**** BEGIN LOGGING AT Wed Mar 08 10:59:56 2006 Mar 08 11:00:18 good morning Mar 08 11:00:33 hey florian_kc Mar 08 11:00:36 aaargh Mar 08 11:00:56 now I can't build images, since the locale-base package are missing Mar 08 11:01:00 god damnit Mar 08 11:01:21 * koen is very close to reverting that cset Mar 08 11:01:47 mickey|office: so I should also look at the E packages to "correct" the patch I submitted, right? Mar 08 11:02:20 koen: maybe we can try to add latest quemu and check if it build on amd x64 machines Mar 08 11:03:02 florian_kc: qemu builds on x86_64, but it need gcc 3.X Mar 08 11:03:20 Bernardo: which patch are we talking about? Mar 08 11:03:59 XorA: right. for the native packages this is necessary to make them not accidantly pick up foo-config. Mar 08 11:04:01 florian_kc: later than qemu-native_cvs ? Mar 08 11:04:32 mickey|office: the one I sent to the mailing list that I was discussing with XorA , for neon and subversion Mar 08 11:04:48 florian_kc: I apt-get installed qemu and added qemu-native to ASSUME_PROVIDED Mar 08 11:04:50 XorA: although i prefer setting environment variables instead of cluttering the EXTRA_OECONF line Mar 08 11:05:05 XorA: i.e. XINE_CONFIG = "xine-config-native" Mar 08 11:05:14 florian_kc: but this situation is very, very, *very* bad for regular users Mar 08 11:05:43 florian_kc: also bitbake svn is broken (try r384 instead of r390) Mar 08 11:05:47 koen: no i18n is even worse... and this one hits end-users and not developers. Mar 08 11:06:19 florian_kc: ehm, how can it hit end users when developers can't even build it? Mar 08 11:06:53 I have been unable to build images for the last 2 days Mar 08 11:07:12 florian_kc: some users does not care about locale stuff Mar 08 11:07:21 florian_kc: opie does not use them even... Mar 08 11:07:26 koen: Yeah.. there's no known workarounds? Mar 08 11:09:07 koen: normal users with the default setup use 3.x i suppose... at least for me it works pretty fine. Mar 08 11:09:31 florian_kc: what default setup? Mar 08 11:09:53 florian_kc: so OE only works if I use the 'default setup'? Mar 08 11:09:59 florian_kc: debian has 4.0.2 as default gcc Mar 08 11:10:13 FC4 is in same bout, gcc 4.X Mar 08 11:10:19 s/bout/boat/ Mar 08 11:10:26 We have more stuff whoch blows up with 4.x Mar 08 11:10:43 like what? Mar 08 11:10:55 I've been running gpe images built with gcc4 for weeks now Mar 08 11:11:18 as native compiler or crosscompiler? Mar 08 11:11:25 both Mar 08 11:11:58 hrm.. i wonder what was wrong here... Mar 08 11:12:33 the qemu stuff in glibc is broken because: Mar 08 11:12:43 1) you need libsdl-dev on your host Mar 08 11:12:50 2) you need texu2html on your host Mar 08 11:13:00 3) you need a 32bit x86 cpu on your host Mar 08 11:13:14 4) you can't disable it Mar 08 11:13:56 koen: pb_ told you how it is disbled correctly, that problem with the missing packages should be easy to fix with s/DEPENDS/RECOMMENDS Mar 08 11:14:13 * florian_kc would volunteer to do this... Mar 08 11:15:09 florian_kc: If I disable it 'correctly' it will still put locale-base-en stuff in RDEPENDS Mar 08 11:15:12 florian_kc: that'd be cool Mar 08 11:15:23 florian_kc: so do_rootfs fails Mar 08 11:15:35 ergo, you can't disable it Mar 08 11:16:42 /cr Mar 08 11:16:52 ops Mar 08 11:21:02 <_guillermo> can anyone tell me if meta-gpe is still being used? I see that most of the stuff is now in task-gpe Mar 08 11:21:20 _guillermo: meta-gpe == build ALL from task-gpe Mar 08 11:21:34 <_guillermo> ok Mar 08 11:21:47 _guillermo: task-* define stuff, *-image use task-* stuff. meta-* builds ALL task-* stuff Mar 08 11:22:06 <_guillermo> but it was not always like that, was it? Mar 08 11:22:10 yep Mar 08 11:22:23 it was changed with DEPENDS -> RDEPENDS cleanup Mar 08 11:22:50 <_guillermo> ok, I knew the code was different Mar 08 11:24:49 hey tim! Mar 08 11:25:37 mickey|office: back in Australia now! Mar 08 11:26:11 hey mithro Mar 08 11:26:17 hey Tim Mar 08 11:26:27 mithro: excellent. Mar 08 11:26:41 mithro: how long will you stay there? Mar 08 11:26:41 ~lart nelson Mar 08 11:26:42 * ibot pulls out a ClueBat (tm) and thwaps nelson Mar 08 11:27:05 mickey|office: untill i pay of my debt and finish uni Mar 08 11:27:49 03koen 07org.oe.dev * r24080ea4... 10/packages/linux/ep93xx-kernel_2.6.15+2.6.16-rc5.bb: ep93xx-kernel: fix some PV related bugs Mar 08 11:27:58 florian_kc: *nod* Mar 08 11:28:34 koen: what's the dependency problem in detail? the glibc i18n package depends on the locale-base, right? Mar 08 11:28:55 morning all Mar 08 11:29:14 mickey|office: i just answered :-) in fact i expected nelson to comment, but he is the one who should better stay quiet... Mar 08 11:29:20 florian_kc: yes, and when disabling the qemu stuff it doesn't seem to get built, but it still is in RDEPENDS Mar 08 11:29:22 hi do13_ Mar 08 11:29:33 hey do13_ Mar 08 11:29:33 koen: okay Mar 08 11:29:40 hi florian_kc, koen Mar 08 11:29:54 hey Dirk Mar 08 11:30:04 hey Liam Mar 08 11:30:40 hey zecke Mar 08 11:34:00 hi Dirk Mar 08 11:34:13 hey hrw|work Mar 08 11:34:36 zecke: were you asking about buildbot? Mar 08 11:37:10 mickey|lunch: what do you thing about encoding ARCH in the feed uri for angstrom? like http://foo.org/release/2006.9/feed/base/arm/, http://foo.org/release/2006.9/feed/base/sh3/, http://foo.org/release/2006.9/feed/base/armv5te/? Mar 08 11:38:15 koen: feed/ARCH/subfeed/ maybe rather? Mar 08 11:38:43 hrw|work: that's also an option Mar 08 11:39:03 koen: easier to look at when you want only sh3 stuff Mar 08 11:40:00 hi zecke Mar 08 11:41:21 njs: yes Mar 08 11:41:25 koen: is it still broken? Mar 08 11:41:39 koen: bitbake that is Mar 08 11:41:47 zecke: r390 was broken Mar 08 11:41:58 so I reverted to r384 Mar 08 11:42:11 zecke: the monotone buildbot stuff basically just picks a linearization of the graph Mar 08 11:42:29 zecke: success (with qvfb) http://www.do13.de/oz/temp/snapshot.png Mar 08 11:42:52 do13_: it is running on the arm as well? Mar 08 11:43:14 do13_: nice :) Mar 08 11:43:16 zecke: I have to check this. Mar 08 11:44:05 zecke: so basically it keeps a "my current head" pointer, and whenever new stuff arrives that's on the branch and descended from its current head pointer, it greedily advances that pointer Mar 08 11:44:39 zecke: the result is that if there's some temporary divergence, it will arbitrarily pick one side to follow, and then once they are merged it will catch up with the stuff added on the other side Mar 08 11:44:52 zecke:I think it works also on arm, because monday I had a small windows without text, because my tosa lacks the helvetica font. Mar 08 11:45:17 njs: okay Mar 08 11:45:30 zecke: the code is available, or I could also poke Brian Warner about integrating it upstream; I've been letting it slide because we were the only ones using it anyway :-) Mar 08 11:45:44 njs: we use tinderbox3 :} Mar 08 11:45:50 yeah Mar 08 11:46:12 njs: well I want to make more use of it - we have way too many configurations Mar 08 11:46:53 oof, yeah, I can imagine... Mar 08 11:47:02 koen: do you remember what caused bitbake to pull a wrong cvs snapshot? Mar 08 11:48:16 florian_kc: a too old bitbake that doesn't know about SRCDATE Mar 08 11:49:16 zecke: anyway, the code may still be useful for reference, dunno... http://venge.net/monotone/monotone-buildbot-2.tar.bz2 , buildbot/changes/monotone.py Mar 08 11:50:19 koen: hmm... unlikely, the package is using CVSDATE and r384 from svn. Mar 08 11:51:02 ~lart OE to not fallback to CVSDATE... Mar 08 11:51:03 * ibot takes out a cattle prod and gives OE to not fallback to CVSDATE... a good jolt Mar 08 11:53:20 ~praise bitbake -iu Mar 08 11:53:23 All hail bitbake -iu! Mar 08 11:54:11 -u Mar 08 12:05:34 koen: hmm Mar 08 12:05:34 koen: what was broken? Mar 08 12:06:41 zecke: it displays COMPATIBLE_HOST exceptions during parsing and bombs out somewhere in better_exec Mar 08 12:07:35 <_guillermo> diet-x11 doesn't compile, there is some problem patching errordb-keysymdb-path-fix.patch Mar 08 12:08:42 http://bugs.openembedded.org/show_bug.cgi?id=727 Mar 08 12:11:21 hallo Mar 08 12:11:21 hi greentux Mar 08 12:11:28 hey greentux Mar 08 12:11:30 koen: weird, why are you seing parsing errors? :} Mar 08 12:11:55 zecke: because COMPATIBLE_HOST raises an exception Mar 08 12:12:43 a .bb with COMPATIBLE_HOST="x86" while produce debugging spew during parsing when building for an ARM target Mar 08 12:13:00 strange... bitbake crashes when i try to build bootstrap-image at the first package, but I can build them one by one... any ideas? Mar 08 12:13:15 koobla: latest svn? Mar 08 12:13:22 yes Mar 08 12:13:36 koen: ah right Mar 08 12:13:36 koobla: if so, revert to r384 and tell zecke that bitbake is broken Mar 08 12:13:42 koen: what was the error in better_exec? Mar 08 12:14:00 zecke: I suspect that koobla is looking at the error right now Mar 08 12:14:11 and I'm busy generating lcoales using qemu at the moment Mar 08 12:14:11 I get: IndexError: list index out of range Mar 08 12:14:25 and of course a big backtrace :) Mar 08 12:14:27 <_guillermo> koen: but the patch exist: org.openembedded.dev/packages/xlibs/libx11/errordb-keysymdb-path-fix.patch Mar 08 12:14:52 koobla: pastebin? Mar 08 12:16:02 parser still running - post it in a minute Mar 08 12:16:37 those 'error in executing' appear as well, but do no harm Mar 08 12:17:22 gosh humans are too stupid... Mar 08 12:19:25 hmm.. I'll start new .dev build Mar 08 12:20:21 koobla: and please it did not crash... python apps do not crash Mar 08 12:21:02 http://pastebin.com/590630 Mar 08 12:21:18 no, they just don't follow my orders :) Mar 08 12:22:39 i'm reverting to r384, as suggested - should i dump the cache as well? Mar 08 12:25:15 hmm def module-base; is not a good name Mar 08 12:26:03 I'm tempted to revert the better error reporting capabilities... Mar 08 12:26:18 if people can't handle them :( Mar 08 12:27:31 hm. cause? effect? countermeasures? Mar 08 12:27:57 koobla: well if people don't understand ERROR: On Parsing File: foo.bb Mar 08 12:28:15 and say bitbake is crashing people are confused if they see parsing error reports Mar 08 12:28:19 3128 recipes.. nice amount Mar 08 12:28:25 -> do not show parsing errors Mar 08 12:28:39 zecke: add --quiet mode? Mar 08 12:29:37 well SkipPackage is a legitime RuntimeException it should not be reported Mar 08 12:30:04 agreed Mar 08 12:30:35 ok. whats the difference to bitbake -b /media/ppc/oe.org/packages/libtool/libtool-native_1.5.10.bb -- this works pretty well. but I don't want to build every single package by hand :) Mar 08 12:30:57 koobla: the issue is with data_smart.py (cahing) Mar 08 12:31:09 koobla: the method name is not legal (and yes this is a bug) Mar 08 12:31:23 koobla: you can either revert or get my crash test dummy Mar 08 12:31:24 ok. i'll dump the cache again and try with the last released version :) Mar 08 12:32:07 I fear I won't be of much help, as I haven't touched the bitbake code yet Mar 08 12:32:46 koobla: well Microsoft has testers as well, and M$ hopes that they haven't touched their code :) Mar 08 12:33:37 agreed ;) Mar 08 12:33:50 koobla: for current head, turn off the cache :) Mar 08 12:34:38 ahm, how? Mar 08 12:34:54 koobla: see local.conf.sample Mar 08 12:35:03 aah, there -- ok. Mar 08 12:35:11 thought there was a cmdline switch Mar 08 12:36:04 ok, next run started, no cache. Mar 08 12:36:14 SkippedPackages are not reported as runtime errors anymore Mar 08 12:36:30 hmm berlios jut blew away :( Mar 08 12:36:40 zecke: ;) :( Mar 08 12:36:56 koobla: try svn up as well Mar 08 12:37:03 did already Mar 08 12:37:35 MUCH more peaceful :) Mar 08 12:38:01 now I need to find out what is going wrong here Mar 08 12:38:53 hmm it is working here Mar 08 12:39:13 * koobla is away for half an hour -- will report when I'm back Mar 08 12:39:55 and it is not the cache :( Mar 08 12:40:39 heh... Nothing provides runtime dependency diffstat-native -> quilt-native -> ipkg-utils-native Mar 08 12:40:57 hrw|work: do you set an ASSUME_PROVIDED? Mar 08 12:41:19 oz-unstable do it probably Mar 08 12:42:06 bitbake.conf does it as well :( Mar 08 12:43:07 cleaned configs Mar 08 12:44:56 updated to r391 Mar 08 12:45:08 hrw|work: check local.conf Mar 08 12:45:21 ASSUME_PROVIDED = "gcc-crap-version" Mar 08 12:45:34 change the '=' to '+=' Mar 08 12:45:51 koen: did it Mar 08 12:45:59 koen: I removed >50% of site.conf Mar 08 12:48:02 we have now way better debugging of the 'embedded' python functions Mar 08 12:49:10 gosh is parsing slow :( Mar 08 12:50:47 zecke: how is 'your' in deutsch? Mar 08 12:51:41 hrw|work: worse than my english Mar 08 12:52:02 hrw|work: "Dein" Mar 08 12:52:08 mickey|lunch: danke Mar 08 12:52:12 ouch Mar 08 12:52:13 :) Mar 08 12:52:16 mickey|lunch: wb Mar 08 12:52:21 thx Mar 08 12:52:35 mickeyl: magicly sys.exc_info started to work Mar 08 12:53:01 zecke: good to hear Mar 08 12:53:08 apart from the "magical" part :) Mar 08 12:53:55 mickeyl: I think this is the first thing I need to call in my except: Mar 08 12:54:17 mickeyl: I wrote 'My VW' service which we call itnernally 'Mine Volkswagen' (with pseudo-deutch accent) so I wanted to know how 'Your VW' will be ;) Mar 08 12:54:29 hrw|work: ah right :) Mar 08 12:54:40 mickeyl: how was your food? did they serve leffe as well? Mar 08 12:55:14 zecke: no, not quite. it was lousy as usual. i had a Caffe Creme instead. Mar 08 12:55:42 ~lart university cafeteria Mar 08 12:55:43 * ibot rm -rf's university cafeteria Mar 08 12:56:13 mickeyl: how goes OZ build? Mar 08 12:56:51 hrw|work: world build is doing its thing atm. afterwards, i start with the images. Mar 08 12:57:03 i guess the world build will need another day or so Mar 08 12:57:35 mickeyl: did you saw my build-oz.sh script in contrib/openzaurus/ Mar 08 12:57:43 not yet. Mar 08 12:57:44 mickeyl: and feed-split stuff? Mar 08 12:57:45 * mickeyl looks Mar 08 12:58:07 mickeyl: it need multimachine Mar 08 12:58:12 koobla: it is working here?! Mar 08 12:58:34 hrw|work: oh, that's handy. i'll use it after the world build for the individual machines Mar 08 12:59:05 03pH5 07org.oe.dev * rda502db9... 10/packages/xlibs/ (2 files in 2 dirs): libx11, diet-x11: move errordb-keysymdb-path-fix.patch so diet-x11 can find it (closes #727) Mar 08 12:59:13 mickeyl: I used it for 3.5.4-test* builds Mar 08 12:59:30 koobla: libtool was just fetched Mar 08 12:59:59 mickeyl: DEPLOY_DIR_IMAGE = "${DEPLOY_DIR}/images/${DISTRO_VERSION}/${MACHINE}" Mar 08 13:00:17 mickeyl: it was also nice - I do not remember did I add it into oz354.conf or not Mar 08 13:00:45 file-native need updating... Mar 08 13:00:47 hmm so many patches and hacks. Mar 08 13:00:53 hopefully i remember all that stuff Mar 08 13:02:26 ;) Mar 08 13:06:45 * koobla is back Mar 08 13:06:56 hey gints|wrk Mar 08 13:07:03 gints|wrk: how are you? Mar 08 13:08:30 hmm. same error as before. http://pastebin.com/590687 Mar 08 13:09:04 koobla: can you fetch it using wget by hand? Mar 08 13:09:11 sure Mar 08 13:10:47 _guillermo: diet-x11 should build now Mar 08 13:10:53 def module-base;_do_fetch(): Mar 08 13:11:05 that just looks broken Mar 08 13:11:10 <_guillermo> koen: thx Mar 08 13:11:10 hrw|work, which link? Mar 08 13:11:31 koobla: and your OE tree is free of conflicts? Mar 08 13:12:18 its a plain oe.dev tree, I have some board specific packages in an external tree, so yes Mar 08 13:12:23 koobla: I only have the lousy answer: it is working here Mar 08 13:12:38 :) thats why we get paid, right? Mar 08 13:12:52 koobla: SRC_URI of libtool-native Mar 08 13:13:54 hrw|work, don't think thats the prb: http://pastebin.com/590689 -> single package build works just fine Mar 08 13:14:21 zecke, which python version do you use? Mar 08 13:14:40 dunno Mar 08 13:14:54 python -V Mar 08 13:14:56 koobla: I hate to repeat myself so I won't Mar 08 13:15:27 so don't. will dig some other solution then. Mar 08 13:16:07 koobla: recheck your configuration, bitbake is fine now Mar 08 13:16:39 yop - will do so. thanks Mar 08 13:17:21 koobla: ok Mar 08 13:18:43 build.py:exec_func_python Mar 08 13:18:52 so 'func' is already an invalid name Mar 08 13:19:00 pb_: ping Mar 08 13:19:37 koobla: to ease it up go to your bitbake.conf and change include to require Mar 08 13:20:00 ok, will do so Mar 08 13:22:09 XorA: the new evince is working Mar 08 13:22:21 RP: mixer in 2.6.15-r11 is kind of funky Mar 08 13:22:33 RP: exerytime I fast forward in mplayer it turns off headphones Mar 08 13:23:52 hi Mar 08 13:24:07 i have problem to build openldap with oe cross.. Mar 08 13:24:20 ./ucgendat: ./ucgendat: cannot execute binary file Mar 08 13:24:39 XorA: mplayer must be closing the audio device when it does the fast forward. I'm planning to add a timeout after device close to deal with things like this Mar 08 13:24:42 same problem here http://www.openldap.org/lists/openldap-bugs/200601/msg00195.html Mar 08 13:25:00 lrg: why doesn it turn headphone function off though? Mar 08 13:25:44 l1nux: just like any other stupid buildsystem? Mar 08 13:25:53 XorA: powersaving? Mar 08 13:25:55 XorA: the audio driver turns the hp off to save power. mplayer must be calling close() and then open() in between fast forward Mar 08 13:26:06 l1nux: it should not use arm-linuix-gcc for tools Mar 08 13:26:21 XorA: are you using oss emulation ? Mar 08 13:26:25 lrg: yes Mar 08 13:26:47 XorA: ah, I wonder if it does this under native alsa ..... Mar 08 13:27:21 lrg: I will test after this ep has finished Mar 08 13:27:24 zecke, how ? Mar 08 13:27:38 XorA: thanks Mar 08 13:29:10 l1nux: beat the build system? Mar 08 13:29:46 l1nux: look at the code, understand it, fix it, win Mar 08 13:32:29 03koen 07org.oe.dev * r02146a05... 10/packages/gnome/gconf-dbus_svn.bb: gconf-dbus: bump PV up to 2.10 to provide an upgrade path Mar 08 13:32:34 NOTE: package file-native-4.16-r2: task do_build: completed Mar 08 13:33:09 hrw|work: yay. Mar 08 13:33:18 hrw|work: can you take a look at libaal while you're cleaning SRC_URIs? Mar 08 13:33:20 :) Mar 08 13:33:42 mickeyl: I'm building task-bootstrap now and file 4.13 was MIA Mar 08 13:34:28 hi, folks. here is a strange problem with the busybox which I builded by bitbake yesterday. Mar 08 13:35:21 when I do list all file for detailed information, I found the file size is very large . Mar 08 13:35:26 here is the log : http://rafb.net/paste/results/EBAmii23.html Mar 08 13:35:31 any ideas ? Mar 08 13:37:13 leoncamel: no Mar 08 13:37:52 leoncamel: umount this filesystem, fsck it, fsck it again, mount Mar 08 13:38:24 leoncamel: you got fscked filesystem for 80% sure Mar 08 13:40:43 hrw|work, hmm, but I could read and write the file ? should I fsck the file system? Mar 08 13:41:15 leoncamel: I would do it Mar 08 13:42:00 koobla: found anything? Mar 08 13:42:01 hrw|work, I google for it, and somebody said maybe it is a bug of busybox, caused by "stat" system call. Mar 08 13:42:18 03hrw 07org.oe.dev * r92a452f1... 10/packages/file/ (4 files): file: upgraded to 4.16 (4.13 is not fetchable) Mar 08 13:42:44 leoncamel: hm. then hunt for fix Mar 08 13:43:18 hrw|work, hmm, but I can't find any fix for it .. :( Mar 08 13:44:07 no idea then Mar 08 13:44:47 hrw|work, hmm, I am using a iwmmxt -mtune. Is there anybody test it before ? Mar 08 13:45:50 iirc Justin Treon tested it Mar 08 13:46:28 koen, what is his "id" ? Mar 08 13:46:41 leoncamel: I don't think he is on IRC Mar 08 13:46:49 leoncamel: well, if you think it is a busybox bug Mar 08 13:46:57 leoncamel: then look at the implementation? Mar 08 13:47:09 koen, well, how can I contact with he ? Mar 08 13:47:32 leoncamel: I have an older version of busybox (from the summer) and this is working on a PXA270 Mar 08 13:47:49 leoncamel: triton is a company selling boards including PXA270 modules.. Mar 08 13:47:49 leoncamel: send a mail to justin_treon@yahoo.com Mar 08 13:47:57 ah treon :0 Mar 08 13:47:59 gosh Mar 08 13:49:38 thanks . Mar 08 13:52:25 hrw|work: what do think of http://ewi546.ewi.utwente.nl/angstrom/ ? Mar 08 13:52:44 sorry guys, the volume in this room is just too high (physical one) Mar 08 13:54:18 koen: last point looks nice :D Mar 08 13:55:11 koen: what you will use as CMS for Angstrom? Mar 08 13:55:24 koen: drupal/wordpress + wiki? Mar 08 13:55:36 hrw|work: drupal or wordpress + moin probably Mar 08 13:56:12 it will depend on someone having the skills with CMS'es willing to set something up Mar 08 13:57:17 since we all know about hrw's wordpress expertise, i recommend using wordpress :) Mar 08 13:59:53 hey reenoo Mar 08 14:00:05 morning Mar 08 14:00:08 hey mickeyl Mar 08 14:00:36 koen: the glibc-qemu patch doesn't change RDEPENDS in any way Mar 08 14:00:46 koen: so please stop talking nonsense Mar 08 14:01:19 then why does do_rootfs fail on missing locale-base packages when glibc was built? Mar 08 14:04:03 linux-libc-headers-2.6.11.1 - will we update them to more current one? Mar 08 14:04:41 hey guys, warning: cast increases required alignment of target type -> does alignment stands for size of the variable? searched google but didn't get a meaningful answer if anyone could explain or point out good search terms for it, it would be much appreciated Mar 08 14:05:16 awp__: see http://www.handhelds.org/minihowto/porting-software.html Mar 08 14:05:27 actually, it shouldn't stand for size.... too early in the morning for me Mar 08 14:05:30 thanks koen Mar 08 14:09:35 koen: not sure. I'm a bit reluctant to engage with the breakage that current bitbake's dependency handling appears to be these days and can't build image from .dev metadata right now. Mar 08 14:10:03 koen: don't get me wrong. the point is I'd try to get things working again and I don't have the time for that right now. Mar 08 14:10:28 koen: a matter of self-protection schedule wise I guess... Mar 08 14:12:08 speaking of which Mar 08 14:12:36 * reenoo heads off to continue project work... Mar 08 14:19:45 hi tmbinc Mar 08 14:20:22 hi Mar 08 14:29:44 zecke, was afk - still looking. I am cross-testing different bitbake and oe.dev's, but the error stays, so it must be my conf somewhere Mar 08 14:34:47 lrg: behaviour is not present when using alsa output Mar 08 14:36:11 XorA: thanks. It looks like another oss emu situation :-\ Mar 08 14:37:31 lrg: now is there a quick way to turn it on before mplayer runs so I dont have to run alsamixer? Mar 08 14:40:38 XorA: RP created some scripts for changing scenario. What z do you have ? Mar 08 14:41:11 latest zaurusd/devmand should have c7x0 support Mar 08 14:45:04 lrg: a c7x0 Mar 08 14:45:40 lrg: the problem is I set headphone with alsamixer, but as soon as an app closes the device, it stops outputting to headphone at all until I reselect headphone on the mixer Mar 08 14:46:44 XorA: do you mean alsamixer is changing in between apps opening and closing the device ? Mar 08 14:48:09 lrg: alsamixer says headphone output is enabled, no sound comes out, I flick between headphone/mic/headphone and sound output starts again Mar 08 14:49:02 XorA: ah, ok. That may be a bug. :( I'm about to do a new release and will check. Mar 08 14:49:40 lrg: cool Mar 08 14:50:31 hi lrg Mar 08 14:50:42 lrg: i order a new borzoi for u Mar 08 14:50:52 lrg: love the new simpler mixer BTW, great improvement Mar 08 14:51:15 lrg: and you have fixed the bug where pause in alsa in mplayer would make mplayer freeze forever Mar 08 14:51:22 lrg: so thanks for all the work Mar 08 14:51:24 Hmm.. I was silly enough to build opie-image without bitbake in path. Now peek opie-image PATH gives me unpleasant bitbakeless path. What am I to remove now? (except for the whole tmp? :)) Mar 08 14:52:18 hey greentux Mar 08 14:52:41 greentux: thanks. Mar 08 14:55:33 lrg: how long do you need it approx? 2 month? Mar 08 14:56:15 koen: the diet-x11 patch problem is fixed Mar 08 14:56:18 greentux: yes, 2 months seems fine as we are almost there. Just need to simplyfy mixers and add scenarios. (oh and fix bugs) Mar 08 14:56:33 koen: but now the same problem with another patch showed up Mar 08 14:56:43 fix-utf8-wrong-define.patch Mar 08 14:56:50 the problem is exactly the same Mar 08 14:57:20 koen -> sorry to bother you again, but the faq does not clear out what exactly is alignment, it does say that a X-byte quantity must be aligned to a multiple of X memory boundary (being X 2,4 or 8 in that text) but doesn't say exactly what does alignment means, any other hints? Mar 08 14:57:42 hi there. Mar 08 14:57:49 hallo Mar 08 14:57:58 Where should I look to .bb files on the oe structure tree? Mar 08 14:58:54 you mean the packages? org.openembedded.dev/packages/* I think Mar 08 14:59:17 /*.bb on top of that Mar 08 14:59:45 I'll search for that. Thanks Mar 08 14:59:56 Arthur -> if you are looking for the config statement, mine is: BBFILES := "${HOME}/oe_new/org.openembedded.dev/packages/*/*.bb" Mar 08 15:00:17 ok. Mar 08 15:02:47 koen -> found what alignment means, it's cool, thanks (; Mar 08 15:05:54 NOTE: package gcc-cross-3.4.4: started Mar 08 15:05:59 uf.. glibc passe Mar 08 15:06:00 d Mar 08 15:07:35 03xora 07org.oe.dev * re5ee771a... 10/packages/gimp/gimp_2.3.7.bb: gimp_2.3.7.bb : new version Mar 08 15:16:24 ~lart screen Mar 08 15:16:24 * ibot pours hot grits down the front of screen's pants Mar 08 15:18:08 * chouimat is back. Mar 08 15:18:32 morning Mar 08 15:19:54 file-native is broken Mar 08 15:20:11 katossi: I suspect your OE tree is too old Mar 08 15:20:29 I just pulled from vanille Mar 08 15:20:40 koen: no he is right Mar 08 15:21:30 koen: moving fix-utf8-wrong-define.patch from packages/xlibs/libx11 to packages/xlibs/files/ fiex the problem Mar 08 15:21:42 s/fiex/fixed Mar 08 15:21:48 katossi: that was also fixed a while ago :) Mar 08 15:21:56 XorA: hrw should have fixed it this afternoon Mar 08 15:22:06 koen: I thought you did it with the other patch Mar 08 15:22:17 maybe I should clean my tmp Mar 08 15:22:22 koen: http://pastebin.com/590874 Mar 08 15:22:22 file-native 4.16 built for me - I started from scratch anyway Mar 08 15:22:41 anyone having problems with libxine? I been getting http://pastebin.com/590815 on xine-utils/color.c but can't figure out what is wrong, the code seems legal.... Mar 08 15:23:05 koen, hrw: file-native will only build if your host version of file is sufficiently recent relative to the version of file-native Mar 08 15:23:46 my file is 4.16 Mar 08 15:24:04 re Mar 08 15:24:25 mine is 4.12 Mar 08 15:24:30 I gues that's too old Mar 08 15:24:40 katossi: did you look at my error, is it the same as yours? Mar 08 15:24:47 XorA: it is Mar 08 15:25:00 XorA: ah, heh. and file-native breaks. sounds like it has grown new oddities... Mar 08 15:25:12 s/breaks./breaks?/ Mar 08 15:25:23 03rw 07org.oe.dev * r856af962... 10/packages/glibc/glibc-package.bbclass: (log message trimmed) Mar 08 15:25:27 glibc: make use of qemu conditional on ENABLE_BINARY_LOCALE_GENERATION which defaults to disabled. reinstate BINARY_LOCALE_ARCHES and make it a list of regular expressions. Mar 08 15:25:34 - against contrary belief BINARY_LOCALE_ARCHES = "" did disable the new code. Mar 08 15:25:39 - zapping that variable in glibc-package.bbclass as the changeset before Mar 08 15:25:41 this did is not an option since we need to keep track of which arches Mar 08 15:25:43 are supposed to have this enabled. Mar 08 15:25:45 - also, rename USE_BINARY_LOCALE to GLIBC_INTERNAL_USE_BINARY_LOCALE to Mar 08 15:25:51 * obergix[work] is away: I'm busy Mar 08 15:25:52 reenoo: argh.. Mar 08 15:25:55 ~curse me Mar 08 15:25:58 May you be reincarnated as a Windows XP administrator, hrw|work ! Mar 08 15:26:42 * reenoo helps CIA with the final sentence Mar 08 15:26:50 hrw|work: what you forget? Mar 08 15:26:52 can someone revert that changeset then or resolve it better? 4.13 is not fetchable from upstream Mar 08 15:27:07 ~lart javascript Mar 08 15:27:08 * ibot plops javascript into a giant vat of herring Mar 08 15:27:35 - also, rename USE_BINARY_LOCALE to GLIBC_INTERNAL_USE_BINARY_LOCALE to prevent people from shooting themselves in the foot. DONT MESS WITH THAT VARIABLE. setting USE_BINARY_LOCALE = "0" will _enable_ the new code since "0" (being a string) evaluates to True in Python. Mar 08 15:28:11 stupid python Mar 08 15:32:02 with what bitbake parameter can i only see the compile warnings .. Mar 08 15:32:11 -D -D -D iis too heavy Mar 08 15:32:28 reenoo|afk: does ENABLE_BINARY_LOCALE_GENERATION also disable the locale-gen postinsts? Mar 08 15:32:45 rob_w|mis: look into WORKDIR/temp/ of package Mar 08 15:35:25 hrw|work: koen katossi file problem is solved, remove file from staging/${HOST_ARCH}/bin Mar 08 15:37:31 XorA: so do_compile_prepend { rm ${STAGING_BINDIR}/file} is solution? Mar 08 15:37:51 hrw|work: its not sweet, but would work Mar 08 15:38:37 hrw|work: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=7&logfile=20060308163749.log Mar 08 15:38:42 hrw|work: problem is it picks up the old file, which cant parse the new magic file, as reenoo|afk said, but its the old -native version, not the host one thats causing the problem Mar 08 15:38:46 hrw|work: your new file-native makes all nslu2 builds fail Mar 08 15:39:14 shit Mar 08 15:40:06 koen: yes. that's the point behind the new code after all. Mar 08 15:40:15 hmmm Mar 08 15:40:26 I guess I have some stale stuff in the feed Mar 08 15:40:29 koen: assuming you're referring to localdef Mar 08 15:40:40 localedef even Mar 08 15:40:47 after I did 'upkg upgrade' my z has been busy generating locales for the past 2 hours Mar 08 15:40:51 (swap on sd) Mar 08 15:40:51 hrw|work: http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenEmbeddedBuilds Mar 08 15:40:57 pick any of the error reports Mar 08 15:41:33 zecke: thats the bug we just discussing Mar 08 15:42:43 03koen 07org.oe.dev * r7d746d64... 10/conf/distro/angstrom.conf: angstrom: ENABLE_BINARY_LOCALE_GENERATION ?= "1" Mar 08 15:42:46 okay Mar 08 15:43:09 zecke: does tinderbox build clean each time? Mar 08 15:43:25 XorA: that depends Mar 08 15:43:36 XorA: the NSLU2 are clobbered Mar 08 15:44:27 I am thinking the real fix is make the file-native Makefile use src/file as its command and not either the host/old file-native or any other random version of file that happens to be around Mar 08 15:49:52 I love bitbake error reporting... Mar 08 15:50:12 hrw|work: what is the error now? Mar 08 15:50:18 http://pastebin.com/590910 Mar 08 15:51:24 hrw|work: svn up, you shouldn't see the index error anymore Mar 08 15:51:43 hrw|work: maybe it is better to keep thes 'improvements' for me Mar 08 15:52:11 zecke: r391 is too old? Mar 08 15:52:31 ok. r393 now Mar 08 15:53:02 * awp__ hates to post the message again, but... Mar 08 15:53:05 I'm having problems with package libxine-fb-1.0 been getting: http://pastebin.com/590815 (lines are in green on the right side of the statements) on xine-utils/color.c to who to I report it? here on in #xine? the code looks legal to me though S: Mar 08 15:53:24 really strange... not sure what is going on... Mar 08 15:53:44 awp__: I'd say #xine Mar 08 15:54:14 zecke: some of those error messages are usefull Mar 08 15:54:42 awp__: that wouldnt happen to be gcc 4.X you are using? Mar 08 15:54:48 hrw|work: anyway, as I have not the time to revert it - ignore them please Mar 08 15:55:02 hrw|work: lib/bb/utils.py deserves a list with expected exceptions Mar 08 15:55:57 zecke: ok - I'm ignoring them by most of time Mar 08 15:56:12 XorA -> yes, that is it, is the default of the OE .dev branch, isn't it? Mar 08 15:56:54 awp__: I have a feeling under newer C standard that line is no longer legal Mar 08 15:57:39 awp__: the default depends what machine/distro you are using Mar 08 15:58:14 erm... hum... I would ask "why not" but since I bet you are not the one behind these decisions... which standard is this? Mar 08 15:58:52 awp__: you convert a char * to an int * then increment it, who is to tell if there is actually some memory where the pointer now points Mar 08 16:00:52 XorA -> I would say it's the programmers job to assure there is memory where the pointer points now, hum... Mar 08 16:01:09 ../src/file -C -m magic Mar 08 16:01:29 unbroken file-native Mar 08 16:02:21 XorA -> sorry for the not too clever answer, I'm still learning (; Mar 08 16:02:41 XorA -> thanks for the tips, I will search on the new standard, and will downgrade my gcc Mar 08 16:02:50 hrw|work: cool Mar 08 16:03:15 awp__: chances are there are a gcc 4.X patch for xine lib you can google for, to see if my guess is right Mar 08 16:05:49 XorA -> I'll look for it, cheers Mar 08 16:07:10 hrw|work: FuncFailed is not reported anymore Mar 08 16:07:33 hrw|work: and there shouldn't be more errors :} Mar 08 16:08:40 koen, which cross toolchain is the most stable one? Mar 08 16:10:09 leoncamel: 3.4.x is the one that is used by most people here, but I haven't had problems with 4.x Mar 08 16:11:14 does anybody know how the "openembedded/site/armeb-linux" file gets used? Mar 08 16:12:07 HopsNBarley: ./configure --config.cache=site/armeb-linux Mar 08 16:12:17 HopsNBarley: kind of that iirc Mar 08 16:12:22 cool, thanks. Mar 08 16:15:10 I remember time when I had config.cache in GG:etc/ Mar 08 16:15:16 any elegant solution for the native-file problem? Mar 08 16:15:19 man, i am learning alot about build systems (-; Mar 08 16:16:11 katossi: mt pull/merge/update;build Mar 08 16:16:27 hrw|work, this is a great feature. why wasn't i informed? Mar 08 16:17:58 03hrw 07org.oe.dev * rdb1ffce2... 10/packages/file/ (files/native-fix.diff file-native_4.16.bb): file-native 4.16: use just built 'file' to generate magic files - now it should not break with older file-native installed Mar 08 16:18:08 katossi: here it is Mar 08 16:24:22 hrw|work: do you have mmc/sdcard news? Mar 08 16:33:59 http://www.maintech.de/uploads/pics/dayofthetomtom.png Mar 08 16:34:01 cute Mar 08 16:34:16 it just needs support in OE :) Mar 08 16:37:23 greentux: no Mar 08 16:42:19 cu Mar 08 16:46:46 file-native now builds :-) Mar 08 16:51:57 not here Mar 08 16:52:25 and I did a bitbake -cclean file before Mar 08 16:52:26 where get package source (like mplayer + patch) ? Mar 08 16:53:07 katossi: did you get the file-native hrw just checked in? Mar 08 16:53:18 yes I did Mar 08 16:53:52 not found here http://ipkgfind.handhelds.org/details.phtml?package=mplayer&official=&format= :/ Mar 08 16:56:19 any ideas of this one: http://pastebin.com/590999 Mar 08 16:56:29 koobla: did you find the error? Mar 08 16:57:02 NOTE: package glibc-intermediate-2.3.2+cvs20040726: failed: http://pastebin.com/590999 Mar 08 16:57:27 hah... i just managed to reproduce that qemu build problem Mar 08 16:57:52 florian_kc: http://bugs.openembedded.org/show_bug.cgi?id=729 Mar 08 16:58:22 where ask about this ? Mar 08 16:58:39 l1nux: all the patches are in OE Mar 08 16:59:33 koen, where ? Mar 08 16:59:40 koen: interesting... i have a different one Mar 08 17:00:06 l1nux: package/mplayer would be my guess Mar 08 17:01:38 file-native still breaking, do I need any specific file version in the host? Mar 08 17:05:53 'Note that 2 x 2 is _not_ 48, not on this planet at any rate.' Mar 08 17:06:05 RMK mails are getting funnier every day Mar 08 17:09:12 how do I enable the generation of binary locales? Mar 08 17:19:13 03frederic 07org.oe.oz354fam083 * r92a1e1b6... 10/packages/base-files/base-files/simpad/fstab: base-files: add cf and mmc to simpad fstab Mar 08 17:20:05 shouldn't those to into .dev? Mar 08 17:22:00 for those who haven't seen: check the OE mailing list, there is a DoS possible against drop bear Mar 08 17:23:27 if the maintainer of drop bear would be nice enough to update it (if possible), that might be cool. Mar 08 17:24:08 http://www.freelists.org/archives/linux-cirrus/03-2006/msg00012.html <- another candidate for OE ;) Mar 08 17:25:22 koen: hi, i've just moved my build task to another server for a try Mar 08 17:25:52 koen: but this time build process failed during build file-native ... Mar 08 17:26:03 koen: could you plesae help me take a look at http://pastebin.com/591054 Mar 08 17:27:01 file-native still failing here Mar 08 17:28:57 minipanda: hrw should have fixed that an hour ago Mar 08 17:30:00 koen: oh.. thanks, let me sync with latest and see ... Mar 08 17:38:07 minipanda: tell me if after that it works plz Mar 08 17:39:00 katossi: sure... but we need to wait for monotone pull finished , that's painfully slow :-( Mar 08 17:39:19 np Mar 08 17:56:25 good evening Mar 08 17:56:35 hey pH5 Mar 08 17:56:51 hey koen Mar 08 17:58:28 koen: oooops , after sync to lastest , i meet those warnings ... Mar 08 17:58:29 NOTE: Using cache in '/home/celf/netspectrum/celf/OE-20060228_omap5912osk_familiar-unstable_22_20060228/build/tmp/cache' Mar 08 17:58:30 NOTE: Handling BitBake files: \ (0878/3114) [28 %]NOTE: generation of binary locales disabled. this may break i18n! Mar 08 18:07:16 koen: katossi ok, that issue fixed :-) Mar 08 18:30:30 katossi: well, seems you'ved changed TARGET_CC_ARCH = "-march=armv5te -mtune=arm926ej-s" in conf/machine/include/tune-arm926ejs.conf ? Mar 08 18:30:52 s/ katossi / koen , sorry Mar 08 18:32:35 koen: but seems, both -mtune=arm926ej-s and -mtune=arm926ej-s will not work Mar 08 18:34:56 look http://www.rainer-keuchel.de/wince/dirlist.html :P Mar 08 18:35:08 GNUWINCE :P Mar 08 18:54:12 keypad mode is specific to Qtopia Phone, and only enables key presses such as: ...... The applications in opie have not been modified to handle these. Mar 08 18:54:16 :( Mar 08 18:55:49 ya, sorry Mar 08 18:59:34 Opie looked really nice too. Mar 08 19:00:04 Guess I can now investigate GPE, although I expect it's likely to still be the same :( Mar 08 19:00:30 ljp: so QtE filters every key away? Mar 08 19:00:36 ljp: even when it comes over VNC? Mar 08 19:03:29 are you aware that sylpheed is broken? Mar 08 19:03:39 zecke: basically konqe works on the tosa too. http://www.do13.de/oz/temp/konqe.png Mar 08 19:03:54 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=4&logfile=20060308171100.log Mar 08 19:03:56 hmm. hang on, theres more to it than I thought Mar 08 19:05:15 hi ljp Mar 08 19:05:17 do13: awesome Mar 08 19:05:31 zecke: javascript is broken on arm. I guess my configure options are wrong. Mar 08 19:07:37 hey pb_ Mar 08 19:07:55 ooh, gpe-image has completed.. time to try it Mar 08 19:08:41 cya Mar 08 19:10:00 hmm. not much more though. changes in modality/focus when editing Mar 08 19:16:27 hi all ... i got >>>NOTE: generation of binary locales disabled. this may break i18n!<<< whould i update bitbake ??? Mar 08 19:18:09 gremlin[it]: I guess you should always update bitbake, but I don't think it will help with that particular issue. Mar 08 19:19:02 hmm /local/pkg/oe/collie/tmp/staging/i686-linux/bin/edje_cc: Error. :434 token FRAME not one of: RAW COMP LOSSY USER Mar 08 19:19:08 JustinP: does that sound familiar? Mar 08 19:19:13 JustinP: it's the version from branch Mar 08 19:19:39 ooh, GPE bootsplash... Mar 08 19:20:37 hi mickeyl Mar 08 19:20:44 hey pb_! Mar 08 19:22:39 hmm, phone has locked up Mar 08 19:23:02 ashak: probably busy booting up Mar 08 19:23:45 maybe, but the led has stopped flashing and I can't connect to it over the network Mar 08 19:25:30 hi pb_ ... thanks Mar 08 19:25:56 To make xdpyinfo 1.0.1 build against diet-x11, I have to force it to disable xkb support. Does this look sane: http://en.pastebin.ca/44934 ? Mar 08 19:26:59 argh, my eyes... m4 code Mar 08 19:28:16 mickeyl, i'm with ya. i was just wondering why this is still used! Mar 08 19:28:24 mickeyl: you suffer from the same affliction? Mar 08 19:28:33 eyes starting to bleed from looking at m4? Mar 08 19:28:33 mickeyl: ok, I rephrase: "does this look _acceptable_" Mar 08 19:28:35 koen: yeah, it's cronical Mar 08 19:28:50 pH5: if it's correct, it looks right to me Mar 08 19:28:58 read: if it works Mar 08 19:30:03 sleep well, chouimat|nap Mar 08 19:30:33 mickeyl: I hope ... insomnia for 3 weeks now Mar 08 19:30:53 d'oh. that sucks! been to a doctor yet? Mar 08 19:32:55 is there a variable that is filled in with "/path/to/openembedded/packages/package_name/files" ?? Mar 08 19:34:18 ah, nevermind... Mar 08 19:34:51 it would be FILESDIR for those playing along at home (-; Mar 08 19:35:32 iirc its FILESPATH that's actually obeyed. FILESDIR is in that by default Mar 08 19:36:00 thanks for the clarification! Mar 08 19:36:02 np Mar 08 19:36:12 hey kergoth, how are you doing? Mar 08 19:37:32 pretty good. new job is going well, getting things done, pretty productive, good variety of code so i wont go nuts like at ti :) Mar 08 19:37:43 just need to get some cash together and move out of this damne xtended stay hotel and itll be all good Mar 08 19:37:47 you? Mar 08 19:38:02 hey kergoth Mar 08 19:38:31 g'day kergoth Mar 08 19:39:16 hey Mar 08 19:39:51 course i was trying to avoid kernelspace, and here i am backporting a memory management subsystem patch Mar 08 19:39:54 ah well Mar 08 19:40:00 doh, oh well Mar 08 19:40:17 I guess the kernel is okay in small doses Mar 08 19:40:32 indeed Mar 08 19:41:14 koen: hey ... shall i turn off this mtune for build userland for ti5912OSK? Mar 08 19:41:33 koen: i always meet this :"| checking for arm-linux-gcc... arm-linux-gcc -march=armv5te -mtune=arm926ejs Mar 08 19:41:34 | checking for suffix of object files... configure: error: cannot compute suffix of object files: cannot compile Mar 08 19:41:36 kergoth: glad to hear things are going well. i'm pretty busy as time is progressing towards my (self chosen) delivery date of the Ph.D. End of May/Beginning of June is not far away. Otherwise fairly good as well. Mar 08 19:41:38 bbl, dinner Mar 08 19:42:06 minipanda: yeah, you can comment out the include in the machine .conf Mar 08 19:42:11 koen: and i've tried both mtune=arm926ejs and mtune=arm926ej-s, but just the same issue Mar 08 19:42:40 koen: but comment that out will do no harm for userland running, right? Mar 08 19:44:48 hmm, something seems to break when it gets to Configuring locale-base-de-de Mar 08 19:45:31 koen: or in gcc 3.4.4 , we still have this additional mtune, but in gcc 4.x it's vanished? Mar 08 19:47:11 koen: sounds cpu specific option for arm926ejs have minor difference during gcc 3.x to gcc 4.x? Mar 08 19:47:31 sigh - gcc-4.0.2 is pulling in native headers Mar 08 19:50:38 CosmicPenguin: suck Mar 08 19:50:49 minipanda: yes Mar 08 19:56:42 ok... it's dying when /etc/rcS.d/S98configure is being called, that seems to be a script that simply calls ipkg-cl configure... That then tells me it's configuring dbus-1, gdk-pixbuf-loader-gif, gdk.....jpeg, gdk....-png, gdk....-xpm, local-base-de-de, which is where is hangs Mar 08 19:57:15 it doesn't hang, it's just very busy generating locales Mar 08 19:57:37 oh right Mar 08 19:57:47 I hadn'g realised it would take this long Mar 08 19:58:09 koen: well, so could you let me know how can i find the right mtune gcc 4.x ? for a gcc newbie like me :-) Mar 08 19:58:25 minipanda: http://gcc.gnu.org/onlinedocs/gcc-4.0.2/gcc/ARM-Options.html Mar 08 19:58:41 do13: is JavaScript not compilable? Mar 08 19:59:00 it must be seriously hammering the cpu as I can no longer ping the device over usb ethernet. Mar 08 20:00:19 ashak: 100% CPU for a while. will eventually be killed by the OOM killer too. Mar 08 20:01:28 if /tmp is tmpfs, yes, otherwise you have a 50/50 chance it will complete Mar 08 20:01:32 mickey|dinner: grrrr...yes, that is familiar Mar 08 20:01:40 ashak: thus the recent glibc changes that run localedef in qemu at buil time Mar 08 20:01:43 mickey|dinner: it has to do with cpp not working right from what I remember... Mar 08 20:02:02 reenoo: ? Mar 08 20:03:03 isn't it a good idea just not to install locales by default? Mar 08 20:03:21 ashak: is that a question? Mar 08 20:03:46 a question mark :) Mar 08 20:03:48 reenoo: itg was more of a... if that' possible, why aren't we doing it? :) Mar 08 20:04:19 because of stuff like http://bugs.openembedded.org/show_bug.cgi?id=729 Mar 08 20:04:55 koen: but this doesn't work :-( "arm926ej-s" Mar 08 20:04:59 polyonymous: I hadn't realised i'd installed any locales, I just built gpe-image Mar 08 20:05:08 koen: why don't you just fix that? Mar 08 20:05:10 ashak, that's exactly what I was talking about. Mar 08 20:05:31 polyonymous: ah Mar 08 20:05:32 reenoo: my fix would be to remove that line Mar 08 20:05:38 reenoo: which is probably not a good idea Mar 08 20:05:40 ashak, I would prefer *-image a bit less populated. Mar 08 20:05:54 ~stab koen Mar 08 20:05:56 * ibot runs at koen with an origami Swiss Army knife, and inflicts a nasty paper cut. Mar 08 20:06:22 polyonymous: right Mar 08 20:07:02 this is still on the same configure step... still not convinced that it's not locked up :) Mar 08 20:07:28 And I would prefer not to have task-* packages which effectively prevent you from uninstalling tictactoe without force :) Mar 08 20:07:56 mickey|dinner: thought I'd fixed that.... It's quite likely I've fubared the utils stuff.... yeah, I think this probably has to do with the native versions of edje and embryo not getting built correctly.... Mar 08 20:08:01 mickey|dinner: which branch is that? Mar 08 20:08:50 JustinP: .oz Mar 08 20:09:25 And I'd prefer machine/qemuarm in .oz branch ;) Mar 08 20:09:50 koen: hmmm....crap. So this is affecting the "release" builds likely... Mar 08 20:09:54 And the last but not least - I'd prefer not to go to the kitchen to have my tea ;) Mar 08 20:10:05 * JustinP should really uninstall his host e17 to see fi that kills it... Mar 08 20:12:26 my ssh session to the phone just died Mar 08 20:13:08 hrw|gone: I would consider the bitbake logging issue fixed now? Mar 08 20:13:18 all I get in dmesg is ETDEV WATCHDOG: usb0: transmit timed out Mar 08 20:14:11 rwhitby: ping Mar 08 20:14:57 so I guess the next question is, can I stop it trying to configure the locales for now? Mar 08 20:15:41 ashak: sure, don't ask how though Mar 08 20:16:13 zecke: heh, not the response I was hoping for :) Mar 08 20:16:30 ashak, kill S98configure? :) Mar 08 20:17:08 * JustinP unmerges e17 Mar 08 20:17:33 polyonymous: I tried stopping that from running, but gpe didn't seem to start, I assume because something hadn't been configured? Mar 08 20:18:06 ashak, yes, then just forceable uninstall locales before they get configured? Mar 08 20:18:41 or kill the corresponing stuff in ipkg libdir or whatever. Mar 08 20:20:14 ah: /etc/rcS.d/S99login-manager: 118: gpe-dm not found Mar 08 20:20:27 that's probably why gpe isn't starting Mar 08 20:21:06 can you configure packages individually? Mar 08 20:21:28 that's what I was wondering Mar 08 20:21:33 update-alternatives wasn't run for gpe-dm likely Mar 08 20:21:39 errr...hmmm Mar 08 20:21:43 * JustinP rescinds that Mar 08 20:22:09 ashak, ipkg configure packagename ? Mar 08 20:22:27 * ashak tries removing the locale-base packages first. Mar 08 20:22:50 ashak: this looks like a bug in login-manager Mar 08 20:26:03 hmm where could I git push my SHARK tree Mar 08 20:27:40 fear Mar 08 20:27:57 (just one cset in it...) Mar 08 20:28:35 ok.... configure now completes, but I still get the gpe-dm: not found message when starting login-manager Mar 08 20:29:45 how can I do 'for i in range(1,150):' in bash? Mar 08 20:30:59 zecke: use the power of seq Mar 08 20:31:08 (and "for", obviously) Mar 08 20:31:08 for ((i=1 ; i <= 150 ; i++)) IIRC Mar 08 20:31:14 I could be remembering wrong Mar 08 20:31:49 pb_: rock! Mar 08 20:31:49 yeah, or that Mar 08 20:32:57 * johnX never knew about seq Mar 08 20:33:12 for i in `seq 1 250`; do addgroup testgroup$i; done Mar 08 20:33:37 neato Mar 08 20:33:39 what a lot of groups you have Mar 08 20:33:40 hmm, gpe-dm exists, I just ran it. Mar 08 20:33:44 all the better for testing svn, I suppose Mar 08 20:33:48 ldc: ping Mar 08 20:34:08 pb_: What makes you believe that? Mar 08 20:34:42 nothing, I just made it up Mar 08 20:35:00 ashak: the problem is that login-manager doesn't want to start the gpe-dm binary, but the gpe-dm init script that is in /etc/X11/login-managers/gpe-dm. there is a path missing somewhere Mar 08 20:35:01 <_law_> ashak: perhabs an path error in login-manager script Mar 08 20:35:02 pb_: didn't I promise to test svn with lot of groups many moons ago? Mar 08 20:35:07 you did Mar 08 20:35:32 _law_: yes I think so... just looking Mar 08 20:35:38 pb_: and you thought I didn't do it, man I'm an opaque person Mar 08 20:36:09 if it wouldn't be so obvious pb_ could guess the color of my shorts :} Mar 08 20:37:30 file-native_4.16 is still failing for me even with whatever patch hrw added a bit ago? Mar 08 20:37:48 (errr...that "?" shouldn't have been there) Mar 08 20:38:41 johnX: for me too Mar 08 20:40:48 hmm I think I broke my ubuntu :} Mar 08 20:41:08 zecke: That sounds painful Mar 08 20:41:10 Thats gotta hurt Mar 08 20:41:36 * reenoo rotfl Mar 08 20:42:17 bah, why does everything pop up a screen to calibrate the damn touchscreen ;) Mar 08 20:42:40 ashak: at a guess, because your damn touchscreen is uncalibrated Mar 08 20:42:52 pb_: I don't have a touchscreen :( Mar 08 20:42:59 pb_: psst that damn thing doesn't have one :) Mar 08 20:43:00 .) Mar 08 20:43:35 zecke: damn Mar 08 20:43:49 http://oe.pastebin.com/591443 Mar 08 20:43:57 :P Mar 08 20:45:19 :( I should have not updatet my OE tree Mar 08 20:53:38 Hey does anyone know how to compile a pyrex program for the zaurus without have a gcc toolchain on the zaurus? Mar 08 20:56:47 not me Mar 08 20:57:07 dtx: you could run it on the host Mar 08 20:57:20 and make it use your cross compiler instead of your normal one Mar 08 20:57:45 where does openembedded place the cross compiler Mar 08 20:57:56 guillermo_, I cheated by using my distro provided version of file, I'll see if it causes my whole system to explode Mar 08 20:57:57 dtx: wherever you ask it to Mar 08 20:58:10 dtx: normally tmp/cross/ Mar 08 20:58:14 ok Mar 08 20:58:18 that would be it then Mar 08 20:58:20 thanx Mar 08 20:58:34 well,does it see all the libs that would be installed and everything Mar 08 20:58:36 ? Mar 08 20:58:38 dtx: tmp/cross/arm-linux/bin could have gcc,g++ etc Mar 08 20:58:57 dtx: no, libs are in tmp/staging/arm-linux/lib Mar 08 20:59:06 dtx: you will need to set that path as well Mar 08 20:59:20 dtx: maybe you are lucky and pyrex honors CC, CFLAGS, LDFLAGS Mar 08 20:59:23 zecke: 'bitbake devshell' Mar 08 20:59:27 I'm assuming the headers will be in staging as well Mar 08 21:00:00 dtx: sure Mar 08 21:00:37 k Mar 08 21:00:38 cool Mar 08 21:14:43 pb_: My svntest user can create a svn repo, commit, pull to and from it Mar 08 21:14:52 pb_: I cloned ~600 kde changes into it Mar 08 21:15:46 zecke: pong Mar 08 21:16:27 rwhitby: we might need a way to automatically call clean if a packages fails Mar 08 21:16:53 zecke: rock Mar 08 21:16:54 rwhitby: once file-native would be fixed it would start building again Mar 08 21:17:01 Looks like /etc/init.d/login-manager is doign somethign wrong, because it looks like it's setting SELECTED_LM to ' gpe-dm', rather than 'gpe-dm', not sure why yet. Mar 08 21:17:18 now we would need to bump the PR to make all automatic builds work again Mar 08 21:17:38 pb_: hehe, the result is as expected. If svn is using groups somewhere Mar 08 21:17:44 pb_: libapache is doing it Mar 08 21:18:39 We've just bumped the PR in the past to do things like that. Bumping the PR will work for nslu2-linux autobuilds. Mar 08 21:18:40 ashak: I think SELECTED_LM should be /etc/X11/login-managers/gpe-dm Mar 08 21:19:12 what is this login-manager thing anyway> Mar 08 21:19:17 s/>/?/ Mar 08 21:19:40 pH5: yeah, it should, but /etc/X11/login-managers/ seems to be substituted by a spam character. Mar 08 21:19:46 pb_: some strange kind of dm selector that doesn't work like it should Mar 08 21:20:01 sounds weird. is it in the gpe-images? Mar 08 21:20:07 and, if so, how did it get there? Mar 08 21:20:29 pb_: I built gpe-image, it appears to be how gpe-dm is started. Mar 08 21:20:40 hrm Mar 08 21:21:00 well, I guess the question is not so much how it got there, as how we get rid of it again. Mar 08 21:21:01 heh Mar 08 21:21:14 heh Mar 08 21:21:24 gpe-dm used to be perfectly capable of starting itself. Mar 08 21:21:36 and, I imagine, left to its own devices it still is capable of doing that. Mar 08 21:22:13 03freyther 07org.oe.dev * r2d5ad291... 10/classes/tinderclient.bbclass: Mar 08 21:22:17 classes/tinderclient.bbclass: Send the name Mar 08 21:22:19 Send the name of the to be build packages. Now Mar 08 21:22:21 we can see what was intended to be build from Mar 08 21:22:23 within the log Mar 08 21:24:05 good nite guys Mar 08 21:24:11 'night zecke Mar 08 21:24:24 pb_: looks like /etc/X11/login-managers/gpe-dm could simply be used as an init script for gpe-dm Mar 08 21:24:30 I suspect the login manager business was added to stop various managers conflicting (like gpe-dm conflicting with its e counterpart) Mar 08 21:25:41 'night zecke Mar 08 21:25:49 Well, whatever... it currently seems to be broken :) Mar 08 21:27:26 RP: yeah, the e folks are welcome to do whatever they like in their own images. But I would prefer it if they left the gpe-images alone. Mar 08 21:27:53 or, at least, solicited comments from gpe@handhelds.org before making that kind of change. Mar 08 21:28:11 pb_: I think the idea was to make it work with either e, gpe or both installed Mar 08 21:28:30 pb_: I suggest you point this out to whoever made that change... Mar 08 21:28:44 that fell on deaf ears Mar 08 21:28:52 reenoo_ already tried that Mar 08 21:29:43 I know nobody ever answered my request about rewriting meta-gpe. I just assumed that meant nobody cared and went ahead and made the changes... Mar 08 21:29:54 yeah, that's fine Mar 08 21:30:05 if you ask the question and get no response, it is reasonable to assume that nobody minds. Mar 08 21:32:18 reenoo_: actually removing the cp -a libgcc_s line does make it work here, but I'll leave that as a local diff :) Mar 08 21:33:59 * koen sits back and waits for the locales to be generated Mar 08 21:39:30 Wow, the joystick on this device makes the cursor on the New User Details screen for Opie move between text fields. Mar 08 21:43:45 finally file-native worked, I had to get rid of all my tmp Mar 08 21:44:27 guillermo_: wow Mar 08 21:44:40 I wonder why that was Mar 08 21:44:52 <_law_> n8 all Mar 08 21:45:54 hrm Mar 08 21:46:09 * CosmicPenguin slaps gcov around Mar 08 21:46:25 pb_: shouldn't it be "GPE Display Manager" instead of "GPE Desktop Manager"? Mar 08 21:46:41 (DESCRIPTION of gpe-dm) Mar 08 21:47:19 reenoo_: yeah, I guess Mar 08 21:47:31 it isn't really a display manager, but it definitely isn't a desktop manager either Mar 08 21:47:32 heh Mar 08 21:47:39 yeah Mar 08 21:47:52 I'll change it to diplay manager for now Mar 08 21:48:05 can't think of any better description atm :/ Mar 08 21:48:09 ok, cool Mar 08 21:53:07 03rw 07org.oe.dev * r2e402ac2... 10/packages/gpe-dm/ (files/multi-lm.patch gpe-dm_0.48.bb): Mar 08 21:53:09 gpe-dm: revert bf4c4f3acd8633a080b0a3d75420ff47be99e302. slightly improve DESCRIPTION. Mar 08 21:53:11 - please consult gpe@hh.org before making changes like this. Mar 08 21:54:35 RP: what is the problem with the current meta-gpe? Mar 08 22:01:00 03rw 07org.oe.dev * r8ed7bf35... 10/packages/glibc/glibc-package.bbclass: glibc: move some code around to avoid warnings while parsing. hopefully fix #729. Mar 08 22:01:09 koen: may want to give that a try Mar 08 22:01:22 guillermo_: There isn't one I'm aware of... Mar 08 22:02:00 RP: then why do you want to rewrite it? Mar 08 22:02:24 he already rewrote it :) Mar 08 22:02:26 guillermo_: There is no problem now I've rewritten it... Mar 08 22:02:44 RP: is it already commited? Mar 08 22:02:48 guillermo_: yes Mar 08 22:03:33 reenoo_: will test it, thanks Mar 08 22:07:09 zh is the locale for china right? Mar 08 22:07:50 Would something like this be desireable in OE? http://www.rpsys.net/openzaurus/temp/oe-sanity.patch Mar 08 22:07:54 localegen needed some extra sweat on that Mar 08 22:09:34 RP: cute Mar 08 22:12:38 CosmicPenguin, pong Mar 08 22:13:00 CosmicPenguin, for a few minutes more Mar 08 22:13:23 ldc: do you know of any kernel code for the WM8973? Mar 08 22:14:48 CosmicPenguin, nope? Mar 08 22:15:05 CosmicPenguin: It's similar to the WM8750 driver in ASoC Mar 08 22:15:35 CosmicPenguin: do you need a driver for au12xx ? Mar 08 22:17:12 no - for another board Mar 08 22:19:11 CosmicPenguin, take a look at the code for the wm8971 Mar 08 22:21:01 but now it's time to get some sleep.... Mar 08 22:24:50 koen|sleep: NOTE: build 200603100450: completed Mar 08 22:24:51 Build statistics: Mar 08 22:24:51 Attempted builds: 55 Mar 08 22:29:37 night all Mar 08 22:31:14 is it true that gtk 2.8 is slower that 2.6 ? Mar 08 22:32:39 katossi: With floating point emulation, yes Mar 08 22:34:02 RP: no other way with akita I guess Mar 08 22:34:27 katossi: correct Mar 08 22:38:50 how can I tell glibc to only generate the locales that I need? Mar 08 22:39:01 right now, you can't. Mar 08 22:39:08 patches welcome :-} Mar 08 22:39:48 Wow - that seems to be the popular FAQ these days Mar 08 22:39:49 unless you plan on building glibc frequently, though, I wouldn't worry about it too much. Mar 08 22:40:06 CosmicPenguin: It's a conspiracy! Mar 08 22:40:46 Lots of people, particuarly those new to OE tend to wipe tmp a lot :-/ Mar 08 22:41:10 heh Mar 08 22:41:15 Let's hope they learn from their mistakes. Mar 08 22:43:38 well...it's impossible to know if things will build right from scratch unless you wipe tmp Mar 08 22:44:21 and if you're working on new packages you might end up with bad stuff in staging and want to wipe tmp to make sure it's all ok Mar 08 22:44:50 I generally keep a copy of my tmp up to gcc-cross around so that I don't have to rebuild glibc and such when I want to wipe Mar 08 22:45:13 (I tend to wipe tmp whenever I update EFL/E17 as I want to make sure that it works from scratch...) Mar 08 22:45:56 JustinP, isn't it a one night build, anyway? Mar 08 22:46:29 if you disable binary locales, yes Mar 08 22:46:51 Everything revolves around binary locales ;-) Mar 08 22:46:57 but it's wwasted time....glibc doesn't need to be rebuilt most of the time Mar 08 22:47:04 heh Mar 08 22:47:11 on *my* machine at least ;-) Mar 08 22:47:13 I rebuilt glibc a lot, lately. Tried 2.3.6, then HEAD, back to 2.3.5+cvs, and now rebuilt to test host locale generation. Mar 08 22:47:49 If I had a 3Ghz machine I'd be ok....but all I have right now is a lowly 700Mhz Celeron Mar 08 22:50:48 I have a 2200+ athlon and still complain :) Mar 09 00:12:44 this is not possible, diet-x11 has problems again with patches Mar 09 00:20:08 I just did a build fron scratch and diet-x11 definately worked... Mar 09 00:22:23 I had to do the same thing with xim.patch that with the other ones Mar 09 00:22:45 move it to xlibs/files/ Mar 09 00:22:50 and then it works Mar 09 00:38:33 perhaps you're building the Xorg7 version of x? Mar 09 00:43:36 don't know Mar 09 00:47:36 maybe it's required by dillo or gpsdrive Mar 09 01:35:30 katossi: please make sure your OE tree is up to date and file bugs if you still encounter any issues. Mar 09 01:35:33 'night all Mar 09 01:36:14 reenoo: night and yes it's most of the times updated Mar 09 03:09:04 03mwester 07org.oe.dev * r2621e7a7... 10/conf/distro/unslung.conf: Unslung: unslung.conf, DISTRO_VERSION is now 6.7-beta Mar 09 03:48:27 i'm brainstorming about what to do with embedded meta-distro Mar 09 03:52:07 hmm, I build toolchains with linux-2.6.11-glibc-hdr, and the kernel is linux-2.4.20. Is it right ? Mar 09 03:55:07 not sure Mar 09 03:55:14 depends on your distro setting and kernel version Mar 09 03:58:53 hmm, I want setup a OE environment for a linux-2.4 kernel. but the package glibc-linux-header only provided linux-2.6.8-hdr .. Mar 09 03:59:12 hmm, what shall I do to setup that environment ? Mar 09 04:00:45 ? Mar 09 04:00:55 leoncamel: OE will build the 2.4 kernel for you fine Mar 09 04:01:12 leoncamel: just make sure to install the gcc 2.95 arm package Mar 09 04:02:31 JustinP, I want to build a gcc-3.[34] and binuitls-2.15(or more) with "glibc-linux-header 2.4". Mar 09 04:02:56 again....? Mar 09 04:03:09 just set up OE, tell it you want kernel 2.4, and build... Mar 09 04:03:25 JustinP, but, I search the package directory, and the linux-libc-header only provided linux-libc-headers-2.6.8.1.bb.. Mar 09 04:03:44 just BUILD IT! Mar 09 04:03:46 it works Mar 09 04:03:55 JustinP, hmm, you mean the "KERNEL="2.4"" in local.conf ? Mar 09 04:03:58 yes Mar 09 04:04:20 ok. thanks . Mar 09 04:05:39 hmm, not really. Mar 09 04:05:47 what is good sftp client for zaurus? command line or opie.... Mar 09 04:06:01 JustinP, I already set that option for linux-2.4. Mar 09 04:07:57 JustinP, but I find the "iwmmxt-linux/linux-libc-headers-2.6.11.1-r1".. Mar 09 04:08:12 Hmm, wait, I am confused. Mar 09 04:08:37 do we need linux-headers while building cross-gcc ? Mar 09 04:10:30 just let it build Mar 09 04:10:53 stop looking, OE knows better than you what it needs Mar 09 04:14:48 leoncamel: you're welcome to look into and mess with the metadata, of course, but the first step is to just use it Mar 09 04:16:03 is there a pre-existing wiki page for developer names and who everyone is? Mar 09 04:16:56 not AFAIK Mar 09 04:22:29 Guys and gals, if you have commit access please add yourself http://oe.handhelds.org/cgi-bin/moin.cgi/People Mar 09 04:23:15 it's the next best thing to having a /~username/ to put thoughts and ideas related to OE Mar 09 04:30:13 * JustinP edits Mar 09 04:40:24 jnc, most devs have thier own sites/pages Mar 09 04:40:46 including links to donate/support the work they do Mar 09 04:46:33 hmm Mar 09 04:46:37 i'm just broke i guess Mar 09 04:49:12 lol Mar 09 04:52:00 http://www.vanille.de/projects/openembedded.html Mar 09 04:52:01 http://www.hrw.one.pl/articles/donations/ Mar 09 04:52:03 etc Mar 09 04:53:22 interesting .... Mar 09 04:53:35 http://pgpkeys.mit.edu:11371/pks/lookup?op=vindex&search=0x093DF623 Mar 09 04:53:44 the handhelds.org public key server ... Mar 09 06:15:50 hi, folks . how can I build cross-gcc with option "disable-multilib" ? Mar 09 06:17:33 add it to the EXTRA_OECONF Mar 09 06:18:56 JustinP, hmm, where should I add this option into ? Mar 09 06:19:17 copy the gcc-cross bb file that you want, then edit it and build that one Mar 09 06:20:24 JustinP, ok. Mar 09 06:22:51 JustinP, hmm, I am not sure. could you tell me more about it ? Mar 09 06:23:20 EXTRA_OECONF Mar 09 06:23:24 I'm heading to bed now Mar 09 06:24:17 JustinP, hmm, it seems the gcc-cross-initial_3.4.4.bb already use "--disable-multilib" now . Mar 09 06:24:53 JustinP, but the cross-gcc I build, seems use "--enable-multilib" .. :( Mar 09 06:25:26 JustinP, OK. :) here is moon now .. Mar 09 06:30:40 Hmm, very strange here : I find two gcc-cross in my work directory : tmp/work/iwmmxt-linux/gcc-cross-3.4.4-r3 and tmp/work/iwmmxt-linux/gcc-cross-initial-3.4.4-r3. Mar 09 06:31:13 but the arm-linux-gcc in tmp/cross is gcc-cross-3.4.4-r3. Mar 09 06:31:54 why ? and how can I install gcc-cross-initial-3.4.4-r3 into tmp/cross/arm-linux ? Mar 09 06:32:22 arm-linux is *ARM-LINUX* binaries Mar 09 06:32:30 not for running on the host Mar 09 06:32:38 * JustinP really goes to be dnow Mar 09 06:35:27 JustinP, I mean how can I change my current toolchain to "gcc-cross-initial" instead of "gcc-cross" ? Mar 09 06:37:41 You don't want to Mar 09 06:37:49 gcc-cross-initial is just for creating gcc-cross Mar 09 06:37:57 gcc-cross is th eone that actually builds stuff Mar 09 06:42:09 JustinP, I know it is very late for you. But could you mind have a look at this : http://rafb.net/paste/results/kQ6SHv16.html Mar 09 06:43:33 the two cross-gcc got different specs. Mar 09 07:01:50 <_law_> hi Mar 09 07:02:16 <_law_> get some compile errors if i try to build file-native http://pastebin.com/592252 Mar 09 07:07:08 _law_, could you paste more message and your local.conf ? Mar 09 07:10:09 <_law_> leoncamel, http://pastebin.com/592260 Mar 09 07:11:16 <_law_> leoncamel, http://pastebin.com/592263 Mar 09 07:14:07 no idea. try "rebuild" Mar 09 07:56:32 morning Mar 09 08:18:12 good morning all Mar 09 08:18:56 'mornin Mar 09 08:25:54 hi all! Mar 09 08:27:45 Anyone familiar with meta-sdk metadata Mar 09 08:42:39 morning all Mar 09 08:42:53 morning Mar 09 08:44:36 Creepy Mar 09 08:44:56 it seems that I'm starting to understand how the metadata works Mar 09 08:45:17 ^^ you do better than i do... Mar 09 08:45:34 Maybe it was just pure luck Mar 09 08:46:02 I managed to compile meta-sdk target with DISTRO as "familiar-0.8.3" and machine "x86" Mar 09 08:46:47 Too bad that familiar is a kind of a curse word these days Mar 09 08:46:57 morning Mar 09 08:48:47 does OE support EABI toolchains ? Mar 09 08:50:42 leoncamel: If someone adds the right packages, it can support most toolchains Mar 09 08:50:45 morning all Mar 09 08:59:59 this is so weird: I have a custom gpe image that compiles fine, but if I add just one more package, like for example jamvm to the task-gpe, then it always tries to compile samba, and not even the lastest version but the 3.1.14 Mar 09 09:08:55 RP, I want use my own compiler. How ? Mar 09 09:09:01 leoncamel: koen is working on EABI support Mar 09 09:09:54 katossi: I cant remeber what, but there is a hard coded depend somewhere in OE for samba 3.1.14, I fell over that one a couple of says ago, but I also temporarilly killed my Z so Ive forgotten where Mar 09 09:10:29 It seem something wrong with the toolchain which I build with OE. I am using gcc-cross-3.4.4. Mar 09 09:12:34 here is the log : http://rafb.net/paste/results/I1u23266.html Mar 09 09:12:52 RP: did you see discussion on mixer funkiness yesterday? Mar 09 09:20:50 morning Mar 09 09:21:39 if I want to be tinderclient then I need to build from scratch or can I became one in any moment? Mar 09 09:22:16 03koen 07org.oe.dev * r8d0568a4... 10/packages/linux/ep93xx-kernel/defconfig: ep93xx kernel: update defconfig to match 2.6.16-rc5-derevo14 Mar 09 09:22:30 hrw|work: you can become one any moment Mar 09 09:22:55 thx Mar 09 09:26:02 <_law_> hi hrw Mar 09 09:26:23 argh... Mar 09 09:26:39 kopete cannot connect to jabber.org but any other client can ;( Mar 09 09:29:16 <_law_> hrw|work, you have updated file-native ? Mar 09 09:29:49 <_law_> hrw|work, i get the following error, any ideas? http://pastebin.com/592260 Mar 09 09:31:20 _law_: it's the famouse file-native problem Mar 09 09:31:52 _law_: update your tree, clean your tmp and compile 'file'. it worked to me Mar 09 09:32:59 <_law_> katossi, only remove file stuff from tmp? or remove the whole tmp? Mar 09 09:33:52 _law_: file stuff should be enough, but for some reason I had to do a deeper cleaning, try yourself and find out Mar 09 09:35:10 rm tmp/staging/HOST_SYS/bin/file if file-native_4.16-r1 does not build? Mar 09 09:35:11 XorA: it would be nice if we could find all the dependencies in a very easy way, a la gentoo for example Mar 09 09:36:03 hrw|work: didn't work here Mar 09 09:36:05 katossi: bitbake -n something? Mar 09 09:36:29 katossi: should. I build r1 with 0 sized file in staging/bin Mar 09 09:38:10 to be sure that it will use just-built file Mar 09 09:57:55 morning Mar 09 10:20:10 hi greentux_alt2 Mar 09 10:25:30 I'm going crazy with dependencies. Why does gpe-image try to install sylpheed if it's not in task-gpe? Mar 09 10:27:32 katossi: its in conf/distro/*.conf files as extra DEPENDS Mar 09 10:29:36 or conf/machine/*.conf Mar 09 10:29:45 or conf/machine/include/*.conf Mar 09 10:32:05 someone need to clean that mess a bit... Mar 09 10:32:32 rootfs_size, gpe_extra_stuff, bigscreen etc. Mar 09 10:32:43 I'm trying to: http://handhelds.org/hypermail/oe/current/6904.html Mar 09 10:32:46 hey grma Mar 09 10:32:46 yes, sometimes I'm kind of lost Mar 09 10:32:52 oh boy Mar 09 10:33:03 hi Mar 09 10:35:39 ~lart glibc for changing interfaces Mar 09 10:35:40 * ibot hurls dozens of incontinent, insomniac, hungry kittens with tiny little razor-sharp claws and a wide variety of contagious intestinal parasites at glibc for changing interfaces Mar 09 10:36:00 I don't understand this: zaurus-clamshell.conf:GPE_EXTRA_INSTALL += "gaim sylpheed chkhinge" Mar 09 10:36:16 why does a zaurus-clamshell need gaim, sylpheed, etc... Mar 09 10:36:32 katossi: it doesnt need it, its usable because it has a vga screen Mar 09 10:36:37 katossi: Some of us like using IM on the run Mar 09 10:36:44 koen: I hate to bother you with beginner problems, but I did as you suggested few days ago, and managed to build the 'meta-sdk'. So, now I got this oe-sdk-2006xxxxx.tar.bz2 on the tmp/deploy/sdk directory, but I'm not sure what I should do with it's content. :) Mar 09 10:37:15 XorA: how could I prevent it to install? or is it just a DEPEND and not RDEPEND dependency? Mar 09 10:37:26 Cockroach-: unpack to system, add /usr/local/*/bin to PATH and use arm-linux-gcc Mar 09 10:37:30 Cockroach-: untar it somewhere on your computer and play with it :) Mar 09 10:37:40 katossi: comment out that line Mar 09 10:40:56 XorA: just because it's usable doesn't mean we need it. And I think, that should depend more on distro than in machine Mar 09 10:42:14 somebody more info? http://www.linuxdevices.com/news/NS4475809720.html Mar 09 10:45:47 koen & hrw|work: I did figure that much out, I did a meta-sdk build for both h3600 and x86 machines and the idea was to first develope, build and debug under x86 hardware and after everything works, build for arm. However when I tried to build nano (not through OE, but using the sdk) and having set CC, CPP, LDFLAGS and CPPFLAGS environment variables to point where the SDK was unpacked, everything went well except that the resulting nano binary was linked to Mar 09 10:46:15 So I suppose there is no "free lunch" by just chroot:ing to some directory and having everything working by itself ;) Mar 09 10:48:13 I lack experience when working with developing software for linux. Mar 09 10:49:11 greentux_alt2: another montavista users Mar 09 10:49:20 greentux_alt2: I suppose - 2.4.20 kernel Mar 09 10:49:33 hrw|work: ah, ok ... montavista :( Mar 09 10:49:40 Cockroach-: it was linked to wrong libs? Mar 09 10:50:07 Cockroach-: you need to give params to configure script with path to cross-libs, cross-includes Mar 09 10:50:15 greentux_alt2: It would be nice if we could see these devices running 2,6 :-/ Mar 09 10:50:32 RP: will check at CeBit... Mar 09 10:51:01 hrw|work: when setting LDFLAGS="-L/path/to/my/special/libs", it just appends that to the library search path, and not override the default ones? Mar 09 10:52:36 yesterday i got a CF umts card (from japan...)... and: it works! Mar 09 10:52:47 greentux_alt2: Nice.. Mar 09 10:52:52 greentux_alt2: How much, roughly? Mar 09 10:53:28 NAbyss: dont know. sold only with a contract (plan). i assume around $450 without. Mar 09 10:53:41 Ah, not too bad then Mar 09 10:54:22 Cockroach-: unpacked nano. it has ./configure script which probably has --with-libs --with-includes switches which you should use instead of LDFLAGS Mar 09 10:54:32 k Mar 09 10:54:34 NAbyss: will be available in uk next month Mar 09 10:54:35 ~change 450 usd to pln Mar 09 10:54:40 450.00 United States Dollar (USD) makes 1,461.08 Zloty (PLN) (from http://www.xe.com/) Mar 09 10:57:54 where can I get linux-libc-headers for a 2.4.25 kernel? Mar 09 10:58:11 koobla: linux-libc-headers are not related with running kernel Mar 09 10:58:26 good morning Mar 09 10:58:33 hm, well, agreed ;) Mar 09 10:58:35 hey florian_kc Mar 09 10:58:38 hi florian_kc Mar 09 10:59:20 if my opie-image build failed on sharp-aticore-oss then I do not get any opie stuff but opie-image is starting to build. Mar 09 10:59:51 I'm thinking about something like opie-image:DEPENDS = "what MUST be built to even start do_rootfs" **** ENDING LOGGING AT Thu Mar 09 10:59:57 2006