**** BEGIN LOGGING AT Wed Jan 05 02:59:58 2011 Jan 05 06:27:57 mwester: ipkg-utils snapshot used by slugos is not fetchable anymore Jan 05 06:28:20 mwester: can you give me the tar file so we can upload it to oe source mirror Jan 05 07:17:02 gm Jan 05 07:30:51 gm all Jan 05 08:06:15 khem, http://uclibc.org/~aldot/oe/ipkg-utils_anoncvs.handhelds.org__20050930.tar.gz Jan 05 08:07:11 khem, i, personally, would much prefer to just add full .opk support (see my question and non-patch to the list) Jan 05 08:29:23 03Melissa Watkins  07org.openembedded.dev * r596212da20 10openembedded.git/recipes/ti/ (4 files): Jan 05 08:29:23 ti-pru-sw: Update SRC_REV Jan 05 08:29:23 For the ti-pru-sw-examples, ti-pru-sw-app-loader, ti-pru-sw-edma-driver, Jan 05 08:29:23 ti-pru-sw-edma-library: Jan 05 08:29:23 * Update SRC_REV Jan 05 08:29:23 Signed-off-by: Melissa Watkins Jan 05 08:29:23 Signed-off-by: Koen Kooi Jan 05 08:29:49 03Jeff Lance  07org.openembedded.dev * r918ac10ebc 10openembedded.git/recipes/ti/am-benchmarks_1.2.bb: (log message trimmed) Jan 05 08:29:49 am-benchmarks: add platform support and correct makefile Jan 05 08:29:49 * Modify recipe name to include package version Jan 05 08:29:49 * add support for additional platforms am181x-evm and am180x-evm Jan 05 08:29:49 * Update the SRCREV Jan 05 08:29:50 * Fix pathname to Rules.make in each Makefile Jan 05 08:29:50 Signed-off-by: Jeff Lance Jan 05 08:30:02 03Jeff Lance  07org.openembedded.dev * rb1008d9751 10openembedded.git/recipes/ti/am-benchmarks_svn.bb: Jan 05 08:30:03 am-benchmarks: Remove am-benchmarks_svn.bb Jan 05 08:30:03 * Remove am-benchmarks_svn.bb now that the recipe name inlucdes the PV. Jan 05 08:30:03 Signed-off-by: Jeff Lance Jan 05 08:30:03 Signed-off-by: Koen Kooi Jan 05 08:30:14 03chase maupin  07org.openembedded.dev * r2d922abff7 10openembedded.git/recipes/ti/ (matrix-tui.inc matrix-tui_1.1.bb): (log message trimmed) Jan 05 08:30:14 matrix-tui: adding multi platform support Jan 05 08:30:14 * adding am180x, am181x, da850-omapl138, am3517 and am3715 support Jan 05 08:30:14 * Updating SRC rev to pull in matrix-tui project with platform support Jan 05 08:30:14 * removing INSANE_SKIP, it originally handled pre-built binaries which Jan 05 08:30:14 have since been removed Jan 05 08:30:14 * Creating version 1.1 of the recipe Jan 05 08:33:45 03Martin Jansa  07master * r561d3bdaa3 10openembedded.git/recipes/shr/libphone-ui_git.bb: Jan 05 08:33:45 libphone-ui: use branch master again Jan 05 08:33:45 Signed-off-by: Martin Jansa Jan 05 08:33:55 03Martin Jansa  07master * rd7fd75b65b 10openembedded.git/recipes/openmoko-3rdparty/ (iliwi/new.vala.patch iliwi_git.bb): Jan 05 08:33:55 iliwi: add patch to build it with newer vala Jan 05 08:33:55 Signed-off-by: Martin Jansa Jan 05 08:33:58 03Martin Jansa  07master * rdfcd2e6ded 10openembedded.git/recipes/freesmartphone/fsotdld_git.bb: Jan 05 08:33:58 fsotdld: add json-glib to DEPENDS, dependency added in fed5109c117258ac3033df6252920047b299464e Jan 05 08:33:58 * http://git.freesmartphone.org/?p=cornucopia.git;a=commit;h=fed5109c117258ac3033df6252920047b299464e Jan 05 08:33:58 Signed-off-by: Martin Jansa Jan 05 08:58:53 03Michael 'Mickey' Lauer  07org.openembedded.dev * r3a8c8d04f7 10openembedded.git/recipes/freesmartphone/fsotdld_git.bb: fsotdld: add dependency to libsoup-2.4 Jan 05 09:16:54 03Koen Kooi  07org.openembedded.dev * r01ffe4edd1 10openembedded.git/recipes/ti/ (3 files in 2 dirs): Jan 05 09:16:54 ti-c6accel 1.01.00.03: add patch to fix link problem Jan 05 09:16:54 Signed-off-by: Koen Kooi Jan 05 09:17:57 03Martin Jansa  07master * re2c0c14c40 10openembedded.git/recipes/gdbus-binding-tool/gdbus-binding-tool_git.bb: Jan 05 09:17:57 gdbus-binding-tool: fix do_configure in native build Jan 05 09:17:57 * do_configure_prepend without _virtclass-native override is also prepended, that's not what I intended Jan 05 09:17:57 Signed-off-by: Martin Jansa Jan 05 09:35:30 03Martin Jansa  07master * r22d7cfdb56 10openembedded.git/recipes/openmoko-3rdparty/iliwi_git.bb: Jan 05 09:35:31 iliwi: bump SRCREV and add libeflvala to DEPENDS Jan 05 09:35:31 Signed-off-by: Martin Jansa Jan 05 09:41:09 03Martin Jansa  07master * r7649c68030 10openembedded.git/recipes/json-glib/json-glib_0.10.4.bb: (log message trimmed) Jan 05 09:41:09 json-glib_0.10.4: disable introspection for now Jan 05 09:41:09 * otherwise fails to build: Jan 05 09:41:09 | make[4]: Entering directory `/home/shr/shr-unstable/tmp/work/armv6-novfp-oe-linux-gnueabi/json-glib-0.10.4-r0/json-glib-0.10.4/json-glib' Jan 05 09:41:09 | GEN Json-1.0.gir Jan 05 09:41:10 | g-ir-scanner: compile: arm-oe-linux-gnueabi-gcc -march=armv6j Jan 05 09:41:11 -mtune=arm1136jf-s -mthumb-interwork -mno-thumb -pthread Jan 05 10:50:13 03Martin Jansa  07master * r34bbadba5f 10openembedded.git/recipes/linux/ (4 files in 3 dirs): Jan 05 10:50:13 linux(-kexecboot): add 2.6.37 with spitz defconfigs Jan 05 10:50:13 Signed-off-by: Martin Jansa Jan 05 11:10:45 hi xora Jan 05 11:40:21 03Koen Kooi  07org.openembedded.dev * rffdd614909 10openembedded.git/recipes/linux/linux-omap4/omap4430-panda/defconfig: Jan 05 11:40:21 linux-omap4: disable L2 PLE access for panda Jan 05 11:40:21 Signed-off-by: Koen Kooi Jan 05 12:15:32 hi obi Jan 05 12:15:58 hi woglinde! Jan 05 12:22:07 hm damn what was the push command Jan 05 12:22:08 again Jan 05 12:23:52 git push? :-) Jan 05 12:24:49 hm have it Jan 05 12:24:57 that would push all branches Jan 05 12:25:00 dont want that Jan 05 12:25:04 but I found it Jan 05 12:25:11 -n helps to make nothing break Jan 05 12:25:24 I hope Jan 05 12:27:32 you can set "git config push.default current", to always push the current branch only as default Jan 05 12:27:46 yes Jan 05 12:29:24 03Henning Heinold  07org.openembedded.dev * r130843b39a 10openembedded.git/ (4 files in 4 dirs): linux-ac100: support for toshiba AC100/dynabook Jan 05 12:36:44 http://marcin.juszkiewicz.com.pl/2011/01/05/2010-timeline/ Jan 05 12:37:23 As year changes I wrote my usual summary of past year: http://marcin.juszkiewicz.com.pl/2011/01/05/2010-timeline/ Jan 05 12:54:03 hrw: an interesting year :) Jan 05 12:56:13 indeed Jan 05 12:57:31 03Koen Kooi  07org.openembedded.dev * rff1cf74a1d 10openembedded.git/recipes/ti/ (ti-dmai.inc ti-dmai_svn.bb): Jan 05 12:57:31 ti-dmai: fix build with recent kernel headers and recent toolchains Jan 05 12:57:31 Signed-off-by: Koen Kooi Jan 05 13:04:07 03Koen Kooi  07org.openembedded.dev * r89ed3ea727 10openembedded.git/recipes/ti/ (3 files): Jan 05 13:04:07 ti-xgxperf: move SRCREV to common include and bump to r79 Jan 05 13:04:07 Signed-off-by: Koen Kooi Jan 05 13:43:02 gm Jan 05 13:43:14 hi likeweise Jan 05 13:58:12 greetings and salutations Jan 05 14:04:16 03Henning Heinold  07org.openembedded.dev * r4aaf290289 10openembedded.git/conf/machine/include/tune-cortexa9tegra.inc: tegra2: change feed and basearch to armv7a-vfp as suggested by koen Jan 05 14:14:42 gm kergoth Jan 05 14:14:45 hey Jan 05 14:19:33 03Koen Kooi  07org.openembedded.dev * r2bb663abd4 10openembedded.git/recipes/ti/gstreamer-ti.inc: Jan 05 14:19:33 gstreamer-ti: fix build with recent toolchains Jan 05 14:19:33 Signed-off-by: Koen Kooi Jan 05 14:39:48 gm Jan 05 14:42:50 oh god, http://thedailywtf.com/Articles/XMLd-XML.aspx Jan 05 14:45:56 hehe! How to take something unwieldy and difficult to deal with, and make it worse... Jan 05 14:46:10 * kergoth_ shudders Jan 05 14:49:26 Are we missing a layer? MACHINE=nslu2le DISTRO=slugos bitbake slugos-image <--- but MACHINE really is ixp4xx, and nslu2le is more of a platform. Jan 05 14:50:07 dsmg600 is another (similar) platform containing the ixp4xx chipset, but it needs different drivers in the image. Jan 05 14:50:41 so, do I create two different images? or do I make the machine files do the work to pull in the correct drivers? Jan 05 14:53:05 mwester hm look at it Jan 05 14:58:41 anybody knows what is libQtGuiE.so.4 ? Jan 05 14:59:22 sound like a qte lib but is not generated by qt4embedded Jan 05 15:00:18 i know Jan 05 15:00:20 :) Jan 05 15:00:35 It's a lib built in alls qt versions Jan 05 15:00:50 includes basics headers,functions,etc.. about Qt Jan 05 15:01:08 it changes names on some versions Jan 05 15:02:58 mwester: sounds like a machine specific detail to me fwiw... Jan 05 15:12:07 RP, thanks. I'll work on fixing slugos-image.bb to move some of that into the machine then. Jan 05 16:16:28 03Martin Jansa  07master * r81ee5dffa0 10openembedded.git/conf/distro/shr.conf: Jan 05 16:16:28 shr: blacklist libnl Jan 05 16:16:28 Signed-off-by: Martin Jansa Jan 05 16:16:41 03Martin Jansa  07master * rb117b871f4 10openembedded.git/recipes/bluez/bluez4_4.84.bb: (log message trimmed) Jan 05 16:16:41 bluez4: add 4.84 with DP -1 (except SHR) and without libnl dependency Jan 05 16:16:41 * libnl1 and libnl2 create conflicts, fso needs libnl2 so we try keep images without libnl1 depenency Jan 05 16:16:41 * 16:18:27 < mickeyl> commit ecb835fca7e7134f38a3d061d186eb38d83bcb8b Jan 05 16:16:42 16:18:27 < mickeyl> Author: Marcel Holtmann Jan 05 16:16:42 16:18:27 < mickeyl> Date: Wed Oct 6 10:21:11 2010 +0200 Jan 05 16:16:43 16:18:27 < mickeyl> Remove netlink plugin Jan 05 16:17:00 03Martin Jansa  07master * r61b3a0cd21 10openembedded.git/recipes/bison/ (bison-2.4.3/m4.patch bison_2.4.3.bb): Jan 05 16:17:00 bison: add 2.4.3 version needed for libnl2-2.0 to build ok Jan 05 16:17:00 Signed-off-by: Martin Jansa Jan 05 16:17:17 Anyone else seeing things like this: Jan 05 16:17:25 03Martin Jansa  07master * r448e4dcbef 10openembedded.git/recipes/freesmartphone/cornucopia.inc: Jan 05 16:17:25 freesmartphone: bump cornucopia SRCREV again to force upgrade after libnl2 rebuild Jan 05 16:17:25 Signed-off-by: Martin Jansa Jan 05 16:17:34 03Martin Jansa  07master * r86545d9bba 10openembedded.git/recipes/shr/libphone-ui-shr_git.bb: Jan 05 16:17:34 libphone-ui-shr: bump PR to relink to libnl2 again Jan 05 16:17:34 Signed-off-by: Martin Jansa Jan 05 16:17:45 03Martin Jansa  07master * rc262a784b9 10openembedded.git/recipes/shr/phoneuid_git.bb: Jan 05 16:17:45 phoneuid: bump PR to relink to libnl2 again Jan 05 16:17:45 Signed-off-by: Martin Jansa Jan 05 16:17:47 http://pastebin.com/1mq4112U Jan 05 16:17:56 03Martin Jansa  07master * rfb68e7c6d9 10openembedded.git/recipes/libnl/ (libnl2/fix-pc-file.patch libnl2_2.0.bb): Jan 05 16:17:56 libnl2: add recipe for released version 2.0 Jan 05 16:17:56 Signed-off-by: Martin Jansa Jan 05 16:31:42 hmm Jan 05 16:31:53 indeed Jan 05 16:32:03 ? Jan 05 16:32:34 he pb hny Jan 05 16:33:06 hi woglinde, same to you Jan 05 16:33:43 * kergoth_ is trying to figure out the cleanest way to add tracking of filename/lineno from parsing into flags in the metadata Jan 05 17:00:27 the issue i'm hitting is, it's the parser that has filename/lineno info, and i pass those into the ast nodes.. but its the metadata class that applies _append/_prepend, and it doesn't have filename/lineno information. I'm thinking maybe we should kill the logic in setVar that rips apart the variable name and does the append/prepend bits, and instead have another method which adds one of these lazy operations more explicitly Jan 05 17:00:43 this means you wouldn't want to d.setVar('foo_append'), but that'd be idiotic anyway, no big loss imo Jan 05 17:02:22 cannt really follow you Jan 05 17:02:55 currently, what happens is, you do d.setVar('somevariable_append', ' bar'), and it rips apart that variable name and stores what's to be appended out of band Jan 05 17:03:00 so it can apply that at finalize time Jan 05 17:03:12 rather than adding the append being a separate method Jan 05 17:03:29 and filename/lineno isn't passed into setVar, so it can't store that when it sets it aside Jan 05 17:03:39 guess i could pass filename/lineno into setvar, but that's pretty ugly Jan 05 17:03:42 hmm Jan 05 17:05:19 thanks for the sms marchin Jan 05 17:05:24 ups Jan 05 17:05:25 hrw Jan 05 17:10:23 kergoth: bitbake doesn't like RDEPENDS_${PN} := "${@oe_filter_out('^(hicolor-icon-theme)$', '${RDEPENDS_${PN}}', d)}", because of the nested variables. is there an alternative way to remove somthing from such a variable, instead of hardcoding it? Jan 05 17:11:31 changing the '${RDEPENDS_${PN}}' to a d.getVar may work Jan 05 17:11:38 or you could save it in another variable and get it from there Jan 05 17:11:51 e.g. ORIG_RDEPS := "${RDEPENDS_${PN}}", maybe Jan 05 17:11:56 unsure on that Jan 05 17:12:15 the issue may be exactly when variable references within variable names are expanded Jan 05 17:12:24 or expand is being stupid Jan 05 17:13:07 ok, i'll try both Jan 05 17:14:01 what is it actually complaining about? Jan 05 17:14:25 good question Jan 05 17:16:28 http://pastebin.com/BBuf7csz Jan 05 17:18:10 oh, right. that is a bit weird. Jan 05 17:18:25 hmm Jan 05 17:21:36 i've seen that error once before, when not setting MACHINE. it freaks expanding OVERRIDES rather than getting to the sanity check Jan 05 17:24:22 hm, it does look to me as though the way __expand_python_regexp__ is used in data_smart is a bit too simple minded Jan 05 17:24:53 but, you'd think, if that was just totally broken then someone would have noticed it long ago Jan 05 17:25:27 lets see if some cake is left over Jan 05 17:25:58 yeah, you'd think so. of course, i'm not sure how common nested expansions are Jan 05 17:34:06 hmm, i wonder. if i use compile() to compile a string into an ast, and then again to compile the ast into a normal executable object, which of those two compile executions needs the dont_inherit argument Jan 05 17:34:12 * kergoth_ tests Jan 05 17:42:13 03Dirk Opfer  07org.openembedded.dev * r8220739f03 10openembedded.git/recipes/eibd/ (4 files in 2 dirs): Jan 05 17:42:13 eibd: Add recipe for git version based on pu branch Jan 05 17:42:13 - Improved packaging, Sourcecode of examples are packaged separately Jan 05 17:42:14 03Dirk Opfer  07org.openembedded.dev * r9372d5698a 10openembedded.git/recipes/liboauth/liboauth_0.8.8.bb: liboauth: add dependency to curl Jan 05 17:42:17 03Dirk Opfer  07org.openembedded.dev * r951c7b7f0d 10openembedded.git/recipes/owfs/owfs_2.8p0.bb: owfs: 2.8p0 -> 2.8p4 Jan 05 17:43:29 03Philip Balister  07org.openembedded.dev * r004275c05b 10openembedded.git/recipes/less/less_436.bb: Jan 05 17:43:29 less : Update to version 436. Jan 05 17:43:29 The older version went runbound when you tried to search for a string. This Jan 05 17:43:29 version corrects that problem. Jan 05 17:43:29 humm Jan 05 17:43:43 ? Jan 05 17:44:21 I'd prefer if busybox links were done during packaging and not postinst. Any reason to do that on postinst? Jan 05 17:44:31 not sure Jan 05 17:44:36 kergoth_, btw.. lineno info from better_exec'ed functions are a bit unintuitive imho. It would be neat if the lineno of the original spot would be mentioned somewhere so i don't have to calculate it myself :) Jan 05 17:44:44 I thought you might have humm'd about my commit :) Jan 05 17:44:53 Crofton|work: ahahha Jan 05 17:44:54 otavio, er, might that get a bit complex w/ the ones that are or are not update-alternative'able Jan 05 17:44:58 or needed, anyhow Jan 05 17:45:28 Tartarus: right but there're some that we ought to have this way. Specially the ones we can have packages for Jan 05 17:45:36 Tartarus: as syslog, udhcpc, linuxrc and like Jan 05 17:46:01 Tartarus: for example, now I am fighting with it since I have no easy way to not provide linuxrc Jan 05 17:46:28 So there's a good case for opt-in'ing a few of the links then, yeah Jan 05 17:46:46 or making it easier for distros to just provide busybox config fragments Jan 05 17:46:59 (or is this a board not distro problem you have?) Jan 05 17:47:03 blindvt`: that's a big part of why i'm doing this Jan 05 17:47:49 oh. heh, great! Jan 05 17:48:08 blindvt`: check out https://gist.github.com/743677 Jan 05 17:48:59 my intent is to use something like that. the line numbers will be right, the filenames will be right, and the function names in the traceback will match the function name in the recipe. no need for a mangled anonymous python name if we retain the filename/lineno with it, it can just be "anonymous" Jan 05 17:49:25 03Koen Kooi  07org.openembedded.dev * rbad92062d3 10openembedded.git/contrib/angstrom/sort.sh: Jan 05 17:49:25 angstrom feed sorter: add support for NEON-less armv7a SoCs Jan 05 17:49:25 Signed-off-by: Koen Kooi Jan 05 17:49:32 03Koen Kooi  07org.openembedded.dev * rd419feeaf2 10openembedded.git/conf/distro/include/angstrom.inc: Jan 05 17:49:32 angstrom: use floating point for tegra2 Jan 05 17:49:32 Signed-off-by: Koen Kooi Jan 05 17:49:57 blindvt`: the nice thing is we dont' have to do as much mangling of the traceback itself, since it will already contain correct numbers Jan 05 17:50:04 we can focus on just adding more context to it Jan 05 17:51:46 kergoth_, sounds good Jan 05 17:52:23 i figured we might want to drop the event handler string compilation crap from bb.event. just register a *function* with it, and let the ast use this to create the function and register it Jan 05 17:56:40 something similar would work for both methodpool functions and regular python tasks. could compile them at parse time and set them aside for later execution Jan 05 17:56:50 hmm Jan 05 17:57:15 either that or just make exec_task/exec_func use location info from a variable flag Jan 05 17:57:24 that info needs adding either way Jan 05 17:59:01 kergoth_, if i find the time during the weekend i mean to see if we can (for certain funcs that are identical in different recipes) use just one run.do_something file instead of per recipe. Should make writing out recipes data faster on my poor old machine Jan 05 17:59:43 (and also waste less space) Jan 05 17:59:59 blindvt`: https://github.com/kergoth/bitbake/compare/master...build-shell-stdin - passes the shell script to the shell we spawn via stdin, and only writes the run scripts when debugging is enabled or the execution failed Jan 05 18:00:05 ;) Jan 05 18:00:16 erm, bad comparison on github Jan 05 18:00:19 look at the commit list instead Jan 05 18:01:55 the main motivation there is due to poky-sync. it adds all exported vars to the env of the worker processes we run, so there's no need to emit them to the shell script at all -- we do want them in the run. script, so it can be sourced, but for the task execution, its not needed Jan 05 18:02:40 * kergoth_ quite likes https://github.com/kergoth/bitbake/commit/310470 too, makes -l Build / -D slightly better Jan 05 18:02:56 yes, that's also a good idea. But still, if not going directly to stdin we should remember the output and just make a symlink to a previously emitted function instead of generating it anew, imho Jan 05 18:03:41 the poky-sync branch helps slightly, in that it only emits the shell functions which are run by the function we're running Jan 05 18:03:49 but yes, that could be useful too :) Jan 05 18:04:22 (there's clearly no shortage of improvements to make ;) Jan 05 18:04:38 kergoth_, common sh setup stanza could also be sourced in individual recipes. Not sure what poky does, but bitbake master leaves quite some room for improvement in this area AFAICS Jan 05 18:05:22 anyway. We'll see if i can spare some time to play with it. Have to run now. Jan 05 18:05:26 well, like i said, in poky the exported vars get into the actual process environment for the worker process, so there's no way to share that piece Jan 05 18:05:26 * blindvt` cheers & Jan 05 18:05:33 later Jan 05 18:06:53 03Simon Busch  07org.openembedded.dev * r7ffc76fcd2 10openembedded.git/conf/machine/palmpre.conf: Jan 05 18:06:53 palmpre machine: switch to .tar.gz format for rootfs Jan 05 18:06:53 Signed-off-by: Simon Busch Jan 05 18:10:23 kergoth: d.getVar returned None, although the variable name was expanded. then i tried again with hardcoded $PN, but even that failed with the same error message as i pasted before Jan 05 18:11:36 * Tartarus sets up a torturiously long path to build in to confirm he finally fixed his klcc issue (related to direct path to perl) Jan 05 18:13:03 gm Jan 05 18:14:20 kergoth: this one is ok (besides being semantically incorrect): RDEPENDS_libsoup-2.4 := "${@oe_filter_out('^(hicolor-icon-theme)$', '${RDEPENDS}', d)}" Jan 05 18:14:37 kergoth: but this one fails with a syntax error: RDEPENDS_libsoup-2.4 := "${@oe_filter_out('^(hicolor-icon-theme)$', '${RDEPENDS_foo}', d)}" Jan 05 18:45:09 obi: strange. Jan 05 18:45:24 obi: would you mind emailing that to the bitbake-dev list, for reference? Jan 05 18:49:26 03Chris Larson  07master * r0b11a3d4ea 10bitbake.git/lib/bb/parse/ast.py: Jan 05 18:49:26 parse.ast: avoid code duplication for inherit Jan 05 18:49:26 Signed-off-by: Chris Larson Jan 05 18:49:34 03Chris Larson  07master * rc8ff0fd3e9 10bitbake.git/lib/bb/event.py: Jan 05 18:49:34 event: use BBLogFormatter in print_ui_queue Jan 05 18:49:34 This ensures that when a failure occurs very early on in bitbake startup, the Jan 05 18:49:34 message formatting ematches that used by the UIs. Jan 05 18:49:35 Signed-off-by: Chris Larson Jan 05 18:49:37 03Chris Larson  07master * r9615c538b8 10bitbake.git/lib/bb/parse/ (ast.py parse_py/BBHandler.py parse_py/ConfHandler.py): Jan 05 18:49:37 parse: pass filename, lineno into the ast Jan 05 18:49:37 We will be needing this information to improve the tracebacks of python code Jan 05 18:49:37 from the metadata, as well as to give the user information about where Jan 05 18:49:37 variables were defined, so they know how it ended up the way it is. Jan 05 18:49:38 Signed-off-by: Chris Larson Jan 05 18:49:46 03Chris Larson  07master * r01f582992f 10bitbake.git/lib/bb/process.py: Jan 05 18:49:46 process: simplify the log option for run Jan 05 18:49:46 Signed-off-by: Chris Larson Jan 05 18:49:47 03Chris Larson  07master * raa9646717b 10bitbake.git/lib/bb/parse/ast.py: Jan 05 18:49:47 parse.ast: drop __word__ regular expression Jan 05 18:49:47 We can use the string split method for this instead. Jan 05 18:49:48 Signed-off-by: Chris Larson Jan 05 18:49:49 03Chris Larson  07master * rdc22643876 10bitbake.git/lib/bb/build.py: Jan 05 18:49:49 build: use logger.isEnabledFor, not getEffectiveLevel Jan 05 18:49:49 I think this is less confusing, and avoids needing to know about the *range* Jan 05 18:49:50 of logging levels, instead simply asking what we really want to know. Jan 05 18:49:50 Signed-off-by: Chris Larson Jan 05 18:50:02 03Koen Kooi  07org.openembedded.dev * r2915ea2ab1 10openembedded.git/recipes/linux/linux-ac100_git.bb: Jan 05 18:50:03 linux-ac100: switch to fetching from git, less flaky Jan 05 18:50:03 19:01 < woglinde> yeah feel free to commit Jan 05 18:50:03 Signed-off-by: Koen Kooi Jan 05 18:51:55 03Simon Busch  07org.openembedded.dev * r443c760bb1 10openembedded.git/recipes/compat-wireless/ (2 files in 2 dirs): Jan 05 18:51:55 compat-wireless-2.6-old: add initial recipe Jan 05 18:51:55 compat-wireless-2.6-old provides more recent wireless drivers than kernels <= 2.6.26 Jan 05 18:51:55 provides on their own. On some machines you are stucked to that quite old kernel version Jan 05 18:51:55 but need a more recent wireless driver to get wifi working. Jan 05 18:51:55 Acked-by: Michael 'Mickey' Lauer Jan 05 18:51:56 Signed-off-by: Simon Busch Jan 05 18:52:07 03Simon Busch  07org.openembedded.dev * rc48983a2ef 10openembedded.git/recipes/compat-wireless/ (4 files in 2 dirs): Jan 05 18:52:07 compat-wireless-2.6-old: make recipe even support the palmpre machine Jan 05 18:52:07 To support compat-wireless-2.6-old for the palmpre machine we need several patches to let Jan 05 18:52:07 it build with the modified 2.6.24 based kernel. Jan 05 18:52:07 Acked-by: Michael 'Mickey' Lauer Jan 05 18:52:07 Signed-off-by: Simon Busch Jan 05 19:46:09 mwester: around ? Jan 05 19:46:24 mwester: how does slugos changes look like ? Jan 05 19:46:37 mwester: I have built a minimal-image successfully with this Jan 05 19:48:31 hrm Jan 05 19:54:15 mwester: and chrooted into the new image too Jan 05 20:00:19 On boot, I get a bunch of errors around nbdxx like "nbd12: Attempted send on closed socket end_request: I/O error, dev nbd12, sector 0" Jan 05 20:00:24 does anyone know what the source of this is? Jan 05 20:00:51 i have done some googling and there are fragments of people asking about this issue but i have seen no answer Jan 05 20:01:24 I also get a bunch of "FAT: unable to read boot sector" errors even though there is no FAT filesystem anywhere on the system Jan 05 20:15:45 martinmeba: seems like your kernel is loading nbd Jan 05 20:15:56 martinmeba: if you dont need it then disable it Jan 05 20:16:33 mickeyl, ping Jan 05 20:19:43 ok - thanks Jan 05 20:19:52 I was looking through init scripts going ??? Jan 05 20:37:37 khem: thank you very much - you were exactly correct. Jan 05 22:20:49 calling it a day, cya Jan 05 22:21:02 nite effem Jan 05 22:48:58 hi, python-numeric is failing to compile here Jan 05 22:49:09 I was compiling shr-lite-image from scratch Jan 05 22:49:25 /usr/lib/libgcc_s.so.1: file not recognized: File format not recognized Jan 05 22:51:25 yeah host lib Jan 05 22:53:04 woglinde, any idea on what do I need to do? Jan 05 22:55:11 pespin: I've had that issue a couple times, it's libtool's fault Jan 05 22:56:22 vadmium, how did you solve it? using different libtool version? Jan 05 22:56:24 See http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-December/027868.html Jan 05 22:57:03 basically get rid of the add_dir="-L$libdir" line in whatever copy of Libtool it is you're using Jan 05 22:58:13 I'm assuming it is the same problem as mine, you could tell because libtool inserts an -L/usr/lib to gcc during a "relink" phase Jan 05 23:01:14 vadmium, yes, you are right Jan 05 23:01:30 there a -L/usr/lib Jan 05 23:01:48 control config.log Jan 05 23:01:59 where it might come from Jan 05 23:02:14 hi woglinde Jan 05 23:02:21 jo ant Jan 05 23:02:32 ah my birthday is over Jan 05 23:02:48 hm configure should use readelf Jan 05 23:03:23 to find out if the lib which the macro is searching is right for the target-arch Jan 05 23:03:27 woglinde, uhm there's no log.do_configure in python-numeric Jan 05 23:04:10 uhm actually there's only a log for do_compile Jan 05 23:04:29 pesping ah right Jan 05 23:05:01 sorry to tired to look at the recipe Jan 05 23:06:27 vadmium, uhmm that add_dir="-L$libdir" should be in ltmain.sh right? Jan 05 23:06:44 possibly Jan 05 23:06:58 there's no ltmain.sh here Jan 05 23:07:07 in python-numeric Jan 05 23:07:15 I think there are different names and stuff in different versions of libtool Jan 05 23:07:53 maybe it's a libtool in sysroot that it's using instead of in the python numeric package? I dunno anything about that package Jan 05 23:07:55 and grep -r ad_dir * doesn't show anything Jan 05 23:08:23 Tartarus: ping Jan 05 23:09:41 pespin can you see what libtool binary it's running from your log file? Jan 05 23:11:02 vadmium, here's the logfile http://paste.pocoo.org/show/315574/ Jan 05 23:11:56 hmm, maybe it's not libtool after all Jan 05 23:13:46 But either way, I think you have to figure out why that -L/usr/lib is getting there. I don't think it has any business there when you're cross compiling Jan 05 23:25:36 vadmium, grep showed me a reference to /usr/lib here: Jan 05 23:25:37 Packages/dotblas/setup.py:blas_dirs_list = ['/usr/lib/'] Jan 05 23:27:33 vadmium, should I try setting it to sysroots/armv4t-blabla/usr/lib ? Jan 05 23:28:42 dunno, i'm not really an expert on python packages Jan 05 23:29:28 me neither Jan 06 00:08:34 anyone around? Jan 06 00:08:49 I am trying to figure out what to add to my recipe for it to install dsplink, lpm, cmem to the image Jan 06 00:08:56 but for some reason, they arent getting installed Jan 06 00:19:59 03Tom Rini  07master * r676cbb54d4 10openembedded.git/recipes/klibc/ (klcc-cross_1.5.20.bb klibc-1.5.20/use-env-for-perl.patch): Jan 06 00:19:59 klcc-cross: Add patch to use /usr/bin/env perl Jan 06 00:19:59 Certain configurations (such as autobuilders) may build in very Jan 06 00:19:59 deep paths (that are longer than the #! mechanism allows) which Jan 06 00:19:59 makes it unsafe to use the direct path for perl. In our case we know Jan 06 00:19:59 that /usr/bin/env perl will always return ours (if it has been built). Jan 06 00:20:00 Signed-off-by: Tom Rini Jan 06 00:20:16 Tartarus: we need to talk about klcc (the one in klibc-dev)... I need to refer to the host Perl and not to sysroot Jan 06 00:20:38 gn **** ENDING LOGGING AT Thu Jan 06 02:59:58 2011