**** BEGIN LOGGING AT Fri Feb 24 02:59:58 2012 Feb 24 05:28:34 hello] Feb 24 05:29:34 i want to build jre for angstrom image of mini2440.... can any one help me.... the recipes for the same is available in oe Feb 24 06:39:27 kergoth: ping Feb 24 08:25:11 good morning ao2, bluelightning, all Feb 24 08:25:25 hi mckoan, ao2, all Feb 24 08:26:07 hi. mckoan put me on CC if you write to linux-media Feb 24 08:34:39 ao2: ok Feb 24 09:51:49 heh double acked :) Feb 24 09:57:35 JaMa|Off: I could build a linux-yocto-tiny for Zaurus. With defaults defconfig and patches got a 1,5MB zImage (our is >2MB) Feb 24 09:57:51 there is some mangling to check but looks promising Feb 24 10:01:03 (I'm using the "compatibility mode") Feb 24 10:02:00 do you have diff of defconfigs? Feb 24 10:10:18 Hello! I have added an recipes which create an proc with some plugins. The plugins will be created with do_split_packages and PACKAGES_DYNAMIC. ipkg's are successful created. how do i add the splited packages to the image? IMAGE_INSTALL didn't work "Nothing PROVIDES..." Feb 24 10:10:27 prog Feb 24 10:14:13 decembersoul: it should work... make sure the packages you expect to be produced are indeed being produced - check TMPDIR/work/*/PN-PV-PR/packages-split/ to see if your packages are there and contain what you expect Feb 24 10:19:14 yes they are there and also at build/tmp-eglibc/work/i586-oe-linux/myprog-1.7.24-r2/deploy-ipks/i586/ Feb 24 10:21:55 decembersoul: so do you get the error straight away? Feb 24 10:22:06 i.e. as soon as bitbake is finished parsing? Feb 24 10:22:33 if so, are you sure PACKAGES_DYNAMIC is set up so that it satisfies the items you're requesting in IMAGE_INSTALL? Feb 24 10:22:41 that's all that PACKAGES_DYNAMIC really does Feb 24 10:24:14 * bluelightning -> office, back shortly Feb 24 10:37:09 It is a bit sad that the build process still seems to be fundamentally a rather serial one. I guess there are just these unavoidable pinch points that everything has to wait for. Feb 24 10:37:17 er, morning all Feb 24 10:37:32 * pb_ stabs ircii Feb 24 10:41:59 good morning :) Feb 24 10:53:21 * florian stabs his RTC - it seems to fall asleep like the rest of the device even if it is supposed to run Feb 24 10:57:21 gdb-cross-sdk_7.2.bb do_fetch failed, any alternative link? Feb 24 10:58:23 actually the problem is gdb-7.2.tar.gz no longer existing Feb 24 11:01:33 ls Feb 24 11:01:42 oops good morning btw ;) Feb 24 11:03:51 florian: we also had a time update problem during suspend last week on a platform, it was solved by a big hack in the kernel ... Feb 24 11:05:27 ericben: here the kernel - the driver at least - does not seem to be guilty. I have turned on debugging and it does to seem to write a wrong value to the rtc Feb 24 11:06:53 thanks to bluelightning. Now it works. fixed a typo and made a clean Feb 24 11:07:59 florian: in our case that was not exactly the RTC in the end but fact that sleep time was not incremented during suspend to ram Feb 24 11:11:25 ericben: sounds familiar... but here its funny because there is a pin for power and some for communications but no dependency on clock signals or such Feb 24 11:11:32 morning Feb 24 11:11:45 hi otavio Feb 24 11:22:13 ericben: Hi :) Feb 24 11:46:18 decembersoul: any luck? Feb 24 12:21:59 bluelightning: Now it works. fixed a typo and made a clean and rebuild it Feb 24 12:22:19 decembersoul: ah ok, cool Feb 24 12:27:26 im updating glib2.0 Feb 24 12:30:38 hi Feb 24 12:31:06 im trying to compile glib-2.0.22 Feb 24 12:31:23 during configuration im getting the error /data/home/anuj/OE_New/build/tmp/work/i586-linux/glib-2.0-2.22.1-r6.3/glib-2.22.1/configure: line 37066: ./libtool: No such file or directory Feb 24 12:43:27 is there a way to check which .bb file is set to virtual/kernel without building anything? Feb 24 12:46:14 mertsas: bitbake -n ? Feb 24 12:49:45 don't give any output, neither -n or -p Feb 24 12:49:57 maybe they expect the dependencies to be built? Feb 24 12:51:02 -s seems to be the one I was after Feb 24 12:56:09 pb_: Are you aware lists.linuxtogo.org is down? Feb 24 12:56:21 RP__1: nope. let me have a look at it. Feb 24 12:56:49 well, hm, the machine isn't down. what exactly is not working for you? Feb 24 12:56:55 pb_: the webpage Feb 24 12:57:38 so it is Feb 24 12:57:47 florian, are you there? Feb 24 13:05:01 how do I tell bitbake to use the 2.6.38 kernel instead of the 3.1 kernel I have made? Tried MACHINE_KERNEL_VERSION = "2.6.38" as well as PREFFERED_VERSION_linux-linaro = "2.6.38" Feb 24 13:05:04 none worked Feb 24 13:08:59 pb_: yes Feb 24 13:10:05 We have some trouble with spammers at projects.linuxtogo.org. Nils disabled apache for that but noone here has time to fix it right now :-( Feb 24 13:12:19 ah, I see Feb 24 13:12:32 can we just disable gforge in apache and turn the server itself back on? Feb 24 13:14:36 mertsas: PREFERRED_VERSION and PREFERRED_PROVIDER should help Feb 24 13:15:48 * RP__ was hoping to be able to point to something in the mailing list archives Feb 24 13:17:12 RP__: so when I have PREFFERED_PROVIDER_virtual/kernel = "linux-linaro" in my machine config, I should have PREFFERED_VERSION_virtual/kernel = "2.6.38" there as well, and it should pick up linux-linaro_2.6.38.bb? Feb 24 13:18:20 mertsas: PREFFERED_VERSION_linux-linaro = "2.6.38" should have worked Feb 24 13:18:39 mertsas: Try running bitbake with -DDD and see if it explains why its picking 3.1 Feb 24 13:19:00 mertsas: Pipe to a file and grep it, there will be a lot of output Feb 24 13:21:06 the RP__ just bitbake -DDD, or do I have to do virtual/kernel? Don't really want to build it if it's wrong:P Feb 24 13:21:54 mertsas: virtual/kernel Use -n if you don't want it to actually do anything Feb 24 13:21:57 btw, if I have OLDEST_KERNEL defined as 2.6.38, I can still use 2.6.38, right? Feb 24 13:22:15 mertsas: that only affects eglibc iirc Feb 24 13:22:19 mertsas: so yes Feb 24 13:25:06 pb_, florian: If we could turn lists back on and leave projects turned off that would be an improvement at least Feb 24 13:25:06 damn, embarassing. Just gave myselv a headache because of a typo!:P Feb 24 13:25:09 thanks for the help Feb 24 13:25:58 RP__: yes indeed, I think I well manage to do it this afternoon Feb 24 13:26:25 florian: ok, thanks! Feb 24 13:27:49 do we support sh3 and sh4 in meta-oe? Feb 24 13:35:48 JaMa|Off: don't see why not... in fact I think we accept sh-related patches in OE-Core even though currently we don't officially support it Feb 24 13:39:14 if bb.data.getVar('TARGET_ARCH', d, 1) in [ 'sh4', 'sh3' ] probably makes task-cli-tools.bb all-oe-linux-gnueabi and depending on eglibc and gcc-runtime Feb 24 13:42:52 hi bluelightning Feb 24 13:42:58 and because it inherits task -> so allarch then it shouldn't depend on TARGET_ARCH value Feb 24 13:43:01 hi GNUtoo|laptop Feb 24 13:44:08 bluelightning, how do you set the display in opie, and what is ODevice Feb 24 13:44:36 weird for qemux86-64 build it is just all-oe-linux Feb 24 13:45:59 GNUtoo|laptop: it's mostly handled by Qt/E, controlled via the QWS_DISPLAY environment variable Feb 24 13:46:31 ah so if I set QWS_DISPLAY it overrides ODevice? Feb 24 13:46:35 GNUtoo|laptop: ODevice is the abstraction class for device-specific stuff Feb 24 13:46:47 ok Feb 24 13:46:52 I just saw the source Feb 24 13:46:58 there is zaurus,simpad etc... Feb 24 13:47:09 GNUtoo|laptop: AFAIK ODevice doesn't do much display-wise, it only really handles things like buttons, LEDs, etc. Feb 24 13:47:49 ahh ok Feb 24 13:48:04 so to set /dev/fb0 I do that: Feb 24 13:48:21 export QWS_DISPLAY=LinuxFb:device=/dev/fb0 ? Feb 24 13:49:54 GNUtoo|laptop: I don't recall I'm afraid Feb 24 13:50:10 ok anyway what I just pasted is for QT4 Feb 24 13:50:15 I must look for qt2 Feb 24 13:50:36 maybe just LinuxFb:/dev/fb0 Feb 24 13:54:29 really ought to document this on the opie wiki Feb 24 13:54:31 doesn't work Feb 24 13:54:32 ok Feb 24 13:54:54 also I've a patch for a recipe in meta-opie Feb 24 13:55:02 I'll try to send it as soon as possible Feb 24 13:56:25 GNUtoo|laptop: great, will look forward to that :) Feb 24 13:57:40 basically it's just adding a kernel and machine extra rrecommands Feb 24 13:57:54 because most machine can't boot without that Feb 24 13:59:23 GNUtoo|laptop: http://doc.qt.nokia.com/2.3/envvars.html Feb 24 13:59:38 GNUtoo|laptop: default is /dev/fb0 already Feb 24 14:00:15 GNUtoo|laptop: if it's machine-specific I'm not sure it really belongs in meta-opie Feb 24 14:00:30 GNUtoo|laptop: which machine is this for? Feb 24 14:01:28 it's a bug20 with a 320x240 display Feb 24 14:01:35 it uses standard framebuffer Feb 24 14:03:23 the default seem to be that: Feb 24 14:03:26 : setting QWS_DISPLAY to 'Transformed:Rot0:0' Feb 24 14:03:43 bluelightning, it's specific to most of the machines Feb 24 14:04:05 I think not having /boot/uImage shipped is the exception Feb 24 14:04:23 and MACHINE_EXTRA_RRECOMMANDS is usually needed too Feb 24 14:04:47 GNUtoo|laptop: MACHINE_EXTRA_RRECOMMENDS is something that should appear only in the machine conf file though... Feb 24 14:05:05 yes but it should be included somehow in the image like that: Feb 24 14:05:11 ${MACHINE_EXTRA_RRECOMMENDS} \ Feb 24 14:05:22 in IMAGE_INSTALL or in a task or whatever Feb 24 14:05:28 ah, now I understand Feb 24 14:05:30 yes Feb 24 14:05:46 this is because currently opie-image does not pull in task-base-extended Feb 24 14:05:52 yes possible Feb 24 14:06:02 (mainly because task-base-extended has way too much in it IMHO) Feb 24 14:06:15 and the oe core task-core-boot doesn't have the kernel image Feb 24 14:06:18 but until we have a chance to re-evaluate that we can at least do as you suggest Feb 24 14:06:21 task-boot from meta-oe does Feb 24 14:06:30 ok Feb 24 14:06:32 yes, that's another thing we need to properly resolve Feb 24 14:07:00 for the moment I've just been adding MACHINE_ESSENTIAL_EXTRA_RDEPENDS to pull in the kernel Feb 24 14:07:11 but that means doing that for every machine which is not ideal Feb 24 14:07:20 the thing is it is kind of machine specific Feb 24 14:07:30 depends on how you're booting Feb 24 14:07:37 also you don't have udev Feb 24 14:08:00 that requires to add modules autoload in each machines Feb 24 14:08:06 I'll do that then Feb 24 14:08:08 RP__: I like externalsrc.bbclass, but it doesn't solve an issue srctree solved. of course, there are other ways around the issue (e.g. using ${B}). The issue is, if you use BBCLASSEXTEND, the various tasks can interleave. Using lockfiles doesn't help, because it's released between tasks. The best method is to use ${B} with a path including an appropriate triplet, though, rather than trying to solve it generically Feb 24 14:08:16 GNUtoo|laptop: udev should be pulled in by other means surely.. ? Feb 24 14:08:27 I think it's not in my image Feb 24 14:09:21 I've only udevadm Feb 24 14:09:24 but not the daemon Feb 24 14:09:30 GNUtoo|laptop: really? because that definitely is pulled in by task-core-boot Feb 24 14:10:03 I bitbaken opie-image Feb 24 14:10:06 let me check Feb 24 14:10:15 GNUtoo|laptop: unless VIRTUAL-RUNTIME_dev_manager is set to something else in your distro config Feb 24 14:10:27 how's the deamon is called? Feb 24 14:10:51 udevd? Feb 24 14:11:10 then I have some part of udev but not udevd Feb 24 14:14:37 kergoth: That is why I just set ${B}. I appreciate it solves a different set of problems Feb 24 14:15:29 kergoth: most autotools stuff will support that at least and if your code doesn't you could probably add it easily enough Feb 24 14:19:46 distro was angstrom Feb 24 14:21:30 how can control which kernel modules are installed into my image and which not? Feb 24 14:33:33 RP__: agreed. Feb 24 14:34:02 GNUtoo|laptop: I can only suggest doing some further digging (e.g. bitbake -e) as to why this doesn't get pulled in by task-core-boot, because that's supposed to be how it gets in there Feb 24 14:34:24 GNUtoo|laptop: in fact I'm positive when I build it with poky that I get udev in opie-image Feb 24 14:35:10 ok Feb 24 14:45:25 bluelightning, it was not the display variable finally , it was the fonts that were lacking Feb 24 14:45:47 /lib/fonts/fontdir was absent Feb 24 14:47:25 so I did that and it started: Feb 24 14:47:38 .../work/armv7a-angstrom-linux-gnueabi/qte-mt-2.3.10-r31/qt-2.3.10$ scp -r lib/fonts/ root@192.168.7.2:/lib/ Feb 24 14:51:45 now I've that: Feb 24 14:51:46 tslib: Selected device uses a different version of the event protocol than tslib was compiled for Feb 24 14:51:49 I've 2.6.35 Feb 24 14:51:54 I think I'll abandon Feb 24 14:57:06 and I guess qt2 doesn't support evdev Feb 24 14:58:30 GNUtoo|laptop: it does Feb 24 14:58:42 wow Feb 24 14:58:46 GNUtoo|laptop: that's entirely handled via tslib Feb 24 14:58:59 ah ouch Feb 24 14:59:33 the tslib message really annoys me, it means that it was compiled with kernel header more recent than the kernel I have Feb 24 15:00:02 I'll send a patch for fixing the boot tough Feb 24 15:00:05 ok Feb 24 15:00:09 thanks Feb 24 15:00:16 but not right now, as soon as I can Feb 24 15:00:33 sure Feb 24 15:04:17 btw we need to fix the font issue too Feb 24 15:07:32 GNUtoo|laptop: hmm, I wonder why you don't get any fonts installed and I do... Feb 24 15:12:10 hmmm Feb 24 15:12:25 I used angstrom as a distro and you used poky Feb 24 15:12:33 maybe that creates a font problem Feb 24 15:14:03 GNUtoo|laptop: FYI we are using truetype fonts in /usr/share/fonts/truetype... do you not have anything in there? Feb 24 15:15:10 yes I do Feb 24 15:15:19 but it was rather the qte fonts that were the probelm Feb 24 15:15:31 you mean qpf? Feb 24 15:15:36 we don't use those anymore Feb 24 15:15:38 yes Feb 24 15:15:40 ah ok Feb 24 15:15:41 but: Feb 24 15:15:46 freetype support is enabled Feb 24 15:15:46 strace looked at: Feb 24 15:16:00 ahh maybe I've an old .inc for opie Feb 24 15:16:03 it still looks for them though (and may even complain they aren't there) Feb 24 15:16:05 I'll look in a moment Feb 24 15:16:12 but it should find and use the ttf fonts Feb 24 15:16:18 ok Feb 24 15:18:12 conf/distro/include/preferred-opie-git-versions.inc Feb 24 15:18:16 that's all I have Feb 24 15:19:55 if that's the one in meta-opie that's the latest Feb 24 15:20:34 03Steffen Sledz  07org.openembedded.dev * r780de23f1f 10openembedded.git/recipes/jakarta-libs/avalon-framework-api_4.3.bb: Feb 24 15:20:34 jakarta-libs: fixed SRC_URI for avalon-framework-api-4.3 Feb 24 15:20:34 Signed-off-by: Steffen Sledz Feb 24 15:20:36 03Steffen Sledz  07org.openembedded.dev * rb010bc47e2 10openembedded.git/recipes/fakeroot/fakeroot-native_1.14.5.bb: Feb 24 15:20:36 fakeroot-native: update to 1.14.5 Feb 24 15:20:36 * version 1.9.6 is no longer fetchable Feb 24 15:20:36 * recipe for version 1.9.6 is not removed yet because of pinnings in Feb 24 15:20:36 various distro configurations Feb 24 15:20:37 Signed-off-by: Steffen Sledz Feb 24 15:20:37 03Steffen Sledz  07org.openembedded.dev * r745cb81dfd 10openembedded.git/recipes/net-snmp/ (net-snmp-5.7.1/net-snmp-libnl.patch net-snmp_5.7.1.bb): (log message trimmed) Feb 24 15:20:38 net-snmp-5.7.1: fix build error "too few arguments to function 'nl_geterror'" Feb 24 15:20:38 Fixes compile errors with libnl-1_2.0: Feb 24 15:20:39 | ... Feb 24 15:20:39 | mibII/tcpTable.c:579: error: too few arguments to function 'nl_geterror' Feb 24 15:22:20 GNUtoo|laptop: here are the packages I have in an opie-image I built just yesterday: http://pastebin.com/MgGLX9LZ Feb 24 15:22:36 GNUtoo|laptop: might be worth seeing if there's anything you don't have Feb 24 15:23:26 hmm, it'd be interesting to try making B usage opt-out rather than opt-in for autotools.bbclass users, and see what blows up Feb 24 15:24:33 what's B? Feb 24 15:49:43 JaMa|Off: separated build/object dir from source Feb 24 16:07:11 hi, I always forgett how to tell SRC_URI to fetch git or svn trough http when the address starts with http:// (code in fetcher2 makes it use http fetcher instead which uses wget and not svn) Feb 24 16:07:43 ah maybe it seem to have worked that time Feb 24 16:07:47 sorry for the noise Feb 24 16:09:45 protocol= in the parameters will tell the git fetcher what url scheme to use Feb 24 16:09:51 ah no sorry it fetched in html Feb 24 16:09:52 ok Feb 24 16:09:58 that's what I have with svn fetcher Feb 24 16:10:19 SRC_URI = "http://src.chromium.org/svn/trunk/tools;rev=123481;protocol=svn;module=depot_tools" Feb 24 16:10:43 the url scheme controls what fetcher is used Feb 24 16:10:49 if you use that, it'll try to fetch it with wget Feb 24 16:10:53 swap the url scheme and protocol Feb 24 16:10:54 indeed Feb 24 16:10:59 you want svn:// and protocol=http Feb 24 16:10:59 ok Feb 24 16:11:16 I'll retry Feb 24 16:11:40 thanks a lot Feb 24 16:11:53 np Feb 24 16:11:57 and quite some time ago there was proto and protocol Feb 24 16:12:01 both existed Feb 24 16:12:07 I think it was for different fetchers Feb 24 16:12:34 now it tries to fetch that: Feb 24 16:12:36 /usr/bin/env svn co -r 123481 svn://src.chromium.org/svn/trunk/tools/depot_tools Feb 24 16:12:49 SRC_URI = "svn://src.chromium.org/svn/trunk/tools;rev=123481;protocol=http;module=depot_tools" Feb 24 16:12:51 hmm, maybe it is proto then, i never remember :) Feb 24 16:12:53 maybe I need proto Feb 24 16:12:54 ok Feb 24 16:13:32 if I remember well proto was for svn and protocol for git Feb 24 16:13:35 or something like that Feb 24 16:13:43 * kergoth shakes head Feb 24 16:14:02 thanks a lot Feb 24 16:14:04 it works now Feb 24 16:14:18 I was just unlucky and didn't get the right combinaison Feb 24 16:16:25 git grep is your friend in such situations :) Feb 24 16:20:36 but I grepped the wrong thing Feb 24 16:20:45 I grepped for http:// Feb 24 16:20:55 then I read the source of fetcher2 tough Feb 24 16:34:34 khem: error seen this issue? https://gist.github.com/1901898 Feb 24 16:37:01 its not really occuring in oe-core - just asking because we are messing with the toolchain Feb 24 16:38:12 msm: yes I have seen that Feb 24 16:38:26 msm: explain to me what changes are you making Feb 24 16:38:40 that will give me an idea why this could be happening Feb 24 16:39:19 msm: thats eglibc I think right ? Feb 24 16:40:01 khem: removing yocto patches from gcc and going to a pristine gcc-4.6.2 Feb 24 16:40:12 khem: yea, eglibc is complaining Feb 24 16:41:14 hmmm, hi again, I wonder how to make a recipe build before fetching another Feb 24 16:41:33 like make do_fetch depend on a fetching tool Feb 24 16:41:43 khem: i dont know a lot… im trying to get someone else to do this exercise so don't waste too much effort on it Feb 24 16:43:00 ahh maybe with the following syntax Feb 24 16:43:04 subversion-native:do_populate_sysroot Feb 24 16:43:11 in case of svn Feb 24 16:43:24 (in my case it's a special google tool) Feb 24 16:44:23 an it didn't work Feb 24 16:44:31 do_fetch already depends on all the needed fetching tools, read base.bbclass to see how it's done there Feb 24 16:45:49 maybe that: Feb 24 16:45:57 do_fetch[depends] += Feb 24 16:46:03 msm: ok in anycase that error means you did not enable 128 bit long double in gcc or its not working correctly Feb 24 16:46:08 I'll try that Feb 24 16:52:46 msm: another thing that is probably happening is fpu settings in eglibc configure are mismatching Feb 24 16:52:48 khem: seems like that should be available on all powerarch Feb 24 16:53:21 msm: if you have config.log from eglibc build that will help Feb 24 16:53:28 khem: well… we just removed some gcc patches and these issues arose Feb 24 16:53:38 khem: right, im trying to repodouce this now Feb 24 16:53:50 you are still using oe to build it ? Feb 24 16:53:55 or building outside Feb 24 16:58:54 khem: i think this is the relenvant bit Feb 24 16:58:54 https://gist.github.com/1902019 Feb 24 16:59:09 khem: yes, we just removed patches and stepped back the SVN revision a bit Feb 24 16:59:42 khem: trying to go back to the version that our toolchain guys tested Feb 24 17:00:16 khem: then add their patches… this is for us internally… we want to upstream their work to OE's gcc too Feb 24 17:04:40 im not even seeing where gcc passes the targets it supports Feb 24 17:06:21 maybe it just supports all the cpus for the current arch? so we don't need to enable it Feb 24 17:10:03 hi JaMa|Off Feb 24 17:10:35 do you remember how to solve that: Feb 24 17:10:35 tslib: Selected device uses a different version of the event protocol than tslib was compiled for Feb 24 17:10:42 I think on palmpre it's solved Feb 24 17:10:56 but I don't see anything on palmpre's bbappend for tslib Feb 24 17:11:09 I mean anything related to the solving of that problem Feb 24 17:12:13 IIRC I had to weaken that check in tslib or xf86-input-tslib code so they it allowed higher version Feb 24 17:13:20 GNUtoo|laptop: check 390e8cd69ee11bcc8a10ea3d887a386eab2579cb in oe-classic if it's the same issue Feb 24 17:20:13 if a recipe is generating staticdev packges… is there a way to make that the default stuff that's installed Feb 24 17:20:32 or do I just do FILES_${PN} = FILES_${PN}-staticdev Feb 24 17:23:38 RDEPENDS_${PN} += "${PN}-staticdev" ? Feb 24 17:25:18 that seems like a better fix Feb 24 17:25:33 will the default package (rpm, ipkg) get created that depends on the other package only? Feb 24 17:25:48 does it have anything in it? Feb 24 17:26:06 if not, you might need to ALLOW_EMPTY_${PN} Feb 24 17:26:25 pb_: k Feb 24 17:26:25 ok Feb 24 17:28:39 JaMa|Off, the problem is the following: Feb 24 17:28:41 if (version < EV_VERSION) { Feb 24 17:28:48 that's what I already have Feb 24 17:29:31 My zenitthink c91 arrived Feb 24 17:34:46 let me try something Feb 24 17:38:14 I'll make a patch Feb 24 17:40:55 pb_: you had some comments for my thumb tune related patches Feb 24 17:41:14 pb_: Do you have some ideas to improve them Feb 24 17:42:43 pb_: I think the issue we are running into is what is thumb tune feature ? does it mean packages are compiled with thumb mode or does it indicate they are thumb interworkable I think we need to account for both Feb 24 17:43:02 pb_: so I think we need two tune features Feb 24 17:45:04 pb_: tis is the thread http://patches.openembedded.org/patch/20409/ Feb 24 17:52:55 bluelightning: using buildhistory it dumps some logs at the end of build like this Feb 24 17:53:53 http://paste.ubuntu.com/855604/ Feb 24 17:54:00 how can I avoid it Feb 24 17:54:30 khem: it's one of the very few differences between bitbake upstream and bitbake in poky Feb 24 17:54:47 khem: bitbake upstream defaults set +x on which means the output gets echoed Feb 24 17:54:59 well, that which you see is echoed Feb 24 17:55:07 kergoth: can we resolve this? Feb 24 18:07:45 bluelightning: hmm I see Feb 24 18:07:49 ok Feb 24 18:08:12 bluelightning, http://www.pastie.org/private/hmm1k2ru36qjdgatzkjqeg Feb 24 18:08:26 sounds like rpath issue Feb 24 18:10:38 hmmm Feb 24 18:10:41 I'll look Feb 24 18:13:06 bluelightning: don't see any reason why not, but the set -x/+x thing gets brought up on a periodic basis, some wanting one behavior, some the other Feb 24 18:13:08 heh Feb 24 18:13:40 kergoth: wanting in what context? for Feb 24 18:13:46 argh @ keyboard Feb 24 18:13:55 for all tasks? or selectable per task? Feb 24 18:14:12 if it is something that people have different ideas about then surely it should be an option? Feb 24 18:14:17 hmmm I think it's a setting in the init script Feb 24 18:15:17 bluelightning: personally, I think it's not particularly useful, as you can always set it explicitly in the function you care about when using it for debugging -- and when you aren't debugging, you probably don't care about it. but others disagree, and want more fleshed out logs (e.g. for license compliance / legal reasons) Feb 24 18:15:31 so yeah, some sort of global option would probably do it Feb 24 18:20:05 kergoth: I just want to get rid of that delta we have in a way that makes the most sense :) Feb 24 18:20:28 agreed Feb 24 18:20:40 I'll see if I can cook something up soon Feb 24 18:45:24 GNUtoo|laptop: if i get time I'll be around over the weekend and will try a few opie-image builds Feb 24 18:46:14 bbl Feb 24 18:46:40 ok Feb 24 19:11:23 fray: how do I set the default arch for rpm? Feb 24 19:11:49 fray: my rpms are powerpc/powerpc64 instead of ppce500/ppc64e5500 Feb 24 20:07:30 fray: think i figured it out TUNE_PKGARCH_tune-${DEFAULTTUNE} is not defined Feb 24 20:07:34 so one needs to use simple overrides Feb 24 20:09:32 but that's not really a good solution Feb 24 20:09:43 i need to alternate TUNE_PKGARCH between three values Feb 24 20:16:41 arr Feb 24 20:18:25 so i set ENABLE_BINARY_LOCALE_GENERATION to "1" and specified a handful of UTF-8 locales in GLIBC_GENERATE_LOCALES but it still builds a bunch of locales i don't want/need Feb 24 20:19:24 either I'm not understanding the config options or something is seriously weird... Feb 24 20:20:05 any tips would be appreciated... Feb 24 21:30:02 hi there Feb 24 21:34:14 i found Olimex at91sam9260 board in my closed, dusted it off and trying to remeber some old hacks... Im trying to flash new u-boot over sam-ba. Some time ago when i played with this board more often usbserial was a kernel module, used for connecting and flashing at91 board... Today in kernel 3.1.6 world usbserial seams to be embedded in the kernel... and refuses to create /dev/ttyUSBX for myBus 002 Device 016: ID 03eb:6124 Atmel Corp. at91sam SAMBA Feb 24 21:34:16 bootloader praty please, help :) Feb 24 21:38:45 mentis: you can pass, by kernel cmdline, the vendor and products ids for the usbserial module Feb 24 21:39:12 otavio: kernel cmdline? Feb 24 21:41:05 otavio: but is kernel cmdline available only at boot time? Feb 24 21:41:10 mentis: your machine kernel (your host) Feb 24 21:41:17 otavio: but is kernel cmdline available only at boot time? Feb 24 21:41:22 mentis: yep Feb 24 21:41:49 otavio: thats bad m8 Feb 24 21:41:50 mentis: I don't know any other way of changing it; it needs to be done at the driver initialization Feb 24 21:42:05 mentis: or use module version of usbserial Feb 24 21:42:38 otavio: i wanted to use module version but cant find 'subserial' in 3.1.6 kernel Feb 24 21:43:10 otavio: 'usbserial'* Feb 24 21:43:33 mentis: so change your boot commandline to add usbserial.vendor=0xXXXX usbserial.product=0xXXXX Feb 24 21:43:44 mentis: and reboot your machine ;P Feb 24 21:44:55 otavio: :/ but myBus 002 Device 016: ID 03eb:6124 Atmel Corp. at91sam SAMBA bootloader is only available when i shorten the pins on the EvalKit :/ Feb 24 21:45:18 >.< Feb 24 21:45:29 mentis: yes but your host needs to know about the 03eb:6124 Feb 24 21:45:37 mentis: am I missing something? Feb 24 21:46:17 otavio: will it be the same when i replug the device ? i mean the '03eb:6124' Feb 24 21:46:27 mentis: I guess so Feb 24 21:46:32 hmmm Feb 24 21:46:53 otavio: ty, i will try... Feb 24 21:47:16 mentis: np :) Feb 24 22:08:30 otavio: I found usbserial ^_^ was in main tree xD how nobish of me ... >.< Feb 24 23:03:28 JaMa|Off: http://paste.debian.net/157503/ Feb 24 23:31:53 * mr_science has the buiild log handy for the python fail Feb 24 23:38:27 http://pastebin.com/2q7xWNEc <= it's hard to tell, but it looks like an autoconf/m4 failure Feb 24 23:38:47 even though the recipe does "autoreconf..." Feb 24 23:56:51 don't everyone jump at once... Feb 25 00:49:55 * mr_science goes home **** ENDING LOGGING AT Sat Feb 25 02:59:58 2012