**** BEGIN LOGGING AT Thu Mar 06 12:37:35 2008 Mar 06 12:38:36 well, if you fix everything, we'll forgive you ;) Mar 06 12:39:05 * * OE Bug 3949 has been created by  Mar 06 12:39:07 * * kexec-tools-1.101-r4-do_compile Mar 06 12:39:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3949 Mar 06 12:41:04 * * OE Bug 3950 has been created by  Mar 06 12:41:06 * * wxwidgets-2.6.3-r0-do_compile Mar 06 12:41:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3950 Mar 06 12:47:17 hrw: ping Mar 06 12:48:42 póng Mar 06 12:48:50 hrw: about the defconfig-c7x0 you provided me: kernel size is around 930kb (no pm, no bl) Mar 06 12:49:16 :=) ...but the module support is disabled Mar 06 12:50:51 ant|work: I got better results... Mar 06 12:51:09 ant|work: I named files config-SIZEWHICHIGOT so.. Mar 06 12:51:27 hrw: it was on 2.6.24: nfs.ko alone is over 230kb! Mar 06 12:52:26 question: would it boot an initramfs from sd without modules support? I don't think ... Mar 06 12:52:45 if sd is in kernel... Mar 06 12:53:04 k, I let only sd :-) Mar 06 12:54:08 the thing is I need a full-featured defconfig to create the modules for initramfs and then a minimal one for the boot-kernel... Mar 06 12:54:18 but do I need those modules??? Mar 06 12:54:48 I decided to not build modules at all - if they have to be in initramfs then they can also be in kernel. Mar 06 12:54:57 module support take space too... Mar 06 12:55:17 I suppose in the bootimage for linux-hh is a "boot+emergency" kernel Mar 06 12:55:31 but on Z we don't have so much space... Mar 06 12:56:58 linux-hh supported devices do not have space limit for kernel Mar 06 12:57:03 they use kernel from rootfs Mar 06 12:58:11 hrw: so, finally, to have a full-fledged boot image we need to launch it with a kexec-image... Mar 06 13:00:38 hrw: should this bootimage reside on /mtd2 or would you use diferently the space in /mtd1? Mar 06 13:02:44 I would keep it in mtd2 if possible Mar 06 13:04:53 k, thx for your advices Mar 06 13:07:57 np Mar 06 13:08:44 RP: ping Mar 06 13:09:10 rschuster: pong Mar 06 13:09:22 RP: am I supposed to set ${MULTIMACH_ARCH}? Mar 06 13:09:45 rschuster: Ah, sorry I forgot to look at that problem. Are you using multimachine or not? Mar 06 13:09:56 it is because packaged-staging uses that to store the ipk control's Architecture field and it is empty for me. Mar 06 13:10:03 no I am not Mar 06 13:10:29 ok, so packaged-staging is breaking without multimachine. Hmm :/ Mar 06 13:10:29 rschuster: chinook-compat? Mar 06 13:10:33 the package's arch is 'all' though. maybe that is related. Mar 06 13:10:38 hrw: yes Mar 06 13:10:48 rschuster: I have few updates for it Mar 06 13:11:43 rschuster: http://bugs.openembedded.net/show_bug.cgi?id=3905 is one of them Mar 06 13:12:23 rschuster: packaged-staging.bbclass needs a MULTIMACH_ARCH ?= ${PACKAGE_ARCH} Mar 06 13:12:42 hrw: do you think the problems I am seeing with packaged staging are related to bugs in chinook-compat.conf? Mar 06 13:15:07 rschuster: chinook-compat do not need multimachine but packaged-staging assumes it Mar 06 13:15:39 hrw: ok Mar 06 13:16:10 rschuster: are you ok with 3905 patch? Mar 06 13:16:15 rschuster: I'm pushing an attempt at a fix Mar 06 13:16:22 rschuster: let me know if it works Mar 06 13:16:44 hrw: yes, that would be very helpful Mar 06 13:17:03 XorA, ScaredyCat: Openmoko-browser2 works ok here for me with angstrom... Mar 06 13:17:50 rschuster: ok, will push it then Mar 06 13:18:13 hrw: that fix looks like very much the right way to handle it :) Mar 06 13:18:54 RP: I know, but rschuster is maintainer of that distro so... Mar 06 13:19:19 hrw: feel free to commit whatever you think is right :) Mar 06 13:20:39 I'm tempted to suggest using multimachine too. It doesn't really hurt anything... Mar 06 13:21:10 sooner or later it will be the default Mar 06 13:21:32 rschuster: pushing - I suggest rebuilding chinook-compat from scratch Mar 06 13:21:46 hrw: ok Mar 06 13:22:02 RP: works here too now, koen I think fixed the problem yesterday Mar 06 13:22:20 XorA: great :) Mar 06 13:23:30 NOTE: package openmoko-browser2-0.0.1+svnr4172: completed Mar 06 13:23:49 * XorA sends napalm to #openmoko to clean it out Mar 06 13:26:27 hrw: BTW I assume that bootloader fix is working for you now, you havent sworn at finnish company for a few days Mar 06 13:26:43 trying a rebuild XorA Mar 06 13:27:00 XorA: thats also because I lacked time to use n810 recently Mar 06 13:27:11 hrw: heh heh Mar 06 13:27:22 hrw: bitbaker nappy-change :-) Mar 06 13:31:29 mickey|zzZZzz: Perhaps you could have a look at the pycairo .pc problem? Its corretly setting prefix in the .pc file based on whats passed to configure but the distutils stuff does things to prefix :/ Mar 06 13:31:37 * mwester-trip beats out the small fires all over him, and disconnects from #openmoko Mar 06 13:31:47 Mornin' Mar 06 13:33:25 * XorA grins evilly Mar 06 13:34:01 yeah, mickey|zzZZzz - it's 14:32 - get up ffs :P Mar 06 13:42:48 * ant|work thinks about timezones and remebers an odd issue: /etc/localtime is missing in rootfs Mar 06 14:12:57 Sorry Mar 06 14:13:15 Argh, I don't see what it was previously Mar 06 14:13:21 ah you... Mar 06 14:13:23 Can somebody please change it back? Mar 06 14:13:27 yes, me ;-) Mar 06 14:13:45 * Laibsch checks live-logs Mar 06 14:14:27 * XorA chuckles, channel hostile takeover Mar 06 14:15:11 Thanks, hrw ;-) Mar 06 14:15:31 XorA: Sorry, honest mistake Mar 06 14:16:05 Laibsch: easilly done, done it a few times myself Mar 06 14:23:58 does bb 1.8 branch still work with dev or do I need to be using the trunk? Mar 06 14:25:29 1.8 branch should work, that is what I use Mar 06 14:25:29 Laibsch: I few notes on sharprom-compat - I added a patch which stops it building glibc. glib doesn't build since the sharp libs headers don't define SSIZE_MAX and ARG_MAX and also the site config isn't compatible with sharprom. In the site config we say that va_copy is available and for sharprom it is not, it should be using __va_copy instead Mar 06 14:25:49 cbrake: Use 1.8, just latest svn Mar 06 14:26:00 or the last stable 1.8 release Mar 06 14:26:01 ok, thanks Mar 06 14:30:49 RP: Great, thanks Mar 06 14:30:57 Can we close the bug about this, then? Mar 06 14:31:05 I don't remember the exact number Mar 06 14:31:10 Laibsch: I've not fixed glib, I'm just explaining the problem Mar 06 14:31:26 Laibsch: but if there are glibc bugs, they can be closed Mar 06 14:33:17 glib or glib-2.0? Mar 06 14:33:26 Laibsch: glib-2.0 Mar 06 14:33:32 There is both the problem that glib-2.0 does not build Mar 06 14:33:44 and that glibc is built unneccesarily Mar 06 14:33:56 Laibsch: The latter is fixed, the former is not Mar 06 14:34:38 OK Mar 06 14:34:41 !oebug 3772 Mar 06 14:34:43 * * Bug 3772, Status: NEW, Created: 2008-02-13 17:00 Mar 06 14:34:44 * * : SharpROM sometimes compiles glibc which it shouldn't really do Mar 06 14:34:45 * * http://bugs.openembedded.org/show_bug.cgi?id=3772 Mar 06 14:35:42 Laibsch: Some of the glib-2.0 problems may be related to the headers issue since mixing very new kernel headers with a very old glibc isn't a good idea Mar 06 14:36:09 but the problems are not sysroot related Mar 06 14:42:04 * * OE Bug 3772 has been RESOLVED (FIXED) by Mar 06 14:42:06 * * SharpROM sometimes compiles glibc which it shouldn't really do Mar 06 14:42:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3772 Mar 06 14:53:29 <_schurig> ~seen thebohemian Mar 06 14:53:31 thebohemian was last seen on IRC in channel #oe, 787d 17h 3m 30s ago, saying: 's'. Mar 06 14:53:43 <_schurig> ouch, that's long ago Mar 06 14:53:55 _schurig: isnt that rschuster ? Mar 06 14:54:11 XorA: no idea ... Mar 06 14:54:21 rschuster: are you thebohemain? Mar 06 14:54:45 rschuster: in that case, I have a bug for you: !oebug 3946 Mar 06 14:55:23 !oebug 3946 Mar 06 14:55:24 * * Bug 3946, Status: NEW, Created: 2008-03-06 08:27 Mar 06 14:55:25 * * hs4233(AT)mail.mn-solutions.de: swt3.4-gtk-3.3+3.4M5 doesn't find "jni.h" Mar 06 14:55:26 * * http://bugs.openembedded.org/show_bug.cgi?id=3946 Mar 06 15:04:12 03rpurdie 07org.oe.dev * r266509b9... 10/ (1 classes/packaged-staging.bbclass): packaged-staging.bbclass: Set MULTMACH_ARCH if its not been set so we don't depend on multimachine.bbclass Mar 06 15:04:21 03hrw 07org.oe.dev * r21e2b3cf... 10/ (1 conf/distro/chinook-compat.conf): chinook-compat: use libgcrypt 1.4.0 like Maemo4 does Mar 06 15:04:29 03hrw 07org.oe.dev * r5c8f022e... 10/ (5 files in 3 dirs): Mar 06 15:04:29 chinook-compat, gcc-csl: make 3.4.4-csl-2005q3 have other PV for chinook-compat distro Mar 06 15:04:29 It is required to have compatibility for C++ applications as Maemo has other Mar 06 15:04:29 PV then OpenEmbedded use. Mar 06 15:17:06 the webserver at 85.214.40.226 does not respond...like yesterday afternoon CET...I'll suggest to check the crontab (updatedb @ niceness -20 ?) Mar 06 15:22:09 How broken is this pkgconfig file? Mar 06 15:22:10 http://pastebin.ca/930366 Mar 06 15:22:57 looks ok for me Mar 06 15:23:12 hmmm Mar 06 15:24:03 Crofton: Use @includedir@ and @libdir@ Mar 06 15:24:14 ok :) Mar 06 15:25:00 instead of creating them from prefix? Mar 06 15:25:06 Crofton: yes Mar 06 15:27:15 Crofton: Imagine what happens if you tell configure includedir is some weird place... Mar 06 15:44:31 RP, this leads to Mar 06 15:44:34 ant|work: thank you for the alert Mar 06 15:44:37 includedir=${prefix}/include Mar 06 15:44:37 libdir=${exec_prefix}/lib Mar 06 15:44:46 ant|work: This is not a cron problem, though. Mar 06 15:44:58 which fails for local build because it can't find exec_prefix? Mar 06 15:46:48 Crofton: Is this wired into autoconf? Mar 06 15:46:58 Crofton: or manual replacement? Mar 06 15:47:21 I am editing the .in file and rebuilding Mar 06 15:48:28 then I sue the PKG_CHECK_MODULES to look for the package Mar 06 15:48:44 Crofton: You probably need exec_prefix=@exec_prefix@ then Mar 06 15:48:56 heh Mar 06 15:49:07 I am not so good at package config ... Mar 06 15:49:50 I use a pkgconfig call to get the path to some CROAB idl files Mar 06 15:50:08 er CORBA, if I could figure out another way, I would :) Mar 06 15:50:14 Crofton: If should really be a fully expanded path you get back from autofoo though so I don't understand it :/ Mar 06 15:51:27 Crofton: Have a look at another package with pkgconfig, something like gnutls Mar 06 15:51:34 hmm, I may have not rebuilt properly after the first edit Mar 06 15:57:24 rp I needed prefix and exec_prefix Mar 06 15:58:28 Crofton: Curious. I've not noticed that before but I guess it makes sense... Mar 06 15:59:36 Crofton: Are you talking about gcc ? Mar 06 16:00:00 KhemHome: pkgconfig Mar 06 16:00:14 ok Mar 06 16:00:29 Laibsch: www.openembedded.org responds again (slowly), www.angstrom-distribution too (very slowly) Mar 06 16:01:45 khem, no, just a configure file I ahve here hates the sysroot changes Mar 06 16:02:09 SI_PATH=`pkg-config --variable=includedir standardInterfaces` Mar 06 16:02:30 the real problem is this line return /usr/lib... Mar 06 16:02:37 not the path into staging Mar 06 16:03:56 Crofton: SI_PATH=$PKG_CONFIG_SYSROOT_DIR`pkg-config...` Mar 06 16:04:34 or grep it from --cflags where it will have been expanded Mar 06 16:06:51 ant|work: The website should be back to normal Mar 06 16:07:39 * Crofton crosses fingers Mar 06 16:09:28 Laibsch: indeed, seems normal now Mar 06 16:13:31 RP, that change worked on local build, but in oe I got a possibly undefined macro, PKG_CONFIG_SYSROOT_DIR Mar 06 16:13:58 pleaase use m4_pattern_allow Mar 06 16:15:03 Crofton: ick. This is where my autofoo knowledge fails me. You want to extrct the value of that environmental variable Mar 06 16:15:36 hmm Mar 06 16:23:05 * * OE Bug 3951 has been created by  Mar 06 16:23:07 * * openmoko-mediaplayer2-0.1.0+svnr3583-r4-do_install Mar 06 16:23:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3951 Mar 06 16:24:56 ${PKG_CONFIG_SYSROOT_DIR} maybe? Mar 06 16:24:58 rp I had to add Mar 06 16:25:01 m4_pattern_allow([PKG_CONFIG_SYSROOT_DIR]) Mar 06 16:25:05 to the configure.ac Mar 06 16:25:32 I think the var name frightens autofoo because PKG is a prefix used by another set of macros Mar 06 16:26:05 Crofton: hmm, or your pkg-config.m4 is too old Mar 06 16:26:25 Crofton: or you have one hiding in the aclocal/ m4/ directory of your stuff? Mar 06 16:26:34 but I just read the two lines :) Mar 06 16:28:10 I don't think so Mar 06 16:28:21 I jsut need something to work for a few hours :) Mar 06 16:35:34 morning Mar 06 16:36:18 anybody has a stk1000 here? Mar 06 16:41:05 * * OE Bug 3948 has been marked as DUPLICATE of bug 3777 by Mar 06 16:41:07 * * glib-2.0-2.15.6-r2 - task do_configure: failed Mar 06 16:41:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3948 Mar 06 16:47:56 Laibsch: it was not duplicate Mar 06 16:48:03 Laibsch: I fixed that one Mar 06 16:49:17 bye Mar 06 17:15:35 RP: Thanks for the ABI change :) By forcing me to clean out my tmp I just removed 12GB of cruft from my HD :) Mar 06 17:44:40 hvontres|work: hehe :) Mar 06 17:46:00 * Genesis found obsolecence in OE a bit too fast for industrial project ;/ Mar 06 17:46:12 Crofton: You're probably right about the prefix. We probably need to patch the pkgconfig m4 in the sysroot patch. Upstream should have merged that by now though so upgrading pkgconfig might be easier... Mar 06 17:46:35 Genesis: It is hard to strike a balance :/ Mar 06 17:47:04 Genesis: separate stable/devel trees would help that but need manpower. It'll be interesting to see how angstrom stable does... Mar 06 17:49:16 i think the main pb is me Mar 06 17:49:44 and after that companies that call for linux architect too late on a project Mar 06 17:49:58 with more time i could organise better. Mar 06 17:50:38 man who want to build a house call first an architect next some builder Mar 06 17:50:53 word! Mar 06 17:51:13 company who want to build a electronic device call architect when the house is finished ... Mar 06 17:51:39 mickeyl ? Mar 06 17:52:32 Genesis: Whilst I understand your problem, I would happily attempt to design my own house ;-) Mar 06 17:52:34 word! = sound of approval Mar 06 17:53:08 Genesis: I spent too much time on building sites as a child ;-) Mar 06 17:53:27 it's in the road rp Mar 06 17:53:32 but i'm 24 years old Mar 06 17:54:05 * RP is "only" 26... Mar 06 17:54:08 french people are afraid by young people Mar 06 17:54:25 mickeyl: Did you see my comment about pycairo? Mar 06 17:54:33 you show the way RP , you show the way ;) Mar 06 17:54:59 also i need more XP. Mar 06 17:55:11 RP: yeah, I will look into that Mar 06 17:55:35 mickeyl: I can think of various ways to hack it but I'd prefer a neat solution if possible ;-) Mar 06 17:55:53 it's difficult Mar 06 17:56:05 mickeyl: The distutils stuff scares me :/ Mar 06 17:56:24 actually, me too. i'm afraid if i touch it, the whole system comes tumbling down :) Mar 06 17:56:43 i will probably go for a hack first Mar 06 17:57:05 mickeyl: python and perl don't seem to like cross compiling :/ Mar 06 17:57:25 * hvontres|work is feeling like an old geezer.... at 36 Mar 06 17:57:51 RP: indeed. it took me months to come up with something controllable. all the ways to construct compiler command lines are scary Mar 06 17:58:10 hvontres|work: welcome to the club Mar 06 18:00:05 howto prevent upgrade package ? Mar 06 18:00:16 i try to fix my package in distro/conf Mar 06 18:00:32 but when a version is not available any more , cause remove from mtn Mar 06 18:00:55 the package is upgrade Mar 06 18:01:18 Genesis: if something is in active use, it shouldn't have really been removed from mtn Mar 06 18:01:35 It is hard to know who is using which versions though :/ Mar 06 18:01:51 the main pb is revision Mar 06 18:03:31 " I rmember when we didn't have any fancy browsers... we used archie and ftp....and we liked it " Mar 06 18:13:13 XorA|gone:Thanks again for sending the podle over :) Mar 06 18:15:05 * * OE Bug 3952 has been created by vtzan(AT)teilam.gr Mar 06 18:15:07 * * gnash doesn't compile Mar 06 18:15:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3952 Mar 06 18:22:56 ~seen Crofton Mar 06 18:22:59 crofton is currently on #oe (23h 38m 15s) #edev (23h 38m 15s). Has said a total of 34 messages. Is idling for 1h 33m 12s, last said: 'ce_geek, it has a2d convertors that can run 100ksps or so'. Mar 06 18:25:26 hhmm Mar 06 18:25:41 RP: -f no longer works, you renamed bbfile_data, right? Mar 06 18:25:56 RP: it's bailing out while trying to delete the stamp Mar 06 19:10:58 good morning Mar 06 19:11:19 mickey|bbiab: -f works here :/ Mar 06 19:12:09 RP: is there a way to disable the "so" part for oe_libinstall for all packages ? Mar 06 19:12:19 so only static stuff is processed and so stuff is ignored? Mar 06 19:12:41 Jin^eLD: I don't know that code well but I'd suspect not Mar 06 19:13:40 hmm Mar 06 19:13:56 I'd have to check that shared flat library format, but I suspect it will also not be "so" Mar 06 19:14:25 I managed to build something in bflt using a prebuilt csl toolchain, need at least something working for the project and did not get too far with building everything from source Mar 06 19:15:05 Jin^eLD: You might be interested in what I did with poky for darwin Mar 06 19:15:32 you can build for darwin in poky? Mar 06 19:15:50 that might be interesting indeed, what should I look at, i.e. what part handles it? Mar 06 19:15:53 Jin^eLD: yes Mar 06 19:15:56 http://svn.o-hand.com/view/poky/branches/experimental/meta-darwin/package.bbclass.patch?rev=3848&view=markup Mar 06 19:16:09 Note that experimental means experimenal ;-) Mar 06 19:16:23 well, it's probably less experimental than my uclinux attempts ;) Mar 06 19:16:49 Basically it uses libtool to get dependencies Mar 06 19:16:57 its horrid Mar 06 19:17:22 nevertheless it might be exactly what I am looking for Mar 06 19:17:29 thanks Mar 06 19:26:15 re Mar 06 19:26:20 RP: read jabber? Mar 06 19:32:40 hrw: Not at that moment in time ;-) Mar 06 19:33:16 ;D Mar 06 19:34:21 03rpurdie 07org.oe.dev * ra098d198... 10/ (4 files in 2 dirs): kismet: Disable stripping and parallel make Mar 06 19:34:34 03mickeyl 07org.oe.dev * r7c95723b... 10/ (6 files in 3 dirs): dbench 3.04 update to 4.0 and apply patch to make it compile again Mar 06 19:34:45 03mickeyl 07org.oe.dev * re86cbb0a... 10/ (4 files in 3 dirs): python 2.5.2 ship _sqlite3.so binary module in python-sqlite3 (thanks emdete for spotting) Mar 06 19:34:53 03hrw 07org.oe.dev * r2b134fb4... 10/ (1 site/ix86-common): ix86-common: added ac_cv_func_posix_getgrgid_r=yes required by glib 2.15.6 Mar 06 19:35:04 03mickeyl 07org.oe.dev * r86618f96... 10/ (1 packages/python/python-pycairo_1.4.0.bb): Mar 06 19:35:04 python-pycairo 1.4.0 fix pkgconfig file. if we have more of these problems, then we need a general Mar 06 19:35:04 solution in distutils.bbclass Mar 06 19:41:46 bye Mar 06 19:41:59 bye hrw Mar 06 19:54:31 good evening Mar 06 19:55:39 gm Mar 06 20:00:04 * * OE Bug 3953 has been created by  Mar 06 20:00:06 * * libtool-cross-1.5.10-r9-do_configure Mar 06 20:00:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3953 Mar 06 20:02:04 * * OE Bug 3954 has been created by  Mar 06 20:02:07 * * openssl-0.9.7g-r6-do_compile Mar 06 20:02:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3954 Mar 06 20:36:20 http://pl.youtube.com/watch?v=ls4F_QXmMso Mar 06 21:04:52 hi all Mar 06 21:08:35 03likewise 07org.oe.dev * r5d8c39ff... 10/ (1 packages/linux/linux_2.6.24.bb): linux-2.6.24: Unbreak, misses patches which are in SRC_URI, will commit those later. Mar 06 21:15:08 hello flo_lap Mar 06 21:15:47 hrw|gone: thanks for the ncdu tip Mar 06 21:39:01 ipkg install python-pygtk-dbg_2.10.4-ml3_armv5te.ipk Mar 06 21:39:01 ipkg: invalid magic Mar 06 21:39:13 anyone know what this ipkg error means? Mar 06 21:40:30 you need to be a level 7 wizard Mar 06 21:41:28 http://www.handhelds.org/pipermail/familiar/302/30275.html Mar 06 21:49:42 hi everybody Mar 06 21:59:47 hi The_Niz Mar 06 21:59:54 eh thesing Mar 06 22:00:04 ~lart xchat Mar 06 22:00:04 * ibot DoSes xchat Mar 06 22:47:05 * * OE Bug 3879 has been RESOLVED (FIXED) by rpurdie(AT)rpsys.net Mar 06 22:47:07 * * Updated tosa support Mar 06 22:47:08 03mickeyl 07org.oe.dev * r3c8bec0c... 10/ (3 files in 3 dirs): python-pygtk 2.10.4 add nodocs.patch to make it _really_ skip building the docs Mar 06 22:47:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3879 Mar 06 22:47:15 03Dmitry 07org.oe.dev * reb1c40b4... 10/ (5 files in 3 dirs): linux-rp-2.6.24: Add tosa supprot from Dmitry Baryshkov (#3879) Mar 06 22:48:40 03Dmitry 07org.oe.dev * r682b7f9b... 10/ (3 files in 3 dirs): keymaps: Update tosa keymap from Dmitry Baryshkov (#3879) Mar 06 23:05:59 Hi all, I own a Dell Axim x51v and noticed that progress on it has pretty much stopped for a while now it seems. Since I haven't really been using mine much, I think I'd like to jump on board and work on the porting effort for it if possible. Has progress slowed due to it being impractical, or did it just lose developers? Mar 06 23:06:53 RP: something strange is happening here... bitbake -c clean no longer removes stamps for me Mar 06 23:07:19 mickeyl: Which bitbake are you using, bitbake 1.8 svn? Mar 06 23:07:23 yes Mar 06 23:07:31 the latest? Mar 06 23:07:32 1.8.x trunk Mar 06 23:07:36 yes Mar 06 23:07:59 is this with the latest OE.dev? Mar 06 23:08:04 yes Mar 06 23:08:17 should work :/ Mar 06 23:08:20 it does here... Mar 06 23:08:26 hmm Mar 06 23:08:45 sounds like rm -rf time again Mar 06 23:08:47 mickeyl: Whats the exact failure? Mar 06 23:08:54 IOError: [Errno 2] No such file or directory: '/local/pkg/oe/fic-gta01/tmp/work/fic-gta01-angstrom-linux-gnueabi/openmoko-dialer2-1_0.1.0+svnr4176-r7/install/openmoko-dialer2-dbg.lock' Mar 06 23:09:06 then i try to bitbake -c clean Mar 06 23:09:21 and once i bitbake again, it goes straight to the task package_write Mar 06 23:09:26 bailing out Mar 06 23:09:28 permissions? Mar 06 23:09:58 everything ok Mar 06 23:10:12 is there some jammed bitbake process left behind holding a lock? Mar 06 23:10:32 no undead bitbakes Mar 06 23:11:00 mickeyl: and if you try and remove that work directory by hand? Mar 06 23:11:09 well this directory is not existing in the first place Mar 06 23:11:42 hmm Mar 06 23:11:45 and another artefact Mar 06 23:11:49 libgsmd is built everytime Mar 06 23:11:56 seems my stamp handling is completely not working here Mar 06 23:12:06 ok, what does the stamps files for openmoko-dialer2 look like? Mar 06 23:12:41 they are not ther Mar 06 23:12:41 is this over nfs? Mar 06 23:12:45 local Mar 06 23:12:48 i removed all stamps Mar 06 23:12:53 then i bitbake openmoko-dialer2 Mar 06 23:12:59 and it jumps straight into package_write Mar 06 23:13:09 i did increase BB_MULTIPLE threads lately Mar 06 23:13:12 thats insane :/ Mar 06 23:13:16 indeed Mar 06 23:13:25 have you tried removing the cache? Mar 06 23:13:35 no Mar 06 23:13:36 *trying* Mar 06 23:14:04 I don't understand how is can possibly just to a task like that when no stamps exist :/ Mar 06 23:14:25 yeah Mar 06 23:14:29 must be a bogus cache Mar 06 23:14:34 i did not think of that Mar 06 23:14:40 reparsing... Mar 06 23:15:17 Also is there any generic PocketPC build I can try? I set up the openembedded environment and all that already but haven't tried building anything yet because I don't know what I should try first for the DISTRO and/or MACHINE configuration options Mar 06 23:18:58 loudawg: I think the userland is not the problem. I don't realy know the Axim, but If there is a working linux kernel the userland is easy. Mar 06 23:21:19 thesing: Ok, that's what I'm trying to understand from reading around on the angstrom and openembedded sites. I know there must be enough similar hardware between the axim and HP's or others to make this possible. Or at least I think. So I just want to start with some generic kernel or something that I can start with to see what works out of the box and what needs work. Mar 06 23:22:20 thesing: I just haven't found a clear starting point for hardware that's not listed on the openembedded site Mar 06 23:22:37 RP: oh that's new: Mar 06 23:22:44 ERROR: check_stamps fatal internal error Mar 06 23:23:16 mickeyl: ouch. What are you doing to poor bitbake? :) Mar 06 23:23:22 heh Mar 06 23:23:23 i have no idea Mar 06 23:24:13 That means that there were stamps left over when it had finished checking them and there shouldn't have been Mar 06 23:24:29 mickeyl: Was this with BB_STAMP_POLICY set? Mar 06 23:25:16 just using the default bitbake.conf Mar 06 23:25:51 mickeyl: ok. There is a load of debug code commented out in check_stamps, try uncommenting that and then send me the log... Mar 06 23:25:58 k Mar 06 23:26:36 loudawg: here is some information about the linux port http://handhelds.org/moin/moin.cgi/DellAximX51 Mar 06 23:26:59 mickeyl: I should have tied the debug code to the sanity check... Mar 06 23:27:39 loudawg: try adding this kernel to OE and create a machine that uses this kernel. Mar 06 23:28:51 thesing: Thanks, that page was my original starting point but I wasn't sure if it's just abandoned or what. I'll see what I can do with that kernel Mar 06 23:32:44 loudawg: Read their Mailinglist archives and ask on the ML. Mar 06 23:33:29 RP: http://pastebin.ca/931117 Mar 06 23:34:11 thesing: thanks, I subscribed a little earlier. What exactly is the current relationship between handhelds.org and oe.org? Mar 06 23:34:43 well Mar 06 23:34:55 there's hardly anything handhelds.org does Mar 06 23:35:02 except some folks working on the kernel Mar 06 23:35:21 so, there is no kind of relationship at all Mar 06 23:35:53 mickeyl: So does OE use kernels from handhelds.org, or does it separately do it's own kernel work? Mar 06 23:36:25 mickeyl: That is weird, [8: ''] doesn't look right at all... Mar 06 23:36:29 loudawg: some machines use kernels that are still hosted in the handhelds.org repository Mar 06 23:36:42 loudawg: there's no duplication of efforts there Mar 06 23:38:01 mickeyl: ok, cool. I'm trying to sort it all out. In the past few nights I've come across so many different sites related to this all. linuxtogo, handhelds.org, angstrom, OE, etc. I'm just trying to understand it all :-) Mar 06 23:38:15 *nod* can be confusing at first Mar 06 23:38:24 it's easy though Mar 06 23:38:40 OE is the build system Mar 06 23:38:44 angstrom is a distribution Mar 06 23:38:47 linuxtogo is a software project hoster Mar 06 23:38:52 handhelds.org the same Mar 06 23:38:57 except that they went evil Mar 06 23:39:01 which is why linuxtogo exists Mar 06 23:39:54 haha ok, thanks so much for clearing that up :-P Mar 06 23:41:05 that doesn't mean that all these share this opinion ;) Mar 06 23:41:31 you mean g**** ? Mar 06 23:41:36 mickeyl: I can tell you what isn't happening but I don't know why :/ Mar 06 23:41:45 RP: heh, ok Mar 06 23:42:12 mickeyl: That function decides whether each stamp is current or not so two piles. stamp number 8 isn't being put in either pile Mar 06 23:42:52 hmm Mar 06 23:42:55 mickeyl: It checks by recursing the dependency graph and only checks each stamp when it becomes "buildable" Mar 06 23:42:58 this is the populate_staging stamp Mar 06 23:43:07 earlier i tried bitbake -f -c populate_staging Mar 06 23:43:08 loudawg: aximx50v is already in oe. you could use this as a starting point. Mar 06 23:43:10 which failed with a traceback Mar 06 23:43:31 mickeyl: Is there something odd about that recipe? Mar 06 23:44:08 didn't touch that for a while, let me check Mar 06 23:44:18 mickeyl: yeah Mar 06 23:44:18 mickeyl: Its as if bitbake's dependency graph is corrupt somehow... Mar 06 23:44:43 mickeyl: If that were the case there should be sanity checks we'd hit before check_stamps though Mar 06 23:44:50 mickeyl: and check_stamps is really new code... Mar 06 23:44:57 hmm, looks usual Mar 06 23:45:52 RP: I tried poky on my 3100. You guys at OH did a very good job. We should use that gui for angstrom. Mar 06 23:45:57 urgs... uncompressed tiff files suck Mar 06 23:46:05 RP: i will just rm -rf tmp and rebuild over night Mar 06 23:46:08 lets see Mar 06 23:46:17 once i run into it again, I'll debug Mar 06 23:46:45 mickeyl: ok. I don't see how tmp should break bitbake though Mar 06 23:47:00 mickeyl: if it can break bitbake like this it is a bitbake bug :/ Mar 06 23:47:44 mmh, yeah Mar 06 23:47:54 mickeyl: From what I understood you'd wiped all the openmoko-dialer2 stamps? Mar 06 23:47:59 yes Mar 06 23:48:12 mickeyl: The first thing that code does is checks for stamps that don't exist and marks as not current Mar 06 23:49:39 mickeyl: so ignoring the complex stuff later, that stamp should have been a simple case :/ Mar 06 23:49:57 right Mar 06 23:50:08 The "#If the stamp is missing its not current" bit... Mar 06 23:51:15 thesing: fantastic, I didn't realize that. I don't remember seeing it listed on the angstrom wiki so I didn't think there was anything at all to start with Mar 06 23:53:25 loudawg: angstrom is not oe and does not support every machine oe knows. Mar 06 23:53:50 It seems that you can just pich aximx50v as MACHINE and start building images. Mar 06 23:56:26 thesing: thanks, pleased you like it :) Mar 06 23:57:45 thesing: ok, I figured that for the most part whatever hardware was supported by OE would most likely have at least a stub on angstrom's wiki and would have some limited support. Mar 06 23:57:52 lou, get back to work Mar 06 23:57:55 :P Mar 06 23:58:52 thesing: do I still set DISTRO to angstrom? Mar 06 23:59:02 yes Mar 06 23:59:34 great, thanks. I think that should get me started, thanks for the help Mar 06 23:59:42 mat: don't make me come over to your office :-P j/k Mar 06 23:59:42 why does it take so long for ipkg to configure the python ipks? Mar 07 00:00:08 loudawg: to angstrom-2008.1 to be exact. Mar 07 00:00:13 loudawg: now much is your little device? maybe i'll contribute to your port Mar 07 00:00:17 unless you use the stable branch Mar 07 00:00:43 loudawg: *how much Mar 07 00:01:05 thesing: right, I just didn't feel like looking up the exact string :-) Mar 07 00:01:38 mat, probably more than you might wanna spend. I think I paid close to $400 for mine if I remember right Mar 07 00:02:10 or maybe it was more than that considering they seem to be going for over $400 right now Mar 07 00:03:28 g'niught Mar 07 00:04:00 'night mickey|zzZZzz Mar 07 00:04:39 loudawg: you're right, i'll just get a chumby Mar 07 00:16:22 RP: can i have a dependency on a task of another package for a task? like do_link_initramfs depends on $(INITRAMFSPACKET)_do_build? Mar 07 00:16:57 thesing: sure Mar 07 00:17:49 do_link_initramfs[depends] = "${INIT}:do_build" Mar 07 00:18:06 See rootfs_ipk.bbclass (top of) for an example Mar 07 00:18:42 very nice. ;) Mar 07 00:19:38 Thank you. Mar 07 00:20:43 gn mickey|zzZZzz Mar 07 00:21:05 RP, any clue why ipkg configuring step is so slow? Mar 07 00:21:13 I'm about to shoot this gumstix :) Mar 07 00:21:18 and I am getting bored Mar 07 00:21:20 :) Mar 07 00:21:57 Crofton|cshool: What are you installing? Mar 07 00:22:00 kernel modules? Mar 07 00:22:05 python-pygtk Mar 07 00:22:07 no Mar 07 00:22:14 and dependencies Mar 07 00:22:14 Is it running ldconfig each time? Mar 07 00:22:21 python-core etc Mar 07 00:22:51 itt doesn't look cpu bound Mar 07 00:22:57 root is on microsd Mar 07 00:23:12 ldconfig is probably disk bound Mar 07 00:23:24 what does top say its doing? Mar 07 00:23:45 mmc process is hi Mar 07 00:23:58 I'm guessing disk bound, but it jsut seems strange Mar 07 00:24:25 ipkg does weird and wacky things Mar 07 00:24:25 don't worry too much, it just gives me time to bitch :) Mar 07 00:24:29 yeah Mar 07 00:24:36 this is what I am thinking Mar 07 00:24:53 the configure step is taking liek 10 minutes Mar 07 00:24:57 this is ridiculous Mar 07 00:25:21 hmm, something sounds wrong... Mar 07 00:25:27 all so we can have 4 buttons on a screen Mar 07 00:25:50 maybe an ipkg guru will be around in the morning Mar 07 00:25:57 no such thing :( Mar 07 00:26:01 yeah Mar 07 00:26:07 ipkg sufferer Mar 07 00:26:50 psolovosky is the only other person I know whos dared to patch it in recent times in OE other than me... Mar 07 00:27:13 ah, you are the guru then :) Mar 07 00:27:32 next time I get a chance I'll run on the OSK and see what happens Mar 07 00:27:45 maybe it is related to root on microsd Mar 07 00:27:57 sakoman only just got this working Mar 07 00:28:14 it isn't mounted sync is it? Mar 07 00:28:28 hmm Mar 07 00:28:39 that would almost certainly do it Mar 07 00:28:56 bullseye Mar 07 00:28:58 thanks Mar 07 00:30:02 ~lart sakoman for taking vacation Mar 07 00:30:02 * ibot rm -rf's sakoman for taking vacation Mar 07 00:30:13 I'll poke at it in the morning Mar 07 00:31:03 I wonder why he mounted it sync? Mar 07 00:32:15 fstab has root as defaults,sync,noatime Mar 07 00:34:40 Crofton|cshool: sync is safe, async is dangerous, at least in theory Mar 07 00:34:57 painfully safe Mar 07 00:35:13 I guess he's afraid of people poppping the card out and making a mess Mar 07 00:36:09 I can't decide what is more annoyint a corrupt file systm or ipk's taking 10 minutes to install :) Mar 07 00:36:29 I go the async route ;-) Mar 07 00:36:44 * RP -> Zzzz. 'night all Mar 07 00:36:53 I think we will beat our heads against the wall and hope the pain ends soon Mar 07 00:36:54 nite rp Mar 07 00:36:55 good night Mar 07 00:36:55 gn Mar 07 00:38:43 good night Mar 07 00:44:07 a/win 18 Mar 07 01:03:04 * * OE Bug 3955 has been created by  Mar 07 01:03:07 * * wlan-ng-modules-0.2.1-pre26-r2-do_compile Mar 07 01:03:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3955 Mar 07 01:45:56 night all Mar 07 03:05:04 * * OE Bug 3956 has been created by  Mar 07 03:05:06 * * openmoko-calculator2-0.1.0+svnr3655-r0-do_install Mar 07 03:05:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3956 Mar 07 04:11:05 * * OE Bug 3957 has been created by  Mar 07 04:11:07 * * libeventdb-0.90-r1-do_compile Mar 07 04:11:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3957 Mar 07 06:11:45 ~oemirrors **** BEGIN LOGGING AT Fri Mar 07 06:24:49 2008 Mar 07 07:21:50 morning all Mar 07 07:22:07 looks like monotone.openembedded.org is down again Mar 07 07:22:54 RP: it looks like the linux-rp-2.6.24/tosa directory is empty Mar 07 07:34:42 hvontres|home: it works here Mar 07 07:39:48 morning Mar 07 07:47:45 ynezz: hmm, somebody must have kicked it :) Mar 07 07:52:31 hrw|gone: morning :) Mar 07 07:56:41 hi all ! Mar 07 08:16:53 night all Mar 07 08:42:33 morning Mar 07 08:50:19 bonjour Mar 07 08:51:55 morning Mar 07 08:52:10 morning Mar 07 09:01:36 is anyone actually working on webkit? Mar 07 09:01:48 Koen upgraded it recently Mar 07 09:15:36 * schurig wonders who will tend to all of those automatically or manually bugs ... don't they just pile up? Mar 07 09:49:35 03xora 07org.oe.dev * r38138654... 10/ (1 packages/dbus/dbus-native_1.0.3.bb): dbus-native_1.0.3.bb : fix SRC_URI Mar 07 09:49:47 03koen 07org.oe.angstrom-2007.12-stable * r768b9b6a... 10/ (55 files in 4 dirs): Mar 07 09:49:47 linux-rp: backports from .dev: Mar 07 09:49:47 * make .23 default for poodle Mar 07 09:49:47 * misc fixes for .23 Mar 07 09:49:47 * merge .24 as well, but disable it by default Mar 07 09:49:52 03koen 07org.oe.angstrom-2007.12-stable * rb84a4ddb... 10/ (1 contrib/angstrom/build-feeds.sh): angstrom feed builder: add gpsbabel and sudo Mar 07 09:50:01 03koen 07org.oe.angstrom-2007.12-stable * r63fee798... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007: bump to r9 Mar 07 09:51:41 morning all Mar 07 09:55:29 schurig: good question... Mar 07 09:59:19 03rpurdie 07org.oe.dev * rb7110cfd... 10/ (69 files in 2 dirs): linux-rp-2.6.24: Add missing patches from previous commit Mar 07 10:06:07 03xora 07org.oe.dev * r586ec17f... 10/ (3 files in 2 dirs): Mar 07 10:06:07 angstrom-2007-for-openmoko-versions.inc, preferred-om-2008-versions.inc : Mar 07 10:06:07 update some versions to what is currently available in mtn Mar 07 10:34:48 good monring Mar 07 10:35:00 monring florian Mar 07 10:41:59 03hrw 07org.oe.dev * r49e64703... 10/ (4 files in 2 dirs): less: upgrade to 418 (from Poky), drop 382 Mar 07 10:42:05 03xora 07org.oe.dev * r13e3b90d... 10/ (2 files in 2 dirs): angstrom-2007-for-openmoko-versions.inc : one more version change Mar 07 10:42:13 03koen 07org.oe.angstrom-2007.12-stable * r584d00fe... 10/ (69 files in 2 dirs): linx-rp 2.6.24: repair last commit by ading missing patches Mar 07 10:44:05 03hrw 07org.oe.dev * r35db89c4... 10/ (1 packages/ncdu packages/ncdu/ncdu_1.3.bb): ncdu: added 1.3 version of ncurses disk usage tool Mar 07 10:49:45 i've a question for a new "platform" Mar 07 10:50:03 diferences between glibc & ucLinux **** ENDING LOGGING AT Fri Mar 07 10:59:56 2008