**** BEGIN LOGGING AT Wed Dec 17 02:59:57 2008 Dec 17 03:26:59 hi.all. when I build midpath-opengles_0.2+0.3rc1.bb , I get an error: Dec 17 03:27:55 the error message says: cpproc: In function 'cpproc_forkAndExec': Dec 17 03:27:55 cpproc.c:89: error: ignoring return value of 'chdir', declared with attribute warn_unused_result Dec 17 03:28:21 can anybody help me to solve my problem? Dec 17 03:36:20 favor: check upstream for patches/fixes/bugreports. That sort of error tells you that there is a code problem; a function was called and its return value ignored, but the coder instructed the compiler to warn if such a thing was ever done. Dec 17 03:41:50 Hi everyone... I'm having trouble with a JFFS2 root file system image and was wondering if anyone has seen this before Dec 17 03:42:08 I get the following once the kernel mounts the partition: Dec 17 03:42:09 JFFS2 error: (1) jffs2_build_inode_pass1: child dir "lib" (ino #427) of dir ino #403 appears to be a hard link Dec 17 04:05:36 I build gstreamer bb-file and also Qt embedded phonon model rootfs, but I still can't run phonon example application? Dec 17 04:06:50 I was frustrated by this. do somebody have any suggestion to me? Dec 17 05:03:30 hi. I come across many problem in using oe. I doubt I do not really know how do oe work. I get an erro when i build qte with phonon: The path does not exist:/home/user/oe/beagleboard/tmp/work/armv7a-angstrom-linux-gnueabi/qt-embedded-4.4.3-r1/image/usr/lib/qtopia/plugins/codecs/. please help me. I am a newbie on oe Dec 17 06:09:21 03Robert Schuster 07org.openembedded.dev * r03d5c900d9 10openembedded.git/packages/midpath/files/fix-openfile.patch: midpath 0.2rc2: Fixed patch. Dec 17 07:02:04 I can never seem to get OE to work Dec 17 07:02:15 my builds always fail for one reason or another even when I follow the directions Dec 17 07:02:51 example: ERROR: '/home/mutex/openembedded/packages/linux/mx31ads-kernel_2.6.19rc6.bb' failed Dec 17 07:03:04 is that normal ? Dec 17 07:12:03 mutex: are you working on the dev branch? If so, yes, it's quite normal for things to be broken rather regularly. Dec 17 07:13:39 but in any case, that message says nothing, the preceding messages either gave you the snippet of the log file that contained the error it encountered, or it called out the pathname for that log file. THat will tell you what went wrong. Dec 17 07:36:16 how come /etc/resolv.conf gets cleared everytime I boot? Dec 17 08:56:55 morning Dec 17 09:56:45 genext2fs fails: http://codepad.org/MCuBW5VF Dec 17 09:57:28 I've built the base-image without any errors, but the console-image fails. I checked the logs, and this is the command that fails. Dec 17 09:58:53 anyone has git remove branch right online? Dec 17 09:59:19 git push origin :john_lee/openmoko-merges Dec 17 09:59:19 *** Deleting a branch is not allowed in this repository Dec 17 10:01:39 zecke: ping Dec 17 10:14:13 Should networking work with the base-image? Dec 17 10:15:15 (angstrom dist, x86 target) Dec 17 10:26:30 Perhaps I should switch to the stable branch =) Dec 17 10:32:44 johncylee: pong Dec 17 11:01:07 good morning Dec 17 11:03:20 hi florian Dec 17 11:53:22 raster: hey Dec 17 11:53:39 raster: bossa guys created an EFL port of WebKit... it looks like a evas smartobject... Dec 17 12:07:58 zecke: I have no permission to delete a branch Dec 17 12:08:28 zecke: and I rebase -i my branch now I cannot force update Dec 17 12:26:50 crap... Dec 17 12:27:11 hosting git via http require possibility to run shell scripts on http server Dec 17 12:27:16 hi hrw Dec 17 12:27:27 hrw hm Dec 17 12:33:05 * * OE Bug 4940 has been created by Dec 17 12:33:08 * * ffmpeg-0.4.9+r36+gitr48b4495cdae4a4bc2bc65902f3214c1b084be01d-autobuild Dec 17 12:33:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4940 Dec 17 12:37:12 so I cant share my git tree on my web hosting ;( Dec 17 12:47:03 johncylee: oh you can force the push?! Dec 17 12:47:16 johncylee: you are right that you can not delete a branch.. no one can Dec 17 12:52:29 zecke: I've not read the history of this conversation, but I think you can always delete a branch with "git branch -D", even in a bare-repository that has been designated as "master" one. However, if only one does "git push --all", it's back --- except if you write a hook that checks for this and aborts the incoming push. Dec 17 12:53:32 zecke: To git@git.openembedded.net:openembedded Dec 17 12:53:32 ! [rejected] org.openembedded.dev -> org.openembedded.dev (non-fast forward) Dec 17 12:54:02 johncylee: --force Dec 17 12:54:16 schurig: right, it is an ACL that forbids deleting any branch Dec 17 12:54:45 schurig: sure thing, it could allow to delete any other branch than org.open* Dec 17 12:55:21 03Chia-I Wu 07john_lee/openmoko-merges * r64b844ec5c 10openembedded.git/packages/udev/ (4 files in 4 dirs): fastboot: udev-static-devices: New package to provide static dev nodes. Dec 17 12:55:21 03Chia-I Wu 07john_lee/openmoko-merges * rc6c0673757 10openembedded.git/packages/wireless-tools/ (2 files in 2 dirs): fastboot: wireless-tools: Make zzz-wireless.if-pre-up only for dhcp. Dec 17 12:55:21 03Chia-I Wu 07john_lee/openmoko-merges * rc0041ba98c 10openembedded.git/packages/alsa/ (alsa-state.bb alsa-state/openmoko/alsa-state): fastboot: alsa-state: Run init script earlier and in the background. Dec 17 12:55:23 03Chia-I Wu 07john_lee/openmoko-merges * r288a7b1dc8 10openembedded.git/conf/machine/om-gta02.conf: fastboot: om-gta02.conf: Set USE_VT to 0. Dec 17 12:55:26 03Chia-I Wu 07john_lee/openmoko-merges * r27952ca6f4 10openembedded.git/packages/udev/ (files/links.conf udev-118/openmoko/init udev_118.bb): fastboot: udev: Update init script to support udev-static-devices. Dec 17 12:55:27 zecke: okay Dec 17 12:55:30 03Chia-I Wu 07john_lee/openmoko-merges * r8c97946061 10openembedded.git/packages/modutils/ (files/openmoko/modutils.sh modutils-initscripts.bb): fastboot: modutils: Do not run depmod. Dec 17 12:55:35 03Chia-I Wu 07john_lee/openmoko-merges * r0eaaabc53b 10openembedded.git/packages/procps/ (4 files in 3 dirs): fastboot: procps: Install and use /etc/sysctl.conf. Dec 17 12:55:40 03Chia-I Wu 07john_lee/openmoko-merges * rb3b5fee1f8 10openembedded.git/packages/tasks/task-openmoko-linux.bb: Dec 17 12:55:43 fastboot: task-openmoko-linux: Cleanup. Dec 17 12:55:45 Remove packages already in task-base. Remove splash support. Add Dec 17 12:55:47 udev-static-devices. Dec 17 12:55:49 03Chia-I Wu 07john_lee/openmoko-merges * rf1b6139604 10openembedded.git/packages/netbase/netbase/openmoko/ (init options): fastboot: netbase: No longer parse /etc/network/options. Dec 17 12:55:52 03Chia-I Wu 07john_lee/openmoko-merges * r4974eeafa3 10openembedded.git/packages/sysvinit/sysvinit/openmoko/ (rc rcS): Dec 17 12:55:55 fastboot: sysvinit: Simplify rc and rcS. Dec 17 12:55:59 Remove support for boot splash. Source initscripts end with .sh. Dec 17 12:56:01 johncylee: did you find it? :) Dec 17 12:56:01 03Chia-I Wu 07john_lee/openmoko-merges * r85e6bbce9f 10openembedded.git/packages/xserver-kdrive-common/ (xserver-nodm-init.bb xserver-nodm-init/xserver-nodm): fastboot: xserver-nodm-init: Start at 05. Dec 17 12:56:12 03John Lee 07john_lee/openmoko-merges * rf42a49dc8c 10openembedded.git/packages/avahi/ (7 files): Dec 17 12:56:15 avahi: fix postinst script error Dec 17 12:56:17 The previous version of dbus-1 will not stop clean if there Dec 17 12:56:19 exists running 'dbus-daemon --session' in the system. That Dec 17 12:56:21 will cause the script to believe it should ask for a Dec 17 12:56:23 force-reload of dbus, which leads to an error exit because the Dec 17 12:56:25 'dbus-daemon --system' wasn't really running. Dec 17 12:56:31 03Chia-I Wu 07john_lee/openmoko-merges * rf10d679508 10openembedded.git/packages/base-files/ (3 files in 3 dirs): fastboot: base-files: Simplify fstab. Dec 17 12:56:34 03John Lee 07john_lee/openmoko-merges * r18f748e076 10openembedded.git/packages/xserver-kdrive-common/xserver-nodm-init.bb: xserver-nodm-init.bb: really update INITSCRIPT_PARAMS during upgrade. Dec 17 12:56:43 03John Lee 07john_lee/openmoko-merges * rbc71c24a51 10openembedded.git/packages/alsa/alsa-state.bb: Dec 17 12:56:46 alsa-state.bb: fix wrong usage of update-rc.d.bbclass Dec 17 12:56:48 Should use override ${PN} first Dec 17 12:56:50 03John Lee 07john_lee/openmoko-merges * r3476d7ae48 10openembedded.git/packages/alsa/alsa-lib_1.0.15.bb: Dec 17 12:56:53 alsa-lib_1.0.15.bb: fix PR, should be "r1" instead of "1" Dec 17 12:56:57 since the default is 'r0' Dec 17 12:56:59 03John Lee 07john_lee/openmoko-merges * r72deb5e545 10openembedded.git/packages/xserver-kdrive-common/xserver-nodm-init.bb: Dec 17 12:57:02 xserver-nodm-init.bb: fix wrong usage of update-rc.d.bbclass Dec 17 12:57:04 Should use override ${PN} first Dec 17 12:57:06 03erin 07john_lee/openmoko-merges * rf51ad6650e 10openembedded.git/packages/ (3 files in 3 dirs): Dec 17 12:57:13 (91 lines omitted) Dec 17 12:59:38 is it possible that u cant overlay the glibc? i copied the glibc from -dev to my overlay directory for use with -stable but this overlay is NOT used Dec 17 12:59:41 others are Dec 17 13:01:09 nik0n: hihgly unlikely, are your bb collections setup correctly? Dec 17 13:01:51 yes, other overlays work as stated Dec 17 13:02:19 because somehow it looks like that glibc is not working as libc for arm targets in stable Dec 17 13:02:27 it is indeed working in dev ... Dec 17 13:03:09 do you get parse errors? Dec 17 13:03:33 no parse errors, but i see that the files got parsed Dec 17 13:05:47 and you want to replace glibc x.y.z with glibc x.y.z or a different version? Dec 17 13:12:32 hi ! Dec 17 13:17:47 johncylee: so we only miss julian's toolchain work and QtE? Dec 17 13:19:49 I build angstrom 2008.1 gpe-image for Motorola a1200e phone. System booted, ts calibrated. But when I try add new user at init dm screen. Chars dose not appear when I press on virtual keyboard keys. Keys are blue when I press on it. So ts is working. Dec 17 13:22:23 ~curse git http hosting Dec 17 13:22:25 May the fleas of a thousand camels infest your most sensitive regions, git http hosting ! Dec 17 15:03:08 ho ho ho, merry freakin morning OE people Dec 17 15:08:56 prpplague: gm ug ug ug Dec 17 15:09:08 likewise: greetings Dec 17 15:09:14 likewise: whats cookin today? Dec 17 15:09:30 cookin PCI Express drivers still, you beagling along? Dec 17 15:10:10 03Koen Kooi 07org.openembedded.dev * r8d61ddd9d9 10openembedded.git/packages/ffmpeg/ffmpeg.inc: Dec 17 15:10:10 ffmpeg: ARM_INSTRUCTION_SET = "arm" Dec 17 15:10:10 * some archs (e.g. armv4t) compile fine with thumb, but other don't (e.g. armv5te). Dec 17 15:10:10 * ffmpeg is also one of the libraries that go faster in ARM mode Dec 17 15:10:10 * note that ffmpeg has assembler that will use e.g. thumb2 options where it makes sense, so this is more nuanced than it seems Dec 17 15:10:32 hi kergoth Dec 17 15:11:12 likewise: yea, boss has me working on an expansion board for the beagle, tbh the beagle doesn't really spark my interest due to the lack of hardware hacking features Dec 17 15:11:15 03Koen Kooi 07org.openembedded.dev * r0fb4b44bb3 10openembedded.git/packages/povray/povray_3.6.1+3.7.0-beta25b.bb: povrayL fix up build a bit Dec 17 15:11:46 hey Dec 17 15:11:58 prpplague: nr of headers underrun error? Dec 17 15:13:36 likewise: indeed Dec 17 15:13:59 kergoth: just missed my friendly morning greeting to the channel Dec 17 15:14:16 re Dec 17 15:38:42 anyone made progress on the gpe-info segfaults and other gpe alignment traps errors? Dec 17 15:40:49 prpplague: I just noticed that I can reproduce this here. If you remind me this evening I'll install the debug package and search for the reason. Dec 17 15:41:10 florian: ok thanks Dec 17 16:02:36 florian: hmm, seems mb-applet-wireless is segfaulting tooo Dec 17 16:03:06 * * OE Bug 4941 has been created by Dec 17 16:03:08 * * gdm-2.24.1-autobuild Dec 17 16:03:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4941 Dec 17 16:03:44 prpplague: that thingy is part of the old matchbox stuff and not maintained any more. I hope we can update to the later matchbox stuff easily... need to work on this in a free minute. Dec 17 16:04:21 florian: oh? what is the alternative to mb-applet-wireless? Dec 17 16:08:33 prpplague: hum well, not sure if there is a direct replacement but there should be two applets that handle all the networking stuff. Dec 17 16:09:04 florian: hmm, that is not good Dec 17 16:09:45 prpplague: depends on what you have in mind what the applet should do Dec 17 16:10:28 florian: hehe, pretty much exactly what the mb-applet-wireless does Dec 17 16:11:17 we need more developers here! Dec 17 16:11:53 florian: well, my problem is i don't normally do userland stuff Dec 17 16:12:24 florian: which is why i was totally surprised to see the pda environment still in shambles after years of linux on handhelds Dec 17 16:12:33 prpplague: I ususally do more userland but this year seems to be the year of linux ports :) Dec 17 16:12:50 hehe Dec 17 16:13:24 * kergoth kind of misses application and tools development Dec 17 16:13:30 prpplague: I guess most of these problems are caused by changes in the layers below... mb-applet-wireless used to be quite stable Dec 17 16:13:54 * chouimat|work looks at all the perl crap he's had to maintaint ... man MFC look nice Dec 17 16:13:56 kergoth: I wouldn't complain about a userland project too :) Dec 17 16:14:20 development... Dec 17 16:14:22 * mwester mutters that the problem is the wholesale replacement of normal config tools with bloated monstrosities like dbus. Dec 17 16:14:46 * florian would like to work on Pim synchronisation and an easy to use frontend for OE Dec 17 16:14:47 I had to debug device remotely today. but someone unplug serial cable... Dec 17 16:14:54 wireless *used* to work on amost everything. Then redhat came along with NetworkMangler Dec 17 16:15:03 and with 6 time difference it is hard to cooperate sometimes Dec 17 16:15:15 mwester I agree with you Dec 17 16:15:27 hrw: heh Dec 17 16:15:34 NM... please do not use that curse here Dec 17 16:15:39 And now bluetooth has followed suit, except they don't even HAVE an app to turn on bluetooth anymore. "Write your own - it's easy, just send dbus messages!" Yeah, sure -- 8MB of total flash, dbus is bloat. Dec 17 16:15:42 my daughter is few meters from me Dec 17 16:15:55 mwester lol Dec 17 16:16:09 hrw: ;) Apologies! :P{ Dec 17 16:16:17 mwester: amen Dec 17 16:16:27 mwester: i hate dbus Dec 17 16:16:50 mwester I'm not looking at the day I will have to hook up the media server in the product we work on with dbus Dec 17 16:19:08 poky was the only thing that started looking organized, and then intel scouped them up, i expect they will continue doing dev, but not so much oriented towards non-intel devices Dec 17 16:21:06 prpplague: I hope that Poky will be maintained still Dec 17 16:22:17 hrw: yea Dec 17 16:22:30 it was really nice to play with it Dec 17 16:22:56 today I maintain one distro based on Poky 'pinky' (last official release) and one based on OE Dec 17 16:22:59 hrw: my problem is that we are trying to compete with wince style userland, and the linux userland is basically terrible compared to wince Dec 17 16:23:24 this on Poky just works. that on OE... kind of works if xserver 1.3.0.0 is used Dec 17 16:23:47 prpplague: is there something called 'linux userland'? Dec 17 16:24:05 Most of the OE developers were not as financially motivated as the Poky developers Dec 17 16:24:21 CosmicPenguin: indeed Dec 17 16:27:20 its also a lot easier to maintain a minimal metadata repo than one the size of OpenEmbedded Dec 17 16:27:25 indeed Dec 17 16:27:37 meta/ in Poky passes 'bitbake world' Dec 17 16:27:51 OE + bitbake world == fail on start Dec 17 16:28:18 kergoth: agreed Dec 17 16:29:06 hrw: hehe, that is like comparing OpenWRT and OE just by testing if typing "make" is working Dec 17 16:30:04 world is a bi tmessy due to the way we handle providers Dec 17 16:30:08 heh Dec 17 16:31:16 time for assignments Dec 17 16:32:18 fun fun Dec 17 16:32:29 * kergoth decides to read some more in this c# book Dec 17 16:33:25 zecke are you/have you taken information theory? Dec 17 16:34:02 kergoth: i'm just surprised at the lack of consolidation of the applications designed for a pda/handheld Dec 17 16:34:09 Crofton|work: a year ago :) Dec 17 16:34:10 man, i really should go back to school one of these days.. hard to get the cash and time together when working full time though, with lots of stuff to save for Dec 17 16:34:30 kergoth, do it Dec 17 16:34:33 even part time Dec 17 16:34:36 prpplague: shouldn't be, thats the nature of open source, more forks and splits than consolidation Dec 17 16:34:43 prpplague: if you feel audacious, try 'connman', the moblin's connection manager...the holy grail Dec 17 16:34:53 zecke, what sort of degree will you end up with Dec 17 16:35:02 Crofton|work: yeah... itll take many yeras.. but if i dont, many years from now i'll have nothing instead of a degree, and will be kicking myself :) Dec 17 16:35:11 Crofton|work: I'm always irritated by users that do not understand what is going on, but then still filter the log file Dec 17 16:35:14 right Dec 17 16:35:35 :) Dec 17 16:35:51 Crofton|work: no idea, I might not finish university at all. So anything between B Sc., M Sc, Diplom Informatiker and Moron is possible Dec 17 16:35:52 kergoth: yea, i understand forks, but i figured there would be some core distros that were fairly robust, i.e. familar started off with what i would have expected Dec 17 16:35:59 I was amused you managed to work information theory into the reply :) Dec 17 16:36:17 zecke, get some form of degree before quitting Dec 17 16:37:07 Crofton|work: I think the reason I'm thinking of it again is that stefan_schmidt gave me "Cryptonomicon" of Neal Stephenson Dec 17 16:37:14 prpplague, I suspect there are some, but they are not feeding back Dec 17 16:37:18 prpplague: heh.. yeah, there is that... i expect there are a lot of minimal OpenEmbedded based distros floating around, but most are probably kept internal by companies using them, rather than being like poky Dec 17 16:37:21 Crofton|work: heh :) Dec 17 16:37:22 ah Dec 17 16:38:15 OE is provding 90% of the work, the final 10% is very product specific Dec 17 16:38:54 even in OE, discussuons get "harder" as people try to converge on something very focused Dec 17 16:39:07 ok lunch Dec 17 16:39:40 distros like angstrom could probably use more QA. most of our qa is centered around the builds.. but itd be cool to set up, say, a test using qemu to bitch and moan if the distro doesnt successfully boot to a login prompt with a console-image Dec 17 16:40:14 kergoth: i don't find that that is really where the QA needs to be Dec 17 16:40:21 03Jeremy Lainé 07org.openembedded.dev * r9dd5f2f30a 10openembedded.git/packages/linux/ (9 files in 2 dirs): linux-2.6.26: update patches for boc01 Dec 17 16:40:45 its also "hard" trying to support a range of hw Dec 17 16:40:58 kergoth: heck my angstrom build went the very first time with no errors, plus it booted, but there tons of apps that simply didn't work or were misconfigured Dec 17 16:41:44 so, do we delete them, ot hope someone provides patches? Dec 17 16:41:55 zecke: oh that book is great Dec 17 16:42:05 the ultimate would probably be a board farm, not unlike the setup handhelds.org guys had with the ipaqs, but with a wide variety of dev boards, all running the distro in question, preferably with some sort of testing by remote controlling the units... Dec 17 16:42:07 heh Dec 17 16:42:09 Dec 17 16:42:20 florian: Hell yeah it is! Dec 17 16:42:26 and with the people to support it :) Dec 17 16:42:33 florian: hehe, I met Stefan in the train from Frankfurt -> Berlin, I came back from Taipei, he was back from Australia :} Dec 17 16:42:54 kergoth: hehe, sounds like a startup biz Dec 17 16:43:12 Kensan: well, with things like Openmoko, we have such a distributed cluster... Dec 17 16:43:19 kergoth: well, with things like Openmoko, we have such a distributed cluster... Dec 17 16:43:25 ah, nice Dec 17 16:43:35 morning Dec 17 16:43:41 kergoth: so with many packages we get user reports, fixing stuff all over the place Dec 17 16:43:54 kergoth: the question is what about regressions, and then again we have users :} Dec 17 16:44:03 prpplague, kergoth: I guess the main reason for lacking userlad software quality is that devices come and go too fast for the community to react. And in fact the support from companies is not the best in this area. Dec 17 16:44:27 could someone explain me what is the difference between TOOLCHAIN_TARGET_TASK and TOOLCHAIN_HOST_TASK ??? one is supposed to run on the target ? Dec 17 16:44:30 zecke: okay, that's a good one :) Dec 17 16:45:03 florian: indeed, our usual product life span is about 5 years, maybe it would be a good idea to base testing on our products, hehe Dec 17 16:45:13 kergoth: e.g. with Om we simply had to upgrade hal... before there was 0.5.9 and it leaked like hell, user based QA noticed... Dec 17 16:45:14 zecke: I met Stallman in Beijing... and we did not go to the same event. Dec 17 16:45:23 hehe Dec 17 16:45:40 florian: recently? that is even better than seeing germans Dec 17 16:45:57 zecke: no, about one year ago Dec 17 16:48:18 Longfield: TARGET_TASK defines which target related stuff to add into sdk, HOST_TASK defines host ones Dec 17 16:49:10 prpplague: I would be happy with that - any kind of reference hardware that is available for some time would be good. Dec 17 16:49:40 would be nice to cover arm4t/5t/6/7 Dec 17 16:50:36 prpplague: I currently do a linux port to a similar kind of device - but my experience is that most of the companies are not interested in a gui framework but in a single custom application. Dec 17 16:51:18 or contentrate more on getting devices working ten on userland Dec 17 16:52:56 someone knows UI for git which allow to export single changesets as format-patch? Dec 17 16:52:58 kergoth, hrw: That would be a nice project - we have a lot devices here we could use for this purpose. Dec 17 16:53:35 hi mickeyl Dec 17 16:53:39 yo hrw Dec 17 16:54:01 florian: there are few other nice things which we should do to make OE look better Dec 17 16:54:03 localtime(), pplz :) Dec 17 16:54:21 Jay7: 17:54 Dec 17 16:54:46 hrw: yes, we need an easy to use frontend... that's why that many people use buildroot. Dec 17 16:54:46 hrw: good evening :) Dec 17 16:55:01 hi mickeyl Dec 17 16:55:08 florian: *.openembedded.org != *.org.openembedded.net, we still not have e.V. etc Dec 17 16:55:21 hey flor Dec 17 16:55:26 ian Dec 17 16:56:33 oh good point Dec 17 16:56:41 florian: yea, i've got to get a unit to hrw as well Dec 17 16:56:50 florian: people also tend to use b uildroot for bringup more than oe, due to the lesser overhead, faster builds, and its easier to build rootfs's that are even smaller than we can.. no matter how granular our packaging is, you can't construct it file by file Dec 17 16:56:53 florian: we are finishing up a pre-production run now Dec 17 16:56:56 hrw: ah I see. For instance we could imagine that some additional build packages (like cmake, qmake, etc ...) would rather be in the HOST_TASK Dec 17 16:56:58 imo thats why, anyway Dec 17 16:57:06 mickeyl: did you receive the document for printing? Dec 17 16:58:02 florian: I cant count how many times I git cloned from not existing git.openembedded.org Dec 17 16:58:22 florian: yes, going to sign it and send it forward to you tomorrow. Send me your snailmail address, please Dec 17 16:58:26 prpplague: yes, would be good too... Dec 17 16:58:43 florian: http://www.elinux.org/M8050 Dec 17 16:58:54 florian: basically same specs as the gta-02 Dec 17 16:59:59 kergoth: imho the most important point is that it is very easy to understand how to get started with buildroot. Dec 17 17:00:03 Sharp has closed Zaurus line officially.. Dec 17 17:00:14 Jay7: finally Dec 17 17:00:16 kergoth: it wouldn't be that complicated for us to do somethign similar Dec 17 17:00:45 Jay7: http://marcin.juszkiewicz.com.pl/2007/01/12/zaurus-line-officially-ended/ Dec 17 17:00:51 hrw: yes we need to fix this domain stuff then Dec 17 17:01:07 frontends would help for some.. but OpenEmbedded is still a great deal more complex.. its more difficult to grasp how the thing is put together than it is to grasp the whole of buildroot Dec 17 17:01:09 R.I.P Dec 17 17:01:14 mickeyl: address... is somewhere... waait Dec 17 17:01:15 imo that could still be a barrier Dec 17 17:01:37 mickeyl: http://www.bsystems.de/index.php?cat=03_Contact Dec 17 17:02:05 * Crofton|work has never figured out buildroot :) Dec 17 17:02:22 I tried to use buildroot nearly 5 years ago. Dec 17 17:02:22 kergoth: the complexity is of course an advantage, but we could help the user that wants to get started quite a lot. Dec 17 17:02:26 Crofton|work: which is funny, because i've always had trouble with OE Dec 17 17:02:39 (and never with buildroot) Dec 17 17:02:39 :) Dec 17 17:02:47 but two guys from one irc channel shouted at me so I used Oe Dec 17 17:03:11 hrw: heh Dec 17 17:03:21 florian: aye. we could definately better document aspects of our metadata to help people understand things. even to what base.bbclass does, the variables used in the machine and distro configuration files.. Dec 17 17:03:24 * kergoth shrugs Dec 17 17:05:08 kergoth: yes that would be an important thing for people who want to go more into detail. Dec 17 17:05:33 03Marcin Juszkiewicz 07org.openembedded.dev * rae62c45784 10openembedded.git/packages/php/php_5.2.6.bb: Dec 17 17:05:33 php: package php-cgi into separate package (from BugLabs) Dec 17 17:05:33 git-svn-id: svn://svn.buglabs.net/bug/trunk/com.buglabs.build.oe@7711 35ad6e01-222d-0410-b96a-b41bba574936 Dec 17 17:05:33 03Marcin Juszkiewicz 07org.openembedded.dev * r0e080813b4 10openembedded.git/packages/uicmoc/uicmoc4-native.inc: Dec 17 17:05:35 uicmoc4-native: unbreak building on Gentoo - closes BL#290 (from BugLabs) Dec 17 17:05:37 git-svn-id: svn://svn.buglabs.net/bug/trunk/com.buglabs.build.oe@7667 35ad6e01-222d-0410-b96a-b41bba574936 Dec 17 17:05:43 03Marcin Juszkiewicz 07org.openembedded.dev * r160c32de43 10openembedded.git/packages/php/php_5.2.6.bb: Dec 17 17:05:43 php: depend on lemon-native to build sqlite stuff (from BugLabs) Dec 17 17:05:44 git-svn-id: svn://svn.buglabs.net/bug/trunk/com.buglabs.build.oe@7712 35ad6e01-222d-0410-b96a-b41bba574936 Dec 17 17:05:58 it would help to point them to simple things too, like the fact that its bitbake.conf that includes local.conf, it wouldnt be used otherwise.. its not "special", its just included Dec 17 17:06:11 heh Dec 17 17:06:20 kergoth: I would love to read such a document ... most of my (noob) problems with OE could be solved by reading such doc Dec 17 17:10:16 kergoth: more important: engineers who want to get somethign done do not want to read much of documentation before they can start. things like setting the correct environment and putting the pieces for local.conf together would be a good task for simplification by some ui or automatism. Dec 17 17:10:30 agreed Dec 17 17:10:45 * florian needs more time Dec 17 17:11:15 i was thinking of a tool that initializes a local build directory. links or clones the OpenEmbedded metadata repo, sets up the local.conf based on -m and -d commandline arguments for machine and distro, with bash and zsh tab completion for what distros and machines exist.. Dec 17 17:11:45 mkdir build; cd build; oecmd init -d angstrom -m beagleboard -t console-image; bitbake; Dec 17 17:11:47 or something Dec 17 17:11:48 heh Dec 17 17:12:56 kergoth: or some kind of "new oe project" wizzard Dec 17 17:13:16 yeah, same idea.. a wrapper around the initial setup, not just setting env vars, but the whole works Dec 17 17:13:30 ... and did I mention that the environment variables must die? ;) Dec 17 17:13:32 florian: I have the same tool in mind ;-) Dec 17 17:13:49 kergoth: something like http://pastebin.com/d49f5a641 script? Dec 17 17:14:00 kergoth: i hate to say this, but you might actually look at some of the build for wince, it really makes getting up and running easy Dec 17 17:14:01 kergoth: The wrapper in poky is probably close ;-) Dec 17 17:14:07 florian: heh, they were always suppsoed to be temporary... bbfiles was never supposed to stick around this long.. Dec 17 17:14:10 RP: nice Dec 17 17:15:02 kergoth: the main problem with the wince build is that the host pc can be horrible unstable Dec 17 17:17:28 * cbrake is impressed with driver quality in Linux compared to WinCE. In Linux, a lot of production code from people actually using it makes it back upstream. Dec 17 17:18:06 there is no such path in WinCE that I know of, so everything comes from MS, or chip vendors BSPs. Dec 17 17:18:14 if Linux driver quality is that much better than wince's, wince's must be REALLY bad :) Dec 17 17:19:10 cbrake: yea, thats what i was finding as well Dec 17 17:19:20 cbrake: very fractured dev Dec 17 17:19:54 kergoth: heh, well when you start with pre-production code, you can only imagine ... Dec 17 17:20:07 hehe Dec 17 17:20:33 kergoth: driver code does not really get wrung out until it has made it into a production product Dec 17 17:20:44 kergoth: yea, the major difference is the userland for wince is pretty rock solid, it is the driver code that isn't Dec 17 17:20:59 prpplague: heh, so just opposite :-) Dec 17 17:21:12 cbrake: exactly Dec 17 17:22:03 prpplague: which userland? where do I send my broken windows mobile phone (still based on a win ce stack)? Dec 17 17:22:08 prpplague: I think it depends. I've been very happy with the Linux basics -- GTK, X, base apps, etc. They always seem to work for me, and is about all most companies need. Dec 17 17:28:01 prpplague: have you tried the enlightenment environment? ... seems to be the preferred environment for embedded Linux (at least in OE) right now. Dec 17 17:28:13 Excuse me? Dec 17 17:28:24 enlightenment is the preferred environment>? Dec 17 17:28:42 preferred by (payed) developers ? Dec 17 17:28:53 ant_work: hi Dec 17 17:28:55 I really haven't been paying attention, have I? Dec 17 17:28:58 CosmicPenguin: :-) well, its what openmoko is using, its what I see on beagleboard demos. My gumstix board came with enlightenment installed, etc. Dec 17 17:29:31 Jay7: hello Dec 17 17:29:41 ant_work: what is right instructions to get git access to kexecboot? :) Dec 17 17:30:08 Jay7: ask me ;) Dec 17 17:30:16 thesing: hi :) Dec 17 17:30:51 Jay7: you need to register at linuxtogo and request to join the collie-kernel project. Dec 17 17:30:55 thesing: I'm asking :) Dec 17 17:31:50 03Jeremy Lainé 07org.openembedded.dev * ra283f38e6e 10openembedded.git/packages/u-boot/ (5 files in 2 dirs): u-boot-1.3.2: update patches for boc01 Dec 17 17:32:03 thesing: can't find 'register' link on www.linuxtogo.org Dec 17 17:32:11 thesing: did you mean that site? Dec 17 17:33:00 Jay7: projects.linuxtogo.org/projects/collie is site of the project Dec 17 17:33:07 bye Dec 17 17:33:45 zecke: windows mobile is a different story Dec 17 17:34:11 cbrake: sorry no, i've been working hard to just get a stable environment Dec 17 17:36:51 thesing: done Dec 17 17:38:55 Jay7: me too. Dec 17 17:39:40 thesing: is here any instructions? Dec 17 17:40:18 about order of committing or vice versa Dec 17 17:41:51 Jay7: we two are the only devs of kexecboot atm. Dec 17 17:42:31 Jay7: until now this repository was little more than my backup in case of a hard disk crash. Dec 17 17:42:40 Jay7: so no strict rules ;9 Dec 17 17:42:49 �� Dec 17 17:42:54 ok Dec 17 17:43:15 what is git url of this project? Dec 17 17:44:06 git://git.linuxtogo.org/home/thesing/kexecboot.git Dec 17 17:47:12 Jay7: beside the points koen mentioned you might want to create a 'run()' function for the fork+exec stuff. I think the code will be easier to read then. Dec 17 17:48:19 thesing: is that now called kexec_execw()? Dec 17 17:50:15 Jay7: btw last build faled to kexec...console=tty0 changes? Dec 17 17:50:39 ant_work: hm.. I don't know Dec 17 17:50:54 I'm back, I'll see later Dec 17 17:50:55 what is diagnostics? Dec 17 17:50:57 ok Dec 17 17:51:01 ant_work: unlikely Dec 17 17:51:34 thesin: I admit I was playing with several versions Dec 17 17:51:52 2.6.26 (our 'alpha') works Dec 17 17:52:00 Jay7: yes exactly. I don't remember reading that ... strange. Dec 17 17:53:27 thesing: should I rename it? Dec 17 17:55:54 Jay7: I don't have a strong opinion here. If we get more utility functions we should split them into an separate file and name them more generic. Dec 17 17:56:21 thesing: I'll do some code review and (may be) commenting.. Dec 17 17:56:55 then we will look again :) Dec 17 18:01:39 thesing: so the agreement is to use standard defconfigs and sed them later? Dec 17 18:02:17 Jay7: thanks for your help. Dec 17 18:02:48 ant_work: its an option. you can create a custom defconfig if you like. Dec 17 18:03:21 thesing: for c7x0 there are already 3 kernels :/ Dec 17 18:03:26 ant_work: in a perfect world we would have some tool in oe to manage kernel configuration based on machine and distribution. Dec 17 18:03:44 I know about one Dec 17 18:03:51