**** BEGIN LOGGING AT Thu Jan 20 02:59:57 2011 Jan 20 03:02:40 * kergoth` happens to think that use of BBVERSIONS is somewhat cleaner than using a .inc -- at least it reduces the number of files you have to look through Jan 20 03:03:04 * kergoth` grumbles Jan 20 03:26:46 I'm writing a recipe and I have a : in the URL of the git I'm pulling. OE interprets the : as port Jan 20 03:26:49 how can I avoid that? Jan 20 06:13:17 hello Jan 20 06:13:56 I m new to openembedded. I had prepared my own distribution bt now i m getting an error Jan 20 06:13:59 ERROR: Could not inherit file classes/rootfs_${IMAGE_PKGTYPE}.bbclass while parsing Jan 20 07:55:03 good morning Jan 20 08:41:14 sakoman__: ping Jan 20 09:36:12 mornign Jan 20 09:49:09 hi hrw Jan 20 10:55:17 mornign everyone Jan 20 11:09:33 hm.. Jan 20 11:09:58 trying to boot minimal/qemux86 in qemu Jan 20 11:10:16 image native-sdk-image Jan 20 11:10:24 have no network at all Jan 20 11:11:37 ah.. no network card in kernel and no modules installed.. Jan 20 11:12:46 anyway, it have same problems with autom4te etc. Jan 20 11:36:12 morning Jan 20 11:38:07 can anyone help me or suggest some example to create our own distribution.... Jan 20 11:38:43 sanket: first thing is do you need your own distro? Jan 20 11:39:00 any git geniouses online? Jan 20 11:39:11 how to create our own distro... give simple example... Jan 20 11:39:57 give me basic sample.conf distro file Jan 20 11:45:25 sanket: micro is about the simplest example there is Jan 20 11:48:20 #XorA: can u explain me how bitbake handles it... Jan 20 11:48:51 sanket: go read the manuals, they explane it all Jan 20 11:50:17 I hd gone through the manual.... bt still I m getting some error... Jan 20 11:53:12 actually I want create my own distro which contains features from ....chapter 5 .Constructing a Temporary System .... http://www.linuxfromscratch.org/lfs/view/development/ Jan 20 11:56:26 sanket: do you really want new own distro? may be just right image would be sufficient? Jan 20 11:59:55 bt how build filesystem with same version as they provided in Temporary system.... since I m not getting it how build filesystem with same version Jan 20 12:01:22 where to modify my version so that bitbake vll take only that version package to build a filesystem... Jan 20 12:12:30 sanket: first do build of micro or angstrom to get used to system. then try Jan 20 12:13:43 I hd already tried.. now I wnt to knw how bitbake handles it.... there are lots of variable... Jan 20 12:14:39 I hd prepared a sample-image.bb in /recipes/images/ ... without defining any package into it... Jan 20 12:28:12 03Stefan Schmidt  07master * rb203a051cc 10openembedded.git/recipes/buglabs/ (bug-avahi.bb bug-avahi/bugdevice.service): bug-avahi: Add avahi service file for bug20 Jan 20 12:28:21 03Stefan Schmidt  07master * r15190c7ca1 10openembedded.git/classes/bug-image.bbclass: bug-image.class: Update version info Jan 20 12:28:22 03Stefan Schmidt  07master * r65e97ee9eb 10openembedded.git/recipes/linux/linux-omap-hah_2.6.31.bb: linux-omap-hah_2.6.31: Bump PR for Option and LTE driver Jan 20 12:41:31 Someone knows about bitbake problem with rxvt? Jan 20 12:41:32 File "/home/otavio/hacking/bitbake/lib/progressbar.py", line 368, in start Jan 20 12:41:32 assert self.maxval > 0 Jan 20 12:53:17 how to set required version in openembedded? Jan 20 13:03:53 ERROR: Function 'Fetch failed: Unable to fetch URL ftp://invisible-island.net/ncurses/5.7/ncurses-5.7-20100501.patch.gz;name=p20100501 from any source.' failed Jan 20 13:03:56 ERROR: Task 812 (virtual:native:/home/otavio/hacking/embedded-linux/openembedded/recipes/ncurses/ncurses_5.7.bb, do_fetch) failed with exit code '1' Jan 20 13:23:06 Can someone put ftp://invisible-island.net/ncurses/5.7/ncurses-5.7-20100501.patch.gz on mirror? Jan 20 13:39:47 Does git automatically convert line endings in patches? Jan 20 13:40:12 I have a patch which won't apply because the line endings are LF, whilst in the working patch they are CRLF Jan 20 13:40:28 I've already pushed the patch twice Jan 20 13:42:41 Might be git-send-email Jan 20 13:44:19 <_alx_> hi all Jan 20 13:44:48 B_Lizzard: aer you generating patches under windows? Jan 20 13:44:59 Nope Jan 20 13:45:13 But the original file had CRLF line endings Jan 20 13:45:15 mckoan: hmmm... Jan 20 13:45:20 fromdos Jan 20 13:45:24 <_alx_> does anyone is working with the qt4-x11-base-image or demo-image ? I'm having some troubles (it used to be the ts_calibrate but now it's missing) and how to autostart Xserver ? Jan 20 13:45:33 03Stefan Schmidt  07master * r904e9ad416 10openembedded.git/recipes/buglabs/ (5 files in 3 dirs): matchbox-theme-bug: Add matchbox themes for buglabs Jan 20 13:45:38 <_alx_> I think I'm missing some base features of OE :D Jan 20 13:45:47 So diff -u produces a patch with CRLF line endings for the file's lines Jan 20 13:45:55 _alx_: AFAIK nowadays we use xcalibrate Jan 20 13:46:15 <_alx_> mckoan, ok this solves part of my questions :D Jan 20 13:47:03 _alx_: autostart Xserver adding it in init.d (rc5.d) Jan 20 13:47:23 The patch doesn't apply with LF line endings Jan 20 13:47:37 Bloody java crap Jan 20 13:47:41 <_alx_> mckoan, i don't have any xcalibrate binary on my qt4-x11-base-image Jan 20 13:48:30 I suspect you have autocrlf enabled or something on the git server Jan 20 13:48:52 <_alx_> mckoan, what I'm supposed to write (and where) in order to obtain an image which autostarts Xserver ? Jan 20 13:51:10 _alx_: modify/create your own recipe (in xorg-* ?) Jan 20 13:51:14 not sure Jan 20 13:52:22 _alx_: if I knew I would have already done it :-) Jan 20 13:52:28 mckoan, it would probably have to be fixed from the inside or something. Jan 20 14:08:52 _alx_: check qt4-x11-demo-image is launch a screen calibration and launch X + qtdemo Jan 20 14:09:26 "it launch" sorry Jan 20 14:09:38 <_alx_> oh Jan 20 14:10:11 <_alx_> just back and read (thanks ericben), going to read (just launched bitbake qt4-x11-demo-image) Jan 20 14:26:58 ERROR: Function 'Fetch failed: Unable to fetch URL cvs://anoncvs:anoncvs@anoncvs.handhelds.org/cvs;module=ipkg-utils from any source.' failed Jan 20 14:27:01 ERROR: Task 193 (/home/otavio/hacking/embedded-linux/openembedded/recipes/ipkg-utils/ipkg-utils-native_1.6+cvs20050404.bb, do_fetch) failed with exit Jan 20 14:28:13 why would a image not load a firmware for wifi card ? do i miss a package which does the loading ( but i assumed udev will do that ) Jan 20 14:33:39 03Marco Cavallini  07master * r8b3b078b21 10openembedded.git/recipes/psplash/psplash-kaeilos_svn.bb: psplash-kaeilos_svn.bb: Added KaeilOS specific psplash recipe Jan 20 14:33:53 03Marco Cavallini  07master * rdc63b49ec3 10openembedded.git/recipes/psplash/files/psplash-24bpp-BGR666.patch: psplash-24bpp-BGR666.patch: psplash patch to support 24bpp BGR666 Jan 20 14:33:54 03Marco Cavallini  07master * r8c448a064e 10openembedded.git/conf/distro/kaeilos.conf: Jan 20 14:33:54 kaeilos.conf: changed KaeilOS specific psplash-kaeilos Jan 20 14:33:54 * bump PR Jan 20 14:44:47 hi all Jan 20 14:45:26 any idea why glibc-2.9-r37 throws me this error: configure: error: C preprocessor "arm-angstrom-linux-gnueabi-gcc -E" fails sanity check Jan 20 14:46:34 <_alx_> could someone help me in configuring a local overlay where to put all my own files (from conf/machine/bla bla bla) to recipes ? Jan 20 14:47:08 hi ericben Jan 20 14:48:17 _alx_: try this http://kom.kaeilos.com/ Jan 20 14:53:19 <_alx_> mckoan, tried but it goes beyond my comprehension :( Jan 20 14:54:12 _alx_: np, too complex? Jan 20 14:54:52 <_alx_> mckoan, don't know, tried to, but didn't understand well how it works Jan 20 14:55:48 _alx_: hmmm.... exaples and documents didn't help? Jan 20 14:57:14 <_alx_> mckoan, well, I'm reading docs but my feeling is near to trial and error :( Jan 20 15:00:21 _alx_, what about the manual way? Jan 20 15:00:30 trough local.conf Jan 20 15:00:50 ceyusa, look at the logs Jan 20 15:00:55 like config.log Jan 20 15:01:22 <_alx_> GNUtoo|laptop, I'm tring http://wiki.openmoko.org/wiki/OpenEmbedded#BitBake_Overlays Jan 20 15:01:45 <_alx_> but bitbake doesn't eat my customized machine.conf Jan 20 15:02:02 GNUtoo|laptop: done, it says that the preproc couldn't find the Jan 20 15:02:19 _alx_: because machine is not under recipes Jan 20 15:02:40 <_alx_> mckoan, I created a local/conf/machine/myowndevice.conf Jan 20 15:03:06 <_alx_> just copying the original qemux86.conf to test to myowndevice.conf Jan 20 15:03:09 <_alx_> to test Jan 20 15:03:25 _alx_: that's the standard way, must work Jan 20 15:03:27 <_alx_> but it complains that it cannot find endianess o.O Jan 20 15:03:31 ceyusa, which glibc is it, initial? cross? Jan 20 15:03:46 <_alx_> let me do some more "trial and error" :P Jan 20 15:04:09 _alx_, I suspect you have setup the overlay for recipes Jan 20 15:04:17 like they do in the manual Jan 20 15:04:31 * GNUtoo|laptop looks at the link Jan 20 15:04:43 <_alx_> GNUtoo|laptop, I don't really know what I'm doing Jan 20 15:04:45 GNUtoo|laptop: none, I want to fulfill the virtual-libc-dev for the task-sdk-native.bb Jan 20 15:04:58 _alx_, ah ok bitbake overlays, sorry my bad Jan 20 15:05:20 GNUtoo|laptop: so it's the glibc_2.9.bb Jan 20 15:05:47 ceyusa, yes but what do you see in console when trying to compile it? Jan 20 15:05:55 glibc-2.9-r37 ? Jan 20 15:06:14 GNUtoo|laptop: glibc-2.9-r37.4 Jan 20 15:06:25 ok Jan 20 15:08:51 ceyusa, if no one knows ask in the mailing list but do include the interesting part of config.log Jan 20 15:09:17 GNUtoo|laptop: ok, thanks :) Jan 20 15:11:08 <_alx_> GNUtoo|laptop, ERROR: Unable to determine endianness for architecture 'INVALID' :( Jan 20 15:11:51 _alx_, yes that's because it was not overriden Jan 20 15:12:59 ericben, on what mailing list should I send my machine support for the imx31 bug base? Jan 20 15:13:15 <_alx_> GNUtoo|laptop, I got it, but how to solve ? Jan 20 15:13:36 basically it works that way Jan 20 15:13:48 the machine config is read Jan 20 15:13:54 which normally set the architecture Jan 20 15:13:57 or the cpu Jan 20 15:14:01 or something like that Jan 20 15:14:05 that overrides the endianess Jan 20 15:14:26 so or your machine config is not read Jan 20 15:14:35 or your machine config has an issue Jan 20 15:15:08 <_alx_> it is not read! solved part of the problem adding local overlay directory to BBPATH Jan 20 15:15:12 <_alx_> now it seems to work Jan 20 15:15:14 ok Jan 20 15:15:55 03Stefan Schmidt  07master * rd3167963e9 10openembedded.git/recipes/icedtea/ (13 files in 2 dirs): Jan 20 15:15:56 icedtea6-native: Add version 1.7.6 Jan 20 15:15:56 Refresh icedtea-fix-jar-path.patch for this release. Jan 20 15:16:42 Is autocrlf enabled on the oe git server? Jan 20 15:16:57 I need to push a patch that has CRLF line endings Jan 20 15:17:16 Actually, the patch is already up on openembedded, but it doesn't apply Jan 20 15:17:37 Because the line endings are automatically converted to LF Jan 20 15:17:51 B_Lizzard: what about to call 'fromdos patchname' before? Jan 20 15:18:04 Hmmm Jan 20 15:18:07 it's a patch Jan 20 15:18:21 he should apply it manually etc... Jan 20 15:18:23 that's ugly Jan 20 15:18:38 or maybe add a task before do_patch Jan 20 15:18:45 that seem ugly Jan 20 15:19:00 GNUtoo|laptop: ah... now I understand Jan 20 15:19:05 I'm trying to find a solution with the least issues. Jan 20 15:19:25 As far as I can tell I cannot push a patch with CRLF line endings Jan 20 15:19:39 So I may end up replacing the entire file or something like that Jan 20 15:19:41 GNUtoo|laptop: B_Lizzard said he needed to push, so I thought to push in the repo Jan 20 15:19:54 Yes, push in the repo Jan 20 15:20:10 The patch in question is under midpath/files/sdl-fixes.patch Jan 20 15:20:33 is the patch somewhere on the internet? Jan 20 15:20:42 because you could fetch the patch Jan 20 15:20:45 Um, in OE Jan 20 15:21:17 if it was on the internet you would just have to add to SRC_URI and the checksums Jan 20 15:21:23 As I said, the patch is already in OE but doesn't apply because of the LF line endings Jan 20 15:21:25 and it would solve the whole issue Jan 20 15:21:30 Ah Jan 20 15:21:36 That makes sense Jan 20 15:21:38 and remove the present oe patch Jan 20 15:21:41 I'll do that Jan 20 15:21:46 B_Lizzard: I see only /n there, like any patch Jan 20 15:21:51 Great idea, yeah. Jan 20 15:22:05 With Geany you can see what kind of newline it is Jan 20 15:22:20 hexdump -c filename Jan 20 15:22:28 and maybe mirror it somewhere too in case it disappears from the internet Jan 20 15:22:41 I'll upload it to jlime.com Jan 20 15:22:45 ok Jan 20 15:22:53 you're the jlime dev? Jan 20 15:22:53 That idea should work, thanks for that. Jan 20 15:22:57 Aye Jan 20 15:23:09 nice, how's nanonote support? Jan 20 15:23:20 Great, actually. Jan 20 15:23:29 everything is framebuffer based? Jan 20 15:23:34 or is there some kind of X Jan 20 15:23:42 Xfbdev Jan 20 15:23:43 http://jlime.com/wiki/news/muffinman_preview_release Jan 20 15:23:48 ohh nice kdrive Jan 20 15:23:58 unmaintained but that was the best solution Jan 20 15:23:59 A preview of what I'm doing Jan 20 15:24:10 Yeap, Xorg takes up 10% more RAM Jan 20 15:24:19 wow only 10% Jan 20 15:24:28 About 4MB Jan 20 15:24:34 ok Jan 20 15:24:36 acceptable Jan 20 15:24:51 you have 32M Jan 20 15:24:52 Workable, but I'd rather not have it bog down the system Jan 20 15:24:58 ok Jan 20 15:25:01 Yeah :) Jan 20 15:26:14 Everything's in OE Jan 20 15:28:59 03Koen Kooi  07org.openembedded.dev * r3eca7c2641 10openembedded.git/recipes/buglabs/matchbox-theme-bug.inc: Jan 20 15:28:59 matchbox-theme-bug: remove AUTOREV to fix parsing Jan 20 15:28:59 Signed-off-by: Koen Kooi Jan 20 15:35:48 hi Jan 20 15:35:58 ping stefan_schmidt Jan 20 15:36:13 hi Jan 20 15:36:19 hi gnutoo Jan 20 15:44:44 B_Lizzard, wow it seem incredible, you can see videos with nanonote..... Jan 20 15:44:53 what are they encoded in? Jan 20 15:45:07 I think MPEG2 has the best speed Jan 20 15:45:17 But I've gotten some good results with Xvid Jan 20 15:45:25 ok Jan 20 15:45:30 that's what I used on htc dream Jan 20 15:45:32 mpeg2 Jan 20 15:45:36 Yeah Jan 20 15:45:47 what's your mips cpu speed? Jan 20 15:45:47 Haven't tried any h.264, I should Jan 20 15:45:54 not sure.... Jan 20 15:45:56 lol Jan 20 15:46:07 maybe h264 ultra low profile Jan 20 15:46:09 if that exists Jan 20 15:46:13 Three hundred something MHz Jan 20 15:46:17 324? Jan 20 15:46:20 Can't remember Jan 20 15:46:24 ok Jan 20 15:46:31 how can it view videos so fast.... Jan 20 15:46:40 on my freerunner I've not that speed Jan 20 15:46:45 and it's armv4 400Mhz Jan 20 15:46:59 Well, there exists an optimized version of mplayer for the specific CPU Jan 20 15:47:12 ahh Jan 20 15:47:32 I use mainline mplayer builds which work fine but I guess the optimized build would get better results all over the place Jan 20 15:47:38 the cpu has something like SMID or something like that Jan 20 15:48:07 maybe the resolution is lowrer in your case: Jan 20 15:48:09 03Stefan Schmidt  07master * rd03d153a2b 10openembedded.git/recipes/buglabs/matchbox-theme-bug.inc: Jan 20 15:48:10 matchbox-theme-bug: Set SRCREV to correct value. Jan 20 15:48:10 Sorry for the AUTOREV problem I introduced. Slipped through my review. Thanks Jan 20 15:48:10 for fixing. Jan 20 15:48:10 320x240 Jan 20 15:48:23 pong woglinde Jan 20 15:48:29 woglinde: how are you? Jan 20 15:48:43 stefan hm you could use filespath to reduce the patch copy Jan 20 15:48:59 stefan fine at least Jan 20 15:49:11 but behind my timetable Jan 20 15:49:13 woglinde: I thought about it. At least one patch needed to get refreshed. Jan 20 15:49:41 woglinde: Any good way to use filepath and still have one patch differ? Jan 20 15:50:15 it will first lookup in the new-rev-dir Jan 20 15:50:21 and than in the other dir Jan 20 15:50:41 woglinde: I also was thinking about removing 1.7.5 as it has security problems Jan 20 15:50:44 so you only need to renew this one patch Jan 20 15:51:01 woglinde: ah, did not know that. That would make it cleaner Jan 20 15:51:10 hm I will commit my 1.8.3 recipe this evening Jan 20 15:51:21 which uses this feature Jan 20 15:51:26 woglinde: fine with me as well Jan 20 15:51:56 woglinde: So I will wait on your stuff and adjust 1.7.6 afterwards Jan 20 15:52:12 hm now its okay Jan 20 15:52:17 but for the next version Jan 20 15:52:49 well, if we can save so many duplicates I should take care of this one as well :) Jan 20 15:53:00 Will do if we don't want to drop 1.7.5 Jan 20 15:53:46 woglinde: What else do you have in the pipeline? Only updated icedtea6-native or also newer openjdk6? Jan 20 15:54:25 B_Lizzard, did you publish the script you used for compressing the videos that runs on the nanonote? Jan 20 15:55:08 stefan ah no its openjdk 1.8.3 Jan 20 15:55:22 did not looked at native for a long time Jan 20 15:55:51 woglinde: so no update on native from your side? Jan 20 15:56:03 woglinde: will look into filepath usage then for it Jan 20 15:56:22 GNUtoo|laptop, this is for the Jornada but I think this is the script we use: http://jlime.com/wiki/documentation/user/7xx/howtos/converting_videos Jan 20 15:56:39 woglinde: will you also update the openjdk build to b20? Jan 20 15:57:25 ok Jan 20 15:57:38 hm isnt that needed for 1.9 openjdk? Jan 20 15:57:40 hmm, does b20 actually build with 1.8.3? Jan 20 15:57:42 heh Jan 20 15:57:49 yeah, got it wrong Jan 20 15:58:22 I was thinking about adding a build with 1.9.x and b20 for some testing. Jan 20 15:58:52 http://wiki.openmoko.org/wiki/Video_Player#Encoding was the script I used on htcdream but I never got video working fullscreen on om-gta02 Jan 20 15:58:54 stefan I will commit 1.8.3 with def_pref -1 Jan 20 15:59:04 stefan when its running fine for the bug Jan 20 15:59:14 you can remove it Jan 20 15:59:19 woglinde: ok Jan 20 15:59:36 hm or wait Jan 20 15:59:37 please report Jan 20 15:59:56 I will write to the ml than and remove 1.8 for security issues Jan 20 16:00:06 so distro maintainers can pin the stuff Jan 20 16:00:25 hm okay Jan 20 16:00:30 will go now Jan 20 16:00:38 the session for my son is nearly over Jan 20 16:00:41 till later Jan 20 16:00:45 cu Jan 20 16:00:59 * GNUtoo|laptop should test java on armv6 at some point Jan 20 17:14:36 gm Jan 20 17:15:02 cbrake: haven't had time to update my testing entry yet, will do that in the next 15 mins or so. Jan 20 17:15:13 ok, thanks Jan 20 17:25:51 cbrake, already done, was a good run Jan 20 17:26:00 great Jan 20 17:26:57 actually the issue is sometimes not running the tests but thinking of and finding time to report the results Jan 20 17:27:00 eFfeM: I'm still scrambling for disk space, but I should be up and running for the next round Jan 20 17:27:06 cool Jan 20 17:27:18 eFfeM: newegg send me a DOA disk, so that set me back a week Jan 20 17:27:31 will you also be starting to produce the weekly log ? Jan 20 17:27:42 eFfeM: yes Jan 20 17:27:55 anyone familiar with gtk-doc? for several weeks in the weekly test for neek I get Jan 20 17:27:56 ERROR: QA Issue with staging: libsupc++.la failed sanity test (workdir) in path /home/hudson/jobs/OE_test/workspace/tmp/sysroots/nios2-linux/usr/lib ERROR: QA Issue with staging: libstdc++.la failed sanity test (workdir) in path /home/hudson/jobs/OE_test/workspace/tmp/sysroots/nios2-linux/usr/lib Jan 20 17:28:06 but all other targets build fine Jan 20 17:32:48 eFfeM, are you building for x86 at all? do we still have that x264 issue with it? Jan 20 17:32:57 i'll have to kick off a build tonight Jan 20 17:35:08 ~seen grg Jan 20 17:35:14 grg <~grg@eth7090.sa.adsl.internode.on.net> was last seen on IRC in channel #oe, 1d 14h 47m 20s ago, saying: 'that could probably be added though... Both libcurl and wget support such things.'. Jan 20 17:36:01 tharvey: haven't build x264 for quite a while, so no idea Jan 20 17:45:01 eFfeM, not something I care about either but I threw x86 machine into my test matrix just because it seemed like something people would use Jan 20 17:45:08 don't use it myself Jan 20 17:51:33 neither do I at the moment Jan 20 17:59:23 Do someone knows why SDK fails to build with: Jan 20 17:59:24 | /home/otavio/hacking/embedded-linux/tmp/work/x86_64-i586-sdk-oe-linux-uclibc/gcc-cross-sdk-4.5-r31.1+svnr168622/gcc-4_5-branch/build.x86_64-linux.i586-oe-linux-uclibc/i586-oe-linux-uclibc/libstdc++-v3/include/bits/locale_facets_nonio.tcc:594:57: error: there are no arguments to '_S_get_c_locale' that depend on a template parameter, so a declaration of '_S_get_c_locale' must be available Jan 20 17:59:30 | make[4]: *** [i586-oe-linux-uclibc/bits/stdc++.h.gch/O2g.gch] Error 1 Jan 20 18:03:21 wiki editing seems to be broken "Input error: Invalid referer " Jan 20 18:03:33 Crofton|work: ping Jan 20 18:05:31 otavio, yeah, been seeing that for forever Jan 20 18:05:39 XorA: I managed to update the testing page about 40 minutes ago Jan 20 18:05:45 not sure why only gcc-cross-sdk does it Jan 20 18:06:09 otavio: Tartarus its locale related. Jan 20 18:06:22 will bitbake gcc-cross-sdk reproduce it ? Jan 20 18:06:37 khem: I guess so; but against uclibc distro Jan 20 18:06:44 khem: I did it using micro-uclibc Jan 20 18:09:06 nopes wiki is certainly broken for editing :-( Jan 20 18:12:03 XorA, you need an account and account creation is locked Jan 20 18:12:09 the spammers are winning ... Jan 20 18:12:17 I just edited it ,... Jan 20 18:12:40 I am not at the Crofton|work machine for a week ... Jan 20 18:12:48 Crofton: I have an account Jan 20 18:13:02 Crofton: logged in and everything Jan 20 18:13:09 check with ka6sox Jan 20 18:13:14 Crofton: the error message is non sensible Jan 20 18:13:16 maybe he needs to do something .... Jan 20 18:13:45 ka6sox: hi, I cant edit the wiki, error message is "Input error: Invalid referer " which is non sensible as an error Jan 20 18:14:11 Crofton: actually is that page locked to board only? Jan 20 18:14:54 hmm Jan 20 18:14:56 not sure Jan 20 18:15:02 it should be :-) Jan 20 18:15:10 true Jan 20 18:37:36 hello Jan 20 18:38:25 does oe support the ka-ro tx25 board? Jan 20 18:40:32 I dont think so Jan 20 18:40:53 karo kernel support is a bit ropey from memory of messing with their pxa ones Jan 20 18:41:48 Reason I am asking is that it seems to have a Bitbake machine Jan 20 18:44:22 * XorA cant see it in meta data Jan 20 18:44:26 but i could be blind Jan 20 19:06:22 bertszoghy, hi do you have details on this board? Jan 20 19:07:00 tx25.conf Jan 20 19:07:29 bertszoghy, yes oe supports it Jan 20 19:08:05 http://cgit.openembedded.org/cgit.cgi/openembedded/tree/conf/machine/tx25.conf Jan 20 19:08:26 GNUtoo|laptop, very interesting, thank you. Jan 20 19:08:50 Is there a way to avoid the expansion of ${variable} when bitbake parses recipe files? Jan 20 19:09:05 mario-goulart, yes using special quoting Jan 20 19:09:15 I don't remember how Jan 20 19:09:19 you will have to grep Jan 20 19:09:35 let me try to remember Jan 20 19:09:38 Nice. Thanks, GNUtoo|laptop Jan 20 19:09:52 maybe it was with passing flags Jan 20 19:09:57 to Make Jan 20 19:10:06 grep for EXTRA_OEMAKE Jan 20 19:10:16 playing with " and ' could do it Jan 20 19:10:23 else you have the \ Jan 20 19:11:17 beep/beep_1.2.2.bb:EXTRA_OEMAKE += 'CC="${CC}"' Jan 20 19:11:45 try that Jan 20 19:11:48 and look if it works Jan 20 19:11:58 I'm not sure if it works Jan 20 19:12:54 Hmmm. I'm not sure they work, since I've been reading some funky substitutions (via sed) to workaround this behavior. Things like 's#I/usr/include#Iincludedir}#g; s#Iin#I${in#g' Jan 20 19:13:12 fbreader/fbreader_0.12.1.bb:EXTRA_OEMAKE = "CC='${CXX}' Jan 20 19:32:10 is it possible to modify a file in do_rootfs from a package's pkg_postinst(), e.g. add a line to inetd.conf so that it doesn't need to be added on the first boot? Jan 20 19:32:52 i tried ${D}${sysconfdir}/etc/inetd.conf, but ${D} points to the package's work directory Jan 20 19:33:40 obi, there was a way to do that but I don't remember how, it went on the mailing list tough Jan 20 19:33:48 I've no idea on what keyword to search Jan 20 19:35:41 i'll try using $D instead of ${D}... maybe the latter gets subtituted too early Jan 20 19:39:12 :-) Jan 20 19:39:22 it worked, thanks! i found this on the ml ;) Jan 20 19:48:46 03Khem Raj  07master * r576df0d476 10openembedded.git/recipes/uclibc/ (7 files in 2 dirs): Jan 20 19:48:47 uclibc_git.bb : Update to 0.9.32-rc2 Jan 20 19:48:47 Signed-off-by: Khem Raj Jan 20 19:49:59 GNUtoo|laptop, thank you I was able to bitbake a base image and kernel for the TX25 Jan 20 19:50:34 ok nice, you have a quick machine Jan 20 19:50:40 or an already setup build system Jan 20 19:51:28 does someone know how to transfert files trough ymodem, specially the target part Jan 20 19:51:36 yes, I've bitbaked a lot for the TI AM1808 Jan 20 19:51:42 ok Jan 20 19:53:39 ymodem file send on http://quadmore.com/blog.php#ipaq Jan 20 19:54:02 ok I've looked but I've no rz Jan 20 19:54:07 I'll look at your article Jan 20 19:54:16 re Jan 20 19:54:22 Proceed to send or "upload" the jffs2 file (from the tarball that you downloaded earlier) with ymodem Jan 20 19:54:36 With minicom, use the key sequence: Jan 20 19:54:41 ah they use lrsz Jan 20 19:54:49 I need that on target Jan 20 19:54:49 ctrl-A-z-s Jan 20 19:54:54 basically here's my issue Jan 20 19:55:00 I'm porting a machine to linux-next Jan 20 19:55:09 it boots Jan 20 19:55:15 to initarmfs.... Jan 20 19:55:23 but it ask for the wifi firmware Jan 20 19:55:30 I don't want to re-do an initarmfs Jan 20 19:55:39 so I wondered if I could send something with serial Jan 20 19:55:53 the issue is that I've no rz on the target Jan 20 19:56:47 no tftp? Jan 20 19:56:57 nothing Jan 20 19:56:59 only serial Jan 20 19:57:11 ok I'll re-do an initramfs then Jan 20 19:58:11 no removable storage either I guess? Jan 20 19:58:20 no Jan 20 19:58:26 else it would be easy Jan 20 19:58:34 basically it fails with an error Jan 20 19:58:37 -84 I think Jan 20 19:58:39 which means: Jan 20 19:58:52 #define EILSEQ 84 /* Illegal byte sequence */ Jan 20 20:02:31 maybe I should partitionate the NAND better Jan 20 20:02:42 but as it's dangerous to mess with partitons.... Jan 20 20:14:24 GNUtoo|laptop, haven't done that. Jan 20 20:15:08 other question: is there a reference on how to add new kernel module code to a Bitbake recipe? Jan 20 20:26:53 I have sent mail about wrong paths inside autoconf package files (contain sysroots path) Jan 20 20:27:34 not sure what images are affected, but native-sdk-image is Jan 20 20:28:42 bertszoghy, kernel module code? Jan 20 20:28:48 you mean ship a module in rootfs? Jan 20 20:33:51 GNUtoo|laptop: yes, add a new .c file to be compiled into a .ko Jan 20 20:34:00 ah ok Jan 20 20:34:13 so you want a separate module Jan 20 20:34:29 so make your separate module Jan 20 20:34:37 create the Makefiles for it Jan 20 20:34:45 Yes. I'm at the "hello world" stage right now, looking for bitbake docs. Jan 20 20:34:51 and look at the recipes that uses separate modules Jan 20 20:34:53 ah ok Jan 20 20:34:56 that's simple then Jan 20 20:34:59 use the devshell Jan 20 20:35:02 bitbake devshell Jan 20 20:35:06 and source it Jan 20 20:35:17 mine is at: ~/embedded/oe/oetmps/bug/deploy/glibc/addons/arm-angstrom-linux-gnueabi-angstrom-bug-devshell Jan 20 20:38:52 GNUtoo|laptop: you're saying start up a command line shell right in the middle of the bitbake operation and manually compile the kernel module? Am assuming Bitbake remembers its cross-compiler variables? Jan 20 20:39:09 no Jan 20 20:39:16 I'm saying to bitbake a devshell Jan 20 20:39:20 which create a shell script Jan 20 20:39:25 that you source after compiling it Jan 20 20:41:59 was reading http://docs.openembedded.org/usermanual/html/usage_devshell.html Jan 20 20:48:11 oe_runmake Jan 20 20:48:14 is like make Jan 20 20:48:24 but do LDFLAGS="" Jan 20 20:48:29 in the case of kernel builds Jan 20 20:48:36 or unset ldflags Jan 20 20:48:38 like that Jan 20 20:48:40 unset LDFLAGS Jan 20 20:49:28 Am quite a bit confused. So I'll do MACHINE=tx25 bitbake base_image -c devshell and then? Jan 20 20:49:38 source it Jan 20 20:49:44 source the devshell Jan 20 20:50:03 and you'll have the good path exported and some other shell commands Jan 20 20:50:15 for instance arm-foo-gcc Jan 20 20:50:18 will be in the path Jan 20 20:50:35 and you're wrong Jan 20 20:50:36 yes Jan 20 20:50:42 do that: Jan 20 20:50:49 MACHINE=tx25 bitbake devshell Jan 20 20:50:57 devshell is a target not a task Jan 20 20:51:13 k Jan 20 20:58:16 devshell is also a task bitbake -c listtasks foo ;) Jan 20 21:04:36 What's the OE hotel for FOSDEM this year? Jan 20 21:12:00 so did MACHINE=tx25 bitbake devshell and it compiled. I see I have ./arago-tmp/deploy/glibc/addons/arm-none-linux-gnueabi-angstrom-tx25-devshell and timestamp for that is a few minutes ago. Jan 20 21:15:59 Do I modify an image bb file next or a package bb file (???) Jan 20 21:18:09 and where do I put my helloworld.c file? Jan 20 21:26:30 OK, ./arago-tmp/deploy/glibc/addons/arm-none-linux-gnueabi-angstrom-tx25-devshell is executable Jan 20 21:26:41 and echo $CC gives me Jan 20 21:27:07 arm-none-linux-gnueabi-gcc -march=armv5te -mtune=arm926ej-s -mthumb-interwork -mno-thu Jan 20 21:29:37 I do export LDFLAGS="" Jan 20 21:30:02 cd to my demoARM dir where I have a helloworld.c file and a Makefile Jan 20 21:30:37 I do oe_runmake Jan 20 21:33:07 For compiling a module on my host PC in my Makefile I do Jan 20 21:33:19 default Jan 20 21:33:31 make -C /lib/modules/`uname -r`/build M=/home/bertrand/demo modules Jan 20 21:33:40 What do I change it to for oe_runmake? Jan 20 21:36:39 03Khem Raj  07master * rb86b1570bd 10openembedded.git/recipes/gcc/ (44 files): Jan 20 21:36:40 gcc-cross,gcc-cross-sdk,gcc-cross-canadian-sdk: Disable nls when using uclibc Jan 20 21:36:40 Signed-off-by: Khem Raj Jan 20 21:39:56 bertszoghy, unset LDFLAGS Jan 20 21:40:03 s/make/oe_runmake Jan 20 21:40:08 and you should be ok Jan 20 21:40:23 else it exports rpath etc... that are for userspace Jan 20 21:42:42 Yes indeed Jan 20 21:42:47 It worked. Jan 20 21:42:53 hello.ko Jan 20 21:42:57 amazing. Jan 20 21:43:07 Thank you very much! Jan 20 21:45:18 btw why not building an internal module Jan 20 21:45:21 rather than external Jan 20 21:45:36 you just have to add to Kconfig and Makefiles Jan 20 21:53:07 I'm not quite all there yet. Still figuring my way around. Jan 20 22:01:55 Wrong, I compiled a .ko for ther host not the target. What should I put in the Makefile, -C is wrong: make -C /lib/modules/${KDIR}/build M=/home/bertrand/demoARM modules ?? Jan 20 22:03:02 Does Bitbake have a variable I can replace that with? I checked with set in my [OE::arm-none-linux-gnueabi-angstrom-tx25]:~ command line Jan 20 22:10:43 bertszoghy: it seems it get the CROSS_COMPILE bits wrong I suppose Jan 20 22:10:46 or CC bits Jan 20 22:14:42 Kernel configuration is invalid on my external dir, where is it in Bitbake? Jan 20 22:15:05 huhu Jan 20 22:15:14 hehe Jan 20 22:18:01 mwester: uslugc is functional now :) Jan 20 22:18:09 Wow. Jan 20 22:18:14 mwester: image size is 2.6M Jan 20 22:18:46 mwester: would you mind if I switch the perl to 5.10.1 for slugos distros Jan 20 22:19:03 I was planning to do that, actually -- so that would be good. Jan 20 22:19:05 mwester: and its boots well from nfs so far Jan 20 22:19:12 cool Jan 20 22:19:20 so I have your pseudo ack Jan 20 22:19:39 :) I'm here, so I can give it a real ack too. Jan 20 22:19:47 ok lemme post Jan 20 22:20:59 sent Jan 20 22:21:17 mwester: you should plan on releasing a newer version of slugos Jan 20 22:21:38 I am planning to do so. Need to fix a few issues in turnup, and then do some testing. Jan 20 22:21:48 cool Jan 20 22:21:51 testing? Jan 20 22:21:54 :) Jan 20 22:21:56 please consider ucslugc Jan 20 22:22:08 Crofton: slugos is a real distro Jan 20 22:22:09 :) Jan 20 22:22:35 mwester: I could not get opkg to install from my local feeds Jan 20 22:22:47 make -C /home/bertrand/oe/arago-tmp/sysroots/tx25-none-linux-gnueabi/kernel M=/home/bertrand/demoARM modules Jan 20 22:22:51 I'll probably release a version for the NSLU2 LE only, without support for RAID, and minimal testing -- just to try to pull in a few community members to test a "full" release later on (adding the missing stuff). But I'll see what feedback I get. Jan 20 22:22:54 mwester: that stopped me from running gcc cross tests using nslu Jan 20 22:22:56 kergoth: why would you need opkg on windows? Jan 20 22:22:59 make[1]: Entering directory `/home/bertrand/oe/arago-tmp/sysroots/tx25-none-linux-gnueabi/kernel' Jan 20 22:23:08 make[1]: *** No rule to make target `modules'. Stop. Jan 20 22:23:30 mwester: hmmm all my testing so far has been on BE Jan 20 22:23:40 Excellent -- mine has been on LE. Jan 20 22:23:44 So both work still! Jan 20 22:23:46 but LE should just work TM Jan 20 22:23:58 That's why I switched to LE. Jan 20 22:24:06 yes one incentive for me to use nslu is BE Jan 20 22:24:12 so I can test BE gcc Jan 20 22:24:16 Ok. Jan 20 22:24:21 We can release both, that's easy. Jan 20 22:24:53 2.6M of jffs2 arent you impressed :) Jan 20 22:25:15 thats slugos-image size btw Jan 20 22:25:31 Jin^eLD: I've worked on getting OE builds working under cygwin and msys on multiple occasions. Jan 20 22:25:39 As an FYI, the reason for BE SlugOS at all is historical -- originally the NIC would only run in BE mode, so the Linksys firmware, and Unslung, and early SLugOS all ran BE. Jan 20 22:25:45 heh cygwin Jan 20 22:25:50 kergoth: I was just curious about the use case :) Jan 20 22:25:57 lot of folks still use it I know for sure Jan 20 22:25:59 * kergoth would prefer msys over cygwin, but bitbake itself has portability issues there Jan 20 22:26:01 khem, I should compare that to openwrt on NSLU2 -- if its close, that would be excellend. Jan 20 22:26:06 er, excellent. Jan 20 22:26:27 Does that use busybox-mdev, or udev? Jan 20 22:26:30 I think OE/ucslugc will be superior to openwrt Jan 20 22:26:39 mwester: busybox-mdev Jan 20 22:26:43 err no Jan 20 22:26:43 kergoth: but from what I know from poking in the sourced, and a couple of my bugfixes went upstream, I have not really seen something "evil" there in terms of portability Jan 20 22:26:43 neat. Jan 20 22:26:49 kergoth: does windows support lock files? Jan 20 22:26:53 the one I booted still used udev Jan 20 22:26:57 or rather, does cygwin support it? Jan 20 22:27:04 but the next image will have mdev Jan 20 22:27:25 cygwin should, not sure though. i know windows does, but its a different mechanism. from python you have to use the win32api stuff, or implement it with directories or so Jan 20 22:27:36 looked into it because lockfiles are one of the areas where bitbake isn't portable to windows either Jan 20 22:27:43 mwester: I was thinking of turning off some of unused eglibc features too Jan 20 22:28:02 mwester: now we can turn on/off eglibc stuff too in OE :) Jan 20 22:28:02 another area, at least for bitbake, was forking Jan 20 22:28:12 can't just fork and not exec in windows Jan 20 22:28:29 kergoth: was, opkg uses system() calls for the postinst and so on hooks, but I think cygwin supports that Jan 20 22:28:34 s/was/well/ Jan 20 22:28:35 yeah, that should be fine Jan 20 22:28:39 For ucslugc that would be good; i'd be concerned about that for SlugOS since users may wish to turnup to an external HDD and add the full set of packages -- hate to have to switch LIBC in order to do that. Jan 20 22:28:40 dont they have clone2 syscall in windoze :) Jan 20 22:28:44 will give it a try and see how it goes Jan 20 22:28:59 mwester: hmm ok Jan 20 22:29:07 mwester: then we can leave it so Jan 20 22:29:40 mwester: infact I tried the same tried to get mediatomb working on nslu :) Jan 20 22:30:24 Well, a review of the features would be useful -- I haven't looked to see if all of them are even applicable. Some features may never be useful on a machine with only 32 or 64 MB of RAM. Jan 20 22:30:40 Mediatomb works, but transcoding isn't so good! Jan 20 22:30:52 yeah :) Jan 20 22:31:09 transcoding needs a powerful cpu Jan 20 22:31:31 may be a cortex-a8 with neon wont be bad Jan 20 22:34:17 03Michael 'Mickey' Lauer  07org.openembedded.dev * r6db340758e 10openembedded.git/recipes/freesmartphone/dbus-hlid_git.bb: dbus-hlid: bump Jan 20 22:36:36 ok now I guess is the moment where I have to be ashamed that I did not update the mediatomb packages to 0.12.1 ;) Jan 20 22:36:53 but since I got that car I more or less stopped coding in my spare time :P Jan 20 23:06:06 sounds like we take a trip to austria and flatten your tires then Jan 20 23:08:19 lol Jan 20 23:08:40 Crofton: I wouldn't :) I'm ready to kill for that car :> Jan 20 23:10:03 a soviet UAZ 469 from the year 1976 :> I did a three month non-stop restoration madness session in 2009, but still finding plenty of things to tune and improve Jan 20 23:10:35 I think I have not coded a line on mediatomb since I got it :P Jan 20 23:23:41 pain pain pain Jan 20 23:25:36 ? :) Jan 20 23:25:46 suffering from Thai Boxing class Jan 20 23:26:26 wow Jan 20 23:26:27 :) Jan 20 23:28:10 first week back in proper training after xmas/new year break and muscles complain :-) Jan 20 23:28:27 I should force myself to do some sports too :P Jan 20 23:29:53 I thought Thai Boxing would help with eV decision making meetings :-) Jan 20 23:31:04 lol :) Jan 21 00:21:10 rofl Jan 21 00:25:17 NOTE: multiple providers are available for runtime libstdc++-dev (gcc, external-toolchain-csl, external-toolchain-generic, external-toolchain) Jan 21 00:25:20 NOTE: consider defining a PREFERRED_PROVIDER entry to match libstdc++-dev Jan 21 00:25:28 anyone knows why this happens? Jan 21 00:27:28 Annoying answer Jan 21 00:27:41 It's always been there, it just hasn't been printed due to a bug until semi recently Jan 21 00:28:15 The fix, for a runtime provider is to add P_P_libstdc++-dev = "libstdc++-dev" Jan 21 00:28:18 iirc Jan 21 00:28:28 Some of the others that are printed aren't as obvious Jan 21 00:28:43 humm OK Jan 21 00:28:50 (ie the hotplug one would be P_P_udev = "udev") Jan 21 00:29:22 khem, thanks for fixing up meta-toolchain on uclibc :) Jan 21 00:29:28 kicking off the autobuilder again momentarily Jan 21 00:29:41 Tartarus: it works here. Jan 21 00:29:52 Tartarus: I am just failing to link with C++ apps Jan 21 00:30:27 different problem? heh Jan 21 00:30:31 still doing micro-uclibc? Jan 21 00:31:39 Tartarus: yes, micro-uclib Jan 21 00:31:45 Tartarus: it fails on the sdk itself Jan 21 00:32:19 k Jan 21 00:32:26 i'll see what I see here semi soon Jan 21 00:46:17 Tartarus: seems good now Jan 21 00:46:28 Tartarus: but I still have the warning about -dev Jan 21 00:48:00 maybe i recalled the fix wrong then Jan 21 00:48:40 but in fact it ought to be commited. Jan 21 00:50:58 well, rfc'd and in that case also cover the external toolchain ones Jan 21 00:51:39 didn't work.. otherwise I'd have queued it up ;-) Jan 21 00:52:32 heh Jan 21 01:58:38 03Chris Larson  07master * r088d516e02 10bitbake.git/lib/bb/cache.py: Jan 21 01:58:38 cache: fix docstring usage Jan 21 01:58:38 Use comments instead of docstrings where appropriate in CacheData. Jan 21 01:58:38 Signed-off-by: Chris Larson Jan 21 01:58:43 03Chris Larson  07master * r5bc773d461 10bitbake.git/.gitignore: Jan 21 01:58:43 Add a few items to .gitignore Jan 21 01:58:43 Signed-off-by: Chris Larson **** ENDING LOGGING AT Fri Jan 21 02:59:57 2011