**** BEGIN LOGGING AT Fri May 18 02:59:56 2007 May 18 07:31:21 good morning May 18 07:39:33 morning :) May 18 08:11:33 03koen 07org.oe.dev * rdc6962ae... 10/ (21 files in 12 dirs): pointercal: merge from poky, add some more from http://linuxtogo.org/gowiki/AngstromPointercals May 18 08:11:39 03koen 07org.oe.dev * r970eed5b... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: prefer pango 1.16.4 May 18 08:11:50 03koen 07org.oe.dev * r8b7c04c6... 10/ (1 packages/tslib/tslib_1.0.bb): tslib: ship pointercal files for angstrom May 18 08:12:10 03koen 07org.oe.dev * rc0d6b6ea... 10/ (1 packages/tslib/tslib_1.0.bb): tslib: fix RRECOMMENDS May 18 08:13:29 erk, someone stole the #gpe channel May 18 08:13:59 that donkeys arse is at it again May 18 08:16:59 morning koen, XorA May 18 08:17:05 hey likewise May 18 08:18:29 yo likewise May 18 08:33:58 a wish list thingy: I don't know how to call it, the task sequences of building a big target, for example opie-image, could be better optimized for people baking on the road with limited space on laptop May 18 08:34:24 right now, it downloads and unpacks nearly everything first May 18 08:35:28 it's better if I can clean up some stuff in the work dir after those are built May 18 08:39:15 INHERIT += "rm_work" May 18 08:39:45 in local.conf? May 18 08:39:54 yes May 18 08:42:01 well, i don't think you quite get my point. maybe it's just an opie thing. it tries to download everything before configure/compile, rather than finish compile one by one. May 18 08:42:40 that's the multithreaded bitbake May 18 08:42:55 koen: I should use a sane mailer, but I won't :} May 18 08:43:54 xjqian: does it download more than what is needed? May 18 08:44:58 likewise: very likely. I don't think opie-mindbreaker depends on opie-snake May 18 08:45:01 etc. May 18 08:45:10 you can try a bitbake opie-image yourself May 18 08:46:30 xjqian: I'm running out of disk space myself :-) May 18 08:46:31 it's checking out cvs for > 5min now after all the key qte libs are built and before doing a single configure/compile May 18 08:47:04 keepings checking out cvs .... May 18 08:48:58 I need to take a look at preferred-opie-versions-1.2.3-pre.inc May 18 08:55:50 nothing there, I think i need to understand how bitbake handle image targets May 18 09:06:47 ibot: test May 18 09:06:59 Oh, no! There's a test and I haven't studied! May 18 09:08:01 koen: I traced opie.bbclass->palmtop.bbclass->palmtop-base.bbclass, but can't find any indication of turn-on multithreading May 18 09:08:30 koen: and I'm sure i don't have multithreading on my own conf May 18 09:08:53 xjqian: you skipped the qmake bbclass May 18 09:08:54 don't know where the multithreaded bitbake kicks in May 18 09:10:54 zecke, I meant opie->palmtop->qmake->qmake-base May 18 09:11:23 xjqian: I have not followed your discussion. But multithreaded as in Qt Multithreaded, or BitBake forking and having more than one job? May 18 09:13:33 koen: I'm just following koen's suggested diagnosis about why bitbake opie-image not compiling opie user space programs sequentially May 18 09:13:43 sorry, to zecke May 18 09:14:32 xjqian: well it is a sequence but it is not linear (or whatever) May 18 09:16:16 zecke: yeah, that's my original question. well, I'm not complaining much, just a quibble May 18 09:16:35 03koen 07org.oe.dev * rb940cb01... 10/ (3 files in 3 dirs): xserver-kdrive-common: fix DPI for fic-gta01 May 18 09:31:48 ibot: test May 18 09:31:58 Oh, no! There's a test and I haven't studied! May 18 09:34:46 xjqian: bitbake needs a better multithreading scheduler ;-) May 18 09:36:54 03koen 07org.oe.dev * r68277678... 10/ (1 packages/gpe-bluetooth/gpe-bluetooth_svn.bb): gpe-bluetooth svn: lower default pref May 18 09:48:18 03koen 07org.oe.dev * r7d83827b... 10/ (1 conf/machine/fic-gta01.conf): fic-gta01: add 'phone' machine feature May 18 09:50:51 03koen 07org.oe.dev * r36fc0f3a... 10/ (1 packages/angstrom/angstrom-x11-image.bb): angstrom-x11-image: install openmoko-dialer for devices with 'phone' machine feature May 18 09:51:43 hello everyone, a new day and new problems on my way to run a complete bitbake cycle :) May 18 09:52:06 package gqview-2.1.1-r1: task do_compile: failed May 18 10:45:44 does anyone know why an ohci_readl function might return -1? May 18 10:46:04 I'm assuming because it's unable to read the appropriate register May 18 11:03:31 03koen 07org.oe.dev * r978fb790... 10/ (1 packages/linux/linux-gta01/defconfig-2.6.21.1-fic-gta01): linux-gta01: update defconfig for 2.6.21.1 May 18 11:18:34 does anyone know how to get sound working on hx2000? May 18 11:35:26 03koen 07org.oe.dev * r6feeea12... 10/ (3 files in 3 dirs): xserver-kdrive-common: merge n800 support from poky May 18 12:11:16 hello, i am fat-man and i weigh 400 pounds. you can see video of my supreme fatness at www.fat-man.org May 18 12:18:32 03koen 07org.oe.dev * rb5c064a1... 10/ (1 packages/xcalibrateext/xcalibrateext_git.bb): xcalibrateext: provice compatibility with old cvs version May 18 12:28:27 03koen 07org.oe.dev * r806fd6cb... 10/ (18 files in 3 dirs): xserver-kdrive-xomap: merge from poky May 18 12:51:36 03koen 07org.oe.dev * r8986b41c... 10/ (2 files in 2 dirs): linux-nokia800: update defconfig for nokia800 May 18 13:25:44 03koen 07org.oe.dev * r3a8f0e74... 10/ (2 files in 2 dirs): nokia 800: fix defconfig May 18 13:29:11 lrg_: ping May 18 13:33:27 um guyw May 18 13:33:29 guys May 18 13:33:45 WTF happened to #gpe and #opie?? May 18 13:33:51 CoreDump|afk: handhelds.org stole them May 18 13:33:58 CoreDump|afk: and trademarked them in the US May 18 13:34:12 I don't bloody believe it May 18 13:34:34 XorA: I was afraid that was the case when #gpe was shut down =( May 18 13:34:40 CoreDump|afk: if you see Mr France I know we will give you an alabi while you make use of a heavy handheld in a threatenig maner May 18 13:35:12 this behaviour is unacceptable May 18 13:35:33 wtf is this guy thinking? What makes him "own" GPE? May 18 13:35:53 hehe I'm banned from #handhelds-gpe May 18 13:36:11 how you manage that? May 18 13:36:17 I dunno May 18 13:36:30 tried to join the channels, said I'm banned May 18 13:36:41 Im not May 18 13:36:55 you must be in a littleblack book May 18 13:37:19 actually my relation to france and hh.org were on rather good terms May 18 13:37:37 ( until they began acting like idiots of course ) May 18 13:37:43 CoreDump|afk: you do host hentges/opie May 18 13:37:55 CoreDump|afk: that probably qualifies you as an enemy May 18 13:37:58 and hentge/gpe =) May 18 13:38:03 XorA: ahh lol May 18 13:38:21 lets see if I can join w/ another nick May 18 13:38:54 Im guessing as florian isnt logged in he is currently exploding May 18 13:39:27 yeah May 18 13:39:44 btw, my bot is banned as well, looks like they banned hentges.net =( May 18 13:40:34 Ill guess thats for hosting stuff then, someone should try from linuxtogo May 18 13:41:59 I'll fire a salvo into #gpe-handhelds, lets see what happens =) May 18 14:23:12 hi all. May 18 14:25:30 hey thesing May 18 14:27:17 koen: could you look at bugs 2334 and 2335 and push them? May 18 14:32:12 hi, all! May 18 14:32:21 hrw, ping May 18 14:43:49 koen: pong May 18 14:44:00 lrg_: I sent you a mail :) May 18 14:44:09 * lrg_ checks May 18 14:44:10 lrg_: about the travel to guadec May 18 14:44:44 koen, hi! May 18 14:44:49 03thommy 07org.oe.dev * re0433003... 10/ (8 files in 3 dirs): mono: update to 1.2.4 May 18 14:44:54 03koen 07org.oe.dev * r25448aa7... 10/ (1 site/mipsel-linux): mipsel-linux site: add some more entries (from #2234) May 18 14:45:04 * lrg_ is interrupted again by builders in house. bbl May 18 14:48:03 should TARGET_OS = "linux-uclibc" be enough in angstrom-2007.1 be enough to generate a uclib based image? May 18 14:48:12 no May 18 14:48:25 see http://www.angstrom-distribution.org/building-angstrom May 18 14:49:14 thanks.. I must admit I did not looks at that site yet May 18 14:49:55 ANGSTROM_MODE :P May 18 14:50:33 :D May 18 14:50:45 have fun at guadec! May 18 14:50:51 building for uclibc requires some more magic than setting target os May 18 14:51:04 koen, about commit access... May 18 14:51:09 keesj: guadec is some months away May 18 14:51:14 I know you where bussy with that last week(s) May 18 14:51:52 slapin: that's still under review by the core team May 18 14:52:34 cbrake: ping May 18 14:57:11 koen, thanks a lot! May 18 15:00:30 anyone here have experience with OE on a pxa270? May 18 15:00:48 03koen 07org.oe.dev * r047d0bb5... 10/ (3 files in 2 dirs): mono: fix require statements May 18 15:10:52 xuumbi: hello May 18 15:12:00 cbrake: hey Cliff! just wanted to ask you a quick question, have you pursued writing to the NOR flash (for example, the kernel) from within OE on the CM-x270? May 18 15:12:29 cbrake: compulab provides a normod.ko kernel module, but no source :( May 18 15:18:33 xuumbi: yes, just map MTD partitions May 18 15:18:39 xuumbi: and then you can " May 18 15:18:56 cat kernel.img > /dev/mtd1 May 18 15:19:10 cbrake: ah, excellent May 18 15:19:40 cbrake: thanks for the info May 18 15:19:46 xuumbi: I'm completely free of all the compulab flash stuff -- its all MTD :-) May 18 15:20:16 cbrake: I wonder if kernel boot parameters can be changed, I'm guessing not since ARMmon's config area can't be written to from outside ARMmon May 18 15:20:38 cbrake: yeah that's a good thing May 18 15:21:01 xuumbi: what I do is just compile the CMDLINE into the kernel. That way if I want to change it, I just update the kernel image. May 18 15:22:09 xuumbi: crude, but works May 18 15:23:07 cbrake: yeah that's what I'm thinking we'll need to do as well May 18 16:38:31 cbrake: do you have any intention of making a howto for your jffs2 flashing system for the cm-x270? May 18 16:39:02 xuumbi: yes, its been on my list as part of the 2.6.20 kernel upgrade. May 18 16:39:23 cbrake: oh cool May 18 16:39:59 xuumbi: I've implemented jffs2 on compulab for two customers so far -- everyone has different requirements, so its kind of difficult to extract something general. May 18 16:40:01 cbrake: I just tried to compile in mtd support to write the kernel to nor (following http://compulab-systems.com/web/linux/x270/x270_mtd.html), but I get a kernel panic May 18 16:40:45 cbrake: I'm assuming that the binary compulab nand flash driver conflicts with mtd somehow (we're still using the closed source nand driver) May 18 16:41:09 xuumbi: you can use mtd on NOR, and the compulab driver for NAND May 18 16:41:22 cbrake: so if that's the case, we'll definitely have to switch to jffs2 since being able to update the kernel is one of our requirements May 18 16:41:58 cbrake: really? I thought so, but that error made me second guess myself, I wonder why I got the kernel panic then... May 18 16:42:22 xuumbi: the directions for the above doc is for MTD on NAND, not NOR. May 18 16:42:26 cbrake: I'll try compiling as modules and inserting one by one instead of into the kernel, that should give me a better idea of the problem May 18 16:42:47 cbrake: but don't you see the "Enabling MTD support for NOR flash" section? May 18 16:42:53 xuumbi: for NOR, you'll want to set up a PHYS mapping, and use the cmdlineparts. May 18 16:43:06 xuumbi: this can all be done through the cmdline May 18 16:43:12 xuumbi: and kernel config. May 18 16:43:13 cbrake: I followed the "Enabling MTD support for NOR flash" section May 18 16:43:34 xuumbi: oops, I did not scroll down far enough :-\ May 18 16:44:26 xuumbi: yeah, that should work. Additionally you'll have to use the kernel cmdline to set up partitions on the NOR flash. May 18 16:45:18 xuumbi: something like: mtdparts=physmap-flash.0:256k(boot)ro,0x180000(kernel),-(root) May 18 16:45:38 xuumbi: that is for 2.6.20, the name is different for 2.6.16 May 18 16:45:56 xuumbi: SIGLUNCH -- be back in a bit May 18 16:46:10 cbrake: ah, I wasn't sure if that was needed, thanks! May 18 16:48:06 cbrake: it's great to know of another cm-x270 developer using OE, thanks for all the help May 18 16:55:05 hi all May 18 17:36:59 03koen 07org.oe.dev * rf97b052c... 10/ (8 files in 2 dirs): libx11: add a new -native and stop removing xlib.h May 18 17:38:33 zecke: ping May 18 17:39:18 pong May 18 17:42:14 pretty good RTT ::) May 18 17:42:32 it was just an accident May 18 17:42:52 how's the webkit stuff coming along? May 18 17:43:42 slowly. Let us get icu into OE May 18 17:44:01 the stuff that made alp work, and I slowly improve should be good for the feed-reader May 18 17:44:19 I can surf google but there is a lot of left todo May 18 17:46:49 hello everyone May 18 17:46:59 hey likewise May 18 17:48:22 is the idea about OE foundation that is a legal entity? May 18 17:49:41 zecke: rm -Rf packages/icu May 18 17:49:59 * likewise feels like working on upstart a bit May 18 17:50:00 is it working? awesome, then add WebKit ;) May 18 17:50:11 zecke: not yet, I pushed my WIP May 18 17:50:11 likewise: did koen send you the document? May 18 17:50:37 no (not that I know at least), but I'm not a core dev I guess. I noticed some discussion on this channel. May 18 17:50:47 Hello to everyone! Many thanks to developers of OE. That is what I have been searching for guite a long time. May 18 17:51:17 DeeKey: How did you find OE? May 18 17:52:09 likewise: well, this is nothing that needs to be kept secret May 18 17:52:12 03koen 07org.oe.dev * r13cfed81... 10/ (5 files in 3 dirs): icu: add icu-native and a broken icu May 18 17:52:44 zecke: the commit message says it all May 18 17:52:57 zecke: just curious. Having a legal entity is interesting when dealing with sponsorship, companies etc. May 18 17:53:07 koen: right, these lazy dutch guys... May 18 17:53:18 Hmm... I am keen in robotics. And was searching for a minimal size distribution.... Somehow I googled to OE... May 18 17:53:19 * likewise hides May 18 17:53:48 DeeKey: OK, nice. Let us know how it works for you (or not). May 18 17:57:35 likewise: biff May 18 17:58:43 hi all, I faced the next prob :) Is there any known reason why gnumeric compilation may fail ? I thought a "apt-get build-dep gnumeric" will help but until know it still fails with a enormous error log. May 18 18:02:20 after apt getting a bunch of libs and stuff I ended up apt-get install gnumeric the bitbake -c clean and rebuilding May 18 18:02:36 that fixed it May 18 18:02:53 * AsLan^ is using kubuntu by the way May 18 18:03:08 AsLan^: I don't like the "I burned a voodoo doll and it helped" aproach May 18 18:03:25 AsLan^: Identify build dependencies, fix the .bb file May 18 18:04:02 identify is good just need to sniff my 2.1 MB error-log ;) May 18 18:04:02 well, unfortuneatley I wasn't in a position to identify the dependencies etc. I just wanted gnumeric built May 18 18:04:45 likewise: http://openembedded.org/~zecke/foundation-considerations.odt May 18 18:05:06 likewise: it is incomplete, inaccurate and has lot of errors May 18 18:05:07 I should clarify, I wasn't able to identify the dependancy I needed or the package that contained it May 18 18:05:18 AsLan^: Since I'm using kubuntu as well it might be a kubuntu specific proplem May 18 18:05:53 I think so, before install gnumeric I apt-getted just about every gtk lib suffixed with dev I could find to no avail May 18 18:06:16 * zecke shakes his head... May 18 18:06:28 * AsLan^ shakes zeckes head :D May 18 18:07:11 seriously though, the error was something weird May 18 18:07:18 zecke: hmm, good points in there, is harder than it seems May 18 18:07:25 do you still have it up Zuse? May 18 18:07:51 yep, I try to find the key-sentence in the error log May 18 18:07:54 likewise: even when koen and I mention to get flithy rich, the Foundation shouldn't be the tool ;) May 18 18:08:24 zecke: no because that way its traceable :-) May 18 18:08:43 likewise: tracable to kaymen islands May 18 18:08:52 likewise , that is why i came to IRC chat :) I am trying to make a toolchain for rapid robot prototyping. It is aimed for students who are in computer control (MATLAB/Simulink and etc). May 18 18:08:52 I need to make a linux distro based on hard real time linux kernel + some specific libs for data acquisition board. The GNU part of GNU/Linux should be as small as possible. I hope that with OE I will be able to use my distro for various hardware targets. Am I right? May 18 18:09:22 DeeKey: yes May 18 18:09:34 * koen aaarghs at simulink May 18 18:09:43 DeeKey: yes(^2), and what hard real time kernel? May 18 18:09:48 RTAI May 18 18:10:24 DeeKey: I'm not sure if we tested our toolchain for RTAI, I think it's the same toolchain as vanilla Linux, unlike for example RTEMS. May 18 18:10:54 DeeKey: I am about to test the upcoming real-time patch for linux (2.6.21-rt2) May 18 18:11:03 yes , RTAI is a patch for vanilla May 18 18:11:12 DeeKey: not rtai May 18 18:11:17 DeeKey: what type of real-time reponse do you need? May 18 18:11:28 DeeKey: http://people.redhat.com/mingo/realtime-preempt/ May 18 18:11:41 hi cbrake May 18 18:11:46 likewise: hello May 18 18:12:24 cdbrake what do you mean by that? Type of response? maybe time of response? As always the less the better! :) May 18 18:12:48 no, I need hard real time. May 18 18:12:58 DeeKey: stock linux is getting pretty good if 10's of ms is good enough May 18 18:13:19 cbrake: rt2 should be us, not ms May 18 18:14:28 likewise: what is "rt2"? May 18 18:14:58 cbrake: sorry, I mean the upcoming real-time support for vanilla Linux. May 18 18:15:05 with RTAI you can get up to 1 ms in userspace and near 1 us in kernel space May 18 18:15:35 DeeKey: the -rt patch reaches ~10 us in kernel space, depending on h/w May 18 18:16:13 DeeKey: Do you have sample rates higher then 1MS ? May 18 18:16:27 DeeKey: especially in combination with real-time scheduled device driver tasks (the new user space device driver model) this is interesting also for user space. May 18 18:16:27 likewise: impressive May 18 18:16:36 in userspce? May 18 18:17:05 RTAI have own sheduler both for kernel space and userspace.. May 18 18:17:23 DeeKey: I have no numbers for userspace, but if the real-time scheduler is used, it should be in sub ms range May 18 18:17:27 Linux is runned as a non priority task May 18 18:17:41 DeeKey: I know the RTAI approach, I just don't like it May 18 18:18:00 DeeKey: I very much like the QNX approach, but that's a microkernel. May 18 18:18:58 The point is that using rtai you can use both linux apps nad hard real time stuff in one May 18 18:19:19 DeeKey: If you use on the other side matlab / simulink (on which OS), due you really need 1 us reactions from your system ? May 18 18:21:03 The question do yo need is not really good question... It is better to ask what we can get ;) May 18 18:21:54 DeeKey: Right, just as usual there is no limit if time ans money is not critical ;) May 18 18:21:57 As I said the quicker response the better! May 18 18:22:54 I have mentioned Matlab/Simulink because it is easy for students to teach the basics of control theory... May 18 18:23:20 that reminds me.... May 18 18:23:39 * koen kicks his lab buddy for the control engineering assignment May 18 18:24:32 DeeKey: good luck using OE. We hope it is "stable" enough for your systems :-) May 18 18:24:34 Yes, but if you run e.g. matlab / simulink on a Windows-based machine without RT- toolbox and get your data from a CAN-bus or even USB bus. Don't worry to much about real time of your system :D. The bus queues and windows will make your RT useless May 18 18:24:56 DeeKey: BIBO stable, that is. May 18 18:25:28 likewise: 'make oldconfig' isn't BIBO ;) May 18 18:26:33 koen: for the I or the O part? May 18 18:26:39 zecke: I'm compiling icu now May 18 18:26:43 likewise: O May 18 18:27:14 I wonder why stuff stopped working :} May 18 18:27:29 and the answer is the usual one, because I'm a moron May 18 18:27:43 So enough this University stuff :) Lets rock ! :) May 18 18:27:47 DeeKey: if you start to use OE you also like to swich to scilab and comedi May 18 18:27:58 zecke: if we can replace ../bin with ${STAGING_BINDIR_NATIVE} .... May 18 18:28:07 bah May 18 18:28:18 adding comedi to OE has been on my todo for 3+ years now May 18 18:28:19 Zuse !! Just to clarify: May 18 18:28:22 Deekey: saves your lab a few thousand euro for licences :D May 18 18:28:38 koen: I can add some proper "if cross compiling" constructs to icu May 18 18:29:34 I can use both MATLAB/SCILAB and Scilab/Scicos to create code which will be later executed in userspace under RTAI target machine. May 18 18:30:30 My aim is to make really FREE system, so that there will be no limitation in software part! Students will be able to use anythink they want! May 18 18:31:12 DeeKey: O.K. I guess I got it. I thought you like to simulate a system with the help of a RT-Linux and control it by Matlab. But you like to use Matlab to create code for your RT-system. Sorry wrong assumptions on my side :) May 18 18:31:14 OE should give them freedom in choosing hardware (target) part. May 18 18:31:55 Zuse, wait I will give you a link to a document with picture, you will get what I want in no time ;) May 18 18:32:40 zecke: source/data/Makefile:145 May 18 18:32:45 this is general discription:: https://www.rtai.org/index.php?module=pagemaster&PAGE_user_op=view_page&PAGE_id=8 May 18 18:33:24 ehm source/data/Makefile.in:145 May 18 18:33:32 and here is a document, if you are curious in it https://www.rtai.org/RTAILAB/RTAI-Lab-tutorial.pdf May 18 18:35:51 DeeKey: Thanks for the info. As I said i started from a wrong point of view. One of our labs is using CAN to control different sample systems with Matlab / Simulink. They have massive probs due to the use of Microsoft :D. Therefore I was asking May 18 18:36:21 I would like to say that there should be a special place in hell for .... May 18 18:36:33 vendors who do not provide printed documentation May 18 18:36:56 Crofton: which vendor would be there, with honours? May 18 18:37:01 TI May 18 18:37:01 Crofton: and hi there :-) May 18 18:37:04 hi May 18 18:37:26 I am off to print 350 pages May 18 18:37:28 Crofton: they have options for multiple places in hell. May 18 18:37:33 of DV docs :) May 18 18:38:06 I need to translate the DDR controller values from the values for the EVM to the Lyrtech board I have May 18 18:38:10 Crofton: I'm still a bit reluctant to spend time on DV, as it's only really useful when combined with the DSP part. May 18 18:38:17 anyway, time to baby sit the printer May 18 18:38:23 :-) May 18 18:38:39 I have a DV board that has a FPGA coupled with it May 18 18:38:49 We are very interested in the overall system May 18 18:38:50 Crofton: Make your life easier and just send the very few vendors which give a printed documentation to heaven. Much less work ;) May 18 18:41:01 Lets return back to OE. May 18 18:41:02 What I have: May 18 18:41:02 Hardware: old P2 diskless PC with an PCI input output card (one of NI). I boot the kernel via ethernet by PXE. The file system is on NFS folder May 18 18:41:02 I have compiled hard real time linux kernel (vanilla + RTAI patch) May 18 18:41:02 I use statically compiled busybox for a GNU part. May 18 18:41:03 No I want to use OE to compile my distro. I want May 18 18:41:05 Busybox, dropbear and comedy libs. May 18 18:42:02 Zuse if you are interested in the subject U are welcome :) May 18 18:43:39 DeeKey: I'm always interested but time is limited :(. As far as I understood koen comedy will be not available for OE yet May 18 18:43:43 DeeKey : If you use the PII to build OE i have only one suggestion.... May 18 18:43:52 DeeKey : make a lot of coffee May 18 18:44:43 NO!!! That is the point why I want to use OE! I will compile everything on the host PC nad just put it to NFS folder! May 18 18:44:57 that is how everything is working right now! May 18 18:45:50 that works. i boot most of my boards over ethernet also... May 18 18:48:02 But right now I am totally new to OE. And just dont know where to start! I have stuck in the beginning... :( May 18 18:48:02 So I am building for x86. It should be easy.. Ok we can skip comedy for now. I will compile it myself. Actually it is easy. May 18 18:49:13 But how to compile dropbear for instance? May 18 18:49:56 first make sure you set you local.conf correctly May 18 18:50:01 zecke: found this in the log for gnumeric-error. /bin/sh: /usr/local/OE/oetmp/staging/arm-angstrom-linux-gnueabi/lib/../local/OE/oetmp/staging/i686-linux/bin/orbit-idl-2: No such file or directory May 18 18:50:01 OE repository does have dropbear. But how to compile it for my machine May 18 18:50:25 zecke: http://rafb.net/p/GQYAIO33.html May 18 18:50:43 DeeKey : bitbake dropbear May 18 18:50:55 DeeKey: do you have already a working OE environment ? May 18 18:51:06 koen: hehe May 18 18:51:18 zecke: it works :) May 18 18:52:23 okay... whom to argue with May 18 18:52:34 ifaistos: first make sure you set you local.conf correctly I think here is my problem... May 18 18:53:07 DeeKey there is a how-to did you get this ? May 18 18:53:10 DeeKey: You followed the Documentation / Wiki / Getting Started May 18 18:53:23 * likewise question mark May 18 18:54:12 likewise yes, but smthing went wrong... So wait I will summarize.. May 18 18:54:47 DeeKey: Maybe you miss to remove the last line (I did ;) ) May 18 18:57:20 So here is my current situation: May 18 18:57:22 bitbake dropbear May 18 18:57:22 NOTE: Handling BitBake files: | (0001/0001) [100 %] May 18 18:57:22 NOTE: Parsing finished. 1 cached, 0 parsed, 0 skipped, 0 masked. May 18 18:57:22 NOTE: build 200705182156: started May 18 18:57:22 OE Build Configuration: May 18 18:57:22 BB_VERSION = "1.8.3" May 18 18:57:24 OE_REVISION = "e70ceff9620fd2551611e5c4fdaef7c7ddbde010" May 18 18:57:26 TARGET_ARCH = "i486" May 18 18:57:29 TARGET_OS = "linux" May 18 18:57:30 MACHINE = "x86" May 18 18:57:32 DISTRO = "generic" May 18 18:57:35 DISTRO_VERSION = ".dev-snapshot-20070518" May 18 18:57:36 TARGET_FPU = "" May 18 18:57:39 NOTE: Resolving missing task queue dependencies May 18 18:57:40 ERROR: No buildable providers available for required build target dropbear May 18 18:57:50 DeeKey: one BBFILE? May 18 18:58:00 wooooo! http://www.burtonini.com/blog/computers/doomed-2007-05-18-19-30 May 18 18:58:03 ~pastebin May 18 18:58:05 methinks pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste, or http://rafb.net/paste/, or http://pastebin.com is usually painfully too slow and unresponsive to use, use one of the other pastebin sites, or dpaste.com is a very nice pastebin as well May 18 18:58:20 DeeKey: please use a pastebin for longer pastes May 18 18:58:28 DeeKey: you might need to set BBPATH May 18 18:58:31 sorry. I will... May 18 18:58:36 BBPATH is set May 18 18:59:05 export BBPATH=/media/sda9/oe/build/:/media/sda9/oe/org.openembedded.dev/ May 18 18:59:16 hmmm May 18 18:59:20 DISTRO="generic" May 18 18:59:24 does that still work? May 18 19:00:11 not sure but it should get the toolchain to start building May 18 19:00:41 NOTE: Applying patch 'fixchicken.patch' May 18 19:00:45 :-) May 18 19:01:34 DeeKey: also you should set BBFILES to more than just dropbear May 18 19:01:55 DeeKey: also -vvv and -DDD can help to get a slight idea May 18 19:02:40 koen: now, how to package 300MB of sources? May 18 19:03:07 likewise : Do chickens need fixing or fixing is for the chicken ? :) May 18 19:03:24 ifaistos: hehe, you don't know the chicken language? May 18 19:05:16 zecke: I am in a town, no chickens around here to here their language... Dead ones i find the supermarket are too cold to talk :) May 18 19:06:23 ifaistos: I think it needs egg.patch :-) May 18 19:07:01 zecke: Im tidying up the recipe and will push the changes May 18 19:07:30 hmmm..... Some problems with pastebin... Is is really pastebin's problems or my? Query failure: Can't open file: 'recent.MYI'. (errno: 145) May 18 19:07:31 lock tables recent write May 18 19:07:33 koen: I might make a smaller tarball available... May 18 19:07:47 koen: it will be fun to upload the code to google May 18 19:08:02 did they change the rules? May 18 19:08:08 last year they only needed a link iirc May 18 19:08:22 btw i am thinking of upgrading my workstation (again) to a dual or quad core. any recommendations withing a reasonable budget <=2000 Euro May 18 19:08:23 I need to reread the stuff May 18 19:08:26 mtn: misuse: base revision 7c47b3d7781fa237457801831ad1710394e5e0da does not exist in database May 18 19:10:19 5.5M ../../ipk/armv6/icu_3.6-r0_armv6.ipk May 18 19:10:41 Ifaistos: look at the apply mac pro's, they are cheaper than dells May 18 19:10:42 koen: okay, I can split the package a bit more May 18 19:11:09 ifaistos: dell precision 390 w/ quad? May 18 19:11:31 zecke: NOTE: Multiple libraries (libicudata.so.36, libicuuc.so.36, libicui18n.so.36, libicule.so.36, libiculx.so.36, libicutu.so.36, libicuio.so.36) found and LEAD_SONAME not defined May 18 19:11:46 ifaistos: scrap that. Mac pro :-) May 18 19:11:47 /space/dports/lib/libicui18n.dylib.34 (compatibility version 34.0.0, current version 34.1.0) May 18 19:11:50 /space/dports/lib/libicuuc.dylib.34 (compatibility version 34.0.0, current version 34.1.0) May 18 19:11:53 /space/dports/lib/libicudata.dylib.34 (compatibility version 34.0.0, current version 34.1.0) May 18 19:13:03 koen: WebKit is currently a monster, but it can be easily slimmed down May 18 19:13:28 zecke: commit pushed, have fun with FILES_ :) May 18 19:13:42 03koen 07org.oe.dev * rb327d3b1... 10/ (1 packages/icu/icu_3.6.bb): icu: add some sed to remove non-crosscomile friendly invocations May 18 19:14:06 koen: will do, I'm currently playing with the rss reader again May 18 19:14:16 koen: but I need to study tomorrow... and on sunday May 18 19:20:05 interesting.... 4 channel 512 taps FIR filter in a 32 pin lqfp package -> http://www.quickfiltertech.com/html/qf4a512-psc.php May 18 19:21:16 zecke: http://rafb.net/p/V6yyJc14.html May 18 19:21:50 4.2MB of unicode data, I will start to cry May 18 19:22:43 ifaistos: AC accuracy does not look professional though. May 18 19:22:57 ~-89dB at best May 18 19:24:03 likewise : well i was looking at the a fir filter part... not so much at the analog May 18 19:24:08 ifaistos, are you the buildbot expert? May 18 19:24:52 * likewise is listening in :-) May 18 19:25:04 03koen 07org.oe.dev * r7a50fb84... 10/ (1 packages/icu/icu_3.6.bb): (log message trimmed) May 18 19:25:04 icu: more granular packaging: May 18 19:25:04 140K armv6/icu_3.6-r0_armv6.ipk May 18 19:25:04 36K armv6/icu-dbg_3.6-r0_armv6.ipk May 18 19:25:04 548K armv6/icu-dev_3.6-r0_armv6.ipk May 18 19:25:04 16K armv6/icu-doc_3.6-r0_armv6.ipk May 18 19:25:06 4.2M armv6/libicudata36_3.6-r0_armv6.ipk May 18 19:25:13 Crofton : Since i got married i am not expert in anything... that's a wife's job :) May 18 19:25:21 of course May 18 19:25:28 zecke: there you go May 18 19:25:29 Crofton : just icecc May 18 19:25:39 koen: now slim icudata ;) May 18 19:25:40 who setup buildbot? May 18 19:25:44 used buildbot once May 18 19:25:46 hmm May 18 19:25:52 can I ask a question? May 18 19:25:55 Crofton: RP May 18 19:26:10 well, here is the questin May 18 19:26:19 yeah RP is running it for sometime now May 18 19:26:24 I have a library in svn I want to build May 18 19:26:29 then I need to install it May 18 19:26:42 so I can build a program from another part of svn May 18 19:26:50 can I do this with build bot? May 18 19:27:08 and set the lib as a dependency on the prgram May 18 19:27:15 OE can do that :) May 18 19:27:23 I know that:) May 18 19:27:41 but oe does not have nagging and svn monitor capability May 18 19:27:51 nor will it skip failed builds May 18 19:27:56 :) May 18 19:28:03 import buildbot; May 18 19:28:17 As far i understoof buildbot you can set different "events" to trigger a build slave May 18 19:28:25 yeah May 18 19:28:36 the trivky bit is installing the libs on the build host I think May 18 19:28:40 just don't ask me how to do it ;) May 18 19:29:29 adding buildbot like capability to oe would be "useful" May 18 19:30:21 *maybe* if you could modify the makefile to do an "make install" instead of a simple make ? May 18 19:30:23 use a tinderbox :) May 18 19:30:49 heh May 18 19:31:05 zecke, which tinderbox? May 18 19:31:09 (10:01:36 PM) zecke: DeeKey: also you should set BBFILES to more than just dropbear May 18 19:31:09 Thanks zecke ! It was the problem! Now I am advanced to the next level :) Running task 1 of 247 May 18 19:32:36 Crofton: like http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenEmbeddedBuilds May 18 19:33:02 Hmmm... Is it OK, that this operation take so much time? NOTE: Running task 1 of 247 (ID: 45, /media/sda9/oe/org.openembedded.dev/packages/shasum/shasum-native.bb, do_fetch) May 18 19:33:03 likewise : Have you seen anything in a single chip solution that can do echo cancellation on a "longish" tail (100-150 msec) May 18 19:33:08 I looked at tinderbox once, and found several versions, all with crappy instructions May 18 19:34:06 ifaistos: I know the blackfin guys do open-sourced echo cancellation on the blackfin dsp, not sure for what lengths though. Would a FPGA be allowed? May 18 19:34:27 DeeKey: it is stuck there? May 18 19:34:31 yes May 18 19:34:37 koen,likewise : Mac pro looks interesting... anyone using it ? May 18 19:34:51 ifaistos: not me, would wish though. May 18 19:34:53 I wish May 18 19:35:01 Ifaistos: iirc mickey|moving is going to buy one May 18 19:35:21 Ifaistos: buy your own ram, though May 18 19:37:42 likewise : no fpga/dsp looking for something like the octasic -> http://www.octasic.com/en/products/oct6100.php May 18 19:38:00 koen : why ? any problem with the apple stock ram ? May 18 19:38:15 DeeKey: no May 18 19:38:35 Ifaistos: it cost 3x as much as regular ram May 18 19:39:04 DeeKey: you need to call stty, let me find the mail May 18 19:39:21 what's the command to get the exchange rates on irc ? May 18 19:39:38 ~change 5 gbb to usd May 18 19:39:46 i got some error trying that May 18 19:39:47 ~change 5 gbp to usd May 18 19:39:53 i got some error trying that May 18 19:40:05 anyway, it uses xe.com May 18 19:40:09 DeeKey: stty -tostop May 18 19:40:22 O.K. back to the gnumeric prob. I found that the path to ORBIT_IDL seems wrong ORBIT_IDL = /usr/local/OE/oetmp/staging/arm-angstrom-linux-gnueabi/lib/../local/OE/oetmp/staging/i686-linux/bin/orbit-idl-2. just need to find were it was set. May 18 19:40:29 ~change 2500 usd to euro May 18 19:40:35 i got some error trying that May 18 19:40:45 bank is closed ..... May 18 19:41:32 Zuse: do you have a orbit-idl-2? May 18 19:42:19 well using the list price on apple's site for the MAC pro is around 1850 Euros... not bad May 18 19:43:20 zecke: yes i have (i guess ;) ). The file is at ../oetmp/staging/i686-linux/bin/orbit-idl-2. But look the path cannot be right, i guess there are to variables concatenated wrong May 18 19:43:22 TNKS zecke :) No everything works! GREAT! Hope to get the output.... May 18 19:44:12 Zuse: now find the autoconf check for ORBIT_IDL May 18 19:44:23 ibot says: "I got some error trying that" -- now *there's* a useful error message! May 18 19:44:37 zecke: Thanks for guiding me May 18 19:45:05 * koen points to the reference local.conf at http://www.angstrom-distribution.org/building-angstrom May 18 19:47:01 * chouimat|away kicks perl-native for not compiling May 18 19:47:19 chouimat|away: welcome to dash May 18 19:47:31 zecke: hmm? May 18 19:47:43 chouimat: ls /bin/sh is dash? May 18 19:48:02 no bash May 18 19:50:03 did we update libiconv lately? May 18 19:50:41 I will continue tomorrow. So close to finish ... but Wife is complaning already ;) Bye everone. May 18 19:50:51 buy! May 18 19:51:12 zecke: no, but there where some bogus csets adding DEPENDS = "libiconv" May 18 19:53:26 koen: should we disapprove them? May 18 19:54:11 zecke: I should have fixed them, see my mails to the oe-devel list May 18 19:56:23 * zecke is tempted to hack the gcc frontend to remove all -rpath's May 18 19:56:52 zecke: as a workaround? May 18 19:57:07 there is not reason for them ;) May 18 19:57:35 llvm 2.0 is nearing a release May 18 19:57:40 zecke: I know, at least not within the way we use them. May 18 19:57:52 03koen 07org.oe.dev * r3eeaba61... 10/ (1 packages/xorg-lib/libx11-native_1.1.1.bb): libx11-native_1.1.1.bb: remove bogus file May 18 19:57:59 zecke: although I wonder if our host gcc-cross needs a rpath for its libraries May 18 19:58:16 zecke: as it shouldn't use the host's glibc, should it? May 18 19:58:30 likewise: -rpath-link is sane May 18 19:58:41 zecke: I know. May 18 19:58:42 likewise: setting LD_LIBRARY_PATH is sane for native packages May 18 20:06:53 hmmm.... after quite a long time after executing #bitbake dropbear I have received following errors: http://pastebin.ca/495511 What next? Rerun it again? May 18 20:07:20 DeeKey: DISTRO="angstrom-2007.1" May 18 20:08:23 koen: dont get you :( May 18 20:09:02 don't use DISTRO='generic' May 18 20:09:38 DeeKey : Well it could not find a mirror to download the source file May 18 20:11:28 ifaistos: but how to fix this problem (just curious) ? Need to find and fix bogus .bb receipts? May 18 20:16:19 '''[[[[[[5'r May 18 20:16:33 * likewise removes kitten again... May 18 20:16:56 * likewise closes 100 print screen requesters... May 18 20:22:00 mtn: this is almost certainly a bug in monotone. May 18 20:33:16 hi all May 18 20:35:46 thesing: hi, only for your information - touchscreen-bug (on simpad) after suspend/resume solved now May 18 20:36:20 mrdata: cool. what did cause it? May 18 20:36:32 koen: I have libtool! May 18 20:36:37 koen: and I hate it May 18 20:37:07 thesing: i found out, that you wakup the ktsd-thread 10 times per second May 18 20:38:22 thesing: for original code it should only wakeup after MAX_SCHEDULE_TIMEOUT or interrupt while touchscreen pressed May 18 20:38:56 mrdata: not me. I never looked into the ts-stuff. ;) May 18 20:39:54 thesing: the original code works only befor suspend, after suspend no interrupt was occured after ts pressed May 18 20:40:25 mrdata: did you find out why? May 18 20:40:55 waking up at 10Hz is going to kill battery life May 18 20:41:04 thesing: not realy May 18 20:41:17 mrdata: maybe the ucb must be reinitialized after suspend. that would be a better fix. May 18 20:41:56 thesing: yes, i use a self repair May 18 20:42:34 thesing: the thread must restart only one time alone (short schedule time) May 18 20:42:59 thesing: then all works like a charme May 18 20:43:49 mrdata: could you mail me your ucb1x00-ts.c ? (if thats the file you patched) May 18 20:43:54 thesing: two extra lines of code inside the ktsd-thread and all works well May 18 20:44:30 thesing: yes, the ucb1x00-ts.c May 18 20:44:51 koen: could you look again at 2334? I changed the machine file. May 18 20:46:11 the kernel patch as well? May 18 20:47:59 yes I changed that too. May 18 20:51:36 thesing: mail send ;-) May 18 21:06:03 cbrake: ping May 18 21:06:06 03thommy 07org.oe.dev * r18e7d83f... 10/ (1 conf/machine/wgt634u.conf): wgt634u: add machine, part of #2334 May 18 21:06:13 03koen 07org.oe.dev * r9901b7d8... 10/ (1 packages/xorg-lib/libx11-native_X11R7.1-1.0.1.bb): libx11-native: somehow the makekeys patch got lost, readded it May 18 21:06:14 cbrake: I just came across this: http://tree.celinuxforum.org/CelfPubWiki/ELC2007Presentations?action=AttachFile&do=get&target=preempt070418celfelc.pdf May 18 21:06:54 another presentation I didn't watch May 18 21:07:28 koen: last page is interesting. It (again) shows real-time performance of the -rt patch May 18 21:08:06 03likewise 07org.oe.dev * rbdb1d549... 10/ (4 files in 3 dirs): linux-efika: Add kernel from GIT tree. Supports deep sleep. May 18 21:08:57 likewise: page 26? May 18 21:09:29 koen: yes, not quite the last.. May 18 21:10:07 koen: a bit chaotic graph, but it shows the idea. Better is to keep track of the worst latency measured, but hey... May 18 21:21:21 quick question about building python-imaging in open embedded without TIFF, JPEG and TRUETYPE support (becuase they bloat my image too much) May 18 21:21:29 Is there an easy way to configure the PIL to not build the JPE,TIFF, TT components. May 18 21:21:47 hey mgross May 18 21:22:17 Hay! May 18 21:22:26 It was good seeing you at ELC May 18 21:22:33 likewise May 18 21:22:49 I'm trying to finish my robot magellen entry. and I need to do some image processing based on PIL May 18 21:23:20 packages/python/python-imaging_1.1.5.bb May 18 21:23:48 It fills up my slug with TIFF, JPEG and true type cruft. :( May 18 21:23:55 hmmm, distutils.... May 18 21:24:40 I'm toing abuild righ now where I took the DEPENDS = "freetype jpeg tiff" May 18 21:24:40 out May 18 21:25:03 distutils? May 18 21:25:22 koen: is still somesthing wrong with the kernel.diff in #2334? May 18 21:25:29 http://selenic.com/repo/pagemap/raw-file/tip/memory-profiling.html#(1) May 18 21:25:33 thesing: haven't looked at it May 18 21:25:51 Who did you memory usage talk at FOSDEM? May 18 21:25:53 mgross: you'd have to ask mickey|moving, best bet is to ask in the mailinglist and CC him May 18 21:26:18 the oe mailing list or the nslug list? May 18 21:26:26 mgross: oe-devel May 18 21:30:07 Crofton: kernel development is moving faster than ever, if you ask me. May 18 21:30:16 yeah May 18 21:30:20 interesting stuff May 18 21:31:03 all areas of the kernel are worked on in innovative ways May 18 21:34:59 Received the following error: http://pastebin.ca/495646 I think due to my modification of x86.conf. I put OLDEST_KERNEL= "2.6.19". How to accustom OE for using 2.6.19 kernel with a custom patch and .config ? May 18 21:37:46 drat May 18 21:38:03 I tricked bitbake into forgetting all PREFERRED_PROVIDERS during the build May 18 21:38:27 I don't want to debug that now May 18 21:38:29 'night all May 18 21:38:49 I'll go too. CU May 18 21:39:49 is there a way to get a new password for the bugtracker? May 18 21:42:08 dcordes: You can have bugzilla send you your old one. May 18 21:43:33 night all May 18 21:44:19 Crofton: is it "system architect" or "systems architect" in the US? May 18 21:45:11 likewise, I didn't do well in English :) May 18 21:45:55 Crofton: :-) May 18 21:46:39 Does anyone know anything about setting up DDR controllers? May 18 21:50:47 Crofton: hmm, so far, I have setup SDR controllers, DDR is a bit different (other than the double clock) May 18 21:51:30 Crofton: what's the problem you are facing? May 18 21:51:35 I wish they would just say, for this part# set the device clock to this number! May 18 21:55:09 I'm about to crash. May 18 21:55:15 good nite all May 18 21:55:20 good night May 18 21:56:20 hi May 18 21:56:28 CoreDump|home: ping May 18 22:25:56 03Laibsch 07org.oe.dev * r4a45071b... 10/ (5 files in 2 dirs): gdb-cross: unify May 18 22:26:00 03Laibsch 07org.oe.dev * r4df275d1... 10/ (5 files in 2 dirs): gdb: unify May 18 22:26:36 * Bernardo is away: Ausente por agora. May 18 23:37:50 Hello May 18 23:38:35 I'd like to ask a question. May 18 23:38:44 Why does finch depend on pidgin? May 18 23:40:40 B_Lizzard: I guess your best bet to find out is to check with "mtn log --diffs" to find out who made the commit for that depends May 18 23:40:48 that DEPENDS May 18 23:41:34 Actually, finch and pidgin (and libpurple) are built from the same bb file May 18 23:42:06 libpurple is the backend, finch is the text-based frontend, and pidgin is the GTK2-based frontend May 18 23:43:21 I just want the text-based frontend, but pidgin gets included **** ENDING LOGGING AT Sat May 19 02:59:56 2007