**** BEGIN LOGGING AT Mon Jun 28 02:59:57 2010 Jun 28 03:10:05 03Denys Dmytriyenko  07org.openembedded.dev * rc93524aff2 10openembedded.git/recipes/ti/matrix-tui_svn.bb: Jun 28 03:10:05 matrix-tui: depend on libxml2 Jun 28 03:10:05 Signed-off-by: Denys Dmytriyenko Jun 28 06:45:23 03Martin Jansa  07org.openembedded.dev * r036a0fcada 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Jun 28 06:45:23 linux-openmoko-2.6.32: add patch with possible WS fix Jun 28 06:45:23 Signed-off-by: Martin Jansa Jun 28 06:59:38 hi Jun 28 07:08:59 good morning Jun 28 07:09:30 (for germans would be more good) Jun 28 07:15:54 h emckoan Jun 28 07:15:58 ups mckoan Jun 28 07:16:06 yeah was the best wm-match yet Jun 28 07:16:29 args worldcup Jun 28 07:16:30 sorry Jun 28 07:16:56 woglinde: too excited ? ;-) Jun 28 07:19:33 woglinde: hi Jun 28 07:23:03 hi zecke Jun 28 07:23:36 ji hineld Jun 28 07:23:38 args Jun 28 07:23:44 hi jin^eld Jun 28 07:35:52 morning Jun 28 07:37:04 hrw: gm Jun 28 07:37:37 hi hrw Jun 28 08:40:42 hi florian Jun 28 08:44:04 03Martin Jansa  07org.openembedded.dev * r9e3c69bd62 10openembedded.git/ (6 files in 6 dirs): Jun 28 08:44:04 xorg: bump versions xorg-sgml-doctools xf86-video-intel libxfont util-macros Jun 28 08:44:04 * adjust bumping script to handle -native versions better with BBCLASSEXTEND Jun 28 08:44:04 Signed-off-by: Martin Jansa Jun 28 08:44:15 03Martin Jansa  07org.openembedded.dev * rc5d324ae7c 10openembedded.git/recipes/dri/ (4 files in 2 dirs): Jun 28 08:44:15 libdrm: add 2.4.21, use the same PE as libdrm_git had Jun 28 08:44:15 Signed-off-by: Martin Jansa Jun 28 08:44:15 03Martin Jansa  07org.openembedded.dev * rf9e5577949 10openembedded.git/recipes/xorg-doc/xorg-doc-common.inc: Jun 28 08:44:16 xorg-doc-common: use BPN instead of XORG_PN Jun 28 08:44:20 good morning Jun 28 08:44:23 Signed-off-by: Martin Jansa Jun 28 08:44:24 03Martin Jansa  07org.openembedded.dev * r9c6d528574 10openembedded.git/recipes/xorg-doc/ (10 files): Jun 28 08:44:25 xorg-doc: use INC_PR Jun 28 08:44:25 Signed-off-by: Martin Jansa Jun 28 08:44:25 03Martin Jansa  07org.openembedded.dev * r4a112bf47f 10openembedded.git/recipes/xorg-doc/ (9 files): Jun 28 08:44:26 xorg-doc: oe-stylize Jun 28 08:44:26 Signed-off-by: Martin Jansa Jun 28 08:45:54 morning Jun 28 08:46:06 hi jama Jun 28 08:46:15 big xorg cleanup Jun 28 08:46:17 *g* Jun 28 08:49:51 jo ant Jun 28 08:50:27 hm someone has an example for listinig for evdev events Jun 28 08:50:31 hello Jun 28 08:50:32 ? Jun 28 08:50:37 hmm..yes Jun 28 08:50:40 mom Jun 28 08:50:58 without xorg-stuff Jun 28 08:51:01 http://atrey.karlin.mff.cuni.cz/~metan/evfilter/ Jun 28 08:52:21 hm Jun 28 08:52:22 thanks Jun 28 08:52:28 perhaps is not yet 100% Jun 28 08:52:33 why the hell they checked in a .so Jun 28 08:52:43 or if youwant a quick-and dirty C solution: Jun 28 08:53:11 see http://git.linuxtogo.org/?p=groups/kexecboot/kexecboot.git Jun 28 08:54:17 hm I only want to listen on the dev Jun 28 08:54:28 and react on keycode-event Jun 28 08:54:44 enum actions_t process_events(struct ev_params_t *ev) Jun 28 08:55:08 ant which file I have to look up? Jun 28 08:55:14 oh..sorry Jun 28 08:55:22 kexecboot.c Jun 28 08:55:37 pasted half link... Jun 28 08:56:43 hm is this a callback? Jun 28 08:57:51 hm do while loop Jun 28 08:57:58 I think there was a single commit adding event readings + some patches Jun 28 08:58:58 ???? Jun 28 08:59:17 in .git Jun 28 08:59:23 anyway, yes it's a loop Jun 28 08:59:25 hm meeting Jun 28 08:59:27 till later Jun 28 08:59:31 bye Jun 28 09:03:35 woglinde: devicescan.c Jun 28 09:03:40 bottom half :) Jun 28 09:03:46 morning Jun 28 09:05:13 hey Jun 28 09:13:51 03Koen Kooi  07org.openembedded.dev * r76adf8ac3f 10openembedded.git/recipes/freondemo/freondemo-media-files.bb: freondemo-media-files: add sha256sum as well Jun 28 09:19:10 03Martin Jansa  07org.openembedded.dev * r9d4a494cdd 10openembedded.git/recipes/gnuchess/ (files/gnuchess.desktop gnuchess_5.021.bb): Jun 28 09:19:10 gnuchess: fix fatal QA issue GNU_HASH and minor QA issues Jun 28 09:19:10 * misinterpreted Version field Jun 28 09:19:10 * deprecated Encoding field Jun 28 09:19:10 Signed-off-by: Martin Jansa Jun 28 09:22:09 03Martin Jansa  07org.openembedded.dev * r3249005fd5 10openembedded.git/recipes/ipkg-utils/ipkg-utils_1.6+cvs20050404.bb: Jun 28 09:22:09 ipkg-utils: respect LDFLAGS, fix QA GNU_HASH error Jun 28 09:22:09 Signed-off-by: Martin Jansa Jun 28 09:32:24 03Koen Kooi  07org.openembedded.dev * r54c5c004e7 10openembedded.git/recipes/qt4/ (qt-4.6.2+4.7.0-beta1.inc qt4.inc): qt4: move xmlpatterns to 4.7 build, it breaks 4.6.x Jun 28 09:57:49 re Jun 28 09:58:12 hm launch now Jun 28 09:58:17 woglinde: hey Jun 28 09:59:37 he jay7 Jun 28 09:59:46 hm thats only the half-task Jun 28 09:59:52 open the device Jun 28 09:59:52 woglinde: what do you needed about event devices? Jun 28 10:00:22 jay7 open the device Jun 28 10:00:30 wait for a special keycode Jun 28 10:00:43 react on the keycode Jun 28 10:00:57 and I want do the waiting right Jun 28 10:01:08 in devicescan.c we are selecting and opening right devices Jun 28 10:01:13 without a forever loop Jun 28 10:01:15 in kexecboot.c is main processing Jun 28 10:01:35 inside select() Jun 28 10:02:38 okay lunch now Jun 28 10:02:51 if you wish to blocking wait you can just read() it Jun 28 10:03:04 but only one device at time then Jun 28 10:03:25 no Jun 28 10:03:28 none blocking Jun 28 10:03:36 hm okay I read this right Jun 28 10:03:45 you are waiting for a evebt Jun 28 10:03:48 ups event Jun 28 10:03:58 and the get the event from the right eventdevice Jun 28 10:04:31 okay have to go Jun 28 10:04:33 till later Jun 28 10:09:36 well.. I should go now Jun 28 10:09:45 afk for 1 day :) Jun 28 10:10:10 enjoy Jun 28 10:33:15 re Jun 28 11:02:44 woglinde, a simple acpi listener/actor sample is e.g. here FYI: http://git.busybox.net/busybox/tree/util-linux/acpid.c Jun 28 11:03:11 03Martin Jansa  07org.openembedded.dev * r9925d751e4 10openembedded.git/conf/distro/include/preferred-shr-versions.inc: Jun 28 11:03:11 shr: remove P_V for boost 1.40.0 -> prefer newer boost 1.41.0 Jun 28 11:03:11 Signed-off-by: Martin Jansa Jun 28 11:04:54 hm my co-worker showed my gizmod Jun 28 11:04:59 ups me Jun 28 11:05:18 acpi would only work for x86 Jun 28 11:16:46 I have IMAGE_ROOTFS_SIZE_ext2 = "100000" in my image recpie but bitbake keeps ignoring it Jun 28 11:16:54 falling back to 65535 Jun 28 11:17:20 is an image recipe an appropriate place to specifiy the size of the image? Jun 28 11:25:09 hmm, when I have an init script ipk, and I create a new ipk where I only change the INITSCRIPT_PARAMS so that the script gets run earlier in the sequence, and I then update the ipk - the rc.d link does not get updated.... any ideas? Jun 28 11:25:35 it says "System startup links for /etc/bleh.. already exist." Jun 28 11:29:34 Jin^eLD do a update-rc.d -f remove bla before Jun 28 11:29:37 ups an Jun 28 11:29:59 woglinde: well, it's supposed to get updated along with everything while doing an opkg update Jun 28 11:30:10 i.e. when updating via a feed Jun 28 11:30:54 when just doing opkg remove package it does remove the startup links too Jun 28 11:31:04 but it does not do so during an opkg update Jun 28 11:32:07 it seems that the postrm checks for the purge parameter and only then removes the scripts Jun 28 11:32:54 so I guess there's a little flaw in the logic of the update-rc.d class, which assumes that init order will never change between package version Jun 28 11:34:01 jepp Jun 28 11:34:13 thats why debian uses update-rc.d -f remove to Jun 28 11:34:24 for versions pined Jun 28 11:34:32 there is in some deb something like Jun 28 11:34:45 if version == bla foo Jun 28 11:34:52 update-rc.d -f remove bla Jun 28 11:35:27 but anyway the future should be upstart or systemd Jun 28 11:36:51 hm better <= Jun 28 11:50:14 jo raster Jun 28 11:51:02 woglinde: boop Jun 28 12:42:39 03Martin Jansa  07org.openembedded.dev * r6eabc9e5c6 10openembedded.git/recipes/pingus/pingus_0.7.2.bb: Jun 28 12:42:40 pingus: Workaround QA issue GNU HASH Jun 28 12:42:40 Signed-off-by: Martin Jansa Jun 28 12:42:50 03Martin Jansa  07org.openembedded.dev * r1959671957 10openembedded.git/recipes/fbreader/ (3 files in 2 dirs): Jun 28 12:42:51 fbreader: add 0.12.10 Jun 28 12:42:51 Signed-off-by: Martin Jansa Jun 28 12:45:45 Hi all folks! Jun 28 12:47:15 I have a question, which may be simple, for for which I can't get soutions. Jun 28 12:47:57 I would like to upgrade my small embedded distro (ARM) to eglibc 2.12.90. I noticed that the soname of the ld-linux interpreter has been changed. Jun 28 12:48:59 Thank to a helpful guy in IRC, I discovered that the soname changed after eglibc release 2.12-r11.2+svnr10689.5. My executables are searching for ld-linux.so.2, but now .so.3 is being built. When this happens, what should I do? Re-build all my executables? Jun 28 12:49:43 mrkiko jupp Jun 28 12:49:48 best solution Jun 28 12:49:58 not so good soltuion but may work Jun 28 12:50:03 make a link Jun 28 12:51:03 woglinde: thank you!!! Jun 28 12:51:40 woglinde: I wanted to make a link also; but I told myself: "don't be evil. If they changed the soname, a reason may exist. I think I'll read the svn log...". Thank youuuu!!! Jun 28 12:52:13 hm a link is easy to remove Jun 28 13:05:03 ok ... thank you Jun 28 13:30:08 bye! Jun 28 13:30:11 oops wrong window Jun 28 13:44:37 03Koen Kooi  07org.openembedded.dev * r7ef921f191 10openembedded.git/recipes/dri/libdrm_2.4.21.bb: Jun 28 13:44:37 libdrm: temporarely disable radeon for angstrom Jun 28 13:44:37 * I have a radeon based card attached to my efika and I'd like to use that :) Jun 28 15:11:23 03Dmitry Eremin-Solenikov  07org.openembedded.dev * rb409fdad28 10openembedded.git/recipes/opie-taskbar/opie-taskbar-images.inc: Jun 28 15:11:23 opie-taskbar-images: unbreak for non-default extensions Jun 28 15:11:23 The automagic change for RPROVIDES -> RPROVIDES_${PN} broke Jun 28 15:11:23 opie-taskbar-images-WWWxHHH providing opie-taskbar-images. Unbreak opie Jun 28 15:11:23 for most of the platforms (e.g. tosa). Jun 28 15:54:42 hi all Jun 28 16:10:56 03Stefan Schmidt  07org.openembedded.dev * r6c332eb3f1 10openembedded.git/recipes/shadow/shadow.inc: Jun 28 16:10:57 shadow: Disable email checking completely Jun 28 16:10:57 This allows faster login due to not checking for new emails. Jun 28 16:10:57 03Stefan Schmidt  07org.openembedded.dev * r09a8755408 10openembedded.git/recipes/ecj/libecj-bootstrap_3.4.bb: libecj-bootstrap_3.4.bb: Fix SRC_URI for new location Jun 28 17:06:09 Anyone have experience with the power management facilities for OMAP under OE? Jun 28 17:12:44 robtow, isn't that kernel related? Jun 28 17:13:17 gnutoo - largely. Some userland facilities. Jun 28 17:13:25 ok Jun 28 17:13:45 gnutool - apparently large changes are included in the .34 kernel. Jun 28 17:13:58 err, "gnutoo". Jun 28 17:14:11 nice Jun 28 17:14:30 so the pm tree/branch is not necessary anymore? Jun 28 17:15:07 gnutoo - apparently. However, do you have any experience with the pm branch tree under OE? Jun 28 17:38:06 robtow: OE does not provide special provisions for powermanagement Jun 28 17:38:37 khem - didn't think there was any special provision; but there may be experts in here who have twiddled with it on OMAP. Jun 28 17:38:41 robtow: it could be supporting utililities you need to get it working Jun 28 17:39:22 robtow: hmm your earlier questions did not say that explicitly Jun 28 17:40:02 robtow: what areas are you looking info on Jun 28 17:40:05 khem - true, therefore some ambiguity. Jun 28 17:40:39 robtow: pm is a vastly arch/soc and sometimes board specific Jun 28 17:40:49 khem - my toplevel goal is to see how low I can draw down the average power consumption on an OMAP 3450; specifically, the Gumstix Overo, using Angstyrom. Jun 28 17:40:55 err, "Angstrom". Jun 28 17:40:59 kevin hilman has done some work in this regard on omaps Jun 28 17:41:20 you can ask your question on omap kernel mailing list I am sure he will replu Jun 28 17:41:21 khem - I found a reference to some of his patches. Jun 28 17:41:30 khem - good suggestion. Jun 28 17:41:51 mostly if its a userlevel question then you should study the documents Jun 28 17:42:00 there is a proc interface to control it Jun 28 17:42:25 and you can have a userspace utility to tinker it Jun 28 17:42:36 Hmmm. Got some errors in executing Jun 28 17:42:36 and do your experiment Jun 28 17:43:06 opkg is getting fragile day after day Jun 28 18:00:15 03Khem Raj  07org.openembedded.dev * r9dff5ded22 10openembedded.git/recipes/uclibc/uclibc_git.bb: Jun 28 18:00:15 uclibc_git.bb: Bump SRCREV Jun 28 18:00:15 Signed-off-by: Khem Raj Jun 28 18:06:46 03Khem Raj  07org.openembedded.dev * r4e252b4700 10openembedded.git/recipes/eglibc/ (eglibc_2.12.bb eglibc_svn.bb): Jun 28 18:06:46 eglibc_svn, eglibc_2.12: Bump SRCREV Jun 28 18:06:46 Signed-off-by: Khem Raj Jun 28 18:10:02 03Khem Raj  07org.openembedded.dev * rff23c9810a 10openembedded.git/recipes/gcc/gcc-4.5.inc: Jun 28 18:10:02 gcc-4.5.inc: Bump SRCREV Jun 28 18:10:02 Signed-off-by: Khem Raj Jun 28 18:24:58 hi Jun 28 18:25:51 blindvt: hey Jun 28 18:26:12 khem, hey Jun 28 18:27:21 khem, given mips n32, what ISA would be ok/common to pick? gcc-4.5 branch yells at me that it doesn't like isa3 (which we have as default, from the looks)? Jun 28 18:27:46 khem, i don't want to go o32 Jun 28 18:29:39 hmz. Should we pick n32 for mips-whatever-*gnueabi? Did mips ever honor eabi in the triplet? Jun 28 18:29:59 * blindvt goes to look at gnu-config git Jun 28 18:30:14 re Jun 28 18:30:26 for arm yes Jun 28 18:32:41 woglinde, for arm it always did, yea. Don't know offhand if mipsel64-foo-bar would pick n64 eabi LE, mips-foo-bar would pick o32 EB and mipsel-foo-bargnueabi would pick n32 EL though. Jun 28 18:35:26 woglinde, i do remember that buildroot always patched gnuconfig to handle sh*eb-*-* proper. Can't seem to find my clone in the first place. wahhh Jun 28 18:36:40 *g* me wonders why you now work on oe Jun 28 18:36:47 well screw it. o32 is good enough for them. *shrug* Jun 28 18:38:21 woglinde, i've done my share for a couple of years but it doesn't make sense given oe. The latter just needs smallish configury tweaks and the former would look just as oe if it was ever fixed for good ;) Jun 28 18:39:26 aeh??? Jun 28 18:43:56 woglinde, oe is technically superior compared to buildroot. Nowadays i prefer to tweak oe rather than reinventing the wheel (i.e. fixing buildroot) Jun 28 18:44:06 *g* Jun 28 18:44:07 fine Jun 28 18:45:05 * Jay7 is back.. things was quicker Jun 28 18:45:26 lol Jun 28 18:45:38 welcomeback jay7 Jun 28 18:46:24 woglinde: how are your event devices? :) Jun 28 18:46:42 hm was only eval Jun 28 18:46:44 for now Jun 28 18:47:36 hm cve for png Jun 28 18:47:40 who will fix it? Jun 28 18:47:55 libtiff too Jun 28 18:52:21 hi florian Jun 28 18:53:22 re Jun 28 19:02:00 03Koen Kooi  07org.openembedded.dev * r17cf1e72de 10openembedded.git/recipes/ti/ti-dsplib_2.10.bb: ti-dsplib: fix up broken seperators in header Jun 28 19:02:56 blindvt: hey, was away for a while Jun 28 19:03:23 blindvt: gnueabi is only for arm Jun 28 19:03:58 blindvt: gcc should configure for o32 by default though you can configure it to chose say n32 Jun 28 19:04:05 he khem 4:1 Jun 28 19:04:09 or n64 in mips64-linux case Jun 28 19:04:25 woglinde: yeah :) although it could have been 2:2 Jun 28 19:04:31 yes Jun 28 19:04:40 woglinde: and Klose's first goal wasnt valid either Jun 28 19:04:50 so I would say Germany was lucky Jun 28 19:05:20 khem, is there anything i could do to help in getting those configury bits in? I've just pinged them again fwiw Jun 28 19:05:47 blindvt: they will be in I am still struggling to fix errno handing on mips/nptl Jun 28 19:06:24 blindvt: point me to patches in patchwork Jun 28 19:06:34 there are 2 I guess Jun 28 19:06:38 I have only one Jun 28 19:13:36 khem, i think those are all, in about this order: http://patchwork.openembedded.org/patch/{2203,2194,2196,2200,2204}/ Jun 28 19:14:50 khem, it could be that 2200 was in another patch too (i don't think so, but if it was, drop that duplicate hunk) Jun 28 19:15:48 khem, if you're about to touch mips then please have a look at the abi/arch selection in uclibc-config.inc for mips (see the TODO/FIXME) Jun 28 19:17:10 khem, (i don't know offhand how to distinguish n32 from o32 given just the target triplet.) Jun 28 19:22:16 blindvt: there is no way to distinguish Jun 28 19:22:42 blindvt: its only decided by what options you default gcc to Jun 28 19:22:58 or use multilib and pass right commandline options Jun 28 19:23:25 choice for o32 is due to many old devices not supporting n32|n64 Jun 28 19:24:38 khem: any idea if alignment traps from eglibc loader are fixed in latest gcc/eglibc bump? Jun 28 19:24:39 khem, ld-2.20.1.20100303 seems to segfault on me for n32 anyway, thus my screw them note above Jun 28 19:25:30 JaMa: its not a gcc problem rather a gcc problem Jun 28 19:25:52 JaMa: the problem is due to unaligned relocations in eh_frame section Jun 28 19:26:42 blindvt: hmm. Jun 28 19:27:13 blindvt: I was thinking of either creating a new arch for n32 or use n32 as default for OE's mips toolchain Jun 28 19:28:35 khem: but I got those only in images built with gcc-4.5, same eglibc built with gcc-4.4.4 doesn't show them Jun 28 19:29:08 khem, yea, i was also switching the default to n32. Perhaps i did something wrong so don't take my word for it Jun 28 19:29:22 khem: so from what you told me last time, I understand that it happens from time to time with "bad" combination of toolchain/eglibc Jun 28 19:36:42 JaMa|AFK: yes Jun 28 19:36:49 JaMa|AFK: its definitely a gcc bug Jun 28 19:36:56 but since how long I dont know Jun 28 19:37:18 JaMa|AFK: it could be that gcc-4.4.4 just hides it Jun 28 19:39:20 hi pb__ Jun 28 19:40:43 hi pb Jun 28 19:41:23 hi blindvt Jun 28 19:42:11 hi woglinde Jun 28 19:43:26 blindvt: hmm above patches do not apply through patchwork anymore. Can you refresh them against current master and resend that would be awsome Jun 28 19:44:29 s/awsome/awesome Jun 28 19:45:19 khem, sure, i'll do that after having dealt with mips' odd libc-tls handling in the buildsys. Jun 28 19:46:52 khem, btw.. I didn't look, but what (if any yet) is the --enable-gold knob in oe? Jun 28 19:47:41 khem, i have a note in my TODO.binutils that warn-once is NIH (or was at the time i added that todo) Jun 28 19:48:53 khem, same file suggests that i missed sort-common back then Jun 28 19:53:15 blindvt: no we dont have knob for gold yet Jun 28 19:53:16 I don't think there is a gold knob at the moment. If you wanted to add one then that'd be cool. Jun 28 19:53:18 khem, so if good old ld segfaults on me on mips n32 i'd tend to bring gold forward before touching ld legacy (which in turn begs the question about gold support in oe) Jun 28 19:53:41 blindvt: gold is not so well supported on all arches Jun 28 19:53:53 blindvt: x86 and arm are one;s where it would work better Jun 28 19:54:02 but usually ld is still preferred Jun 28 19:54:36 blindvt: ld should work well with n32 I wonder what error do you get Jun 28 19:55:05 03Khem Raj  07org.openembedded.dev * r2eb1d58ffc 10openembedded.git/recipes/file/file_5.04.bb: Jun 28 19:55:05 file_5.04.bb: Add zlib-native dependency to file-native Jun 28 19:55:05 Signed-off-by: Khem Raj Jun 28 19:55:57 khem, i've only tried gold on x86 back then and it only needed little patching to behave properly. But yea, mips could turn out to be a bit involved from the looks Jun 28 19:56:16 it will I am sure :) Jun 28 19:56:31 so far it only supports arches that google cares Jun 28 19:57:56 khem, i don't get why you'd need explicit listing of that zlib-native dep. Wasn't zlib optional and shouldn't the native depends be automagically deduced off plain depends? Jun 28 19:58:57 blindvt: look into file.inc Jun 28 19:59:00 DEPENDS_virtclass-native = "" Jun 28 19:59:07 that clears it Jun 28 20:00:32 I want to setup and internal gcc git repo where others can commit and this repo is tracking upstream gcc.git repo Jun 28 20:00:41 anybody setup such a thing Jun 28 20:00:47 may be for some other project Jun 28 20:01:17 khem, that's just silly, isn't it. ISTR that the dependency resolver drops deps to self, so resetting those native deps is the bug, not the missing dep to zlib-native, no? Jun 28 20:01:23 I still would want to pull changes from gcc.git at the same time share it with other devs Jun 28 20:02:44 woglinde: Ever seen this one? http://pastebin.com/5bULMW2K Jun 28 20:03:11 stefan_schmidt yes Jun 28 20:03:28 woglinde: good :) Jun 28 20:05:04 woglinde: David will fill an redmine ticket about it. Just to have it noted somewhere. Jun 28 20:05:17 Need to go now. bye Jun 28 20:05:32 hm Jun 28 20:06:18 try -Xms 1024 -Xmx 1024 in javac from bootstrap/openjdk-1.6.0/ Jun 28 20:06:32 blindvt: it will create dep on file-native-native Jun 28 20:06:37 prolly Jun 28 20:06:40 or did so Jun 28 20:06:46 woglinde: ok, will try tomorrow Jun 28 20:07:16 khem, wasn't that fixed already? Jun 28 20:08:31 blindvt: I dont know Jun 28 20:08:32 khem, you should rather poke kergoth to fix that for real (or tag a release where this is fixed) instread of adding that kind of duct tape, shouldn't you Jun 28 20:09:26 blindvt: factually I should remove the deps and build file-native and test Jun 28 20:09:38 but that will need a fresh build Jun 28 20:09:48 to be error free Jun 28 20:10:17 zlib-native might be pulled in by someone else before file native is built and you wouldnt get the issue Jun 28 20:10:18 khem, yep Jun 28 20:11:00 I think that using explicit dependency list is a good thing as well because you know it as you see it Jun 28 20:11:24 blindvt: I will wait for your refreshed patches Jun 28 20:12:08 * khem has to see how tls is copied during thread creation Jun 28 20:13:38 khem, as said, i think that file deals nicely without support for compressed magic (i could be wrong though. That would arguably be a bug in file though). If it indeed needs zlib nowadays then the native package can just use the deps from the cross package. If bitbake erroneously constructs -native-native deps then bitbake needs fixing, not the poor, innocent package. That DEPENDS_virtclass_file="" is a hack, as well as any addition ontop Jun 28 20:13:38 of it Jun 28 20:14:24 * blindvt goes to cook & Jun 28 20:17:26 someone (eldis) recently had a problem compiling the kernel (http://pastebin.com/FL9SUJX2). I have the same problem, was a solution ever stated? Jun 28 20:18:49 sure Jun 28 20:18:52 tommd: do you have it built one time ? Jun 28 20:18:57 update your defconfig Jun 28 20:19:07 tommd: you might be missing defconfig update too Jun 28 20:19:41 I do have it built once, but due to the funky way the ai repository (alwaysinnovating.com) is setup, I must build the kernel twice to get a bootable uImage. Jun 28 20:20:09 Not sure how the build system really works, but evidently the first time it doesn't build the initramfs into the uImage Jun 28 21:13:59 03Koen Kooi  07org.openembedded.dev * r1a6a5f2ca9 10openembedded.git/recipes/xcb/libxcb_1.6.bb: libxcb: bump PR Jun 28 21:15:46 gm Jun 28 21:16:01 hi likewise Jun 28 21:16:04 hi likewise Jun 28 21:16:10 yes libxcb Jun 28 21:16:17 more should converted to it Jun 28 21:16:21 +soft Jun 28 21:16:48 hi khem, woglinde et al Jun 28 21:20:11 eds-dbus do_fetch fails http://tinderbox.openembedded.net/packages/624281/ Jun 28 21:20:15 anybody able to confirm ? Jun 28 21:25:06 dcordes: works well here Jun 28 21:25:10 without problem Jun 28 21:26:22 hmmm I am getting it from http://www.angstrom-distribution.org/unstable/sources/trunk_svn.o-hand.com_.repos.eds-dbus_736_.tar.gz Jun 28 21:26:30 not from svn though Jun 28 21:29:01 khem: 23:08:01 < JaMa> dcordes: svn: Could not open the requested SVN filesystem Jun 28 21:29:07 svn co http://svn.o-hand.com/repos/eds-dbus/trunk eds-dbus fails Jun 28 21:29:55 so I wonder if it was moved to some other server or directory or is just dead Jun 28 21:30:33 good nite Jun 28 21:31:37 I've had similar problem Jun 28 21:31:54 with different p Jun 28 21:32:13 -SRC_URI = "http://gypsy.freedesktop.org/gypsy-releases/gypsy-0.7.tar.gz;name=gypsy \ Jun 28 21:32:16 +SRC_URI = "http://gypsy.freedesktop.org/releases/gypsy-0.7.tar.gz;name=gypsy \ Jun 28 21:32:29 gypsy source moved Jun 28 21:33:54 ah it moved to git Jun 28 21:34:10 khem: eds-dbus ? Jun 28 21:34:24 yes Jun 28 21:34:37 maybe other recipes are affacted ? Jun 28 21:35:35 khem: can you push a fix for the eds-dbus recipe ? Jun 28 21:36:16 ok Jun 28 21:38:35 khem: nice. can you check http://bugs.openembedded.org/show_bug.cgi?id=5435#c19 if you have a minute ? It's about the armv7a toolchain version Jun 28 21:47:18 khem: FYI: alignment traps still here with bumped gcc-4.5+eglibc-2.12 (just in case you will be bored in near future :)) Jun 28 21:48:18 JaMa: ok Jun 28 21:48:22 thx for info Jun 28 21:48:53 dcordes: yes I read your comment Jun 28 21:49:04 I have no problem bumping gcc to 4.4.4 for armv7 Jun 28 22:03:45 khem: ok you think it's good to push ? Jun 28 22:10:40 03Martin Jansa  07org.openembedded.dev * r7e87614466 10openembedded.git/recipes/matchbox-keyboard/ (2 files): Jun 28 22:10:40 matchbox-keyboard*: drop RPROVIDES Jun 28 22:10:40 * RCONFLICTS is enough to block packages providing conflicting files Jun 28 22:10:40 * distributions can still decide if they need only some or all 3 parts of Jun 28 22:10:40 matchbox-keyboard-inputmethod or if matchbox-keyboard (like Jun 28 22:10:51 03Martin Jansa  07org.openembedded.dev * r4fabbb2fc8 10openembedded.git/recipes/opie-taskbar/opie-taskbar-images.inc: Jun 28 22:10:51 opie-taskbar-images: fix RPROVIDES Jun 28 22:10:51 Signed-off-by: Martin Jansa Jun 28 22:10:51 Acked-by: Andrea Adami Jun 28 22:10:51 Acked-by: Paul Eggleton Jun 28 22:12:44 03Martin Jansa  07org.openembedded.dev * rd12d783046 10openembedded.git/recipes/e17/elmdentica_svn.bb: Jun 28 22:12:44 elmdentica: temporary bump EFL_SRCREV only for elmdentica, please remove with next EFL_BUMP Jun 28 22:12:44 Signed-off-by: Martin Jansa Jun 28 22:16:23 dcordes: can you try this patch for eds plz http://uclibc.org/~kraj/0001-eds-dbus-Update-to-follow-git-repository.patch Jun 28 22:16:42 dcordes: I have no way to verify its build atm as my machine are busy Jun 28 22:17:03 dcordes: if you can confirm it work for you then I will push it Jun 28 22:17:46 with pleasure thanks Jun 28 22:32:13 khem, the errno problem on mips could perhaps be caused by mips using the wrong tls handler for nptl (it should use the arch impl but links in both the generic one and the specialized arch variant). I've fixed this locally for now Jun 28 22:33:14 g'night all Jun 28 22:37:50 khem: you're a beast :) works well Jun 28 22:38:16 dcordes: cool thx Jun 28 22:38:22 blindvt: send me the patch Jun 28 22:38:25 before you sleep Jun 28 22:38:34 I can test it on my oven as it is hot Jun 28 22:45:18 03Khem Raj  07org.openembedded.dev * rb918ca1061 10openembedded.git/recipes/eds/ (10 files in 2 dirs): Jun 28 22:45:18 eds-dbus: Update to follow git repository. Jun 28 22:45:18 * Remove unused patches. Jun 28 22:45:18 Thanks to initial patch from Ilyes Gouta Jun 28 22:45:18 Signed-off-by: Khem Raj Jun 28 23:02:39 ah blindvt thats not the only problem. I have fixes for that Jun 28 23:02:42 too Jun 28 23:35:10 * likewise needs sleep Jun 28 23:35:14 see you all Jun 28 23:35:17 g'nite for later Jun 28 23:35:39 can somebody check if they can build gypsy ? Jun 28 23:39:30 didn't expect my angstrom-gnome-image mission to be that productive :D Jun 29 00:06:57 dcordes: what problem do u see Jun 29 00:12:59 khem: gypsy didn't fetch for me Jun 29 00:15:33 hmm Jun 29 02:46:42 03Khem Raj  07org.openembedded.dev * re614547faf 10openembedded.git/recipes/orinoco/spectrum-fw.bb: Jun 29 02:46:42 spectrum-fw.bb: Change SRC_URI to somewhere fetchable. Jun 29 02:46:42 Signed-off-by: Khem Raj Jun 29 02:46:53 03Khem Raj  07org.openembedded.dev * rc5aee636a8 10openembedded.git/recipes/gdb/gdb-common.inc: Jun 29 02:46:53 gdb-common.inc: Use FILESPATHPKG in place of FILESDIR. Jun 29 02:46:53 Signed-off-by: Khem Raj Jun 29 02:46:54 03Khem Raj  07org.openembedded.dev * r4daefb7e39 10openembedded.git/recipes/realpath/ (5 files in 3 dirs): Jun 29 02:46:54 realpath: Add recipe for version 1.15 **** ENDING LOGGING AT Tue Jun 29 02:59:57 2010