**** BEGIN LOGGING AT Tue Oct 03 02:59:57 2006 Oct 03 05:46:45 * v8jlene is back Oct 03 06:36:36 good morning all Oct 03 06:48:36 morning all Oct 03 06:58:52 did any mega speed improvements happen to bitbake last night? Oct 03 07:01:08 koen: I have a patch which seemingly gives a 40% improvment parsing poky Oct 03 07:01:37 let's start a packaged-parsing branch :) Oct 03 07:01:52 someone preparses everything and distributes the results Oct 03 07:02:20 If we got every combination of variables into the cache filename... Oct 03 07:03:39 I have a Japanese pdaxrom user emailing me pcmcia IDs as 2.6 doesn't work with his network card :) Oct 03 07:14:05 forward the ids to pcmcia-master hrw|gone Oct 03 07:14:07 :) Oct 03 07:21:09 morning Oct 03 07:27:12 hi hrw|work Oct 03 07:29:43 hey hrw|work Oct 03 07:30:10 morning Oct 03 07:30:59 hey XorA Oct 03 07:38:06 hrw|work: could you build http://www.angstrom-distribution.org/~koen/OE/build-stuff_0.1.bb and put the ipk in the oz feeds? Oct 03 07:38:30 hrw|work: that should silence all those whiny "installing gcc is too hard" morons Oct 03 07:42:49 hmm.. where did I put task-sdk-native.bb... Oct 03 07:44:44 * XorA whines that installing gcc is too easy instead then :-) Oct 03 07:49:02 morning all Oct 03 08:21:49 "No, we can't send you your paycheck this month because the only person capable of doing that is in holiday" Oct 03 08:21:53 * koen sighs Oct 03 08:32:09 koen: numpties Oct 03 08:35:12 hrw-pcmcia-ids-r5 sent to Linus Oct 03 08:37:44 hrw@bitbake:~/devel/build/3541-feed$ bitbake task-native-sdk Oct 03 08:47:18 CIA-9: ping Oct 03 08:48:52 03hrw 07org.oe.oz354x * r2b11f3f4... 10/ (1 conf/distro/openzaurus-3.5.4.conf): oz.conf: switch to binutils 2.16 - this version works OK on device Oct 03 08:49:09 forgot to push that thing Oct 03 08:49:13 03koen 07org.oe.dev * r38492872... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: don't use linux-oz as provider for hostap-modules for ipaq-pxa270 Oct 03 08:58:35 I have to prepare list of things to get to oedem... Oct 03 08:58:52 * koen looks at the big box with t-shirts Oct 03 08:58:59 that will be fun in the train Oct 03 08:59:37 yep - have to go to exchange to 'buy' some euros.. Oct 03 08:59:54 ~lart Poland for not being in Euroland yet Oct 03 08:59:54 * ibot drops a truckload of VAXen on Poland for not being in Euroland yet Oct 03 09:16:06 Depends: gcc-symlinks, g++-symlinks, binutils-symlinks, make, libc6-dev, flex, flex-dev, bison, gawk, sed, grep, autoconf, automake, make, patch, patchutils, diffstat, diffutils Oct 03 09:16:10 Source: file:///home/hrw/devel/oe/branch.oz354x/packages/meta/task-native-sdk.bb Oct 03 09:19:48 hi zecke Oct 03 09:20:34 ho Oct 03 09:20:35 03hrw 07org.oe.oz354x * re119a23d... 10/ (1 packages/meta/task-native-sdk.bb): task-native-sdk: one simple package to install all needed to build some stuff on device Oct 03 09:21:39 Depends: gcc-symlinks, g++-symlinks, binutils-symlinks, make, libc6-dev, flex, flex-dev, bison, gawk, sed, grep, autoconf, automake, make, patch, patchutils, diffstat, diffutils, libgcc-s-dev, libstdc++-dev Oct 03 09:24:47 03lrg 07org.oe.dev * r20a77bb6... 10/ (1 conf/machine/mx31ads.conf): mx31ads: Added VFP hard float support Oct 03 09:47:10 koen: when you will have time you can piss oesf people with 'just install task-native-sdk' package messages Oct 03 09:50:09 add it to the wiki Oct 03 09:50:35 "koen saus: 'ipkg install task-native-sdk or get larted'" Oct 03 09:51:13 hrw|work: could you pluck that libgcc_s fix to .dev? Oct 03 09:51:57 holy cow, ipkg need >300MB ram for ipkg -o . install Oct 03 09:54:30 koen: will Oct 03 10:05:33 03hrw 07org.oe.dev * rdaa17321... 10/ (1 packages/tasks/task-native-sdk.bb): Oct 03 10:05:33 task-native-sdk: one simple package to install all needed to build some stuff on device Oct 03 10:05:33 applied changes from 2b11f3f4bb122d1defb22c4476aa035bed79c754 (.oz354x) Oct 03 10:05:33 through e119a23d0f38807146843c508cad69961c2ccdf7 Oct 03 10:05:39 03hrw 07org.oe.dev * rd8437cda... 10/ (1 conf/distro/openzaurus-unstable.conf): oz.conf: switch to binutils 2.16 - this version works OK on device Oct 03 10:05:44 03hrw 07org.oe.dev * r3c6681af... 10/ (3 files in 2 dirs): Oct 03 10:05:44 gcc-cross: package libgcc-dev (named libgcc-s-dev :( Oct 03 10:05:44 applied changes from 515587ac4598d7ab78bd946632a31cf94fa79bf8 (.oz354x) Oct 03 10:05:44 through 33d804e082e0702e05818192cfab8a8610605d88 Oct 03 10:07:27 should be all Oct 03 10:07:49 thanks Oct 03 10:12:36 03koen 07org.oe.dev * r52221d20... 10/ (4 files in 2 dirs): gcc-cross: bump PR Oct 03 10:24:19 moin Oct 03 10:26:24 yo mickeyl Oct 03 10:27:26 hi mickeyl Oct 03 10:38:51 he mickeyl Oct 03 10:47:59 mickeyl: enjoying your holiday? Oct 03 10:48:07 hah Oct 03 10:48:16 i'm a freelancer. what is holidy? Oct 03 10:48:21 holiday, even Oct 03 10:48:56 mickeyl: holiday is a day when freelancers has more time for work because clients does not interrupt Oct 03 10:49:04 ah that's what you mean. Oct 03 10:49:05 so get back to work Oct 03 10:49:09 in that case i enjoy holiday ;) Oct 03 10:49:37 since my client is in Taiwan a german holidy really doesn't count anyway though :D Oct 03 10:50:09 yep Oct 03 10:50:27 my client will wait until after oedem Oct 03 10:53:32 RP: why there is no versioned git in OE? you added snapshot one only Oct 03 10:53:43 apropos Oct 03 10:53:51 is it really not possible to checkout a version by date using git? Oct 03 10:53:56 even svn can do taht Oct 03 10:54:00 * koen has no clients :) Oct 03 10:54:39 according to Polish tax law I will work in US Oct 03 10:55:14 mickeyl: dates don't make sense in distributed systems with a dag Oct 03 10:55:27 mickeyl: you can't Oct 03 10:55:28 good point Oct 03 10:55:34 it sucks a bit anyway though Oct 03 10:55:55 mickeyl: you first need to get all revs to find something by date Oct 03 10:55:56 even a fuzzy mapping from a date to a revision would help Oct 03 10:56:51 mickeyl: SHA1 hash does not have this property ;) Oct 03 10:57:05 darn Oct 03 10:57:07 I need git in OE.. so will try to add 1.4.2.3 Oct 03 10:57:31 git over http.. I'm curios does montavista has git over http.. Oct 03 10:58:16 http://127.0.0.1:8000/bonsai/rev/509/ yeah my first webapp works :) Oct 03 10:58:16 mickeyl: throw away your cvs (and svn) paradigma Oct 03 10:58:16 you can specify a revision Oct 03 10:58:16 and revision don't suffer from time resets after reboot Oct 03 10:58:16 mickeyl: mtn can checkout using dates :) Oct 03 10:58:31 mickeyl: I wish OE devs would use http://venge.net/monotone/wiki/DaggyFixes Oct 03 10:58:42 zecke: doesn't work over here :) Oct 03 10:59:05 koen: you didn't use daggy fixes yesterday. Be a good role model Oct 03 11:00:13 zecke: right :) Oct 03 11:00:54 parsing 4K recipes takes 6 minutes... Oct 03 11:01:11 4 minutes with RP's patch Oct 03 11:01:41 hmm.. time to dig celinux-test distro and use testlab Oct 03 11:02:25 what ABI is celinux-test going to use? Oct 03 11:02:45 normal Oct 03 11:02:51 'normal'? Oct 03 11:02:54 OABI Oct 03 11:03:06 what OABI? Oct 03 11:03:14 last celin Oct 03 11:03:25 hardfpa, softfpa, hardvfp, softvfp? Oct 03 11:03:29 koen: now they use familiar distro Oct 03 11:03:57 right, so celinux will be obsoleted once debian/ARM switches Oct 03 11:04:17 Debian will not be CE compatible then? Oct 03 11:04:26 heh Oct 03 11:04:38 I think debian doesn't care about celinux Oct 03 11:05:01 does anyone care about celinux, except for the managerfolk? Oct 03 11:05:31 mvista doesn't care, wind river doesn't care, etc Oct 03 11:06:54 there is little value for community in celinux if they use such old systems Oct 03 11:07:41 koen: OABI has one plus above EABI: it works Oct 03 11:07:41 hrw|work: EABI works as well Oct 03 11:07:41 hrw|work: nokia is shipping a product with it Oct 03 11:07:58 well, "EABI" Oct 03 11:07:58 koen: then why angstrom images has so many problems? Oct 03 11:08:31 hrw|work: I'm only aware of shm problems with X, and that's a glibc problem Oct 03 11:08:52 koen: you tracked it down more? Oct 03 11:09:02 XorA: o-hand did Oct 03 11:09:08 looks like one of my first new doc addons will be long awaited 'how to use OE to build native tools for development' Oct 03 11:09:09 koen: cool, patches? Oct 03 11:09:39 XorA: not really, they have a csl glibc that works Oct 03 11:09:51 and I tend to avoid using CSL stuff like the plague Oct 03 11:09:54 koen: ah yes, I had meant to import that into OE Oct 03 11:10:23 hrw|work: very cool. Oct 03 11:10:41 there is still a HUGE bug in OE Oct 03 11:10:57 libgcc has the wrong architecture Oct 03 11:11:18 hrw|work: i have this problem in the current project as well. it's nearly impossible to get all the (x86) versions right on the development machine, so i need to give the developers an x86-chroot'able directory with the proper versions Oct 03 11:11:46 mickeyl: its easier to build apps with OE then waiting for admin to install them Oct 03 11:11:55 hrw|work: no kidding, right Oct 03 11:12:35 mickeyl, hrw|work: http://bugs.openembedded.org/show_bug.cgi?id=1265 Oct 03 11:13:53 hmm Oct 03 11:19:58 PACKAGE_ARCH would be armv5te in that situation? Oct 03 11:21:02 hrw|work: iirc no, which is the problem Oct 03 11:22:20 okay, later Oct 03 11:22:29 time for the gym Oct 03 11:26:24 koen: btw. EABI need >2.6.16 and currently we have working 2.6.12-rc for omap and non-working recent one Oct 03 11:26:48 hrw|work: it needs 3 patches which apply against 2.6.12 Oct 03 11:27:34 good to know Oct 03 11:34:22 koen: poky doesnt seem to have a csl glibc 2.4 Oct 03 11:35:12 XorA: correct, it's a patched to hell 2.3.x Oct 03 11:35:21 koen: ah right Oct 03 11:36:18 XorA: check youy dp@ mail in a few secs Oct 03 11:37:12 XorA: https://garage.maemo.org/scm/?group_id=54 might interest you ass well Oct 03 11:37:16 s/ass/as/ Oct 03 11:37:52 s/you/your Oct 03 11:38:16 :D Oct 03 11:39:19 koen: not another fork :-( Oct 03 11:46:21 :/ Oct 03 11:46:32 * mickeyl shakes head Oct 03 11:49:43 fixchicken.patch makes me chuckles everytime Oct 03 11:50:07 heh.. 30 minutes of fetching snapshot for 0.30 to get version which need to be pull synced Oct 03 11:50:33 ROTFL! Oct 03 11:50:39 [hrw@host1 org.openembedded.dev]$ mtn pull monotone.openembedded.org Oct 03 11:50:39 mtn: doing anonymous pull; use -kKEYNAME if you need authentication Oct 03 11:50:39 mtn: connecting to monotone.openembedded.org Oct 03 11:50:39 mtn: finding items to synchronize: Oct 03 11:51:16 329 revs to pull Oct 03 11:51:19 the firewall isn't blocked? Oct 03 11:51:39 XorA: hehe Oct 03 11:51:44 XorA: that was my patch Oct 03 11:51:54 koen: atleast not monotone Oct 03 11:52:17 hrw|work: I am thinking something about CELF and it contains the word 'clueless' Oct 03 11:52:32 so CELF was wasting our bandwidth for nothing Oct 03 11:53:56 sent mail Oct 03 11:58:52 anyone got a second? i'm trying to 'make gconfig' from within the kernel source dir that bitbake got for me to see what i need to add to the kernel bb file. i've installed gtk, glib, and libglade (dependencies) via bitbake, but when i try 'make gconfig' it says it can't find them. do i have to figure out where bb put them and add them to the path? Oct 03 11:59:33 bigd0g: why not use menuconfig? Oct 03 11:59:45 koen: what's that do? Oct 03 12:00:03 the same, but without using gtk/qt Oct 03 12:00:18 'make menuconfig' requires only ncurses Oct 03 12:00:31 hrm. well, my application is going to require gtk sooner or later :) Oct 03 12:00:41 and works even on dump terminals Oct 03 12:00:42 it's a PyGTK-based gui Oct 03 12:00:48 dumb I mean Oct 03 12:00:57 right, but that will run on your target Oct 03 12:01:04 'make config' runs on your host Oct 03 12:02:20 hrm. so if i do 'make menuconfig' and then look at the conf files, i should be able to pull what i need to add to the kernel bb file to get display drivers running in the kernel build? Oct 03 12:04:19 oh, i've seen this (menuconfig) before. hrm. Oct 03 12:05:06 ok, i see what you were trying to tell me. sorry. Oct 03 12:13:58 someone know why we use fakeroot 1.2.13? Oct 03 12:14:09 other then 'because no one updated it'? Oct 03 12:14:33 I think you answered it yourself Oct 03 12:15:06 NOTE: fetch http://openzaurus.org/mirror/fakeroot_1.2.13.tar.gz Oct 03 12:15:13 worst possible place... Oct 03 12:23:35 ~lart IRQs Oct 03 12:23:35 * ibot slaps a compatible dib on IRQs's head Oct 03 12:31:15 | /home/hrw/devel/build/dev/tmp/work/armv5te-angstrom-linux-gnueabi/fakeroot-1.5.10-r0/fakeroot-1.5.10/configure: line 19223: LT_INIT: command not found Oct 03 12:31:34 LT_INIT and LT_LANG(C) block me from using fakeroot 1.5.10 Oct 03 12:48:46 ahah, smdk2440 booted OZ image :-) Oct 03 12:54:38 is "linux-omap1-2.6.12-rc2" as stable as "linux-omap1-2.6.x+git" ? Oct 03 12:54:53 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Oct 03 12:56:09 bigd0g, kernels that follow git are inherently unstable Oct 03 12:56:31 git isn't booting on my OSK right now, when I get in, I will boot a version build with LL debug Oct 03 12:57:08 ok, then i guess i'll stick with what i have Oct 03 12:57:22 from what i can tell, it seems that everything needed for framebuffer write support was enabled Oct 03 12:57:31 some of it was set to "modularize", though Oct 03 12:57:38 once I get a recent git that boots, I am going to see if we can build from a fixed git rev Oct 03 12:57:40 ok Oct 03 12:57:48 i'm not sure how that effects things -- the "modularize" part Oct 03 12:58:00 I would prefer display be modules since not all osk have displays Oct 03 12:58:33 right. so now, i guess the question is... how do i enable the module on my board? Oct 03 12:58:49 modprobe :) Oct 03 12:59:20 I am not an expert at the boot stuff and getting modules loaded early Oct 03 13:01:36 hrm. Oct 03 13:02:49 i guess i'll have to find a newbie guide for loading this module correctly Oct 03 13:02:55 I am going to test soon Oct 03 13:03:04 boot it and do a modprobe Oct 03 13:04:10 ok. i'm getting things set up now Oct 03 13:11:41 re Oct 03 13:13:37 hrm... "modprobe: could not parse modules.dep" Oct 03 13:14:00 depmod -ae? Oct 03 13:14:39 uh oh... "WARNING: Couldn't open directory /lib/modules/2.6.18-rc6-omap1: No such file or directory" Oct 03 13:15:57 bigd0g: lack of KERNEL_RELEASE Oct 03 13:18:46 hrm. not sure what your hint is pointing me towards ;) Oct 03 13:20:40 bigd0g, wgen did you build the kernel? Oct 03 13:21:18 ~lart arm926ejs Oct 03 13:21:18 * ibot strangles arm926ejs with a 9-pole serial cable Oct 03 13:22:11 Crofton: looks like 06-Sep Oct 03 13:22:20 time to rebuild, methinks Oct 03 13:22:49 hrw|work: with the new pkgmaps kergoth added we might be able to see which gcc version we are using Oct 03 13:23:17 TARGET_CC_ARCH_omap5912osk = "-march=armv5te -mtune=arm926ejs" Oct 03 13:23:45 workaround in distro config Oct 03 13:23:50 heh Oct 03 13:24:56 any good reason for using gcc 3.x? Oct 03 13:25:00 lots of c++ apps? Oct 03 13:25:07 koen: machine, distro order force me to it. distro, machine would allow to get it in other way Oct 03 13:25:40 koen: KISS method? Oct 03 13:25:59 with all the cross building bugs in 3.x? Oct 03 13:26:10 ~lart opensync Oct 03 13:26:10 * ibot takes out a seltzer bottle and sprays opensync in the face. You know, one of those old-school seltzer bottles clowns have? Yeah those. Anyway, consider yourself spritzed Oct 03 13:29:58 bigd0g, I think the bb file has KERNEL_RELEASE, but current git may not boot on the OSK ... Oct 03 13:30:31 methinks I should change defconfig for git build to leave low level debug on Oct 03 13:31:03 Crofton: you don't think my build is just outdated? Oct 03 13:31:39 I think we added the kernel release var Oct 03 13:31:44 that should help with modules Oct 03 13:31:49 BUT Oct 03 13:31:58 changes in git tree may have messed up OSK Oct 03 13:32:01 ..... Oct 03 13:32:38 I'll know more in a bit, just need to get trash ot and go over to school and flash kernel with low level debug enabled so I can see where it hangs Oct 03 13:34:28 ok, noted. Oct 03 13:39:10 fix one issue, another crops up Oct 03 13:39:15 bbiab Oct 03 13:39:19 heh.. I forgot how fast glibc package is without binary locales Oct 03 13:40:23 news flash, independent confirmation git is busted on omap linux list Oct 03 13:41:21 bigd0g, I just forwarded you an emial from the omap list Oct 03 13:41:33 bye again Oct 03 13:45:37 hrm. oh, dear. broken stuff is not good Oct 03 13:45:54 are some of you compile redboot/ecos with OE ? Oct 03 14:03:24 ~hail RP for git in OE Oct 03 14:03:26 * ibot bows down to RP for git in OE and chants, "I'M NOT WORTHY!!" Oct 03 14:04:18 would be nice if fetcher would check does fetch command is available and if not then add proper tool into DEPENDS and force build of it Oct 03 14:04:38 re Oct 03 14:04:47 what is the syntax in python for "not equal" Oct 03 14:05:02 not a == b Oct 03 14:05:02 " not equal " Oct 03 14:05:16 simpel a != b does not work? Oct 03 14:05:19 for two strings? Oct 03 14:05:22 dunno Oct 03 14:05:25 the same as other languages, but with more annoying indenting ;) Oct 03 14:05:29 guy behind me is asking :) Oct 03 14:05:46 thanks guys :) Oct 03 14:05:51 he went away Oct 03 14:08:15 i seem to think he'd have to compare the strings first... but meh Oct 03 14:08:55 bother Oct 03 14:09:04 I seem to have lost u-boot again Oct 03 14:09:23 doh Oct 03 14:09:25 anyone shed any light on this hostap-modules wierdness? Oct 03 14:09:28 http://www.rafb.net/paste/results/xiO5CR53.html Oct 03 14:10:23 XorA: do not build out-of-kernel hostap-modules for >2.6.15 kernels Oct 03 14:10:33 hrw|work: ah of course Oct 03 14:10:44 hrw|work: cheers Oct 03 14:10:57 wew u-bbot still home Oct 03 14:12:33 Crofton: want me to send you my build of u-boot or something? Oct 03 14:12:46 it was another problem Oct 03 14:12:58 network connection to machine with serial port is flaky Oct 03 14:13:19 last time this happened it took me longer to realize I had lost connection to minicom :) Oct 03 14:13:52 stupid newer machine has *no* serial port Oct 03 14:13:53 file transfer over minicom to the osk was apparently not that good Oct 03 14:13:58 that's why i switched to c-kermit Oct 03 14:14:02 which works beautifully Oct 03 14:14:09 I use tftp for file transfer :) Oct 03 14:15:34 i'll have to set that up eventually Oct 03 14:16:00 much faster I suspect Oct 03 14:21:05 * mrz80 is back (gone 01:26:13) Oct 03 14:37:51 mickeyl: does rxvt-unicode build for you with x86? Oct 03 14:39:04 ~lart ipkg Oct 03 14:39:04 * ibot judo chops ipkg Oct 03 14:39:08 DiskStation> ipkg -force-depends install ./strace_4.5.12-r0_armeb.ipk Oct 03 14:39:08 Installing strace (4.5.14-3) to root... Oct 03 14:39:08 Downloading http://ipkg.nslu2-linux.org/feeds/optware/ds101/cross/stable/strace_4.5.14-3_armeb.ipk Oct 03 14:39:33 koen: let me try building within the next 2h Oct 03 14:44:05 NAiL: ipkg will always pick the latest, unless you use -force-downgrade Oct 03 14:44:48 koen: I did that too Oct 03 14:45:00 --force-downgrade --force-depends Oct 03 14:45:10 still fetched the stupid ipkg from the feed :-\ Oct 03 14:45:15 FYI: the debian/ARM dudes are now staring at an OE-build scrolling by so they can start with the EABI port :) Oct 03 14:47:44 woot Oct 03 14:49:27 mickeyl: except they get paid by arm and have arm SMP hardware sponsored Oct 03 14:49:55 s/arm/ARM/ Oct 03 14:50:33 koen: maybe we should rant about Debian and compare them to Canonical? Oct 03 14:52:41 zecke: I'm thinking how to get good PR out of it Oct 03 14:54:58 koen: post about it on LinuxDevices would be great. Oct 03 14:57:11 the debconf video of emdebian has some nice quotes about OE as well :) Oct 03 14:58:29 koen join #oe-marketing Oct 03 14:59:18 pft Oct 03 14:59:40 suihkulokki: Oct 03 15:00:15 suihkulokki: you are from movial? wow more and more dvb-h dudes turn up Oct 03 15:00:56 zecke: I'm from movial yes, but I've never done anything dvb-h related Oct 03 15:03:23 the arm<->eabi<->debian<->alephone deal annoys me, but mostly because they guy doing doesn̈́'t seem to accomplising anything Oct 03 15:04:14 it doesn't annoy me at all Oct 03 15:04:30 suihkulokki: "leveraging the community" Oct 03 15:04:39 * koen practices his manager speak Oct 03 15:05:10 suihkulokki: you forgot to include csl in your annoyance diagram ;) Oct 03 15:07:40 * chouimat gives a 4.5 out of 10 to koen ... need improvement :) Oct 03 15:08:24 :) Oct 03 15:08:55 koen I have a job interview in 2.5 hours ... and I'm cleaning my apartment ... maybe something is wrong with me :) Oct 03 15:09:44 koen: well CSL actually manages to do usefull stuff :) Oct 03 15:10:39 * koen -> food Oct 03 15:12:10 zecke: Did you give that patch I posted last night any more thought? Oct 03 15:12:39 zecke: I don't think it should have any problems with CoW as we wipe the cache upon writes... Oct 03 15:12:46 RP: I only wondered when you reset the cache Oct 03 15:12:59 zecke: Any write to the data Oct 03 15:13:20 zecke: Not optimal but should be safe Oct 03 15:13:31 RP: the questions is which 'cache' with wipe :) Oct 03 15:14:05 RP: I would need some more context to decide on this issue. But besides that I think your patch is safe Oct 03 15:15:18 zecke: ok. I might commit it and see what breaks ;-) Oct 03 15:15:33 zecke: Can you see why we'd spend 15% of our time in /bin/bitbake:713(parse_bbfiles) ? Oct 03 15:16:00 I'm tired ATM Oct 03 15:16:08 zecke: ok, np :) Oct 03 15:16:09 yeah, I hate ATM machines too Oct 03 15:16:19 * CosmicPenguin shows himself out Oct 03 15:16:35 * RP hands CosmicPenguin is coat Oct 03 15:16:40 s/is/his/ Oct 03 15:16:41 gah Oct 03 15:18:52 RP: 15% from parse_bbfiles or within parse_bbfiles Oct 03 15:19:06 RP: but my thinking would be, if we spend 15% there everything is fine :) Oct 03 15:19:31 http://www.bash.org/?694821 Oct 03 15:19:32 seems to be a little bug in umount initscript Oct 03 15:19:34 the line: Oct 03 15:19:34 umount -f -a -r Oct 03 15:19:46 just gives a busybox error Oct 03 15:19:51 incorrect usage Oct 03 15:20:32 hmm I'm always tired after eating Oct 03 15:20:59 zecke: eat less Oct 03 15:21:45 chouimat: I will starve when doing this :) Oct 03 15:22:21 zecke: the other solution is not to wait when you're starving to eat ... because all your remaining energy goes into digestion Oct 03 15:24:04 I think it should be just `umount -a -r` Oct 03 15:24:14 since -f doesn't seem to be supported by busybox Oct 03 15:24:48 koen: why do the debian people need a SMP ARM when they could use their Quad Xeon's which are so much faster than a NSLu2? Oct 03 15:25:24 cu Oct 03 15:26:20 zecke: cross-compile is EVIL!! Oct 03 15:28:02 actually... scratch that... busybox umount doesn't work fullstop! Oct 03 15:29:24 tkp: angstrom uses util-linux-(u)mount as the initscripts dont work with busybox Oct 03 15:29:52 XorA: ah ok.. nice to know! Oct 03 15:30:21 I seem to remember having to use util-linux-mount already to fix some otherr problems... Oct 03 15:30:32 * tkp installs util-linux-umount Oct 03 15:30:56 I'm sure that a patch to busybox would be gladly accepted Oct 03 15:31:03 I hate having to install the full utilities for stupid stuff like that Oct 03 15:31:26 anyother areas of busybox that are know to be busted? Oct 03 15:31:28 CosmicPenguin: the problem with patching busybox is at what point do you stop? when busybox reaches the size of the original utils? :-) Oct 03 15:31:51 heh - they're discussing that right now Oct 03 15:32:46 Its a fine line - if the script is using some stupid feature added to the fully leaded version for no good reason then you re-write the script Oct 03 15:33:04 I patched ps -e x support in (well it ignores the args but default behaviour is -e x) but I dont like the idea of that patch so it sits on my machine Oct 03 15:33:21 but if the script appears sane, then it might just be that somebody was lazy in busybox land Oct 03 15:34:34 mount.c got some love in svn lately Oct 03 15:35:44 chouimat: no, qemu on Quad Xeon ;) Oct 03 15:37:56 zecke: :) Oct 03 15:40:53 anyone have any suggestions for enabling low level debug so I can see messages on the console at boot time? Oct 03 15:41:13 I swear I changed my defconfig, but nothing is coming out Oct 03 15:58:29 why can't I get low level debug messages out of my OSK? Oct 03 16:02:33 can I build a specific git rev? Oct 03 16:12:38 good evening Oct 03 16:14:16 hey pH5 Oct 03 16:16:30 pH5: I accidentily made llh 2.6.18 standard for angstrom yesterday Oct 03 16:16:55 http://www-128.ibm.com/developerworks/linux/library/l-lobmediamvp.html?ca=dgr-lnxw07LinuxMediaMVP Oct 03 16:17:49 Hmmn, no HDTV support though... Oct 03 16:18:26 a+ Oct 03 16:19:50 koen: I noticed :) Did anything else break so far? Oct 03 16:19:59 Crofton: you can. See the oz git kernel .bb Oct 03 16:22:05 yeah Oct 03 16:22:09 I checked :) Oct 03 16:22:11 pH5: not that I can see Oct 03 16:22:26 I am having t binary search omap git for a working kernel Oct 03 16:22:42 the string is the commit string? Oct 03 16:22:54 I ahve commit, tree, and parent Oct 03 16:24:48 Crofton: 'git bisect'? Oct 03 16:25:49 Crofton: why not point S to a local checkout? Oct 03 16:33:50 koen:\\ Oct 03 16:34:12 Why does angstrom-gpe-image create libqpe-opie? Oct 03 16:36:55 it doesn't over here Oct 03 16:37:04 you probably have a bogus provides for sdl Oct 03 16:37:11 ok Oct 03 16:37:43 * koen doesn't even have qmake in his OE tree Oct 03 16:39:43 Newbie question: Why is there a qmake.bbclass but not (GNU) make.bbclass? Is it built-in, or somewhere else? Oct 03 16:40:35 likewise: afaik make 'just works' without any tricks needed Oct 03 16:41:54 koen: Thus for cross-compilation, bitbake sets env. variables such as CC, LD (for cross-compilation), and the Makefile should use these?] Oct 03 16:42:05 afaik, yes Oct 03 16:49:54 likewise: although a make.bbclass could sed out stuff like 'CC = gcc' 'LD = ld', etc Oct 03 16:50:44 koen: yes, for some packages that hard-code their tools, this would be an option Oct 03 16:51:14 koen: I was just thinking about what the easiest way is, to add some custom applications to a OE build, relatively easy. Oct 03 16:51:18 how you pass in variables into gnu make is a giant pain in the ass Oct 03 16:51:25 we cheat, really Oct 03 16:52:00 kergoth: where can I find how make gets its variables in BitBake/OE? Oct 03 16:52:16 read bitbake.conf Oct 03 16:52:20 its one line. Oct 03 16:52:29 well, mostly. Oct 03 16:52:36 base.bbclass too. Oct 03 16:52:50 we call make with the ${EXTRA_OEMAKE} var as its commandline arguments Oct 03 16:52:57 which are -e MAKEFLAGS="" by default Oct 03 16:53:09 which means obey the environment over the versions set in files, which make doesnt do by default Oct 03 16:53:22 but it resets makeflags, which means that env override only happens at toplevel, not in submakes Oct 03 16:54:30 kergoth: a submake is a make called from within a Makefile? Oct 03 16:54:39 right Oct 03 16:54:49 ok, thanks Oct 03 16:55:14 example situation: toplevel make sets cflags, submake appends something to cflags... if you use -e without resetting makeflags, that appended value wont be added to the cflags we supplied Oct 03 16:58:50 kergoth: good, I will paste this in stone and sleep on it (or something) Oct 03 16:59:02 Where can i get this file. Angstrom-gpe-image needs it. And it is not in the ewi546.ewi.utwente.nl or oesources. : file: git_anongit.freedesktop.org.xorg.proto.calibrateproto.tar.gz Oct 03 16:59:40 s/paste/chisel Oct 03 16:59:44 hehe Oct 03 16:59:47 ~lart gmake Oct 03 16:59:47 * ibot takes out a cattle prod and gives gmake a good jolt Oct 03 17:02:37 goxboxlive: it gets that from a git repository at freedesktop.org Oct 03 17:02:55 hmm, doesnt workfor me. Oct 03 17:03:37 koen: http://pastebin.ca/190031 Oct 03 17:04:12 I am going to catch a movie sneak preview at the cinema. Thanks for all the fish^H^H^H^H popcorn guys. Oct 03 17:05:29 goxboxlive: git is not installed correctly on your system Oct 03 17:05:38 i c Oct 03 17:06:05 i am running over ssh is that a problem? There is no gui is there? Oct 03 17:07:45 your debian has 'git' tied to 'gitfm' Oct 03 17:07:45 you want the linus' git, not the filemanager Oct 03 17:07:53 as the error message indicates :) Oct 03 17:13:26 thx :-) Oct 03 17:15:16 ~lart po4a Oct 03 17:15:16 * ibot chops po4a in half with a free Solaris 7 CD Oct 03 17:16:23 "Unable to get Terminal Size. The TIOCGWINSZ ioctl didn't work. The COLUMNS and LINES environment variables didn't work. The resize program didn't work. at /usr/lib/perl5/Term/ReadKey.pm line 362." Oct 03 17:16:36 * koen wonders why it needs terminal size to translate stuff Oct 03 17:17:17 Crofton: how's it going? Oct 03 17:32:00 03koen 07org.oe.dev * rdf1b9d9f... 10/ (7 files in 3 dirs): dpkg: sync with poky Oct 03 17:32:30 03koen 07org.oe.dev * re59abee0... 10/ (12 files in 4 dirs): apt: sync with poky Oct 03 17:45:29 03koen 07org.oe.dev * r0280346d... 10/ (7 files in 2 dirs): Oct 03 17:45:29 apt: update to 0.6.46, hack do_install to work around missing .so files Oct 03 17:45:29 * not tested on device Oct 03 17:47:04 bigd0g, I am building tags from git util I found one that compiles and boots ... Oct 03 17:47:28 sounds like fun :/ Oct 03 17:49:38 Crofton: will you be available tomorrow to help me try and work through getting LCD support on my board? i'm not finding a whole lot of documentation for this. Oct 03 17:50:21 sorry, I'll be travlling tomorrow Oct 03 17:50:37 there is a kernel config option for Mistral QVGA support Oct 03 17:50:48 in the osk config section I think Oct 03 17:50:53 you have acustom board? Oct 03 17:51:01 yeah, custom board + lcd Oct 03 17:51:13 connects via expansion port C Oct 03 17:51:27 I wouldn't be much help Oct 03 17:51:35 are you on the linux-omap list? Oct 03 17:51:49 no... where is it? Oct 03 17:53:43 bigd0g, look at linux.omap.com Oct 03 17:53:47 hi all Oct 03 17:57:51 Crofton: ok, i've subscribed and will ask there Oct 03 18:01:22 time to go. lot of work ahead. hrm... Oct 03 18:03:43 good nite Oct 03 18:13:54 "to empower innovative action-items, seize extensible ROI or facilitate scalable architectures" Oct 03 18:14:46 error, buzzword overload Oct 03 18:14:58 right :) Oct 03 18:15:01 that's from http://www.wantstofly.org/~buytenh/intel.suxorz Oct 03 18:15:54 kergoth: btw, your deps code works pretty neat: http://www.angstrom-distribution.org/repo/?action=details&pnm=gtk%2B-dev Oct 03 18:16:28 nice Oct 03 18:16:49 one thing we should fix.. the -dev version of a package should probably Depends on the specific version of the main package Oct 03 18:20:42 "gtk+, gtk+ (>= 2.8.9)," Oct 03 18:20:52 it seems it does both Oct 03 18:21:08 hehe Oct 03 18:21:08 weird Oct 03 18:21:11 not all packages are though Oct 03 18:21:22 see http://www.angstrom-distribution.org/repo/?action=details&pnm=libcairo-dev for example Oct 03 18:21:25 right: http://www.angstrom-distribution.org/repo/?action=details&pnm=libgcc-s-dev Oct 03 18:21:25 not sure whats up there Oct 03 18:21:29 yeah Oct 03 18:22:59 kergoth: do you know how dpkg gets its list of supported architectures? Oct 03 18:23:10 e.g. the dpkg version of /etc/ipkg/arch.conf Oct 03 18:23:14 hehe, libcairo-dev depends on libcairo-dev? thats always fun Oct 03 18:23:37 nope, i ended up cheating when i added dpkg images, i broke up the feeds and pinned them with priorities based on our archs list Oct 03 18:23:41 got the job done :) Oct 03 18:24:22 I noticed that apt and/or dpkg doesn't really know about architectures Oct 03 18:24:42 apt happily downloaded the 'wrong' archs with maemo-sardine Oct 03 18:24:53 it barfed at the moment dpkg tried to install it Oct 03 18:25:06 hmm thats odd, dpkg usually freaks out if i try to install packages with the wrong arch, and apt says the package doesnt even exist at all if its the wrong arch Oct 03 18:25:09 * kergoth ponders Oct 03 18:25:42 it was armel vs gnueabi-arm-linux Oct 03 18:25:48 so that might have hit some edge case Oct 03 18:26:07 --force-architecture is scary Oct 03 18:26:29 hehe Oct 03 18:26:43 did you look at the dpkg/apt bits in poky? Oct 03 18:26:54 * kergoth should push his latest stuff back to oe one of these days Oct 03 18:27:50 kergoth: see http://www.openembedded.org/aggregator/sources Oct 03 18:28:27 that's only the recipes, not the bbclass changes Oct 03 18:28:28 ah, excellent Oct 03 18:28:34 so all that hasnt been pushed probably is the image stuff Oct 03 18:28:48 it has some weirdnesses left, but it's a start Oct 03 18:29:13 especially since debian/ARM is using OE built images to bootstrap their EABI port :) Oct 03 18:29:38 they are? hah Oct 03 18:29:40 thats cool Oct 03 18:31:45 unfortunately the rootfs_deb.bbclass isnt as clean as it could be Oct 03 18:32:17 ended up patching apt-native to not configure packages, then the class renames the .dpkg-new versions of conffiles to the correct filenames, and manually runs the {pre,post}inst Oct 03 18:32:28 RP and I have been pondering about absorbing ipkg-build into OE and share it between ipkg and dpkg Oct 03 18:33:00 RP: I can't find why my .dev bootstrap image doesn't set the root partition to rw, here the initscript comparaison from latest oz release and .dev --> http://pastebin.ca/190118 Oct 03 18:33:08 but first we want depchains support in packaged-staging Oct 03 18:33:40 cedric: try adding util-linux-mount to your rootfs Oct 03 18:35:13 koen: how come it's not part of the bootstrap by default? Oct 03 18:35:42 cedric: because busybox provides mount Oct 03 18:35:54 but it seems the mount applet in busybox 1.2.x is a bit broken Oct 03 18:37:09 so busybox 1.2.1-r9 has the same issue? Oct 03 18:37:48 yes Oct 03 18:38:58 koen: using ipkg-build for both wouldnt be too bad, would avoid building dpkg-native until the image creation.. but you'd have to account for the fact that dpkg is pickier about its control file contents than ipkg is Oct 03 18:39:37 kergoth: yeah, that's why it's at the bottom of the TODO :) Oct 03 18:39:57 hehe Oct 03 18:40:37 koen: what's the easy way to add it to the bootstrap rootfs? I can't install an ipk because it's read-only... Oct 03 18:41:03 add it to your image RDEPENDS Oct 03 18:50:19 koen: so in task-bootstrap.bb, I'm adding util-linux? Oct 03 18:51:03 util-linux in DEPENDS and util-linux-mount in RDEPENDS Oct 03 18:51:13 hrw: wb Oct 03 18:51:15 hi Oct 03 18:51:34 I'll look from time to time - impress is waiting Oct 03 18:51:58 struggling with the angstrom-gpe-image. What's this then: http://pastebin.ca/190132 Oct 03 18:52:01 NOTE: package bootstrap-image-1.0-r0: task do_build: completed Oct 03 18:52:11 hi all Oct 03 18:52:16 first build at celf Oct 03 18:52:16 hey florian Oct 03 18:52:26 shouldnt be necessary to add util-linux to depends afaik... it should pick up that util-linux provides util-linux-mount and build that Oct 03 18:52:33 goxboxlive: right, I'm seeing the same. Someone broke xcalibrate for xorg a while ago Oct 03 18:52:38 course who knows waht bitbake version that requires Oct 03 18:52:38 heh Oct 03 18:53:09 kergoth: 1.6.0 should work iirc Oct 03 18:53:16 koen: So it's not me this time. Good to hear :-) Is there a easy way to override this, or do i have to wait for the fix ? Oct 03 18:54:09 goxboxlive: I have no idea, I just noticed that error an hour ago Oct 03 18:54:15 i c Oct 03 18:56:15 ~botmail for psokolovsky: I still get errors when building xcalibrate_cvs.bb, any idea? Oct 03 19:01:58 03hrw 07org.oe.dev * r617c0aa8... 10/ (1 packages/tasks/task-base.bb): task-base: add usbutils for usbhost machines Oct 03 19:05:29 I like idea of task-base Oct 03 19:05:37 me too! Oct 03 19:05:49 very clean images Oct 03 19:06:50 Crofton: I want to move PREF_VER of linux-omap1 from omap5912osk into openomap distro - so celf guys will be able to build 2.6.12-rc which they use Oct 03 19:07:01 ok Oct 03 19:07:33 can I override in my local.conf? Oct 03 19:07:43 friggin git kernel is driving me nuts :) Oct 03 19:08:21 someone seriously use OE for the OSK :) Oct 03 19:09:11 Crofton: I will do it so it can be overwritten Oct 03 19:09:35 that would be best for now Oct 03 19:09:39 Crofton: doesn't 2.6.17 work? Oct 03 19:09:49 Crofton: same stuf should be probably done for u-boot Oct 03 19:09:49 Crofton: but thats left for you Oct 03 19:10:00 I haven't backed up that far :) Oct 03 19:10:17 I think the u-boot I have selected is best Oct 03 19:10:25 I don't think 1.1.4 built for the osk Oct 03 19:10:38 ok Oct 03 19:10:41 and there have been entries suggesting OSK bugs are corrected Oct 03 19:18:13 wow.. 2.6.12-rc has cross-kernel 3.3.4 set Oct 03 19:18:21 first time I will build gcc-cross-kernel Oct 03 19:24:16 god.. I hate OO.org Impress Oct 03 19:24:37 'do all charts in OO Calc and then copy/paste them into OO Impress' way suxx Oct 03 19:25:49 each chart inserted into Impress is not a chart then Oct 03 19:30:11 rebooted Oct 03 19:31:42 hrw, that is there because it was in the bb file I copied from originally :) Oct 03 19:33:05 ah Oct 03 19:39:00 can that era kernel build with 4.1? Oct 03 19:39:23 or whatever the celf guys are using? Oct 03 19:39:32 most kernel can, but external patches tend to screw it up Oct 03 19:39:36 kernels* Oct 03 19:48:44 Crofton: I will build it with 3.4.4 later Oct 03 19:49:12 ok Oct 03 20:28:32 hrw, I am trying 2.6.17 ... Oct 03 20:28:40 something weird is going on Oct 03 20:30:24 * v8jlene is back Oct 03 20:32:24 v8jlene: wb Oct 03 20:32:35 morning all Oct 03 20:33:55 koen: got mail? Oct 03 20:35:36 now trying the defconfig from the kernel .... Oct 03 20:36:43 hrw: I already replied iirc Oct 03 20:38:16 koen: to forwarded one? Oct 03 20:38:26 yes Oct 03 20:38:46 kone: I have succesfully booted the Angstrom-bootstrap-image on the HTC Universal. It seems fast, and i didnt had any problem with UDEV like we have had with Familiar. Realy looking forward to test gpe and opie. Oct 03 20:39:04 goxboxlive: that's great news! Oct 03 20:39:19 goxboxlive: I'll poke florian to test gpe some more on his universal Oct 03 20:39:20 03lenehan 07org.oe.dev * r3d618cbb... 10/ (7 files in 3 dirs): Oct 03 20:39:20 perl: Another sh3/sh4 shared library fix. Modules build built with Oct 03 20:39:20 cpan.bbclass were being linked using ld instead of gcc, which often Oct 03 20:39:20 breaks on sh. Override the LD settings so any cpan modules that Oct 03 20:39:20 compile c/c++ code use gcc to link. Bump the PR on the effected Oct 03 20:39:20 modules. Oct 03 20:39:23 koen: then I did not yet get it Oct 03 20:39:43 hrw: sent it 34 minutes ago, but the uni mailserver can be a bit brackish Oct 03 20:40:08 v8jlene: and I just build a lot of perl modules :) Oct 03 20:41:12 * koen rebuilds Oct 03 20:41:22 koen: I had to do some changes, like remove pcmcia and hostap. And i have also changed preffered kernel to 2.17.1-hh1 in the Angstrom-2007.1.conf file Oct 03 20:41:59 goxboxlive: I'm working on getting pcmcia out of machines that don't need it Oct 03 20:42:15 koen: This change is for sh3/sh4 only... later today I'll probably push the sh3/sh4 changes for all targets though, since in theory the change from LD to GCC should work everywhere and it'll remove all the sh special cases.. Oct 03 20:42:49 v8jlene: we've have problems before with perl and missing symbols due to linking with LD Oct 03 20:43:01 v8jlene: a quick fix is to LD_PRELOAD libgcc Oct 03 20:43:23 koen: great. I can send you my mtn diff so you can see what i have changed. Oct 03 20:43:39 I'll push thing which change few things maybe not best but usefull Oct 03 20:43:45 goxboxlive: could you attach it to bugzilla? Oct 03 20:43:52 goxboxlive: I'm about to go to sleep :) Oct 03 20:44:00 yes, i can do that. Oct 03 20:44:08 koen: Good thinking. I've fixed it all properly now though ;) Oct 03 20:44:51 koen: 5a0ba6276c9c5a55e17b2681a1d2220b37d09582 will interest you probably Oct 03 20:44:52 perl was broken for ages due to that, till DaKa fixed it Oct 03 20:45:09 03hrw 07org.oe.dev * r5a0ba627... 10/ (3 files in 2 dirs): libusb: package libusbpp into separate package so libusb will not RDEPEND on libstdc++ Oct 03 20:45:23 forgot to add: taken from .oz354x Oct 03 20:45:29 ah, right Oct 03 20:46:24 I'll think I'll look into fixing bash (/etc/shells) tomorrow Oct 03 20:47:04 v8jlene: I have someone that wants a whole lot of cpan modules for his rootfs, so your commits make me happy :) Oct 03 20:47:11 03lenehan 07org.oe.dev * r10d887a6... 10/ (1 packages/perl/perl-5.8.7/config.sh-sh4-linux): (log message trimmed) Oct 03 20:47:11 perl 5.8.7: Enable threading in perl for sh4 to fix modules that Oct 03 20:47:11 check for threading availability. Oct 03 20:47:11 Since perl-native is used to compile modules, and perl-native is Oct 03 20:47:11 built with threading enabled, any modules that check for threading Oct 03 20:47:12 think that it's available. However when you attempt to use the module Oct 03 20:47:16 on the target it complains about missing symbols (those related to Oct 03 20:50:02 koen: I've almost finished with perl, the application I'm trying to get to run almost works! (Just found another missing module...) Oct 03 21:07:55 *sleep* Oct 03 21:07:55 cu koen Oct 03 21:28:28 Hi, I have a problem when trying to build a drembox image using makefile. Oct 03 21:28:48 It reports the it cant find the source @treke.net Oct 03 21:29:12 Someone have experienced this Oct 03 21:29:29 ement: oesources.org? this service need reimplementation Oct 03 21:29:44 it stopped working some time ago Oct 03 21:30:03 we will discuss this this week during oedem Oct 03 21:30:03 ok, are there a new ver of makefile? Oct 03 21:30:18 ement: I do not know about makefile from dreambox branch - sorry Oct 03 21:30:37 ement: look in conf/*.conf files for CVS_TARBALL_STASH Oct 03 21:30:50 ok, do you know where I can find the tarbalkl stash Oct 03 21:31:06 echo 'CVS_TARBALL_STASH = "http://www.treke.net/oe/source/"' >> build/conf/local.conf Oct 03 21:31:19 That url isn't valid Oct 03 21:31:33 and hasn't been for two years. Oct 03 21:31:37 I discovered that too Oct 03 21:31:42 hmm Oct 03 21:32:05 old source of building manual then Oct 03 21:33:27 someone has a new manual for building dreambox images? Oct 03 21:45:23 03kristoffer 07org.oe.dev * rd97c3915... 10/ (5 files in 3 dirs): Oct 03 21:45:23 packages/linux/linux-jlime-jornada7xx : Addition of 2.6.18 Package Oct 03 21:45:23 * Addition of jornada7xx 2.6.18 package with patches Oct 03 21:47:33 03kristoffer 07org.oe.dev * rc23ea977... 10/ (1 packages/linux/linux-jlime-jornada7xx_2.6.17.bb): Oct 03 21:47:33 linux/linux-jlime-jornada7xx_2.6.17.bb : Fix typo Oct 03 21:47:33 * Removed typo in src_uri -"m" Oct 03 22:00:33 cu Oct 03 22:14:59 03rpurdie * r614 10bitbake/lib/bb/data_smart.py: data_smart.py: Add expand_cache, caching expanded data (but wiped on any write to the data) - gives a performance improvement Oct 03 22:16:24 ~praise RP Oct 03 22:16:26 03rpurdie * r615 10bitbake/lib/bb/parse/parse_py/ConfHandler.py: parse_py/ConfHandler.py: Remove broken default options (as identified by mithro) Oct 03 22:16:27 All hail RP! Oct 03 22:21:23 03rpurdie * r616 10bitbake/ (bin/bitbake lib/bb/cache.py lib/bb/shell.py): cache.py: Stop passing cooker around as its no longer needed Oct 03 22:22:38 wow Oct 03 22:22:38 sexy Oct 03 22:27:06 03rpurdie * r617 10bitbake/bin/bitbake: bin/bitbake: Move handle_data() into parse_bbfiles instead of the callback Oct 03 22:27:38 03lenehan 07org.oe.dev * r0bd5ac16... 10/ (1 classes/cpan.bbclass): Oct 03 22:27:38 cpan.bbclass: If a module attempts to use LDDFLAGS to link then it will be Oct 03 22:27:38 linking against the host libraries. Modify LDDFLAGS as part of the other Oct 03 22:27:38 modifications that already occur. Oct 03 22:30:24 03lenehan 07org.oe.dev * re4335807... 10/ (1 packages/perl/perl_5.8.7.bb): Oct 03 22:30:24 perl 5.8.7: Manually add some dependencies for the math-bigint Oct 03 22:30:24 modules which are not picked up by the script which was used to Oct 03 22:30:24 generate the automatic dependencies. They way they are used makes it Oct 03 22:30:24 difficult to automatically detect that these are needed. Oct 03 22:33:07 03lenehan 07org.oe.dev * r87347ec2... 10/ (1 packages/perl/libnet-ip-perl_1.25.bb): Oct 03 22:33:07 libnet-ip-perl: Net::IP required Math::BigInt to function, so add to Oct 03 22:33:07 RDEPENDS. Oct 03 22:35:41 03lenehan 07org.oe.dev * r44a955a7... 10/ (1 packages/perl/libmath-bigint-gmp-perl_1.18.bb): Oct 03 22:35:41 perl: Add the Math::BigInt::GMP cpan module. This provides a faster, gmp Oct 03 22:35:41 based, implementation of Math::BigInt. Oct 03 22:42:08 'night all Oct 03 22:52:42 03lenehan 07org.oe.dev * re7282633... 10/ (1 packages/perl/libperl6-junction-perl_1.10.bb): Oct 03 22:52:42 perl: Add Perl6::Junction perl module. This provides support for the Oct 03 22:52:42 perl6 style junction operators in perl5. Oct 04 02:03:44 03lenehan 07org.oe.dev * re669ea3e... 10/ (5 files in 2 dirs): Oct 04 02:03:44 perl: Add the perl Text::Diff module and the Algorithm::Diff module it Oct 04 02:03:44 requires. Oct 04 02:03:50 03lenehan 07org.oe.dev * ra29ae4f5... 10/ (27 files in 2 dirs): Oct 04 02:03:50 perl: For native perl module packages determine the actual non-native Oct 04 02:03:50 package name to require from the name of the native package. This Oct 04 02:03:50 makes all native files identical which makes adding new ones easier. Oct 04 02:03:57 03lenehan 07org.oe.dev * rfe0f9746... 10/ (1 packages/perl/libalgorithm-diff-perl_1.1902.bb): Oct 04 02:03:57 libalgorithm-diff-perl: Fix up RDEPENDS, there was one " too many and one Oct 04 02:03:57 dependency listed twice. Oct 04 02:04:18 03lenehan 07org.oe.dev * r65201cf2... 10/ (16 files in 2 dirs): perl: Add a new set of perl modules related to testing. Oct 04 02:20:57 03lenehan 07org.oe.dev * raf4d9707... 10/ (25 files in 2 dirs): Oct 04 02:20:57 perl modules: Update the SRC_URI on all the modules to use PV to refer to Oct 04 02:20:57 the version so upgrades are easier. **** ENDING LOGGING AT Wed Oct 04 02:59:57 2006