Aug 06 14:40:49 hrw, well, maybe someone else can build openzaurus but I cannot. IMHO, one huge issue with OE that is a point of confusion is this: Yes, there are a host of conf/distro + conf/machine files, but there are no local.conf samples that tell you what to use with what. Aug 06 14:41:23 so the openzaurus machine may build, but with what distro. I have yet to build anything other than my own project Aug 06 14:41:46 not angry with OE, just trying to point out what I consider a weakness Aug 06 14:42:22 tasks are just as bad, but, I've been able to muddle through those Aug 06 14:43:00 T0mW: my standard local.conf is just DISTRO + MACHINE Aug 06 14:43:17 rest of vars are BBINCLUDELOGS or DL_DIR or mirrors Aug 06 14:44:46 03koen 07org.oe.dev * r0b702cfd... 10/ (1 packages/angstrom/angstrom-x11-base-depends.bb): angstrom-x11-base-depends: remove RDEPENDS on virtual/x11, apps link to it anyway Aug 06 14:45:50 steliosk: there you go Aug 06 14:46:11 wait for it Aug 06 14:46:16 crap Aug 06 14:46:31 I can compile libisb with standard pkconfig/libtool Aug 06 14:47:31 * T0mW renames 'libc-dev' reference to 'glibc-dev' in distro conf Aug 06 14:52:54 * koen still doesn't get why people are so intent on creating their own distro instead creating a customized image Aug 06 14:53:11 koen: angstrom does not support 2.4 kernels Aug 06 14:53:23 angstrom-legacy does Aug 06 14:53:23 and T0mW device lack 2.6 kernel Aug 06 14:53:32 or rather, can do Aug 06 14:53:49 require conf/distro/angstrom-2007.1.conf Aug 06 14:53:50 # We'll have to fill this in as we go. Aug 06 14:53:58 it does not support 2.4 at all Aug 06 14:54:08 * T0mW is creating zipit-install-libc6-dev_1.0.bb Aug 06 15:06:43 koen: have you heard anything from zottd about the src for his psx emu yet? Aug 06 15:20:36 mtn: 3 heads on branch 'org.openembedded.dev' Aug 06 15:20:41 not often situation ;) Aug 06 15:20:44 * hrw merged Aug 06 15:20:50 scary\ Aug 06 15:25:09 i can't find a bug tied into libts-1.0 on the tracker, has anyone here had problems installing the new version? Aug 06 15:25:46 03hrw 07org.oe.dev * r55f602d6... 10/ (1 packages/gstreamer/gst-meta-base_0.10.bb): gstreamer: added meta package from Poky to get plugins grouped Aug 06 15:25:53 03hrw 07org.oe.dev * r96bb64a3... 10/ (1 packages/openmoko2/openmoko-dialer2_svn.bb): openmoko-dialer2: depend on gstreamer for ringtones Aug 06 15:25:57 03florian 07org.oe.dev * r7594ca77... 10/ (1 packages/libmimedir/libmimedir_0.4.2.bb): libmimedir: Add 0.4 2 Aug 06 15:29:10 wirelessdreamer: not yet Aug 06 15:29:46 koen: he just released a new version 3 days ago, its supposed to do much better then the last one, once again :) Aug 06 15:30:52 wirelessdreamer: maybe you can bother him for handing out the source to selected people, so the can build it? Aug 06 15:30:58 I don't have an interest in gaming Aug 06 15:32:19 hrw: don't forget to set COMPATIBLE_MACHINE = "fic-gtao1" on openmoko-dialer Aug 06 15:32:30 i asked in the thread on the fourms how long it'll be, from what i understand it needs sdl running on the console, i've been working on getting it running, built 1.2.9 but it needed mouse support, just found a usbmouse and touchscreen patch i'm going to try, i'd like to up it to 1.2.12 also, but i'm having build problems with that version i didn't figure out Aug 06 15:37:32 koen: why? Aug 06 15:37:53 hrw: sorry, I read the code wrong Aug 06 15:38:09 hrw: http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=687 is non-fatal Aug 06 15:38:43 hi Aug 06 15:38:51 hey chouimat Aug 06 15:39:01 koen: when openmoko will get internal software team... Aug 06 15:39:17 hrw: when hell freezes over? Aug 06 15:39:56 ;D Aug 06 15:40:04 when koen wins the lottery Aug 06 15:40:38 I'd use my money for other goals when I win the lottery :) Aug 06 15:40:52 selfish Aug 06 15:40:56 I'd trademark everything! Aug 06 15:40:59 * koen runs Aug 06 15:41:56 I need STAGING_LIBDIR_NATIVE --- so I don't have to put export LD_LIBRARY_PATH = "${STAGING_BINDIR_NATIVE}/../lib" in ncurses bitbake recipe... :( Aug 06 15:42:09 !oebug 666 Aug 06 15:42:10 * * Bug 666, Status: RESOLVED (WONTFIX), Created: 2006-02-08 02:16 Aug 06 15:42:11 * * pachkov(AT)gmail.com: OZ 3.5.4test2: SD card does not appear in cardmanager. Aug 06 15:42:12 * * http://bugs.openembedded.org/show_bug.cgi?id=666 Aug 06 15:42:23 Nope, NOT that one! Aug 06 15:43:00 * mwester ncurses loudly, and goes to find the ncurses bug Aug 06 15:43:44 !oebug 2554 Aug 06 15:43:45 * * Bug 2554, Status: NEW, Created: 2007-06-23 11:51 Aug 06 15:43:46 * * utx(AT)penguin.cz: broken cross build environment / fix ncurses build Aug 06 15:43:47 * * http://bugs.openembedded.org/show_bug.cgi?id=2554 Aug 06 15:44:51 That's the one. I have a working fix, but as above, it's pretty ugly. Commit it as is, or does somebody have any objections? (other than general asthetic objections) Aug 06 15:48:43 mwester: STAGING_LIBDIR_NATIVE sounds good for me - we have some bugreports from fedora and opensuse users Aug 06 16:09:14 bye Aug 06 16:11:24 anyone know what mtn: error: you must clean up and remove the _MTN/detached directory means? Aug 06 16:11:44 nm found it Aug 06 16:32:33 ~seen lardman Aug 06 16:32:40 lardman was last seen on IRC in channel #maemo, 2h 2m 43s ago, saying: 'http://fanoush.wz.cz/maemo/'. Aug 06 16:50:03 gm khem Aug 06 16:50:23 Hey howdy Crofton Aug 06 16:50:47 thanks for taking care of the patch Aug 06 16:51:07 no problem Aug 06 16:51:25 at the time I did that, I was still figuring out alot of embedded stuff Aug 06 16:51:46 now I have OE compiling and booting on OSK5912 Aug 06 16:51:56 next I am gonna try OMAP2430 Aug 06 16:51:58 excellent Aug 06 16:52:05 awesome Aug 06 16:52:17 is there support for omap2 already in Aug 06 16:52:23 nope Aug 06 16:52:33 I think omap1 should be able to handle 95% Aug 06 16:52:37 we have no boards to test with Aug 06 16:52:41 yeah Aug 06 16:52:47 cool I am happy I can help Aug 06 16:52:54 there is a bb fiel to pul from omap git Aug 06 16:53:03 which one Aug 06 16:53:28 it has omap1 and git in the name Aug 06 16:53:30 :) Aug 06 16:55:32 hehe ok Aug 06 16:55:44 I would also want to boot with eglibc Aug 06 16:55:47 sorry looking up some other stuff :) Aug 06 16:55:55 and do the locale stuff Aug 06 16:55:59 you get what you pay for from us Aug 06 16:56:53 Crofton: means ? Aug 06 16:57:31 im tring make my first BB file for openbox but i got some erros Aug 06 16:57:32 http://paste.debian.net/34124 Aug 06 16:57:34 we are working for free Aug 06 16:57:57 so won't necessarily jump when some asks a question Aug 06 16:58:02 i can set automake options ? Aug 06 16:58:20 Crofton|home: yeah thats true Aug 06 16:58:29 is cygwin a supported OE environment? Aug 06 16:58:31 linux-omap1_2.6.x+git.bb Aug 06 16:58:42 jsheldon, I have not heard of anyone trying it Aug 06 16:59:03 psycoman: i think if your remove --warnings=cross it should get past that Aug 06 16:59:21 Crofton: interesting, I just kicked of bitbake angstrom-console-image, I'll report back with the results :) Aug 06 16:59:30 khem, the git bb file should likely just use the default config based on machine Aug 06 16:59:31 wirelessdreamer: the question is how ? rs Aug 06 17:00:22 Crofton: 2430 support is pretty new in kernel so it would require relatively new kernel something like 2.6.22 to be on safeside I will figure out Aug 06 17:00:47 that's why I suggest extending the git recipe to pull from git Aug 06 17:00:51 as a start Aug 06 17:01:11 why are trying to redcue "kernel recipe proliferation" Aug 06 17:01:17 Crofton|home: OK Aug 06 17:01:29 makes sense Aug 06 17:01:49 The holy grail is one basic recipe for all kernels Aug 06 17:01:52 linux.inc Aug 06 17:02:00 kind of like one ring to rule them :) Aug 06 17:02:37 what is the command to select a machines defconfig? Aug 06 17:03:46 make omap_2430sdp_defconfig Aug 06 17:03:55 will do it for omap2430 e.g. Aug 06 17:04:24 for all available arm configs you can look into arch/arm/configs/ Aug 06 17:04:30 in kernel srcs Aug 06 17:04:42 wirelessdreamer: removing --warnings=cross dont work Aug 06 17:04:46 so we can add something like Aug 06 17:05:22 make ${MACHINE}_defconfig for saces where we do not override kconfig file Aug 06 17:05:47 there is also some efforts to select EABI versus OABI at configure time Aug 06 17:06:30 is 2430 considered omap2? Aug 06 17:06:37 hmmm we have to change $MACHINE to match kernels defs Aug 06 17:06:42 yep Aug 06 17:06:47 that is best Aug 06 17:06:51 thats why it starts with 2xxx Aug 06 17:07:01 there is omap3 already out there Aug 06 17:07:10 maybe we can end up with linux-omap.inc that works for omap{1,2,3} Aug 06 17:07:59 and davinci :) Aug 06 17:08:14 yes it could actually Aug 06 17:08:28 I am hacking at a davinci boars atm Aug 06 17:08:39 I use the OSK toolchain in my uboot attempts Aug 06 17:09:29 Every now and then we beg TI for dev boards :) Aug 06 17:09:42 I am lucky to have the OSK through some work we do at school Aug 06 17:10:21 and I have one of these Aug 06 17:10:23 http://www.lyrtech.com/DSP-development/dsp_fpga/sffsdrdevelopmentplatform.php Aug 06 17:10:53 anyway, time for lunch bbiab Aug 06 17:14:21 psycoman: can i see your new log Aug 06 17:18:47 wirelessdreamer: still only this error : akefile.am:466: `dist_doc_DATA' is used but `docdir' is undefined Aug 06 17:18:50 psycoman: either remove dist_doc_DATA or define docdir Aug 06 17:18:54 yea just saw that one Aug 06 17:18:55 if i comment 466 line all works Aug 06 17:19:15 then just make a patch for it and add a line to apply it in your bb file Aug 06 17:19:18 but this is not the right way ^^ Aug 06 17:19:38 psycoman: dont exist a --without-doc to handle this ? Aug 06 17:19:43 psycoman: i'm not familiar with dist_doc_DATA Aug 06 17:19:49 me too Aug 06 17:20:12 * psycoman psycoman is talk with himself @@ Aug 06 17:26:16 Crofton|home: any news from rishi? Aug 06 17:26:30 wirelessdreamer: well compile work fine, jsut still the doc issue Aug 06 17:31:12 psycoman: are you really going to install the docs on your system? Aug 06 17:31:59 psycoman: what package are you building btw? Aug 06 17:32:11 wirelessdreamer:i dont want install it, im make patch like you said Aug 06 17:32:15 wirelessdreamer: openbox Aug 06 17:35:55 psycoman: it isn't going to break anything, if someone cares enough they can fix it Aug 06 17:36:17 psycoman: now there will be another diecient wm along with e17 Aug 06 17:36:34 psycoman: you should try e17 though, runs really well :) Aug 06 17:39:41 wirelessdreamer: im trying make a "openbox/fbpanel" mini-desktop Aug 06 17:40:16 wirelessdreamer: e17 is fine but still unstable and i cannot find a task-applet for opened windows Aug 06 17:40:41 wirelessdreamer: any it is just for study bb system Aug 06 17:40:49 s/any/any qay Aug 06 17:40:56 s/qay/way Aug 06 17:46:35 e17 unstable eh? Aug 06 17:50:06 to make a diff for bb, need more option then -u ? Aug 06 17:50:23 i usually use -Nurd Aug 06 17:50:47 e17 is very stable Aug 06 17:50:52 have been using it on my z for a few months Aug 06 17:51:46 wirelessdreamer: unstable as brach dont in use i told Aug 06 17:51:55 wirelessdreamer: you have a zaurus ? Aug 06 17:54:40 i have tried e17 but i get shifted screen :( Aug 06 17:55:49 does e17 draw to the fb then? Aug 06 18:11:25 re Aug 06 18:15:07 psycoman: yea i have a z c3000 Aug 06 18:15:24 the startup files in the feed are bad Aug 06 18:15:42 it use that file in the xterm type xrandr -o right to straighten in Aug 06 18:16:18 i looked around at the configs before but didn't see how angstrom was trying to start x, the normal .xinitrc and .xserverrc don't work as they should Aug 06 18:16:31 startx script acts a little strange Aug 06 18:16:43 i don't have autologin enabled either, just console Aug 06 18:24:43 could someone look at bug #2711 for consideration to merge? :) Aug 06 18:52:50 someone had mentioned a patch floating around to enable xrandr -s 640x480, does anyone here know where i could find it? Aug 06 18:53:05 err 320,240 that is Aug 06 18:54:36 koen: ping Aug 06 18:54:52 I was looking at the FAQ in the wiki Aug 06 18:55:21 I would like to add a question and also I might be able to answer one or two Aug 06 18:55:53 is it possible for me to edit wiki or in other words how does this work Aug 06 18:56:25 * khem kicks eglibc bases angstrom build Aug 06 19:01:40 hi everybody, Aug 06 19:02:00 mrdata: ping Aug 06 19:05:57 khem: the wiki should be editable by everyone Aug 06 19:06:15 * koen tries to get the glue of his fingertips Aug 06 19:09:45 mrdata: could you test if this http://lkml.org/lkml/2007/6/27/1719 touchscreen driver works on simpad? Aug 06 19:12:24 thesing: remove the trailing '9' from the uri Aug 06 19:12:28 hey zecke Aug 06 19:13:20 zecke: feedreader2 compiles, but webkit breaks somewhere in kjs :( Aug 06 19:14:23 koen: then use a older version Aug 06 19:14:30 thesing: pong Aug 06 19:14:45 zecke: that's what I did to check the compile Aug 06 19:14:53 koen: http://build.webkit.org/ looks okay now Aug 06 19:15:01 bitbake -b Aug 06 19:15:07 (webkit is already in staging) Aug 06 19:15:13 koen: http://build.webkit.org/post-commit-linux-gdk/builds/1315/step-compile-release/0 Aug 06 19:15:21 koen: was this your error? Aug 06 19:15:31 yes Aug 06 19:15:44 it is fixed ;) Aug 06 19:15:50 nice Aug 06 19:15:55 http://build.webkit.org/changes/2405 Aug 06 19:16:31 * zecke considers to finally start sleeping at the new flat Aug 06 19:17:12 thesing: had made tests over the weekend, changed the driver to work also on simpad Aug 06 19:18:34 thesing: but i'am not happy with it, one touchscreen press is okay, but press and hold works not Aug 06 19:20:12 thesing: with changes in old driver, i have more luck Aug 06 19:21:31 thesing: but the concept of using the interrupt method to wait for propper adc-sampling sounds not realy bad Aug 06 19:24:31 thesing: i would try to merge the best part from both and look at the result Aug 06 19:29:12 can we pinpoint git ? Aug 06 19:29:28 +v Aug 06 19:29:33 can we pinpoint a git revision? Aug 06 19:29:54 mickeyl: check the autobuilder for the last succesfull revision? Aug 06 19:30:25 no, i meant Aug 06 19:30:29 how to pinpoint a git revision Aug 06 19:30:33 given that we know one Aug 06 19:30:39 ouch wrong channel :) Aug 06 19:30:44 mickeyl: git-bisect Aug 06 19:30:44 mickeyl: yes, set tag in the SRU_URI Aug 06 19:30:53 sorry again :) Aug 06 19:31:19 RP: thanks Aug 06 19:34:42 mickeyl: ;tag= Aug 06 19:36:42 nod Aug 06 19:36:49 crap, looks like libtool-1.5.24 fails for Efika, but not OSK Aug 06 19:37:17 03mickeyl 07org.oe.dev * ra770244d... 10/ (1 packages/mrxvt/mrxvt_0.5.2.bb): mrxvt: add no-debug-x patch as per mrxvt author's recommendation Aug 06 19:37:25 03mickeyl 07org.oe.dev * re3b95b3f... 10/ (8 files in 4 dirs): add openmoko-terminal2, using mrxvt instead of gtkterm2 Aug 06 19:37:32 03mickeyl 07org.oe.dev * rd486b07f... 10/ (1 conf/distro/include/sane-srcdates.inc): sane-srcdates.inc: add sane srcdate for webkit (it breaks shortly after that) Aug 06 19:41:17 hi Aug 06 19:41:20 gm Aug 06 19:41:22 mickeyl: could you remove that SRCDATE? webkit has already been fixed Aug 06 19:41:24 hey likewise Aug 06 19:41:31 hey koen, Crofton Aug 06 19:41:34 likewise, in an annoying development Aug 06 19:41:37 koen: oh, good. that was quick Aug 06 19:41:50 it would appear I can build EFika with libtool 1.5.10, but not 1.5.24 Aug 06 19:42:02 mickeyl: scroll up 30 minutes :) Aug 06 19:42:04 but, I can build OSK with 1.5.24 .... Aug 06 19:42:12 Crofton|home: both are in current? Aug 06 19:42:24 yes Aug 06 19:42:31 1.5.24 in DEF_PREF="-1" Aug 06 19:43:23 Crofton|home: your local.conf looks like mine? I.e. do I need to set any specifics to reproduce here? Aug 06 19:43:40 hang on Aug 06 19:44:38 http://pastebin.ca/647835 Aug 06 19:45:19 I've saved a couple copies of tmp, so I can compare commnd lines Aug 06 19:45:23 I'll do this tomorrow Aug 06 19:45:24 AM Aug 06 19:45:38 03mickeyl 07org.oe.dev * rda728408... 10/ (1 conf/distro/include/sane-srcdates.inc): sane-srcdates.inc: webkit builds again Aug 06 19:46:08 mickeyl: thanks to the wonders of SRCDATE the fix will be available in ~3 hours Aug 06 19:46:18 or when someone implements SRC_FORMAT :) Aug 06 19:47:27 sadly, I have a new bb file that needs the later libtool Aug 06 19:59:15 hmm Aug 06 19:59:25 anyone know offhand how to get the revision of a checked out tree? Aug 06 19:59:28 git, that is Aug 06 20:00:12 git-log and see the last rev? Aug 06 20:00:12 /first Aug 06 20:02:35 Crofton|home: which bitbake version do you use? Aug 06 20:02:48 1.8.7 Aug 06 20:02:56 hmm Aug 06 20:02:56 oe@mahavidya:/space/fic/openmoko-oe.dev/fic-gta01/tmp/work/armv4t-angstrom-linux-gnueabi/uboot-gta01-1.2.0+svn20070804-r11/git$ git log Aug 06 20:02:56 fatal: Not a git repository Aug 06 20:02:57 mickeyl: you could look at refs/master as well **** BEGIN LOGGING AT Mon Aug 06 20:10:29 2007 Aug 06 20:16:16 Crofton|home: could you use diff -Nurd? Aug 06 20:16:46 sure Aug 06 20:17:53 http://pastebin.ca/647891 Aug 06 20:17:58 Crofton|home: "powerpc-angstrom-linux-g++: -lstdc++: linker input file unused because linking not done" looks suspicious, though Aug 06 20:18:27 yeah, I am just frustrated because it worked for teh OSK Aug 06 20:19:01 the error seems like it missed libstdc++ Aug 06 20:19:44 I am going to spend time tomorrow AM looking at it, need to do web monkey work atm Aug 06 20:19:47 and look at thesis Aug 06 20:20:02 if you guys figure it out b4 then, awesome :) Aug 06 20:20:21 It would be really nice to move to current libtool with pkgconfig and autoconf Aug 06 20:20:32 koen hm mwester wants STAGING_LIBDIR_NATIVE Aug 06 20:22:47 mickeyl: NOTE: package webkit-0.0+svn20070806: completed :) Aug 06 20:23:47 * mwester looks around, puts his hands in his pockets, and whistles innocently... Aug 06 20:25:59 koen: good Aug 06 20:26:38 There are a few recipes that set LD_LIBRARY_PATH to include STAGING_LIBDIR -- I think they are technically wrong, as that only works iff ${BUILD_SYS} == ${HOST_SYS}... Aug 06 20:31:29 mrdata: will you post this on lkml? Aug 06 20:31:51 hi thesing Aug 06 20:31:59 hi woglinde, Aug 06 20:33:37 thesing: i'am not shure, my tests are not fully complete, and i must subscribe before Aug 06 20:34:26 mrdata: But it would help to get a working unified driver. Aug 06 20:34:56 thesing: yes, but only for two tested devices Aug 06 20:35:05 mrdata then describe Aug 06 20:35:13 aeh subscribe Aug 06 20:36:20 one driver for two devices is better than two drivers for two devices. Aug 06 20:37:10 yes, this make sense Aug 06 20:37:24 koen http://www.intercal.entadsl.com/download/CLC-INTERCAL-1.-94.-3.ddsh.gz <--- you need to read this shell script Aug 06 20:38:10 an if the driver works for two devices chances are better that it will work for other devices. Aug 06 20:38:23 on arm we have to many drivers for at least one device Aug 06 20:39:10 okay, first i will also make tests for suspend/resume with the new driver Aug 06 20:39:47 but not possible with rootfs over nfs ;-) Aug 06 20:40:46 mrdata hm you could try it with an initramfs Aug 06 20:42:31 sorry, have this not done before Aug 06 20:43:21 initramfs are great file with all files inside and could start with bootloader? Aug 06 20:43:50 mrdata yes Aug 06 20:44:12 mrdata I put the howto in the opensimpad wiki Aug 06 20:44:12 okay, but could this also done with redboot? Aug 06 20:44:20 hm Aug 06 20:44:24 I think so Aug 06 20:45:00 also transfer initramfs over tftp and then start it on simpad Aug 06 20:47:08 chouimat: wth? Aug 06 20:47:13 dd foo | grep Aug 06 20:47:23 why not 'grep foo ' Aug 06 20:47:57 koen because it's an archive file and dd is there to extract files Aug 06 20:48:22 wirelessdreamer: zottd just landed in #angstrom Aug 06 20:52:31 koen: I am getting this error when building with eglibc Aug 06 20:52:32 http://pastebin.ca/647944 Aug 06 20:52:57 koen: This does not happen on my builds that I do differently for toolchain Aug 06 20:53:18 koen: What is the trick to find out the build order Aug 06 20:55:02 khem, i have this issue too Aug 06 20:55:05 anyway, time to go sleep now, night all Aug 06 20:55:27 night. Aug 06 20:55:29 psycoman: I can try to fix it Aug 06 20:55:48 psycoman: I need to understand to build order of OE Aug 06 20:55:59 * khem is only a 5 day old OE baby Aug 06 20:56:10 khem bitbake -g generates a dependency graph Aug 06 20:56:23 keesj, me too Aug 06 20:56:26 likewise: Thanks Aug 06 20:56:28 khem: do you have glibc in staging by chance? Aug 06 20:56:30 khem: 'bitbake -nv' to see a dryrun Aug 06 20:56:41 koen: thats the problem Aug 06 20:58:34 koen: NOTE: Handling BitBake files: \ (4521/4521) [100 %] Aug 06 20:58:34 NOTE: Parsing finished. 4304 cached, 0 parsed, 217 skipped, 0 masked. Aug 06 20:58:34 Nothing to do. Use 'bitbake world' to build everything, or run 'bitbake --help' Aug 06 20:58:34 for usage information. Aug 06 20:59:04 sorry wrong window Aug 06 20:59:09 mia culpa Aug 06 21:01:10 koen:NOTE: Selecting already-staged glibc-intermediate (0:2.5-r6) to satisfy virtual/arm-angstrom-linux-gnueabi-libc-for-gcc Aug 06 21:01:10 NOTE: selecting glibc-intermediate to satisfy virtual/arm-angstrom-linux-gnueabi-libc-for-gcc due to PREFERRED_PROVIDERS Aug 06 21:01:15 could this be problem Aug 06 21:04:00 yes Aug 06 21:04:15 khem: yes, I need to add eglibc-intermediate and eglibc-initial Aug 06 21:04:15 I'll do that tomorrow Aug 06 21:04:50 koen: OK I will create a local copy if I can Aug 06 21:04:59 e17, eglibc, what is next? elinux? Aug 06 21:05:03 koen: it must be late there Aug 06 21:05:39 I am sure eglibc will be right choice in OE Aug 06 21:06:16 hmmm looking in logs Aug 06 21:06:19 I see NOTE: Selecting already-staged glibc (0:2.5-r6) to satisfy virtual/libc Aug 06 21:06:34 So it seems I need to change this in my local.conf too Aug 06 21:06:37 23:06 here Aug 06 21:07:21 Ah so you in CET Aug 06 21:08:22 khem, it helps to get up early for OE work :) Aug 06 21:08:38 heh Aug 06 21:08:54 but Crofton|home will be happy to help you for the rest of your day :-) Aug 06 21:09:41 likewise: yeah I will keep writing my frustrations out to channel :) Aug 06 21:10:11 khem: it helps to cheer inbetween :-) Aug 06 21:10:27 khem: OE has been a real work horse for us Aug 06 21:10:35 build monkey that is Aug 06 21:10:36 cest Aug 06 21:10:44 khem Aug 06 21:10:46 like when it breaks down, you shoot it and turn it into glue Aug 06 21:10:49 summertime Aug 06 21:11:59 woglinde: I remember once it was 35C in Ulm and it was very hot Aug 06 21:12:11 I use to swim in Donau Aug 06 21:12:24 khem ;) Aug 06 21:12:28 30C here today Aug 06 21:12:40 seems we get a little bit summer back Aug 06 21:12:48 july was cold as autmn Aug 06 21:12:51 and rainy Aug 06 21:13:20 woglinde: global cooling Aug 06 21:14:37 hm is there a way for .bb recipe to include dirs with subir in FILES_${PN}-bla = "..." Aug 06 21:14:43 subdirs Aug 06 21:22:02 khem: I just pushed two updates that make eglibc support a bit better Aug 06 21:22:17 khem: it needs a third bigger change which I need to test first Aug 06 21:22:29 03mickeyl 07org.oe.dev * r2e5ca879... 10/ (1 packages/uboot/uboot-openmoko_svn.bb): uboot-openmoko: pinpoint GIT revision Aug 06 21:22:33 03mickeyl 07org.oe.dev * rff743fd6... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: add openmoko-feedreader2 and neod Aug 06 21:22:37 03mickeyl 07org.oe.dev * r6855225d... 10/ (1 packages/gstreamer/gst-meta-base_0.10.bb): gst-meta-base: remove gnomevfs stuff Aug 06 21:22:42 03koen 07org.oe.dev * r3af6b758... 10/ (1 packages/glibc/eglibc-initial_svn.bb): eglibc-initial: add completely untested recipe (copy of glibc-initial 2.6) Aug 06 21:22:46 03koen 07org.oe.dev * r1e75cab4... 10/ (1 conf/distro/include/angstrom-eglibc.inc): angstrom-eglibc: add more defines to stop glibc from ending up in staging Aug 06 21:23:52 * koen heads to bed now Aug 06 21:24:03 nite koen Aug 06 21:24:55 g'night koen Aug 06 21:31:39 nite koen Aug 06 21:37:05 koen: good nigh Aug 06 21:37:23 khem, how much will eglibc buy me? Aug 06 21:37:50 Crofton|home: I dont know but definitely more than glibc Aug 06 21:38:01 you mean less :) Aug 06 21:38:01 for OE Aug 06 21:38:35 I'm assuming the intent is to create a smaller version of glibc? Aug 06 21:38:44 I think it will be nice you can do cross testing of library and build locales in cross build env Aug 06 21:39:14 and wait a while it is going to become configurable that means you can throw out the bulk you do not need Aug 06 21:39:52 yep and cross buildable and support arches which are mainly used in embedded world like ARM MIPS etc etc. Aug 06 21:40:10 which are not maintained upstream in glibc they are in ports Aug 06 21:40:26 sounds interesting Aug 06 21:49:14 channel #angstrom **** ENDING LOGGING AT Tue Aug 07 02:59:57 2007