**** BEGIN LOGGING AT Thu May 13 02:59:56 2010 May 13 07:10:44 hi May 13 07:13:43 Hi. I have a problem: oe rebuild busybox everytime I rebuild rootfs image. Can someone help me? May 13 08:33:33 03Stefan Schmidt  07org.openembedded.dev * r93c655e4f6 10openembedded.git/conf/machine/imote2.conf: May 13 08:33:33 imote2.conf: Restore old kernel provider. May 13 08:33:33 Thanks Koen for spotting. May 13 10:51:04 03Koen Kooi  07org.openembedded.dev * r9f68124263 10openembedded.git/recipes/xorg-lib/ (pixman-0.18.0/565-scanline.patch pixman_0.18.0.bb): pixman: add generic 565 scanline fetcher May 13 11:17:17 03Martin Jansa  07org.openembedded.dev * r93a439a42b 10openembedded.git/recipes/qemu/ (2 files in 2 dirs): May 13 11:17:18 qemu: add workarround to fallback to safe mmap_min_addr May 13 11:17:18 * in case /proc/sys/vm/mmap_min_addr cannot be read because of security May 13 11:17:18 of host system May 13 11:17:18 Signed-off-by: Martin Jansa May 13 11:30:26 what's the best way to get utf-8 input working in a shell in oe? May 13 13:17:12 03Shane Volpe  07org.openembedded.dev * r21600c8a33 10openembedded.git/recipes/mono/ (41 files in 7 dirs): May 13 13:17:12 mono: Add version 2.6.3 May 13 13:17:12 * Removed old versions of mono as they were broken. May 13 13:17:12 * Fixed mono packages so mono now builds with new native staging. May 13 13:17:12 * Changed mono floating point to ARM_FPU_NONE so it works with all architectures. May 13 13:17:13 * Added libtoolize --force --copy to mono.inc to allow any libtool version to work. May 13 13:54:23 is there any way to find out which packet in the dependency chain triggers building of x11 stuff? May 13 13:54:55 bitbake -g footarget May 13 13:55:07 look at the generated .dot files May 13 13:55:47 woglinde: thanks. I'll try May 13 14:00:48 hi woglinde May 13 14:00:56 btw can I add an init script in XSERVER of a machine? May 13 14:01:11 the init script recipe is for calibrating the 2 screens at once May 13 14:01:20 using ts_calibrate May 13 14:01:52 also if I do so should I put something like COMPATIBLE_MACHINE in the tslib-init recipe? May 13 14:02:28 because it has only the tslib-init/bug/ts_calibrate.sh which works only on bug 1.x May 13 14:02:41 but maybe kgilmer would be interested in adding support for bug 2.0 May 13 14:02:57 because I heard that they use tslib and ts_calibrate too May 13 14:03:10 sorry May 13 14:03:12 dont know May 13 14:03:16 ask on the ml May 13 14:16:38 hms May 13 14:17:01 how do I staging correct for a shell script generated inside the recipe? May 13 14:19:33 good morning woglinde GNUtoo May 13 14:19:39 hi May 13 14:20:06 woglinde, try installing in do_install_append May 13 14:20:21 hi kgilmer May 13 14:20:35 sorry its native package May 13 14:20:48 look at jikes-inital May 13 14:20:51 kgilmer, I'm confused between matchbox 1.x and matchbox 2.x May 13 14:20:58 which one is better? May 13 14:21:06 is 2.x develloped? May 13 14:21:25 on bug 1.x and 2.x which matchbox is used? May 13 14:22:06 woglinde, ok thanks btw May 13 14:22:22 hm maybee we will go for icepick May 13 14:22:26 to build the java stuff May 13 14:22:36 icepick? May 13 14:22:46 http://icedtea.classpath.org/wiki/IcePick May 13 14:22:57 GNUtoo, i don't know :< May 13 14:23:08 re hrw May 13 14:23:10 kgilmer, ah ok May 13 14:23:24 but hrw-uds probably would GNUtoo :) May 13 14:23:32 wow icepick May 13 14:23:37 ok May 13 14:23:56 I mean the native java stuff May 13 14:24:00 to build May 13 14:24:32 yes I heard that java build was not modular May 13 14:24:49 with icepick it becomes modular May 13 14:24:56 => more easily buildable May 13 14:25:02 *buildable more easily May 13 14:25:05 hm not really May 13 14:25:11 icepick is only the langtools May 13 14:25:22 ah ok May 13 14:25:27 javac javadoc and stuff with out the classlibs May 13 14:25:35 ah ok :( May 13 14:25:47 sigsaw you mean May 13 14:25:57 and its sheduled for jdk7 May 13 14:26:04 ok I'll look May 13 14:26:10 I hope jdk7 stay free May 13 14:26:22 s/free/floss May 13 14:26:34 openjdk7 has same license as openjdk6 May 13 14:27:22 kgilmer: ? May 13 14:28:15 ok nice May 13 14:28:29 hrw-uds, hi kgilmer told me that you knew well matchbox May 13 14:28:35 GNUtoo was wondering about matchbox 1.x vs 2.x in the bug1.3 vs bug20 builds May 13 14:28:48 is 2.x maintained? May 13 14:29:00 I bet 1.x isn't anymore May 13 14:29:13 catch me next hour please then May 13 14:29:22 ok thanks a lot May 13 14:29:26 inside of session now May 13 14:31:09 GNUtoo: mb1.x is not maintained, mb2 is rather not maintained too. mbwm2 is used by nokia in maemo so maybe they do something. May 13 14:31:27 ok thanks May 13 14:31:31 GNUtoo: OH people moved to Clutter powered devices May 13 14:31:39 * GNUtoo wonder which one to use May 13 14:31:50 panel2, desktop2, wm1 or wm2 May 13 14:31:59 ok thanks a lot May 13 14:32:07 because 1.x state in oe isn't great May 13 14:32:13 lots of things to fix May 13 14:40:12 Has something happened to patchwork? I'm sure there's been some patches in the last 8 dyas! May 13 14:40:41 s/dyas/days/ May 13 14:56:51 hm someone remember how to do do_unpack into a specific dir? May 13 14:59:55 woglinde: 'subdir' uri parameter, its relative to the WORKDIR, it seems May 13 15:00:12 see base.bbclass, the end of oe_unpack_file May 13 15:00:19 kergoth thanks May 13 15:00:58 np May 13 15:02:15 hm cannt find a recipe using it May 13 15:02:20 anyone have some examples of amend.inc files they can point me to or share? Is this in the manual? May 13 15:02:33 hi kergoth May 13 15:02:37 hey May 13 15:02:44 woglinde: wouldn't surprise me if none did :) May 13 15:02:45 hi pb May 13 15:02:49 hi woglinde May 13 15:02:50 should work though, in theory May 13 15:02:55 will test it May 13 15:03:06 dickelbeck: an amend.inc is the same as any other .inc, its bitbake metadata May 13 15:03:15 dickelbeck: just applied after the regular recipe is May 13 15:03:46 just avoid BBCLASSEXTEND and anonymous python usage in it, those make it unhappy May 13 15:04:51 to enable it, just add the amend class to the distro file? Would be nice to have a summarizing manual page. I could maybe do that if I can get to a point of understanding. May 13 15:05:55 kergoth hm subdir is not what I wanted May 13 15:06:17 kergoth I have 2 sources packages which should be untared in the same dir May 13 15:06:48 args May 13 15:06:56 I meant I have 2 tar's May 13 15:06:59 not sourcepackages May 13 15:07:02 *sigh* May 13 15:07:30 what do you mean by "untarred in the same dir" exactly? May 13 15:07:46 if you just specify them both in SRC_URI, bitbake will effectively do "cd ${WORKDIR}; tar xf one.tar; tar xf two.tar" May 13 15:07:57 pb right May 13 15:08:06 but the content shoud go into one dir May 13 15:08:09 hm May 13 15:08:13 do you mean that you want to unpack them and then somehow munge the contents together? May 13 15:08:17 seems I have to work with cp May 13 15:08:19 pb yes May 13 15:08:29 i.e. one.tar unpacks into ${WORKDIR}/one, two.tar unpacks into ${WORKDIR}/two, but you want all the files in one/? May 13 15:08:36 pb right May 13 15:08:38 in that case, you will need to add a do_munge() method to move the bits around May 13 15:09:08 or teach the tar unpacker to pass --strip-components to tar, which I don't think it currently does May 13 15:09:08 yes May 13 15:09:21 indeed like emacs or glibc May 13 15:09:31 (for the do munge) May 13 15:09:39 if it did, then you could say "strip_components=1;subdir=woglinde_is_cool" and it would do the right thing May 13 15:10:02 that would probably be the most elegant solution May 13 15:10:57 pb jupp May 13 15:12:25 pb is it in bitbake or oe? May 13 15:12:45 in oe. base.bbclass May 13 15:12:55 okay thanks May 13 15:14:06 looking at do_unpack May 13 15:14:24 I am a little bit ashamed May 13 15:15:45 god, i really wish the pydoc on functions/methods would document the exceptions they might raise May 13 15:19:14 pb haha you pointed me to do_unpack to clean it up May 13 15:19:19 heh May 13 15:20:42 woglinde: i started trying to do that - http://github.com/kergoth/OpenEmbedded/commits/kergoth/unpack May 13 15:20:49 haven't had time to polish it May 13 15:21:00 find the time May 13 15:21:00 woglinde: ignore the top commit, that's a no-go May 13 15:21:03 *g* May 13 15:21:08 http://github.com/kergoth/OpenEmbedded/commit/2c130ea5209eadd902a698db2ac4229186424cb1 is the cleanup May 13 15:23:08 hm I will split it up May 13 15:23:11 in cmd_tar May 13 15:23:15 woglinde: mainly got bit by the implicit globbing May 13 15:23:23 old method concatenated into a big command string May 13 15:23:28 so '*' in uri got expanded there May 13 15:23:43 but * in file:// semantics are unclear, i think it shouldn't be supported May 13 15:24:32 kergot we have 10 times the string tar x --no-same-owner -f May 13 15:25:00 hehe May 13 15:25:06 that's a bit much May 13 15:25:15 sure May 13 15:25:29 and you can't easily append options to it May 13 15:25:57 likely better to use a list and pass the list directly to subprocess as the arguments May 13 15:26:01 * kergoth shrugs May 13 15:26:20 *g* May 13 15:26:42 yeah. or split it up, use one subprocess for the decompressing, then an explicit pipe to another subprocess to do the untarring May 13 15:27:52 * kergoth looks over his oe branches and wonders if he ever got some of this merged.. is a slacker, clearly May 13 15:33:46 did I ever merge this first pass removal of the 'bluez4' recipe? May 13 15:33:48 * kergoth scratches head May 13 15:34:17 not that I recall May 13 15:34:35 * woglinde wonders too why we dont use j option for tar and bz2? May 13 15:35:05 It's changed names (admittedly years ago, but still). May 13 15:35:42 dunno, might be a sop for non-gnu tar or something May 13 15:35:55 not that I imagine there are many folks using oe with a tar that doesn't understand -j May 13 15:36:38 or might have been done that way with a view to a future refactor along the lines that I suggested above :-} May 13 15:36:54 matchbox-desktop2 is great...it works with :0.0 and :0.1 May 13 15:37:27 so, two out of the possible infinite universe of displays and screens? good work there. May 13 15:37:30 pb_: http://git.openembedded.org/cgit.cgi/openembedded/commit/?h=kergoth/bluez&id=a053da8743966445342b495307e532178a1294e9 - thoughts? based on your stuff :) May 13 15:38:58 looks fairly decent to me, though I wonder whether it would be better to stick with "bluez4-libs" rather than calling the 4.x ones bluez-libs as well. May 13 15:39:30 iirc, the 4.x and 3.x libraries are not even close to api compatible with each other, and (absent versioned dependencies of course) it would probably be easier for apps if they could stipulate which they wanted. May 13 15:40:38 apart from that, though, it looks good (at least as far as I can tell from reading the diff) May 13 15:42:49 I wonder when some fixes May 13 15:42:51 NOTE: Handling BitBake files: | (1554/8421) [18 %]ERROR: Error Method already seen: get_glibc_fpu_setting in' glibc-package.bbclass' now in 'glibc-package.inc' May 13 15:43:53 hmm, good point May 13 15:49:58 mickeyl: good morning May 13 15:50:05 good morning pb_ May 13 15:50:11 pb_: already moved? May 13 15:50:11 hi mickeyl May 13 15:50:14 hi woglinde May 13 15:50:26 mickeyl: yup, last monday May 13 15:50:44 cool, congrats. can you settle now or is there lots of work left in the new house? May 13 15:50:51 no, lots of work still to do May 13 15:51:01 we basically need to move out of the house into a caravan for a few months May 13 15:51:29 gotta demolish and rebuild part of the building, and the bit that we aren't demolishing requires quite a lot of remedial work May 13 15:51:48 new insulation, new wiring, new doors, new windows, new staircase, new floors May 13 15:52:04 wow, that sounds like a plan for ther remainder of this year May 13 15:52:11 yup, pretty much May 13 15:52:30 hope to have the heavy work finished by about october, so that we don't have to stay in the caravan during the really cold bit of winter May 13 15:52:54 but yeah, it will probably keep our builder occupied for several months May 13 15:53:13 right May 13 15:55:22 first step is to get the electricity company to install a new supply, which annoyingly involves digging up the road and hence costs rather a lot May 13 15:56:20 then, gotta get the bloke with the hiab truck back to install the caravan. then connect services up to it, then we can move in and start the work on the actual house. May 13 15:59:01 uh oh May 13 15:59:14 diggin up the road sounds expensive May 13 16:04:54 yeah, it is rather annoying May 13 16:05:04 no way to avoid it though, really May 13 16:05:11 the supply cable is on the other side :-} May 13 16:17:27 any plans for OEDEM later this year? May 13 16:19:12 good question. no plans whatsoever so far. wouldn't hurt to run a doodle poll to find a popular weekend within the Sept.->Nov. timeframe May 13 16:20:57 there is an ELC-Europe in Cambridge, UK end of October... not sure PB would want to host another OEDEM though... :) May 13 16:22:16 as much as I enjoyed Cambridge I wouldn't mind less of a trip... May 13 16:22:29 ~seen zecke May 13 16:22:33 zecke <~ich@123-192-188-67.dynamic.kbronet.com.tw> was last seen on IRC in channel #oe, 6d 3h 59m 13s ago, saying: 'ant_work: no, it is not broadcast.'. May 13 16:24:41 denix yes saw this too May 13 16:25:29 zecke will be in iceland this or next week according to harald's blog May 13 16:26:36 I would be happy enough to host another OEDEM in Cambridge if that was desirable. We have a new, and slightly smaller, office now but I think we could still fit everyone in. May 13 16:30:18 woglinde: I think he is there right now May 13 16:30:25 woglinde: vulcan tourism. ;) May 13 16:30:45 pb hm where do I have strip_components to be announced so that Exception: Message:global name 'strip_components' is not defined May 13 16:30:55 dont occur anymore May 13 16:33:54 anyone else? May 13 16:34:00 name and subdir are known May 13 16:39:25 hmmmm May 13 16:39:47 i wonder how to handle exceptions raised by python snippets. i'd really rather not do what we do today and ignore them other than messages May 13 16:39:50 * kergoth thinks May 13 16:40:13 hms May 13 16:40:17 dont get it May 13 16:40:42 *sigh* May 13 16:44:38 anyone using the 2mmc at91sam9g20ek ? May 13 16:45:21 ojay fixed May 13 16:46:41 anyone done any work on speeding up OE builds - wondering if there is any info out there: recomendations on disk configuration/filesystem, ramdisk for cache etc? May 13 16:47:29 and what is the general consensus on PARALLEL_MAKE and BB_NUMBER_THREADS - I've had issues using this on some machines but not others May 13 16:47:43 tharvey: have you looked into packaged staging? http://marcin.juszkiewicz.com.pl/2008/07/01/packaged-staging-and-what-it-gives/ May 13 16:48:17 was just looking at the RFC for that May 13 16:51:32 I've used packaged-staging in the past with arago (OE derived) and found that while it did speed things up it made certain updates difficult to debug. I ran into situations where something in package staging was out of date and causing build issues resulting in my having to do a full clean May 13 16:51:52 spent a lot of time thrashing around before I tried the clean build - left a sour taste... May 13 16:53:13 pstage packages contain pr/pv/pn.. the only reason itd be out of date is if you changed a global cflag or something.. those cases you have to wipe it, the same way you'd wipe tmp May 13 16:53:26 which is why we need to get sane metadata hashing for it May 13 16:53:32 to avoid that need in the future May 13 16:54:41 noted - I was working with a dev version of arago and its likely I or they were not bumping things correctly at the time May 13 16:54:53 * kergoth nods, thatd do it May 13 16:55:30 I can chalk it up to confusion - I wasn't familiar with the package staging at the time and it caught me by surprise that there was another thing to clean May 13 16:56:10 yeah, understandable :\ May 13 16:56:28 tharvey: usually I'm paying attention to pv/pr, but I could have missed some before... May 13 16:56:46 * kergoth is trying to resolve this madness via http://github.com/kergoth/OE-Signatures May 13 16:56:46 what filesystems are being used for build efficiency? I'm using ext4 on ubuntu 10.4 (standard) - wondering if any of the other filesystems perform much better for builds May 13 16:57:10 denix, could just have easily been something I was doing :) May 13 16:59:08 tharvey: since arago does not maintain upgrade path for binary packages between releases, it's easy to miss some pv/pr badness May 13 16:59:59 tharvey: I use ext4 over RAID0 on Gentoo for my builds May 13 17:01:13 denix, its been many months since I did some ti-gstreamer testing with arago, do you know if the org.openembedded.dev ti-gstreamer/ti packages have been updated? May 13 17:02:21 tharvey yes May 13 17:02:25 koen takes care May 13 17:02:37 but you need binutils 2.18 May 13 17:03:06 all above didnt compile the codecs package for me May 13 17:03:23 tharvey do you have a beagleboard? May 13 17:03:38 woglinde, yes, a revC - haven't used it in a while May 13 17:03:45 hm okay May 13 17:03:52 beacuse with 2.6.32 psp kernel May 13 17:04:02 I need an external powerplug May 13 17:04:12 because of powermangament stuff May 13 17:04:18 usb didnt work May 13 17:04:21 anymore May 13 17:04:30 woglinde, your saying current dev angstrom uses something other than binutils 2.18 and you had to override that? May 13 17:04:42 tharvey yes May 13 17:04:47 but maybee it works for you May 13 17:04:53 just try a normal build May 13 17:05:20 will do, its high time that I re-sync'd with OE May 13 17:18:27 tharvey: all the recipes from recipes/ti/ were synced between arago and OE.dev couple month ago. May 13 17:21:24 denix, cool - thanks! Looks like the kernel's in dev are perhaps out of date with the PSP patches though? May 13 17:21:38 ie, I don't see cameraisp patches in 2.6.32 kernel May 13 17:22:57 woglinde: btw, koen is not the only one... May 13 17:23:58 denix *g* May 13 17:24:03 tharvey: linux-omap-psp in oe.dev (or linux-omap3 in arago) are the latest. it's slightly behind upstream l-o kernel though May 13 17:24:53 I hope my external power plug will arrive at the weekend May 13 17:25:01 so I can use beagleboard again May 13 17:25:54 woglinde: power-over-usb? May 13 17:26:18 denix jupp May 13 17:26:42 denix psp-kernel 2.6.32 dont works here May 13 17:26:52 because of the powermanagment patche May 13 17:27:05 thats I was told in #beagleboard May 13 17:27:43 denix, good to know - I'll see if that has camera-isp in it. Who maintains the linux-omap-psp recipes? - I'm curious in how up to date they are kept May 13 17:28:06 tharvey koen again takes care May 13 17:28:10 tharvey: that would be koen May 13 17:28:13 woglinde: :-P May 13 17:28:27 denix *g* May 13 17:28:59 and the diff between the linux-omap and linux-omap-psp kernels is? the psp ones simply have TI's PSP patches applied? May 13 17:29:25 tharvey: in arago I have linux-omap3 recipe. but they both pull from arago repo... May 13 17:29:25 tharvey you cannt really use the dsp without psp May 13 17:29:47 dont know about gles stuff May 13 17:29:52 tharvey: pretty much May 13 17:30:03 woglinde: gles is external May 13 17:30:50 woglinde: as long as you link modules and libs against the kenrel you are using, you should be fine May 13 17:31:31 denix okay May 13 17:31:38 as I said wasnt sure May 13 17:32:03 ok - I'll give the linux-omap-psp kernel a shot for my overo + camera May 13 17:35:40 tharvey - I'm also using an Overo and a camera; a Canon G10. What sort of camera setup are you working with? May 13 17:37:38 robtow, I'm using an image sensor, via the omap ISP - not an end-user camera May 13 17:40:57 I'm finding some references to openwrt and openembedded together - are people just using the openebedded toolchain to build OE recipes or something else? May 13 17:44:40 khem: You're sure you don't have something switching everything on ppc from -Os to -O2 locally, right? May 13 17:46:01 fixes incomming May 13 17:46:02 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * raaf2edbcf7 10openembedded.git/recipes/gnome/gconf_2.28.0.bb: (log message trimmed) May 13 17:46:02 gconf 2.28.0: depending on dbus-x11 fixes ORBit issues and bump PR May 13 17:46:02 If we don't depend on dbus-x11 starting /etc/x11/xserver-nodm output messages like: May 13 17:46:02 GConf Error: Failed to contact configuration server; some possible causes are that you May 13 17:46:02 need to enable TCP/IP networking for ORBit, or you have stale NFS locks due to a system May 13 17:46:02 crash. May 13 17:46:03 I think gconf is the right package to depend on dbus-x11 because: May 13 17:46:06 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r6205a271b4 10openembedded.git/recipes/matchbox-desktop/matchbox-desktop_0.9.1.bb: (log message trimmed) May 13 17:46:06 matchbox-desktop 0.9.1: fix segfault at startup and convert to new staging May 13 17:46:06 If matchbox-common is not present,matchbox-desktop segfault May 13 17:46:06 (the debugging was done with strace) May 13 17:46:06 Also convert to new staging: May 13 17:46:07 I had packaged staging,removed the custom do_stage and got: May 13 17:46:07 staging/armv6-angstrom-linux-gnueabi/usr/include/matchbox-desktop $ ls May 13 17:46:08 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r0e617baf29 10openembedded.git/recipes/initscripts/initscripts_1.0.bb: (log message trimmed) May 13 17:46:08 initscripts 1.0 : change it to MACHINE_ARCH and bump PR May 13 17:46:09 This recipe is clearly MACHINE_ARCH because: May 13 17:46:20 matchbox common contains matchbox-session: May 13 17:46:20 root@bug:~# opkg files matchbox-common | sort May 13 17:46:20 /usr/bin/matchbox-session May 13 17:46:21 ... May 13 17:46:21 And matchbox-session depends on matchbox-panel: May 13 17:46:22 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * rd306e0ebad 10openembedded.git/recipes/tslib/tslib.inc: tslib: convert to new staging and bump PR May 13 17:46:25 03John Connolly  07org.openembedded.dev * r8f7e4a5c3d 10openembedded.git/conf/machine/bug20.conf: May 13 17:46:25 conf/machine/bug20.conf, add xf86-input-tslib for BUG2.0 touchscreen May 13 17:46:25 *BUG20 is using tslib for xorg touchscreen. May 13 17:46:25 *would like this to get into stable May 13 17:46:25 *tested on BUG2.0 May 13 17:48:46 oops I reviewed everything etc...but forgott to add an acked by in John Connolly org.openembedded.dev * r8f7e4a5c3d openembedded.git/conf/machine/bug20.conf May 13 17:59:43 Tartarus: I found out that I was using DEBUG_BUILD = "1" in my local.conf which changes the OPTFLAGS from FULL_OPTIMIZATION to DEBUG_OPTIMIZATION May 13 17:59:52 which means I was compiling with -O -g May 13 17:59:57 and not with -Os -g May 13 18:00:07 Hi all May 13 18:00:33 Tartarus: I think we could improve two losely tied variables May 13 18:00:43 DISTRO_TYPE and DEBUG_BUILD May 13 18:01:43 so if DISTRO_TYPE="debug" then OPTLEVEL is set to DEBUG_OPTIMIZATION automatically and if DISTRO_TYPE="release" then its set to FULL_OPTIMIZATION May 13 18:02:39 so if I override DEBUG_BUILD in my local.conf and select DISTRO="release" it should ignore it May 13 18:02:49 right now it does not happen that way May 13 18:02:54 I plan to put a patch for that May 13 18:03:16 Tartarus: now I will try to rebuild gnutls with -Os May 13 18:08:12 khem: There we go May 13 18:08:14 thanks :) May 13 18:10:22 Tartarus: yeah you are not crazy :) May 13 18:16:25 re May 13 18:16:28 hi khem May 13 18:19:31 hey woglinde May 13 18:19:50 woglinde: have you had chance to try my qemuarm patch May 13 18:19:55 it boots well May 13 18:20:05 and all segfaults during boot are gone May 13 18:20:09 I am so happy May 13 18:20:30 with 2.6.25+26 I use to get quite a few segfaults May 13 18:20:37 didnt look nice in demos :) May 13 18:21:37 hm? May 13 18:21:43 sorry May 13 18:21:45 no time May 13 18:26:41 ...the days of rday are upon us... :) May 13 18:28:25 rday? May 13 18:29:15 Robert P.J. Day. May 13 18:30:23 broonie hm you were little bit absent the last time May 13 18:32:40 touching machine.conf means a 10min parsing/coffee break May 13 18:33:21 he makes us fix our silly bugs :) May 13 18:33:24 re May 13 18:33:26 hi May 13 18:33:32 woglinde: Hrm? May 13 18:35:28 broonie I meant I didnt see you a long time May 13 18:41:17 Crofton: could you catch hold of someone who could fix patchwork May 13 18:41:44 it seems its not getting emails from the mailing list. I dont know how it trools May 13 18:41:47 trolls May 13 18:41:49 woglinde: Ah, right. "recently"Â rather than "last time" May 13 18:42:04 woglinde: I've been here, just quiet :) May 13 18:45:23 hmmm http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44091 May 13 18:45:34 affects up as we use -Os and -mthumb May 13 18:45:41 s/up/us May 13 18:48:39 khem and no fix May 13 19:11:26 yes May 13 19:11:28 sadly May 13 19:15:37 khem, I'll try May 13 19:15:55 we need to figure out what email account is feeding it and see if we can figure out what is stuck May 13 19:16:29 we are going to a friends wedding this weekend, so I don't hvae much time May 13 19:16:55 Crofton|work: ok May 13 19:17:05 who set patchwork for us May 13 19:17:30 someone gave koen a vm or something May 13 19:17:54 Crofton|work: ok May 13 19:18:08 need to figure out where it runs and get a long in next week May 13 19:18:23 this is becoming annoyinh May 13 19:35:07 03Martin Jansa  07org.openembedded.dev * rf01994c18e 10openembedded.git/recipes/linux/linux-openmoko-shr-devel_git.bb: May 13 19:35:07 linux-openmoko-shr-devel: bump SRCREV for ar6000 fixes May 13 19:35:07 Signed-off-by: Martin Jansa May 13 19:40:31 Tartarus: as expected now I can see the same error May 13 19:41:59 yay May 13 19:44:09 libgnutlsxx.so.26.14.12 contains the refrence May 13 19:47:20 libgnutlsxx_la-gnutlsxx.o brings it in May 13 19:50:22 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r1625946510 10openembedded.git/recipes/initscripts/initscripts_1.0.bb: May 13 19:50:22 Revert "initscripts 1.0 : change it to MACHINE_ARCH and bump PR" May 13 19:50:22 This reverts commit 0e617baf2916269bbe02cb7b2d4eb20161c86547. May 13 19:50:22 This commit broke initscripts is in practically all images May 13 19:50:22 Thanks a lot Koen for notifying it to me. May 13 19:52:56 Tartarus: This patch seems to be a workaround. http://pastebin.com/HFbbBrJn May 13 19:53:12 heh May 13 19:53:30 but there might be more problems May 13 19:53:31 I'd almost rather just switch from -Os to -O2 for powerpc everywhere with a big old comment on why May 13 19:53:39 Or at least most ppc subarches May 13 19:53:56 Make the bandaid big and obvious so someone is more motivated to fix it May 13 19:54:20 I will send an email to ml asking for switch and if someone cares for size May 13 19:54:25 should speak up May 13 19:55:06 libtool thinks that you are compiling shared so he filters out libgcc and instead links iwth libgcc_s May 13 19:55:23 which is right otherwise PIC libraries may not work May 13 19:55:36 but with ppc -Os he has to link with both May 13 19:55:57 order being link all libgcc_s and remaining from libgcc.a May 13 19:58:29 So we either do -O2 or fix libtool (and fixup by hand any stragglers) May 13 19:58:34 To use -lgcc_s -lgcc or so? May 13 20:45:01 re kergoth May 13 21:10:25 hi. is there a way to get all the required/inherited structure for a recipe - similar to bitbake -g ...? May 13 21:11:05 ???? May 13 21:12:12 i dont want to see the dependencies DEPENDS... in a dot file, but the files included with require/inherit May 13 21:13:10 nothing like that exists today May 13 21:13:23 there is a list of the files loaded for a given recipe in a variable, __depends May 13 21:13:25 you can try looking at that May 13 21:13:27 python list May 13 21:15:01 kergoth: thanks. i will check this list. May 13 21:15:18 might be able to see it with bitbake -e, otherwise programmatically from a python task or anonymous python snippet May 13 21:23:13 is there a way to disbale do_stage for native packages? May 13 21:37:48 woglinde: i guess an empty do_stage task in your recipe will do it. May 13 22:43:12 re May 13 22:44:52 hmm May 13 22:46:13 * mwester-laptop hums along May 13 23:00:04 Tartarus: we need to fix libtool certainly May 13 23:27:09 03Khem Raj  07org.openembedded.dev * r682e67793d 10openembedded.git/ (3 files in 3 dirs): May 13 23:27:09 machine/qemuarm: Move kernel version to 2.6.33 May 13 23:27:09 * Tested using gcc 4.4.4 DISTRO=minimal May 13 23:27:09 Signed-off-by: Khem Raj May 13 23:27:20 03Khem Raj  07org.openembedded.dev * r63f6566a8f 10openembedded.git/conf/distro/include/sane-toolchain.inc: May 13 23:27:20 sane-toolchain.inc: Bump preferred versions for binutils, gcc and eglibc May 13 23:27:20 * Move gcc to 4.4.4, binutils to 2.20.1 and eglibc to 2.11 May 13 23:27:20 Signed-off-by: Khem Raj May 13 23:28:00 03Khem Raj  07org.openembedded.dev * r3b41668fc1 10openembedded.git/recipes/eglibc/ (eglibc_2.11.bb eglibc_svn.bb): May 13 23:28:01 eglibc: Bump SRCREV for 2.11 and trunk recipes May 13 23:28:01 Signed-off-by: Khem Raj May 14 02:33:48 PythonExpansionError: invalid literal for int() with base 10: 'test' while resolving PythonSnippet(["int('test')"]) via Value(['foo value ', VariableRef(['BAR'])]) -> Value(['bar value ', PythonSnippet(["int('test')"])]) -> PythonSnippet(["int('test')"]) May 14 02:33:53 thoughts? May 14 02:34:00 hmm May 14 02:34:21 yeah, think this'll do May 14 02:34:26 gives more context than what we do today, anyway May 14 02:43:35 Hi, ALL, May 14 02:44:03 I tried to compare the dynamic link dependencies on x86 and ARM May 14 02:44:44 the result of running "ldd " are different May 14 02:45:57 On x86, there are 3 more dependencies: libXi.so.6, libXrandr.so.2 and linux-gate.so.1 May 14 02:46:20 Can someone tell me what those libraries for? Thank u May 14 02:47:27 google is your friend **** ENDING LOGGING AT Fri May 14 02:59:56 2010