**** BEGIN LOGGING AT Mon Oct 04 02:59:57 2010 Oct 04 05:04:54 NAMES Oct 04 05:05:55 hi Aarti , GM :) Oct 04 05:06:22 harsh, hi,GM :) Oct 04 05:07:12 Aarti, are you going to prepare the OE paper today ? Oct 04 05:07:30 harsh,yup Oct 04 05:07:41 ok Oct 04 05:22:22 NAMES Oct 04 05:29:19 NAMES Oct 04 06:01:55 NAMES #oe Oct 04 07:27:30 gm Oct 04 07:28:05 gn :D Oct 04 07:28:14 :-) Oct 04 07:30:14 good morning Oct 04 07:42:24 03Koen Kooi  07org.openembedded.dev * r0849a1dca4 10openembedded.git/recipes/v4l2apps/ (4 files): Oct 04 07:42:24 libv4l: add 0.8.1, delete old version Oct 04 07:42:24 v4l-utils: add combined libs/utils recipe for 0.8.1, keep old packages split for compatibility Oct 04 07:47:02 morning Oct 04 07:49:27 hi hrw Oct 04 08:06:50 hm, how one should report eglibc bugs? http://eglibc.org/issues/ seems dead Oct 04 08:13:55 filip: looks like you're right eglibc-initial_2.12.bb' failed Oct 04 08:15:25 mckoan: I am having: | ../ports/sysdeps/unix/sysv/linux/arm/eabi/libc-do-syscall.S:28: Error: selected processor does not support THUMB opcodes Oct 04 08:16:00 mckoan: which is introduced by http://repo.or.cz/w/glibc-ports.git/commitdiff/2dcd7ed8fd149a5aefc8c7a7ff041ff389348b4f Oct 04 08:16:43 filip: for which MACHINE are you building it? Oct 04 08:19:01 moin. i get the following error if i try to build openssl-native: http://pastebin.com/QvkgWj02 Oct 04 08:19:38 a touch Makefile helps, but this is a little bit "hackish" Oct 04 08:23:45 JaMa: Jornada 720 Oct 04 08:23:49 JaMa: SA1110 CPU Oct 04 08:25:05 JaMa: jornada7xx.conf Oct 04 08:47:48 * mckoan needs to create a new personal branch Oct 04 08:50:18 * mckoan is is tired of chasing problems generated by patches Oct 04 08:55:53 filip: sa1110 does not support thumb indeed Oct 04 09:01:43 hrw: yes, but this git commit introduces thumb instruction regardless of the ARM cpu family Oct 04 09:05:02 filip: thats 2.12 eglibc or newer? Oct 04 09:18:07 hrw: 2.12 Oct 04 09:21:34 filip: check was it fixed in git tree or start writing own fix. Oct 04 09:21:47 filip: other option would be using 2.11 or older eglibc Oct 04 09:31:00 in fact ansgtrom avoided this pain setting PREFERRED_VERSION_eglibc ?= "2.11" Oct 04 09:34:22 at first glance, compilation of the libc-do-syscall.S needs only to be conditionalized (e.g. by a #if defined(__thumb__) block) as it is called from inside of a #if defined(__thumb__) block only Oct 04 12:03:47 03Pierluigi Passaro  07org.openembedded.dev * rc31636c6a9 10openembedded.git/recipes/mysql/mysql5_5.1.40.inc: Oct 04 12:03:48 mysql5: adopt to new style sysroot Oct 04 12:03:48 Signed-off-by: Koen Kooi Oct 04 13:22:43 03Marco Cavallini  07org.openembedded.dev * rf6b92a2ef1 10openembedded.git/conf/distro/kaeilos-2010.conf: Oct 04 13:22:43 kaeilos-2010.conf: set PREFERRED_VERSION_eglibc to 2.11 Oct 04 13:22:43 * bump PR Oct 04 13:41:03 cbrake: if the testing branch is not made can you hold off for a short while Oct 04 13:41:08 have a fix that needs to go into it Oct 04 13:51:50 03Frans Meulenbroeks  07org.openembedded.dev * r98037f1062 10openembedded.git/classes/kernel.bbclass: (log message trimmed) Oct 04 13:51:50 kernel.bbclass: set kerneldir Oct 04 13:51:50 commit b6cc4bc217e695801db060a1366915a870cab5db moved work Oct 04 13:51:50 from kernel_do_install to sysroot_stage_all_append Oct 04 13:51:50 This change changed kerneldir and staged directy from Oct 04 13:51:51 ${SYSROOT_DESTDIR}${STAGING_KERNEL_DIR}. This works fine but Oct 04 13:51:52 the patch removed the line kerneldir=${D}/kernel/ Oct 04 13:53:02 morrning.... I have a quick questions do variables like THUMB_INTERWORK = "no" work from the local config override distro configurations Oct 04 14:06:56 cargoudel: The easy way to test is bitbake -e some-recipe | grep ^THUMB_INTERWORK Oct 04 14:07:43 thanks ... even though I set my my arch to armv4 .... no thumb bx intructions .... that flag by bx's in all my libraries Oct 04 14:07:54 by = put* Oct 04 14:08:31 so also .. how do I rebuild all non-native packages Oct 04 14:09:16 03Koen Kooi  07org.openembedded.dev * r2eee26fdda 10openembedded.git/classes/module_strip.bbclass: module_strip bbclass: inherit class to get the kernel version method Oct 04 14:09:48 cbrake: that was the patch I wanted to get through Oct 04 14:14:45 (the one on kernel.bbclass, not the one from koen) Oct 04 14:17:11 isn't adding gstreamer-ti to IMAGE_INSTALL all that's needed to get the ti gstreamer plugins? Oct 04 14:35:05 03Tom Rini  07org.openembedded.dev * r6e5fb2f553 10openembedded.git/recipes/ipsec-tools/ipsec-tools.inc: Oct 04 14:35:05 ipsec-tools: Fix build with gcc 4.4+ Oct 04 14:35:05 Signed-off-by: Tom Rini Oct 04 14:35:17 03Tom Rini  07org.openembedded.dev * rfa008da10d 10openembedded.git/recipes/bonnie/bonnie++.inc: Oct 04 14:35:17 Bonnie++: One more PN -> BPN Oct 04 14:35:17 Signed-off-by: Tom Rini Oct 04 14:37:36 seems the answer is no, but adding task-gstreamer-ti helped, though I don't know why.. Oct 04 14:41:58 what file in oe would have a list of a recipes installed so far Oct 04 14:51:44 "installed? Oct 04 14:51:54 recipes aren't installed Oct 04 14:52:08 true Oct 04 14:52:26 i mean compiled and merged into my sysroot Oct 04 14:52:28 and installed *where*? Oct 04 14:52:43 okay, that makes it more clear Oct 04 14:52:51 given there are multiple places things get installed (e.g. images..) Oct 04 14:53:14 look at the tmp/ dir Oct 04 14:53:25 you should see either opkg/ipkg or usr/ or lib/ or something.. i don't remember the exact path Oct 04 14:53:33 but nowadays sysroot is managed by opkg with packaged staging Oct 04 14:53:45 so there's actually a package manager database for it Oct 04 14:53:54 it installs the packages into TMPDIR as though TMPDIR was / Oct 04 14:54:06 basically all my arm stuff got compiled with thumb_interwork (bx instructions) .... i want to keep the native x86 stuff and rebuild my arm stuff Oct 04 14:54:10 you can look at the status file to see what's been installed Oct 04 14:54:32 oh, i'd probably just list the stamps dir and grep out the natives, and clean what's left with -c clean Oct 04 14:54:37 tmp/stamps/ Oct 04 14:54:44 ahh great cool Oct 04 14:54:56 stamps are how bitbake knows what tasks its completed, so they don't need to be rerun Oct 04 14:55:02 so there's one per task it's finished Oct 04 14:55:09 thats easy Oct 04 14:57:31 thanks Oct 04 14:57:57 np Oct 04 14:58:56 i did an objdump and saw all that thumb_interwork .. was like noo....... cuz i have a armv4 processor and that gives a instruction exception Oct 04 15:00:23 i didn't notice some of the angstrom variables that must be set to not get thumb code don't check whether the architecture supports it before it sets it Oct 04 15:01:14 sorry for the typos ..... not get thumb code. It doesn't check* Oct 04 15:05:20 03Steffen Sledz  07org.openembedded.dev * r20ce3847a8 10openembedded.git/recipes/gnome/epiphany_2.30.2.bb: Oct 04 15:05:20 epiphany-2.30.2: unbreak build by adding ca-certificates dependencies Oct 04 15:05:20 Signed-off-by: Steffen Sledz Oct 04 15:05:20 Acked-by: Koen Kooi Oct 04 15:11:06 * kergoth_ kicks oe Oct 04 15:48:52 RP: ping Oct 04 16:01:39 zecke: ping Oct 04 16:06:10 kergoth_: hi Oct 04 16:06:35 wondering if you might know if this behavior is intentional Oct 04 16:06:49 bitbake forcibly checks RDEPENDS_${PN} even if PN isn't in PACKAGES Oct 04 16:07:20 which results in things like hicolor-icon-theme being in the runtime deps of native recipes, whcih in turn results in building a crosscompiler when you bitbake gtk+-native Oct 04 16:08:24 i'm wondering if it was by design that natives should be able to rdepend Oct 04 16:08:27 it seems pretty useless to me Oct 04 16:10:10 kergoth_: hmm.. yes, seems useles but I don't know if it was intentional, you would have to check with RP Oct 04 16:10:19 okay, figured, but thought i'd check Oct 04 16:10:33 it seems silly to have to go and _virtclass-native the rdepends/rprovides/etc Oct 04 16:10:52 there are other cases, gtk+-native is just one -- any time a recrdeptask task is run, it will obey those.. Oct 04 16:11:02 e.g. rm_work_all, package_stage_all, etc Oct 04 16:11:15 * kergoth_ spent a lot of quality time in .dot files this weekend :P Oct 04 16:11:26 thanks Oct 04 16:11:49 kergoth_: hmmm, git blame showed anything interesting on that code? Oct 04 16:11:58 good idea, will see Oct 04 16:12:07 itd be nice if our testing included growth of deps Oct 04 16:12:16 like capture the # of tasks from a regular build and note when it increases Oct 04 16:12:18 or something.. Oct 04 16:12:42 right now deps can grow and often we just don't notice Oct 04 16:13:31 anyway.. Oct 04 16:14:54 i hate trying to chase blame across code reorgs / renames / moves Oct 04 16:16:07 b03a424879367fd4118f60abda7f78bc5f27b957 added it Oct 04 16:16:17 prior to that it only checked packages, that made it check packages + [pn] Oct 04 16:16:27 april 15th, 2006.. Oct 04 16:16:32 heh Oct 04 16:34:24 03Tom Rini  07org.openembedded.dev * rb4c0cc2e7f 10openembedded.git/classes/insane.bbclass: Oct 04 16:34:24 insane.bbclass: Skip '.pc' in qa_configure Oct 04 16:34:24 The files in the .pc directory won't have anything useful in them and are Oct 04 16:34:24 often unreadable so lets just skip them. Oct 04 16:34:24 Signed-off-by: Tom Rini Oct 04 16:34:25 Acked-by: Khem Raj Oct 04 16:46:28 gm Oct 04 16:50:27 gm Oct 04 16:51:52 hey eFfeM, ka6sox Oct 04 16:52:10 morning kergoth_ Oct 04 16:52:42 eFfeM, I think we are down to the last lib on freenx..hopefully by tommorrow. Oct 04 16:53:23 ka6sox, cool Oct 04 16:53:46 interested to see what you changed, will try to give it a try too Oct 04 17:00:49 03Frans Meulenbroeks  07org.openembedded.dev * rc7e15931b0 10openembedded.git/recipes/sane/sane-backends-1.0.19/sane-plustek.patch: Oct 04 17:00:49 sane : moved unused files to obsolete dir Oct 04 17:00:49 Signed-off-by: Frans Meulenbroeks Oct 04 17:00:51 03Frans Meulenbroeks  07org.openembedded.dev * rc43ee6836f 10openembedded.git/recipes/sdcc/files/gcc4.patch: Oct 04 17:00:51 sdcc : moved unused files to obsolete dir Oct 04 17:00:51 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:11 03Frans Meulenbroeks  07org.openembedded.dev * rd139b31810 10openembedded.git/recipes/sdlperl/Makefile.patch: Oct 04 17:01:11 sdlperl : moved unused files to obsolete dir Oct 04 17:01:11 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:11 03Frans Meulenbroeks  07org.openembedded.dev * r99903dc5ce 10openembedded.git/recipes/socketcan/files/install-can_netlink.h.patch: Oct 04 17:01:11 socketcan : moved unused files to obsolete dir Oct 04 17:01:12 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:12 03Frans Meulenbroeks  07org.openembedded.dev * r5c44690dfc 10openembedded.git/recipes/sylpheed/files/streamline-ui.patch: Oct 04 17:01:16 sylpheed : moved unused files to obsolete dir Oct 04 17:01:16 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:16 03Frans Meulenbroeks  07org.openembedded.dev * r2e06fca314 10openembedded.git/recipes/strace/ (3 files in 2 dirs): Oct 04 17:01:16 strace : moved unused files to obsolete dir Oct 04 17:01:16 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:16 03Frans Meulenbroeks  07org.openembedded.dev * r8a66c84c7f 10openembedded.git/recipes/sysfsutils/sysfsutils-1.2.0/libsysfs-write-attribute.patch: Oct 04 17:01:16 sysfsutils : moved unused files to obsolete dir Oct 04 17:01:17 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:17 03Frans Meulenbroeks  07org.openembedded.dev * r3640339a41 10openembedded.git/recipes/swt/files/make_linux-fix.patch: Oct 04 17:01:18 swt : moved unused files to obsolete dir Oct 04 17:01:59 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:59 03Frans Meulenbroeks  07org.openembedded.dev * r3fe2c955d8 10openembedded.git/recipes/squashfs-tools/squashfs-tools-4.0/ (lzma-support.patch portability.patch): Oct 04 17:01:59 squashfs-tools : moved unused files to obsolete dir Oct 04 17:01:59 Signed-off-by: Frans Meulenbroeks Oct 04 17:01:59 03Frans Meulenbroeks  07org.openembedded.dev * r819adc6878 10openembedded.git/recipes/sqlite/sqlite3/ (cross-compile.patch ldflags.patch): Oct 04 17:01:59 sqlite : moved unused files to obsolete dir Oct 04 17:01:59 Signed-off-by: Frans Meulenbroeks Oct 04 17:04:39 eFfeM we will commit everything we had to change. Oct 04 17:07:46 ka6sox thanks Oct 04 17:07:50 03Frans Meulenbroeks  07org.openembedded.dev * rdda5eab30f 10openembedded.git/recipes/squashfs-tools/ (squashfs-tools-3.3/Makefile.patch squashfs-tools_3.3.bb): Oct 04 17:07:51 squashfs-tools_3.3: removed recipe Oct 04 17:07:51 doesn't patch and we have 4.0 for about a year Oct 04 17:07:51 not pinned, no DP Oct 04 17:07:51 Signed-off-by: Frans Meulenbroeks Oct 04 17:19:40 hey mickeyl Oct 04 17:23:58 hi eFfeM, how are things? Oct 04 17:25:00 fine thank you, and you ? Oct 04 17:25:36 will you be going to elc/oedem? Oct 04 17:26:11 good good. vacation was too short and i'm pretty busy, so that relaxing feeling is vanishing fast. I'm afraid I can't make either one, as I have some important meeting on 30th of Oct :/ Oct 04 17:26:27 oh pity Oct 04 17:26:40 * eFfeM still has a short vacation ahead :-) Oct 04 17:26:45 vacation is always too short Oct 04 17:28:04 right Oct 04 17:28:34 yeah, I'm sad not being able to make it, but I can't clone(2) myself Oct 04 17:30:25 BTW, the OEDEM is only for the core developers? Oct 04 17:30:50 ynezz: read http://wiki.openembedded.net/index.php/Oedem/2010 Oct 04 17:31:38 I've read it, and if I understand it correctly the friday is not, right? Oct 04 17:32:36 fri+sat Oct 04 17:33:06 I'm definitely should switch nationality to some european country Oct 04 17:33:35 anyone else having problems building anything? (sqlite3-native bails out for me) Oct 04 17:33:36 * Jay7 wish to have possibility to attend on any european conference Oct 04 17:34:06 i have no problems building other stuff haven't tried sqlite3-native today Oct 04 17:35:13 i've been having random build problems since friday, but i think its me that's cursed Oct 04 17:35:39 hm strange, what special happened that day? Oct 04 17:35:57 murphy, i suspect Oct 04 17:36:37 03Frans Meulenbroeks  07org.openembedded.dev * rd2c10e2989 10openembedded.git/recipes/mythtv/mythplugins_0.23+fixes.bb: Oct 04 17:36:37 mythplugins: fix include directive for lighty Oct 04 17:36:37 Signed-off-by: Frans Meulenbroeks Oct 04 17:36:53 well, as soon as cbrake makes a new testing branch I'll give things a stab again Oct 04 17:37:53 any opinions on the 1000 or so files that JaMa tried to fetch in april and that still do not fetch ? Oct 04 17:38:08 (or on the recipes that fail in patch, I've also bumped into a few of these) Oct 04 17:39:45 move 'em to broken Oct 04 17:40:27 for both were broken then and still broken now Oct 04 17:40:55 and expect a 5-10% of "oh, lemme go fix ... so it fetches" sever side things Oct 04 17:41:10 and maybe <5% "oh, lemme go point that over at ..." differnt URIs Oct 04 17:41:35 but 6 months of it being broken mean it needs some prodding to get fixed Oct 04 17:41:50 eh, if I move 1000 recipes to obsolete someone will explode ;-) Oct 04 17:41:57 * JaMa had already tried to find those 5% but found&fixed maybe 1% Oct 04 17:42:07 Was it really 1k tho? Oct 04 17:42:22 Maybe i'm just thinking checksum mismatch Oct 04 17:42:22 if its been broken that long, i doubt anyone is building it anyway, so they don't have much room to complain Oct 04 17:42:23 But even so Oct 04 17:42:25 heh Oct 04 17:42:41 JaMa's figure, I got something like 55 in a small test Oct 04 17:42:53 kergoth, how does image features work again? Oct 04 17:43:04 IMAGE_FEATUIRES += "dbg" ? Oct 04 17:43:41 Crofton|work, it's essentially just a package grouping mechanism. IMAGE_FEATURES is a space separated list of "features" which are groups of packages, and there are a few that are special, like dbg, as it sucks up a bunch of info from pkgdata to get all the packages dynamically Oct 04 17:43:43 yeah Oct 04 17:44:12 eFfeM: the diff is probably that I was downloadig really ALL recipes for ALL archs (in case some urls were behind machine override) Oct 04 17:44:18 kergoth_, I agree, but it is the same as with the obsolete files, instead of doing it in a weekend I was forced to split it over 4 months Oct 04 17:44:23 * kergoth` nods Oct 04 17:44:26 sad but true Oct 04 17:45:08 JaMa: agree, still the question remains how to tackle this Oct 04 17:45:19 eFfeM, I am getting a little tired of the constant complaining .... Oct 04 17:45:39 (and also the ones that do not patch, bumped e.g. into it for an old squashfs-tools version, so removed it) Oct 04 17:45:45 btw, new BBCLASSEXTEND-ed sqlite-native doesn't build... Oct 04 17:46:00 Crofton|work: well I suggest that you help cleaning up, the sooner that is done the sooner I stop complaining Oct 04 17:46:27 http://tinderbox.openembedded.org/public/logs/task/8357331.txt Oct 04 17:46:30 I really feel it is pointless to have lots of recipes around for which newer versions exist, which are broken and unmaintained Oct 04 17:46:30 eFfeM, I have actual work to do Oct 04 17:46:36 so have i Oct 04 17:46:51 well, obviously I have a lot more of it Oct 04 17:46:58 and meanwhile I do try to keep our tree from becoming a junkyard Oct 04 17:47:12 do you ? Oct 04 17:47:25 i like the idea of keeping the most recent minor version of each major version around (or most recent micro for a minor, depending on the versioning scheme and which corresponds to compatibility break) Oct 04 17:47:36 life is a matter of priorities :-) Oct 04 17:47:53 kergoth, the issue is keeping all distros up to date Oct 04 17:48:00 * kergoth_ hugs schroot Oct 04 17:48:02 * kergoth_ nods Oct 04 17:48:09 and impacting people doing out of tree work Oct 04 17:48:45 kergoth, i agree (and probably deprecate very old major versions every once in a while, unless there is a compelling reason to keep them (eg license chagne from gplv2 to v3) Oct 04 17:48:49 Crofton: amen! :) Oct 04 17:48:52 yeah, out of tree is a tough one, especially now that bbappend exists -- if they get their minor version bumped out from under them, their local changes are no longer being applied to what they're building. hopefully they lock everything down Oct 04 17:49:22 * eFfeM does out of tree work too, but we do not base them on dev head Oct 04 17:50:07 kergoth, see you are contributing to the problem :) Oct 04 17:50:15 :) Oct 04 17:51:28 hmm Oct 04 17:51:30 where the heck "i64" type definition comes from? Oct 04 17:51:46 maybe could do something in the metadata, at parse time, which uses persist_data to store a mapping of PN to bbappended state Oct 04 17:52:00 which alerts if the current recipe is the same PN as another recipe which is bbappended, but this one isn't Oct 04 17:52:02 or similar Oct 04 17:52:18 * kergoth` ponders Oct 04 17:52:41 well an append might also stop to work if the recipe itself gets an update Oct 04 17:53:23 eFfeM_work: thaks, I'll branch testing-next now Oct 04 17:53:59 cbrake: ok, i'll kick of a build after my coffee break Oct 04 17:54:08 I tried adding gstreamer-ti to my beagleboard earlier today. Adding gstreamer-ti to IMAGE_INSTALL gave me CMEM, DSPLINK ond probably more, but no gstreamer plugins. Adding task-gstreamer-ti gave me the plugins, and after looking at the two recipes I can't say I understand why. I sense this is a sign of something I really should understand to work with OE, so I call upon the expertise to enlighten me :) Oct 04 17:55:11 effeM: right, i'm just concerned that you may, say, be applying a local patch to an upstream recipe, and upstream bumps the default version, and all of a sudden you lost the patch you require -- maybe applying it to current requires manual work, so it shouldn't be applied automatically, but you should at least have a way to know if this occurs Oct 04 17:55:18 tasslehoff: guess gstreamer-ti has no DEPENDS or RDEPENDS on the plubins Oct 04 17:55:21 plugins Oct 04 17:56:27 kergoth true; a way around that would be automatically stash a copy of the recipe you append to somewhere and use that as a fallback or so Oct 04 17:56:58 eFfeM: that is it. by adding gstreamer-ti I get its ipk, right? I thought the plugins were inside that one, but that's probably wrong then. Oct 04 17:57:05 gstreamer-ti is a plugin on its own Oct 04 18:00:15 denix: but that's not where the codecs are? Oct 04 18:00:32 tasslehoff: no, codecs are separate Oct 04 18:01:52 denix: downloaded when build gstreamer-ti, and added to my image because I RDEPEND on gst-plugins-*-meta? Oct 04 18:02:35 tasslehoff: is that a question or a statement? Oct 04 18:03:13 denix: a question, therefore the ? ;) Oct 04 18:03:30 aha, i64 supposed to come from staged sqlite3.h, but mine is empty... Oct 04 18:04:52 trying to understand what I depend on to get the plugins, and why I only get them if I have added gstreamer-ti Oct 04 18:06:41 tasslehoff: check the task-gstreamer-ti, which DEPENDS/RDEPENDS on other plugins - good/bad/ugly Oct 04 18:07:53 otherwise you need to bring them by yourself in the image... think about gstreamer-ti as not a complete app or a bundle, but just as another plugin to gstreamer Oct 04 18:10:41 denix: that's what I've done earlier, brought in just the plugins I needed by adding them specifically to my "image.bb", but I could not find the ti-plugins in my deploy-folder. perhaps I just didn't search for the right ipk-names. Oct 04 18:11:12 denix: anyway, things are getting clearer now, thanks :) Oct 04 18:15:34 denix, the previous sqlite3-native, as far as i could tell, was installing absolutely nothing into staging. it emptied do_install, but the main sqlite3 recipe no longer defined do_stage. that's why i switched it to bbclassextend Oct 04 18:15:45 denix, didn't seem to cause any problems for me Oct 04 18:22:24 Is it OK to distribute .desktop files along with their icons for recipes that don't provide them? Oct 04 18:22:52 I updated the rxvt-unicode recipe and I added a .desktop file with an icon from the tango icon theme Oct 04 18:23:02 Which I guess is the baseline for Linux these days Oct 04 18:23:13 Is that OK in OE? Oct 04 18:24:52 I suspect it is. Oct 04 18:25:36 tasslehoff: what is ti-plugins? Oct 04 18:27:29 denix: I meant for instance TIViddec2 and the other ones with "TI" in them that show up in gst-inspect Oct 04 18:40:06 re Oct 04 19:18:29 hey all.. I am getting a error when running bitbake minimal-image -- > openembedded/recipes/expat/expat_2.0.1.bb --> fails -- anyone seen this issue? Oct 04 19:19:16 orbarron: this? https://bugs.launchpad.net/ubuntu/+source/gzip/+bug/524366 Oct 04 19:19:24 If so, yes :) Oct 04 19:19:50 * orbarron looks Oct 04 19:20:44 yup.. that looks like the issue Oct 04 19:21:48 RP, hi do I bother you, we've got some difficulties making CONFIG_MTD_OOPS work on nokia900 machine(SHR distro) Oct 04 19:21:58 orbarron: update your ubuntu :) Oct 04 19:22:27 we have: Oct 04 19:22:28 CONFIG_MTD_OOPS=y Oct 04 19:22:44 CONFIG_CMDLINE="snd-soc-rx51.hp_lim=42 snd-soc-tlv320aic3x.hp_dac_lim=6 console=ttyMTD,log root=/dev/mmcblk1p1 rootdelay=10 panic=20" Oct 04 19:22:56 and we see nothing on /dev/mtd2 Oct 04 19:22:57 denix: thanks... Oct 04 19:23:01 with hexdump -c Oct 04 19:23:04 do I miss something Oct 04 19:23:12 do I need mtdblock? Oct 04 19:26:20 yes Oct 04 19:26:33 mmcblk1p1 is 2nd sd slot first partition Oct 04 19:27:01 for nand or nor use mtdblock0 or so Oct 04 19:27:13 but ofc you need to have something there Oct 04 19:28:54 hi Oct 04 19:29:31 ynezz: for http://patchwork.openembedded.org/patch/3144/ did you check other machine using this recipe still build (or at least their patch still apply) ? Oct 04 19:31:14 just c7x0 patch to test in fact Oct 04 19:31:21 ericben: nope Oct 04 19:31:32 will do it and push your serie Oct 04 19:31:41 thanks a lot! :) Oct 04 19:31:56 it's hard for me to test all the possible configurations, slow build machine Oct 04 19:32:33 yes for kernel recipe building foor everymachine is not possible but at least testing patches still work is possible Oct 04 19:34:37 ericben: but anyway if I find some breakage what would I do? rebase the patches? it's time consuming and even after that I can't test boot it anyway... Oct 04 19:36:17 hm seems ti throws more and more people on oe and opensource Oct 04 19:36:59 ericben: and updating the stable patchset shouldn't break the stuff that much, or am I missing something? well it's from nothing right to the .7, but... Oct 04 19:37:11 woglinde: that's good, isn't it? Oct 04 19:37:22 yeah Oct 04 19:37:46 woglinde: looking strange :) Oct 04 19:37:55 Jay7: why? Oct 04 19:38:13 koen is doing great work with OE on Beagleboard Oct 04 19:38:14 linaro should grab a lot of OE's clients/market Oct 04 19:38:25 pff :p Oct 04 19:38:28 ynezz: in the present case no. In fact I aasked you before looking at the recipe :-) Oct 04 19:38:51 Jay7: you belive peoples using OE will use linaro ? Oct 04 19:38:57 Jay7: try to update some package in linaro/ubuntu Oct 04 19:39:09 Jay7: it's not that straight forward Oct 04 19:39:09 Jay7: just look at the size of their minimal image ! Oct 04 19:39:24 linaro is meego Oct 04 19:39:26 class Oct 04 19:39:28 do not mix peoples using OE and big companies :) Oct 04 19:39:32 for arm Oct 04 19:39:35 *g* Oct 04 19:39:41 and try to use linaro/ubuntu on something else then arm7 Oct 04 19:39:46 we have no stable branch again :) Oct 04 19:40:05 ynezz: I'm fine with OE and .dev branch :) Oct 04 19:40:06 woglinde: I like meego, but I hate rpm Oct 04 19:40:09 but.. Oct 04 19:40:38 rpm is a no go Oct 04 19:40:40 woglinde: hope that somebody, someday would make meego buildable and usable with OE :) Oct 04 19:40:42 woglinde: :) Oct 04 19:40:53 yneez since I tried to build a kernel from rpm-src Oct 04 19:41:15 denix yes? Oct 04 19:41:22 I wonder what made them switch from apt, even that scratchbox was quite usable Oct 04 19:41:33 denix I only calculate from people which joins here and beagle Oct 04 19:41:43 woglinde: just smiled at your comment :) Oct 04 19:41:49 ynezz rpm comes from intel Oct 04 19:41:57 moblin Oct 04 19:42:05 I don't like atoms either :p Oct 04 19:42:34 * ynezz hopes for ARM ownage Oct 04 19:43:21 zecke: ping Oct 04 19:43:27 denix which comment? Oct 04 19:43:29 *g* Oct 04 19:43:34 :) Oct 04 19:43:54 03Denys Dmytriyenko  07org.openembedded.dev * r9f503dbde0 10openembedded.git/recipes/sqlite/sqlite3.inc: Oct 04 19:43:55 sqlite: fix BBCLASSEXTEND breakage Oct 04 19:43:55 Old version of sqlite3-native used 'sed' to create sqlite3.h, but the new Oct 04 19:43:55 version uses 'tclsh'. Fix DEPENDS on tcl-native, as target version of 'tcl' Oct 04 19:43:55 is not needed due to --disable-tcl, but native is needed to stage 'tclsh'. Oct 04 19:43:55 Signed-off-by: Denys Dmytriyenko Oct 04 19:44:07 woglinde: the one you started with about 8-9 minutes ago... :-P Oct 04 19:44:15 haha Oct 04 19:44:29 your contract dont allow to talk about it Oct 04 19:44:32 denix: nicely done, thanks for looking into it Oct 04 19:44:34 *g* Oct 04 19:44:42 btw Oct 04 19:44:45 where is sqlite 3.7 Oct 04 19:44:53 is mupdf fixed now? :) Oct 04 19:45:06 woglinde: in qt 4.7? :p Oct 04 19:45:14 we should profit from the new engine Oct 04 19:45:31 hm, there's new engine? Oct 04 19:45:40 ynezz where comes qt4.7 into play? Oct 04 19:46:09 * Jay7 is waiting for power supply unit and pc case now.. Oct 04 19:46:16 all my life is awaiting.. Oct 04 19:46:24 woglinde: what do you mean? there's sqlite driver in qt Oct 04 19:46:31 ynezz -> Added support for write-ahead logging. Oct 04 19:46:45 no sqlite database self Oct 04 19:46:46 kergoth: btw, is there any difference between VAR_append_virtclass-native and VAR_virtclass-native_append? Oct 04 19:46:56 yes, a big difference. Oct 04 19:47:08 the latter appends to VAR_virtclass-native, which then *replaces* VAR Oct 04 19:47:12 the former is a conditional append Oct 04 19:47:17 the latter is an append to a conditional definition Oct 04 19:47:36 the VAR_virtclass-native_append is almost never the right answer Oct 04 19:47:44 copy this to manual please :) Oct 04 19:47:46 denix: hi Oct 04 19:47:54 hi zecke Oct 04 19:48:06 woglinde: http://www.xtranormal.com/watch/7023615/ and http://www.xtranormal.com/watch/6995033/ Oct 04 19:48:28 ericben saw your patch on cups, looks fine to me Oct 04 19:48:30 porn? Oct 04 19:48:51 ericben actually was thinking of something like that Oct 04 19:49:11 cool kids use ruby Oct 04 19:49:15 ericben but was worried that people would add columns to busybox Oct 04 19:49:19 zecke: you've added "-embedded" flag to qt4-tools-native - which version have you tried it with? Oct 04 19:49:21 kergoth: I don't know if my wget patcher in bitbake made it to the mailing list, can't see it in the archive... Oct 04 19:49:29 denix: 4.7 Oct 04 19:49:35 denix: you see failures with 4.6? Oct 04 19:49:36 kergoth: I think, that I've sent it like two times :p Oct 04 19:49:53 haven't seen it. are you on the list? Oct 04 19:50:02 nope, that's why I'm asking Oct 04 19:50:03 many mailing lists don't allow posts from outside the list Oct 04 19:50:08 kergoth: I see both "appends" being used in different recipes. I guess we should unify Oct 04 19:50:11 denix: -embedded should be there since forever (3.x) Oct 04 19:50:11 ah, d'oh Oct 04 19:50:14 zecke: fails for me on 4.6.2 Oct 04 19:50:26 denix wrong ones Oct 04 19:50:27 denix: the question is what the original intent was. there *is* a case where you might want to use the other one, it's just much much more rare Oct 04 19:50:49 kergoth: is the list that quiet? https://lists.berlios.de/pipermail/bitbake-dev/2010-September/thread.html Oct 04 19:50:52 zecke: instead, I use -no-x11 Oct 04 19:50:57 kergoth: or the archiver doesn't work? Oct 04 19:51:06 nokia should make a seperate tarball for qmake Oct 04 19:51:12 the list is extremely quiet Oct 04 19:51:17 ah, ok Oct 04 19:51:23 very very little work is done on bitbake in general, only RP and I really do much on it Oct 04 19:51:23 denix: can you paste me the code? and use -help? Oct 04 19:51:27 (at the moment, anyway) Oct 04 19:51:51 no wonder :p it's hard to follow :) Oct 04 19:53:20 * ynezz should probably offer some beers on CELF to someone who would make bitbake --interactive working again :) Oct 04 19:53:24 hm I should write a qtjambi recipe Oct 04 19:53:42 isn't jambi death? Oct 04 19:53:51 its community now Oct 04 19:54:03 http://qtjambi.sourceforge.net/ Oct 04 19:54:05 ah so Oct 04 19:54:20 4.7 will be out till end of the month Oct 04 19:54:42 I've used it only once for some fast mockup of gui for one java library Oct 04 19:54:52 woglinde: what did you mean by "wrong ones"? Oct 04 19:55:02 denix virtclass_append Oct 04 19:55:06 args Oct 04 19:55:09 othrere way Oct 04 19:55:13 append_virtclass Oct 04 19:55:18 hm Oct 04 19:55:19 damn Oct 04 19:55:23 let me see Oct 04 19:56:08 03Frans Meulenbroeks  07org.openembedded.dev * r3dac3ac87f 10openembedded.git/recipes/musicpd/mpd_0.15.12.bb: Oct 04 19:56:08 mpd_0.15.12: added creation of /var/lib/mpd/playlists dir Oct 04 19:56:08 Signed-off-by: Frans Meulenbroeks Oct 04 19:56:30 denix as kergoth stated out most of append_virtclass maybee wrong Oct 04 19:57:23 woglinde: no, you got it backwards Oct 04 19:57:23 woglinde: I thought the other way around... now I'm confused :) Oct 04 19:57:36 _append_ is usually what people want Oct 04 19:57:44 a conditional append, rather than appending to a conditional version of a variable Oct 04 19:58:17 kergoth: so, this is not right http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=9f503dbde0f2f6e9d4a94f6ab4be1b04dfc99950 Oct 04 19:58:22 FOO = "bar"; FOO_local = "baz"; FOO_local_append = " test" Oct 04 19:58:27 this results in FOO being 'baz test' Oct 04 19:58:34 note: 'bar' is gone Oct 04 19:58:42 this is the case whether FOO_local itself is defined or not Oct 04 19:58:51 FOO = "bar"; FOO_local_append = 'baz' -> FOO is baz Oct 04 19:58:55 03Petr Štetiar  07org.openembedded.dev * r5786f08ba6 10openembedded.git/recipes/linux/linux_2.6.34.bb: Oct 04 19:58:56 linux-2.6.34: upgrade to latest stablepatch 2.6.34.7 Oct 04 19:58:56 Signed-off-by: Petr Štetiar Oct 04 19:58:56 Signed-off-by: Eric Bénard Oct 04 19:58:59 this is why it's almost never what you want Oct 04 19:59:01 03Petr Štetiar  07org.openembedded.dev * r88ff1fbba5 10openembedded.git/recipes/linux/ (20 files in 2 dirs): Oct 04 19:59:02 linux-2.6.34: add support for ts72xx and make it default kernel Oct 04 19:59:02 Signed-off-by: Petr Štetiar Oct 04 19:59:02 Signed-off-by: Eric Bénard Oct 04 19:59:02 03Petr Štetiar  07org.openembedded.dev * r66888364f0 10openembedded.git/recipes/netbase/ (netbase/ts72xx/interfaces netbase_4.21.bb): Oct 04 19:59:02 netbase: add interface file for ts72xx machine and bump PR Oct 04 19:59:03 Signed-off-by: Petr Štetiar Oct 04 19:59:03 Signed-off-by: Eric Bénard Oct 04 19:59:23 denix: right, that'll result in only having tcl-native in the native deps Oct 04 19:59:31 ynezz: please check the keernel still builds. Patchwork didn't took your 2nd kernel patch so I had to hand apply it Oct 04 19:59:31 03Petr Štetiar  07org.openembedded.dev * re2f261e5e0 10openembedded.git/recipes/librs232/librs232_1.0.0.bb: Oct 04 19:59:32 librs232: add new recipe Oct 04 19:59:32 Signed-off-by: Petr Štetiar Oct 04 19:59:32 Signed-off-by: Eric Bénard Oct 04 19:59:32 03Petr Štetiar  07org.openembedded.dev * r996f7110f8 10openembedded.git/recipes/lua/lua-rs232_1.0.0.bb: Oct 04 19:59:32 lua-rs232: add new recipe Oct 04 19:59:33 Signed-off-by: Petr Štetiar Oct 04 19:59:33 Signed-off-by: Eric Bénard Oct 04 19:59:34 03Petr Štetiar  07org.openembedded.dev * re19cd914f7 10openembedded.git/recipes/lua/ (lua-filesystem-1.5.0/makefile.patch lua-filesystem_1.5.0.bb): Oct 04 19:59:34 lua-filesystem: add new recipe Oct 04 19:59:35 Signed-off-by: Petr Štetiar Oct 04 19:59:35 Signed-off-by: Eric Bénard Oct 04 19:59:36 03Petr Štetiar  07org.openembedded.dev * r1d03ea798e 10openembedded.git/recipes/linux/linux_2.6.27.bb: Oct 04 19:59:37 linux-2.6.27: don't use as default kernel for ts72xx Oct 04 19:59:37 Signed-off-by: Petr Štetiar Oct 04 19:59:38 Signed-off-by: Eric Bénard Oct 04 19:59:49 but it depends on evaluation order between the overrides and the anonymous python function in native.bbclass which adjusts DEPENDS Oct 04 19:59:54 best to change it Oct 04 20:00:05 kergoth: thanks! Oct 04 20:00:09 np Oct 04 20:00:17 this should be in an FAQ somewhere Oct 04 20:00:19 heh Oct 04 20:00:38 kergoth ah now I see Oct 04 20:00:47 yes both can be usefull Oct 04 20:01:21 of course, you could just use FOO_local += instead of FOO_local_append, which would greatly reduce the confusion Oct 04 20:01:33 but really, _append needs to go away entirely :) Oct 04 20:01:41 hm Oct 04 20:01:42 hehe Oct 04 20:02:00 we just need a new syntax for doing conditional variable definitions and appends, to be able to remove them Oct 04 20:04:41 ynezz hm only problem with qtjambi seems you cannt use the qt-designer ui files the same easy way as in python Oct 04 20:05:39 I did used ui, tho Oct 04 20:05:52 hm with jui Oct 04 20:05:56 ah Oct 04 20:06:02 but seems not directly Oct 04 20:06:21 with python-pyqt you can use the ui file directly Oct 04 20:06:42 even without precompiling Oct 04 20:06:46 this rockz Oct 04 20:07:21 03Petr Štetiar  07org.openembedded.dev * rb0bcd6021f 10openembedded.git/recipes/ntpclient/ (files/fix-adjtimex-uclibc.patch ntpclient_2007_365.bb): (log message trimmed) Oct 04 20:07:21 ntpclient: fix adjtimex compile error on uclibc Oct 04 20:07:21 ntpclient.o: In function `set_freq': Oct 04 20:07:21 ntpclient-2007_365-r3/ntpclient-2007/ntpclient.c:166: undefined reference to `__adj timex' Oct 04 20:07:21 ntpclient.o: In function `get_current_freq': Oct 04 20:07:22 ntpclient-2007_365-r3/ntpclient-2007/ntpclient.c:149: undefined reference to `__adj timex' Oct 04 20:07:23 Signed-off-by: Petr Štetiar Oct 04 20:07:35 hm? Oct 04 20:07:37 woglinde: looking at the source, yes it's jui, but it's similar to ui Oct 04 20:07:48 wasnt the timex stuff fixed longtime agao? Oct 04 20:08:08 I remember dark I made some patches for it Oct 04 20:08:15 on uclibc? Oct 04 20:08:18 yes Oct 04 20:08:33 hm, maybe something strange here :p Oct 04 20:08:50 ah ntpclient Oct 04 20:08:52 okay Oct 04 20:08:56 I made it for ntpd Oct 04 20:09:01 uff :) Oct 04 20:09:04 re gnutoo Oct 04 20:09:44 woglinde, hi Oct 04 20:09:56 not enough ram.... Oct 04 20:10:10 I should increase my swap Oct 04 20:10:42 * ynezz is happy, with bbappend has made first image in his life without tweaking any stuff in the image manually Oct 04 20:12:10 woglinde: btw what's the difference between ui and jui? Oct 04 20:14:33 woglinde: just the ? otherwise it looks same Oct 04 20:14:44 but i'm not ui expert :) Oct 04 20:14:44 ynezz hm strange Oct 04 20:14:52 why they called it jui than Oct 04 20:15:15 no, it's my mistake :) Oct 04 20:15:34 overlooked com.trolltech.qt.core.Qt.Orientation.Horizontal and friends :) Oct 04 20:16:36 hm Oct 04 20:17:01 anyway it might be possible to convert it somehow Oct 04 20:17:12 sure Oct 04 20:17:22 but who cares, I try to stay far away from java stuff Oct 04 20:17:27 :) Oct 04 20:17:48 only bytecode I love is Lua's Oct 04 20:18:50 haha Oct 04 20:18:58 hmm, even building tcl-native doesn't let sqlite3-native succeed here. does it needs special compilers or what? Oct 04 20:19:34 | In file included from ./src/alter.c:15: Oct 04 20:19:34 | ./src/sqliteInt.h:418: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'i64' Oct 04 20:20:38 ericben: thank you for the time spent on my patches and I hope it didn't break much stuff :p Oct 04 20:30:51 mickey|bbl: hi! Oct 04 20:31:07 ynezz: time will say :-) Oct 04 20:31:19 zecke: hi z. Oct 04 20:31:41 fwiw, 2.6.13 builds, 2.6.23 bails out Oct 04 20:32:27 zecke: how are things? when will you be back in .de? Oct 04 20:32:42 mickey|bbl: -20 days or such. :) Oct 04 20:33:13 ah, already. nice, welcome back Oct 04 20:33:35 mickey|bbl: yeah, that's because you have 0-size sqlite3.h already staged from previous failure... Oct 04 20:34:03 denix: hmm, i built from scratch Oct 04 20:35:41 mickey|bbl: I saw that error on the second run - first it failed to "tclsh sqlite3.h", but staged empty header. then those types would not be defined, due to empty header Oct 04 20:36:41 denix: ah, that's indeed possible. so is it "just" a missing dependency to tcl-native or more? Oct 04 20:36:58 zecke, will you make OEDEM? Oct 04 20:37:37 mickey|bbl: it worked for me after I fixed the dependency on tcl-native, but I had to clean staging dir Oct 04 20:39:50 no Oct 04 20:40:58 hm, what's the right syntax of doing something like DEPENDS -= "bluez-libs" ? Oct 04 20:41:49 in the docs there's bunch of the examples for appending/prepending but nothing for removing :) Oct 04 20:42:09 ynezz: you will need to use the filter_out syntax Oct 04 20:42:17 ynezz: filter it Oct 04 20:42:47 oe_filter_out() Oct 04 20:43:04 thanks, git grep shows some examples Oct 04 20:43:41 ynezz: used a _lot_ here - http://arago-project.org/git/?p=arago.git;a=blob;f=recipes/qt4/qt4-embedded-4.6.2/amend.inc;hb=HEAD Oct 04 20:43:42 :) Oct 04 20:43:53 hehe Oct 04 20:45:49 would this work? DEPENDS_ts72xx := "${@oe_filter_out('bluez-libs', bb.data.getVar('DEPENDS', d, 1), d)}" Oct 04 20:47:32 I mean just the DEPENDS_ts72xx part Oct 04 20:48:51 looks fine Oct 04 20:49:00 hm and can I even use this with image name ? Like DEPENDS_my-cutom-image = "" ? Oct 04 20:50:07 ynezz: what do you want to do with that? Oct 04 20:50:22 denix shorten the buildtime Oct 04 20:50:33 because bluze drags in half of gtk Oct 04 20:50:46 I don't mean exactly the bluez-libs use case now Oct 04 20:51:03 if you want to alter your image's DEPENDS, use DEPENDS_pn- = "" Oct 04 20:51:14 woglinde: I was asking about the image question Oct 04 20:51:20 thanksoh sorry Oct 04 20:51:27 but for example I would like to tweak some stuff based on image name Oct 04 20:51:40 ynezz: but that's not what you want Oct 04 20:52:11 yes, wrong example Oct 04 20:52:15 no, you can only use machines, architecture, distro in overrides, not the image you are building :) Oct 04 20:52:33 ah, that's the answer I was looking for :) Oct 04 20:54:27 what is the right LICENSE for Apache v2 ? we have : Apache License V2.0 , Apache License, Version 2.0 , Apache-2.0 ? Oct 04 21:00:12 03Dmitry Artamonow  07org.openembedded.dev * r51f101e880 10openembedded.git/recipes/ttf-fonts/ttf-droid_git.bb: Oct 04 21:00:12 ttf-droid: new recipe Oct 04 21:00:12 Add recipe for Droid fonts (specifically developed by Ascender for Oct 04 21:00:12 Google's Android platform). Since there's no canonical tarball URL for Oct 04 21:00:12 these fonts, we fetch them directly from Android git repo. Oct 04 21:00:12 Signed-off-by: Dmitry Artamonow Oct 04 21:00:13 Signed-off-by: Eric Bénard Oct 04 21:00:13 03Ben Gardiner  07org.openembedded.dev * r72816b2be9 10openembedded.git/recipes/latencytop/ (6 files in 2 dirs): Oct 04 21:00:14 latencytop: add version 0.5 Oct 04 21:00:14 Add a new recipe to for latencytop. Oct 04 21:00:15 * patch the makefile to disable GTK support, use the LDFLAGS and use ncurses Oct 04 21:00:15 instead of the 'w' variant. Oct 04 21:00:16 Signed-off-by: Ben Gardiner Oct 04 21:00:17 Signed-off-by: Eric Bénard Oct 04 21:02:33 do_populate_staging) failed with 256 Oct 04 21:02:40 NOTE: the following files were installed but not shipped in any package: Oct 04 21:02:45 NOTE: /lib/modules/2.6.35.4/modules.builtin Oct 04 21:02:49 NOTE: /lib/modules/2.6.35.4/modules.order Oct 04 21:02:56 ERROR: Function do_stage failed Oct 04 21:03:13 where can I include this files into my package? Oct 04 21:12:49 dv: I think your error was in a different place... Oct 04 21:16:29 denix: I think, I built too new kernel version and this files just not included into the package. where can I look for package script to include them? Oct 04 21:16:56 dv: that would be kernel.bbclass Oct 04 21:18:53 denix: thank you so much! I see! Oct 04 21:25:44 denix: could you paste me the output of your ./configure -help? Oct 04 21:26:21 denix: I can fix that tomorrow... and use -no-x11 if it is necessary for older Qt versions Oct 04 21:27:30 zecke: have not been able to do that yet - in the middle of a build right now... Oct 04 21:51:31 http://www.cheeseorfont.com/ Oct 04 21:57:36 so sysroot in libtool is working but not so well yet need to steer this libtool Oct 04 22:00:14 There's no reason for a recipe to just do do_stage () { autotools_stage_all } right? Oct 04 22:00:23 * Tartarus sees effem's commit that nuked 500+ of them Oct 04 22:03:59 kergoth: replacing DEPENDS_virtclass-native_append with DEPENDS_virtclass-native += doesn't seem to work... Oct 04 22:08:07 does anyone know if GIt trunk is broken? When I run bitbake console image, there is a linker error on _video_sdl.la because it can't find lesd. Oct 04 22:11:34 mekoniak, it's often broken for some combinaison Oct 04 22:11:50 kergoth: (repeat) replacing DEPENDS_virtclass-native_append with DEPENDS_virtclass-native += doesn't seem to work... Oct 04 22:11:51 mekoniak, but you could report on the mailing list if you get no answer here Oct 04 22:12:19 DEPENDS_append_virtclass-native works though Oct 04 22:12:30 uh, i didn't tell you to use _virtclass-native += Oct 04 22:12:36 that's broken just as badly as the one i told you not to use Oct 04 22:12:41 i said that makes it more *clear* whats happening Oct 04 22:12:45 but its still NOT what you want Oct 04 22:14:09 :) Oct 04 22:15:01 ugghhh too many emails after a long weekend away Oct 04 22:16:26 thanks GNUtoo. Oct 04 22:18:43 03Denys Dmytriyenko  07org.openembedded.dev * r9189f87363 10openembedded.git/recipes/sqlite/sqlite3.inc: Oct 04 22:18:43 sqlite: use conditional append, instead of appending to conditional variable :) Oct 04 22:18:43 While both seem to work in this case, this one is more correct. Oct 04 22:18:43 Signed-off-by: Denys Dmytriyenko Oct 04 22:23:21 03Jaap de Jong  07org.openembedded.dev * r86f4623bf5 10openembedded.git/recipes/xc3sprog/xc3sprog_461.bb: Oct 04 22:23:21 xc3sprog: new recipe Oct 04 22:23:21 Author: Jaap de Jong Oct 04 22:23:21 Signed-off-by: Eric Bénard Oct 04 22:30:06 03Tom Rini  07org.openembedded.dev * r33afdd5a47 10openembedded.git/recipes/dtc/dtc-native.inc: Oct 04 22:30:06 dtc-native: Switch from legacy staging Oct 04 22:30:06 Signed-off-by: Tom Rini Oct 04 22:30:09 03Tom Rini  07org.openembedded.dev * r2c7b3abb44 10openembedded.git/recipes/libtool/ (libtool-sdk_2.2.6a.bb libtool-sdk_2.2.6b.bb): Oct 04 22:30:09 libtool-sdk: Convert from legacy staging Oct 04 22:30:09 Since these are ${HOST_SYS} specific we make ${HOST_SYS}-libtool here. Oct 04 22:30:09 Signed-off-by: Tom Rini Oct 04 22:30:10 03Tom Rini  07org.openembedded.dev * r4b559b091b 10openembedded.git/recipes/oprofile/oprofile.inc: Oct 04 22:30:11 oprofile: Drop do_stage() { autotools_stage_all } Oct 04 22:30:11 Signed-off-by: Tom Rini Oct 04 22:30:12 03Tom Rini  07org.openembedded.dev * r9975b23563 10openembedded.git/recipes/dt/dt.inc: Oct 04 22:30:12 dt: Fix 'S' as it's no longer dt.d-WIP Oct 04 22:30:12 Signed-off-by: Tom Rini Oct 04 22:30:15 03Tom Rini  07org.openembedded.dev * ra44217d83e 10openembedded.git/recipes/gdb/gdb-cross-sdk.inc: Oct 04 22:30:15 gdb-cross-sdk: Only fixup sim Makefiles that exist Oct 04 22:30:15 This can fail for versions that just pass in --disable-sim Oct 04 22:30:15 Signed-off-by: Tom Rini Oct 04 22:38:09 grg: fyi I've also seen that gcc blowup on 4.3.3, just updated to 4.4.1 so here's hoping I don't see it anymore again either :) Oct 04 22:39:15 Tartarus, maybe it was fixed upstream... could be useful to find the commit Oct 04 22:43:23 yeah Oct 05 00:02:15 Tartarus, what distro is your build system where you've seen the gcc/strsignal build failure? Oct 05 00:02:53 seen it on u804, u910 and centos5 Oct 05 00:19:37 Anybody can help-me with this: http://pastebin.com/fMYyzySk | Angstrom Distro | A1200 Machine Oct 05 00:25:05 I'm having issues with a cmake based recipe and bbthreads Oct 05 00:25:14 03Tom Rini  07org.openembedded.dev * r1e2e6295da 10openembedded.git/recipes/gdb/gdb-common.inc: Oct 05 00:25:14 gdb-common.inc: Nuke a few more possible conflicts. Oct 05 00:25:14 Signed-off-by: Tom Rini Oct 05 00:25:21 is there a way to let a recipe finish, before other tasks run Oct 05 00:25:44 Beyond DEPENDS? Oct 05 00:27:08 who? Oct 05 00:32:56 That was to crofton, sorry Oct 05 00:33:36 angelox_123, I guess e-image isn't supported right now by angstrom or your machine needs updating to have a good X server picked out Oct 05 00:33:41 heh Oct 05 00:33:48 it is weird problem Oct 05 00:33:50 any fix? Oct 05 00:34:39 angelox_123, see openembedded/conf/machine/omap4430-sdp.conf as an example that doesn't use kdrive for an xserver Oct 05 00:35:13 No idea if that'll work on your actual hw Oct 05 00:53:12 grg, thanks for tracking that down. I don't have a pref either way really Oct 05 00:53:40 I've always been a patch N versions of gcc/etc person, but that's less popular now (and I'm also fine with that, really) Oct 05 01:37:37 03Chris Larson  07master * r653f9f0f69 10openembedded.git/recipes/gtk+/gtk+.inc: (log message trimmed) Oct 05 01:37:37 gtk+-native: fix relocation Oct 05 01:37:37 In a sysroot process hook, to deal with relocation: Oct 05 01:37:37 - Re-generate the gdk-pixbuf.loaders file to contain the current sysroot path Oct 05 01:37:37 to the pixbuf loaders directory. Oct 05 01:37:38 - Create wrappers for gdk-pixbuf-csource and gdk-pixbuf-query-loaders Oct 05 01:37:39 utilities, passing along the appropriate environment variables to ensure **** ENDING LOGGING AT Tue Oct 05 02:59:57 2010