**** BEGIN LOGGING AT Wed Jan 16 02:59:57 2008 Jan 16 04:21:29 any body awake Jan 16 04:22:08 yep Jan 16 04:22:25 can't guarantee I'll be of any use though Jan 16 04:33:09 summatusmentis hi Jan 16 04:34:06 hey rushadrenaline how goes it? Jan 16 04:34:32 just corect me if im wrong ::: the purpose of oe is to build a package along with its dependencies based on a specific arch Jan 16 04:35:06 essentially, it's a build environment. Jan 16 04:35:10 a package, or an image Jan 16 04:35:49 but how does it create a rootfs Jan 16 04:36:03 i didnt come across anything like that in the manuals Jan 16 04:36:20 while my t.l. is saying that it can handle rootfs too Jan 16 04:40:35 rushadrenaline: You can use any toolset you wish to create the rootfs; it's just another bitbake recipe. The most common technique in OE is to use fakeroot Jan 16 08:16:24 morning Jan 16 08:48:37 hi Jan 16 08:48:40 can someone explain to me what ccache does? Jan 16 08:49:44 03florian 07org.oe.dev * r3a9c0f4a... 10/ (18 files in 2 dirs): gpephone: Unify all SVN recipes, fix several issues. Jan 16 08:49:49 03florian 07org.oe.dev * r54380a91... 10/ (1 packages/gpephone/libmsgenabler_svn.bb): libmsgenabler: add svn recipe Jan 16 08:49:54 03florian 07org.oe.dev * ref88c515... 10/ (1 packages/gpephone/libmsgenabler2_svn.bb): libmsgenabler2: add svn recipe Jan 16 08:50:00 03florian 07org.oe.dev * r29df705d... 10/ (4 files in 3 dirs): libim2: add svn recipe Jan 16 08:50:06 03florian 07org.oe.dev * r57ab5aa9... 10/ (4 files in 3 dirs): libiac2: add svn recipe Jan 16 08:50:24 03florian 07org.oe.dev * rea198c70... 10/ (1 packages/gpephone/libiac_svn.bb): libiac: add svn recipe Jan 16 08:50:30 03florian 07org.oe.dev * r8a0a3f2b... 10/ (1 packages/gpephone/libgemwidget_svn.bb): libgemwidget: add svn recipe Jan 16 08:50:38 03florian 07org.oe.dev * re850b717... 10/ (1 packages/gpephone/calendar2_svn.bb): calendar2: add svn recipe Jan 16 08:50:48 03florian 07org.oe.dev * rd2784f86... 10/ (1 packages/gpephone/liblipsevent2_svn.bb): liblipsevent2: add svn recipe Jan 16 08:50:55 03florian 07org.oe.dev * rdc68e3b4... 10/ (1 packages/gpephone/libalmmgr_svn.bb): libalmmgr: add svn recipe Jan 16 08:51:03 03florian 07org.oe.dev * rda498130... 10/ (1 packages/gpephone/almmgrd_svn.bb): almmgrd: add svn recipe Jan 16 08:51:10 03florian 07org.oe.dev * r1f9075ae... 10/ (1 packages/gpephone/dialer_svn.bb): dialer: add svn recipe Jan 16 08:51:18 03mickeyl 07org.oe.dev * r2462bf83... 10/ (3 files in 3 dirs): add gsm0710muxd Jan 16 08:51:29 03mickeyl 07org.oe.dev * rb7d3d5dc... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs.inc: add gsm0710muxd entry Jan 16 08:51:43 03mickeyl 07org.oe.dev * r7942b108... 10/ (3 files in 3 dirs): epdf cvs disable ewl test widget (which no longer compiles), bring back to task-openmoko-feed Jan 16 08:51:51 03Laibsch 07org.oe.dev * r0e55c1c0... 10/ (4 files in 2 dirs): ncurses: really fix packaging now. bump PR. Jan 16 08:52:05 03rwhitby 07org.oe.dev * r2817dae2... 10/ (5 files in 4 dirs): linux-ixp4xx: Updated to 2.6.23.14 Jan 16 08:52:12 03rwhitby 07org.oe.dev * r3d85f32a... 10/ (4 files in 3 dirs): linux-ixp4xx: Updated defconfigs for 2.6.23.14 Jan 16 08:52:17 03freyther 07org.oe.dev * rc0856cf5... 10/ (3 files in 3 dirs): (log message trimmed) Jan 16 08:52:17 packages/gcc/gcc-cross_4.2.2: Attempt to fix the fortran issue Jan 16 08:52:17 f951 was not executable because libmpfr and libgmp which were build by Jan 16 08:52:17 OpenEmbedded were not found. They resist in the staging lib directory and Jan 16 08:52:18 we do not have the LD_LIBRARY_PATH set. Jan 16 08:52:20 --enable-rpath for gcc was not an option, as we would have difficulties to Jan 16 08:52:22 say which path would be checked (we install to cross and not staging). Jan 16 08:52:24 03awilcox 07org.oe.dev * r2a876e0a... 10/ (1 packages/sccd/files/scc.h packages/sccd/sccd_1.0.bb): convert tty to udev name, start a bit earlier to avoid a race condition Jan 16 08:52:28 03awilcox 07org.oe.dev * r5c727796... 10/ (5 files in 4 dirs): Jan 16 08:52:31 bump storcenter kernel to 2.6.15.7, improved config, much Jan 16 08:52:33 smaller kernel patch. Jan 16 08:52:35 03awilcox 07org.oe.dev * r6dc2b7f1... 10/ (1 classes/concatenated-image.bbclass): A simple concatenated image class. Jan 16 08:52:42 03awilcox 07org.oe.dev * r3caac13a... 10/ (19 files in 14 dirs): Jan 16 08:52:42 Numerous openprotium fixes as we transition to udev based root Jan 16 08:52:42 and task based build Jan 16 09:12:36 cache compiler outputs Jan 16 09:12:57 so if you built xyz.o before it will get it from cache instead of building it again Jan 16 09:32:05 hrw: XorA: isn't that what make is for? to /not/ compile the same stuff again? Jan 16 09:32:12 no Jan 16 09:32:20 no? Jan 16 09:32:22 do 'make clean' and build then Jan 16 09:32:31 without ccache it will compile all Jan 16 09:32:39 with ccache it will use from cache Jan 16 09:32:40 okay, after a make clean i want a complete compile Jan 16 09:33:00 then do not use ccache Jan 16 09:33:04 : Jan 16 09:33:05 :) Jan 16 09:33:34 I am doing lot of builds and most of them have glibc for armv5te. without ccache it would take days Jan 16 10:17:31 good morning Jan 16 10:39:10 hi florian Jan 16 10:49:57 hi everybody Jan 16 10:51:51 yo thesing_ Jan 16 11:03:03 hey thesing Jan 16 11:03:08 how is it going? Jan 16 11:04:30 thesing_: Would it be OK for you if I applied my changes from bug 3664, but made them eject hostap cards only? Jan 16 11:06:26 Laibsch: better make it consistent and eject all cards. Those who care about connections which survive suspend can probably delete the file in apm.d. Jan 16 11:06:46 OK, fine with me Jan 16 11:10:16 Laibsch, for bug 3693: one only needs to add locomo-kbd to module_autoload in linux-rp.inc (I didn't test but loading module with serial console enables keyboard.) Jan 16 11:10:19 when build xfonts-xorg i encounter the following problem Jan 16 11:10:20 | ../../../exports/bin/ucs2any 4x6.bdf ../../../fonts/util/map-ISO8859-1 ISO8859-1 Jan 16 11:10:20 | ../../../exports/bin/ucs2any: No STARTFONT line found in '4x6.bdf'! Jan 16 11:10:20 | make[2]: *** [4x6-ISO8859-1.bdf] Error 1 Jan 16 11:10:25 anyone ever saw this? Jan 16 11:10:51 thesing_: Thanks for the hint Jan 16 11:10:53 but the 4x6.bdf font file contains on top a STARTFONT 2.1 line Jan 16 11:11:22 thesing_: I will give it a shot. 2.6.23 git version works fine for you? Jan 16 11:11:41 thesing_: no connection can survive a suspend of more than 2 mins anyway Jan 16 11:12:04 Laibsch, mmc does not work yet. Jan 16 11:13:19 bummer Jan 16 11:13:54 thesing_: I have not yet personally experienced bug 3693 Jan 16 11:14:05 Need to compile the latest kernel and flash Jan 16 11:14:18 XorA, maybe not if you use tcp. But udp should work if the if the layer 5 timeouts are high enough. Jan 16 11:14:49 thesing_: and as most apps open 0.0.0.0 they will survive the eject :-) Jan 16 11:15:08 RP: Can you add locomo-kbd to module_autoload in linux-rp.inc as suggested by thesing_ five minutes ago and close bug 3693. If you give me the go-ahead, I can try and do it, too. Jan 16 11:16:23 hi thesing_ Jan 16 11:16:37 hi hrw Jan 16 11:17:30 Laibsch: does that not mean keyboard does not work in altboot Jan 16 11:17:36 Laibsch: thats gonna get some moaning Jan 16 11:17:59 btw. I build an x11-image with squashfs-lzma and its small enough to fit on internal flash of collie. So we should think about a squashfs-lzma + unionfs solution for small flash machines IMO. Jan 16 11:18:43 XorA: I admit to not understanding the inner workings of altboot? Jan 16 11:18:54 Laibsch: altboot happens before init Jan 16 11:19:08 Laibsch: so /etc/modules doesnt get done yet Jan 16 11:19:32 Laibsch: so you'll have to make sure CoreDump|afk knows to add that to altboot for collie to at least load keyboard :-) Jan 16 11:19:46 XorA: maybe one could readd the module to the kernel. But I put everthing into modules that was possible. Jan 16 11:20:14 thesing_: Im just noting problems early rather than when the first altbooter starts raging :-) Jan 16 11:20:46 thesing_: until you finish collie work I doubt it matter anyway Jan 16 11:21:19 collie conversations tend to go "XXX: Collie doesnt work" "XorA: yes" Jan 16 11:39:38 Laibsch: Feel free to do it Jan 16 11:47:19 RP: what do you think about keeping machine related autoloads in machine configs? Jan 16 11:47:42 but now it would anyway require kernel bump ;( Jan 16 11:48:15 hrw: I agree its probably better than the kernel bb file... Jan 16 11:48:25 The whole thing needs rethinking though Jan 16 11:59:20 mickey|sports: thx Jan 16 12:25:32 mickey|sports: made me a bb and made my project "autotools" but it does not work - it's missing Makefile.in. autogen.sh did not run... does someone know what to do? Jan 16 12:51:36 thesing_, RP: module_autoload_locomo-kbd_collie = "locomo-kbd" Jan 16 12:51:40 correct? Jan 16 12:52:25 Is a PR bump for the 2.6.23 git version only sufficient? Jan 16 12:53:30 Laibsch, looks correct and PR bump should be enough. But better test it before committing ;) Jan 16 12:54:09 thesing_: I was wondering if other kernel bb files also need a PR bump Jan 16 12:54:29 thesing_: I cannot test yet since I always use altboot which still needs to be fixed Jan 16 12:54:48 collie is not yet working and you can't break something if it ain't fixed ;-) Jan 16 12:56:12 Laibsch, older linux-rp kernel recipes don't compile locomo-kbd as a module. Jan 16 12:57:00 OK Jan 16 13:02:08 03Laibsch 07org.oe.dev * r1b0acc21... 10/ (3 files in 3 dirs): apm: turn off wifi cards before suspend so they are fully reloaded upon resume. closes 3664. Jan 16 13:02:20 03mickeyl 07org.oe.dev * r0460728e... 10/ (1 conf/distro/include/sane-srcdates.inc): sane-srcdates.inc: bump EFL Jan 16 13:12:19 shoot, forgot to commit the PR bump Jan 16 13:12:53 hehe Jan 16 13:13:23 ggilbert: i did not receive anything wrt. bugtracker yet Jan 16 13:15:04 Laibsch: consider updating 2.6.23+git in OE to 2.6.23+2.6.2.4-rc6 from Poky Jan 16 13:15:20 thesing_: What do you say? Jan 16 13:15:26 You are collie kernel master Jan 16 13:16:32 Laibsch: I think we should do this. It's another step towards .24 Jan 16 13:17:03 I can commit it, but I don't have time or possibility to test it Jan 16 13:17:20 I have lots of other things still in the queue Jan 16 13:17:28 Commit as is - untested? Jan 16 13:19:13 its not default Jan 16 13:20:03 OK Jan 16 13:20:47 03Laibsch 07org.oe.dev * re7739cea... 10/ (1 packages/linux/linux-rp.inc): linux-rp.inc: autoload locomo-kbd on collie. partly closes 3693. Jan 16 13:21:54 03Laibsch 07org.oe.dev * re7739cea... 10/ (1 packages/linux/linux-rp.inc): linux-rp.inc: autoload locomo-kbd on collie. partly closes 3693. Jan 16 13:21:57 hrw: it is default for collie Jan 16 13:22:13 03Laibsch 07org.oe.dev * re7739cea... 10/ (1 packages/linux/linux-rp.inc): linux-rp.inc: autoload locomo-kbd on collie. partly closes 3693. Jan 16 13:22:18 03Laibsch 07org.oe.dev * r525bef44... 10/ (3 files in 2 dirs): linux-rp: bump PR for 2.6.23 kernels to go in line with e7739cea697d3013eb0d05d3812297cb7911994b Jan 16 13:23:45 Laibsch: not 2.6.20/21 anymore? Jan 16 13:25:03 it could be that angstrom-preferred-versions.inc still lists 2.6.20 Jan 16 13:25:06 not sure Jan 16 13:25:31 but the 2.6.23 kernel bb file in .dev has DEFAULT_PREFERENCE_collie = "1" Jan 16 13:25:59 I guess "you can't break it if it ain't fixed" applies, though Jan 16 13:26:05 collie is still broken Jan 16 13:27:05 <_saurabh_> Help needed with building a toolchain using oe Jan 16 13:27:21 <_saurabh_> when doing a bitbake meta-toolchain Jan 16 13:27:24 <_saurabh_> the Jan 16 13:27:44 <_saurabh_> gcc-initial is done to give the c compiler Jan 16 13:27:59 <_saurabh_> but the glibc build after that fails due to absence of a c++ compiler Jan 16 13:28:07 <_saurabh_> (arm-linux-g++ not found) Jan 16 13:28:15 <_saurabh_> have other ppl also faced this problem Jan 16 13:39:59 re Jan 16 13:46:23 thesing_: Please take a look at c794a426d2a55b0c4b5f1f6fee90c8272bd266e9, especially http://rafb.net/p/2frEpM33.html Jan 16 13:46:23 RP, maybe you want to use thesing_'s patches in poky instead of those from Dirk? Jan 16 13:47:50 thesing_: Do you want any of those SRC_URI changes reverted? Jan 16 13:47:55 Laibsch: collie isn't an officially supported machine in poky Jan 16 13:48:03 OK Jan 16 13:48:19 Laibsch: yes all of them. Jan 16 13:48:24 OK Jan 16 13:48:25 Laibsch: I do keep the kernels roughly in sync though :) Jan 16 13:48:36 thesing_: Anything else? Jan 16 13:48:55 wait a moment Jan 16 13:48:59 RP: Wouldn't it make sense to go with thesing_'s patches despite collie not being supporetd in poky? Jan 16 13:49:13 thesing_: I'll give you an mtn diff before the next commit Jan 16 13:49:48 Laibsch: Like I said, it will happen next time I sync things Jan 16 13:50:09 cool Jan 16 13:50:19 I did not get that drift from your comment Jan 16 13:52:33 Laibsch: ok. Jan 16 13:53:32 thesing_: Anything else you want reverted in that bb file? Jan 16 13:54:58 Laibsch: did you already commit something? didn't get a mail on oe-commits yet. Jan 16 13:55:29 just a minute ago Jan 16 13:55:42 It will probably take just a little bit more time Jan 16 13:56:02 or maybe five to ten minutes ago Jan 16 13:56:15 It should be available now Jan 16 13:56:39 thesing_: c794a426d2a55b0c4b5f1f6fee90c8272bd266e9 Jan 16 13:56:44 pb_: around? Jan 16 13:57:28 RP: off hand, do you know the difference between gcc-cross and gcc-cross-sdk (besides the inherit sdk) Jan 16 13:57:42 RP: is gcc-cross-sdk just meant to be packaged and installed into a different path? Jan 16 13:59:32 Laibsch: how can I get mtn to show me this change? Jan 16 14:03:07 03Laibsch 07org.oe.dev * rc794a426... 10/ (15 files in 3 dirs): linux-rp: update linux-rp-2.6.23+2.6.24-rc0+git to rc6 (from poky) Jan 16 14:03:29 anyway, I will patch gcc again and use -static for linking, this will break on OSX but who cares :) Jan 16 14:04:30 1) monotone-viz (very nice) Jan 16 14:04:37 thesing_: 1) monotone-viz (very nice) Jan 16 14:05:06 2) search for monotone on my blog http://blog.leggewie.org, I have some details about doing it with monotone alone Jan 16 14:05:48 3) http://www.openembedded.org, the viewmtn link, if necessary replace the URL with the rev you need Jan 16 14:06:00 Laibsch: ok if you revert the SRC_URI_append_collie stuff everthing should be fine. Jan 16 14:06:01 I usually use monotone-viz Jan 16 14:06:33 thesing_: Just reverting http://rafb.net/p/2frEpM33.html it will be, then. Jan 16 14:07:12 yes Jan 16 14:18:16 03florian 07org.oe.dev * r3f2d00b3... 10/ (3 files in 3 dirs): libgsmd-lips: Add modified gsmd with LiPS API Jan 16 14:18:28 03florian 07org.oe.dev * rbbbe0748... 10/ (1 packages/linux/linux-tornado-omap2/htctornado/defconfig): linux-tornado-omap2: Enable IP configuration on boot (patch by Patrick Steiner) Jan 16 14:18:37 03florian 07org.oe.dev * rbac6bab8... 10/ (6 files in 2 dirs): meta-toolchain*: Introduce name suffix for toolchain packages to avoid toolchain targets overwriting each other. Jan 16 14:18:42 03florian 07org.oe.dev * r0c2875b5... 10/ (1 packages/gpephone/libmsgenabler2_svn.bb): libmsgenabler2: Improve packaging Jan 16 14:18:47 03florian 07org.oe.dev * r9d6e0600... 10/ (1 packages/gpephone/libmsgenabler2_svn.bb): libmsgenabler2: More packaging improvements. Jan 16 14:34:29 morning Jan 16 14:35:23 BTW: anyone played with usb-host-mode of a N8X0? Jan 16 14:36:32 not yet, dont have the cable Jan 16 14:36:46 zecke|sad: Yes, it installs into SDK_PREFIX Jan 16 14:36:53 they used a non stadard connector Jan 16 14:37:32 zecke|sad: Only the 770 Jan 16 14:37:58 why is zecke sad anyway? Jan 16 14:38:16 XorA: The usb-host cable for the n800 is different than the usb-host cable for the Neo? Jan 16 14:38:41 mwester: the n810 is not a standard mini socket, its even smaller Jan 16 14:38:48 XorA: I was just about to ask that... Jan 16 14:39:28 Sigh. Yet another reason to stay with my n800. Why did Nokia go *backwards* in several areas with the n810?? Argh! Jan 16 14:40:40 yeah, miniSD choice was insane Jan 16 14:40:48 miniSD isnt even made anymore Jan 16 14:42:01 XorA: but they might still be made when they started the design :) Jan 16 14:42:38 chouimat|busy: they stopped being made before n800 was even in the works, I suspect they got the connectors cheap as they are useless to anyone else Jan 16 14:42:54 :) Jan 16 14:43:18 well, at least there are microSD -> miniSD adapters aren't there? Jan 16 14:43:45 mickeyl: yes, but they are rare and difficult to get, I had to buy a 2G microSD with adapter, just to be able to buy a 4g SDHC microSD Jan 16 14:44:12 mickeyl: purely because almost zero devices used miniSD, they went straight to micro Jan 16 14:56:58 RP: the best way to fix gcc, gcc-cross and gcc-cross-sdk is to link f951 statically Jan 16 14:57:06 RP: and i'm testing a patch to do now Jan 16 15:02:17 is any of you guys using gdbserver? Jan 16 15:03:32 thesing_, RP: fc51ba6c6f78e661b51a328e391e78d3fa49d2d0 are the current changes between poky and .dev for 2.6.24-rc6 Jan 16 15:04:42 Jin^eLD: I believe poky is Jan 16 15:05:01 have you seen anything like this? Jan 16 15:05:08 Program terminated with signal SIGTRAP, Trace/breakpoint trap. Jan 16 15:05:08 The program no longer exists. Jan 16 15:05:31 there was some post from 2006 in the redhat forums with a patch to glibc which supposedly fixed that problem, I applied it but it did not help Jan 16 15:05:37 I wonder if anyone knows more... Jan 16 15:05:55 03osas 07org.oe.dev * r5fd39dd5... 10/ (1 packages/meta/slugos-packages.bb): Jan 16 15:05:55 slugos: enable asterisk sound files for 1.4 version Jan 16 15:05:55 slugos: disable asterisk sound files for 1.2 version Jan 16 15:05:55 slugos: enable netatalk and ngrep Jan 16 15:06:00 03osas 07org.oe.dev * rcb269c66... 10/ (8 files in 6 dirs): add monit Jan 16 15:06:16 03osas 07org.oe.dev * r16c96161... 10/ (1 packages/meta/slugos-packages.bb): unrar promoted for slugos Jan 16 15:06:43 03florian 07org.oe.dev * rb54b6efc... 10/ (1 packages/linux/linux-tornado-omap2_2.6.16.16.bb): linux-tornado-omap2: bump PR Jan 16 15:06:48 03Laibsch 07org.oe.dev * rfc51ba6c... 10/ (1 packages/linux/linux-rp_2.6.23+2.6.24-rc6+git.bb): linux-rp: for collie, use thesing's patches instead of Dirk's in 2.6.24-rc6 Jan 16 15:28:02 Shouldn't collie have PACKAGE_EXTRA_ARCHS = "armv4 " and possibly others as well? Jan 16 15:28:57 Laibsch: probably Jan 16 15:29:10 anyone found the n810 linux sources? Jan 16 15:29:21 thesing_: which others? Jan 16 15:29:28 armv4t? Jan 16 15:30:53 I think t=thumb mode. And afaik collie doesn't support this. Jan 16 15:30:58 zecke|sad, for the kernel? Jan 16 15:34:13 johnx: yes, linux kernel, I doubt it is the *-rx-34 package Jan 16 15:34:47 in http://repository.maemo.org/dists/chinook/free/source/Sources it mentions the kernel source package Jan 16 15:35:15 and the N800 and N810 are the same hardware to the point that the N810 kernel will boot on an N800 Jan 16 15:39:11 thanks Jan 16 15:39:28 sure Jan 16 15:40:19 http://maemo.org/development/documentation/how-tos/4-x/kernel_guide_for_maemo.htmlhttp://maemo.org/development/documentation/how-tos/4-x/kernel_guide_for_maemo.html is bad and I will compile the kernel the usual way... Jan 16 15:40:42 zecke|sad: nice url :) Jan 16 15:41:17 pasted twice? ;) Jan 16 15:43:21 Laibsch: collie should have only 'arm' so no PACKAGE_EXTRA_ARCHS at all Jan 16 15:43:56 Laibsch: we do not support arm<4 so 'arm'==armv4l Jan 16 15:44:12 Laibsch: and 'arm' is TARGET_ARCH which is in PACKAGE_ARCHS by deafult Jan 16 15:44:32 OK, thanks for clarification Jan 16 15:45:20 wow Sun bought MySQL AB Jan 16 15:45:36 zecke|sad: yes. for 1blnUSD Jan 16 15:45:53 heh, that's the gdbserver problem.. http://sourceware.org/ml/gdb/2007-11/msg00274.html Jan 16 15:45:59 I guess I should try to compile from head Jan 16 15:47:04 zecke|away: at your service Jan 16 15:47:07 oh, now you're away Jan 16 15:47:42 pb_, zecke|sad has multiple personalities today Jan 16 15:48:34 ah Jan 16 15:48:36 pb_: f951 is a tool used by gfortran Jan 16 15:48:39 zecke|sad: why are you sad? Jan 16 15:48:43 pb_: it is using mpfr and gmp Jan 16 15:48:47 pb_: long story Jan 16 15:49:25 pb_: now our cross toolchain resists in cross/ and mpfr and gmp in staging/ Jan 16 15:49:45 pb_: adding a rpath in f951 to staging/i686-linux/lib would work for gcc-cross but is ugly Jan 16 15:50:02 pb_: but then we have gcc and gcc-cross-sdk left Jan 16 15:50:17 zecke|away: doesn't gcc-cross already have that rpath? Jan 16 15:50:18 pb_: my current preferred fix is to link mpfr and gmp statically Jan 16 15:50:33 gcc isn't executable on the build host so presumably the rpath (and the whole of staging) is irrelevant there Jan 16 15:50:41 likewise gcc-cross-sdk Jan 16 15:51:36 pb_: right, I wonder what gcc will do but I'm still compiling :) Jan 16 15:52:24 pb_: for gcc-cross and gcc-cross-sdk a copy of f951 is executed (at least for an autoconf test) and can fail if no suitable mpfr and gmp can be resolved by the linker, the rpath to staging is 'fixing' that Jan 16 15:52:55 oh, right Jan 16 15:52:57 pb_: for gcc I'm currently building and waiting on a failure, so my preferred solution is -static when linking f951 Jan 16 15:53:15 don't you have staging/BUILD_SYS/lib in LD_LIBRARY_PATH when configure runs? Jan 16 15:53:25 I thought that was always the case, otherwise anything that depends on a staged library would lose. Jan 16 15:53:32 it isn't obvious to me why f951 is special Jan 16 15:53:51 pb_: we don't have LD_LIBRARY_PATH and most (all) tools have rpath to the staging/lib Jan 16 15:54:10 ah, fair enough Jan 16 15:54:15 I must have remembered it wrong Jan 16 15:54:32 Do I have to be logged in to report a bug on the openmoko gforge bug tracker? Jan 16 15:54:42 not the openmoko bugzilla Jan 16 15:54:42 if the other tools all use rpath then I guess f951 should do the same Jan 16 15:55:01 but http://projects.openmoko.org/tracker/?atid=329&group_id=62&func=browse Jan 16 15:55:21 that sounds like a question for the openmoko people. this isn't a distro support channel, right? Jan 16 15:56:05 well, you are right of course Jan 16 15:56:15 But the reason I ask here is because I have nothing to do with OM Jan 16 15:56:21 and thus no login for example Jan 16 15:56:32 I wonder if I should report it to bugs.oe.org Jan 16 15:56:41 but it might not get noticed in there Jan 16 16:05:39 03Laibsch 07org.oe.dev * r7be93ec1... 10/ (1 packages/pwsafe packages/pwsafe/pwsafe_0.2.0.bb): Jan 16 16:05:39 add pwsafe to .dev Jan 16 16:05:39 closes 1065 Jan 16 16:05:42 bye Jan 16 16:05:53 * RP has wondered if we shouldn't add a LD_LIBRBARY_PATH="staging/BUILD_SYS/lib" to native.bbclass Jan 16 16:07:17 * mwester seems to recall that he tried that, and soemthing about it didn't work... Jan 16 16:08:35 I mean bitbake.conf... Jan 16 16:08:45 native.bbclass wouldn't be enough... Jan 16 16:10:26 I definitely tried bitbake.conf; it didn't work there. The reason seemed clear upon investigation, and I think I came to the conclusion that bitbake code would need to be changed to support it. Jan 16 16:14:15 mwester: What kind of changes? Jan 16 16:17:35 RP: yeah, I had thought that bitbake.conf had that already (same as it does for PATH). It's obviously not risk free, in that the libraries from staging would then override the host system's own ones from /usr/lib (with which they aren't necessarily all that compatible) but I guess it would work OK most of the time. Jan 16 16:18:04 :( I don't recall specifically, but it made sense at the time. I think the problem was that setting an env. variable in a bbclass and having it "stick" through other bbclasses was not working, IIRC it had to do with scope. Jan 16 16:20:19 RP: all that said, though, I think zecke's approach of just putting an appropriate -rpath into all the staged binaries is probably a better one. I would have thought it would be possible to fiddle the linker in native.bbclass so that all the staged binaries were automatically rpathed without needing any per-package hackery. Jan 16 16:21:26 pb_: Yes, that would make more sense than the LD_LIBRARY_PATH sledgehammer... Jan 16 16:22:58 * mwester agrees Jan 16 16:57:23 is there a way to disable psplash without changeing the kernel boot args? Jan 16 17:01:24 bye, flying to oslo soon(tm) Jan 16 17:03:33 03thebohemian 07org.oe.dev * r776a345f... 10/ (4 files in 3 dirs): Jan 16 17:03:33 cacaoh-native: Small cleanup in include file. Jan 16 17:03:33 cacao: Added N800 quirk and a patch to fix AttachCurrentThread behavior. Jan 16 17:06:16 Morning Jan 16 17:07:40 kergoth: ping Jan 16 17:07:49 ggilbert_: pong Jan 16 17:09:49 kergoth, I have a temporary host set up for the bug tracker at 76.74.156.109. It's very unlikely to be permanent, but it'll be around for about a month until I find a new permanent home. Jan 16 17:15:17 It doesnt do any emailing at this point. I just wanted to put up something that'd function in an emergency and not end up leaving me with a thousand dollar bill if I massively misguessed on how much traffic I do :p Jan 16 17:17:55 ggilbert: no mail? Jan 16 17:18:05 that is unfortunate Jan 16 17:18:13 Can that be fixed in the next few days? Jan 16 17:18:24 Sure sure. Jan 16 17:18:39 Nice Jan 16 17:20:53 03thebohemian 07org.oe.dev * r776a345f... 10/ (4 files in 3 dirs): Jan 16 17:20:53 cacaoh-native: Small cleanup in include file. Jan 16 17:20:53 cacao: Added N800 quirk and a patch to fix AttachCurrentThread behavior. Jan 16 17:27:47 03thebohemian 07org.oe.dev * r776a345f... 10/ (4 files in 3 dirs): Jan 16 17:27:47 cacaoh-native: Small cleanup in include file. Jan 16 17:27:47 cacao: Added N800 quirk and a patch to fix AttachCurrentThread behavior. Jan 16 17:36:14 hello zecke told me to compile angstrom-2007.1-legacy for a start(because my distro config file fails and i'm a newbee) it worked...(bitbake boostrap-image) but now what should i do to run it with qemu because it doesn't work(mabe the wrong arm eabi?) Jan 16 17:36:44 here's my command: qemu-system-arm -m 250 -kernel zImage-2.6.23-qemuarm-20080116140313.bin -hda Angstrom-bootstrap-image-glibc-ipk-2008.1-test-20080116-qemuarm.rootfs.ext3 Jan 16 17:40:08 found it wrong -M Jan 16 17:43:41 now i have a problem with the filesystem...no partitions tables...i must create them Jan 16 17:56:34 after creating them i saw no offset option to mkfs.ext3... Jan 16 17:59:51 found losetup offset option Jan 16 18:10:34 qt4 recipe looks broken. qt4 when build puts a wrong qmake QT_INSTALL_LIBS information Jan 16 18:10:49 any idea where it might to be? Jan 16 18:49:10 anyone seen that? Jan 16 18:49:11 | paranoia.o:(.text+0x25c): relocation truncated to fit: R_ARM_PLT32 against symbol `free@@GLIBC_2.4' Jan 16 18:49:21 cdparanoia does not really want to compile Jan 16 18:59:51 hello, i've sucessfully built angstrom but i have I/O errors on qemu... Jan 16 19:05:09 ok that's odd Jan 16 19:05:23 after a couple of retries -c rebuild cdparanoia did build after all Jan 16 19:05:24 but Jan 16 19:05:27 I did not change anything! Jan 16 19:05:33 do you have gdb installed? Jan 16 19:05:45 oops gdb is for a different window :) Jan 16 19:08:21 hmm avahi just failed to buidl Jan 16 19:09:00 I think cdparanoia needs a PARALLEL_MAKE = "" Jan 16 19:09:05 after adding that it always builds Jan 16 19:13:15 hello Jan 16 19:13:19 hey likewise Jan 16 19:13:26 Jin^eLD: hey Jan 16 19:13:32 likewise: btw, regarding this PR, zecke had an idea Jan 16 19:13:53 he said i should try using FILES or FILESDIR or something to point to the real device table Jan 16 19:14:02 I did not try that yet though Jan 16 19:14:06 Jin^eLD: yes, I see. Jan 16 19:14:39 Jin^eLD: it's not a very transparent solution though. Jan 16 19:15:19 Jin^eLD: one package would then refer to another on the level of bitbakes, without even knowing if both packages are applied. Jan 16 19:15:42 true Jan 16 19:16:47 can't the .bb file access the devtable variable that is used by the rootfs.bbclass ? Jan 16 19:17:27 Jin^eLD: the dev table variable must be assumed to be globally available in OE. that's not a problem. Jan 16 19:18:33 Jin^eLD: what was the OE bug nr? I need to refresh my memory on this. Jan 16 19:18:49 bugtracker is down Jan 16 19:19:13 Bug tracker ip is in the topic Jan 16 19:19:17 3593, found it in the email Jan 16 19:19:24 oh, thanks Jan 16 19:20:48 configure: error: GMP 4.1 and MPFR 2.2.1 or newer versions required by fortran| FATAL: oe_runconf failed Jan 16 19:20:53 gcc-4.2.2-r3 is failing when doing meta-toolchain Jan 16 19:21:03 wasn't there some fortran discussion on the OEML? Jan 16 19:21:50 Jin^eLD: you might need to add that ip to your /etc/hosts file... otherwise re-directs will fail until the DNS change can propagate through Jan 16 19:22:27 k thx Jan 16 19:22:50 It's not going to be the permanent IP, so make sure you take it out when we move again :p Jan 16 19:22:51 hmm, the error above.. funny since gmp and mpfr seem to be available and the versions are ok Jan 16 19:23:04 did anyone try to build the meta-toolchain lately? Jan 16 19:28:14 Jin^eLD, have you looked at how slugos solves this problem? I just duplicated their approach in openprotium. Jan 16 19:29:00 http://76.74.156.109/show_bug.cgi?id=3636 Jan 16 19:29:19 HopsNBarley: no, what do they do? Jan 16 19:31:32 ggilbert: i did not receive anything wrt. bugtracker yet. did you send something? Jan 16 19:31:34 wtf Jan 16 19:31:38 mickeyl: nope Jan 16 19:31:49 ah, ok. thought it may be stuck somewhere Jan 16 19:32:00 http://76.74.156.109/ Jan 16 19:32:20 oh, there it is again! Jan 16 19:32:24 where is that hosted? Jan 16 19:32:33 kergoth: ping Jan 16 19:32:39 leached wireless :) Jan 16 19:32:48 I'm renting a server for a month while I sort out what my permanent hosting plans will be Jan 16 19:33:10 ok, cool Jan 16 19:33:10 heh, gcc-cross-sdk tries to create a directory in /opt/xxx (sdk prefix probably) Jan 16 19:33:17 and of course gets a permission denied Jan 16 19:33:37 mickeyl: This time we have notice though, so we can do a relatively sane transfer of the site :p Jan 16 19:34:03 right :) Jan 16 19:37:44 anyone good with the cross-sdk things? Jan 16 19:38:32 gcc4-build-sdk.inc - first line in do_configure says # Work around Hardcoded path assumptions in gcc and then it tries to create a directory in /opt Jan 16 19:38:48 well, actually it uses ${SYSROOT} Jan 16 19:38:51 but that can't work out Jan 16 19:41:30 are OEML also down? Jan 16 19:42:29 If it is, it's totally not my fault :) Jan 16 19:42:43 :)) Jan 16 19:42:49 no I just wanted to check before posting Jan 16 19:43:16 each time I try to build the meta toolchain something fails, I'd like to help to work it out but I need help in doing that :) Jan 16 19:43:17 DAMN Jan 16 19:43:44 Why do ipks randomly end up in different deploy folders Jan 16 19:43:45 what's failing? Jan 16 19:43:46 ? Jan 16 19:44:03 Not a do_fetch, right? Jan 16 19:44:16 ggilbert: with the sdk stuff? no, no Jan 16 19:44:20 kk Jan 16 19:44:28 always issues related to compiling it Jan 16 19:44:35 And -native builds getting built in work/sh3-linux instead of work/i686-linux Jan 16 19:52:46 hi ue Jan 16 20:30:09 ~seen RP Jan 16 20:30:14 rp is currently on #maemo (2h 21m 56s) #handhelds (2h 21m 56s) #oe (2h 21m 56s) #openmoko (2h 21m 56s) #openzaurus (2h 21m 56s), last said: 'pb_: Yes, that would make more sense than the LD_LIBRARY_PATH sledgehammer...'. Jan 16 20:32:23 damn ... trough expedia seem astrid hotel is full booked :( Jan 16 20:32:53 how about on their website? Jan 16 20:33:54 it seem free but it cost more and isn't possible to book without breakfast ... Jan 16 20:34:15 maybe call and ask? Jan 16 20:34:44 yes .. now i'm looking with other booking site Jan 16 20:53:10 mickeyl: is the astrid the official team opie hotel again for fosdem this year? Jan 16 21:01:25 pb_, yes Jan 16 21:01:45 well, team OE :) Jan 16 21:38:56 re Jan 16 22:34:27 pb__: absolutely! we'd be glad to share breakfast with you! Jan 16 23:06:49 HelgeD: Moin :) Jan 16 23:08:35 * Crofton reads about templates Jan 16 23:15:05 Someone knows freyther@openembedded.org's nickname? Jan 16 23:15:25 otavio: zecke Jan 16 23:15:33 hvontres|work: Hi! Jan 16 23:16:03 zecke|away: it looks like you forgot something while changing qt recipes. Jan 16 23:16:22 zecke|away: qmake -query QT_INSTALL_LIBS return a wrong value Jan 16 23:17:36 mickeyl: thank you! Jan 16 23:17:47 np Jan 16 23:20:14 HelgeD: Thanks for all the poodle bug reports :) Any chance you would like to be co-mentor? Jan 16 23:31:13 zecke|away: the problem isn't qt but qmake2-native... fixing it and will send you the patch Jan 16 23:39:19 otavio: please mail to zecke@selfish.org, i'm not sure whether freyther@openembedded.org is valid as email address Jan 17 00:29:52 morning all Jan 17 00:30:30 morning Jan 17 01:07:03 03likewise 07org.oe.dev * rcb0f0a06... 10/ (1 packages/linux/linux_2.6.23.bb): linux-2.6.23: Apply kallsyms-missing-include.patch. Jan 17 01:07:08 03mickeyl 07org.oe.dev * r9d982d51... 10/ (3 files in 3 dirs): gsm0710muxd svn fix dependencies, bump rev Jan 17 01:07:15 03awilcox 07org.oe.dev * r6c9956ec... 10/ (1 conf/distro/openprotium.conf): Make the FEED_ARCH selection a bit more intelligent. Jan 17 01:07:19 03thebohemian 07org.oe.dev * r06c77b62... 10/ (1 conf/machine/nokia770.conf conf/machine/nokia800.conf): Jan 17 01:07:19 nokia800: Add screen size. Jan 17 01:07:19 nokia770: Add screen size. Jan 17 01:07:34 03jluebbe 07org.oe.dev * r7de9efd2... 10/ (5 files in 4 dirs): chumby: Add machine definition and kernel recipe Jan 17 01:08:16 yay chumby is in! **** ENDING LOGGING AT Thu Jan 17 02:59:56 2008