**** BEGIN LOGGING AT Thu Jan 17 02:59:56 2008 Jan 17 08:29:39 good morning Jan 17 08:43:35 morning all Jan 17 08:43:42 good morning, RP Jan 17 08:43:57 http://rafb.net/p/et5GDl26.html is incorrectly packaged, right? Jan 17 08:44:32 the contents of that package should have gone to libopie2 instead of libopie2-dev or shouldn't they? Jan 17 08:44:43 libopie2 is currently empty Jan 17 08:45:18 Laibsch: No, those symlinks look like they're in the right place Jan 17 08:46:11 Why -dev instead of ${PN}? Jan 17 08:46:21 Again, libopie2 itself is emtpy now Jan 17 08:50:02 Laibsch: You only need those symlinks if you're compiling things against libopie Jan 17 08:50:24 Laibsch: You want the actual library in ${PN}, not a bunch of symlinks Jan 17 08:52:37 OK Jan 17 08:52:48 Problem is there is *nothing* in ${PN} Jan 17 08:52:53 The package does not even exist Jan 17 08:53:10 But that does not seem be the problem I was trying to fix, it seems Jan 17 08:54:12 Which package does the library file end up in then? Jan 17 08:54:17 or is it not packaged? Jan 17 09:04:19 RP: Looks like they all end up in individual packages Jan 17 09:04:30 and it does seem to be the intended way Jan 17 09:45:33 Laibsch: Now you mention it that does some familiar... Jan 17 09:47:19 Is it to be expected that an installed packaged with version 1.2.3+cvs20070727-r11 will not be upgraded to one with 1.2.3-r3 ? Jan 17 09:47:41 the latter seems to be the more recent one Jan 17 09:47:50 Laibsch: I think that's how it is designed yes Jan 17 09:49:01 bluelightning: bummer Jan 17 09:49:28 Because that prevents me from testing your fine SL-CXXX key rotate patch Jan 17 09:50:08 I dislike ipkg... a lot of the time it seems to just launch headlong into breaking your system and provides no way to unbreak it Jan 17 09:50:17 yes Jan 17 09:50:23 it is evil Jan 17 09:51:21 bluelightning: Wouldn't your patch to libopie2 also apply to the cvs version? Jan 17 09:51:47 Laibsch: which one is that? Jan 17 09:52:00 the citytime path one? Jan 17 09:52:05 if so, yes Jan 17 09:54:56 no Jan 17 09:55:41 !oebug 3693 Jan 17 09:55:57 CoreDump|Zzzz: Where is the bot? Jan 17 09:56:16 bluelightning: You get the picture Jan 17 09:56:33 bluelightning: rubbish Jan 17 09:57:06 bugs.openembedded.org down for anyone else? Jan 17 09:57:20 hasn't been working for me since yesterday Jan 17 09:58:12 bluelightning: look at topic Jan 17 09:58:21 use the IP or adjust /etc/hosts Jan 17 09:58:28 heh, missed that, thx Jan 17 09:58:30 the server is moving Jan 17 09:58:48 bluelightning: the correct bug numbes is 3678 Jan 17 10:01:41 Laibsch: well, it will apply to CVS until I commit it to CVS, so it's probably not worth adding it to the cvs version ,bb Jan 17 10:01:48 s/,/. Jan 17 10:02:11 can you commit it to cvs soonish? Jan 17 10:02:26 Like, right after I told you it fixes the problem here Jan 17 10:08:03 bonjour Jan 17 10:41:12 zecke: there? Jan 17 10:41:35 zecke: i found problems between qt4 and qmake2 recipes. Jan 17 10:50:14 morning Jan 17 10:51:26 morning Jan 17 10:52:12 zecke: to me, looks like qmake2 ought to be moved to qt4 dir so it could reuse all variables and like, making it simple to maintain Jan 17 10:55:54 otavio: besides possible sharing, what problems? Jan 17 10:58:03 zecke: QT_INSTALL_DIR and HEADERS point to the wrong place Jan 17 10:58:17 zecke: INSTALL_LIBS, sorry Jan 17 10:58:50 otavio: which recipe, when does it point to the wrong place? Jan 17 10:58:51 zecke: .pc files put libdir to lib but you set lib/qt4 Jan 17 10:59:05 otavio: do you use QT_INSTALL_DIR to install files? Jan 17 10:59:23 zecke: it puts staging/qt4/lib instead of stagin/lib/qt4 Jan 17 10:59:52 zecke: cmake uses qmake to query paths Jan 17 11:00:24 otavio: ah right, now I get the picture :) Jan 17 11:00:48 zecke: I could get out of it, making cmake use pkgconfig ... however it is a regression agaist previous recipes Jan 17 11:01:16 zecke: it took a day to me to get it Jan 17 11:01:32 zecke: you're much faster >:P Jan 17 11:09:49 zecke: that's why it looks better to share the settings. So it doesn't happen anymore Jan 17 11:34:53 hi zecke Jan 17 11:44:30 hi everybody Jan 17 11:44:48 hi thesing Jan 17 11:46:21 yo thesing Jan 17 11:46:34 thesing: any news about your mainboard? Jan 17 11:48:56 hrw: they said that it could take up to six weeks. So I should get it back at the end of January. Jan 17 11:50:24 thesing: good to know to not shop for computer stuff in germany Jan 17 11:50:51 warranty repair should not take more then 2-3 weeks Jan 17 11:52:32 hrw: I agree. Maybe its just because my board is not been produced since nearly two years ago. Jan 17 11:55:09 thesing: but it was on warranty still? Jan 17 11:56:11 hrw: yes. We have two years warranty here and my board was 22 months old ;) Jan 17 11:57:20 hrw: I already got the money from my original board back. Its now a board from ebay (that also still had warranty) thats in repair now. Jan 17 11:59:36 totally bad luck Jan 17 12:01:47 re Jan 17 12:03:00 thesing: Can you fix the rc6 patch issue? Jan 17 12:03:08 hrw: computers just don't like me ;9 Jan 17 12:03:19 Laibsch: already done Jan 17 12:03:26 * XorA waits for thesing to brick all the worlds collie Jan 17 12:04:08 XorA: unless I start some bootloader work that won't happen ;) Jan 17 12:05:06 thesing: Great Jan 17 12:05:49 thesing: did they change at all from rc$previous? Jan 17 12:06:33 I was amazed to see that bitbake did not care to download them again but just failed do_patch on an older sharp_sl*2*$something.patch Jan 17 12:06:35 btw a collie kernel with just jffs2 and flash ist ca. 600 KB. One could easily create a klibc based initramfs. Jan 17 12:07:00 Laibsch: I didn't build a new kernel yet. Jan 17 12:07:16 But I will try. Jan 17 12:07:16 OK, let me try, then Jan 17 12:07:39 But, bitbake should reload those patches because the URL changed, right? Jan 17 12:07:44 I think it did not here Jan 17 12:07:57 thesing: No changes to any of the patches, yet, right? Jan 17 12:08:03 I don't know. Jan 17 12:08:07 Laibsch: right. Jan 17 12:08:07 Just renaming the directory on the server? Jan 17 12:08:20 actually a soft link. Jan 17 12:08:24 sure Jan 17 12:08:42 Maybe we should consider adding md5 in the bb file for those patches Jan 17 12:08:43 But I probably should just put them all in one directory. Jan 17 12:08:55 * thesing -> lunch Jan 17 12:10:26 They change frequently and thus IMHO it would unnecessarily bloat checksums.ini Jan 17 12:12:42 XorA: heh, how many collies are there left in the world anyway? Jan 17 12:13:31 pb_: a surprising number Jan 17 12:13:56 what collies stands for? Jan 17 12:14:00 * otavio curious! Jan 17 12:14:06 sl5500 zaurus Jan 17 12:14:09 ahh Jan 17 12:16:12 what's that? Jan 17 12:16:14 oops wrong window Jan 17 12:16:24 ~collie Jan 17 12:16:25 [collie] sharp zaurus SL-5000 (often called SL-5000D because it was developer version) or SL-5500 (also SL-5500G with QWERTZ keyboard) or a dog Jan 17 12:17:30 thesing|lunch: put in one dir and bump each time when they change Jan 17 12:17:58 RP: ping Jan 17 12:18:45 XorA: "psplash_svn.bb : rediff openmoko patch" commit looks to have broken psplash. Could you take a look? Jan 17 12:18:51 XorA: or a patch is missing Jan 17 12:19:28 tmp/deploy/glibc/ipk/armv4t/psplash_0.0+svnr249-r12_armv4t.ipk Jan 17 12:19:31 builds here Jan 17 12:19:53 XorA: please ... see if you haven't forgot to commit another patch Jan 17 12:20:29 no, and that commit was months ago so I suspect someone would have seen the fauilt, please at least give me a log to work with Jan 17 12:22:10 File "patch_do_patch", line 98, in patch_do_patch Jan 17 12:22:11 File "", line 339, in Resolve Jan 17 12:22:11 __builtin__.CmdError: Command Error: exit status: 1 Output: Jan 17 12:22:11 Applying patch configurability.patch Jan 17 12:22:11 patching file psplash.c Jan 17 12:22:13 Hunk #11 FAILED at 253. Jan 17 12:22:43 please pastbin the full log, thats just the last error Jan 17 12:22:52 RP: the sysroot change that you did for gcc 4.2.2 - I wonder how it works in poky... for me, in OE it tried to mkdir a directory in / Jan 17 12:24:13 viewmtn is currently down? Jan 17 12:24:36 http://paste.debian.net/47090 Jan 17 12:26:29 otavio: hmm, my only thought is for some reason the angstrom patch didn't apply Jan 17 12:26:46 the openmoko patch is on top of the angstrom one Jan 17 12:26:49 Is there some easy way to test if a shell script is indeed free from bashims? Something like an HTML validator? Jan 17 12:27:17 otavio@neumann:~/hacking/ossystems/osthinclient/oe$ ls packages/psplash/files/angstrom/ -l Jan 17 12:27:20 total 112 Jan 17 12:27:23 -rw-r--r-- 1 otavio otavio 110849 2008-01-09 15:36 psplash-hand-img.h Jan 17 12:28:44 or someone has rewound time in svn Jan 17 12:29:07 XorA: are you using dev branch? Jan 17 12:29:13 yes Jan 17 12:29:23 otavio: show us 'mtn head' output Jan 17 12:29:43 or yeah, someone forgot to do mtn update for 4 months Jan 17 12:30:51 SRC_URI_append_angstrom = " file://logo-math.patch;patch=1 " Jan 17 12:30:52 SRC_URI_append_openmoko = " file://configurability.patch;patch=1 " Jan 17 12:31:07 otavio: show us 'mtn head' output - then we will talk more Jan 17 12:31:14 logo ought to be on openmoko too, no? Jan 17 12:31:27 otavio: show us 'mtn head' output - then we will talk more Jan 17 12:32:03 hrw: i'm using git converted repository but mine head is> Jan 17 12:32:04 Author: Jan 17 12:32:04 Date: Wed Jan 16 15:11:59 2008 +0000 Jan 17 12:32:04 add pwsafe to .dev closes 1065 Jan 17 12:32:11 otavio: pspash recipe is fully working for everyone here, thus the error is at your end Jan 17 12:32:11 thx Jan 17 12:32:35 otavio: we had too many people using old repository Jan 17 12:32:39 * XorA guese the git conversion doesnt work yet Jan 17 12:35:18 XorA: have you used angstorm or openmoko disrtro? Jan 17 12:35:27 otavio: both Jan 17 12:35:35 !seen psokolovsky Jan 17 12:35:36 psokolovsky (n=psokolov@nat-users.is.ua) was last seen quitting from #oe 4 days, 20 hours, 14 minutes ago stating (Read error: 110 (Connection timed out)). Jan 17 12:36:29 I would guess git has the new psplash_svn.bb file, but the patch change didnt migrate properly, compare your file with the ones in OE Jan 17 12:36:54 XorA: could you pastebin it? Jan 17 12:37:03 XorA: lastest bb, please. Jan 17 12:38:15 http://www.openembedded.org/filebrowser/org.openembedded.dev/packages/psplash Jan 17 12:38:30 check the configuribility patch verses the one you have Jan 17 12:40:53 XorA: same. I got it here but fail to see how it could work for you Jan 17 12:41:08 XorA: I need logo-math first ... Jan 17 12:41:16 XorA: then configurability Jan 17 12:41:31 XorA: however, openmoko, there, only applies configurability Jan 17 12:41:39 XorA: how it works in your side? Jan 17 12:41:45 openmoko applies both Jan 17 12:41:51 ~oemirror Jan 17 12:42:02 XorA: and angstorm doesn't apply configurability Jan 17 12:42:20 NOTE: package psplash-0.0+svnr249-r12: task do_patch: started Jan 17 12:42:20 NOTE: Applying patch 'logo-math.patch' Jan 17 12:42:20 NOTE: Applying patch 'configurability.patch' Jan 17 12:42:20 NOTE: package psplash-0.0+svnr249-r12: task do_patch: completed Jan 17 12:42:24 XorA: have any pointer to that conf? I don't see where Jan 17 12:42:44 DISTRO="openmoko" Jan 17 12:42:59 SRC_URI_append_openmoko = " file://configurability.patch;patch=1 " Jan 17 12:43:00 only one Jan 17 12:43:26 I see that's my fault but I fail to see how it works Jan 17 12:43:45 git migration was fine. It worked... same files.. Jan 17 12:44:33 are you building openmoko? Jan 17 12:45:56 XorA: no. I'm using a distro that uses angstorm as basis, as openmoko Jan 17 12:46:14 XorA: but in my case, it doesn't apply the angstorm things, too. Jan 17 12:46:20 otavio: well dont have openmoko as an override then Jan 17 12:46:24 Laibsch: bugzilla is down, so from where should he get the data Jan 17 12:46:26 XorA: that's why it failes Jan 17 12:46:48 remove all traces of openmoko from your distro Jan 17 12:47:01 XorA: sorry, I wasn't clear Jan 17 12:47:21 XorA: I've written a new distro conf that includes angstorm. This works. Jan 17 12:47:47 XorA: looks like openmoko says to bitbake to apply also things specific to angstorm ... Jan 17 12:48:03 openmoko is an angstrom derivative Jan 17 12:48:04 XorA: and I want to know how openmoko does Jan 17 12:48:09 thesing: oesources.org seems to have been the same server as bugzilla. it is unreachable at the moment Jan 17 12:48:09 XorA: mine too Jan 17 12:48:18 otavio: it add angstrom to its overrides Jan 17 12:48:41 Laibsch: too bad. Jan 17 12:48:44 OVERRIDES = "local:${MACHINE}:${DISTRO}:angstrom:${TARGET_OS}:${TARGET_ARCH}:build-${BUILD_OS}:fail-fast:pn-${PN}" Jan 17 12:48:54 ahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh Jan 17 12:48:58 and includes some the the preferred version stuff Jan 17 12:49:01 living and learning Jan 17 12:49:10 ~smartquestions Jan 17 12:49:11 i guess smartquestions is http://www.catb.org/~esr/faqs/smart-questions.html Jan 17 12:49:29 anyway high priotiry task of feed XorA is starting, back in a bit Jan 17 13:40:58 thesing: 2 out of 6 hunks FAILED -- rejects in file arch/arm/common/sharpsl_pm.c Jan 17 13:41:16 from sharpsl_pm-2.patch Jan 17 13:41:41 I will look into this. Jan 17 13:41:46 I'd support hrw's suggestion of versioning your patches Jan 17 13:42:14 Otherwise people with the faulty patch already in $SRCDIR are screwed Jan 17 13:42:17 thesing: thanks Jan 17 13:42:43 thesing: In fact, why not have your patches right in OE? Jan 17 13:43:46 Laibsch: because most other linux-rp patches aren't in OE. I don't know if there is some policy about this. Jan 17 13:46:37 not that I am aware of Jan 17 13:48:58 hrw: you have locked down http://www.openembedded.org/wiki/CommitPolicy? Jan 17 13:59:53 re Jan 17 14:00:00 Laibsch: not that I remember Jan 17 14:01:37 I can edit other pages, but not that one Jan 17 14:03:12 let me loging Jan 17 14:05:46 ok guys... this external toolchain stuff again, really drives me crazy... DEBUG: providers for virtual/libc are: ['external-toolchain', 'eglibc', 'glibc'], I do have: PREFERRED_PROVIDER_virtual/libc = "glibc" in my distro conf, log says: DEBUG: selecting /org.openembedded.dev/packages/glibc/glibc_2.6.1.bb as PREFERRED_VERSION 2.6.1 of package glibc (for item virtual/libc) Jan 17 14:05:48 but! Jan 17 14:06:09 couple of lines later: DEBUG: adding /org.openembedded.dev/packages/meta/external-toolchain.bb to satisfy virtual/libc Jan 17 14:06:17 why?! :( Jan 17 14:06:53 there must be a way to get rid of the damn external-toolchain.bb package without having to delete it Jan 17 14:07:02 please, anyone? Jan 17 14:07:02 :> Jan 17 14:08:33 heh, seems a more meaningful description can be found here: http://76.74.156.109/show_bug.cgi?id=3509 Jan 17 14:10:47 dev: I am still not convinced in that Jan 17 14:10:50 oops wrong window Jan 17 14:20:50 * hrw wants wordpress and more normal wiki... Jan 17 14:21:37 Laibsch: I do not see any option for that wiki page to be locked Jan 17 14:24:29 Can you edit the page? Jan 17 14:24:51 your wish is going to incur the wrath of TheGod[tm] ;-) Jan 17 14:34:42 Jin^eLD: Have you deleted your tmp directory before you tried to rebuild after setting the PREFERRED_PROVIDER? Jan 17 14:34:52 Jin^eLD: Add it to bbmask Jan 17 14:34:55 mwester: no.. Jan 17 14:35:11 florian: well, that's what I am going to do but it's really stupid Jan 17 14:36:02 and I don't get it why DEBUG says that its selecting my preferred provider, but later on still adds the external toolchain for the same dependency! Jan 17 14:36:13 Jin^eLD: The problem is that it sets the "PREFERRED" ----- not the "REQUIRED"! So if it sees that there are present packages in your existing build tree, I've observed that it will use the existing dependencies instead of the preferred ones. Jan 17 14:36:57 I see.. Jan 17 14:40:50 You can try just blasting away your cache, but I'm not convinced that will work. Perhap? Jan 17 14:41:02 er, "perhaps" even. Jan 17 14:43:05 hrw: Can you edit the page? Jan 17 14:43:56 mwester: I will just try a clean build... Jan 17 14:47:55 Laibsch: edited Jan 17 14:47:59 * hrw -> afk Jan 17 15:22:00 sakoman: please fix wifistix-modules... Jan 17 15:22:09 sakoman: it lack sf_buff.patch Jan 17 15:25:19 hrw, when I get a chance I will copy from his collection Jan 17 15:25:48 I really should be more agressive moving in fixes that are not intrusive to .dev ... Jan 17 15:25:55 but I am swamped again ... Jan 17 15:26:46 strange bitbake doesn't find cacao http://pastebin.com/m579e9b02 Jan 17 15:26:50 Crofton: then give me url to that collection Jan 17 15:27:49 hrw, standby Jan 17 15:28:32 http://www.gumstix.net/Documentation/view/Getting-started/Setting-up-a-build-environment/29.html Jan 17 15:28:43 He copies the mtn repo into svn .... Jan 17 15:28:52 and adds some collections Jan 17 15:30:05 thx Jan 17 15:31:49 GNUtoo: thats really strange. Its working for me: http://scap.linuxtogo.org/ Jan 17 15:33:12 NOTE: package wifistix-modules-5.0.16.p0-r0: task do_build: completed Jan 17 15:33:16 thx Crofton Jan 17 15:33:54 what app is this http://scap.linuxtogo.org/files/6bb82ec6aae8b57a828037062eb5b782.png ? Jan 17 15:34:03 hrw, btw it looks like he is doing quite well with the gumstix crowd and OE Jan 17 15:34:42 guys, any clue why this is happening? Jan 17 15:34:44 http://pastebin.mozilla.org/306116 Jan 17 15:34:56 Crofton: I am pushing missing patch now Jan 17 15:35:03 thanks Jan 17 15:35:25 rschuster: its 'settingsgui' - python based Jan 17 15:35:30 if I copy *curl* from ../arm-angstrom-linux-gnueabi to all-angstrom-linux-gnueabi then it runs through Jan 17 15:35:31 I must have forgot the mtn add it when I copied it over last time :( Jan 17 15:35:45 question is - why is it searching for those packages in the wrong place? Jan 17 15:35:51 or rather for the package data Jan 17 15:36:04 op P4VuL0N Jan 17 15:36:26 NineX: you are not from Łódź... Jan 17 15:37:13 nice wierd characters Jan 17 15:37:57 hrw: nope Jan 17 15:38:44 Hi guys. How can I tell oe to include a kernel module in the rootfs? Jan 17 15:38:49 hrw: but pavulon rulez ;) Jan 17 15:39:03 * hrw -> afk Jan 17 15:54:44 alphaone: just make the image depend on it Jan 17 15:55:39 thesing: I tried MACHINE_EXTRA_RRECOMMENDS, but that doesn't seem to work... Jan 17 15:56:30 alphaone: it's also the wrong way. what image are you building? Jan 17 16:04:17 mwester: I don't remember your commit Jan 17 16:04:27 Did it turn it a problem? Jan 17 16:15:58 hello everybody! Jan 17 16:16:22 I badly need help for a uclibc compiling Jan 17 16:16:26 problem Jan 17 16:17:15 I'm a newby with openembedded, i moved from buildroot recently Jan 17 16:17:47 my problem is that I can't succeed in compiling uclibc on x86 arch Jan 17 16:18:08 the error I get is: Jan 17 16:18:54 openembedded/build/tmp/cross/lib/gcc/i486-linux-uclibc/4.1.1/../../../../i486-linux-uclibc/bin/ld: crt1.o: No such file: No such file or directory Jan 17 16:20:58 seems that the "populate_staging" task is not executed Jan 17 16:21:34 i tried "bitbake uclibc -c populate_staging" but no success Jan 17 16:22:24 xlele: did you try to "find" the file? Jan 17 16:22:39 yes Jan 17 16:22:50 the file is present, but not in the right path Jan 17 16:23:13 xlele: please pastebin the error log Jan 17 16:23:42 of course! Jan 17 16:25:55 thesing: just launched the bitbake Jan 17 16:34:44 thesing: sorry, but mac os x terminal garbled all the text... I had to use another terminal Jan 17 16:34:55 thesing: here it is: http://pastebin.com/m44724ec2 Jan 17 16:37:56 xlele: try the stable branch with angstrom-2007.1 as distro. the generic distros are broken. you could also try .dev and angstrom-2008.1 Jan 17 16:38:54 thesing: sorry to ask for it, but Jan 17 16:40:00 thesing: in order to change to the stable branch, I just need to change the "mtn --db=/openembedded/OE.mtn pull monotone.openembedded.org org.openembedded.dev" coomand I issue in Jan 17 16:40:15 thesing: mtn --db=/openembedded/OE.mtn pull monotone.openembedded.org org.openembedded.stable Jan 17 16:40:17 ? Jan 17 16:40:58 xlele: mtn --db=/path/to/OE.mtn checkout --branch=org.openembedded.angstrom-2007.12-stable Jan 17 16:41:21 RP: hmm... is it possible that we lost the Scratchbox rootstrap images moving from meta-sdk to meta-toolchain? Jan 17 16:41:45 thesing: great! thank you so much! I'll give it a try Jan 17 16:42:06 xlele: see also www.angstrom-distribution.org/building-angstrom for necessary local.conf (esp. ANGSTROM_MODE ) Jan 17 16:42:20 hi florian Jan 17 16:42:23 thesing: openembedded is very, very different from buildroot Jan 17 16:42:29 hi pb_ Jan 17 16:42:49 * florian feels sleepy today... Jan 17 16:43:03 thesing: but I think i'll get good results once I'll get more confident with it Jan 17 16:43:41 xlele: its just make vs. python. In buildroot you know where the config files are. Here you still have to learn. Jan 17 16:45:13 thesing: yes, that's the problem! Jan 17 16:46:01 thesing: but does angstrom support x86 arch built on uclibc? Jan 17 16:48:06 I think so. But never tried x86. I know other people do x86 builds but not I they use uclibc. Jan 17 16:48:19 XorA: Did you manage to do anythign good for gsmd yesterday? Jan 17 16:48:30 florian: not yet, still on my todo Jan 17 16:48:37 xlele: What is your target device? Jan 17 16:48:44 XorA: ok :) Jan 17 16:48:57 thesing: soekris 4801 Jan 17 16:49:02 florian: curse this actually having to work business :-) Jan 17 16:49:20 XorA, agreed Jan 17 16:49:22 XorA: same here... Jan 17 16:49:31 thesing: dectop, and norhtec Jan 17 16:50:39 thesing: basically, under buildroot I successfully run all those platform using just 3.5mb Jan 17 16:51:23 thesing: the same flash can be used on all of them (and also on other x86 embedded platform I tried) Jan 17 16:52:51 thesing: but i decided to move to openembedded recently while trying to make libhal and libdbus working on buildroot.. really painful! Jan 17 16:54:06 xlele: I think the oe-build images will be bigger. Because some packages aren't packaged very fine-granulated Jan 17 16:54:26 xlele: But you can work on that once they build ;) Jan 17 16:55:43 thesing: yes, having a working cross-compile environment should be helpful :) Jan 17 16:56:37 thesing: what I really like to do would be to create a distro for x86 devices Jan 17 16:57:30 thesing: they are really similar and once you have a working environment you can run it on a great number of devices without modifications Jan 17 16:57:38 fixing the generic distro would be good strategy Jan 17 16:58:07 xlele: most probably you want to create an image and not a distro. Jan 17 16:58:17 yes Jan 17 16:58:48 mickey: yes, but I don't have a good knowledge of oe architecture to do that Jan 17 16:58:52 mickey|zzZZzz: maybe once he knows oe better. Jan 17 16:58:59 indeed Jan 17 16:59:08 right Jan 17 16:59:54 xlele: there are a lot of low cost devices out there based on x86 arch Jan 17 17:00:40 for example the nortech costs around 100$ and it can be considered a real pc Jan 17 17:00:54 xlele: you should realy start with angstrom create an new image if you need it and than think if you realy need an own distro Jan 17 17:00:57 a la one-desktop-per-child Jan 17 17:02:19 thesing: yes, it's what I'm going to do... I'll try to build a minimal image Jan 17 17:02:44 thesing: then I'll try to expand it until all the feature I need will be there Jan 17 17:05:04 thesing: do you think angstrom is suitable for my needs? Jan 17 17:07:16 xlele: what a distro in oe realy does is to decide what versions of packages to use and some tweaks to some packages. (like configure options) Jan 17 17:08:02 xlele: images decide what ends up in the image. Jan 17 17:09:20 xlele: just in case you didn't know there is an user manual for oe available: http://www.openembedded.org/user-manual Jan 17 17:14:05 thesing: :D Jan 17 17:14:38 thesing: yes, I went into the manual Jan 17 17:15:17 thesing: but sometimes, for a buildroot-dummy, it can be disorienting! Jan 17 17:16:18 feel free to ask. I'm no oe-expert but I will answer your question as good as I can. Jan 17 17:16:39 thesing: Trying to build console-image for the chumby, but it's missing kernel modules Jan 17 17:17:00 Sorry. was afk for a bit Jan 17 17:17:21 thesing: thank you again! I'll let you know if I get good result! Jan 17 17:18:01 thesing: keep in mind that your help saved me hours of headache! :D Jan 17 17:18:33 thesing: thank you again and see you soon. bye! Jan 17 17:18:44 xlele: bye Jan 17 17:20:31 alphaone: MACHINE_EXTRA_RDEPENDS should do then. Jan 17 17:22:11 Thanks Jan 17 17:22:32 does OE or bitbake have any kind of code that could be (re)used for creating an RDEPENDS-complete subset of ipk files from $DEPLOYDIR? Jan 17 17:27:18 Laibsch: why would you want to do this? Jan 17 17:27:57 thesing: a I see what you mean. Jan 17 17:28:45 Laibsch: I would say it has because this is it what image generation does. Jan 17 17:35:30 image generation relies on ipkg to do that. Jan 17 17:51:40 RP: quick quesion about the module_autoload command in linux-rp.inc. Would it be bad if a driver was listed but actually built in instead? Jan 17 17:52:20 hvontres|work: then there wont be an ipk to do a postinst in anyway so should be fine Jan 17 17:53:31 XorA: thanks. It looks like somehow the rtc-sa1100 driver got turned into a module for poodle but wasn't being autoloaded....kinda freezes time on suspend..:) Jan 17 18:09:35 pb, thesing: thanks. You happen to know where that code would be? image.bbclass? Jan 17 18:09:43 if there is such a thing Jan 17 18:15:43 Hello Jan 17 18:23:40 yo sirfred Jan 17 18:37:16 Laibsch: image-ipk.bbclass, I guess Jan 17 18:37:23 or rather rootfs, not image Jan 17 18:38:31 pb_Thanks, I will take a look Jan 17 18:52:04 * * OE Bug 1065 has been RESOLVED (FIXED) by Jan 17 18:52:06 * *  pwsafe - password keyring utility Jan 17 18:52:08 * * http://bugs.openembedded.org/show_bug.cgi?id=1065 Jan 17 18:52:16 * * OE Bug 3705 has been created by  Jan 17 18:52:18 * * ipkgorphan does not work as expected Jan 17 18:52:20 * * http://bugs.openembedded.org/show_bug.cgi?id=3705 Jan 17 18:52:28 * * OE Bug has been RESOLVED by xcp(AT)mixed-mode.de Jan 17 18:52:30 * *  Package xfonts-xorg 6.8 fails to compile Jan 17 18:52:32 * * http://bugs.openembedded.org/show_bug.cgi?id= Jan 17 18:52:40 * * OE Bug 3664 has been RESOLVED (FIXED) by Jan 17 18:52:42 * *  WLAN cards and suspend sometimes don't mix Jan 17 18:52:44 * * http://bugs.openembedded.org/show_bug.cgi?id=3664 Jan 17 18:59:22 GNUtoo-laptop: "bitbake cacao" seems to work fine for me Jan 17 19:02:32 Laibsch: I fixed the patches so far, that it compiles. But the kernel is to big. Jan 17 19:03:20 big as is outrageously big? Jan 17 19:03:34 or just a little to big Jan 17 19:04:27 Crofton: about 10K Jan 17 19:04:31 thesing: Can you rename them or put them into OE? Jan 17 19:05:03 Laibsch: I will disable netfilter and make a diff then. Jan 17 19:05:55 not the 4M kernel probelm then :) Jan 17 19:11:06 bbl Jan 17 19:58:46 Laibsch: http://www.pastebin.ca/858910 is the diff. Kernel is small enough now. Jan 17 20:18:58 what do i need to do to get something like module_autoload_chumby_accel = "chumby_accel" to work? Jan 17 20:19:14 it just produces files in /etc/modutils/ Jan 17 20:19:21 but the modules are not loaded Jan 17 20:24:47 Shoragan: try running update-modules on device Jan 17 20:24:55 it's an cramfs Jan 17 20:25:34 was it run on image generation? (see log files) Jan 17 20:27:59 Configuring update-modules Jan 17 20:28:06 (offline root mode: not running update-modules.postinst) Jan 17 20:33:25 Shoragan: maybe its run on first boot normally. Jan 17 20:34:44 yes, it would run, but it cant write on the rootfs Jan 17 20:35:42 I don't know if anybody besides you builds ro filesystems. Jan 17 20:36:00 hmm, ok Jan 17 20:37:46 Shoragan: you could open a bug report. Jan 17 20:53:07 If I have a package that depends on multiple svn links how would I go about downloading from two different svn repositories in OE inside of 1 bb file? I tried adding multiple svn: entries to the SRC_URI but that did not seem to work. Jan 17 21:06:29 Should the kdrive xserver use the kernel keymap ? Jan 17 21:07:08 It's weird. I've defined a compose key in the kernel keymap, it worked, but the compose combinations I'm defining aren't Jan 17 21:55:02 re Jan 17 21:58:13 flo_lap: We shouldn't have lost any of the scratchbox functionality with the meta-sdk changes although I don't use the scratchbox stuff... Jan 17 21:58:26 hey guys Jan 17 21:58:35 hi Jin^eLD Jan 17 21:58:41 hi rp and florian Jan 17 21:58:44 RP: why did you change the SYSROOT variable in gcc 4.2.2cross sdk? Jan 17 21:58:46 hi jin^eld Jan 17 21:58:49 hey woglinde Jan 17 21:59:06 Jin^eLD: you asked about the earlier in relation to poky? Jan 17 21:59:19 RP: yes, but you were away Jan 17 21:59:24 that was quite some hours ago Jan 17 21:59:32 I read your mtn log and the comment was that you synced that with poky Jan 17 21:59:38 so I wonder how or how it is working there Jan 17 21:59:54 I think i did send a mail to the OEML regarding this Jan 17 21:59:55 poky was altered a bit to switch to using sysroots Jan 17 22:00:03 RP: well, in OE it breaks... Jan 17 22:00:12 it ries to mkdir in / which of course does not work out Jan 17 22:00:13 Its possible something has been synced to OE without accounting for that Jan 17 22:01:10 this meta-toolchain thing is really an endless pain Jan 17 22:01:11 :) Jan 17 22:01:22 lots of issues seem to come together there Jan 17 22:01:41 Jin^eLD: meta-toolchain is at the cutting edge... Jan 17 22:01:58 RP: that may be, but it's often very much needed... Jan 17 22:02:25 I was looking at the external-toolchain.bb today, tried to define PREFERRED_PROVIDERs for my distro to make sure that external-toolchain is not being selected Jan 17 22:02:28 no chance Jan 17 22:02:48 Which distro? Jan 17 22:03:00 in DEBUG I see, for example, that glibc is picked as provider for virtual/libc - but then later on external-toolchain is still being added Jan 17 22:03:08 RP: angstrom 2008 Jan 17 22:03:30 Ok, one problem at a time Jan 17 22:03:47 I have a 3rd problem in my pocket ready ;) Jan 17 22:03:52 but indeed... Jan 17 22:03:54 gcc4-build-sdk.inc does play with sysroot but its not related to the poky sysroot changes Jan 17 22:04:38 RP: http://lists.linuxtogo.org/pipermail/openembedded-devel/2008-January/004046.html Jan 17 22:04:43 It does indeed do a mkdir ${SYSROOT}/usr but it also sets SYSROOT = "${WORKDIR}/sysroot" Jan 17 22:05:05 RP: no, that's the thing - this is commented out Jan 17 22:05:14 and instead it uses SYSROOT = ${prefix}/${TARGET_SYS} Jan 17 22:05:20 well, at least it did so yesterday evening when I tried Jan 17 22:06:28 It is not commented out here Jan 17 22:06:59 The trouble here is ${SYSROOT} is being used for two different things... Jan 17 22:07:05 let me do a pull/up again Jan 17 22:07:10 maybe someone changed it after my email Jan 17 22:08:29 Jin^eLD: where is this other definition of SYSROOT? Jan 17 22:09:52 RP: yesterday evening it was in gcc4-build-sdk.inc Jan 17 22:10:01 I am doing a pull now to see if its still there Jan 17 22:10:51 * RP too Jan 17 22:14:29 aha, gcc 4.2.2 packages were indeed modified Jan 17 22:15:12 yeah, its changed back now Jan 17 22:15:38 who by and why though? :/ Jan 17 22:15:53 hmm Jan 17 22:16:04 ah, zecke created it I think Jan 17 22:17:22 meta-toolchain creates toolchain so people can build SW outside of OE? Jan 17 22:17:31 Crofton: yes Jan 17 22:17:36 +10 from me Jan 17 22:17:57 ::) Jan 17 22:19:12 Crofton: problem is - it never works out of the box ;) Jan 17 22:20:26 * Crofton hopes enough people care so it gets fixed Jan 17 22:20:42 well, meta-toolchain in poky is used heavily and works Jan 17 22:20:59 Crofton: I was trying to use it and kept trying to figure it all out but so far it was way over my head Jan 17 22:21:04 apart from me breaking it today but that was only to be execpted given what I did... Jan 17 22:22:26 damn, have to leave now Jan 17 22:22:35 missing a good moment to pester you guys with that problem :) Jan 17 22:23:07 mtn log say I was the last person to touch the file and that was 4 months ago Jan 17 22:23:34 I must be going nuts or something heh Jan 17 22:23:55 but I did a fresh pull yesterday Jan 17 22:24:05 well Jan 17 22:24:10 gcc 4.2.2 turned out to be buggy anyway Jan 17 22:24:14 check you don't have anything committed locally... Jan 17 22:24:14 so I went back to 4.1 Jan 17 22:24:25 What was broken out of interest? Jan 17 22:24:35 you mean other things? Jan 17 22:24:37 oh Jan 17 22:24:39 in 4.2.2? Jan 17 22:24:43 yes Jan 17 22:24:59 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31899 Jan 17 22:25:33 gotta run, RP thanks Jan 17 22:25:53 and beware, I will pester you tomorrow regarding other meta-toolchain stuff :) Jan 17 22:26:18 * RP makes a note to hide :) Jan 17 22:26:21 does anybody know how to copy a /dev- file? Jan 17 22:27:58 thesing: use mknod? Jan 17 22:28:50 NAbyss: There is a way to copy them around. But I always forget how. Jan 17 22:29:02 thesing: cp won't do it? Jan 17 22:30:46 NAbyss: well it does... But why do I remember something special about that? Jan 17 22:37:08 thecan_: you can always create new device entries with mknod too Jan 17 22:38:25 RP: It looks like kernel.bbclass defines module_autoload_sa100-rtc but in 2.6.23 that has been renamed to rtc-sa1100. Should we add that to kernel.bbclass? Also, right now only defconfig-poodle has that as a module in linux-rp_2.6.213 Jan 17 22:38:37 err linux-rp-2.6.23 Jan 17 22:38:53 hvontres|work: We should probably add the new name to kernel.bbclass Jan 17 22:39:20 RP: OK. I'll file a patch when I get home... unless you want to beat me to it :) Jan 17 22:39:49 RP: should we make it a module for the other linux-rp divices as well? Jan 17 22:40:07 hvontres|work: No, its only a module for poodle due to space contraints Jan 17 22:41:34 RP: well, I think the other Zaurus Devices would have similar issues Jan 17 22:43:08 hvontres|work: I'm working on the space issues, I have some ideas Jan 17 22:44:31 RP: Np. I just stumbled over this yesterday while trying to figure out why time stood still during suspend Jan 17 22:55:10 RP: what ideas regarding space do you have? I'm trying to create an initramfs that will kexec a kernel from flash. Jan 17 22:55:46 thesing: Note ones that good, I'm just trying to shrink the networking stack in the standard kernels a bit Jan 17 22:58:39 RP: with the initramfs we could put the kernel in rootfs and make upgrade easier, use similar config for kernels etc. Jan 17 23:00:38 thesing: I know what you mean but I just want to get the kernel size down atm **** ENDING LOGGING AT Fri Jan 18 02:59:57 2008