**** BEGIN LOGGING AT Wed Apr 04 02:59:57 2007 Apr 04 05:03:55 03lenehan 07org.oe.dev * r7584fa59... 10/ (1 packages/perl/perl.inc): Apr 04 05:03:55 perl: Fix perl build for sh3/sh4. The recent change to support powerpc was Apr 04 05:03:55 wiping out the options that were set for sh3/sh4 - we only need to init the Apr 04 05:03:55 OPTIONS once. Apr 04 05:42:54 Hi everyone Apr 04 05:44:17 I want to ask which one should I use in bb and bbclass, "TAB" or "SPACE" ? Apr 04 05:45:59 I have been asked to submit bb file to bugzilla that use "SPACE" instead of TAB, but there are alot bbclasses contain "TAB" Apr 04 05:46:36 sometimes it results in parsing failure Apr 04 05:46:38 too young to build OE Apr 04 05:54:30 O_Neil: http://www.openembedded.org/wiki/StyleGuide Apr 04 06:03:30 thanks mwester Apr 04 06:04:34 good morning all Apr 04 06:13:48 hi koen Apr 04 07:12:26 morning all Apr 04 07:15:38 morning Apr 04 07:25:19 03rwhitby 07org.oe.dev * r364a5bb4... 10/ (1 packages/meta/slugos-packages.bb): slugos-packages: added task-native-sdk Apr 04 08:05:58 morning Apr 04 08:36:30 morning all Apr 04 08:36:44 morning RP Apr 04 08:56:22 hey RP Apr 04 08:56:53 !oebug 2080 Apr 04 08:56:55 * * Bug 2080, Status: NEW, Created: 2007-04-04 00:04 Apr 04 08:56:56 * * koen(AT)dominion.kabel.utwente.nl: Bitbake trunk (1.9) doesn't remember errors Apr 04 08:56:56 * * http://bugs.openembedded.org/show_bug.cgi?id=2080 Apr 04 08:59:10 morning all Apr 04 08:59:19 hey do13_ Apr 04 08:59:28 hi koen|lab Apr 04 08:59:46 * koen|lab waits for the simulation to finish Apr 04 08:59:58 koen|lab: what kind of lab? Apr 04 09:00:02 * koen|lab wonders why the sysadmin disabled the second cpu Apr 04 09:00:09 do13_: uni EE lab Apr 04 09:00:52 koen|lab: can't remember these times :) Apr 04 09:02:44 I walked around my old uni a couple of weeks ago (doing a recruitment presentation to the final year students). It felt almost the same as when I was there 20 years ago. Only some new buildings. Apr 04 09:08:34 hi Apr 04 09:15:04 koen|lab: That doesn't sound good :/ Apr 04 09:19:19 hi rp Apr 04 09:28:00 bm Apr 04 09:28:52 hi crofton Apr 04 09:33:25 A newby question: I want to install a packages, though with some different configuration paths when it is executing on my target device Apr 04 09:34:47 I gues it's best to create a seperate bb file that depends on the original package, but how do I then path the configuration of the target device? Apr 04 09:35:13 Just by overwriting the original installed init script from my newly made bb file? Apr 04 09:37:06 RP: the vernier thing looks cute, btw Apr 04 09:39:52 koen|lab: and developed with poky too :) Apr 04 09:40:02 yes :) Apr 04 09:41:28 it was my first OH work ;D Apr 04 09:44:34 NOTE: package util-linux-2.12r-r10: task do_compile: failed Apr 04 09:44:58 umount.c:45: error: expected declaration specifiers or '...' before 'umount2' Apr 04 09:45:20 wonder what's caused that to suddenly break ... Apr 04 09:52:58 last change to util-linux recipe was 12 dec 2006 - how can it be suddenly failing now (there was an independent report of this in #openmoko today) Apr 04 09:53:51 rwhitby: new kernel headers, probably. I had the same problem when upgrading to 2.6.18 - now where was that fix, I can't find it in util-linux/files... Apr 04 10:05:36 3 years of OE passed Apr 04 10:14:31 hi Apr 04 10:33:01 has anyone seen anything like that with bitbake ->http://rafb.net/p/mfPHc958.html Apr 04 10:34:59 hi all Apr 04 10:45:43 hi, all! Apr 04 10:46:02 hi all Apr 04 10:46:06 which target will allow me to build kernel zImage? Apr 04 10:47:17 virtual/kernel Apr 04 10:47:23 (depending on your machine Apr 04 10:47:24 ) Apr 04 10:50:47 hmmm it seems that if bitbake has a problem with the patch it leaves a thread working Apr 04 10:52:07 21903 pts/1 T 0:08 python /proj/oplinux-0.2/bitbake/bin/bitbake opsis-asterisk Apr 04 10:52:41 woglinde: hi, I have tryed the speed of a cf card myself now. what is "slow" for you? and how did you tested it? Apr 04 10:57:54 is non-working puzzles news, or was it fixed? Apr 04 10:58:50 koen|lab, thanks! Apr 04 11:32:24 Ifaistos: You probably need the interactive tasks patch I mentioned on the mailing list. I need to commit that (and some others) Apr 04 11:34:28 mr_nice he Apr 04 11:34:39 mr_nice the first ls was really slow Apr 04 11:34:47 becaus the cf had many files Apr 04 11:34:53 but now I am happy Apr 04 11:36:42 woglinde: :) Apr 04 11:38:18 woglinde: that's fine. I got xynth running on my pad :) http://opensimpad.org/index.php/SLACKpad#xynth Apr 04 11:41:30 mr_nice can you dig at the ipc error? Apr 04 11:42:13 it happens with syslog too besides opie Apr 04 11:42:59 !oebug 2081 Apr 04 11:43:00 * * Bug 2081, Status: NEW, Created: 2007-04-04 04:48 Apr 04 11:43:01 * * slapinid(AT)gmail.com: puzzles package fetch failure Apr 04 11:43:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2081 Apr 04 11:43:14 any ideas? Apr 04 11:44:57 slapin did yous earch with google where the new location is? Apr 04 11:45:07 slapin_nb: upgrade to available snapshot Apr 04 11:45:20 ah okay Apr 04 11:45:22 hi hrw Apr 04 11:48:10 mr_nice next thing I will do is to check if rp's lzo patches works on simpad too Apr 04 11:49:16 hrw, why not use http://www.chiark.greenend.org.uk/~sgtatham/puzzles/puzzles.tar.gz in package? Apr 04 11:49:28 03nail 07org.oe.dev * r3066a734... 10/ (6 files in 5 dirs): madfu: Add package madfu, to load firmware for m-audio devices Apr 04 11:49:31 slapin_nb: because we then have no or the wrong version information Apr 04 11:49:46 I just updated puzzles Apr 04 11:50:18 03pH5 07org.oe.dev * r89b237f2... 10/ (3 files in 2 dirs): puzzles: update to r7436 Apr 04 11:53:09 woglinde: I will try to build a opie image to see the error. lzo patch? sorry I really need to go to the launderette now else my girfriend will kill me ;) see you later Apr 04 11:53:23 mr_nice no, syslogd is enough Apr 04 11:53:30 pH5, probably SVN checkouts could be better solution here? Apr 04 11:53:54 mr_nice okay will see you next probaly I am on vaction than Apr 04 11:54:02 args Apr 04 11:54:35 slapin_nb: this is a very good idea. Apr 04 11:55:45 pH5, svn://ixion.tartarus.org/main/puzzles Apr 04 11:56:16 mtn: 3066a734f02184d60ed86216f8356de08970aefe nail@nslu2-linux.org 2007-04-04T11:48:40 Apr 04 11:56:16 mtn: 89b237f2bb444f1ebbbd18dfa8c97fadd289e6ca pH5@openembedded.org 2007-04-04T11:48:40 Apr 04 11:56:25 2 heads :( Apr 04 11:56:52 NAiL: I'll merge. Apr 04 11:58:13 I already did Apr 04 11:58:27 * pH5 Ctrl+C Apr 04 11:59:12 pH5, slapin_nb: I keep meaning to switch it to svn checkouts... Apr 04 11:59:27 slapin_nb: heads merged Apr 04 11:59:37 * NAiL just committed another change Apr 04 11:59:51 NAiL, thanks! Apr 04 11:59:56 03nail 07org.oe.dev * r5a224ad4... 10/ (1 packages/madfu/madfu_1.2.bb): madfu: Clean up so it matches oe style guide Apr 04 12:01:18 (oe has an automerger as well nowadays) Apr 04 12:10:13 mickey|bbl: Your removal of "outdated" ipkg versions broke the slugos builds. Apr 04 12:11:19 ipkg-native fails on install Apr 04 12:11:31 make[2]: Entering directory `/home/slug/slugosle/tmp/work/i686-linux/ipkg-native-0.99.163-r1/ipkg-0.99.163' Apr 04 12:11:35 make[2]: *** No rule to make target `install'. Stop. Apr 04 12:12:42 03florian 07org.oe.dev * r487b26a1... 10/ (1 conf/machine/netvista.conf): netvista.conf: We have a big screen. Apr 04 12:12:50 03florian 07org.oe.dev * r0dce271d... 10/ (3 files in 3 dirs): gtk+: Apply patch to 2.6.10 to allow background images in iconview styles. Apr 04 12:38:43 <_bambam> hi Apr 04 12:39:36 <_bambam> i'm having an issue on buildin openmoko-image on package libxcb which is asking for dep xsltproc, which still isn't present after build libxslt Apr 04 12:39:44 <_bambam> anyone has an idea? Apr 04 12:40:29 apt-get install xsltproc Apr 04 12:40:56 <_bambam> koen|lab: so i need it on the native side? Apr 04 12:42:45 ljp: ping Apr 04 12:46:34 Hi folks! Apr 04 12:47:00 FYI, I just committed fix for ipkg-utils-native which was discussed last week on the list. Apr 04 12:47:17 please ping me in case of any issue, or reopen: Apr 04 12:47:21 !oebug 2061 Apr 04 12:47:23 * * Bug 2061, Status: NEW, Created: 2007-03-29 05:42 Apr 04 12:47:24 * * pmiscml(AT)gmail.com: Fix broken version comparison in ipkg-make-index Apr 04 12:47:25 * * http://bugs.openembedded.org/show_bug.cgi?id=2061 Apr 04 12:47:38 !oebug 2061 Apr 04 12:47:39 * * Bug 2061, Status: RESOLVED (FIXED), Created: 2007-03-29 05:42 Apr 04 12:47:40 * * pmiscml(AT)gmail.com: Fix broken version comparison in ipkg-make-index Apr 04 12:47:41 * * http://bugs.openembedded.org/show_bug.cgi?id=2061 Apr 04 12:56:36 03pfalcon 07org.oe.dev * rb8bba4b9... 10/ (3 files in 3 dirs): Apr 04 12:56:36 ipkg-utils-native 1.6+cvs20050404: Fix version comparision algorithm, now Apr 04 12:56:36 it treats version as structured notion, and compares it component by component Apr 04 12:56:36 (epoch (was before), then upstream version, then package revision). Apr 04 12:56:36 * Closes #2061. Apr 04 12:56:40 03pH5 07org.oe.dev * r1854e2f3... 10/ (1 packages/puzzles/puzzles_svn.bb): puzzles: add svn version Apr 04 12:56:44 03cbrake 07org.oe.dev * r6cded7d2... 10/ (10 files in 3 dirs): compulab-pxa270: add 2.6.20 version of compulab cm-x270 kernel Apr 04 12:56:49 03cbrake 07org.oe.dev * r2c22bea4... 10/ (1 packages/linux/compulab-pxa270_2.6.16.bb): compulab-pxa270 2.6.16: remove image creation code as it does not work on all distros Apr 04 13:17:24 psokolovsky: very nice! Apr 04 13:17:49 florian: works for you? Apr 04 13:17:58 * florian just noticed that minimo builds with bitbake 1.6.6 but not with 1.8 Apr 04 13:18:30 psokolovsky: i didn't try yet... Apr 04 13:18:41 ok ;-) Apr 04 13:18:51 it should work, and w/o any issues of course ;-) Apr 04 13:19:03 * florian just return from lunch and needs to do some productive work Apr 04 14:08:40 * koen|lab reads http://labs.morpheuz.eng.br/blog/04/04/2007/prison-break-freedom-for-your-videos/ Apr 04 14:20:04 morning Apr 04 14:35:18 Anybody remembers was this actually done on 2.6.20: What: Video4Linux API 1 ioctls and video_decoder.h from Video devices. Apr 04 14:35:18 When: December 2006 Apr 04 14:35:40 I mean feature removal from Kernel. Apr 04 14:41:11 v4l2 is the way to go Apr 04 14:43:16 slapin_nb: kernel.org/git and look ;-) Apr 04 14:44:04 koen|lab, I know, just trying step by step approach. Apr 04 14:45:02 RP, I'm afraid to die of information intoxication, thanks! Apr 04 14:45:20 15 minutes till the deadline Apr 04 14:45:26 koen|lab: bttv is still v4l1 Apr 04 14:45:39 koen|lab: what deadline? Apr 04 14:45:44 hrw: I removed my bttv card last month Apr 04 14:45:52 robtaylor: the project that makes me be in this lab Apr 04 14:45:59 koen|lab: it is shit but works Apr 04 14:46:47 my pvr150 doesn't work to well in 2.6.20 Apr 04 14:47:04 it claims that it can't change channels, which it can do Apr 04 14:47:11 but mythtv trusts the error code.... Apr 04 14:55:24 cu Apr 04 14:58:01 there Apr 04 14:58:05 * koen|lab hands in report Apr 04 14:58:11 time to drink beer Apr 04 15:39:05 hey, random question, has anyone got any suggestions for a board with a dvb tuner and an ide interface? Apr 04 15:39:34 (and video out, of course) Apr 04 15:40:36 robtaylor, why you need video out? Apr 04 15:41:03 * slapin_nb uses large CRT monitor from DEC era to watch TV Apr 04 15:41:32 slapin_nb: cos i want to make a settopbox Apr 04 15:42:23 robtaylor, ah you need miniitx style then... Apr 04 15:42:47 kann mir jemand sagen wo ich nachlesen kann wie ich zusätzliche kernelmodule (pvrusb2) in mein image integrieren kann? Apr 04 15:43:15 slapin_nb: yeah, or pc104 or somesuch Apr 04 15:44:31 robtaylor, I tried once - they are too hot. I used good ole PC approach (custom case though). Apr 04 15:44:49 yeah, i ideally want to steer clear of x86 Apr 04 15:45:08 mono: what distro/kernel are you using? Apr 04 15:45:18 it doesnt need a hd though, so that should cut the heat a lot Apr 04 15:45:20 well sorry - english :) -- does someone know where I can get informations for adding kernel modules to my image? Apr 04 15:45:47 im using a dreambox Apr 04 15:46:05 and my toolchain runs on debian Apr 04 15:47:28 Hello Apr 04 15:48:03 yo sirfred Apr 04 15:48:13 mono: hmm, not too familiar with that one. You might want to take a look at the linux-rp .bb's in the .dev branch. They are set up to generate seperate .ipk's for each module Apr 04 15:48:19 Hey XorA Apr 04 15:48:39 XorA: A question vaguely related with idct, mplayer,... Apr 04 15:48:49 * XorA hides Apr 04 15:48:54 XorA: :) Apr 04 15:49:13 XorA: Do you know if there's any support for DSP instructions in ffmpeg, gcc ? Apr 04 15:49:37 ok - thanks hvontres|poodle :) Apr 04 15:49:45 sirfred: gcc no idea, ffmpeg, I think thats being worked on in SVN Apr 04 15:49:49 XorA: I don't mean iwmmx, but the dsp instructions in pxa25x for the 40 bits acumulator. Apr 04 15:50:12 sirfred: oh then not a clue Apr 04 15:50:19 I've googling for a while, but have found hardly nothing Apr 04 15:50:54 XorA: Yes, I've seen some things about iwmmx, but my husky hides a pxa255 Apr 04 15:51:31 XorA: ok. Thanks. Apr 04 15:52:08 that sounds like freaky ARM asm stuff, that I dont get chance to get down to that level Apr 04 15:52:35 XorA: Well, I wondered if gcc could have some support to use it automatically Apr 04 15:53:06 * XorA looks towards RP who seems to know silly amounts of random stuff Apr 04 15:53:22 XorA: jeje Apr 04 15:56:53 time for train Apr 04 16:29:33 mono: each kernel build in OE split kernel modules into separate packages. Apr 04 16:30:04 mono: so check dreambox feeds (or tmp/deploy/*ipk*/ after your build of kernel) for kernel-module-* packages Apr 04 16:33:33 thx hrw!!!! Apr 04 16:33:53 np Apr 04 16:43:38 how many of OE distros use l-l-h < 2.6.18? Apr 04 16:46:12 l-l-h ? Apr 04 16:46:51 linux-libc-headers Apr 04 16:47:23 Ah, thx. Unslung does, but I don't think anyone but I cares about Unslung anymore. Apr 04 16:48:32 PREFERRED_VERSION_linux-libc-headers ?= "2.6.15.99" Apr 04 16:48:55 that's from angstrom-2007.1.conf Apr 04 16:55:02 Unslung: PREFERRED_VERSION_linux-libc-headers ?= "2.6.11.1" Apr 04 16:55:29 (Newer than that broke things) Apr 04 16:55:58 Which might have something to do with the fact that Unslung uses a Linux 2.4.22 kernel (!) Apr 04 17:03:28 -lgcc_eh again.. Apr 04 17:07:05 03pfalcon 07org.oe.dev * rb02c390e... 10/ (4 files in 2 dirs): linux-handhelds-2.6: Add 2.6.20-hh6, drop 2.6.20-hh3. Apr 04 17:07:12 03pfalcon 07org.oe.dev * r2dedd269... 10/ (1 MAINTAINERS): Apr 04 17:07:12 MAINTAINERS: Looking after linux-handhelds-2.6 for more than half-year now, Apr 04 17:07:12 add to maintained recipes. Apr 04 17:14:16 pH5: what's this: Apr 04 17:14:17 -PREFERRED_VERSION_linux-handhelds-2.6 ?= "2.6.16-hh9" Apr 04 17:14:17 +#PREFERRED_VERSION_linux-handhelds-2.6 ?= "2.6.16-hh9" Apr 04 17:14:37 It's a called a 'comment' Apr 04 17:14:51 something that is usually a sign of not being used to a SCM ;) Apr 04 17:17:33 koen: well, for some reason default kernel version was killed. I assume that's stray change and going to undo it. Apr 04 17:19:53 * koen manages to get llvm-gcc4 past configure Apr 04 17:22:52 03pfalcon 07org.oe.dev * re48d2599... 10/ (3 files in 3 dirs): Apr 04 17:22:52 linux-handhelds-2.6: Update defconfigs for h3900 & h4000. Apr 04 17:22:52 * Defconfigman-generated static configs from HH.org CVS. Apr 04 17:22:56 03pfalcon 07org.oe.dev * r0199eb16... 10/ (1 conf/distro/angstrom-2007.1.conf): Apr 04 17:22:56 angstrom-2007.1.conf: Bump h3900, h4000, hx4700, magician kernel version to 2.6.20-hh6. Apr 04 17:22:56 * Fixes "TS driver spams syslog" issue. Apr 04 17:22:57 * Approved by Koen Kooi. Apr 04 17:23:00 03pfalcon 07org.oe.dev * r16807008... 10/ (1 conf/distro/angstrom-2007.1.conf): Apr 04 17:23:01 angstrom-2007.1.conf: Put back default linux-handhelds-2.6 kernel version to use. Apr 04 17:23:03 * It's a question if it should be still 2.6.16-hh9, but there should be default Apr 04 17:23:05 version for sure. Apr 04 17:31:14 XorA: pong Apr 04 17:31:49 2cu Apr 04 17:31:50 cu Apr 04 17:31:58 hi ljp Apr 04 17:32:12 hi Apr 04 17:38:10 03pfalcon 07org.oe.dev * r3d375d6f... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.20-hh6.bb): (log message trimmed) Apr 04 17:38:10 linux-handhelds-2.6 2.6.20-hh6: Fake commit to add proper changelog for -hh6. Apr 04 17:38:10 * Changes since -hh5: Apr 04 17:38:10 1. First set of defconfigman-generated configs was committed, for h3900, Apr 04 17:38:10 h4000, hx4700. These go to *_static_defconfig namespace. Together with them, Apr 04 17:38:12 reference bootlogs were imported, to fight regression stream. Apr 04 17:38:16 2. htchimalaya, htcblueangel, htcalpine: Revamps for the current kernel. Apr 04 17:41:35 bleh Apr 04 17:41:45 I'm re-learning gcc's buildsystem all over again Apr 04 17:41:55 it hasn't improved in the last 6 years Apr 04 17:45:27 koen why change something that "work" Apr 04 17:45:37 heh Apr 04 17:45:46 may break down and cry now? Apr 04 17:46:11 koen or they are like 99.5% of developers they don't understand autocrap and don't want to play with it ;) Apr 04 17:46:43 chouimat: I was amazed how well designed the llvm buildsystem is Apr 04 17:46:56 koen :) Apr 04 17:46:56 sadly llvm-gcc4 uses the gcc buildsystem Apr 04 17:48:12 I can suggest a solution Apr 04 17:48:33 pay me to go to the indt lab to work with their llvm guru :) Apr 04 17:48:39 koen I'm currently trying to make an update of a product here ... the buildsystem is an ugly mix of make shell script, perl and autotools ... and I'm not paid enough to want to rewrite it ... Apr 04 17:51:48 koen and I will stop using this thing soon so ... (last contractual think to do and I won't renew) Apr 04 17:53:17 heh Apr 04 17:54:28 koen the product is somewhat good ... but the owner of the code is an ass and won't let me fix the problems in it because he want to ditch it :) Apr 04 17:58:31 * joshin just figured out that CosmicPenguin is the person he needs to bribe to fix V4L support for the CS5536 so that he can move his PP3 from 2.6.12 to a more recent kernel and still have everything work. Apr 04 17:59:19 Wow - thats a ginormous bribe Apr 04 17:59:39 :) Apr 04 18:01:15 hi all Apr 04 18:01:16 I almost have everything working on my PP3 in 2.6.20. The last thing is the camera which requires V4L... Then I can ditch the Fedora 4 based OS they use and go with something else. Apr 04 18:01:22 joshin: basically, its going to be a while Apr 04 18:01:22 mr_nice: hi Apr 04 18:01:28 I figured. :) Apr 04 18:01:58 I was just amused when I realized that you were in #oe... Apr 04 18:02:05 we're slowly making our way down the list, X is in testing, the framebuffer will be hitting today or tommorrow Apr 04 18:02:28 then its a tie over which bear we'll try to wrestle first - V4L or UDC Apr 04 18:02:39 Yeah, I've been using an older build of the amd driver that seemed stable. Apr 04 18:03:41 I'm the only guy we've got, and I'm typing as fast as I can Apr 04 18:03:51 But I do take time out for lunch breaks... like now Apr 04 18:03:52 bbl Apr 04 18:03:59 Have a good one. Apr 04 18:04:31 mrdata: hi Apr 04 18:24:27 Hello Apr 04 18:24:53 hey sirfred Apr 04 18:25:07 koen: Hey, I've got the hx4700 Apr 04 18:25:18 ah, cool Apr 04 18:25:51 koen: Not that cool. Apr 04 18:26:28 koen: cursor issue is easy to fix, but kernel framebuffer driver seems to not handle properly mode changes. Apr 04 18:26:42 koen: Trying to set it to 640x480 using framebuffer doesn't work well. Apr 04 18:27:45 Cursor issue seems easy to fix, it seems that w3220 cursor is 32x32 instead of 16x16 Apr 04 18:31:22 hi Apr 04 18:32:36 sirfred: so "don't rotate" and "make cursor bigger" would get w3220 working X? Apr 04 18:33:09 koen: I don't thing "don't rotate" could be an option, considering that Xfbdev is able to handle rotation. Apr 04 18:33:19 sirfred: did you get the links I e-mailed you? Apr 04 18:33:35 hvontres|poodle: I think I didn't Apr 04 18:33:57 psokolovsky: I've been unable to build working opie images in dev/angstrom, I always get a kernel panic (can't find init). I'm going to try to build a x11 image for my akita now, to see if I'm at least one of those will boot. Apr 04 18:34:15 hvontres|poodle: What were the links about? Apr 04 18:34:37 koen: I'm going to try to correct the issues these days. Apr 04 18:35:01 koen: Cursor problem can be corrected, I think, just being aware of the different size. Apr 04 18:35:20 sirfred: just some stuff from the intel site about using the DSP instructions on pxa25x Apr 04 18:35:32 hvontres|poodle: Humm, no. Apr 04 18:36:15 sirfred: here is the first one: http://isnwiki.jot.com/WikiHome/Articles/168863 Apr 04 18:36:40 sirfred: and the second one is for the pxa assembler manual: http://www.intel.com/software/products/compilers/embedded/toolsuite/docs/asxscale.pdf Apr 04 18:36:56 hvontres|poodle: Thanks. Apr 04 18:37:15 hvontres|poodle: It seems it have to be done using assembler, no gcc automatic usage of that features. Apr 04 18:37:23 sirfred: np. you had just left when I found them Apr 04 18:37:37 sirfred: unless you want to try to add it :) Apr 04 18:37:46 hvontres|poodle: I don't know if the idct transformations could benefit of that 40 bit register Apr 04 18:38:05 hvontres|poodle: Perhaps the improvement doesn't pay the effort. Apr 04 18:39:12 sirfred: i would hazard a guess that IPP use some of those instructions... Apr 04 18:39:27 hvontres|poodle: It should. Can IPP be used with gcc ? Apr 04 18:40:37 sirfred: only 3.3.2 on 2.4 kernels I think...:( Apr 04 18:40:55 Bernardo: well, opie has little to do with init booting. kernel boots init, and that's where to start looking. Apr 04 18:41:31 03pfalcon 07org.oe.dev * r7d7fa329... 10/ (1 packages/linux/linux-handhelds-2.6/h5000): Apr 04 18:41:31 linux-handhelds-2.6: Add defconfig for h5000. Apr 04 18:41:31 * Defconfigman-generated, from kernel CVS. Apr 04 18:41:38 03pfalcon 07org.oe.dev * rea691c5d... 10/ (1 conf/distro/angstrom-2007.1.conf): Apr 04 18:41:38 angstrom-2007.1.conf: Bump h5000 kernel version to 2.6.20-hh6. Apr 04 18:41:38 * Approved by Koen Kooi Apr 04 18:41:42 Bernardo: you might want to try a console image first... less stuff to go wrong :) Apr 04 18:42:09 Bernardo: plus it builds faster :) Apr 04 18:44:10 sirfred: try asking ssvb about MUL and DIV Apr 04 18:45:15 koen: Aren't those regular arm instructions ? Apr 04 18:45:52 sirfred: you might also want to look at this doc: http://cache-www.intel.com/cd/00/00/09/55/95527_95527.pdf Apr 04 18:45:57 iirc they are the 'e' in armv5te Apr 04 18:46:11 but I don't know very much about ARM isns Apr 04 18:46:17 "The thread starts with "[PATCH 0/90] Post 2.6.21 OMAP update" Apr 04 18:46:38 seems tony has been busy :) Apr 04 18:46:55 hvontres|poodle: thanks Apr 04 18:47:07 hvontres|poodle: Now, I only need time. ;-) Apr 04 18:47:26 sirfred: hehe...seems to be a common thread around here .. Apr 04 18:47:49 So, I only need to develop a time machine, and then, go back everytime I need. Apr 04 18:48:13 psokolovsky: how is pocketpc looking with the latest angstrom images? Apr 04 18:48:23 * sirfred bumps up priority for the time machine to 1 Apr 04 18:49:26 koen: err, I'd need yet to try latest ones ;-). but I tested that X11R7.2 runs and seems to work well. Apr 04 18:50:02 psokolovsky: including keylaunch? I didn't get that to work yet. Apr 04 18:50:23 hi * Apr 04 18:50:31 pH5: nope, keylaunch didn't work, and I tested last time when even bl issues weer present Apr 04 18:50:51 I mean it did start and run ;-) Apr 04 18:50:52 i try to make rootfs for my device (AT91rm9200) w/ openembedded Apr 04 18:51:37 03philippe 07org.oe.dev * rcd87faaf... 10/ (5 files in 3 dirs): (log message trimmed) Apr 04 18:51:37 packages/gtk-webcore/files/build_silence.patch Apr 04 18:51:37 * remove now obsolete build fix Apr 04 18:51:37 packages/gtk-webcore/osb-nrcit_svn.bb Apr 04 18:51:37 * updated svn recipe to add some more patches Apr 04 18:51:37 packages/gtk-webcore/files/NRCit_HTTP_auth.diff Apr 04 18:51:39 i boot from nfs but this kernel hang on "Freeing init memeory : 88K", it seem to be error when run init/sysvinti Apr 04 18:51:39 * new patch that fix HTTP AUTH issues Apr 04 18:52:19 can i have some suggestion ? Apr 04 18:53:24 hvontres|poodle: ok, it's a good idea Apr 04 18:53:38 might be something in my build env that is borked Apr 04 18:53:53 it wouldn't be the first time Apr 04 18:54:01 (dash vs bash, etc.= Apr 04 18:54:02 ) Apr 04 18:55:18 sirfred: and,just to add some more confusion, here is some info from the evil empire: http://msdn2.microsoft.com/en-us/library/aa448595.aspx Apr 04 18:55:26 hvontres|poodle: The functions? Apr 04 18:55:54 hvontres|poodle: I found them this morning, it seems that the microsoft compiler have some builtin functions using the 40 bit acumulator Apr 04 18:56:02 sirfred: i think so. it's the documetaton for the funcitons in armintr.h Apr 04 18:56:42 hvontres|poodle: Yes, they are. Apr 04 19:01:27 sirfred: sirfred : according to this http://shareit.intel.com/WikiHome/Articles/168675 they are just intrinsic functions that map directly to the assembly because the MS compiler can't hadle inlines :) Apr 04 19:01:53 hvontres|poodle: Interesting.:) Apr 04 19:03:48 sirfred: and the ms docs give you the arm assembly equivalents :) Apr 04 19:04:43 Bernardo: heh. Well, I use that trick because of my *slow* build machine :) Apr 04 19:23:53 hi koen, after an repository update the file started to broke: 16:10:26 (152.73 KB/s) - `/indt/mamona/openembedded/sources/file-4.20.tar.gz' saved [548393 NOTE: Task failed: Fetch failed: ftp://ftp.astron.com/pub/file/file-4.20.tar.gz Apr 04 19:23:57 how can I fix it? Apr 04 19:26:53 * sirfred starts to hate pocketpc calibrate and copy/paste tutorial Apr 04 19:27:41 sirfred: what, you mean you already *KNOW* how to copy/paste ;) Apr 04 19:28:29 hvontres|poodle: I would kill for that option. ;-) Apr 04 19:29:03 * hvontres|poodle thankfully knows nothing about pocketpc Apr 04 19:29:35 sirfred: install wm2005 if you're not afraid that the owner will hate you. Apr 04 19:29:52 sirfred: I love the skip tutorial on my palm :) Apr 04 19:30:00 pH5: He's a good friend. I don't want to touch his machine Apr 04 19:30:19 pH5: I want him to keep being a good friend. ;-) Apr 04 19:30:21 ah, then you'll have to suffer ;) Apr 04 19:31:01 Hello ppl! Apr 04 19:31:06 At least I've learn that w3220 doesn't like ENG_CNTL to be touched, at least in the way we used to touch it Apr 04 19:36:09 pH5: can you give any hints what I can read to get into avahi coolness? Apr 04 19:37:18 documentation on http://www.avahi.org/ ? Apr 04 19:38:40 pH5: well, I want a dump howto, like "run avahi-daemon and now you magically can ping .local from anywhere on your network!" Apr 04 19:39:01 pH5: docs tell about boring things like some mDNS, and I need magic ;-) Apr 04 19:39:27 "install angstrom and you can ping .local from your broadcast domain" Apr 04 19:39:29 psokolovsky_: well, that'd be installing avahi-daemon and libnss-mdns Apr 04 19:40:02 psokolovsky_: http://www.avahi.org/wiki/Avah4users#Documentation is for you Apr 04 19:40:04 koen: as hostname, machine conf name is used, right? Apr 04 19:40:12 yes Apr 04 19:40:33 psokolovsky_: unless you change it of course :) Apr 04 19:40:38 pH5: ok, thanks, will wade thru it ;-) Apr 04 19:40:46 koen: do we have the nss-mdns resolver in avahi by default? Apr 04 19:40:54 Id on't want to chaneg anything, i want magic! Apr 04 19:41:30 pH5: not sure, but it does seems to broadcast it Apr 04 19:42:19 koen: right, avahi-daemon does that. Apr 04 19:42:33 I just realized that I never tried to ping laptop.local from the pda :) Apr 04 19:43:47 hi all, I'm having problems with new checksum scheme in base.bbclass, and I believe that there is a bug there... Then I'm trying to debug it adding some echo messages to figure out what is happening.... My doubt is: every time that I add a print statement in a bbclass file do I really need to reparse all files? and if it is not necessary is there an easy way to force bitbake avoid this parse? Apr 04 19:45:03 vivi|note: just touch the bbclass with a time in the past. Apr 04 19:47:22 thanks Apr 04 19:49:29 re Apr 04 20:04:48 03hrw 07org.oe.dev * rca890abe... 10/ (4 files in 2 dirs): zip: added DESCRIPTION, cleaned recipes (from Poky) Apr 04 20:04:53 03hrw 07org.oe.dev * r3b9a62d1... 10/ (1 packages/libol/libol_0.3.18.bb): libol: use binconfig class to install libol-config (from Poky) Apr 04 21:01:46 03philippe 07org.oe.dev * r65d33d65... 10/ (4 files in 3 dirs): Apr 04 21:01:46 packages/gpe-filemanager/gpe-filemanager_svn.bb Apr 04 21:01:46 * add new build recipe for svn gpe-filemanager Apr 04 21:01:46 packages/gpe-filemanager/files/svn-build.patch Apr 04 21:01:46 * patch needed to make gpe-filemanager_svn build Apr 04 21:11:36 03philippe 07org.oe.dev * rea9ae858... 10/ (1 packages/gpe-filemanager/gpe-filemanager_svn.bb): Apr 04 21:11:36 packages/gpe-filemanager/gpe-filemanager_svn.bb Apr 04 21:11:36 * use correct default preference of -1 Apr 04 21:15:34 03philippe 07org.oe.dev * r99b5d2f2... 10/ (4 files in 3 dirs): Apr 04 21:15:34 packages/gpe-sketchbook/gpe-sketchbook_svn.bb Apr 04 21:15:34 * add new svn build recipe Apr 04 21:15:34 packages/gpe-sketchbook/files/svn-build.patch Apr 04 21:15:34 * add build patch Apr 04 21:22:17 night Apr 04 21:24:42 03philippe 07org.oe.dev * r4d695f19... 10/ (1 packages/gpe-sketchbook/gpe-sketchbook_svn.bb): Apr 04 21:24:42 packages/gpe-sketchbook/gpe-sketchbook_svn.bb Apr 04 21:24:42 * add PV to recognise it as an svn build Apr 04 21:49:15 philippe: i need to do the same you do with gpe for gpephone tomorrow :) Apr 04 21:59:05 Command to import private key? Apr 04 22:00:05 03pfalcon 07org.oe.dev * reb15a9f3... 10/ (1 conf/machine/h4000.conf): h4000.conf: Install BT and WiFi firmware. Apr 04 22:03:10 florian: hopefully you have less work... Apr 04 22:03:20 well some of it has already been done now of course :-) Apr 04 22:04:02 philippe: i have, only a few thing need it now. the ptim stuff and gpe-phonepanel mostly. Apr 04 22:04:20 i wonder if i have the bb for the email client somewhere Apr 04 22:05:44 florian: calendar, contacts and sketchbook are already done Apr 04 22:06:01 see above :-) Apr 04 22:06:31 btw. i have an almost working scratchbox sdk built with oe now Apr 04 22:06:46 only autotools i18n stuff fails somehow Apr 04 22:06:55 florian: well I guessed ptim was a type for pim Apr 04 22:07:21 03kristoffer 07org.oe.dev * r88e7fc4e... 10/ (5 files in 3 dirs): Apr 04 22:07:21 packages/linux/linux-jlime-jornada6xx-2.6.21rc5: Apr 04 22:07:21 * Addition of new kernel for the hp6xx platform. Apr 04 22:07:21 Updates since last is mainly focused on adjusting to Apr 04 22:07:22 changed API's and also moving from ide-sh into Pata. Apr 04 22:08:46 philippe: no, its an input framework Apr 04 22:08:55 xkbd+++++ Apr 04 22:08:57 ;) Apr 04 22:09:24 florian: ok :-) Apr 04 22:14:03 florian: Regarding bug 2037, did you try inheriting debian.bbclass with bitbake 1.8? Apr 04 22:14:12 florian: Also, which DISTRO was it? Apr 04 22:16:29 RP: yes, i built meta-sdk-sbox which inherts debian. Apr 04 22:16:55 RP: generic distro built for a generic x86 target. Apr 04 22:17:19 florian: A .bb file shouldn't inherit debian, ever - it should be a .conf setting or it doesn't work properly Apr 04 22:19:09 florian: Looking at the output from "bitbake meta-gpe -g", I can't see a problem in the depends :/ Apr 04 22:20:36 RP: hmm.. ok... but isn't a bad idea if we need to inherit debian in order to be able to build the sdk packages? Apr 04 22:20:49 florian: This is even without debian.bbclass Apr 04 22:20:58 florian: Did you have packaging enabled? Apr 04 22:21:41 RP: how would i deactivate it? i got packages at least Apr 04 22:22:12 florian: If you didn't inherit a packaging class it would deactivate certain code Apr 04 22:23:50 florian: Looking at the differences between meta-gpe and meta-sdk, it could be something to do with the lack of DEPENDS I guess Apr 04 22:23:52 RP: from generic.conf: "INHERIT += "package_ipk"" Apr 04 22:24:45 RP: well... that does not give clue about the fact that it works with 1.6.6 Apr 04 22:24:45 florian: If I drop the DEPENDS from meta-gpe, it all goes a bit wrong which hints at the problem I guess Apr 04 22:25:31 Shouldn't the RDEPENDS magic solve this? Apr 04 22:25:38 florian: Dependency handling was totally rewritten between 1.6 and 1.8, not least that 1.8 is task based rather than package based Apr 04 22:25:49 florian: Yes, it should. I'm just trying to pin down the problem Apr 04 22:27:08 RP: I could try to add all the depends to the package and see if it works then. Apr 04 22:28:34 florian: Better still, I think a do_build[recrdeptask] = "do_build" will do it Apr 04 22:28:58 florian: I'll test this and I might add a meta.bbclass Apr 04 22:30:23 RP: ok cool Apr 04 22:30:28 florian: That looks like it should work to me - can you test and confirm? Apr 04 22:31:27 * florian needs to admit that he does not really have a clue why this is necessary Apr 04 22:31:54 RP: i can just add this line to the bb? Apr 04 22:31:59 florian: yes Apr 04 22:32:21 ok, but i need to switch back back first.. one sec Apr 04 22:32:41 florian: bitbake 1.8 removed the BUILD_ALL_DEPS handling as it was flawed. Instead it requires that you specify which tasks depend on which exactly Apr 04 22:34:08 florian: In this case nothing says that it should run all dependencies, hence the problem. Other meta files have by chance worked since they had DEPENDS set Apr 04 22:34:28 RP: ah ok, i got the point Apr 04 22:35:28 For the sdk builds, you might need do_populate_sdk[recrdeptask] = "do_package_write" Apr 04 22:36:08 florian: i.e. tell bitbake that it should run all the rdepends up to their package_write task before trying to populate the sdk Apr 04 22:40:11 RP: I'm having problems w/ kexec -e freezing the nslu2, is that supposed to even work on that machine? Apr 04 22:41:00 CoreDump|home: Does that work on the Z? If so, it should work on the nslu2 Apr 04 22:41:22 hmm good question Apr 04 22:44:01 RP: Do you see any differences between 1.6.6 and 1.8.0 that might influence environment variable handling? I just noticed minimo does not build with 1.8 Apr 04 22:44:37 For some reason it builds and runs binaries that should not be called while crosscompiling. Apr 04 22:44:38 florian: In theory no... Apr 04 22:45:48 I need to take a more close look, but with a really similar setup (but one machine with the new bitbake) it failed one the one and built pretty well on the other. Apr 04 22:46:43 florian: That is rather worrying and I'd appreciate it if you could try and find out why... Apr 04 22:47:19 CoreDump|home: yes. kexec -e works just fine on my poodle :) Apr 04 22:47:38 hvontres|poodle: thanks Apr 04 22:48:03 it's just hanging the nslu2 here Apr 04 22:48:57 CoreDump|home: Which kernel version are you using? I use 2.6.17 for my inital boot. so far I have not had much luck with 2.6.20 Apr 04 22:49:33 2.6.20 is installed and I try to kexec an identical kerneöl Apr 04 22:50:32 RP: I get a nice effect now: if i want to clean a package it seems to try to build the dependencies Apr 04 22:51:16 florian: Not good :/ Apr 04 22:51:21 CoreDump|home: did you rebuild kexec-tools? IIRC, the kecex syscall changed when kexec-arm went into mainline Apr 04 22:51:30 florian: With 1.6 or 1.8? Apr 04 22:51:47 RP: with 1.8 Apr 04 22:52:28 hvontres|poodle: Hmmm right, I'll rebuild and retry Apr 04 22:52:55 CoreDump|home: good luck. Time for me to go home :) Apr 04 22:53:06 hvontres|poodle: ;) cya Apr 04 22:53:19 RP: but i can confirm that it builds the dependencies now Apr 04 22:54:06 florian: Heading the right way. That clean behaviour puzzles me though. What was the exact command? Apr 04 22:55:20 root@mhnas01:~$ kexec -e Apr 04 22:55:20 Starting new kernel Apr 04 22:55:20 Bye! Apr 04 22:55:32 it just hangs there Apr 04 22:55:36 * CoreDump|home scratches his head Apr 04 22:55:43 RP: bitbake meta-sdk-sbox -c clean Apr 04 22:55:48 CoreDump|home: Can the nslu2 reboot properly? Apr 04 22:56:04 hah Apr 04 22:56:05 RP: yes, no problems there otherwise Apr 04 22:56:14 ~lart mozilla build system Apr 04 22:56:15 * ibot hurls dozens of incontinent, insomniac, hungry kittens with tiny little razor-sharp claws and a wide variety of contagious intestinal parasites at mozilla build system Apr 04 22:56:33 jesus, what a lart Apr 04 22:56:43 :-) Apr 04 22:56:55 which sick person is inventing them?? =) Apr 04 22:56:58 florian: When I run that, I see "Running task 1 of 1" Apr 04 22:57:44 RP: with do_populate_sdk[recrdeptask] = "do_package_write" or without? and do you maybe just have all dependencies packaged already? Apr 04 22:57:56 CoreDump|home: i wondered the same multiple times :-) Apr 04 22:58:29 RP: minimo builds some stuff with the host compiler and links it against staging libraries Apr 04 22:58:54 staging for the target to be more precise Apr 04 22:59:31 florian: I don't have half the dependencies built and I have that line of metadata added Apr 04 23:00:27 florian: I'd guess some package is installed which is confusing it which isn't present on the other system Apr 04 23:00:40 RP: hmm... maybe Apr 04 23:01:13 RP: where exactly did you add the line? even if it is quite unlikely to be important Apr 04 23:01:37 florian: At the end of the file. I'll commit my changes and you can see if you still see the same thing Apr 04 23:02:34 RP: ok Apr 04 23:02:51 * florian just switched back to 1.6 to verify his minimo issue Apr 04 23:03:01 i need more fast machines Apr 04 23:04:18 I wonder about the possible reactions of my wife if i bought this: http://shop.itsco.de/shop/default.aspx?TY=item&ST=1&IT=681&CT=40 Apr 04 23:06:17 I guess the major problem would be the neighbours... Apr 04 23:07:03 I have a Dell dual Xeon and its very very loud :/ Apr 04 23:08:01 * mwester_ has an ancient 4U HP server in a rack in his basement. Very loud, but only the cat sleeps there. Apr 04 23:09:03 RP: We have one of these at kc, but still empty. It is in a separate room at the ground floor... it still sounds like the sea in the first floor :) Apr 04 23:12:13 03rpurdie 07org.oe.dev * r4fa76278... 10/ (15 files in 3 dirs): meta packages: Create a new class, clean up the code and add correct dependencies for bitbake 1.8 Apr 04 23:16:21 RP: hmm... i just tried on a single machine. its not bitbake that causes the minimo build failure. Apr 04 23:16:46 florian: To me, thats a relief although it might not be for you ;-) Apr 04 23:16:58 florian: I committed the changes as you'll note Apr 04 23:17:57 right, i'm pulleing them right now to my local machine here... i can use that for testing as well. Apr 04 23:19:26 RP: in the local.conf where it builds i do not have fail-fast and use icecc Apr 04 23:48:17 'night all Apr 04 23:49:03 n8 RP Apr 05 02:30:45 jacques: hello Apr 05 02:41:31 howdy T0mW Apr 05 02:41:53 jacques: what you been up to lately? Apr 05 02:42:45 I think that the last time I talked with you you were working on an ARM Emulator (from japan?) Apr 05 02:52:32 probably skyeye Apr 05 02:52:39 I had some success with softgun Apr 05 02:52:50 nod Apr 05 02:52:57 qemu would be nice but it doesn't really support big-endian Apr 05 02:53:16 softgun does but is much slower than qemu Apr 05 02:53:18 I've been doing a lot of ARM7TDMI work lately, all of it with Philip LPC2000 chips Apr 05 02:53:35 too much of it, I need a break Apr 05 02:54:15 I've gotten to the point that I can now code pretty good in ARM assembler, heh Apr 05 02:54:32 cool :-) Apr 05 02:54:42 jacques: you still in oregon? Apr 05 02:55:24 nope, in texas again Apr 05 02:56:14 I took a vacation two years ago and was up in Wyoming (yellowstone). I was tempted to keep going west and see the pacific northwest. Apr 05 02:56:40 yeah it's nice up there Apr 05 02:56:56 but, that extra 2000 mile trip wasn't to be done. I had to get back east here to take care of things. Apr 05 02:57:08 jacques: what do they have you doing in TX? Apr 05 02:57:54 mostly linux programming on small systems Apr 05 02:58:02 jacques: arm? Apr 05 02:58:22 it's hard to call a 300MHz ppc with 64MB RAM "embedded" Apr 05 02:58:24 powerpc Apr 05 02:58:27 lol Apr 05 02:58:32 jacques: yeah Apr 05 02:58:39 it is a custom board tho Apr 05 02:59:24 jacques: I've got a project to get back to, asap (as the customer says) where I'll be doing my first large scale linux "embedded" system (samsung s3c2410 @ 200mhz) **** ENDING LOGGING AT Thu Apr 05 02:59:56 2007