**** BEGIN LOGGING AT Fri Aug 31 02:59:56 2007 Aug 31 05:22:21 Hmm, how do I restore the original Linux on a SL-C3200? I have installed Angstrom yesterday and played with it, but have to return it today and I need to restore the original UI. I have a 64Mb ROM image, what to do with it? Aug 31 05:39:55 . Aug 31 05:57:06 Hi OE team, how is it going on with the staged package feature? Aug 31 06:00:55 What's the status? I implemented this feature personally, do you interest in it? Aug 31 06:11:17 O_Neil: I would suggest to wait about 3 or 4 hours and talk to koen when he wakes up Aug 31 06:13:32 OK Aug 31 06:14:18 koen: ping Aug 31 06:28:36 good morning all Aug 31 06:33:57 * * OE Bug 2899 has been created by dan.j.williams(AT)intel.com Aug 31 06:33:59 * * Fixup autotools.bbclass for samba-3.0.25c Aug 31 06:34:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2899 Aug 31 06:43:57 * * OE Bug 2900 has been created by dan.j.williams(AT)intel.com Aug 31 06:43:59 * * samba-3.0.25c support Aug 31 06:44:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2900 Aug 31 06:47:24 Strange, my Zaurus boots up to the "Sharp" logo and then says "HD error!" and that's all Aug 31 06:47:51 what can I do at this point? It seems I have access to command-line since 'poweroff' works, but it's 1-line height only Aug 31 06:49:33 [D]+[B] doesn't work too Aug 31 07:09:13 zap: only few people here as zaurus pda Aug 31 07:09:38 zap: better ask on openzaurus/angstrom/cacko/zaurus channels of look into oesf forums Aug 31 07:11:41 aha, thank you. Sorry I used the wrong channel Aug 31 07:26:14 hi mickey|zzZZzz Aug 31 07:27:40 O_Neil: you can also write mail to OE ML Aug 31 07:30:53 hi greentux Aug 31 07:31:22 hi hrw Aug 31 07:33:18 koen: arm crunch patchset is probably broken Aug 31 07:33:36 morning guys Aug 31 07:34:09 mickey|MDD07: how goes mdd? Aug 31 07:42:42 cool. first session right now. listening to Vladi from TT Aug 31 07:42:49 after him, Stefan will talk about OpenEZX Aug 31 07:42:51 and then me Aug 31 07:44:15 you about openmoko? Aug 31 07:46:41 mickey|MDD07: Do they record videos of the talks? Aug 31 07:59:17 CM: unfortunately not. it is a tiny conference Aug 31 07:59:25 hrw: yes, talking about openmoko in a few Aug 31 07:59:35 Ok, too bad Aug 31 08:01:05 well, not mentioning anything new today anyways ;) Aug 31 08:01:21 real news come in next presentation in Zurich Aug 31 08:01:31 When is that? Aug 31 08:01:41 19.-20. September Aug 31 08:01:55 Ok, thanks :) Aug 31 08:02:08 n Aug 31 08:02:09 p Aug 31 08:05:32 mickey|MDD07: when you will announce openmoko OE maintainer? Aug 31 08:06:09 mid / late september Aug 31 08:06:26 so in a couple of weeks Aug 31 08:06:44 bbiab, need to talk in a bit Aug 31 08:07:20 I'm looking forward to that. openmoko-image has been almost impossible to build lately Aug 31 08:07:43 morning Aug 31 08:11:46 CM: fails on gtk? Aug 31 08:12:05 hrw: No, kernel right now Aug 31 08:12:18 CM: pastebin log please Aug 31 08:13:32 hrw: http://rafb.net/p/JIEaUy45.html Aug 31 08:14:00 It's not just me, rwhitby, yness, ScaredyCat.. All get the same error Aug 31 08:14:29 its those patches koen added to gcc, they are bolloxed Aug 31 08:14:52 Yes, seems to be something with gcc Aug 31 08:15:08 revert, revert, revert, ... Aug 31 08:15:49 or were they glibc, but the FPU ones Aug 31 08:17:19 * hrw is disapproving it Aug 31 08:18:07 XorA: gcc ones Aug 31 08:18:29 althpugh I cant see where they are pulled in Aug 31 08:19:41 NOTE: Applying patch 'arm-crunch-cirrus-bugfixes.patch' Aug 31 08:19:41 NOTE: Applying patch 'zecke-no-host-includes.patch' Aug 31 08:19:41 NOTE: package gcc-cross-4.1.2-r5: task do_patch: completed Aug 31 08:25:31 hi zecke Aug 31 08:30:30 ah, stupid me, I had a gcc before crunching :-) Aug 31 08:32:12 lucky you Aug 31 08:32:21 hrw: you will have to dissaprove it then push a r6 PR bump Aug 31 08:33:05 I did Aug 31 08:34:22 hrw: I see it, musta missed it by seconds in first sync :-) Aug 31 08:35:33 brb Aug 31 08:37:17 CM: update and try again Aug 31 08:55:48 argh.. neo has to be connected to host not hub to dfu ;( Aug 31 08:56:36 hrw: yes Aug 31 08:57:35 I am out of host ports ;( Aug 31 08:58:09 remind me never to install resolvconf again Aug 31 08:58:11 good morning Aug 31 08:58:16 hi florian Aug 31 09:00:35 03hrw 07org.oe.dev * rd7c8a05b... 10/ (50 files in 3 dirs): disapproval of revision '4fb1a7447c5d8ace96a5ac6358c7fe10f0d951c4' Aug 31 09:00:42 03hrw 07org.oe.dev * r2a756f49... 10/ (1 packages/gcc/gcc-cross_4.1.2.bb packages/gcc/gcc_4.1.2.bb): gcc/gcc-cross 4.1.2: bump PR to build working compiler after broken ARM Crunch Aug 31 09:15:45 hrw: Thanks, that did it. package linux-gta01-2.6.22.5: completed Aug 31 09:17:37 np Aug 31 09:18:58 hi lrg Aug 31 09:20:18 hey florian Aug 31 09:26:30 ~fish lrg Aug 31 09:26:31 * ibot slaps lrg around with a large trout Aug 31 09:37:31 ~lart linux-gta01 Aug 31 09:37:31 * ibot chops linux-gta01 in half with a free AOL CD Aug 31 09:38:14 heh Aug 31 09:38:18 ~lart hrw Aug 31 09:38:18 * ibot breaks out the Hoover and sucks up hrw Aug 31 09:38:23 don't you chop our kernel :D Aug 31 09:39:12 mickey|MDD07: it has wrong deps Aug 31 09:39:30 quilt-native is build always so no need for it in recipe Aug 31 09:39:35 mickey|MDD07: hey Aug 31 09:39:53 uboot-openmoko is not needed to build linux-gta01. u-boot-mkimage-openmoko-native is enough Aug 31 09:40:17 hi zecke ! Aug 31 09:40:35 hrw: let me fix it. have some other change as well Aug 31 09:41:09 don't lart because of historic cruft Aug 31 09:41:14 mickey|MDD07: and let it fetch linux-2.6.22.tar.bz2 and stable patches to it Aug 31 09:41:16 this dependency comes from the days where we used patcher a lot Aug 31 09:41:28 hrw: no time for that now Aug 31 09:41:38 in october we have our dedicated guy :) Aug 31 09:41:45 mickey|MDD07: each time you (openmoko) change stable to stable each dev has to fetch 40M archive instead of 40K patc Aug 31 09:41:48 h Aug 31 09:41:49 hey mickey|MDD07 Aug 31 09:41:54 hi XorA Aug 31 09:42:07 * XorA gets the popcorn for the OE smackdown Aug 31 09:42:17 XorA: haha Aug 31 09:42:39 mickey|MDD07: any crime you commit is low compared to koens today anyway :-) Aug 31 09:42:50 what's he doing? Aug 31 09:43:02 mickey|MDD07: broke gcc :-) Aug 31 09:43:08 he broke ARM toolchain Aug 31 09:43:41 argh Aug 31 09:43:49 anyone fixed it already>? Aug 31 09:43:52 already reverted Aug 31 09:43:57 mickey|MDD07: yes, hrw disapproved Aug 31 09:44:01 thanks hrw Aug 31 09:44:10 guess he was too busy with blaming me for sometimes forgetting --author Aug 31 09:44:14 grmbl Aug 31 09:45:26 mickey|MDD07: BTW midori, I can launch from ssh, segfaults if I launch from rxvt, seen anything like that? Aug 31 09:46:12 oh strange Aug 31 09:46:18 didn't try from rxvt. will do Aug 31 09:46:31 env changes Aug 31 09:47:55 likely, ya Aug 31 09:48:53 mickey|MDD07: is ok, Ill investigate sometime, just wondered if you had seen it Aug 31 09:49:16 sweet browser though Aug 31 09:49:26 yeay Aug 31 09:49:31 zecke's will outperform that though Aug 31 09:49:36 * mickey|MDD07 eyes zecke Aug 31 09:49:59 * zecke will use midori and use red buttons ;) Aug 31 09:50:15 zecke: speed it up old man :-D Aug 31 09:53:37 hehehe Aug 31 10:00:13 ~hail koen for his good Blackfin work. Aug 31 10:00:14 * ibot bows down to koen for his good Blackfin work. and chants, "I'M NOT WORTHY!!" Aug 31 10:05:20 ~lart 2.6.22.65 archive fetching Aug 31 10:05:21 * ibot squeezes 2.6.22.65 archive fetching till 2.6.22.65 archive fetching turns blue like papa smurf Aug 31 10:12:55 someone remember when OE switch to SRCREV? Aug 31 10:15:30 hrw: 04092007 Aug 31 10:15:50 ok Aug 31 10:15:58 ~hail SRCREV Aug 31 10:15:58 * ibot bows down to SRCREV and chants, "I'M NOT WORTHY!!" Aug 31 10:16:14 * stefan|mdd is really looking forward to it. :) Aug 31 10:16:46 stefan|mdd: ;D Aug 31 10:16:58 stefan|mdd: you have microsdhc card? Aug 31 10:17:17 hrw: yeah, works great. :) Aug 31 10:17:38 stefan|mdd: can you check does it works in pxa machine with 2.6? Aug 31 10:18:19 as zaurus users are asking from time to time and I have no idea how to respond ;) Aug 31 10:18:26 hrw: Will test it on one of my EZX phones, but I fear not in the next week Aug 31 10:18:49 hrw: I add it to my TODO list, perhaps I can find time for it sooner. Aug 31 10:18:59 thx Aug 31 10:20:11 * XorA should get some SDHC cards for testing Aug 31 10:37:53 * koen returns from work Aug 31 10:38:00 sorry about breaking gcc 4.1.2 Aug 31 10:38:52 at least you knew it was me by looking at the author cert :) Aug 31 10:38:58 03mickeyl 07org.oe.dev * rd95b16d3... 10/ (3 files in 3 dirs): linux-gta01: make it compatible w/ gta02, adjust DEPENDS Aug 31 10:39:03 03mickeyl 07org.oe.dev * r177c90ad... 10/ (1 conf/machine/fic-gta02.conf): linux-gta01: use linux-gta02 as kernel Aug 31 10:39:10 03mickeyl 07org.oe.dev * r1da9580d... 10/ (1 packages/linux/linux-gta01.inc): linux-gta01: use gta01 defconfig for gta02 Aug 31 10:39:17 03mickeyl 07org.oe.dev * r735f3c43... 10/ (13 files in 4 dirs): Aug 31 10:39:17 rename linux-gta01 to linux-openmoko, this kernel will be useful on a family of devices, Aug 31 10:39:17 GTA01, GTA02 being the first two models. Aug 31 10:47:28 hmm.. new work to do then Aug 31 10:49:01 at we can stop complaining about linux-gta01 :) Aug 31 10:49:52 mickey|MDD07: will you or your mistic new maintainer complain much if do_prepatch will be changed to do_postpatch? Aug 31 10:50:19 and using linux-2.6.00 + minor patch? Aug 31 10:50:29 s/00/foo/ Aug 31 10:50:34 koen: exactly Aug 31 10:50:38 hi mallum Aug 31 10:50:43 re Aug 31 10:50:51 hey bearded super hero from london Aug 31 10:51:04 hey hrw :-) Aug 31 10:51:13 hey koen ;-) Aug 31 10:52:26 hehe Aug 31 10:56:27 rootfs_deb almost works in OE Aug 31 10:56:36 koen: cool! Aug 31 10:56:40 it needs some small tweaks from poky Aug 31 10:56:53 koen: what host distro are you running? Aug 31 10:57:11 my problem was (apart from some others), that it also did not work because I had a non-deb system Aug 31 10:57:11 ubuntu one, debian on another Aug 31 10:57:20 it seemed to reuse some of the host deb stuff Aug 31 10:57:41 zecke: I heard you may be going to the CELF conference in Linz? Aug 31 10:58:01 Jin^eLD: it depends. Aug 31 10:58:32 well, let me know if you go, would be nice to have a chat or something Aug 31 10:58:43 I live in vienna so I will for sure be there Aug 31 11:15:35 hmm, I tried to abort bitbake using ctrl-c and somehow I am back in the shell but wget seems to be still working in the background, writing over my terminal Aug 31 11:15:36 :) Aug 31 11:15:55 that does not seem right somehow? Aug 31 11:31:28 In case someone is bored: http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-August/002929.html Aug 31 11:36:23 good afternoon all Aug 31 11:38:20 I have a problem with the bzImage I built with bitbake. It can't load its root file system .. Aug 31 11:39:12 I have checked the defconfig, but it seems that all required options are set to 'y'. Aug 31 11:39:58 A kernel that I built two month ago with oe works. And its 1.4 MB instead of 1.2 MB for the new one. Aug 31 11:40:41 Any idea what I might be missing here ? Aug 31 11:41:01 jan_js: sl-a300? Aug 31 11:42:37 hrw: i586, a amd geode board .. Aug 31 11:43:03 jan_js: which OE machine it is? Aug 31 11:43:45 hrw: generic x86, modified slightly Aug 31 11:46:04 hrw: sorry, i586-generic, not x86. Aug 31 11:47:03 hrw: I am using grub, which works fine with the older kernel. Aug 31 11:48:28 re Aug 31 11:48:51 jan_js: which board you have? Aug 31 11:50:48 d/win 1 Aug 31 11:50:49 fsck.. my desktop machine has connections for 4 usb ports on board (via pins). Aug 31 11:50:56 and 3 of them do not work ;( Aug 31 11:51:34 Jin^eLD: kill %1 Aug 31 11:52:53 reboot. Aug 31 11:54:47 hrw: It's a advantech-pcm-3353F Aug 31 11:55:49 It has a CF slot, from where I boot the system Aug 31 12:02:16 jan_js: I use alix1c board - geode lx800 Aug 31 12:02:57 hrw: so what OE machine do you use - i586 ? Aug 31 12:03:27 jan_js: alix Aug 31 12:03:57 jan_js: but it is now i686 and will be switched to 'geode' which gets k6-2 gcc optimizations Aug 31 12:05:04 hrw: My CPU is a geode gx-mmx - maybee similar ? Aug 31 12:05:39 gx is older geode Aug 31 12:06:04 geode gx is based on cyrix MediaGX, then geodeLX was born Aug 31 12:08:13 hrw: So I might better stick to i586 then. I see in the alix machine conf that it uses "linux" as the kernel instead of linux-x86 as I do. Aug 31 12:08:34 I will check for the differences, I think, especially in the defconf. Aug 31 12:11:01 jan_js: one day I will merge linux-x86 kernels into linux one Aug 31 12:12:14 hrw: whats the difference right now ? Aug 31 12:14:52 name only rather Aug 31 12:21:16 jan_js: and version - linux-x86 is 2.6.20, linux is 2.6.21/22/23-rc3 Aug 31 12:25:37 torpor: well, I was kind of wondering why bitbake did not clean up the dl processes when exiting, that was the main question I guess Aug 31 12:29:34 Jin^eLD: ? Aug 31 12:40:19 yes? Aug 31 12:40:39 "wondering why bitbake did not clean up the dl processes when exiting"? Aug 31 12:40:57 oh.. you mean my talk with torpor? well I started a build, pressed ctrl-c, and bitbake seemed to have quit but wget was stil running Aug 31 12:41:22 torpor suggested to kill %1 but I asked not because of not knowing hot to kill it but mostly wondering why bitbake did not take care of it when exiting Aug 31 12:41:22 multihtreaded bitbake forks of new shells for stuff Aug 31 12:49:20 hmm Aug 31 12:49:22 NOTE: package u-boot-git-r1: task do_fetch: failed Aug 31 12:49:42 fresh pull && update, trying Crofton's davinci image, angstrom Aug 31 12:51:19 somebody should teach apt some funny response to "*u-boot-git*do_fetch*failed" :) Aug 31 12:52:30 Jin^eLD: the denx git repo is being down Aug 31 12:52:39 probably upgrading uboot on the server ;) Aug 31 12:52:42 doh.. :( Aug 31 12:52:43 hehe Aug 31 12:53:12 ~lart uboot-git fetching Aug 31 12:53:13 * ibot pulls out his louisville slugger and uses uboot-git fetching's head to break the homerun record Aug 31 12:54:47 that crunch patches are those mentioned by martin guy? Aug 31 12:54:52 ^koen Aug 31 12:57:50 ynezz: yes, and you need gcc 4.2.1 for it Aug 31 12:58:00 since they break 4.1.2 for non-trivial apps Aug 31 12:58:52 * koen does a build from scratch with the merged classes Aug 31 13:04:30 poky ones? Aug 31 13:05:17 yes Aug 31 13:05:21 the ones I sent to the list Aug 31 13:05:29 * koen hugs meld Aug 31 13:17:26 hrw:what was the gcc-cross fix you put in? I was making a new build last night and it was on the gcc-cross build all night and still on it when I got in. would that have anything to do with it? Aug 31 13:17:36 hrw: that is, I got it before your change Aug 31 13:20:17 christopher: gcc 4.1.2-r6 is current one Aug 31 13:22:35 hrw: so, only to update, not to address/fix Aug 31 13:23:12 hrw: I restarted a build and enabled debug output so I can be sure it is still running and compiling. this is on a new machine that should go through compiles fast so that was why I was suprised with a build taking overnight. Aug 31 13:23:19 and still on the same thing Aug 31 13:27:28 in -r5 were applied some patches which broken arm builds, so -r6 address this issue -- removed those offending patches Aug 31 13:29:09 and I think, that itsn't cause of your neverending build :) Aug 31 13:30:56 ynezz: yeah, didn't think so. :( I was at r4 so I didn't have those patches. Aug 31 13:31:21 ynezz: I have on the parallel make and this is a dual core but guess it is no match for these builds. Aug 31 13:32:10 i'm running 10 threads in parallel make on dual core T7100 Aug 31 13:32:35 no problems so far Aug 31 13:32:47 ynezz: and how long does an image build take? is there a way to pass off more to both or will that happen automatically? Aug 31 13:33:17 i don't measure it, but it's fast enough for me Aug 31 13:34:50 angstrom-minimal-image under 30min or so Aug 31 13:34:58 * * OE Bug 2901 has been created by autobuild(AT)openembedded.org Aug 31 13:35:00 * * expect-5.44.1-r0-do_configure Aug 31 13:35:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2901 Aug 31 13:35:04 ok. fair enough. On the old machine, it would have certainly been done overnight and that's why I am concerned. oh geez, under 30 minutes? Aug 31 13:35:25 as i said and don't know exactly :) Aug 31 13:35:37 but it's not noticeable as openmoko-image Aug 31 13:35:51 s/and/I/ Aug 31 13:35:58 * * OE Bug 2902 has been created by autobuild(AT)openembedded.org Aug 31 13:35:59 * * expect-5.44.1-r0-do_install Aug 31 13:36:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2902 Aug 31 13:36:50 ynezz: ok, I'll kick off a new build and see how long it takes. Aug 31 13:36:57 * * OE Bug 2901 has been RESOLVED (INVALID) by Aug 31 13:36:59 * *  expect-5.44.1-r0-do_configure Aug 31 13:37:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2901 Aug 31 13:37:43 christopher: what are you trying to build? Aug 31 13:39:56 ynezz: I was building angstrom-console-image to try and get back to a working console. I had one working two days ago and luckily kept an older one around to proove it. when I build now, I get the no console found error Aug 31 13:40:22 ynezz: so, I got a new computer and have been setting it up for builds on it (as opposed to the old one) so I can verify builds etc. Aug 31 13:40:42 ynezz: just using angstrom and a sarge-at91 machine since it uses the same proc Aug 31 13:41:54 christopher: good to know that sarge config is usable for someone ;) Aug 31 13:42:18 hrw: didn't you add it in there ? Aug 31 13:42:24 christopher: I did Aug 31 13:42:41 hrw: yeah, I thought I remembered that right. figured I'd go with something that 'worked' Aug 31 13:43:14 hrw: although I am getting a little frustrated because I am not sure how to configure the kernel with oe (may have to patch in the oe kernel)... Aug 31 13:43:35 with it not loading any of the modules.dep / creating it and with it not running other things that are supposed to be done as postinsts... Aug 31 13:43:48 I just don't know what I am doing wrong since I am using the code everyone else is straight from OE. Aug 31 13:44:38 with buildroot, the *1* good thing I can say is that I could take a look at the rootfs that it built and everything was there so I could tell if something was ok or not. I'm just not used to it running all of those scripts at startup. Aug 31 13:44:57 * * OE Bug 2902 has been RESOLVED (INVALID) by Aug 31 13:44:59 * *  expect-5.44.1-r0-do_install Aug 31 13:45:00 and am concerned about boot-up time anyway so those add another 10-15 seconds which is not so good Aug 31 13:45:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2902 Aug 31 13:45:35 christopher: add your machine into OE and generate config for it. that will resolve first problem. Aug 31 13:46:45 hrw: meaning the kernel .config ? or the kernel itself with my changes? Aug 31 13:48:35 config + eventual patches Aug 31 13:49:40 hrw: ok. i do have the patch. it is enough to just add it as a file:// and it will know to patch it? I think I saw that in a bb file Aug 31 13:50:05 hrw: same with .config? do I just make that into a defconfig as others have done? like for busybox etc Aug 31 13:50:29 hrw: I'd like to have more options enabled like command history etc for busybox, so that's one thing I'll want to change on mine Aug 31 13:51:12 christopher: look how sarge-at91 patches are handled in kernel recipe and follow it Aug 31 13:51:22 hrw: ok. I'll follow those along Aug 31 13:51:25 for busybox it is other thing Aug 31 13:51:28 ~hail linux.inc Aug 31 13:51:29 * ibot bows down to linux.inc and chants, "I'M NOT WORTHY!!" Aug 31 13:51:43 * hrw just generated modules.dep on host for target device Aug 31 13:52:17 more nokia FUD about community software: http://jaaksi.blogspot.com/2007/08/summer-teasers.html Aug 31 13:52:25 no wonder hildon is such a POS Aug 31 13:52:36 morning Aug 31 13:53:05 hey chouimat Aug 31 13:53:42 ../cross/bin/arm-poky-linux-gnueabi-depmod-2.6 -a -b /home/hrw/devel/OH/poky/build/tmp-poky/rootfs/ -F ../staging/fic-gta01-poky-linux-gnueabi/kernel/System.map-2.6.21.6-moko11 2.6.21.6-moko11 Aug 31 13:55:32 koen: come on Aug 31 13:56:21 koen: the Fear part comes from the Neo being almost all the n800 is, but with capability to phone as well Aug 31 13:56:35 koen: so if it succeeds it proves Nokias analysts wrong Aug 31 13:56:47 yes Aug 31 13:57:00 and nokia is almost certainly doing a linux phone Aug 31 13:57:18 that become evident with Jaksis speach as GUADEC Aug 31 13:57:34 N95 + linux?? Aug 31 13:57:44 bah nokia Aug 31 14:04:03 XorA: there were rumours about n95 linux Aug 31 14:04:37 hrw: its the most obvious path Aug 31 14:05:01 make[2]: Entering directory `/home/mranostay/openembedded/tmp/work/centrality-linux-gnueabi/linux-omap-2.6-2.6.9-omap1-r0/linux-2.6.9' Aug 31 14:05:01 CC scripts/mod/empty.o Aug 31 14:05:01 cc1: error: unrecognized command line option "-mapcs-32" Aug 31 14:05:01 cc1: error: unrecognized command line option "-mshort-load-bytes" Aug 31 14:05:12 Anyone have any clue on this? Aug 31 14:05:17 2.6.9!?!?!?!? Aug 31 14:05:20 Oh and hi all :) Aug 31 14:05:27 that's like 2 years old Aug 31 14:05:44 koen: I know the openmoko-image picked that for some reason Aug 31 14:05:48 mranostay_work: switch to 2.6.2x please Aug 31 14:05:55 mranostay_work: which machhine you build for? Aug 31 14:06:25 centrality Aug 31 14:06:54 which lack kernel in OE? Aug 31 14:07:52 no it built linux-2.6.21 Aug 31 14:08:13 christopher: I do not know why you have problem with modules.dep - all my images here have them generated during creating Aug 31 14:10:07 koen: last year I had to "service" a server .... they were still running linux 1.2 on it :( Aug 31 14:10:37 chouimat: wow your kidding right? Aug 31 14:11:06 no Aug 31 14:11:27 mranostay_work: and another one was running freebsd 2.1/2,.2 Aug 31 14:11:28 that is what like ~10+ years old? Aug 31 14:11:29 in some areas linux 2.0.X was more featureful than 2.6 Aug 31 14:12:04 mranostay_work: close to 13 Aug 31 14:12:20 if it works, dont fix it way Aug 31 14:12:41 and the script kiddiez cant hack it if its not 0day Aug 31 14:12:50 hehe I was in second grade then :) Aug 31 14:13:03 true ... and to get the apps running on those working on something more modern will require a complete rewrite Aug 31 14:16:02 NOTE: package libgsmd-0.1+svn20070831-r22: task do_install: started Aug 31 14:16:04 ERROR: function do_install failed Aug 31 14:18:35 hi Cliff Aug 31 14:19:32 hrw: yeah, Ido have that. so, when you look in lib/modules under the rootfs, you see modules.dep ? Aug 31 14:19:46 yes Aug 31 14:19:55 christopher: as long image contain any modules Aug 31 14:20:58 hrw: well, that may be an issue. In my kernel, I don't have modules and I think I copied over my config to oe because the image wouldn't build due to lack of machine definition. Aug 31 14:21:26 hrw: since oe depends on modules, is there any way to have it...not depend on modules and have them simply built into the kernel? Aug 31 14:23:13 hrw: In other words, if I incorporate my kernel in oe, it will have my config (no modules), but then things built with oe are trying to load modules.dep but I don't know what modules they are trying to load Aug 31 14:24:48 christopher: during start of OE generated images there is step when modules are loaded. if there are no modules then it is still ok Aug 31 14:25:58 hmm Aug 31 14:26:09 with the latest .dev qt no longer compiles Aug 31 14:26:13 this worked yesterday Aug 31 14:26:22 mickeyl: compiles or stage? Aug 31 14:26:29 compiles Aug 31 14:26:30 mickeyl: and how it breaks? Aug 31 14:26:31 3.0/config.tests/unix/compile.test: line 54: 25682 Illegal instruction Aug 31 14:26:41 mickeyl: gcc 4.1.2-r6? Aug 31 14:27:02 hrw: well, I guess it depends on what you mean by ok. there are programs that won't load because it can't find modules.dep Aug 31 14:27:08 christopher: we prefer to have modules as then kernels are smaller so they boot faster Aug 31 14:27:20 hrw: yes, 4.1.2-r6 Aug 31 14:27:30 what kind of changes did we have? Aug 31 14:27:42 mickeyl: NOTE: package qt4-x11-free-4.3.0-r6: task do_compile: started Aug 31 14:27:47 anyone been working on libgsmd today? Aug 31 14:27:58 mickeyl: r6 == r4 Aug 31 14:28:07 well, it breaks in configure here Aug 31 14:28:13 hmm, strange Aug 31 14:28:17 *sigh* Aug 31 14:28:52 mickeyl: will openmoko provide any external toolchain? Aug 31 14:28:59 sure Aug 31 14:29:37 this would lower amount of problems which devs have Aug 31 14:29:40 hmm, somehow the PREFERRED_PROVIDER_virtual/kernel in my machine conf is ignored; trying to build angstrom, -DDD does show that the correct confs are being parsed Aug 31 14:29:46 any hints on why this could be happening Aug 31 14:29:47 ? Aug 31 14:30:08 Jin^eLD: oe kernel recipe or own? Aug 31 14:30:12 own Aug 31 14:30:26 Jin^eLD: COMPATIBLE_* flags checked? Aug 31 14:30:38 hrw: what's the status of toolchain packages in OE ? Aug 31 14:30:42 well the thing is, it works for me here, I replicated that on a slightly older host distro on my collegues PC Aug 31 14:30:52 for me it works, for him it does not, allthough he is using my setup and same bitbake version Aug 31 14:31:07 mickeyl: meta-toolchain should work. for openmoko you need to extend it with more libs Aug 31 14:32:01 I'm merging in poky bbclasses so we can use their toolchain stuff Aug 31 14:32:07 see RFC on oe-devel Aug 31 14:32:28 our toolchains works :D Aug 31 14:32:59 hrw: I think the problem is that the collection is being ignored... but with same bitbake version.. hmm, that should be ok I guess? Aug 31 14:33:42 I do not use collections so hard to tell Aug 31 14:33:49 ~hail screen -x Aug 31 14:33:50 mhm Aug 31 14:33:50 * ibot bows down to screen -x and chants, "I'M NOT WORTHY!!" Aug 31 14:34:25 hmm Aug 31 14:35:13 * hrw -> lunch finally Aug 31 14:37:05 ~hail Linus Torvalds Aug 31 14:37:06 * ibot bows down to Linus Torvalds and chants, "I'M NOT WORTHY!!" Aug 31 14:39:50 how can i force bitbake to use a different kernel .bb? Aug 31 14:48:32 PREFERRED_PROVIDER_virtual/kernel "" Aug 31 14:51:13 mickeyl: I'd just RTFM, thanks :) Aug 31 15:32:57 * * OE Bug 2903 has been created by autobuild(AT)openembedded.org Aug 31 15:32:59 * * claws-plugin-maildir-0.24.4-r0-do_configure Aug 31 15:33:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2903 Aug 31 15:42:24 ~hail ibot Aug 31 15:42:25 * ibot bows down to ibot and chants, "I'M NOT WORTHY!!" Aug 31 15:47:42 r Aug 31 15:47:43 e Aug 31 15:47:57 * * OE Bug 2904 has been created by lukas.gorris(AT)gmx.de Aug 31 15:47:59 * * Dependencies not matched installing openmoko stuff Aug 31 15:48:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2904 Aug 31 15:48:43 mickeyl: NOTE: package qt4-x11-free-4.3.0-r6: task do_package: completed Aug 31 15:49:22 on my x86_64 machine, qmake-native ends up w/ illegal statements when qt detects the machine as x86_64 Aug 31 15:49:27 with my uname hack it works again Aug 31 15:51:42 mickeyl: that is on x86 host Aug 31 16:04:58 is shasum-native-1.0 failing to build for anyone else? Aug 31 16:05:11 not here Aug 31 16:23:03 * mwester shouts a stream of obscenities as gcc rebuilds, and what what was supposed to be a simple update is not going to be an all-day rebuild. Aug 31 16:23:30 lart OE for not understanding the benefits of locking-down a toolchain. Aug 31 16:23:57 mwester: Locking down the toolchain is a distro decision Aug 31 16:24:59 RP: no - he is talking about PR bumping Aug 31 16:25:42 mwester: you can make own branch where you will cherrypick updates from .dev Aug 31 16:30:36 RP: yes, it is, but seldom done because OE doesn't have a good built-in model for handling such a thing. Can it even deal with a package that has a different version for the -native.bb than the target .bb? Aug 31 16:32:30 hrw: yes, I can. But that's the wrong way around. Normal operation, once a distro gets to a certain development stability, should be for everything -native to be isolated, and the typical developer should _choose_ when they are ready for a mass update of the toolchain and native packages. Aug 31 16:33:32 mwester: but .dev is not distro related Aug 31 16:34:38 Ok. I fail to see the significance of that, though. Aug 31 16:36:06 mwester: when OZ was ready to release I branched it and then only needed stuff came to branch Aug 31 16:37:06 From a practical point-of-view, someone will implement a solution, because mass rebuilds for hundreds of users because someone changed something trivial in a bb somewhere is simply unacceptable for most developers. Aug 31 16:37:26 If it's a branch, and "cherry picking", that's fine. If it's a lockdown of the -native packages, that's fine too. Aug 31 16:38:51 It's philosophy I guess. (I just happen to have nothing better to do than contemplate philosophy for the next 8 hours or so until my build completes). :-D Aug 31 16:38:52 mwester: gcc change in general was 'trivial' but needed. we already blamed koen on it Aug 31 16:39:22 hrw: I understand that one, and it's regrettable. But the general case still stands. Aug 31 16:40:08 If my project is at a critical juncture, how do I as a developer with a time-limit, protect myself from checkins of code that I do not expect? Aug 31 16:40:57 mwester: you branch then Aug 31 16:41:01 Every project in the world deals with this issue, and I think that OE is unique in that it chooses a default which has no barriers or guards at all. Aug 31 16:41:25 hrw: my point is that the ORIGINAL change should have been on a dev branch. Aug 31 16:41:38 mwester: and we lack !dev branch Aug 31 16:42:03 Precisely. That's what makes OE unique from every other embedded project I've seen. Aug 31 16:42:26 There is no "stable" branch, or "mainline" that has any guarantee of remaining stable. Aug 31 16:42:34 is there a list of maintainers for packages/metadata in the different distros? Aug 31 16:42:44 dcordes: read MAINTAINERS file Aug 31 16:42:51 ok thank you Aug 31 16:43:15 mwester: we can probably do stable branch but then the problem is who will maintain it Aug 31 16:44:24 Yep. That too is the same issue faced by every other SCM environment. Everyone wants the changes they desire to be merged automatically, and changes they do not wish to see to be ignored -- but nobody is every willing to do the hard work to manage same. Aug 31 16:45:11 obviously koen|away is the maintainer for openmoko recipes in angstrom Aug 31 16:45:16 koen|away: you around? Aug 31 16:45:45 dcordes: no Aug 31 16:45:52 The key is as I said before: "hundreds of users" -- sheer scale will overcome that objection, and somebody somewhere will do it. My only question is if they will do it in OE, or if they will simply copy the toolchain and -native stuff out of OE. Aug 31 16:45:59 dcordes: openmoko project is maintainer of openmoko recipes Aug 31 16:46:05 dcordes: angstrom just use them Aug 31 16:47:01 mwester: I think that OE lack three things now: 1. autobuilders 2. stable branch 3. maintainers for 2 Aug 31 16:47:18 I completely agree. Aug 31 16:48:24 hrw: I don' think the problem I experience with the openmoko recipes is present in openmoko itself but only in angstrom. Aug 31 16:48:27 !oebug 2904 Aug 31 16:48:29 * * Bug 2904, Status: NEW, Created: 2007-08-31 08:47 Aug 31 16:48:30 * * lukas.gorris(AT)gmx.de: Dependencies not matched installing openmoko stuff Aug 31 16:48:31 * * http://bugs.openembedded.org/show_bug.cgi?id=2904 Aug 31 16:48:46 doing stable branch without autobuilders is useless as there will be no way to autocheck does all is ok. Aug 31 16:48:59 and lack of maintainers == no commits fixing stable branch Aug 31 16:49:16 dcordes: patches welcome Aug 31 16:50:42 hrw: is openmoko working with the openembedded mainline anyway? Aug 31 16:51:06 yes Aug 31 16:51:46 Thing is I believe that 2904 is caused by somebody providing only partially upgrade to the openmoko metadata in oe so the openmoko-task-* things are missing the latest branch versions from openmoko Aug 31 16:51:47 ~lart gsmd Aug 31 16:51:47 * ibot urinates on gsmd Aug 31 16:57:30 hrw: I seriously would like to patch this but I don't know where to start Aug 31 16:59:07 mwester: but then, even in the stable branch you may have tool chain updates ... I guess there could be advance notification though ... Aug 31 16:59:49 mwester: I usually just lock down a stable version of OE for my projects, and only move forward when I know I'll have time to work through any issues that come up. Aug 31 17:00:40 bye Aug 31 17:01:04 dcordes: for me openmoko-devel-image builds for fic-gta01 and for alix Aug 31 17:01:39 I'm not building openmoko-devel-image but angstrom-x11-image for akita Aug 31 17:05:07 maybe it would be wise to just build openmoko-devel-image for akita for my purpose :D Aug 31 17:24:33 is the openmoko binary repo upgraded manually or with buildbot? Aug 31 17:24:43 sorry I mean angstrom repo Aug 31 17:26:48 manually by feed maintainers (check the angstrom site) Aug 31 17:31:52 RP, ping Aug 31 17:41:32 I just started building openmoko-devel-image for machine akita distro openmoko. Yesterday I finished a complete angstrom-x11-image also for akita. Now why does bitbake have to rerun all the tasks? Aug 31 17:42:03 I somehow expected the moko image to build faster because I'm building for the same arch Aug 31 18:13:13 hi Aug 31 18:25:59 Hi! I'd like o update http://www.openembedded.org/wiki/OEandYourDistro but I can not see any way to create an account to do so. Aug 31 18:26:18 sort-of wondering what's the use of a closed wiki? Aug 31 18:30:33 why do you think it's closed? Aug 31 18:30:39 you don't need an account to edit it Aug 31 18:32:27 because "Edit" shows me a little text input to make a comment about thet change, but no input to actually make the change? Aug 31 18:32:56 lol, drupal is... Aug 31 18:33:30 it does show a Preview button tho. Aug 31 18:34:51 agreed, our drupal installation is still messed up Aug 31 18:38:51 ok. well. Just wanted to edit the debian-unstable info. monotone 0.36 works from today's mokomakefile, vs 0.35 bailing out (three weeks ago). specifically, 0.33 is *NOT* 'current debian unstable', but at least a month old, if not much older. Aug 31 18:39:30 i'll post to an openmoko mailing list, might occasionally check the OE wiki, whether I can edit (but don't trust I will actually do that...) Aug 31 18:45:20 03koen 07org.oe.dev * r3f5ee415... 10/ (3 files in 2 dirs): angstrom: more .deb integration Aug 31 19:18:30 how can I make an *.ext2 image out of an image I build with OE? (I only get .jffs2) Aug 31 19:21:43 dcordes: should be in your machine conf file with file types. just change 'jffs2' to 'jffs2 ext2' Aug 31 19:21:52 dcordes: and rebuild Aug 31 19:21:55 ok Aug 31 19:24:06 hrw: I tried building the image again and it did not load up modules.dep. what distro/machine are you using? is it angstrom and ? Aug 31 19:48:05 somehow I'm getting no zImage after building angstrom-console-image and angstrom-minimal-image Aug 31 19:48:17 is that normal or did I setup something ina wrong way? Aug 31 19:49:27 docrdes hm Aug 31 19:49:30 which machine? Aug 31 19:52:05 RP: hrw|gone : Is anyone of your crew online to answer some basic clutter API questions? Aug 31 19:57:23 woglinde: I built for htctornado Aug 31 19:58:29 somehow there was one zImage in there but I didn't know what I build in order to get it Aug 31 19:58:50 I removed the htctornado image builddir before I rebuilt minimal and console Aug 31 19:59:12 dcordes zImage is normale under work/arch/linux-vbla/arch/boot Aug 31 19:59:16 could it be that only angstrom-x11-image is creating a zImage for that device? Aug 31 20:01:08 can i put custom cflags in a .bb package? Aug 31 20:01:31 mranostay yes Aug 31 20:01:41 search the packages after oemake Aug 31 20:01:50 than you will find examples Aug 31 20:02:04 s/after/for Aug 31 20:02:23 woglinde: can't find zimage there Aug 31 20:02:54 but I'm used to see the zImage for angstrom in my build/tmp/angstrom/deploy/glibc/images/machine folder Aug 31 20:03:00 like with my akita builds Aug 31 20:03:06 dcordes hm Aug 31 20:03:14 maybe some deps are wrong Aug 31 20:03:29 you can try bitbake imagebla -g Aug 31 20:03:42 to let you make the dependcies graph Aug 31 20:03:49 and look at the .dot files Aug 31 20:03:56 if there is a kernel entry Aug 31 20:05:16 hm it is just building bla then Aug 31 20:05:29 is there a way to force creation of zImage? Aug 31 20:05:42 oh sorry got the depends.dot Aug 31 20:06:32 theres soemthing about kernel-image style = bash and such Aug 31 20:06:39 what exactly am I looking for in that file? Aug 31 20:08:26 dcordes kernel-image should be on the left side Aug 31 20:08:49 args Aug 31 20:08:53 I am to dull Aug 31 20:09:00 what is dull? Aug 31 20:09:12 you already mentioned that you have a work/arch/linux-vbla dir Aug 31 20:09:15 dumb Aug 31 20:09:47 is there a way to turn off uboot building when building angstrom? Aug 31 20:09:59 hm so possibly really angstrom-x11-image is only way to make zImage for htctornado then? Aug 31 20:10:15 dcordes okay look at work/arch/linux-vbla/temp/log.do_compile Aug 31 20:13:27 is -Os the only gcc flag to make smaller binaries BTW? Aug 31 20:13:31 ah maybe it doesn't rebuild zImage because it thinks there is already one from previous build? Aug 31 20:13:44 right that could be Aug 31 20:14:06 how can I avoid that (w/o rebuilding _everything_) ? Aug 31 20:14:30 bitbake kernel-foo-bla -c rebuild Aug 31 20:14:47 where kernel-foo-bla is the .bb recpie name for htc Aug 31 20:14:51 kernel-minimal-image? Aug 31 20:14:51 kernels Aug 31 20:15:00 ok Aug 31 20:16:50 /packages/linux/linux-tornado-omap2_2.6.16.16.bb Aug 31 20:17:04 are you sure? Aug 31 20:17:08 not sure Aug 31 20:17:24 whats the name of the work/arch/linux-bla dir? Aug 31 20:17:36 or look at the .dot files Aug 31 20:17:42 there the name stands too Aug 31 20:18:07 I think that is the one because it is only recipe in packages/kernel containing tornado Aug 31 20:18:38 hm dont know whats angstroem is choosing Aug 31 20:18:49 let me see in the conf/machine dirs Aug 31 20:19:14 which htc you have? Aug 31 20:19:20 tornado really? Aug 31 20:19:26 actually I have no htc at all Aug 31 20:19:37 my friend has a (also omap850) based prophet Aug 31 20:19:51 on which we succesfully tried one ancient wizard kernel Aug 31 20:20:01 ah okay Aug 31 20:20:02 PREFERRED_PROVIDER_virtual/kernel = "linux-tornado-omap2" Aug 31 20:20:30 bitbake linux-tornado-omap2 -c rebuild should doo Aug 31 20:20:50 ok Aug 31 20:21:37 woglinde: do you also have a htc? Aug 31 20:21:48 dcordes no Aug 31 20:22:00 I only have a simpad and a greenphone Aug 31 20:22:10 my neo has zecke Aug 31 20:22:21 because he bourought his Aug 31 20:22:28 simpad? isn't this that huge touchscreen thing? Aug 31 20:22:33 yes Aug 31 20:22:35 800x600 Aug 31 20:22:42 sa1110 strongarm Aug 31 20:22:45 lol he gave away his own so you gave him yours? Aug 31 20:22:47 220 mhz ;( Aug 31 20:22:49 heh Aug 31 20:23:03 dcordes: he did. And I gave away not only one but two ;) Aug 31 20:23:05 he gives one greenphone Aug 31 20:23:16 so its okay Aug 31 20:23:23 I only have akita to offer Aug 31 20:23:31 anybody want to swap with gta02? :D Aug 31 20:24:25 * ljp 's simpad digitizer went wonky Aug 31 20:24:25 lol Aug 31 20:24:39 what is digitizer? Aug 31 20:24:45 touchscreen Aug 31 20:24:58 ljp hm Aug 31 20:25:02 * mwester has a dual P-III server with a pair of 9GByte hard drives to trade... Aug 31 20:25:09 lol Aug 31 20:25:19 my current workstation is P-III Aug 31 20:25:41 500mh with 256megs ram Aug 31 20:25:53 jesus Aug 31 20:26:03 thats over 5 years old Aug 31 20:26:03 It probably works great -- except for OE builds, of course. Aug 31 20:26:10 that's 10 years old Aug 31 20:26:18 no Aug 31 20:26:20 close Aug 31 20:26:25 cant be Aug 31 20:26:34 okay it had that year 2000 ready sticker Aug 31 20:26:42 so somewhat late 1999 Aug 31 20:26:42 10 years would be dual pentium pro with 200 mhz Aug 31 20:27:09 never obsolete machine Aug 31 20:27:11 I remeber exactly Aug 31 20:27:12 :D Aug 31 20:27:25 bought at a well known german supermarket Aug 31 20:29:35 anyway I need something that can build a bit faster Aug 31 20:33:40 me is considering hijacking his wife's quad-core Pentium Extreme for OE builds. Aug 31 20:33:49 lol Aug 31 20:33:58 why your wife need a quad core? Aug 31 20:34:06 Photoshop Aug 31 20:34:06 word using only one proc Aug 31 20:34:26 hehe Aug 31 20:34:44 But she doesn't use it at night, so maybe dual-boot isn't a bad idea... Aug 31 20:36:45 mwester: or simply use a live CD Aug 31 20:38:37 hmm... even easier. I'll still need some local disk, but tossing in a spare 120GB drive is easy (it already has 5, nobody will notice a sixth in there) Aug 31 20:42:11 christopher: angstrom-2007.1/alix angstrom-2007.1/fic-gta01, poky/fic-gta01, poky/c7x0 Aug 31 20:42:36 dcordes: many machines has kernel images in rootfs only - not in deploy Aug 31 20:42:51 this kernel build is taking ages Aug 31 20:45:58 mwester: colinux maybe? or even vmware should fly on it Aug 31 20:46:05 * hrw|gone -> sleep Aug 31 20:46:11 nite hrw Aug 31 20:47:03 openmoko-devel-image is insane - I am building it since yesterday and it has ~200 tasks to do.. on 1.6GHz pentium mobile with 1GB ram.. Aug 31 20:47:07 cu Aug 31 20:55:12 hrw|gone: rebuilding made it- thanks Aug 31 20:55:16 got the zImage now Aug 31 20:55:37 hrw|gone: nite Aug 31 20:57:40 mdd? Aug 31 20:57:53 mobile developer days 07 Aug 31 20:57:54 mobile developer something Aug 31 20:57:57 ah Aug 31 20:57:57 denmark Aug 31 21:00:04 he florian Aug 31 21:00:50 it seems I need to recompile monotone for centos5 Aug 31 21:00:57 mickey|mdd: man you always on the road Aug 31 21:01:02 liked your froscon speech btw Aug 31 21:01:12 dcordes: that's right. i travel too much lately. this will change in 2008 Aug 31 21:01:16 dcordes: thans :) Aug 31 21:01:53 you asked me about some more picture of my zaurus running mokoui but my camera was larted Aug 31 21:03:08 re Aug 31 21:08:00 The correct way to add packages to an openembedded image: http://bec-systems.com/web/content/view/79/9/ Aug 31 21:08:06 anyone building OE on osx? It looks like I'm having trouble when bitbake tries to write the tmp/cache sqlite DB Aug 31 21:08:28 dcordes: don't worry, i'm going to do it on my zaurii one of these days Aug 31 21:08:31 trying to undo the damage from my first article, so please tell me if there is anything incorrect this time ... Aug 31 21:08:39 like maybe I don't have the right python sqlite lib or something Aug 31 21:10:10 hughescr try python2.5 as default 2.4 makes trouble Aug 31 21:10:25 woglinde: I'm using python 2.5.1 Aug 31 21:10:28 afaict Aug 31 21:10:31 hm Aug 31 21:10:31 from fink Aug 31 21:10:37 oh macosy Aug 31 21:10:39 oh macosx Aug 31 21:10:39 python --version says 2.5.1 anyway Aug 31 21:10:53 than install sql bindings for it Aug 31 21:11:00 sqlite even Aug 31 21:11:03 woglinde: yeah, how Aug 31 21:11:13 I installed everything from fink which looked like python sqlite bindings Aug 31 21:11:19 didn't help Aug 31 21:11:24 mom Aug 31 21:12:17 hm you have python-sqlite? Aug 31 21:12:22 that shoud be enough Aug 31 21:45:03 zecke: Most of OH appears to be offline now... Aug 31 21:45:47 zecke: If you have specific questions, ask on the mailing list and I suspect someone will get back to you quickly, or just mail mallum ;-) Aug 31 21:46:19 what does the '?=' in the bb files mean? conditional assignment? Aug 31 21:46:30 Cyberdeck: yes Aug 31 21:47:08 Cyberdeck: Set if not already set. See the bitbake manual, its actually documented (for a change!) :) Aug 31 21:47:56 a one html bitbake manual would be great! Hard to do a full text search at the single pages. at least i did not found the '?=' there. Aug 31 21:48:01 but thank you! Aug 31 21:50:37 I have a question on images: Is there some logic as to what should go in conf/distro vs packages/angstrom vs packages/images vs packages/tasks? Aug 31 21:50:47 I guess also vs conf/machine Aug 31 21:50:54 vs build/conf/local.conf Aug 31 21:51:35 I *think* I just want to create a new thing in packages/images, but it's not clear Aug 31 21:52:10 I guess my biggest question is "what is a distro, as opposed to an image" Aug 31 21:53:21 is all that stuff doc'd anywhere? Aug 31 21:53:24 hughescr: An image is a collection of packages for a certain purpose. A distro is more of a kind of policy determination (what versions to use, which ABI etc.) Aug 31 21:54:13 hughescr: packages/angstrom shouldn't exist ideally and probably won't in the end Aug 31 21:54:37 RP: ok, that packages/angstrom should probably just be in packages/images? Aug 31 21:54:44 hughescr: right Aug 31 21:55:03 RP: and conf/distro then would have stuff like "use uclibc rather than glibc, and use version 1.2.3 of package foo, if you use package foo" Aug 31 21:55:20 hughescr: right Aug 31 21:55:31 RP: whereas packages/images would say "when you're building a small system, then use packages foo, bar, and baz" Aug 31 21:55:56 RP: and packages/tasks is basically macros saying "If you want a bootable system, then use the following 19 packages" Aug 31 21:56:02 hughescr: yes Aug 31 21:56:02 ? Aug 31 21:56:26 tasks are just convinient groups of packages to simplify images Aug 31 21:57:22 RP: OK. So for gumstix, if I want to specify that basically it wants to base things off angstrom, but specify uclibc (where angstrom lets you choose), then I'd maybe make a conf/distro/gumstix which sets ANGSTROM_MODE=uclibc then requires angstrom Aug 31 21:57:43 and a pacakges/image/gumstix-small-image which selects packages for a 4MB gumstix, based on angstrom-minimal or something? Aug 31 21:57:52 hughescr: Yes, you could do that Aug 31 21:58:29 RP: my goal is to get as much stuff out of build/conf/local.conf as possible, so users don't get confused and don't shoot themselves in the foot too much Aug 31 21:58:49 angstrom really wants you to make too many decisions in local.conf Aug 31 21:59:18 RP: ok thanks for confirming what I thought was likely true Aug 31 22:03:36 hughescr: I can understand that aim, poky tries to do that too. Angstrom does cater for a lot of configurations and OE is powerful and it seems a shame to hide that power in a way too though :) Aug 31 22:03:40 I can see both sides... Aug 31 22:11:27 hughescr: thanks for asking that question! Aug 31 22:11:45 I was wondering about the same things Aug 31 22:12:20 * rwhitby sees Angstrom eventually becoming the replacement for OE, by subsuming all the options into the Angstrom distro (glibc, uclibc, etc) and then people creating derivate distros of Angstrom to pin down the distro for support and maintenance reasons. Aug 31 22:12:37 just in time :-) Aug 31 22:13:40 but why not keep Angstrom a distro within OE just as it is now? Aug 31 22:14:20 It's a extremely good reference distro, but I do not see it replacing OE. Aug 31 22:15:47 rwhitby: i don't want that Aug 31 22:16:17 so i try not to allow it to happen Aug 31 22:16:23 mickey|mdd: I agree - but that's the direction it's heading - it already has glibc vs uclibc Aug 31 22:16:59 rwhitby eglibc you forget Aug 31 22:17:01 that's right, but if angstrom becomes too bloated, why not adding something with less options again? that was my idea of DISTRO=generic Aug 31 22:17:28 woglinde: indeed - that is more evidence for my point Aug 31 22:17:30 i see angstrom as a reference distro, not as the only means Aug 31 22:18:04 there is no way that OE has enough resources to _properly_ support *3* C libraries Aug 31 22:18:17 rwhitby? Aug 31 22:18:24 until now it works Aug 31 22:19:12 indeed, cause it's fresh. but soon the one or two individuals driving one of the three variants will loose interest, or move onto the next interesting thing, and the support for the least popular of the three will bitrot. Aug 31 22:19:23 that's just my prediction - take it with a grain of salt. Aug 31 22:20:27 if it works for one distro (Angstrom), C lib supports isn't hard to use in your own distro (like mickey said, a more minimal distro) Aug 31 22:20:41 likewise: fully agree Aug 31 22:21:11 that's why I say that Angstrom will become the OE replacement, and there will be lots of distros based on configurations of Angstrom. Aug 31 22:21:48 (and then there will be a heap of legacy non-Angstrom-based distros which will eventually bitrot) Aug 31 22:21:49 rwhitby: but why does Angstrom replace OE, if it is "only" the distro part of it? Aug 31 22:22:06 rwhitby: agreed, we have lots of bitrotting distro's in there. Aug 31 22:22:28 I would rather see a purpose-driven set of distro's Aug 31 22:22:36 likewise: because if something isn't a derivative of Angstrom, then new changes to OE don't get automatically tested on it, and the legacy things break and bitrot. Aug 31 22:22:41 One distro that targets headless systems, for example Aug 31 22:22:43 angstrom isnt an oe replacement, it just might be the base of a lot of the distros buildable out of oe. Aug 31 22:22:58 rwhitby: in the end, it depends on the people that make it. if enough people want A to become it, then it'll be so. If otherwise, it won't happen Aug 31 22:23:03 hi kergie Aug 31 22:23:10 kergoth: how are things? Aug 31 22:23:13 hi kergoth Aug 31 22:23:31 not too bad, getting a bit tired of arizona heat, but not bad Aug 31 22:23:41 glad to hear Aug 31 22:23:54 kergoth: yes, that's what I mean - I'm just taking it to the extreme to highlight the point that the non-Angstrom distros will probably bitrot, so people who want the least pain should get behind angstrom Aug 31 22:24:09 still burned out on coding, but thus far my job at mvista is just bugfixing, which i can do easily and without getting frustrated, so content with that for now.. think i might go back to school in a year or so Aug 31 22:24:13 heh Aug 31 22:24:27 kergoth: i see. yes, i think that may be the best strategy for you Aug 31 22:24:29 back to school? Aug 31 22:24:30 mickey|mdd: generic distro is reasonably slim, lean and non-rotten? Aug 31 22:24:54 i used to enjoy the architecture of code, but i just dont anymore, cant seem to focus on it.. so bugfixing will do Aug 31 22:24:58 woglinde, never got a college degree Aug 31 22:25:00 likewise: definitely rotten, i didn't work on it since years. but we could work again on it Aug 31 22:25:01 mickey|mdd: is there a bugzilla-submitting autobuilder for generic distro? Aug 31 22:25:09 rwhitby: no idea frankly Aug 31 22:25:26 How can one create such a autobuilder? Aug 31 22:25:34 kergoth hehe I tought I were the only lamer who is going a full study at the age of 28 Aug 31 22:25:46 :) Aug 31 22:26:13 woglinde: I finished University only last year, when I was 32, so no hurries :-) Aug 31 22:26:32 likewise hm do you have to feed a family= Aug 31 22:27:01 I enjoyed learning the theory while I was in the middle of a job / practice. Well, only my girlfriend and the cats :-) Aug 31 22:44:11 03mickeyl 07org.oe.dev * r22b3c11c... 10/ (1 packages/vte/vte_0.11.15.bb packages/vte/vte_0.16.8.bb): vte 0.11.15: update to 0.16.8 Aug 31 23:08:42 gn Aug 31 23:16:47 <_diego__> hi Aug 31 23:33:57 * * OE Bug 2894 has been REOPENED by likewise(AT)gmx.net Aug 31 23:33:59 * *  eglibc should provide glibc packages when chosen as default libc Aug 31 23:34:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2894 Sep 01 00:00:00 I think I narrowed down the keylaunch 'fn' problem on spitz to either the X server intercepting the key or the key syms having changed. Sep 01 00:00:18 the problem is, I don't have enough Xlib knowledge to be sure :/ Sep 01 00:02:10 I strongly suspect xorg 7.1, though. Sep 01 00:02:26 I mean 7.2 ;) Sep 01 00:32:14 yai ! Sep 01 00:32:28 found. the problem is in the xmodmap :) Sep 01 00:32:53 BUT fixing it will bring back another bug with mozilla. Sep 01 00:33:57 * * OE Bug 2905 has been created by kristian(AT)mput.de Sep 01 00:33:59 * * python-pygtk does not provide working "import gtk" Sep 01 00:34:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2905 **** ENDING LOGGING AT Sat Sep 01 02:59:56 2007