**** BEGIN LOGGING AT Sun Aug 27 02:59:57 2006 Aug 27 07:06:02 morning all Aug 27 08:13:35 good morning all Aug 27 08:22:44 03koen 07org.oe.dev * r3ff0f8f7... 10/ (1 classes/package.bbclass): package.bbclass: remove do_install from PACKAGE_FUNCS, as noted by Philipp Zabel Aug 27 08:43:14 03koen 07org.oe.dev * r5b5482ad... 10/ (1 packages/gnome/gconf-dbus_svn.bb): gconf-dbus: depend on dbus-glib Aug 27 08:43:53 koen: qemu-native 0.8.2 built by gcc 3.4.6 still doesn't like to run localedef here. Aug 27 08:43:53 I get signal 4 (Illegal instruction) - probably tune-iwmmxt's fault? Aug 27 08:51:42 I have gcc version 3.3.6 (Debian 1:3.3.6-13) Aug 27 08:51:52 but yes, it could be tune-iwmmxt Aug 27 08:55:58 pH5: do you get those file not found problems with dbus-glib? Aug 27 08:56:20 I have to touch some introspect.xml and some .h to get it past configure Aug 27 09:00:09 koen: oh well, I totally forgot about that. dbus-glib-native has to be fixed. Aug 27 09:05:37 kergoth: pong Aug 27 09:06:12 morning all btw :) Aug 27 09:06:18 hey RP Aug 27 09:06:19 morning RP Aug 27 09:07:51 nifty sci-fy movie idea: have bitbake run the runqueue through graphviz and scroll that over the screen with colours indicating the status of a task Aug 27 09:09:06 graphviz seemed to have trouble with the size of the dependency graph ;-) Aug 27 09:10:08 indeed qemu doesn't support iwmmxt at all and paul brooks doesn't have any interest in it. Is there a way to switch processor optimization on a per-package basis? Aug 27 09:10:37 pH5: Does it support it on x86? Aug 27 09:10:38 zap TARGET_CC flags for glibc? Aug 27 09:12:20 RP: i don't understand. isn't iwmmxt xscale-only? it does support x86's mmx. Aug 27 09:12:38 iwmmxt isn't like mmx Aug 27 09:12:45 ~lart intel marketing Aug 27 09:12:46 * ibot --purges intel marketing Aug 27 09:12:50 ack Aug 27 09:12:56 The instruction sets are the same as I understood it? Aug 27 09:13:58 iirc not Aug 27 09:14:11 but your bearded bossman should have a book on it :) Aug 27 09:14:37 RP: they are similar. there are 8 pages in the iwmmxt dev manual about porting form mmx/sse Aug 27 09:15:05 koen: This was something I heard from my bearded bossman ;-) Aug 27 09:15:12 :) Aug 27 09:16:14 I know mallum would like to see xfbdev using them Aug 27 09:18:39 too bad nico isn't really interested in X Aug 27 09:19:47 He's been very quiet recently Aug 27 09:21:23 * koen reads http://www.openembedded.org/wiki/OEDEM Aug 27 09:23:36 RP: multithreading is working pretty well now Aug 27 09:24:04 koen: Excellent :) Aug 27 09:24:39 I should do some further work on bitbake at some point - try and sort out msy.py and cleanup bin/bitbake some more... Aug 27 09:26:43 morning Aug 27 09:28:21 hey eFfeM Aug 27 09:28:49 back in town again Aug 27 09:29:38 did an update and suddenly everything needs rebuilding and fails, and after a bb -cclean of such a failing package it works Aug 27 09:29:39 very odd Aug 27 09:29:55 I rearranged some tasks Aug 27 09:30:12 eg gcc-cross-initial failed on patching cse.patch then after bb -cclean of gcc-cross-initial it works Aug 27 09:30:24 ah, that's the do_stamp change Aug 27 09:30:33 koen, i did not expect such changes to let things fail Aug 27 09:30:47 currently recompiling glibc Aug 27 09:44:25 pH5: dbus 0.92 seems to be working on my hx, can I unmask it? Aug 27 09:45:54 koen: seems to work here, too. go ahead. I'm currently fiddling around with the dbus-native/dbus-glib-native/dbus-glib stuff Aug 27 09:49:20 hi all Aug 27 09:49:30 hi mr_nice_slacker Aug 27 09:49:45 eFfeM: hi Aug 27 09:49:51 03koen 07org.oe.dev * rf047496c... 10/ (1 packages/dbus/dbus_0.92.bb): dbus 0.92: remove default pref = -1 Aug 27 09:51:19 RP: they are similar. there are 8 pages in the iwmmxt dev manual about porting form mmx/sse Aug 27 09:51:44 whoops, hit the up-arrow button Aug 27 09:52:04 "qemu: uncaught target signal 4 (Illegal instruction) - exiting" Aug 27 09:52:04 that is glibc built with tune-xscale. Aug 27 09:52:08 :-/ Aug 27 09:52:37 does anyone konw where I can download embeddedSTL(http://www.exactcode.de/embeddedSTL/index.html)? On the site I can't find any download option. Aug 27 09:53:24 pH5: Which qemu - 0.8.2? Aug 27 09:53:36 pH5: I know the latest cvs does things like that Aug 27 09:54:07 yes, qemu 0.8.2 compiled with debian's gcc 3.4.6-4 Aug 27 09:55:33 Hmm. Ununtu's 3.4.6 works here... Aug 27 09:56:05 * RP -> back later Aug 27 10:00:26 Where can I read more about switching between building for machine X and Y? Aug 27 10:07:18 ~seen france Aug 27 10:07:21 france is currently on #gpe #familiar #oe #opie #handhelds.org. Has said a total of 26 messages. Is idling for 1d 16h 27m 55s, last said: 'thank u :-D'. Aug 27 10:07:28 Laibsch: if you use a distro that uses the multimachine buildrules (files end up in arch/machine subdirs), just change the MACHINE variable in local.conf or interactive bitbake Aug 27 10:11:17 or use auto.conf for that :) Aug 27 10:12:24 ah, the locale-tree is populated from the staged glibc (or glibc-intermediate) build instead of the current glibc build tree. Aug 27 10:14:07 what I have in staging still is iwmmxt. running qemu-arm localedef -L with the path to ld-linux.so.3 in the current glibc workdir doesn't give the illegal instruction Aug 27 10:17:28 do I feel a patch coming? Aug 27 10:17:50 since staging happends after packaging now Aug 27 10:17:59 -d Aug 27 10:18:23 Laibsch: in the .oz354x branch there's a contrib/buildscripts dir Aug 27 10:25:31 pH5: I guess "openzaurus-unstable" does not do that, right? Judging from looking at the tmp dir. Aug 27 10:26:36 conf/distro/include/openzaurus.inc:8:INHERIT += " package_ipk debian multimachine" Aug 27 10:29:28 Indeed. But the files do not end up in arch/machine subdirs as described it. Aug 27 10:29:46 s/as /as pH5 / Aug 27 10:53:25 mickey|bbl: could you look into http://bugs.openembedded.org/show_bug.cgi?id=1366 ? Aug 27 11:00:06 * pH5 wonders if something like this would make sense: http://en.pastebin.ca/150852 Aug 27 11:00:42 looks ok Aug 27 11:01:41 this is important for people like me who want to compile glibc-intermediate with iwmmxt optimizations ;) Aug 27 11:03:40 and for people that don't use ntpl and don't have glibc-i Aug 27 11:04:04 CoreDump|home: Is hentges.net still the place to go for a decent Opie-based collie image btw? :) Aug 27 11:10:17 pH5: we might have a iwmmxt enabled liboil soon Aug 27 11:10:33 * pH5 excited Aug 27 11:23:34 * pH5 marks the OEDEM weekend in his calendar Aug 27 11:26:45 03pH5 07org.oe.dev * r7217ed94... 10/ (1 packages/glibc/glibc-package.bbclass): glibc: use libs from /lib instead of for binary locale generation Aug 27 11:27:00 03pH5 07org.oe.dev * rcecec739... 10/ (1 packages/dbus/dbus-native_0.92.bb): dbus-native: add 0.92 Aug 27 11:27:08 03pH5 07org.oe.dev * rc2ec9386... 10/ (4 files in 3 dirs): dbus-glib-native-0.71: fix introspection data / binding header generation Aug 27 11:27:13 uh, wrong quotes Aug 27 11:27:33 that should have been 'use libs from ${D}/lib instead of ${STAGING_LIBDIR} for binary locale generation' Aug 27 11:29:58 koen: could you test if building dbus-glib-native works now? Aug 27 11:30:18 ehm Aug 27 11:30:24 not without wiping /tmp Aug 27 11:33:49 koen: oh, ok. I guess we'll notice soon enough if there's something still wrong with it. Aug 27 11:48:35 03koen 07org.oe.dev * r93cd953c... 10/ (1 packages/gstreamer/gstreamer_0.10.8.bb): gstreamer: add 0.10.8 Aug 27 11:54:35 morning Aug 27 11:59:55 hm, what happened to dropbear? install says: Aug 27 11:59:56 ln: creating symbolic link `/home/frans/OpenSlugHead/openslug/tmp/work/dropbear-0.48.1-r0/image/usr/bin/dbclient' to `/usr/sbin/dropbearmulti': File exists Aug 27 12:00:17 you bet it is not allowed to scribble in /usr/sbin on my dev system ... Aug 27 12:03:36 eFfeM: wipe tmp/ Aug 27 12:04:38 koen, for the dropbear stuff? Aug 27 12:04:48 i think this is a bug in dropbear.inc. It reads: Aug 27 12:04:54 ln -s ${sbindir}/dropbearmulti ${D}${bindir}/dbclient Aug 27 12:05:01 where it should read ln -s ${D}${sbindir}/dropbearmulti ${D}${bindir}/dbclient Aug 27 12:05:07 so with a ${D} Aug 27 12:05:14 this is in do_install Aug 27 12:06:34 eFfeM: no, the link should point to ${sbindir}/dropbearmulti, as that's where it will be once the package is installed Aug 27 12:08:03 pH5 this is in do_install, not in postinst ! Aug 27 12:08:11 and later on it does say: ln -s ./dropbearmulti ${D}${sbindir}/$i Aug 27 12:09:22 pH5 and this definitely is in the install script in the temp dir Aug 27 12:19:59 x29a: ping Aug 27 12:29:38 03koen 07org.oe.dev * r1a396604... 10/ (1 packages/gstreamer/gst-plugins-base_0.10.7.bb): gst-plugins base: stage things so good, bad and ugly can proceed to build stuff Aug 27 12:37:06 03koen 07org.oe.dev * r637ca3e6... 10/ (3 files in 2 dirs): Aug 27 12:37:06 angstrom: slide data a teensy bit forward Aug 27 12:37:06 * BIG FAT NOTE: adjust your {local,site,auto}.conf Aug 27 12:37:33 * koen points to the big fat note Aug 27 12:51:17 ibot: botmail for zecke: http://autopackage.org/apbuild-apgcc.php Aug 27 13:23:14 mr_nice_slacker: pong Aug 27 13:23:29 mr_nice_slacker: did you already manage to get the .tar.bz? Aug 27 13:24:40 * RP returns Aug 27 13:38:55 x29a: hi, no unfortunate i wasn't Aug 27 13:39:12 mr_nice_slacker: what do you need again? Aug 27 13:39:17 the data from the debian-package? Aug 27 13:39:29 x29a_: yes - would be nice Aug 27 13:45:47 x29a_: i think i will have to install alien Aug 27 13:46:20 why wont normal compiling with source work? Aug 27 13:46:47 x29a_: I don't understand why i got into compiling errors Aug 27 13:47:14 x29a_: all prerequest are ok - i even tryed the debian patches Aug 27 13:47:52 x29a_: the 0.4 compiles verry well Aug 27 13:48:18 mr_nice_slacker: should dpkg --unpack openjtag.deb not do? Aug 27 13:49:16 x29a_: i have no dpkg i will install have to install it or better install alien Aug 27 13:49:43 mr_nice_slacker: no, for me to get your data out of the .deb Aug 27 13:50:34 x29a_: ah, you have the data allready installed - i think a --unpack is not what it seems to be it could put your package in a other state - if i got the manual right Aug 27 13:51:09 x29a_: but you can do a ??listfiles Aug 27 13:52:34 it says the package is not installed Aug 27 13:54:36 ar x foo.deb ; tar zxvf data.tar.gz Aug 27 13:55:53 koen: :) Aug 27 13:55:57 thx koen Aug 27 13:59:12 I have observed several times now that a package fails to build but upon restarting the same bitbake command it builds fine. In between there have been no changes to the OE tree such as "mtn update" or the like. How can this happen? Aug 27 14:01:11 x29a: thx - but now i have a .deb package *g Aug 27 14:01:27 Laibsch: bad RAM? Aug 27 14:01:49 mr_nice_slacker: what? Aug 27 14:02:10 mr_nice_slacker: look inside, there should be data.tar.gz Aug 27 14:03:42 x29a: ah, ok Aug 27 14:07:21 x29a: fine :) - now i can erase all of my flash and install the image i gave you Aug 27 14:07:31 hehe Aug 27 14:07:34 go ahead Aug 27 14:07:55 x29a: to check if there are some hidden blocks which we need Aug 27 14:09:34 03koen 07org.oe.dev * rce3b6559... 10/ (1 packages/gstreamer/gst-plugins-ugly_0.10.4.bb): gst plugins ugly: add 0.10.4 Aug 27 14:10:50 x29a:my simpad pisses me of a bit ;) Aug 27 14:12:33 x29a: i tryed out the new familiar images and both are not near to be stable on my pad - same on the t-sinus Aug 27 15:12:04 hi Aug 27 15:19:15 x29a: i was not able to get it running - too many differences between the distributions. I now have patched 0.4 with the erasemem patch. I will try to erase the flash now Aug 27 15:25:58 x29a: erasing Aug 27 15:27:06 x29a: btw the money is now on my paypal should i buy that card for you? Aug 27 15:29:06 x29a: flashing ... Aug 27 16:18:43 mr_nice_slacker: will you buy one for yourself? Aug 27 16:19:05 mr_nice_slacker: i created my own paypal account now, dunno if itd be better to do it over this one Aug 27 16:34:45 Anybody else having problems to build libxine? Aug 27 16:43:59 mr_nice_slacker: is yours working? Aug 27 16:47:31 x29a: it boots Aug 27 16:48:12 no way Aug 27 16:48:24 x29a: i have erased the hole first flash and than i have flashed it Aug 27 16:48:27 the same thing u uploaded flashmem 0 image.img Aug 27 16:48:41 mr_nice_slacker: me too, it wont turn on Aug 27 16:48:53 x29a: Test_bootldr.img Aug 27 16:48:59 whatever Aug 27 16:49:02 256kb or so Aug 27 16:50:14 x29a: http://opensimpad.org/images/b/b7/Test_bootldr.img Aug 27 16:51:12 x29a: i will veryfy it with the t-sinus Aug 27 16:51:25 mr_nice_slacker: wait Aug 27 16:51:30 can u first read it again? Aug 27 16:51:42 cause now the rest of the flash will be zero, for sure Aug 27 16:54:54 x29a: i can do, but it will be the same caus i have deleted everything and flashed the image Aug 27 16:55:08 hmm, right Aug 27 16:55:12 no u dont have to Aug 27 16:55:20 did u turn your pad on via the on/off button? Aug 27 16:55:26 x29a: no Aug 27 16:55:34 with the jumper? Aug 27 16:55:56 x29a: i had to disconnect it from the powerline Aug 27 16:56:06 x29a: and than i had to reboot it Aug 27 16:56:26 x29a: cause on the first boot it didn't came up comletely Aug 27 16:57:05 x29a: so try flashing it a second ;) - and disconnect it from the power reboot twice and see what will happen Aug 27 16:57:46 x29a: the jtag tools put the pad in a condition where the reboot switch don't work Aug 27 16:58:39 x29a: a - and close the jtag tools Aug 27 17:04:12 x29a: i think the jumper will work ,too Aug 27 17:08:45 03xora 07org.oe.dev * r31791b51... 10/ (13 files in 5 dirs): Aug 27 17:08:45 links-x11_2.1pre21.bb, links_2.1pre12.bb : patches for cookie saving and Aug 27 17:08:45 preventing segfault on cookies imported from MDK rpms. Aug 27 17:08:45 packages/links/* : drop old versions Aug 27 17:22:18 03koen 07org.oe.dev * r90a1ac7c... 10/ (1 packages/e17/e17-gpe-menu-convert_0.2.bb): e17 menu-convert: set package arch to all Aug 27 17:24:29 * RP -> food Aug 27 17:26:02 03rpurdie 07org.oe.dev * r33001ff2... 10/ (1 classes/gconf.bbclass): gconf.bbclass: Add gconf to DEPENDS (from poky) Aug 27 17:26:08 03rpurdie 07org.oe.dev * rf9004685... 10/ (1 classes/base.bbclass): base.bbclass: Add base_contains and base_both_containg - useful functions for the new task-machine code (from poky) Aug 27 17:26:23 03rpurdie 07org.oe.dev * r8b6cbc02... 10/ (1 classes/gtk-icon-cache.bbclass): gtk-icon-cache.bbclass: Add from poky Aug 27 17:26:33 03rpurdie 07org.oe.dev * r9cef20f4... 10/ (1 conf/machine/include/tune-arm926ejs.conf): tune-arm926ejs.conf: Be a little more specific about gcc version issues (from poky) Aug 27 17:27:00 03rpurdie 07org.oe.dev * r0e722bfb... 10/ (7 files in 4 dirs): bitbake.conf: Remove old style IMAGE_FSTYPE in favour of IMAGE_FSTYPES (which is weakly set to a default allowing machines to override/append) (from poky) Aug 27 17:27:11 03rpurdie 07org.oe.dev * r6d0f4411... 10/ (1 conf/bitbake.conf): bitbake.conf: Add PN to OVERRIDES in the form pn-PN (from poky) Aug 27 17:27:17 03rpurdie 07org.oe.dev * r810a665c... 10/ (7 files in 3 dirs): zaurus machines: Clean up IPKG_EXTRA_ARCHS and IMAGE_FSTPES (from poky) Aug 27 17:27:24 03rpurdie 07org.oe.dev * rd66b0dea... 10/ (1 packages/tasks/task-base.bb): task-base.bb: Add from poky Aug 27 17:27:36 03rpurdie 07org.oe.dev * r8a47f06b... 10/ (16 files in 3 dirs): Add MACHINE_TASK_PROVIDER to remove hardcoded references to task-bootstrap and allow machines/distros to elect an alternative Aug 27 17:27:45 03rpurdie 07org.oe.dev * r47d9bffa... 10/ (4 files in 2 dirs): Switch 2.6 Zaurus models to use task-base Aug 27 17:42:27 RP: yay! Aug 27 17:42:56 I've just realised the distros lack the appropriate config to make it all work but its close :) Aug 27 17:43:04 man.. yesterday one of my shell scripts went wacky and did an rm -rf $HOME Aug 27 17:44:18 kergoth: oops :-( Aug 27 17:46:15 I'd best tend to the fire I just started :) Aug 27 17:47:16 * koen|away will tend to the h2200, ipaq-pxa270, h5xxx, ep93xx and netbook-pro machines tomorrow Aug 27 17:47:36 * koen|away gets some marshmellows and walks to org.oe.dev Aug 27 17:50:02 RP: thank god the dotfiles and such are in svn.. just awfully time consuming checking everything back out again :P Aug 27 17:50:06 * kergoth grumbles Aug 27 17:56:05 koen|away: thanks for the botmail, I remember the rants about autopackage though Aug 27 18:00:25 What has RP done? Aug 27 18:04:06 Crofton: all the 'rpurdie' commits Aug 27 18:04:19 I know :) Aug 27 18:04:31 I'm just not sure where he is going :) Aug 27 18:04:36 zecke: we should investigate the --as-needed thing Aug 27 18:07:51 there, that's better Aug 27 18:18:20 RP: could you send a note to the mailinglist about the changes? Aug 27 18:21:14 koen|away: it breaks on freebsd ;) Aug 27 18:21:19 what differences are there between analog and digital flat panels? Aug 27 18:22:43 Crofton: an extra set of DAC + ADCs Aug 27 18:25:12 heh Aug 27 18:27:52 zecke: freebsd doesn't have a cool logo, so screw it Aug 27 18:40:24 gah Aug 27 18:40:28 sf.net is broken Aug 27 18:40:53 ~lart sourceforge Aug 27 18:40:54 * ibot squishes sourceforge like a bug Aug 27 19:02:27 ~lart XFS for destroying my RAID the second time in one week Aug 27 19:02:28 * ibot strangles XFS with a 9-pole serial cable for destroying my RAID the second time in one week Aug 27 19:03:02 xfs_repair? Aug 27 19:03:05 xfs_db? Aug 27 19:03:38 http://oss.sgi.com/projects/xfs/faq.html#dir2 Aug 27 19:03:41 worse Aug 27 19:03:59 anyway, bbiab Aug 27 19:04:19 ah, that one Aug 27 19:05:14 mkfs.xfs did _not_ get rid of it. Hence the second crash w/ data-loss grrr Aug 27 19:05:37 * CoreDump|bbl now really afk Aug 27 19:06:19 that's why I don't upgrade kernels that often Aug 27 19:45:11 woohoo.. kergoth is still alive. Aug 27 19:50:23 koen: I was kinda forced to upgrade the kernel. Ususally I keep the same (working) kernel for many month Aug 27 19:51:29 +root@mhcln01:..ia/md2/chroot/test-jail2 >cat /proc/cpuinfo |grep "model name"| uniq Aug 27 19:51:30 model name : Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz Aug 27 19:51:38 =) Aug 27 19:55:37 http://printar.fixed.name/index.php Aug 27 19:55:43 local pasty faced nerds Aug 27 19:59:54 lol Aug 27 20:26:06 hi all Aug 27 20:31:46 hi flatronf701C Aug 27 20:31:51 hi florian Aug 27 21:18:20 03rpurdie 07org.oe.dev * r0e65e503... 10/ (4 files in 2 dirs): Add DISTRO_FEATURES to angstrom, familar and openzaurus. At this moment this maintains status quo but distro maintainers are free to adjust it to their needs Aug 27 22:14:07 03rpurdie 07org.oe.dev * rf4b38437... 10/ (1 conf/machine/include/qemu.conf conf/machine/qemuarm.conf): qemuarm: Create common include file and convert to task-base Aug 27 22:14:13 03rpurdie 07org.oe.dev * rf74511a3... 10/ (1 conf/machine/nokia770.conf): nokia770: Convert machine to task-base Aug 27 22:14:18 03rpurdie 07org.oe.dev * r34625a6c... 10/ (4 files in 3 dirs): conf/machine: various task-base conversion misc cleanups Aug 27 22:55:52 03rpurdie 07org.oe.dev * r95b27794... 10/ (1 packages/atk/atk_1.10.1.bb packages/atk/atk_1.10.3.bb): atk: 1.10.1 -> 1.10.3 (from poky) Aug 27 22:55:55 03rpurdie 07org.oe.dev * r3f653813... 10/ (4 files in 2 dirs): autoconf: Drop bogus RRECOMMENDS Aug 27 22:56:04 03rpurdie 07org.oe.dev * rda1a8e6e... 10/ (5 files in 2 dirs): avahi: Fix postinst so we only restart dbus if its already running (it might not have started when the postinst runs) Aug 27 22:56:12 03rpurdie 07org.oe.dev * r72ef3ece... 10/ (4 files in 3 dirs): cairo: Add 1.2.4 (including float conversion speedup) (from poky) Aug 27 22:56:19 03rpurdie 07org.oe.dev * r1e0480e5... 10/ (12 files in 3 dirs): pango: 1.13.2 -> 1.13.3 and add missing PACKAGES_DYNAMIC (from poky). Drop some old versions. Aug 28 00:02:19 03lenehan 07org.oe.documentation * re49f3539... 10/ (1 usermanual/reference/class_autotools.xml): Aug 28 00:02:19 usermanual class_autotools: Document the site file syntax, in particular Aug 28 00:02:19 mention the use of the shell syntax for conditionally setting a variable Aug 28 00:02:19 only if teh variable is not already set. **** ENDING LOGGING AT Mon Aug 28 02:59:56 2006