**** BEGIN LOGGING AT Mon Jan 21 03:00:03 2008 Jan 21 03:34:09 RP, mickey|zzZZzz : there was a period of time where a bitbake version required a bzr python module that isn't available from the normal repos for Fedora. Apparently that is no longer required, so no worries. Jan 21 08:16:53 morning Jan 21 08:17:25 RP: ping Jan 21 08:37:46 good morning all ! Jan 21 09:22:56 rwhitby: pong Jan 21 09:23:29 morning all Jan 21 09:30:35 hey RP Jan 21 09:44:36 RP: is there any way to make bitbake exit with an error status if any task fails? 1.8.8 didn't, and 1.8.10 still doesn't. Jan 21 09:49:22 rwhitby: It should. What kind of failure are we talking about? Jan 21 09:50:12 rwhitby: Come to think of it, my autobuilder scripts use the exit code... Jan 21 09:53:45 RP: let me find the one Jan 21 09:53:58 (it sometimes gives an error exit, and sometimes doesn't) Jan 21 09:55:09 RP: http://pastebin.ca/867317 Jan 21 09:56:19 RP: note that this is using "bitbake -k" Jan 21 09:56:48 rwhitby: ah, -k. This suddenly makes more sense Jan 21 09:58:44 RP: perhaps insert an exit(1) after "NOTE: All possible tasks have been run but build incomplete (--continue mode). See errors above for incomplete tasks." Jan 21 10:04:57 03rpurdie * r1007 10/ (4 files in 4 dirs): runqueue.py: Fix exit code for build failures in --continue mode Jan 21 10:05:08 rwhitby: I've done something equivalent Jan 21 10:05:22 sweet, my autobuilder thanks you. Jan 21 10:07:12 RP, the 1.8.10 tarball is missing the bzr and hg fetchers from the lib/bb/fetch directory Jan 21 10:07:26 RP: next thing, when bitbake does a fail-over to a source mirror, but the original site was a https site (e.g. the current fontconfig source failing), the fail-over should be http, not https. Jan 21 10:07:39 Shoragan: ah, thanks for letting me know, I'll fix that... Jan 21 10:08:05 (since the source mirror site is not https) Jan 21 10:08:19 rwhitby: patches to fix it welcome... Jan 21 10:08:37 RP: will do - you agree with that functionality description? Jan 21 10:09:12 RP: in which bitbake file would I find that functionality? Jan 21 10:10:35 Hey steliosk, how are you doing? How is the new office coming along? Jan 21 10:11:09 rwhitby: Probably uri_replace() in lib/bb/fetch/__init__.py Jan 21 10:11:44 * rwhitby opens an introductory book on python ... Jan 21 10:11:51 rwhitby: be warned that function is subtle :/ Jan 21 10:12:03 RP, when do you think it will be fixed? i'd like to update the debian package before going to work :) Jan 21 10:12:12 Shoragan: 5 mins? Jan 21 10:12:20 ah, ok :) Jan 21 10:13:40 03rpurdie * r1008 10/ (branches/bitbake-1.8/MANIFEST tags/bitbake-1.8.10/MANIFEST): 1.8.x: Fix MANIFEST in branch+release Jan 21 10:14:51 RP: nevermind - I found the magic incantation in angstrom-mirrors.bbclass Jan 21 10:17:08 Shoragan: I've fixed the file on berlios too Jan 21 10:21:36 ok, thanks Jan 21 10:23:30 RP: thx for making me look at the mirrors problem myself, rather than just reporting a bug - I found the problem in the nslu2-linux mirrors class rather than in bitbake ;-) Jan 21 10:24:04 rwhitby: :) Jan 21 10:24:11 r Jan 21 10:24:32 RP, berlios stills says 1.8.10 "no files" Jan 21 10:24:46 BTW, whoever owns nylon-mirrors.bbclass should make the same fix. Jan 21 10:24:58 and oplinux-mirrors Jan 21 10:25:07 and own-mirrors Jan 21 10:26:27 Laibsch : We got a roof over our heads and a chair/desk :) Jan 21 10:26:47 well, ain't that something? Jan 21 10:26:48 ;-) Jan 21 10:26:57 steliosk: you slaves dont have desks? Jan 21 10:27:02 Shoragan: Sorry, should be fixed now Jan 21 10:27:40 XorA : Desks are overrated :) Jan 21 10:28:45 rwhitby: Feel free to fix them, I doubt anyone would mind ;-) Jan 21 10:29:45 brb Jan 21 10:31:25 rschuster: hey, is it you that plays with nokia tablet stuff? Jan 21 10:31:51 XorA: yeah its me Jan 21 10:32:17 rschuster: I assume you are trying to get OE to generate packages that can install into OS2008? Jan 21 10:33:07 XorA: it can do this already :) Jan 21 10:33:07 unfortunately some packages cause the CodeSourcery gcc to break :| Jan 21 10:33:31 rschuster: cool, I was interested in this area since I got my shiney new n810 Jan 21 10:35:25 I am using a customized distro config to generate the packages and do some renaming. I am not sure whether this thing is wanted in OE. it currently lives in our jalimo svn Jan 21 10:36:08 good morning Jan 21 10:36:25 XorA: apart from the distro config everything else is in OE already Jan 21 10:36:30 hi florian Jan 21 10:36:49 rschuster: can you chuck me a URL? Jan 21 10:38:20 XorA: jalimo svn: http://evolvis.org/scm/?group_id=11 Jan 21 10:38:20 the directory with the distro config within: http://evolvis.org/plugins/scmsvn/viewcvs.php/trunk/oe-overlay/conf/distro/?root=jalimo Jan 21 10:39:24 rschuster: awesome thanks Jan 21 10:39:47 hmm Jan 21 10:39:57 play with it and if you think it can be contributed to OE just do so. perhaps we can persuade the CodeSource gcc to work a little better if we apply all their patches Jan 21 10:40:04 somehow my configure does not even find the compiler anymore, I did not change much there Jan 21 10:40:09 did anyone had a similar problem? Jan 21 10:40:37 I tried rebuilding my mediatomb_svn.bb package with the latest revision Jan 21 10:40:40 rschuster: yeah it would be good to knock out packages from OE with no thought Jan 21 10:40:46 rschuster: hi... i remember that I intended to ask you about one of your checkins last week. but I do not remember which one :-) Jan 21 10:41:10 can ccache be borked or would that not be related? Jan 21 10:41:20 florian: something dubious ... ? Jan 21 10:42:11 perhaps this: (from cacao_ ...) Jan 21 10:42:11 # Quirks for Nokia N800: Jan 21 10:42:11 # - with mfloat-abi=softfp the runtime fails to run anything Jan 21 10:42:11 # - without --enable-softfloat it cannot find files within Jars Jan 21 10:42:11 TARGET_CC_ARCH_nokia800 = "-march=armv6j -mtune=arm1136jf-s -mfpu=vfp -mfloat-abi=soft" Jan 21 10:42:12 EXTRA_OECONF_append_nokia800 = "--enable-softfloat" Jan 21 10:43:27 rschuster: unlikely, usually I only notice this if it breaks building something. more likely that it was some patch that might be useful upstream or something like this. Jan 21 10:47:20 florian: I did a little cleanup in gnutls lately. this involved some directory renaming Jan 21 10:47:20 gnutls: Cleanup and fixes. Jan 21 10:47:20 - move 1.6.3-specific patches into respective recipe Jan 21 10:47:20 - renamed patch folder to contain version to which they belong Jan 21 10:47:45 ah and I made gpe-scap work with glade 2.4.0 Jan 21 10:47:54 ah right, that's the one Jan 21 10:48:02 because that is the version used on the N800's OS2008 Jan 21 10:48:44 now I only need an older libxml2 and gpe-scap can be installed without any errors Jan 21 10:49:02 rschuster: yes, I should apply this upstream... gpe-scap is useful :-) Jan 21 10:49:20 florian: indeed. I like it very much! Jan 21 10:50:02 and it is just cool to use OE to build it. :) Jan 21 10:51:07 rschuster: yes indeed... Jan 21 10:51:27 * florian would like to build more maemo stuff with oe... Jan 21 10:52:52 florian: here is your distro config: http://evolvis.org/plugins/scmsvn/viewcvs.php/trunk/oe-overlay/conf/distro/?root=jalimo Jan 21 10:53:29 I pointed XorA to it already. perhaps we can beef it up until it is acceptable to include it in OE Jan 21 10:53:42 I am building all the Java stuff, nano and gpe-scap with this Jan 21 10:53:57 * XorA will try and look at it tonight Jan 21 10:55:35 as said we need an older libxml2 in OE for OOTB-compatibility Jan 21 10:56:17 and some more package renames are possibly needed as well if you dive more into the gnome and gtk bits. Jan 21 10:57:41 rschuster: that's quite nice indeed Jan 21 10:58:47 We need to take a little bit care not to remove all older stuff that fast. Some people are a little bit to ambitious cleaning up here :) Jan 21 11:00:52 03rwhitby 07org.oe.dev * rc9250778... 10/ (1 classes/nslu2-mirrors.bbclass): nslu2-mirrors: Fix https handling Jan 21 11:01:01 03rwhitby 07org.oe.dev * r69a166c7... 10/ (4 files in 2 dirs): *-mirrors.bbclass: Applied the https handling fix to all other mirrors classes Jan 21 11:03:57 hrw|gone, XorA: Man, qte-fonts-common is ugly ;-) re bug 2430, would http://rafb.net/p/Qp3ssG24.txt be the kind of solution you were proposing (qpf and universal part only, X11 part to be done) Jan 21 11:04:38 02qtttffont-update has hard-coded paths, maybe that could be improved, but what do you say as far as general idea is concerned? Jan 21 11:05:27 Laibsch: that looks like what I was saying Jan 21 11:05:31 Laibsch: thanks Jan 21 11:05:41 nice Jan 21 11:05:54 Laibsch: then all fonts just call update-fonts on postinst and should work for both systems Jan 21 11:05:56 Do you know how fc-cache is used? Jan 21 11:06:14 fc-cache with no arguments AFAIR Jan 21 11:06:35 yes, qpf fonts include the qte-fonts-common stuff AFAIK, so they should all have an postinst_append line Jan 21 11:06:53 OK, I'll create the necessary file Jan 21 11:07:35 Which file should RDEPENDS on update-fonts-common for X11 fonts? Jan 21 11:08:30 libfontconfig-utils probably Jan 21 11:08:34 OK Jan 21 11:08:39 I'll take a look Jan 21 11:08:45 thats where fc-cache is I think Jan 21 11:09:08 of course someone will chew you out if you listen to me :-( Jan 21 11:09:22 ;-) Jan 21 11:09:47 Currently, ttf fonts are not being dealt with Jan 21 11:09:58 ttf fonts could be in both opie and X11 Jan 21 11:10:21 But we have update-fonts-common now and can easily add an RDEPENDS here and there as necessary Jan 21 11:23:17 Laibsch: are you certain your local.conf for collie doesn't contain anything special? I'm beginning to think that my build env is just fine... Jan 21 11:23:50 I believe it might be the machine config Jan 21 11:24:05 I keep it in sync with .dev automatically (hard links) Jan 21 11:24:21 But I don't want to commit that until current problems have been worked out Jan 21 11:24:23 I tried building dev... Jan 21 11:24:30 Maybe I should give you a patch Jan 21 11:24:42 pure .dev doesn't build here Jan 21 11:24:46 Right, .dev failed, too Jan 21 11:25:03 Wait a second Jan 21 11:27:04 * * OE Bug has been RESOLVED (FIXED) by Jan 21 11:27:06 * *  classes/storcenter-image.bbclass missing Jan 21 11:27:08 * * http://bugs.openembedded.org/show_bug.cgi?id= Jan 21 11:29:10 CoreDump: BTW, I did clean out tmp yesterday and restarted a build. Maybe things will turn out alright. Jan 21 11:29:16 * Laibsch keeps fingers crossed Jan 21 11:29:36 I reach the point of breakage after ~15m Jan 21 11:29:45 with an empty tmp/ Jan 21 11:31:53 re-trying .dev / angstrom-2008.1 ATM Jan 21 11:32:58 http://rafb.net/p/KQxBgU85.html looks sane to me Jan 21 11:34:10 yeah Jan 21 11:35:40 I wonder why your collie builds go the eabi route Jan 21 11:37:55 I have no idea Jan 21 11:38:16 and it is rather frustrating. My build env is working _perfect_ otherwise Jan 21 11:38:25 CoreDump: http://oz.leggewie.org/wip/sonkei-conf.diff has my current changes against Sonkei for conf/ Jan 21 11:38:31 I'll take a closer look Jan 21 11:38:58 thanks Jan 21 11:39:51 03freyther 07org.oe.dev * rb13ffab7... 10/ (1 classes/cmake.bbclass): classes/cmake.bbclass: Spaces are cool, use them Jan 21 11:39:57 03freyther 07org.oe.dev * r7147319e... 10/ (1 classes/cmake.bbclass): Jan 21 11:39:57 classes/cmake.bbclass: A small explanation why it says inherit autotools Jan 21 11:39:57 CMake's Makefiles know about the DESTDIR. This allows us to use the Jan 21 11:39:57 staging and install functions of the autotools bbclass. Reusing them Jan 21 11:39:57 makes perfect sense and we should consider moving these functions to Jan 21 11:39:59 another base class. Jan 21 11:40:02 03freyther 07org.oe.dev * rf62682df... 10/ (1 classes/cmake.bbclass): classes/cmake: To invoke cmake it has to be present, add cmake-native to the DEPENDS Jan 21 11:40:08 03freyther 07org.oe.dev * r341b3084... 10/ (1 packages/tasks/task-qtopia-core-console.bb): packages/tasks/tasks-qtopia-core-console: Add a task for the core libraries of QtopiaCore/QtE Jan 21 11:40:17 03freyther 07org.oe.dev * r8b72081a... 10/ (1 packages/images/qtopia-core-console-image.bb): Jan 21 11:40:17 packages/images/qtopia-core-console-image.bb: Create a minimal image with the QtopiaCore Core libraries Jan 21 11:40:17 Provide a minimal image to use QtNetwork, QtSql, QtXml and QtCore. Jan 21 11:40:37 and people wonder why cdbot has flood-control ;) Jan 21 11:40:41 CoreDump: I cannot see any changes that would make this come about for you and make it go away for me Jan 21 11:40:58 CoreDump: I like the messages Jan 21 11:41:07 if i want to deply libc on the target Jan 21 11:41:28 what do i specify as IMAGE_INSTALL Jan 21 11:45:20 morning Jan 21 11:45:28 Laibsch: I do not like that way Jan 21 11:45:41 hrw: good Jan 21 11:46:01 * Laibsch would like to hear hrw's points for improvement Jan 21 11:46:26 I assume you are talking about update-fonts-common? Jan 21 11:46:50 yes Jan 21 11:47:01 I have to look at one thing first so moment Jan 21 11:47:41 Sure Jan 21 11:47:43 No rush Jan 21 11:49:12 Laibsch: btw - can we move it to ML? I have rather long text about it to write and prefer to get also other people ideas on it Jan 21 11:49:21 Sure Jan 21 11:49:24 or maybe it is already on ML but I missed it? Jan 21 11:49:33 Hm, no I think it is not Jan 21 11:49:50 Laibsch: it need changing ttf.inc, qpf.bbclass, fontconfig, opie-ttf-support and probably few others Jan 21 11:50:17 uh oh Jan 21 11:50:24 sounds like a lot of work Jan 21 11:50:56 no Jan 21 11:51:03 small work but in many places Jan 21 11:54:59 ouf! Jan 21 11:58:11 Is it a problem if linux-libc-headers are from version 2.6.20 and the kernel is 2.6.23? Jan 21 11:58:32 I am running into problems compiling procps and that is something I was wondering about earlier Jan 21 11:59:42 the collie build used eabi again meh Jan 21 12:00:55 Laibsch: should not be a problem Jan 21 12:02:39 eabi should not be a problem? Jan 21 12:02:49 But collie is oabi Jan 21 12:03:03 * * OE Bug has been RESOLVED (FIXED) by Jan 21 12:03:05 hrw: Oh, I understand now Jan 21 12:03:05 * *  classes/storcenter-image.bbclass missing Jan 21 12:03:07 * * http://bugs.openembedded.org/show_bug.cgi?id= Jan 21 12:09:06 CoreDump: Have you looked at "mtn status && mtn ls unknown" for .dev and Sonkei? Jan 21 12:09:10 Any local changes? Jan 21 12:09:29 it is a fresh checkout Jan 21 12:09:37 the mtn tree I mean Jan 21 12:12:01 mickeyl: good morning Jan 21 12:15:33 CoreDump: MACHINE=collie bitbake -c devshell shasum-native Jan 21 12:15:45 And then try "set|grep -i abi" Jan 21 12:16:03 Maybe we can find out why the environment is incorrectly set up for you Jan 21 12:17:00 good morning pb_ Jan 21 12:20:04 * * OE Bug 2024 has been RESOLVED (FIXED) by Jan 21 12:20:06 * * classes/storcenter-image.bbclass missing Jan 21 12:20:08 * * http://bugs.openembedded.org/show_bug.cgi?id=2024 Jan 21 12:20:48 Laibsch: that didn't work Jan 21 12:21:01 rNOTE: package shasum-native-1.0-r1: task do_devshell: failed **** BEGIN LOGGING AT Mon Jan 21 12:54:47 2008 Jan 21 12:57:11 hrw: Have you already sent a mail to the list? Jan 21 12:57:58 no - I had some work to do Jan 21 13:08:38 Laibsch: sent Jan 21 13:19:14 hey koen, what's your preferred email-address? **** BEGIN LOGGING AT Mon Jan 21 13:23:39 2008 Jan 21 13:33:24 Hi, guys. im just wondering. are there any mentors for the acer n30 fammily of devices? Jan 21 13:45:05 not really Jan 21 13:45:20 we got a code drop from the letux project Jan 21 13:45:28 but i can't remember who got that and what he did with it Jan 21 13:49:30 if at all Jan 21 13:49:40 we have n30 in OE iirc but thats probably all Jan 21 13:50:15 * XorA thinks Chriva just volunteered Jan 21 13:50:17 n30 is even worse then gta01 ;( Jan 21 13:50:32 same cpu but half of ram/flash and 1/4 of screen Jan 21 13:50:41 but at least have keys and dpad ;) Jan 21 13:54:02 it works quite well. but i had to modify the patches for the 6.14 kernel to get them working together with the 6.20 kernel. i even managed to get tiuchscreen up and running after several hours of headscratching and stupid quiestion. no one sems to knew how to do it so i had to play around quite much Jan 21 13:54:44 Chriva: you should have stolen the driver from openmoko kernel Jan 21 13:57:15 there is one step left to do though. the damn sound :/ the touchscreen was not hard to get up and running in the end. i just had to disable mouse support in the kernel and edit tslib.sh to get it working. Jan 21 13:59:36 03jeremy_laine 07org.oe.dev * r87198afa... 10/ (1 packages/linux/linux_2.6.23.bb): Jan 21 13:59:36 linux_2.6.23.bb: fix sources for mpc8313e-rdb, there is no 'defconfig-rt' file Jan 21 13:59:36 * fixes breakage introduced in commit cb0f0a066b9df17408c3017f7954a6b750fd269f Jan 21 14:00:19 Chriva: what codec on the n20? Jan 21 14:00:28 ChanServ: n30/n35 Jan 21 14:00:36 ~lart autocomplete Jan 21 14:00:37 * ibot hereby declares autocomplete a troll Jan 21 14:01:44 XorA: UDA1380 -same as ipaq 36 series of ipaq. probably others too =) Jan 21 14:02:08 Chriva: that should be easy to hook upto a samsung cpu if you use asoc drivers Jan 21 14:02:44 There's already a uda1380 driver in at least the dev branch so you don't even have to write the codec driver :) Jan 21 14:02:59 Though it does need a bit of TLC before mainline. Jan 21 14:03:53 and the idiot who last messed with the samsung driver is easy to find :-D Jan 21 14:04:31 i looked trought the kernel make-files and it sems that the codec driver is bound to ipaq devices only. there are tons of definitions pointing to ipaq stuff inside the source file :S Jan 21 14:04:53 Chriva: http://opensource.wolfsonmicro.com/node/6 Jan 21 14:05:31 Chriva: we dont do it that way anymore, those drivers are dinosaurs Jan 21 14:06:18 The ASoC subsystem splits out the codec/CPU/machine code; like I say there's an unmerged driver in the Wolfson git for the uda1380. Support for the CPU is merged. Jan 21 14:06:23 XorA, are you the father of the dinosaurs? Jan 21 14:06:43 Crofton K hi ! Jan 21 14:06:55 steliosk, wb Jan 21 14:06:59 Crofton: watch it with the age jokes old man :-D Jan 21 14:07:01 thanks ! Jan 21 14:07:12 how is the new office? Jan 21 14:07:14 Walking sticks at dawn! Jan 21 14:07:44 a chair and a desk for the moment.... but its much much better than last week Jan 21 14:08:39 * Crofton need to post a photo of his chair and desk Jan 21 14:08:57 XorA: i could volunteer as n30/n35 mentor. let me think about it. i have to make sure im capable of doing that first. Jan 21 14:09:01 main problem is inet. we could only get a adsl 2mbit circuit and have to wait 30+ days to get our normal service Jan 21 14:09:10 ug Jan 21 14:09:22 but its much better than dial up and gprs ;) Jan 21 14:09:31 :) Jan 21 14:09:34 When we moved office at last work we had 8 of us behind a 64k ISDN link for the first month or so. Jan 21 14:09:58 Could've had 128k except the other channel was our voice line... Jan 21 14:10:07 boonie: same here till lasy friday Jan 21 14:10:13 s/lasy/last Jan 21 14:11:53 one of the side effects of moving was the fact that i had to clear my library and i found all the 'hip' literature about the 'future' i had accumulated over the years.... Jan 21 14:11:57 who need voice :-D Jan 21 14:12:53 funny to see how fast things change to a totaly different directions than people are predicting :) Jan 21 14:12:53 XorA: We needed it to shout at the ISP for being idiots. Jan 21 14:13:19 broonie: should have used Lumison, they were within punching distance Jan 21 14:13:21 XorA : Pesky customers who want tech support ;) Jan 21 14:13:51 XorA: Not quite at the office this was for (on Leith Links) but still... Jan 21 14:15:34 The other thing i found out is that we had a tone (actually) of old equipment sitting in the basement,from pc's and printers to old dev boards,measurement stuff Jan 21 14:17:06 anyone interested for a really old Commodore ? :) Jan 21 14:18:35 http://en.wikipedia.org/wiki/Commodore_PET Jan 21 14:19:03 Hehe those old suckes are quite valuable ^^ Jan 21 14:19:35 i have a couple of the CBM's (not sure what model) Jan 21 14:20:01 plugged them in and still working :) Jan 21 14:20:47 haha 4KB that even worse than my cache memory in my old palm Vx ^^ Jan 21 14:30:11 mm. guys.. is the aso driver really supposed to be 460 Meg in size? :S Jan 21 14:31:20 aso? Jan 21 14:32:05 oh sorry *alsa asoc Jan 21 14:32:33 Ah, you're looking at the diff? Jan 21 14:32:45 460 meg sounds like the whole git repo :-) Jan 21 14:33:03 Or the diff has picked up more stuff than it should Jan 21 14:33:25 its the diff from current asoc to rev 1 in git :-D Jan 21 14:33:29 Hey guys, I am trying to build for collie and as you know, I don't have the faintest clue. But since a few days ago, the builds are giving me problems (might be work commits to conf/machine/) Jan 21 14:33:29 I also seriously lack CPU cycles so I asked mickey to do a build from scratch and his build broke in glibc-intermediate-2.6.1: http://rafb.net/p/2k2qvf98.html Jan 21 14:33:31 Any kind soul around to decipher that for me and suggest a plan of action (IOW, what should be fixed where?) Jan 21 14:33:47 morning Jan 21 14:33:47 mickeyl said it was something about ntpl Jan 21 14:34:36 Nah, not the diff - it's only 960K uncompressed :) Jan 21 14:35:17 Chriva: What are you trying to download here? The asoc git is a complete kernel git plus ASoC development so that does sound a reasonable size. Jan 21 14:35:29 * broonie points at #alsa-soc Jan 21 14:36:39 broonie: i did exactly as the page says (never uset git before ) the downlaod is finnished, but i just wondered if it really should be that big. oh well. time to apply some patches then =D Jan 21 14:37:54 Chriva: Ah, right. A git clone contains a complete history of all the branches (due to the whole distribute RCS thing). Jan 21 14:42:03 broonie: it sems to be the whole 2.6.24 kernel. third time i have to mess around with old patches =) Jan 21 14:43:39 Yes, it is. There's also a separate patch set linked from the page above. Jan 21 14:45:14 XorA, hrw, RP, polyonymous, rschuster, rwhitby, stefan_schmidt, zecke: pinging the usual suspects to get attention to the problem I wrote about 10 minutes ago. Any hardware experts around? Somebody you can suggest with knowledge about oabi/eabi/ntpl? Jan 21 14:46:44 Laibsch: not a clue, havent messed with Angstrom 2008.1 at all Jan 21 14:47:10 XorA: But do you have any knowledge about oabi/eabi/ntpl and those things? Jan 21 14:47:15 I don't Jan 21 14:47:16 Laibsch: I doubt anyone can comment just based on that log. It'll need digging into the code to find out what that symbol does and why its defined for eabi presuably but not oabi Jan 21 14:47:46 Laibsch: if you are building glibc 2.6.x you are in a very experimental area ... Jan 21 14:47:49 Laibsch: nope, I just assume it works Jan 21 14:47:56 Laibsch: When/why did it break? Jan 21 14:48:03 Which glibc did it used to build? Jan 21 14:48:05 RP: Sure. But where would one need to look? Who is knowledgeable about it? Those are my questions now Jan 21 14:48:24 Laibsch: OE lacks people with this kind of knowledge :/ Jan 21 14:48:35 RP: Which was the impression I had Jan 21 14:48:56 RP: It initially broke for me with the errors that do_rootfs would not work anymore Jan 21 14:49:05 This happened around Jan 08 Jan 21 14:49:33 Laibsch: You need to try and isolate the problems, it sounds like there are more than one affecting things Jan 21 14:49:37 * Laibsch suspects that even my changes to conf/machine might have caused this although I paid attention to not make too broad semantical changes Jan 21 14:49:50 RP: I am trying to do this right now Jan 21 14:50:05 I cleaned out tmp now and wait for it to break for me Jan 21 14:50:36 rschuster: glibc2.6, it should be 2.5, right? Jan 21 14:50:47 I think that is what it used to be, but am not sure Jan 21 14:51:13 steliosk: PET? working? Jan 21 14:51:45 steliosk: month ago I would have an use for it - my friend had 0x20 birthday Jan 21 14:51:59 steliosk: I gave him c64dtv ias gift Jan 21 14:52:14 Laibsch: yeah. Is there a strong reason to use angstrom-2008.1? If not use the 2007 releases and make whatever change you want against this. Jan 21 14:52:32 I do Jan 21 14:52:47 And indeed, it seems that I also get glibc 2.5 Jan 21 14:53:06 mickeyl: Can you retry the build against angstrom stable? Jan 21 14:53:35 My build is still running, no breakage so far Jan 21 15:21:23 I'm getting an error from libtool-cross when building apmd for one of my devices, does anyone know what this is about? http://rafb.net/p/e4xZg183.html Jan 21 15:21:38 I'm almost certain it is my fault :) Jan 21 15:26:26 h/win 19 Jan 21 15:30:23 broonie: it works now... well almost -no touchscreen yet Jan 21 15:31:35 cool Jan 21 15:42:13 ok, weird. I just rebuilt libtool-cross for the third time and now apmd built cleanly. Who says you can't recompile and expect different results? Jan 21 15:43:58 * mwester loves determinism Jan 21 15:44:53 maybe libtool only works during business hours Jan 21 15:47:43 what libtool works? Jan 21 15:47:48 this is unpossible\ Jan 21 15:48:50 well, to be fair I don't know that it worked. At least it didn't sabotage the apmd build this time Jan 21 15:48:57 03freyther 07org.oe.dev * r77822f4d... 10/ (3 files in 3 dirs): Jan 21 15:48:57 Add a task for QtopiaCore GUI libraries and related libraries and install this task to a image Jan 21 15:48:57 Provide a minimal image with QtopiaCore GUI available. Jan 21 15:51:37 rschuster: hmm, I hit dbus-1 vs dbus problem with chinook-compat Jan 21 15:53:40 XorA: that means? Jan 21 15:53:49 BTW: fortran is fixed for everyone? Jan 21 15:53:58 XorA: the package has a wrong name? Jan 21 15:54:18 OE generates a dbus-1_1.0.2 package, chinook has dbus_1.0.2 and libdbus-1-3 Jan 21 15:54:53 apaulsen: There is some kind of deep dark voodoo in the libtool recipes :/ Jan 21 15:55:11 zecke: Lets try... :) Jan 21 15:55:27 RP: cool Jan 21 15:56:14 XorA: but what is the error or what should we fix? Jan 21 15:56:19 zecke: btw, did you ever see my qmake change in poky? Jan 21 15:56:31 RP: nope Jan 21 15:56:37 zecke: The mv qmake-native* -> qmake-cross* Jan 21 15:56:40 rschuster: well it looks like you attempt to rename dbus to libdbus-1-3 in your distro.conf Jan 21 15:56:42 RP: looking now Jan 21 15:56:45 rschuster: but that doesnt work Jan 21 15:57:23 zecke: Its do_stage function touches CROSS_DIR... Jan 21 15:57:29 python populate_packages_prepend () { Jan 21 15:57:29 if (bb.data.getVar('DEBIAN_NAMES', d, 1)): Jan 21 15:57:29 bb.data.setVar('PKG_dbus', 'dbus-1', d) Jan 21 15:57:29 } Jan 21 15:57:37 I bet that overrides it Jan 21 15:57:45 zecke: fortran fix with the toolchain is working here; and my system is one of the ones that doesn't have a native mfpr lib Jan 21 15:58:27 XorA: hm. but why not change this in chinook-compat.conf? Jan 21 15:58:45 PKG_dbus = "libdbus-1-3" is there currently Jan 21 15:58:50 rschuster: see above Jan 21 15:59:02 rschuster: that looks like its being overridden by the dbus recipe Jan 21 16:00:28 XorA: ah now I get it. Jan 21 16:00:39 rschuster: Im just double checking Jan 21 16:00:43 XorA: the python script above is part of the recipe? Jan 21 16:00:53 rschuster: yes Jan 21 16:00:55 it should check whether PKG_dbus is already set Jan 21 16:02:00 rschuster: dpkg-deb: building package `libdbus-1-3' in `/home/dp/zaurus/build-n800/tmp/deploy/deb/armv6/libdbus-1-3_1.0.2-r8_armel.deb' Jan 21 16:02:04 rschuster: yeah its that code Jan 21 16:03:16 thats the wrong localtion for that code anyway I feel Jan 21 16:03:20 bugs.oe.org still broken? Jan 21 16:03:27 bugs.openembedded.net Jan 21 16:03:32 rschuster: see the topic Jan 21 16:04:31 XorA: http://bugs.openembedded.net/show_bug.cgi?id=3383 Jan 21 16:05:02 XorA: there is an attempt by me to make custom python code for overriding the debian names unneccessary Jan 21 16:05:57 XorA: if that would be used for dbus it would fix your issue, too Jan 21 16:06:21 hmm, guys, my .bb fetches stuff from a local svn repository, via the http protocol; I did check out from there already so svn should know my username and password Jan 21 16:06:28 rschuster: the problem here is dbus is just doing things the wrong way Jan 21 16:06:40 however the fetcher tells me auth required and refuses to dl the sources Jan 21 16:06:46 rschuster: dicking around in debian.bbclass inside a recipe has got to be bad form Jan 21 16:07:10 any clue on what to do? Jan 21 16:08:12 allthough I think I might have found the reason myself :) Jan 21 16:08:22 yeah, baaah, stupid me =) forget it Jan 21 16:08:29 was using http:// instead of svn:// + proto=http Jan 21 16:09:34 rschuster: so we could add a DEBIAN_PN_dbus_chinook-compat = "libdbus-1-3" ? Jan 21 16:10:13 rschuster: if I understood you code right Jan 21 16:11:55 hm, no Jan 21 16:12:30 XorA: just add DEBIAN_PN_dbus ?= "dbus-1" instead of the python code into the dbus recipe Jan 21 16:12:51 XorA: then the naming override in chinook-compat.conf has the desired effect Jan 21 16:13:20 at least if chinook-compat is build. all other distributions are untouched Jan 21 16:13:57 and this probem is also gone: "fragile because the python code in the recipe and in populate_packages must have the same indentation" Jan 21 16:14:23 rschuster: ah I see Jan 21 16:14:31 rschuster: we should see about getting that in OE then :-) Jan 21 16:15:09 XorA: is mtn down ? Jan 21 16:15:19 lrg: dont think so Jan 21 16:15:41 maybe it's me being tired and dumb today Jan 21 16:16:12 rschuster: ah sorry, I was being a bit think and lost my way in the if statements, its obvious now Jan 21 16:19:32 rschuster: I think that should be commited and have commented to that effect Jan 21 16:20:55 XorA: you mean. ok for bug 3383? Jan 21 16:27:25 rschuster: yes, looks like a good fix to me Jan 21 16:27:36 rschuster: I hate seeing python poking around in recipes Jan 21 16:31:09 XorA, rschuster: Just please don't use PN in variable names Jan 21 16:31:56 lrg: I am also seeing VERY slow connects to mtn Jan 21 16:32:13 RP: DEBIAN_PKG_ instead? Jan 21 16:32:14 I checked ltg, it is humming along Jan 21 16:32:42 rschuster: DEBIANNAME ? Jan 21 16:32:50 RP: sounds good Jan 21 16:32:52 lrg: must be ewi Jan 21 16:33:06 rschuster: It pays to try and aboid _ if you can Jan 21 16:33:23 :) Jan 21 16:33:37 "underscore considered harmful" Jan 21 16:33:44 DEBIANNAME is more readable as a bonus Jan 21 16:34:42 hrw, RP: latest git kernel for collie fails size check (yes, I got this far, maybe I am in luck) Jan 21 16:35:18 Is there anything I can quickly to see if my image builds? I would not care if it failed booting Jan 21 16:35:40 ok Jan 21 16:35:42 lrg: It's not just you... I can connect but then the client sits there doing not a lot. pulling from opensource appears fine. Jan 21 16:35:42 would also be nice to break libdbus out of dbus-1 package Jan 21 16:35:54 Laibsch: Pick something in the defconfig and turn it off ;-) Jan 21 16:35:56 Ill have to look at that later Jan 21 16:36:12 CONFIG_collie = N, that'll fix it Jan 21 16:36:37 Laibsch: I'm in the process of trying to clean up the linux-rp defconfigs, I obviously need to tweak things some more :/ Jan 21 16:36:43 XorA: What does that do? Jan 21 16:36:50 Laibsch: disbale collie support Jan 21 16:36:59 XorA: I figured Jan 21 16:37:03 Laibsch: it may not be a real option :-) Jan 21 16:37:06 British humour, I guess Jan 21 16:37:10 ;-) Jan 21 16:37:23 Laibsch: we feed you beers until you understand us :-D Jan 21 16:37:39 yeah, that might help Jan 21 16:37:51 Hey folks, how do I fix: Fetch failed: ftp://elsie.nci.nih.gov/pub/tzdata2007e.tar.gz Jan 21 16:38:06 drmikecrowe: I think I commited a fix earlier today Jan 21 16:38:14 cool, thanks Jan 21 16:38:14 drmikecrowe: "mtn pull && mtn up" Jan 21 16:38:20 might be enough Jan 21 16:38:22 yep, will do Jan 21 16:38:31 you should then get version 2007k Jan 21 16:38:35 I did that Saturday, but just now getting back Jan 21 16:38:37 (to it) Jan 21 16:38:38 if not, I have not yet pushed Jan 21 16:38:38 Laibsch: There is very little in the collie defconfig left to become a module. You could try CONFIG_FB_SA1100 Jan 21 16:39:22 RP: Can I just hack around the size check for now? Jan 21 16:39:41 My most pressing question right now is whether do_rootfs will fail or succeed Jan 21 16:39:56 I'd just need to remember not to use that kernel Jan 21 16:41:14 RP: Since you are fiddeling with defconfigs, you might want to take a look at http://bugs.openembedded.net/show_bug.cgi?id=3686 Jan 21 16:41:14 Laibsch: Sure, I often do that ;-) Jan 21 16:41:28 Laibsch: Just comment it out in the zaurus-2.6.inc file (I think thats where it was) Jan 21 16:42:10 Laibsch: rtc-sa1100 can also be a module Jan 21 16:42:45 03Laibsch 07org.oe.dev * rb5334f00... 10/ (6 files in 2 dirs): fontconfig: unify Jan 21 16:44:04 hvontres|work: Some of that made this sweep but there are probably more tweaks needed Jan 21 16:45:42 RP: np. I'll try a test tonight then :) Jan 21 16:51:31 RP: looking at the diff for #3686 some more, it looks like we should try to fix loading those modules instead of compiling them back in Jan 21 16:54:22 hvontres|work: Yes, we won't be compiling them in Jan 21 16:55:13 RP: looks like I need to keep an eye on my Co-mentor :) Jan 21 16:55:32 Laibsch: Are you sure you pushed up? I ended up with the same error Jan 21 16:56:40 Can I just rename to ...2007k and go? Jan 21 16:59:07 for Angstrom 2007 stable? Jan 21 17:00:23 bye Jan 21 17:00:28 gn Jan 21 17:03:11 drmikecrowe: I believe you are right Jan 21 17:03:25 I have not yet pushed since I wanted to verify successful compilation first Jan 21 17:03:48 If you rename the first you will later need to manually deal with mtn when the commit comes in Jan 21 17:04:10 "mtn pull && mtn revert packages/tzdata/ && mtn up" Jan 21 17:06:58 If you rename the *file* you will later need to manually deal with mtn when the commit comes in Jan 21 17:14:05 * * OE Bug 3333 has been RESOLVED (FIXED) by Jan 21 17:14:07 * * linux-rp-2.6.23: sd drivers do not load before looking for root fs Jan 21 17:14:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3333 Jan 21 17:29:01 I want to md5 a file in a package before it is packaged. I tried adding the md5sum in the do_install_append but that does not work since the binary I'm md5'ing is stripped later in the OE build process. Were would be the right place to put the md5 logic? Jan 21 17:39:06 svolpe: Or try going the other way: strip in do_install_append, then md5sum Jan 21 17:39:13 zecke: The fortran issue looks fixed here, thanks Jan 21 17:39:29 svolpe: If you want to commit that to OE, you'd need to ask others if that is acceptable Jan 21 17:41:58 Laibsch, I will not commit this package as it is very unique and specific to a product I'm working on. I will commit parts of it that others can use but not this specific bb file since it is very specific. Jan 21 17:42:19 Laibsch: when using strip -o tempfile no harm should be done. otherwise you'd be removing the debug info that should go into the -dbg package. Jan 21 17:43:28 pH5, so you suggest coping the file to a temp file, strip it and then md5sum it? Jan 21 17:44:50 svolpe: Yes, that sounds like an absolutely workable and elegant idea from ph5 Jan 21 17:44:53 svolpe: yes, because runstrip is called from inside populate_packages, so doing the md5sum in or after do_package would probably involve some fragile hackery Jan 21 17:45:22 Laibsch: your idea, I only refined ;) Jan 21 17:45:22 RP: Last comment in bug 3708 Jan 21 17:45:37 team work! Jan 21 17:45:40 ;-) Jan 21 17:46:04 RP: Looks like the kernel passes the size test if those two are put into modules Jan 21 17:47:02 Laibsch: ah, ok Jan 21 17:50:45 drmikecrowe: I just pushed the change after verifying successful compilation of tzdata and tzcode-native Jan 21 17:50:56 great, thanks Jan 21 17:51:08 you're welcome Jan 21 18:00:17 Thanks Laibsch and pH5 Jan 21 18:02:32 03Laibsch 07org.oe.dev * r2a53b1b3... 10/ (3 files in 2 dirs): tzcode: update tzcode-native to 2007k Jan 21 18:02:37 03Laibsch 07org.oe.dev * ra250ef0b... 10/ (3 files in 2 dirs): tzdata: update to version 2007k Jan 21 18:17:04 * * OE Bug 3709 has been created by  Jan 21 18:17:06 * * 2.6.23 git kernel for collie does not pass size test Jan 21 18:17:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3709 Jan 21 18:33:58 is the main repo down ? Jan 21 18:36:27 could be Jan 21 18:36:35 I was having issues earlier Jan 21 18:36:48 But I connected to it successfully about 15 minutes ago Jan 21 18:37:00 but it seems a little flakey today Jan 21 18:37:07 (this is not ltg, AFAICT) Jan 21 18:37:22 i get this -> mtn: network error: name resolution failure for monotone.openembedded.org: Temporary failure in name resolution Jan 21 18:38:11 wow, that sounds strange Jan 21 18:38:18 Laibsch : Any of the mirrors working ? Jan 21 18:38:21 right now I end up at ltg.org and mtn pull works just fine. Jan 21 18:39:08 steliosk: "dig monotone.openembedded.org @NS1.EASYDNS.COM" looks good Jan 21 18:39:28 pH5: That is being rerouted with an ssh tunnel to ewi Jan 21 18:39:45 if ltg goes, monotone.oe.org goes down Jan 21 18:39:52 and ltg had been flakey in the past Jan 21 18:40:02 but today it seems to be ewi, if anything Jan 21 18:40:29 steliosk: I also used the wolfson mirror successfully today Jan 21 18:40:46 steliosk: opensource.wolfsonmicro.com Jan 21 18:40:52 steliosk: Are you sure your DNS is OK? Jan 21 18:40:58 no Jan 21 18:41:07 !mirrors Jan 21 18:41:13 ~mirrors Jan 21 18:41:14 extra, extra, read all about it, mirrors is Official Debian archive mirrors get an address of the form ftp://ftp..debian.org. These are the best advertised and most used sites; the face of Debian for most people. A list is available at http://www.debian.org/mirror/official/ Jan 21 18:41:34 well, this used to contain oe mirrors, too Jan 21 18:41:35 ~oemirrors Jan 21 18:41:36 somebody said oemirrors was extra, extra, read all about it, oemirrors is OpenEmbedded monotone mirrors are ewi546.ewi.utwente.nl, monotone.nslu2-linux.org, opensource.wolfsonmicro.com also see http://www.openembedded.org/wiki/GettingStarted Jan 21 18:41:42 ah, there it is Jan 21 18:42:09 ibot: forget oemirrors Jan 21 18:42:09 i forgot oemirrors, Laibsch Jan 21 18:42:33 ibot: oemirros is extra, extra, read all about it, oemirrors is OpenEmbedded monotone mirrors are ewi546.ewi.utwente.nl, monotone.nslu2-linux.org, opensource.wolfsonmicro.com also see http://www.openembedded.org/wiki/GettingStarted Jan 21 18:42:34 okay, Laibsch Jan 21 18:42:54 Sorry, I thought the wolfsonmicro mirror was not in there Jan 21 18:43:18 Laibsch : ok looks like the machine doing the pull was using the wrong dns Jan 21 18:43:56 OK, guys. It seems like I am back to the do_rootfs problem. Jan 21 18:44:42 * steliosk is completely out of touch with recent OE developments Jan 21 18:45:45 heh just 991 revisions behind ;) Jan 21 18:45:46 Laibsch, pH5 the only issue with doing the strip to a temp file is OE uses the --add-gnu-debuglink= option with objcopy and since I renamed the file to tmpfile then the link name is different thus causing the md5 to be different :-( Jan 21 18:46:33 svolpe: just use strip's -o option instead of renaming the file, doesn't that work? Jan 21 18:48:13 steliosk: wget OE.mtn.bz2 Jan 21 18:49:42 pH5, no and I think it is because of the "objcopy --add-gnu-debuglink=/appdir/.debug/app" operation performed by runstrip in OE. I'm assuming that would change the md5sum of the stripped file. Jan 21 18:50:57 Laibsch : i am too lazy and tired. I'll let it update overnight Jan 21 18:51:02 svolpe: hm. then I'd make a backup, strip/objcopy in place and move the original back afterwards. a bit ugly, but that should work. Jan 21 18:52:10 pH5, that is the direction I was moving, either that or figure out how to pug in a function after the RUNSTRIP but before packaging. Jan 21 18:55:45 http://oe.pastebin.com/m67158bdf lists my problem with do_rootfs. It looks like the packages that are "missing" for the image are all oabi, but ipkg only looks in the collie arch Jan 21 18:55:57 Anybody have an idea where this would have to be fixed? Jan 21 19:47:19 http://www.eetimes.com/showArticle.jhtml?articleID=205602076 Jan 21 19:55:55 does angstrom-2007.1-legacy somehow changes the path for the .bb files ? Jan 21 20:04:20 steliosk, I don't think so Jan 21 20:04:38 kergoth: hehe... so I guess you are "good design practice" :) Jan 21 20:05:10 Crofton : Yeah its probably some old dirs i had setup that are no longer valid Jan 21 20:05:28 bitrot Jan 21 20:05:40 I am debugging some code I wrote about a year ago Jan 21 20:05:47 boy I have learned alot in the past year Jan 21 20:05:57 thats true of most years :) Jan 21 20:06:22 Crofton : Will you be at FOSDEM ? Jan 21 20:06:33 yes Jan 21 20:07:13 it looks like i will not make it (agian) Jan 21 20:07:53 thats too bad, it is loads of fun Jan 21 20:08:04 but as usual it will be a last minute decision Jan 21 20:08:04 even though the crowd is much too large Jan 21 20:08:43 well with the office been in such a flux state after the move i am not sure Jan 21 20:11:45 ~pastebin Jan 21 20:11:45 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Jan 21 20:24:42 Crofton: kergoth is right, you always look back at code like that ;-) Jan 21 20:25:05 * RP shudders to think of his dos turboc apps Jan 21 20:29:21 yeah Jan 21 20:29:45 I need to add lots of asserts, so it dies with some explanation. Jan 21 20:37:13 hi kergoth Jan 21 20:38:50 re Jan 21 20:40:06 hey pb_ Jan 21 20:44:25 RP: http://bugs.openembedded.net/show_bug.cgi?id=3710 -> the rtc-sa1100 change I told you about on friday Jan 21 20:59:59 <_saurabh_> getting a python runtime error: dictionary changed size during iteration Jan 21 21:00:08 <_saurabh_> during the final image creation Jan 21 21:00:17 <_saurabh_> the rootfs looks allright Jan 21 21:00:22 <_saurabh_> but the packaging fails Jan 21 21:00:28 <_saurabh_> any clues Jan 21 21:00:30 <_saurabh_> ? Jan 21 21:19:05 * * OE Bug 3710 has been created by  Jan 21 21:19:07 * * Add rtc-sa1100 to module_autoload list in kernel.bbclass Jan 21 21:19:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3710 Jan 21 21:29:25 hvontres|work: I applied 3710, 3868 has been closed as invalid. Jan 21 21:35:30 RP: sounds good Jan 21 21:36:58 RP: I'll take a look at task-base on the way home tonight Jan 21 21:47:05 * * OE Bug 3710 has been RESOLVED (FIXED) by rpurdie(AT)rpsys.net Jan 21 21:47:07 * * Add rtc-sa1100 to module_autoload list in kernel.bbclass Jan 21 21:47:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3710 Jan 21 22:12:58 03rpurdie 07org.oe.dev * r0b7113a8... 10/ (1 classes/kernel.bbclass): kernel.bbclass: Account for sa1100 rtc renaming (#3710) Jan 21 22:17:04 * * OE Bug 3686 has been RESOLVED (INVALID) by rpurdie(AT)rpsys.net Jan 21 22:17:06 * * Kernel linux-rp-2.6. 23 is missing some kernel options for WPA Wireless Jan 21 22:17:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3686 Jan 21 22:21:27 glibc-intermediate 2.6.1 fails for x86 :( Jan 21 22:50:16 there is a guy on angstrom channel who cant compile for his zaurus collie. he is using the.dev branch but set it to behave as an angstrom distro. where should he ask for help? Jan 21 23:10:47 Chriva: Tell him to try "mtn up -r 37149b3d7b17d35e06e93ecd95c7eb36f4425da6", clean out tmp and start over Jan 21 23:10:52 I think that might fix it Jan 21 23:11:10 Tell him to let me know if he succeeds Jan 21 23:11:18 I am probably having the same problem Jan 21 23:12:52 i tried his settings with my .angstrom and it seems to work but .dev failed almost at once. atleast for him and me. have you tried that branch too? =) Jan 21 23:13:52 yes Jan 21 23:14:01 But I compile against Sonkei Jan 21 23:14:12 I am working on figuring out where the problem is Jan 21 23:14:32 Did you try Angstrom stable or 2008? Jan 21 23:14:53 im using the stable release =) Jan 21 23:15:23 Checked out from org.oe.anstrom-2007.12-stable? Jan 21 23:16:28 thats the one yes =) Jan 21 23:32:21 That would be strange Jan 21 23:36:06 im probably just doing the same misstake as him but i cant get .dev to work. even after removing the tmp folder the problem persist Jan 21 23:38:24 Chriva: Error: selected processor does not support `bx r3' Jan 21 23:38:31 is that your error in .dev? Jan 21 23:38:33 Chriva: What timezone are you in? Jan 21 23:38:42 I will have to go to bed now Jan 21 23:38:50 Catch me in about 8 hours Jan 21 23:39:19 Try "mtn up -r 37149b3d7b17d35e06e93ecd95c7eb36f4425da6", clean out tmp and start over Jan 21 23:39:26 That is what I am trying now Jan 21 23:39:37 Things still worked for me at around Jan 10th Jan 21 23:40:26 Laibsch: gmt +1 =) Jan 21 23:41:17 guess my env isn't f*cked up after all =) Jan 21 23:42:03 Chriva: Good Jan 21 23:42:10 Talk to you tomorrow Jan 21 23:42:17 Looking forward to nail this one Jan 21 23:42:25 CoreDump: no glibc fails att glibc-intermediate caused by an error in the c code. i had the same probleme when i tried .dev for my device and it sopped att the exact same spott whit his settings too Jan 21 23:42:40 For me, it works until image generation Jan 21 23:42:49 do_rootfs doing something strange Jan 21 23:43:04 http://oe.pastebin.com/m67158bdf Jan 21 23:43:07 hmm. are yuo running as root? Jan 21 23:43:16 no, of course not Jan 21 23:43:19 I never do that Jan 21 23:43:43 It seems some packages are created in arm-oabi as they should be Jan 21 23:44:04 But they are not being included for image generation which subsequently fails Jan 21 23:44:06 n8 Jan 21 23:44:06 Chriva: could you pastebin your errorß Jan 21 23:44:09 ? Jan 21 23:44:13 Laibsch: n8 Jan 21 23:45:30 CoreDump: sure. just let me reset me enviroment again. i gave upp after a while and went back to angstrom instead of .dev =) Jan 21 23:45:59 Chriva: no hurry, I read the backlog =) Jan 21 23:48:01 CoreDump: hope you dont get angry at me. i have ben writing quite much today. just trying to help as much as i can Jan 21 23:48:30 of course not Jan 21 23:50:54 later all. Time to catch train home Jan 21 23:50:57 check out the graphic! Jan 21 23:50:58 http://www.gumstix.net/ Jan 21 23:52:18 Crofton: not officially launched yet, but not hidden Jan 21 23:52:27 :) Jan 21 23:52:37 oops Jan 21 23:52:50 sakoman: not much longer :) Jan 21 23:53:08 I get so confused buy the .org,.net, and .com version .... Jan 21 23:53:53 that would be nice to show at FOSDEM :) Jan 21 23:54:42 com = company info & commerce Jan 21 23:54:50 net = developer info Jan 21 23:55:54 there's really no attempt to hide it, current developers know about it Jan 21 23:56:09 it just is anywhere near "done" Jan 21 23:56:18 isn't :-) Jan 21 23:56:24 heh Jan 21 23:56:37 sexy graphic, must be done :) Jan 21 23:56:43 :-) Jan 21 23:56:58 These things are never "done", but there isn't critical mass yet Jan 22 00:33:08 new version of automake ... Jan 22 01:03:46 hi there! I'm coming from buildroot and i'm quite new to oe... can anyone explain me where can I pass flags to a configure script during the do_install phase? Jan 22 01:04:27 I looked inside the .inc and .bb files, but the do_install phase is not defined here... Jan 22 02:40:57 03mickeyl 07org.oe.dev * r43f6b35d... 10/ (3 files in 2 dirs): linux-openmoko: remove some useless versions **** ENDING LOGGING AT Tue Jan 22 02:59:56 2008