**** BEGIN LOGGING AT Sat Aug 28 02:59:57 2010 Aug 28 06:44:47 gm Aug 28 07:00:35 03Frans Meulenbroeks  07org.openembedded.dev * r9d9c578a41 10openembedded.git/recipes/libexif/ (4 files): Aug 28 07:00:35 libexif: removed insecure versions Aug 28 07:00:35 versions 0.6.16 and earlier have several security advisories on them Aug 28 07:00:35 and no one pinned these, also some still had legacy staging Aug 28 07:00:35 so put these to rest Aug 28 07:00:36 Signed-off-by: Frans Meulenbroeks Aug 28 07:00:37 03Frans Meulenbroeks  07org.openembedded.dev * rc4d32c99f5 10openembedded.git/recipes/libexif/ (libexif_0.6.17.bb libexif_0.6.19.bb): Aug 28 07:00:37 libexif: moved to latest version (0.6.19) Aug 28 07:00:38 Signed-off-by: Frans Meulenbroeks Aug 28 07:32:38 eFfeM: hi Aug 28 07:33:04 eFfeM: I fixed the ppc/spe issue in gcc 4.4.4 Aug 28 07:33:36 update and try it out. I could build native-sdk-image for calamari/minimal without issues Aug 28 07:39:55 03Khem Raj  07org.openembedded.dev * ra1fa9b46f5 10openembedded.git/recipes/dspgw-utils/dspgw-utils_3.3.1.bb: Aug 28 07:39:56 dspgw-utils_3.3.1.bb: Remove do_stage and patch=1 Aug 28 07:39:56 Signed-off-by: Khem Raj Aug 28 07:58:01 khem nice, Aug 28 07:58:20 (was afk) Aug 28 07:59:28 khem I think i overlooked the patch this morning, will start a test Aug 28 07:59:51 03Petr Vanek  07org.openembedded.dev * r9c8212bbf9 10openembedded.git/recipes/omnewrotate/omnewrotate_svn.bb: Aug 28 07:59:51 omnewrotate: SRCREV bump to r67 for 2.6.32 kernel paths Aug 28 07:59:51 Signed-off-by: Martin Jansa Aug 28 08:05:27 is there a way with KERNEL_VERSION or so to test in a program to find out the version Aug 28 08:05:37 kernel changed the name of a header field :-( Aug 28 10:29:31 khem, tested the new patch, builds fine for me (can't test as the system is at work), thanks alot! Aug 28 11:31:34 GNUtoo|laptop: hi Aug 28 11:31:42 leviathan, hi Aug 28 11:31:49 oe problem? Aug 28 11:31:54 nope Aug 28 11:31:58 ok Aug 28 11:32:00 wrong channel ^_^ Aug 28 11:32:07 just wanted to say hello Aug 28 11:32:09 mom Aug 28 11:32:14 * leviathan switches channel Aug 28 11:32:24 if no oe problem, you should say hello in #openmoko-cdevel Aug 28 13:31:39 anyone here successfully set up a "production" oestats server using the instructions on the bolloretelecom site? Aug 28 13:33:33 he Aug 28 13:33:36 url? Aug 28 13:33:51 you want a gumstix specific tinderbox? Aug 28 13:34:22 I followed the process, but my installation doesn't quite work :-( Aug 28 13:34:35 http://build.sakoman.com Aug 28 13:34:55 Crofton|work: no, not for gumstix. just for my own use Aug 28 13:35:01 ah Aug 28 13:35:27 it might be nice to have one for gumstix to collect data from everyone using that tree Aug 28 13:36:14 Crofton|work: not sure I am willing to pay for that myself ;-) Aug 28 13:36:25 but I will suggest it to Gumstix Aug 28 13:36:46 wait until you figure out how to do it :) Aug 28 13:38:46 Crofton: I've done it previously using the "development server" instructions with no issues Aug 28 14:38:46 03Philip Balister  07org.openembedded.dev * r12bcc62dba 10openembedded.git/recipes/python/python-cheetah_2.0.1.bb: python-cheetah: Add native support. Aug 28 15:08:16 Crofton|work: do you know who set up the tinderbox oestats server? Aug 28 15:13:58 I need a timezone file for my dreambox, I guess OpenPLI gets all of theirs from you, any way you guys can help me out? Aug 28 15:18:01 zecke I think Aug 28 16:25:24 sakoman: you can also ask Jeremy Laine Aug 28 16:32:26 ah yes Aug 28 16:33:19 * Crofton|work bets many people are winding up their summer holiday though Aug 28 16:48:32 * kergoth really doesn't understand OE folks lately - try to be nice on the mailing lists, and people reply as assholes - screw it Aug 28 17:36:42 hello? Aug 28 17:36:52 Crofton: hello Aug 28 17:36:55 heh Aug 28 17:36:59 Internet had failed Aug 28 17:37:17 did you restored a backup of the internet? Aug 28 17:37:30 yes Aug 28 17:37:49 :) Aug 28 17:38:07 I actually had t read the cmake man page to solve a problem Aug 28 18:04:41 hi, while trying to fix ncurses_5.7 's sdk build, I meet the following problem : I reach a point where I now manage to build ncurses-sdk 's ipk fine but I found that when building meta-toolchain (using task-sdk-base) the ncurses-sdk is not installed (no libncurses & co in usr/local/angstrom/arm/lib) Aug 28 18:04:56 I went backk to ncurses_5.4 and it's the same Aug 28 18:05:30 I'm using bitbake 1.10.0 and angstrom-2010.x distro Aug 28 18:05:44 host it i686, target is armv6 Aug 28 18:06:28 before, meta-toolchain generated a sdk with : libz, libopkg and libncruses in usr/local/angstrom/arm/lib, now libncurses are missing Aug 28 18:06:42 any idea where to dig for this problem ? Aug 28 18:07:27 I can send the changes I did to ncurses_5.7.bb (and class/sdk.bbclass to enable virtclass-sdk override) if that helps Aug 28 18:11:50 03Petr Vanek  07org.openembedded.dev * ref825fcb18 10openembedded.git/recipes/openmoko-3rdparty/atrack_svn.bb: Aug 28 18:11:50 atrack: SRCREV bump to r128, adds aprs symbols, fix missing TODO file Aug 28 18:11:50 Bump Atrack to the latest revision. Adds APRS symbols and some bugfixes. Aug 28 18:11:50 Petr Vanous Vanek Aug 28 18:11:50 Signed-off-by: Petr Vanek Aug 28 18:11:51 Signed-off-by: Martin Jansa Aug 28 18:26:11 ericben: is it only libncruses thats missing or all of libz, libopkg and libncruses Aug 28 18:27:19 so you guys don't make your own "packages_timezones_files_zoneinfo"? Aug 28 18:27:43 JediMaster420: what do you mean Aug 28 18:27:58 JediMaster420: we have tzdata packages Aug 28 18:28:15 right, but it's missing the time zone for where I live Aug 28 18:28:46 well thats sad, but if you know whats missing and even better know to patch it Aug 28 18:28:49 send it to ml Aug 28 18:29:11 all the ones I see are binaly, do you have source for them? Aug 28 18:29:17 biniary* Aug 28 18:29:37 I sent it to the issues list yesterday but hasn't even been added to archive Aug 28 18:33:36 03Khem Raj  07org.openembedded.dev * re9d84bea6e 10openembedded.git/recipes/gdb/ (gdb-7.1/renesas-sh-native-support.patch gdb_7.1.bb): Aug 28 18:33:36 gdb_7.1.bb: Add support to build gdb for SH4 natively. Aug 28 18:33:36 Signed-off-by: Khem Raj Aug 28 18:33:37 03Khem Raj  07org.openembedded.dev * r32f795b407 10openembedded.git/ (2 files in 2 dirs): Aug 28 18:33:38 omap3-pandora.conf/u-boot-omap3pandora_git.bb: Use PREFERRED_PROVIDER_virtual/bootloader to select u-boot recipe to use. Aug 28 18:33:38 * Dont PROVIDES u-boot but let user decide it. Aug 28 18:33:38 Signed-off-by: Khem Raj Aug 28 19:10:15 khem: only libncurses (and associated libs), zlib & libopkg are present Aug 28 19:12:40 03Martin Jansa  07org.openembedded.dev * ra12fa8e6a1 10openembedded.git/recipes/openmoko-3rdparty/omhacks_git.bb: Aug 28 19:12:41 omhacks: bump SRCREV Aug 28 19:12:41 Signed-off-by: Martin Jansa Aug 28 19:50:08 | cp: cannot stat `/home/builds/fresh/jornada7xx/tmp/sysroots/x86_64-linux/usr/armv4/arm-linux-gnueabi/libc/lib/*': No such file or directory Aug 28 19:50:11 NOTE: package external-toolchain-csl-1.0-r6: task do_install: Failed Aug 28 19:50:12 any hints maybe ;)? Aug 28 19:50:59 it happens for ARM and for SH3 Aug 28 19:51:18 for mips too Aug 28 20:07:54 hi, Aug 28 20:08:03 I've an issue with providers Aug 28 20:08:10 there is fso-gpsd and gpsd Aug 28 20:08:16 SHR supports both Aug 28 20:08:45 most recipes use one or the other depending on the recipe or the distro Aug 28 20:08:57 I bet we should convert that to a virtual provider Aug 28 20:09:21 how should I do that? Aug 28 20:11:11 khem, ping Aug 28 20:11:39 the manual is not very clear and I would not like to make a mess Aug 28 20:11:49 of course I'll send a patch Aug 28 20:12:33 anyway it has some issue Aug 28 20:12:42 if let's say I want fso-gpsd for om-gta02 Aug 28 20:12:46 and gpsd for htcderam Aug 28 20:13:02 as JaMa|GoNe pointed out all the respective packages must be machine arch for htcdream Aug 28 20:13:11 if fso-gpsd is the distro's choice Aug 28 20:51:12 do we have a very simple example of srctree inside OE metadata? I.e. the helloworld equivalent, where helloworld.c lives in the recipes/ subdirs? Aug 28 20:51:59 Crofton_|work: hi Aug 28 20:52:32 likewise: look at emails from Cliff Aug 28 20:52:38 he has some exampls Aug 28 20:52:49 likewise: http://cgit.bec-systems.com/cgit.cgi/autotools-demo/tree/ Aug 28 20:53:08 ericben: ok only with libncurses I see Aug 28 20:53:26 ericben: and its same issue no matter if its 5.4 or 5.7 Aug 28 20:53:36 likewise, http://pastebin.com/bNHUVQYh Aug 28 20:53:50 khem: yes : in both cases, nothing from curses-sdk is installed in the sdk Aug 28 20:55:17 khem: in 5.7 there is python populate_packages_prepend which splits the libs in several packages. This is fine for packages which will be installed on the target but for the sdk I think everything should go into ncurses-sdk's ipk : do you have an idea how to bypass this python populate_packages_prepend when building ncureses-sdk ? Aug 28 20:56:34 khem, Crofton_|work: I used cliff's examples, however, when I directly bitbake a srctree recipe, the toolchain isn't built. Aug 28 20:57:35 likewise: now if you create same recipe without srctree does it work Aug 28 20:57:45 I mean does it pull toolchain Aug 28 20:59:20 GNUtoo|laptop: machine choice should override distro choice imo Aug 28 20:59:37 ok Aug 28 20:59:57 I'm implementing the immodule Aug 28 21:00:00 for now Aug 28 21:00:09 I'll do the gpsd just after Aug 28 21:00:15 but is it the good way to go? Aug 28 21:00:16 khem, have you had a chance to think about my c++ issues? Aug 28 21:00:51 Crofton_|work: I am not yet convinced if its really C++ Aug 28 21:00:58 Crofton_|work: you have feed more data Aug 28 21:01:16 you have to feed more data I mean Aug 28 21:01:26 well, lets start with the -dev package not depending on the libstdc++ package ... Aug 28 21:01:27 did you succeed with profiling Aug 28 21:01:31 not yet Aug 28 21:01:34 ok Aug 28 21:01:41 there are a set of problems Aug 28 21:01:46 right Aug 28 21:01:57 I don't understand the gcc recipes enough Aug 28 21:01:57 how did it behave with 4.3.3 Aug 28 21:02:02 and am not sure i want :) Aug 28 21:02:02 did it work Aug 28 21:02:05 yes Aug 28 21:02:09 oh Aug 28 21:02:10 nice Aug 28 21:02:39 so you chose gcc 4.3.3 and everything else remained same and built from scratch all was ok ? Aug 28 21:02:51 well, I won't say that Aug 28 21:03:00 so tell me what you did Aug 28 21:03:12 I the depends worked a couple of months back Aug 28 21:03:26 then I repulled and switched to angstrom-2010.x Aug 28 21:03:31 hmm couple of months is a long time Aug 28 21:03:41 yeah I know Aug 28 21:03:51 we changed toolchain layout etc. Aug 28 21:03:53 although I suspect people would have noticed Aug 28 21:03:59 that could cause such subtle issues Aug 28 21:04:06 but I have to sure where to look at Aug 28 21:04:17 true Aug 28 21:04:19 if there are too many variables things are hard Aug 28 21:04:23 do you have a beagleboard? Aug 28 21:04:27 thats why I asked you Aug 28 21:04:42 choose gcc 4.3.3 once and choose 4.5 other time Aug 28 21:04:55 do two builds from same metadata Aug 28 21:05:04 and then see if it works still with 4.3.3 Aug 28 21:05:08 yeah, I'll need some free time though Aug 28 21:05:13 that is in short supply atm Aug 28 21:05:41 I'll see what I can do Aug 28 21:05:51 Do you have commit id that worked for you Aug 28 21:05:57 sorry no, Aug 28 21:06:02 It was a while back Aug 28 21:06:27 hmm it should not be hard to pin point but will need some info Aug 28 21:06:41 let me try an angstrom2008 build tonight Aug 28 21:06:44 khem: I'm launching the meta-toolchain build using angstrom 2008.1 distro to see if the problem also exists here (vs 2010.x), that was working a few weeks ago Aug 28 21:06:51 hmm, do_compile() is missing in my recipe ... (duh moment) Aug 28 21:07:00 likewise: :) Aug 28 21:07:08 ericben, do you use c++? Aug 28 21:07:26 PEBKAC Aug 28 21:07:43 Crofton_|work: I'm using qt4 so yes I have c++ Aug 28 21:07:55 Crofton_|work: Aug 28 21:08:07 its ok when you do cross build Aug 28 21:08:22 I've had issues with the depends being funny Aug 28 21:08:41 Crofton_|work: btw. do you use bb form git/master Aug 28 21:08:47 I have to force libstdc++ into the image, even if the image install libstdc++-dev Aug 28 21:08:51 if not then I will reccommend Aug 28 21:08:55 1.10.0 here Aug 28 21:09:10 ah so now you descibe the problem nicely Aug 28 21:09:27 heh Aug 28 21:09:29 I have several Aug 28 21:09:38 so you pick libstdc++-dev Isee Aug 28 21:09:47 yes Aug 28 21:09:50 well lemme try that Aug 28 21:09:59 stupid irc :) Aug 28 21:10:26 yeah I have stated using google video Aug 28 21:10:42 cant keep typing with thick fingers on small keyboards Aug 28 21:10:51 heh Aug 28 21:10:54 watchign cricket? Aug 28 21:11:10 heh sure India poopoo'ed today Aug 28 21:11:21 and Pakistan did same against englang Aug 28 21:11:25 england Aug 28 21:12:39 hi, Aug 28 21:12:47 I've to work on a python part that is in: Aug 28 21:13:00 python populate_packages_prepend () { Aug 28 21:13:05 that's far in the tasks Aug 28 21:13:22 is there a way to try without recompiling and cleaning the package every time Aug 28 21:13:23 ? Aug 28 21:14:15 bitbake interactive doesn't exist anymore on 1.10 or MASTER version Aug 28 21:15:46 GNUtoo|laptop: bitbake -f -c -b does that work Aug 28 21:16:17 hmm, even with do_compile() I run into this issue. Aug 28 21:16:47 I'll do that Aug 28 21:16:57 I'm already in -b mode Aug 28 21:17:08 but I fogott about -f Aug 28 21:21:53 hmm likewise do same without srctree Aug 28 21:21:56 does that work Aug 28 21:22:03 Crofton_|work: ok I see the problem Aug 28 21:22:58 khem: yes, that seems to work Aug 28 21:23:12 khem: minimal test case here: http://www.sidebranch.com/openembedded/helloworld2.tar.bz2 Aug 28 21:23:19 extract in openembedded.git checkout Aug 28 21:23:37 bitbake helloworld2 (in a clean OE project) will show the bug Aug 28 21:23:45 Crofton_|work: libstdc++-dev does not note libstdc++ in its Depends Aug 28 21:24:43 Crofton_|work: now if you have a good build around can you poke around the ipk Aug 28 21:24:44 for me Aug 28 21:27:40 likewise: just for my test Aug 28 21:27:51 add inherit autotools to it Aug 28 21:27:55 and see if it works ? Aug 28 21:28:37 I don't have a good build around, but the package in the ansgtrom 2008 feeds should be good? Aug 28 21:28:45 yes Aug 28 21:32:00 angstrom-distribution.org hates me Aug 28 21:32:06 heh why? Aug 28 21:32:09 11 kb/s Aug 28 21:32:21 it is too popular Aug 28 21:32:30 bother, kernel opps Aug 28 21:32:32 thats the speed I get when I download something from mongolia Aug 28 21:34:02 :) Aug 28 21:34:50 hmm its same even there Aug 28 21:36:02 khem: oe git gives me 20kb/s also Aug 28 21:36:15 then oe hates you Aug 28 21:36:22 khem: testing with inherit autotools (alone) -- and yes OE hates me Aug 28 21:36:34 I hate it back sometimes Aug 28 21:36:43 but OE mostly wins Aug 28 21:37:01 likewise: anybody who is not in mongolia and gives me 11 kb/s means it hates me Aug 28 21:37:31 likewise: I wanted you to keep inherit srctree autotools Aug 28 21:38:07 thats OK, nfs is hating me atm Aug 28 21:39:06 03Petr Vanek  07org.openembedded.dev * rdc993ead7e 10openembedded.git/recipes/xorg-xserver/ (xserver-xorg-1.9.0/randr-support.patch xserver-xorg_1.9.0.bb): (log message trimmed) Aug 28 21:39:06 xserver-xorg_1.9.0: fix randr support for input calibration Aug 28 21:39:06 This fix corrects the previous version of this patch and ensures Aug 28 21:39:06 that X output and input stay correctly together after xrandr -o Aug 28 21:39:06 is used. Aug 28 21:39:07 Petr Vanous Vanek Aug 28 21:39:08 Signed-off-by: Petr Vanek Aug 28 21:39:45 Crofton_|work: it should have depended upon libstdc++6 Aug 28 21:40:05 I guess debian .so renaming gets opkg confused may be Aug 28 21:40:20 likewise: I think OE is a nice project Aug 28 21:40:26 it serves many needs Aug 28 21:40:40 khem: agreed, I use it for all my embedded work Aug 28 21:40:48 something like that :) Aug 28 21:40:54 the 6 and renaming confuse me Aug 28 21:41:07 Crofton_|work: how does nfs hate you today? Aug 28 21:41:14 Crofton_|work: let me peek into some other nortmal -dev files Aug 28 21:41:26 nfs is evil Aug 28 21:41:33 khem: inherit autotools does stage the toolchain Aug 28 21:41:35 but I still use nfsboot Aug 28 21:41:39 it is a little easier than scp though Aug 28 21:41:45 likewise: ok Aug 28 21:41:48 hmm Aug 28 21:41:56 it worked earlier Aug 28 21:42:08 now the mount is timing out Aug 28 21:42:20 likewise: your example works without srctree and doesnt with srctree when build from scratch right Aug 28 21:42:52 why don't you do what normal people do and build something simple first :) Aug 28 21:46:02 Crofton_|work: I wonder how libstdc++ got pulled into your old image Aug 28 21:46:12 could be some other package pulled it Aug 28 21:46:21 may be boost ? Aug 28 21:46:22 khem: correct, without any inherit, or with autotools inherit only, the toolchain gets built, with srctree only inherit, it doesn't get built Aug 28 21:46:34 likewise: ok Aug 28 21:46:39 these are mysteries to me Aug 28 21:46:47 that seems to be problem Aug 28 21:47:36 Crofton_|work: bitbake -g doesn't give you a clue Aug 28 21:47:38 ? Aug 28 21:47:39 likewise: ok if I have the 25th hr in dat today I will look into it Aug 28 21:48:06 likewise: its runtime dep Aug 28 21:48:23 khem: no don't, I'll ask the author (Chris) if he can check the depends magic in the bbclass. Deps get merged and it's voodoo to me Aug 28 21:48:43 khem: just enjoy your 25th hour for something relaxing :-) Aug 28 21:51:33 khem, we don't need to fix right away, but before we can widely use gcc-4.5 we need to work out all these issues Aug 28 21:52:10 its not gcc-4.5 issue Aug 28 21:52:17 dont think that way Aug 28 21:52:25 well Aug 28 21:52:36 I am sure you will hit it with 4.3 as well in .dev now Aug 28 21:52:37 it is an issue building a working system with it :) Aug 28 21:52:43 ah Aug 28 21:52:59 you did not do a complete build with other gcc Aug 28 21:53:16 4.3 from a while back worked fine Aug 28 21:53:26 you are thinking this was introduced at some point? Aug 28 21:54:07 could be as I said Aug 28 21:54:34 I just want to make it clear that its not gcc 4.5 problem Aug 28 21:54:53 I understand Aug 28 21:56:57 ah Aug 28 21:57:04 I see what I need to do Aug 28 21:57:19 just look in proc/modules to make sure the module is not loaded Aug 28 21:57:27 before loading the fpga Aug 28 22:08:37 I bet return are not allowed in python functions no? Aug 28 22:26:58 likewise: I fixed srctree recently to I own some moral responsibility too Aug 28 22:27:54 khem: I'm trying to see where it borks. It changes the dependency tree a bit, but I don't see the bug yet. Aug 28 22:28:48 srctree is handy Aug 28 22:29:56 khem: as an extra info point: adding DEPENDS = helloworld is not respected also. Aug 28 22:30:15 Crofton_|work: if it works :-) Aug 28 22:58:25 hmmm Aug 28 22:59:17 in that python function http://pastebin.com/c3pGgGZ3 it fails at | ERROR: 32:SyntaxError:' print 'FILES_' + package' Aug 28 22:59:50 I'll try some double quotes Aug 29 00:15:35 Hello. Aug 29 00:15:58 Im trying compile shr-image to a1200 motorola phone but i get this error! http://pastebin.com/WpjiQvZc Aug 29 00:17:09 can anybody help-me ? Aug 29 00:27:32 eglibc_2.12.bb failed Aug 29 00:27:56 angelox_123: it's not really a distro specific problem but you might ask in #openmoko-cdevel . that's where the SHR crowd hangs Aug 29 00:28:26 angelox_123: there is also somebody else trying to get SHR on a similar old phone :> Aug 29 00:30:54 old? Aug 29 00:31:01 =( Aug 29 02:11:29 hi, one question, how i can run one graphic application in the linux start, without a desktop, i'm work whith qt embedded libraries. Aug 29 02:11:55 Can anybody guide me, please! Aug 29 02:15:10 tuxfan, add it to the init system Aug 29 02:17:40 look at zhone2 in freesmartphone as an example Aug 29 02:17:50 it's an enlightenment application Aug 29 02:17:55 that starts without Xorg Aug 29 02:19:35 yes, a need run over framebuffer like qtopia Aug 29 02:20:28 then look at the recipe I pointed to Aug 29 02:20:50 s/application/recipe/ Aug 29 02:35:10 i have a mini2440 develop boad with qtopia, if i add one binary in the init system, its don't init Aug 29 02:39:24 is there some example to do that in the web Aug 29 02:40:24 one sigle application run on linux start **** ENDING LOGGING AT Sun Aug 29 02:59:57 2010