**** BEGIN LOGGING AT Sat Dec 16 02:59:57 2006 Dec 16 07:17:49 morning all Dec 16 08:44:24 morning Dec 16 08:50:12 good morning all Dec 16 08:50:26 hi koen cliff Dec 16 08:50:30 morning :-) Dec 16 08:55:14 good morning Dec 16 08:56:06 someone know good playable opengl game for linux? Dec 16 08:56:31 and not FPP Dec 16 09:06:35 fpp? Dec 16 09:10:07 first person perpective (doom, quake) Dec 16 09:11:47 ahh, fps :P Dec 16 09:11:59 PlanetPenguin Racer Dec 16 09:13:54 hmm, bzflag Dec 16 09:14:02 neverball too Dec 16 09:18:22 hmmm Dec 16 09:18:35 multimachine doesn't work to well with multiple arm archs Dec 16 09:18:38 ~lart armv6 Dec 16 09:18:39 * ibot executes killall -HUP armv6 Dec 16 09:21:04 morning all Dec 16 09:23:23 koen: what is problem with it? Dec 16 09:24:03 it all ends up in {staging,cross}/arm-foo-linux-gnuabi/ Dec 16 09:24:16 which just gives weird problems Dec 16 09:24:35 TARGET_ARCH instead of other *ARCH? Dec 16 09:25:28 yes Dec 16 09:30:42 here, arm-foo-linux-gnuabi-* ends in cross/bin/ and {ar,as,g++,gcc,..} ends in cross/arm-foo-linux-gnueabi/bin/. staging/arm-foo-linux-gnueabi/bin/ remains empty Dec 16 09:31:16 strangely it does not prevent building images on my debian box, but not on my gentoo box Dec 16 09:32:31 so on my gentoo box I still have to patch bitbake.conf: http://www.rafb.net/paste/results/YU0Wql36.html Dec 16 09:32:52 * cyrilRomain should read the heisenbug thread ... Dec 16 09:35:12 hrw: is there a way to detect if someone plugs in a usbhost cable + device so udev can load the ohci module? Dec 16 09:36:14 koen: there was a patch from pdax for it but we did not used it Dec 16 09:48:07 s/patch/hack/ Dec 16 09:48:27 and iirc that was a regular cable detect instead of usbhost detect Dec 16 09:52:20 iirc it was 'device/host' cable detect Dec 16 09:52:44 I lack original device cable anyway Dec 16 09:53:47 likewise: good morning Dec 16 09:54:04 good morning koen and everyone else Dec 16 09:54:14 hi likewise Dec 16 09:54:15 likewise: I saw a comment from carl worth that cairo has gotten much faster, cairo-clocks hasn't gotten much faster Dec 16 09:55:01 koen: cairo faster, since when, did he mean? Dec 16 09:55:23 cairo 1.2.x vs 1.3.x Dec 16 09:55:26 koen: I think I saw cairo clocks speed up a factor 4 or 5 since 1.0.x Dec 16 09:55:35 macslow didn't see a big improvement Dec 16 09:55:44 but it could be that he didn't test on FPUless hw Dec 16 09:56:19 I'm two days behind on the mailing lists, let me catch up Dec 16 09:56:35 koen: let me commit the gcc patch first, even before coffee. Dec 16 10:01:48 cu Dec 16 10:24:13 hi. just started a bb task-bootstrap with machine=geodelx, distro=generic. will be quite excited to see the result Dec 16 10:25:35 oe instructions were very good and easy to follow. thanks for this Dec 16 10:38:20 likewise: I pushed your gcc patch Dec 16 10:42:31 good morning Dec 16 10:42:51 hey Laibsch Dec 16 10:43:27 If a recipe has a line like "install -d ${D}${bindir} ${D}${mandir}/man1" in do_install, does that mean that man pages will end up in the ipk? I guess that would be a waste. Dec 16 10:43:49 install -d == mkdir -p Dec 16 10:44:44 and FILES specifies what ends up in the packages Dec 16 10:44:58 but default ${mandir} ends up in ${PN}-doc Dec 16 10:45:25 OK Dec 16 10:45:55 s/but/by/ Dec 16 10:45:59 * koen kicks CIA-4 Dec 16 10:46:00 ow Dec 16 10:47:43 koen: thanks. I had a diff and am rebuilding. Not that it matters, its x86 related. Dec 16 10:48:15 It matter for my OE powered workstation :) Dec 16 11:06:21 03likewise 07org.oe.dev * rc1aab6fe... 10/ (4 files in 3 dirs): gcc 4.1.1: add patch to fix PR 13685 Dec 16 11:08:37 what is PR 13685? Dec 16 11:10:56 see the description in the patch Dec 16 11:11:19 Submitted By: Alexander E. Patrakov Dec 16 11:11:20 +Date: 2006-12-11 Dec 16 11:11:20 +Initial Package Version: 4.1.1 Dec 16 11:11:20 +Upstream Status: backport Dec 16 11:11:23 +Origin: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28621 Dec 16 11:11:23 +Description: Fix crash of programs compiled with -Os -ffast-math Dec 16 11:12:21 Well, nothing about PR or 13685. Dec 16 11:13:39 Judging from the info in the URL this is some kind of way to refer to bug numbers in gcc world. Thanks, koen. Dec 16 11:15:21 Laibsch: actually there is, on the bug report page. http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13685 blocks 28621. Dec 16 11:16:50 pH5: yes, as I said in the second statement it is in the URL. There was nothing about PR directly in the stuff that koen posted is what I meant. Dec 16 11:16:56 Can I add files via SRC_URI += "$somefile"? Dec 16 11:17:07 Laibsch: yes Dec 16 11:17:12 nice Dec 16 11:17:34 does will be put into ${WORKDIR} iirc Dec 16 11:18:09 s/does/those/ Dec 16 11:18:13 * koen needs coffee Dec 16 11:20:31 how to create harddisk for qemu with OE? Dec 16 11:28:01 hrw|gone: MACHINE=qemux86 bitbake foo-image Dec 16 11:28:17 ah.. thx Dec 16 11:28:36 that might need a patch or two from poky Dec 16 11:32:13 koen: for example machine ;) Dec 16 11:32:31 heh, right Dec 16 11:34:52 argh.. deps Dec 16 11:42:02 hm. my progear kernel does not fit Dec 16 11:50:18 could someone add "bluetooth ppp wifi" to MACHINE_FEATURES in my machine file htcuniversal.conf? Dec 16 11:50:54 ppp is distro not machine Dec 16 11:51:42 ok, forget ppp then, and just add bluetooth and wifi please. Dec 16 11:52:13 ok. moment Dec 16 11:52:17 I am trying to figure out why all the modules is beeing installed. Dec 16 11:52:22 Thansk a lot hrw Dec 16 11:57:35 monotone server is slooow Dec 16 12:03:10 s/server // ;) Dec 16 12:09:34 * koen is pondering about booking an extra night at fosdem Dec 16 12:10:10 koen: sunday/monday? Dec 16 12:10:16 yes Dec 16 12:10:56 * Crofton is still toying with the idea of going to FOSDEM Dec 16 12:11:13 Crofton: I'll buy you a beer :) Dec 16 12:11:45 with the worthless dollar you may have to ... Dec 16 12:11:45 hmm, bootstrap failed. package glibc-2.3.5+cvs20050627-r13: task do_compile: failed. i guess task-bootstrap was not the right thing to try for geodelx+generic. ok. trying bootstrap-image Dec 16 12:12:27 ~convert 1 usd to euro Dec 16 12:12:30 euro is an invalid unit? Dec 16 12:12:33 Crofton: thats why I use EUR as currency in all calculations Dec 16 12:12:37 ~convert 1 usd to eur Dec 16 12:12:40 eur is an invalid unit? Dec 16 12:12:42 ~change 1 usd to eur Dec 16 12:12:42 ~convert 1 usd to eu Dec 16 12:12:46 eu is an invalid unit? Dec 16 12:12:49 1.00 United States Dollar (USD) makes 0.763595 Euro (EUR) (from http://www.xe.com/) Dec 16 12:12:54 ~convert 1 inch to cm Dec 16 12:12:57 1 inch is approximately 2.54 cm Dec 16 12:13:11 ~convert 1 eu to usd Dec 16 12:13:20 ~change 1 eur to usd Dec 16 12:13:26 1.00 Euro (EUR) makes 1.30959 United States Dollar (USD) (from http://www.xe.com/) Dec 16 12:13:30 grr Dec 16 12:13:39 change for currencies, convert for converting Dec 16 12:13:52 ah Dec 16 12:13:54 hm. 230M feed... Dec 16 12:14:08 need to give net for qemu Dec 16 12:14:45 should i change the part of the oe wiki for gettingstarted that says to do bitbake task-bootstrap? assuming bootstrap-image works Dec 16 12:15:10 right, task-bootstrap is deprecated Dec 16 12:15:18 but that's not the cause of your problems Dec 16 12:16:27 i see. the err for failing glibc was about undefined ref in pthread to _errno Dec 16 12:16:41 geodelx is i486, right? Dec 16 12:16:53 Cleaning up some stuff here. postrm and postinst have slightly evolved over time for avahi bb files. I wonder if I can still move the latest version to the include file and remove it for all other packages? hrw? Dec 16 12:16:54 yup, the geodelx.conf says i486 Dec 16 12:17:08 Laibsch: never used avahi Dec 16 12:17:13 Laibsch: we only need the latest avahi Dec 16 12:17:32 grr. Then why did I do all this work :-( Dec 16 12:17:51 Laibsch: to practise? Dec 16 12:18:16 hrw: you're leaving sunday at fosdem? Dec 16 12:18:48 koen: planned to Dec 16 12:19:00 mickey|zzZZzz: pong Dec 16 12:19:04 ehm Dec 16 12:19:04 koen: and bought place tickets already Dec 16 12:19:07 mickey|zzZZzz: ping Dec 16 12:26:12 hrw: well... Dec 16 12:26:15 Anyways, here is my patch for everybody's kind inspection: http://oz.leggewie.org/unify.a.txt Dec 16 12:28:40 look ok Dec 16 12:29:20 OK, committing. Dec 16 12:30:19 Laibsch: SPLIT it per dir Dec 16 12:30:34 1. avahi 2. autoconf etc... Dec 16 12:56:02 hi zecke Dec 16 13:09:31 hi hrw, hi zecke Dec 16 13:09:42 hi greentux Dec 16 13:10:05 ~lart qemu networking Dec 16 13:10:05 * ibot judo chops qemu networking Dec 16 13:10:13 hrw: do you moved to another local company? or will You mode around? Dec 16 13:10:43 greentux: I will work for my own company here. It will be full time contract job Dec 16 13:11:34 hrw: ok understand, so it will be a community related contract? from outsite polend? Dec 16 13:11:44 yes Dec 16 13:12:02 or I will tell it in different way Dec 16 13:12:31 one company want me to do full time job for them. for me it is better to do it as company then being hired by them Dec 16 13:12:35 taxes.. Dec 16 13:13:06 taxes, schmaxes Dec 16 13:13:17 hrw: yes understand... Dec 16 13:13:18 koen: 19% contra 30-40%... Dec 16 13:13:30 I know Dec 16 13:13:42 19%??? where? Dec 16 13:13:49 greentux: poland Dec 16 13:13:56 you guys are becoming Americanized, always worrying about taxes Dec 16 13:14:14 hrw: ok, will move to poland... (minus 10 % :) Dec 16 13:14:23 hi koen Dec 16 13:14:37 hey greentux Dec 16 13:15:02 some news about the "new" 770? Dec 16 13:15:11 only the FCC docs Dec 16 13:16:11 no xmas surprise from nokia Dec 16 13:16:24 nope Dec 16 13:17:50 koen: but most of the stuff is still confidental? Dec 16 13:18:04 Anything to watch out for when upgrading from monotone 0.28 to 0.31? Dec 16 13:18:05 zecke: yes Dec 16 13:18:41 Laibsch: no Dec 16 13:18:53 ARHG Dec 16 13:18:59 ~lart qemu deeply Dec 16 13:19:00 * ibot takes out a seltzer bottle and sprays qemu deeply in the face. You know, one of those old-school seltzer bottles clowns have? Yeah those. Anyway, consider yourself spritzed Dec 16 13:19:19 sorry Dec 16 13:19:21 Laibsch: http://venge.net/monotone/UPGRADE Dec 16 13:20:32 zecke: Since there was nothing about later version, I was also getting a bit nervous on how current that package is ;-) Dec 16 13:20:47 Better safe than sorry. Dec 16 13:20:55 Laibsch: you need to upgrade stuff Dec 16 13:21:09 http://www.openembedded.org/snapshots/ Dec 16 13:21:28 this is why we have OE.mtn, OE-this-is-for-mtn-0.29 and -0.30 Dec 16 13:21:54 My talk about OpenEmbedded during Pingwinaria was accepted Dec 16 13:22:06 Pingwinaria is Polish Linux meeting in March Dec 16 13:22:40 now need to think what to say then Dec 16 13:22:59 zecke: So I take for-0.30 for version 0.31-version of mtn as well? Dec 16 13:23:52 zecke: what do you think about qpe community version? Dec 16 13:24:18 is it finally there? Dec 16 13:24:36 I have been watching www.trolltech.com for weeks Dec 16 13:25:17 http://www.trolltech.com/developer/downloads/qtopia/coregpl Dec 16 13:26:49 ~qtopia core Dec 16 13:26:51 Qtopia Core, non working QtE for Qt4 Dec 16 13:26:51 methinks qtopia core is qt/embedded. please don't support stupid marketing tricks. Dec 16 13:28:16 RP: Hi! Dec 16 13:28:35 RP: I wonder, do you have suggestions for hx4700 vs w100fb stuff? Dec 16 13:30:09 zecke: not understand: "Qtopia Core is the embedded Linux port of Qt 4 " Dec 16 13:30:26 zecke: "Qt/Embedded was the embedded Linux port of previous Qt versions." Dec 16 13:31:04 greentux: qt/e was renamed to qtopia core ~half year ago Dec 16 13:31:41 yes so i understand there is a communit yversion of QtE for Qt4 aka "Qtopia Core"? Dec 16 13:37:02 hrw: so the qt univerum will be to dark for me... Dec 16 13:41:59 zecke: I guess the proposed "mtn migrate_workspace" is going to take forever? Will I loose my keys or anything like that by starting from scratch with one of the snapshots? Dec 16 13:46:15 hmm. same failure with bb bootstrap-image. package glibc-2.3.5+cvs20050627-r13: task do_compile: failed Dec 16 13:46:26 Laibsch: mtn keys are saved in $HOME/.mtn Dec 16 13:46:41 migrate_workspace is quite fast Dec 16 13:46:45 only changes the _MTN dir Dec 16 13:47:00 regenerating rosters can take ages Dec 16 13:47:50 zecke: would I need to regenerate the rosters? Dec 16 13:47:57 Whatever they are. Dec 16 13:48:10 i guess i should go look at why bb is pulling glibc-2.3 when generic.conf says PREFERRED_VERSION_glibc ?= "2.5" Dec 16 13:48:28 Laibsch: a) grep OE-mtn for 0.30 and only migrate your workspace Dec 16 13:48:46 b) install 0.31 and regenerate your rosters and migrate your workspace (takes longer) Dec 16 13:49:23 koen, any progress getting OE to build for efika? Dec 16 13:50:00 Crofton: http://www.angstrom-distribution.org/unstable/images/20061214/efika/ Dec 16 13:50:47 cool Dec 16 13:51:13 I am hopeful I can get the efika gonig and use the bootloader to mov the ext2 image to the disk Dec 16 13:51:15 hmm Dec 16 13:51:34 * koen boots via tftp Dec 16 13:51:45 hmm Dec 16 13:52:25 I want to get a system going that can compile sdcc code for some wrk I need to do Dec 16 13:52:56 booting via tftp and poking at the disk from there is pretty easy Dec 16 13:53:16 ok Dec 16 13:53:20 that sounds like a plan Dec 16 13:55:29 . Dec 16 13:55:51 .. Dec 16 13:56:42 * Ifaistos tests the new installed dsl modem Dec 16 13:59:50 I am trying to clean up my machine file. Is there another way to aply this modules, because i dont think some of them belongs to MACHINE-FEATURES: http://pastebin.ca/281100 Dec 16 14:00:58 fuse, smbfs aren't machine specific Dec 16 14:01:28 koen: But rest of them is OK that have in my machine file right? Dec 16 14:02:04 not really Dec 16 14:02:22 kernel-module-rtc-pcf8563 should pull in all other rtc modules Dec 16 14:03:12 Ok, i'll try that. Dec 16 14:03:37 anyone have any advice for solving why geodelx+generic pulls glibc-2.3.5 Dec 16 14:03:52 which fails to build with pthread missing _errno Dec 16 14:04:24 goxboxlive: and what do you need bfusb for? Dec 16 14:08:38 after 0.5h qemu parsed 22% of recipes... Dec 16 14:09:34 hmm. i guess i'm on my own figuring this out. Dec 16 14:09:52 NOTE: Couldn't find shared library provider for libc.so.6 Dec 16 14:09:52 NOTE: package gcc-cross-4.1.1-r9: task do_package: completed Dec 16 14:09:54 argh Dec 16 14:09:59 koen: No, thats a leftover. I'll remove it. Dec 16 14:34:10 hrw: that first NOTE seems to be the OBJDUMP/PATH problem all over again Dec 16 14:38:50 I know Dec 16 14:39:25 69% recipes parsed Dec 16 14:39:59 I need 2nd x86 machine (progear does not count - 112M ram hurts) Dec 16 14:40:44 run the qemu image on your Zs ;) Dec 16 14:41:13 gerwinin: wb Dec 16 14:41:20 koen: only if you share your cluster stuff Dec 16 14:41:28 :D Dec 16 14:41:43 cluster-vpn.openembedded.org Dec 16 14:41:44 koen: but building cluster from Z will be intersting thing Dec 16 14:42:00 tosa, c760, spitz as cluster... Dec 16 14:42:27 tosa+512M sd, c760 + 1G sd, spitz with 4G microdrive.. Dec 16 14:42:50 all are angstrom supported.. Dec 16 14:42:52 hmm task-base is failing after gcc update Dec 16 14:43:04 so I only need to provide networking for them Dec 16 14:43:10 goxboxlive: fails? Dec 16 14:43:25 no buildable providers for task-base Dec 16 14:43:37 run with -DDD Dec 16 14:44:33 ~curse ipkg-make-index Dec 16 14:44:35 May the fleas of a thousand camels infest your most sensitive regions, ipkg-make-index ! Dec 16 14:45:51 likewise : hey :) Dec 16 14:46:57 hrw: pastebin.ca/281148 Dec 16 14:47:21 ERROR: Nothing provides runtime dependency # Dec 16 14:47:34 you have # inside somewhere Dec 16 14:47:52 it is not allowed Dec 16 14:48:54 hmm that might be right Dec 16 14:55:52 90% parsed Dec 16 14:58:01 http://www.hrw.one.pl/2006/12/16/flashing-progear-bios/ Dec 16 14:59:02 hi chouimat Dec 16 14:59:20 morning Dec 16 15:03:50 koen, #efika exists Dec 16 15:12:10 ~lart executable stacks Dec 16 15:12:11 * ibot slams executable stacks against a large cement Tux Dec 16 15:14:15 *-native use do_install for staging? Dec 16 15:19:47 anyone got a log.do_configure handy for an armv5te or simiar mplayer build? Dec 16 15:20:06 .dev? Dec 16 15:20:52 I can do build of mplayer for umbaumba Dec 16 15:20:54 hrw: as long as its 1.0pre8 or later it doesnt matter Dec 16 15:21:09 hrw: on arm Dec 16 15:21:09 XorA|gone: I can have one in a few minutes Dec 16 15:21:28 koen: that would be cool Dec 16 15:21:38 started two builds Dec 16 15:21:52 trying to track down why armv4 seems to have executable stacks but not armv54 Dec 16 15:21:56 -4 Dec 16 15:23:48 hrw: When i want something changed inthe oe.dev, like adding a new kernel defcofig file and changing my machine file should i report that as a bug? Dec 16 15:23:57 goxboxlive: yes Dec 16 15:24:08 goxboxlive: yes, that makes sure it doesn't get forgotten Dec 16 15:24:13 ok Dec 16 15:24:16 goxboxlive: was it you that had problems getting /dev/input/touchscreen0 created? Dec 16 15:24:50 goxboxlive: when amount of your good patches will reach some level we will give you r/w to not have to work on them Dec 16 15:26:26 yes Dec 16 15:28:07 someone here use intel core duo? Dec 16 15:28:11 goxboxlive: udev 0.97 on your device? Dec 16 15:28:20 hrw: yes Dec 16 15:28:25 hrw: Xeon one Dec 16 15:28:25 hrw: I dont want r/w access. It's not safe . Dec 16 15:28:30 hrw: but it's a mac Dec 16 15:28:44 koen: its 32 or 64bit? Dec 16 15:28:56 64bit (core 2 duo) Dec 16 15:29:41 I am not sure wich version, but it is the one provided with angstrom. Dec 16 15:29:53 although I suspect OSX 10.4.8 i still 32bit Dec 16 15:30:04 koen: core 1 duo were 32bit? Dec 16 15:30:11 yes Dec 16 15:30:16 thx Dec 16 15:30:46 * XorA|gone likes having 2x 64bit cores Dec 16 15:31:28 wish I had that at home Dec 16 15:31:37 Asus A6M-Q024 looks interesting Dec 16 15:31:54 15.4" (1280x800) with turion x2 for <1000 eur Dec 16 15:31:58 goxboxlive: check, because for me 0.97 is borken, and it accidently made it into Angstrom feeds Dec 16 15:32:13 XorA|gone: http://ewi546.ewi.utwente.nl/tmp/h2200/log.do_configure.26204 Dec 16 15:33:43 hmm, it seems that on armv4 all stacks are set as executable, but armv4t and upwards they are set as non executable, hence the collie mplayer bug Dec 16 15:34:09 my udev version 0.92 Dec 16 15:36:36 CIA-4: kick Dec 16 15:36:56 hrw: cia has been dead since yesterday Dec 16 15:36:59 ok Dec 16 15:37:07 CIA-4: findutils: updated to 4.2.29 Dec 16 15:37:14 4.1.20 was from 2003 iirc Dec 16 15:42:06 koen: cheers for the log Dec 16 15:42:35 !oebug 1692 Dec 16 15:42:37 * * Bug 1692, Status: RESOLVED, Created: 2006-12-16 08:06 Dec 16 15:42:38 * * goxboxlive(AT)gmail.com: htcuniversal.conf and kernel defconfig update. Dec 16 15:42:39 * * http://bugs.openembedded.org/show_bug.cgi?id=1692 Dec 16 15:42:44 hmm. bitbake in qemu has everything needed to start build. Dec 16 15:43:42 time to build gcc-cross inside of OE rootfs Dec 16 15:43:50 using OE ofcourse Dec 16 15:44:00 'self hosting' Dec 16 15:48:12 http://lists.kde.org/?l=kde-devel&m=116628335713678&w=2 Dec 16 15:48:13 lol Dec 16 15:51:32 Hi, anyone knows what happend to oesf.org forums? Dec 16 15:52:11 ~oesf Dec 16 15:52:13 well, oesf is rumour has it, oesf is newer name of ZUG - it provides forum for Zaurus/Ipaq/Simpad/etc users (see zug). It has nothing in common with OpenEmbedded (except for the first two letters), see http://www.oesf.org/ Dec 16 16:05:32 koen: updated, removed fuse and smbfs Dec 16 16:14:10 ~hail task-proper-tools Dec 16 16:14:12 * ibot bows down to task-proper-tools and chants, "I'M NOT WORTHY!!" Dec 16 16:15:36 :D Dec 16 16:17:15 finally started gcc-cross build. but adding rm_work inherit forced reparsing... Dec 16 16:17:25 but only 164 recipes to parse :D Dec 16 16:19:33 koen: add tar to task-proper-tools Dec 16 16:20:09 tar: unrecognized option `--no-same-owner` Dec 16 16:22:08 hrw: option is version-dependent, which tar? Dec 16 16:23:11 busybox one Dec 16 16:23:14 hrw: apr-utils doesn't build for x86 Dec 16 16:23:31 -s Dec 16 16:23:50 argh Dec 16 16:25:43 ping Ifaistos Dec 16 16:26:04 ~lart amd fro dropping socket 939 Dec 16 16:26:04 * ibot throws amd fro dropping socket 939's poor little doggy off a cliff Dec 16 16:32:26 gcc should depends on linux-libc-headers-dev Dec 16 16:32:44 hrw: rrecommend Dec 16 16:33:09 rdepends on libc-dev Dec 16 16:33:43 they collide ;( Dec 16 16:34:40 linux-libc-headers-dev & libc6-dev Dec 16 16:37:09 quilt-native need bash Dec 16 16:39:20 and perl ;( Dec 16 16:40:19 and our perl packaging is.... scary Dec 16 16:41:12 now I understand why having 2 machines at home is handy... Dec 16 16:45:29 koen: Is there still something wrong with my patches/ kernel config ? Dec 16 16:45:54 you have too many rtc devices Dec 16 16:46:15 I tried with the Dec 16 16:46:33 koen, Hi! Dec 16 16:46:46 hey psokolovsky Dec 16 16:46:48 koen, did you think about something like KERNEL_FEATURES? Dec 16 16:46:51 Ifaistos: hi Dec 16 16:46:53 you know something is wrong when you get: -bash: fork: Cannot allocate memory Dec 16 16:47:12 koen, to control kernel config in high-level way and consistent across machines? Dec 16 16:47:22 psokolovsky: yes, we had discussion about that a lot the past 3 years Dec 16 16:47:48 koen, and to what it came? Dec 16 16:48:10 we decided against it Dec 16 16:48:18 koen, why? Dec 16 16:48:28 check the logs :) Dec 16 16:48:34 ;-) Dec 16 16:49:03 You mean that i have both rtc-sa1100 and rtc-pcf8563 ? Or the rtc-sysfv, -dev, and -proc? Dec 16 16:49:08 koen, anyway, I have here solution for that in testing. And for me, it's clearly a good thing ;-) Dec 16 16:50:03 psokolovsky: interesting.. 2.6 or 2.4 kernels? Dec 16 16:50:18 hrw, 2.6 of course ;-) Dec 16 16:50:26 2.6.9 or 2.6.10 or 2.6.11 or 2.6.12 or 2.6.13 Dec 16 16:50:42 hrw, I probably post an ANN when it will haev a bit of doc ;-) Dec 16 16:50:57 k Dec 16 16:51:43 koen, yep, that's the issue. But it's still easy to upgrade one "base" config to a new version manually once, than do it manually for each machine ;-) Dec 16 16:51:59 so far I testing it with .18 and .19 Dec 16 16:53:17 hrw: http://www.angstrom-distribution.org/repo/?action=details&pnm=mc Dec 16 16:53:38 food for me?! Dec 16 16:54:19 koen: thx Dec 16 16:54:40 vim, less tar and the updated task-proper-tools are in as well Dec 16 16:54:47 perl? Dec 16 16:55:11 yep Dec 16 16:55:32 task-self-hosting will be BIG Dec 16 16:55:49 quilt-native 0.45 do_compile STARTED Dec 16 16:56:15 968 packages installed on host... Dec 16 16:56:27 861 are perl ;D Dec 16 16:57:35 http://www.hrw.one.pl/articles/ Dec 16 16:59:56 ok. I go shopping and build will progress... Dec 16 16:59:58 cu Dec 16 17:02:48 morning all Dec 16 17:03:17 RP: hi! Dec 16 17:03:19 RP: I wonder, do you have suggestions for hx4700 vs w100fb stuff? Dec 16 17:05:51 psokolovsky: Have a look at struct w100_gpio_regs. I think there should be something added to this for suspend/resume handling Dec 16 17:08:18 RP: well, of, thanks for hint. But notw that our changes are generally not GPIO-related. they just happen to do sth with w100 gpio on hx4700. on other machine, there will be done sth, else, possibly. Dec 16 17:08:40 psokolovsky: What other changes are you needing? Dec 16 17:10:06 RP: for now, that's all - a handler to tap into w100 chip suspend path. but semetrically, resume would be needed, too. Other changes (not-hx) would be to allow for more flexible init config... Dec 16 17:10:40 psokolovsky: The point is that a w100 is a w100. You shouldn't need to have machine specific bits Dec 16 17:11:12 RP: but you see, they are needed ;-) Dec 16 17:11:22 psokolovsky: What are needed? Dec 16 17:12:10 RP: what I wrote above. at least suspend handler ;-) Dec 16 17:12:34 psokolovsky: Why though? Dec 16 17:12:52 The core should be handling the GPIOs Dec 16 17:13:14 RP: because w100 is w100, but there're "tie up chips/hw" which needs to be controlled too Dec 16 17:13:32 RP: how core can handle GPIOs is they are machine-specific? Dec 16 17:14:40 psokolovsky: You add the suspend values to w100_gpio_regs - the machine specifc gpio configuration. The core can then take care of things Dec 16 17:15:18 Adding the callback just encourages people to hack around problems Dec 16 17:15:31 RP: Well, ok, I will try. But next machine won't use w100 gpios, it will use cpu's gpios, or perephiral controller's gpios. Dec 16 17:16:02 psokolovsky: Do you know what those gpios do? Dec 16 17:17:20 RP: I wrote you in email. Yes and no. But what's important is that it's not important what they do ;-). They do "something". Dec 16 17:20:40 psokolovsky: The problem is I'm not sure you understand what that code in the hx4700 suspend path acutally does. If you understand what it did, you migt understand why it doesn't belong there Dec 16 17:21:29 RP: I twll you what it does: it powers off some external, or maybe even internal machinery related to w100 chip. Dec 16 17:22:57 psokolovsky: but these are connected on the w100's gpios? Dec 16 17:23:51 psokolovsky: If we ever have a case where there are external gpios involved, I will taker a patch to add suspend/resume hooks. Until we need that, I don't see the point in adding them. The internal w100 gpios should be being handled properly and at the monent are not Dec 16 17:24:21 Other devices such as the c7x0 could benefit from these changes as theoretically, it could lower its suspend power consumption Dec 16 17:24:38 RP: Funky, yeah? remember those PXA's 3.8...mhz crystall stuff? You also can turn it on programmatically, but oops, cannot turn off by other measn execpt CPU reset. Dec 16 17:25:07 RP: Ok, I hope that's indeed the reason you don't want to favor taht patch ;-) Dec 16 17:25:14 I will look into that gpio stuff. Dec 16 17:25:28 psokolovsky: What other reason do you think I have? Dec 16 17:27:17 RP: I really had a suspicion that's it's another "port feudalism" case. "I don't need that on Zaurus, so it must not be needed at all" ;-) Dec 16 17:27:38 there is no RMK round here Dec 16 17:28:01 ~seen RMK Dec 16 17:28:20 i haven't seen 'rmk', pH5 Dec 16 17:28:41 XorA|gone: not that I didn't believe you Dec 16 17:30:24 psokolovsky: It more the opposite, I'm wondering if it would help the Zaurus ;-) Dec 16 17:30:45 RP: Ok, perfect then. thanks! Dec 16 17:33:00 ~lart cpus for being too slow Dec 16 17:33:02 * ibot DoSes cpus for being too slow Dec 16 17:39:11 RP: another random question: do you know if new RTC classdev (or old dev for that matter) allows to setup alarm by mere reading/writing files (think: manually from shell)? Dec 16 17:39:38 psokolovsky: As far as I know it needs ioctls Dec 16 17:39:46 ;-( Dec 16 17:39:56 psokolovsky: Have you seen the rtc test app? Dec 16 17:39:59 why shell doesn't offer ioctl command? ;-D Dec 16 17:40:15 psokolovsky: python shell does Dec 16 17:40:19 RP: sure, np. just wante dto know if I can do it solely with bare busybox Dec 16 17:40:36 zecke, lol ;-). I better install atd Dec 16 17:40:43 time to add setalarm to busybox Dec 16 17:40:43 RP: Are you still interested in doing the collie sd driver? Dec 16 17:41:03 JoseJX: Still interested yes although I don't have much spare time Dec 16 17:41:05 XorA|gone, yep ;-) Dec 16 17:41:22 RP: Okay, just checking :) Dec 16 17:41:31 RP: You know where the specs are right? Dec 16 17:43:15 hrw|gone: will the next OZ be generated from the branch? Dec 16 17:50:10 JoseJX: someone leaked the specs? Dec 16 17:51:49 XorA|gone: No, I translated the whole driver from asm to C and wrote my own specs Dec 16 17:52:38 JoseJX: You completed the decoding? Dec 16 17:52:39 XorA|gone: A few things are still missing from my specs, but I'm working on them now, that's why I pinged RP :) Dec 16 17:52:50 JoseJX: amazing Dec 16 17:52:51 RP: Yeah, the driver compiles now actually :) Dec 16 17:53:16 JoseJX: and works? Dec 16 17:53:34 XorA|gone: I did bcm43xx too, after this, I'm going to work on that some more. It's not hard, just time consuming :p Dec 16 17:53:45 RP: I'm fixing bugs now Dec 16 17:54:18 OMG that makes Angstrom on collie a reality Dec 16 17:54:36 RP: Once I get it working, I can play with the clock stuff a bit more so I can document that better Dec 16 17:54:49 XorA: http://josejx.net/collie Dec 16 17:55:24 Im glad I just bought a collie :-) Dec 16 17:55:52 Well, the driver still needs to be written Dec 16 17:58:21 likewise : hi Dec 16 17:58:41 * Ifaistos was fighting with the dsl modem all day Dec 16 18:11:49 re Dec 16 18:14:03 XorA: does mplayer have a --disable-foo switch for executable stacks? Dec 16 18:19:08 koen: no, but the patch on 1666 removes them Dec 16 18:19:16 koen: I am applying that patch for collie only Dec 16 18:23:42 koen: I think mplayer initiated the no exec stacks thing because of the large number of .wmv and .avi based stack bounce attacks Dec 16 18:24:04 that only applies to x86, right? Dec 16 18:25:21 koen: it does until the haxors start nuking mythtv boxes Dec 16 18:38:34 later Dec 16 18:42:11 ok Dec 16 18:42:27 now I know what happens if KITT and the terminator have babies Dec 16 18:43:33 fracking cylons Dec 16 18:43:44 koen: :-D Dec 16 18:43:49 koen: Wich rtc modules are you atlking about. Is it the dev sysfs and core you are taliking about or is it the sa1100 and pcf8563. I'll remove the stuff so that the pacth can be aplied. Because the current machine file isent any good. Dec 16 18:43:50 koen ? Dec 16 18:43:56 koen: how much BSG RI have you watched Dec 16 18:44:54 XorA: prelude + [1-4] Dec 16 18:45:01 watching 5 now Dec 16 18:45:14 chouimat: battlestar galactica Dec 16 18:45:34 koen I know ... /me waiting to watch 3x11 tonight Dec 16 18:45:45 koen: ah, you got a whole 2 series still then Dec 16 18:45:58 oooh, 3x11 is showing cool Dec 16 18:46:14 XorA: someone on the couch next to is protesting against watching more BSG Dec 16 18:46:16 XorA: Fall finale ... back in january Dec 16 18:46:35 chouimat: oh, I already watched that then Dec 16 18:46:56 XorA: yesterday episode for the US ... tonight on cable in Canada Dec 16 18:48:12 chouimat: looks like I havent watched that yet, time for some bittorrent Dec 16 18:48:19 hehe Dec 16 18:48:38 XorA: ah, you have same backup service :) Dec 16 18:48:48 uTorrent backup Dec 16 18:51:51 Hello Dec 16 18:52:51 Koen: If i remove this module (fpcf8563) from the patch, will you commit it then? But if you do, wait for an updated defconfig file to, because i have to change that one to then. Dec 16 19:12:11 Why is it that, even when I install font-adobe-75dpi and all that, and run "fc-cache -fv", and have fc-cache tell me that the 192 fonts in /usr/lib/X11/fonts/75dpi were cached, that xlsfonts persists on not showing the new fonts? Dec 16 19:12:46 Is Xfbdev eternally locked into using misc-fixed fonts? Dec 16 19:15:37 Is everyone avoiding me?:( Dec 16 19:15:40 Even CIA Dec 16 19:19:57 Is there any way to use any other font except for misc-fixed in GTK 1 apps Dec 16 19:20:04 And X-based applications Dec 16 19:20:23 Apps that use LibXaw for widgets and so on Dec 16 19:20:37 ? Dec 16 19:37:04 B_Lizzard: xset fp+ rehash? Dec 16 19:37:14 Hmmmm Dec 16 19:37:47 Lemme check Dec 16 19:37:59 Ah, damn Dec 16 19:38:05 I don't have xset installed Dec 16 19:38:12 Will check your command, thanks Dec 16 19:39:59 gtk1 is ancient, so fancy-schmancy fontconfig commands aren't going to help a lot ;) Dec 16 19:40:19 * koen|away ponders about applying the gtk1 xft patches Dec 16 19:41:42 Aha Dec 16 19:42:41 I'm trying to port siag office to OE, and it needs helvetica fonts Dec 16 19:43:07 I was NOT going to change the whole code to use misc-fixed Dec 16 19:58:20 /join #ubuntu Dec 16 20:20:53 re Dec 16 20:21:01 which package in OE provides makeinfo? Dec 16 20:21:13 texinfo? Dec 16 20:21:42 but we do not have it.. Dec 16 20:22:32 my build failed on binutils Dec 16 20:27:54 time to check how fscked it crowd in Polish tv is Dec 16 20:31:08 texinfo added into OE Dec 16 20:33:44 hrw|tv: is next OZ coming from oz354x branch? Dec 16 20:34:20 XorA|gone: I was thinking about releasing it from .dev but coredump prefer to release it from .oz354x rather Dec 16 20:34:41 hrw|tv: well I was just wondering if you wanted me to port some of my babies into the branch Dec 16 20:35:16 if you can.. Dec 16 20:35:58 hrw|tv: yes, now I have a collie again, I shall run that on OZ so I can test Dec 16 20:37:19 wb koen Dec 16 20:37:38 hrw|tv: and I can clear 1666 with a collie specific mplayer Dec 16 20:37:39 hey hrw|tv Dec 16 20:38:22 great Dec 16 20:38:32 XorA|gone: is collie the only machine affected? Dec 16 20:38:48 or armv4 at all? Dec 16 20:39:25 add a link to #1666 in the .bb Dec 16 20:40:18 woooooo Dec 16 20:40:23 loft support in mainline Dec 16 20:40:29 ~hail [g2] Dec 16 20:40:39 * ibot bows down to [g2] and chants, "I'M NOT WORTHY!!" Dec 16 20:41:54 koen: I plan to leave comment to say if anyone else finds the problem make the overide more generic. I sure as hell dont see it on armv4t Dec 16 20:42:33 * koen ponders sending XorA|gone a nbp Dec 16 20:42:52 or the other solution is to force LDFLAGS=-Wl,-z,-noexestack for whole collie build, but thats bound to break something Dec 16 20:42:53 s/a/lrg's/ Dec 16 20:45:08 koen: that armv4? Dec 16 20:45:27 pxa255 with 256MB ram Dec 16 20:45:39 :-D Dec 16 20:45:56 cf, pcmcia2 and sd Dec 16 20:46:10 but a slooooooooooooooooooooooooooooooooooooooooow framebuffer Dec 16 20:46:18 and a bootloader from hell Dec 16 20:47:41 XorA|gone: http://linuxtogo.org/~anonymous/netbook-booting.txt Dec 16 20:52:17 bloody hell.. Dec 16 20:52:41 hm. binutils-cross still going.. Dec 16 20:53:02 hrw|tv: did you push the makeinfo stuff yet? Dec 16 20:53:55 not into oe Dec 16 20:54:21 http://planetalinux.blog.br/debian - ARGH.. some fscker fetch posts from other blogs and add lot of google ads Dec 16 20:54:43 koen: need to add stuf into recipe first Dec 16 20:55:12 koen: erk Dec 16 20:55:53 and the port from 2.6.9 to 2.6.17 was done by someone without C knowledge Dec 16 20:55:55 (me) Dec 16 20:56:00 ;D Dec 16 20:56:04 koen: any gfx chip on it? Dec 16 20:56:11 koen: kernel hacker! Dec 16 20:56:12 XorA|gone: epson Dec 16 20:56:22 Xepson works with it Dec 16 20:56:44 koen: :-D Dec 16 20:57:09 who needs C knowledge, I though left and middle mouse buttons worked just as well Dec 16 20:57:20 hrw|tv: I will push a hacky git tree and install debian in a chroot to get bash Dec 16 20:57:55 and I will then blame tslib for errors why I overfill a jffs2 partition Dec 16 20:58:26 after that hrw will blog "do not follow koen way" Dec 16 20:58:29 koen: 'do not follow pavel machek way II' Dec 16 20:59:20 ~lart qemu for 'speed' Dec 16 20:59:21 * ibot tries to shut qemu up for 'speed' Dec 16 21:00:42 * hrw|tv want neo phone to be finally released. openmoko-community list is full of shit currently Dec 16 21:01:09 s/shit/utter shit/ Dec 16 21:01:22 koen: Why proper-tools depend on binutils-symlinks? for 'ar'? Dec 16 21:01:27 yes Dec 16 21:01:43 koen: I think that this recipe need comments inside Dec 16 21:01:59 feel free to add some :) Dec 16 21:03:06 description changed Dec 16 21:03:06 I'm thinking about doing DISTRO_EXTRA_RDEPENDS_ = "task-proper-tools" Dec 16 21:03:24 s/=/+=/ Dec 16 21:03:27 you will still have sh == ash Dec 16 21:03:36 yeah, bash needs to get fixed Dec 16 21:04:06 it is fixed already Dec 16 21:04:15 too many people fscked their systems Dec 16 21:04:25 think of 'bash on external media' Dec 16 21:04:41 * koen thinks of PEBKAC ;) Dec 16 21:04:47 binutils-symlinks does not install on busybox systems anyway Dec 16 21:05:01 ah, right Dec 16 21:05:07 koen: and module-init-tools are USELESS on 2.4 machines Dec 16 21:05:08 * koen uses --force-overwrite Dec 16 21:05:14 sucker Dec 16 21:05:18 could you file a bug for that? Dec 16 21:06:03 !oebug 1465 Dec 16 21:06:04 * * Bug 1465, Status: NEW, Created: 2006-10-11 08:30 Dec 16 21:06:05 * * koen(AT)dominion.kabel.utwente.nl: elfutils conflicts with binutils(-symlinks) Dec 16 21:06:06 * * http://bugs.openembedded.org/show_bug.cgi?id=1465 Dec 16 21:07:42 added comment, changed title Dec 16 21:09:29 btw - ipkg works better with busybox wget for me Dec 16 21:09:54 !oebug 1465 Dec 16 21:09:55 * * Bug 1465, Status: NEW, Created: 2006-10-11 08:30 Dec 16 21:09:56 * * koen(AT)dominion.kabel.utwente.nl: binutils(-symlinks) conflicts with other packages Dec 16 21:09:59 * * http://bugs.openembedded.org/show_bug.cgi?id=1465 Dec 16 21:10:34 ~seen [g2] Dec 16 21:10:50 [g2] is currently on #openmoko (5h 59m 49s) #nslu2-linux (5h 59m 49s) #openezx (5h 59m 49s) #openjtag (5h 59m 49s) #openslug (5h 59m 49s), last said: 'alphaone THX and sweet dreams!'. Dec 16 21:14:12 someone played with XEN? Dec 16 21:23:02 looks like it can partially resolve problem of slow qemu Dec 16 21:25:44 qemu --go-fast Dec 16 21:27:36 koen: I do not have qemu 'koen version' Dec 16 21:27:55 you mean g3nt00 version Dec 16 21:28:08 gcc -OMG -funroll-loops Dec 16 21:28:14 I do not know what kind of crap you use ;d Dec 16 21:28:52 Has anyone experimented with 2.6.19? Dec 16 21:29:03 RP: in which way? I use it on desktop Dec 16 21:29:05 (on a Zaurus) Dec 16 21:29:15 XorA did iirc Dec 16 21:32:24 It appears to be full of suspend/resume bugs :-/ Dec 16 21:35:20 RP: when I OOM on 2.6.19 it dies completely Dec 16 21:35:48 XorA|gone: Did you notice any other issues? Dec 16 21:35:57 RP: Well as it was a new image, that was a killer Dec 16 21:36:16 RP: I cant generate LOCALES as I use amd64 Dec 16 21:36:42 XorA|gone: Which qemu was that with? Dec 16 21:37:02 RP: all the ones in OE segfault when running qemu-arm Dec 16 21:37:12 RP: as do the ones in debian Dec 16 21:37:51 XorA|gone: I've not heard that said about the one in poky and I think its being used on amd64... Dec 16 21:38:10 RP: amd64 with 64bit userspace? Dec 16 21:38:14 koen: yes Dec 16 21:38:29 I use qemu on amd64 as well, but with a 32 bit userspace Dec 16 21:38:48 koen: Could have been 32 bit, I'm not sure. You'd have to ask mallum/tf Dec 16 21:39:11 but with gcc 4.1.0 and amd64 Im fairly stuffed Dec 16 21:39:49 binutils-cross BUILT Dec 16 21:40:01 * XorA|gone pops off again to check out Xfactor winner Dec 16 21:40:29 * koen mumbles something about having missed a concert of The Gathering again Dec 16 21:40:36 ~lart public transit Dec 16 21:40:36 * ibot puts on a hockey mask and jumps out at public transit Dec 16 21:41:42 glibc-initial started.. Dec 16 21:45:40 gcc-cross-initial now Dec 16 21:49:11 I wonder how easy that qemu image can be converted to a vmware thingy Dec 16 21:50:19 iirc without problem Dec 16 21:50:27 qemu-img support vmware Dec 16 21:50:48 vmware says it'll run on near-native speed Dec 16 21:54:44 qemux86 is fast Dec 16 21:55:17 haven't tried it yet Dec 16 21:55:36 but zecke and I always laugh about the debian linuxdevices article Dec 16 21:56:04 "OMG!!! a 4 ghz pentium can emulate a slug *faster* as realtime" Dec 16 21:57:10 RP: maybe on x86 with kqemu Dec 16 21:57:59 hrw: I've not tried kqemu yet but qemux86 was faster than on device Dec 16 21:59:03 RP: here qemu 0.8.2 needed ~hour to parse 4000 recipes Dec 16 22:00:41 hrw: I think we're talking cross purposes :-/ Dec 16 22:01:01 RP: qemu-i386 on amd64 Dec 16 22:06:15 RP: eds-dbus still breaks with a recent dbus-bindings tool :( Dec 16 22:06:23 RP: I'll hassle ross later about that Dec 16 22:08:25 2http://aseigo.blogspot.com/2006/12/patchserver.html Dec 16 22:08:25 http://aseigo.blogspot.com/2006/12/patchserver.html even Dec 16 22:08:39 cute Dec 16 22:11:52 ok. time to go sleep Dec 16 22:12:16 maybe during night I will get compiled gcc-cross Dec 16 22:14:14 and maybe glibc Dec 16 22:16:35 to get gcc-cross I need glibc Dec 16 22:19:17 ah right Dec 16 22:19:20 * koen sleepy Dec 16 22:19:29 * hrw too Dec 16 22:19:44 time to go sleep or I will get 4th or 5th bad night Dec 16 22:19:47 in a row Dec 16 22:20:27 RP: qemux86 from poky will get into .dev? please switch it to i686... Dec 16 22:21:20 RP: progear, guiness are i686 and both are angstrom supported (atleast during test releases) Dec 16 22:24:02 cu Dec 16 22:41:47 hrw|gone: qemux86 should get into .dev. It was done as i586, simply as that lead to less conflict between the usual i686 native packages and the i586 cross packages. I have no real fixed preference. Dec 16 23:00:35 Ifaistos: pin Dec 16 23:00:37 Ifaistos: ping Dec 16 23:01:27 Ifaistos: tyour icecc work needs a mention of what version you succesfully used. there are too many incompatible icecc versions around. Dec 16 23:02:02 * koen thinks OE should build icecc itself so people can match that version Dec 16 23:03:09 koen: me too, when i get it working I will also probably replace the shell script with the python script (which was almost correct already). Dec 16 23:03:48 koen: the logic of building and packaging the icecc compiler must be in the .bb Dec 16 23:04:04 right Dec 16 23:04:18 but how do you get that compiler to all the hosts? Dec 16 23:07:58 once the icedd daemon runs, it will deploy the tarballs remotely. Dec 16 23:08:40 koen: icecc.bb (new) will build iceccd(aemon). icecc.bbclass will package it for remote deployment (chroot tarballs). Dec 16 23:09:15 right Dec 16 23:09:29 koen: something else: could you look at this small snippet: http://pastebin.ca/281674 Dec 16 23:09:30 and how does it get to hosts without icecc(d) Dec 16 23:09:47 it can't, probably. Dec 16 23:13:14 hello Dec 16 23:16:44 likewise: that paste is indeed weird Dec 16 23:17:50 koen: the /usr/lib/libc++ part is strange, right? Dec 16 23:18:12 yes Dec 16 23:21:19 koen: I suspect it could be a hardcoded softlink somewhere. (if so, probably to: ../install/libstdc++/usr/lib/libstdc++.so.6) Dec 16 23:21:30 could be Dec 16 23:22:37 why does SRC_URI_append_collie not work? Dec 16 23:22:49 missing space? Dec 16 23:23:05 koen: that#ll be it :-) Dec 16 23:23:34 koen: you are too clever for me Dec 16 23:23:52 no, I have made that mistake far too many times Dec 16 23:26:51 SRC_URI_collie += "foo" is also a classic mistake Dec 16 23:28:40 heh Dec 16 23:28:49 "No more mr nice gaius" Dec 16 23:36:22 koen: :-D Dec 16 23:40:08 'night all Dec 16 23:41:26 * XorA|gone goes to sleep Dec 17 00:41:19 nite all **** ENDING LOGGING AT Sun Dec 17 02:59:59 2006