**** BEGIN LOGGING AT Mon Aug 23 02:59:57 2010 Aug 23 04:44:37 gm Aug 23 05:28:04 Tartarus: tried minimal for ppc e500v2; I have more success than with angstrom (I can build the compiler), but when building libusb I get an error about a hidden symbol and a non representable section) see http://www.pastebin.ca/1923134 Aug 23 05:28:07 any idea on this? Aug 23 05:28:38 morning Aug 23 05:29:04 what browser is better to use on Zaurus now? Aug 23 05:29:10 buildable and usable :) Aug 23 05:32:03 eve Aug 23 05:32:06 hehe.. seems gpe-today is still unbuildable Aug 23 05:32:20 midori is failed for me Aug 23 05:33:30 ah.. that was webkit-gtk: http://tinderbox.openembedded.org/public/logs/task/7364586.txt Aug 23 05:34:25 * Jay7 is trying to build all that gpe-* things.. Aug 23 05:45:39 khem, you have an ack for efika Aug 23 05:49:21 what about efika? Aug 23 05:49:25 * Jay7 have one :) Aug 23 05:55:34 Jay7: khem submitted patches to the ML for supporting efika mx Aug 23 05:55:42 wow.. Aug 23 05:57:25 +1 from me but not sure this is enough :) Aug 23 05:58:14 at least it does no harm as I see Aug 23 05:59:20 I can't check it today but may be will check in this week Aug 23 05:59:25 i already acked it Aug 23 06:01:21 ah.. no.. only after 30 Aug.. Aug 23 06:07:01 peeked at it, nice device, eur 249 though Aug 23 06:09:32 sry usd 249 + shipping + import duty Aug 23 06:13:40 I've received for kexec experiments.. Aug 23 06:13:48 * eFfeM might be more interested in a kindle3 when it is hacked. Aug 23 06:13:55 but still have nothing done :( Aug 23 06:14:22 but I'll show it on exhibition on 28-29 Aug :) Aug 23 06:16:16 eFfeM: what ML are you talking about ? Aug 23 06:17:08 Marex: openembedded mailing list Aug 23 06:17:32 ML here typically refers to oe ml (after all this is #oe) Aug 23 06:17:34 hi khem Aug 23 06:17:39 Hi Aug 23 06:17:51 just ghosted someone who stole my nick Aug 23 06:18:09 ah ok Aug 23 06:18:48 eFfeM: ah ok, thanks Aug 23 06:19:09 eFfeM: you're involved with the efika too ? Aug 23 06:19:11 eFfeM: I posted update to ncurses patch Aug 23 06:19:25 eFfeM: #efika ;-) Aug 23 06:20:23 Marex: i'm not involved, just interested, don't have any Aug 23 06:20:28 khem, ah good Aug 23 06:21:19 its a nice comp. I am going to do a lot of thumb2/armv7 work on it Aug 23 06:21:30 khem do you by any chance have an idea on this libgcc issue? http://www.pastebin.ca/1923134 Aug 23 06:22:05 eFfeM: yes Aug 23 06:22:54 eFfeM: ok Aug 23 06:23:20 gcc-add-t-slibgcc-libgcc.patch should have fixed it Aug 23 06:23:47 eFfeM: you have few options. 1. use gcc 4.5 Aug 23 06:23:47 khem, great, for this build I was not on dev head, seems a good reason to update Aug 23 06:24:00 2. update Aug 23 06:24:01 :) Aug 23 06:24:10 3. Dont use -Os Aug 23 06:24:16 alright. Aug 23 06:24:38 for now I'll update I think, this is for a product we're developing and we haven't really frozen the version we're going to use Aug 23 06:24:58 eFfeM: have a look at the ncurses patch. My build went fine but still more eyeballs are good Aug 23 06:25:06 khem: so the gcc 4.5 problems are sussed now ? Aug 23 06:25:14 I heard there were some problems on ARM with it Aug 23 06:25:24 that's why emdebian didn't switch yet Aug 23 06:25:25 Marex: yes few of them Aug 23 06:25:42 but its better performing on newer arm platforms Aug 23 06:25:56 khem: so it's ready now ? Aug 23 06:26:17 Marex: yes I think so. I have been using it myself for long enough Aug 23 06:26:26 khem: I'll ack ncurses too (as we're using it in SHR already) Aug 23 06:26:31 Marex: but I dont ship products Aug 23 06:26:36 khem is it posted ? when? Aug 23 06:26:38 khem: me neither ;-) Aug 23 06:26:38 JaMa: cool thz Aug 23 06:27:00 Marex: in fact if you use it and find issues I will be happy to fix them Aug 23 06:27:58 now that efika is working here. I will pull in some nice linaro gcc patches which should boost armv7/cortex perf Aug 23 06:28:14 http://tinderbox.openembedded.org/packages/713054/ Aug 23 06:28:19 khem i was triggered by this when moving from angstrom to minimal (as angstrom apparently has issues with spe and triplet) angstrom used 4.3.3. and either had a workaround for this issue or 4.3.3 did not have the problem Aug 23 06:28:20 and also some fixed from trunk Aug 23 06:28:44 I dislike this: arm-angstrom-linux-gnueabi-gcc: language c++ not recognized Aug 23 06:28:53 eFfeM: I think 4.3 also has problem. Aug 23 06:28:54 what it can mean? Aug 23 06:29:00 its 4.2 which does not have it Aug 23 06:29:33 khem maybe angstrom did do the -Os workaround, I know I've seen some hacks somewhere, but not sure if it was there Aug 23 06:29:40 actually problem goes long way back but from user pov 4.4 and 4.5 are good for ppc/spe Aug 23 06:29:53 angstrom uses -O2 Aug 23 06:29:56 * eFfeM does not know too much about abi's and spe Aug 23 06:29:59 so they dont hit this problem Aug 23 06:30:23 ah ok Aug 23 06:30:45 eFfeM: but if you use minimal you will be ok I guess. I make sure that qemuppc keeps working Aug 23 06:30:49 with minimal Aug 23 06:31:05 infact all qemu machines in OE Aug 23 06:31:10 with minimal Aug 23 06:31:20 is my playground Aug 23 06:31:33 alright. time to bed now Aug 23 06:31:35 gn all Aug 23 06:31:41 * khem -> sleep () Aug 23 06:31:56 khem: thanks, I'll update and see Aug 23 06:32:08 gnite Aug 23 06:32:10 khem: btw. I'll have to use OE, not debian, right ? Aug 23 06:32:14 JaMa: cau Aug 23 06:32:19 Marex: cau Aug 23 06:32:20 Marex: yes Aug 23 06:32:49 khem: will update my tree then, thanks Aug 23 06:33:28 khem, nite & thanks for the info Aug 23 06:33:48 Jay7: you need to use g++ not gcc fix CC and CXX env vars Aug 23 06:33:54 eFfeM: np Aug 23 06:33:57 thx Aug 23 06:44:30 hm.. Aug 23 06:44:49 I didn't set any CC/CXX vars Aug 23 06:46:50 | configure: error: iconv() not found. Cannot enable native uconv support. Aug 23 06:47:18 hm.. browsers are against me Aug 23 06:47:21 that was minimo Aug 23 07:19:14 JaMa: wrt ncruses: do recipes with RDEPENDS = "ncurses" need to have a PR bump too? Aug 23 07:20:11 no, unless you update RDEPENDS = "ncurses" ie to RDEPENDS = "libtinfo" Aug 23 07:20:13 morning Aug 23 07:20:22 hi hrw Aug 23 07:20:37 JaMa, ok Aug 23 07:20:42 do not RDEPEND on libraries... Aug 23 07:20:52 eFfeM: many packages which had ncurses in depends before now depend only on libtinfo (ie mc) Aug 23 07:20:59 shlibs do that for you Aug 23 07:21:11 eFfeM: but those are depends automatically generated by bbclasses (shlibs) Aug 23 07:21:38 ok, i don't know too much about the shlibs stuff Aug 23 07:21:42 eFfeM: but some packages depends on ncurses-terminfo now (cannot be found by shlibs) Aug 23 07:22:24 eFfeM: and for packages I know about I've already added it to RSUGGESTS, not RDEPENDS because not every distro had ncurses-terminfo from 5.7 Aug 23 07:22:51 http://git.openembedded.org/cgit.cgi/openembedded/log/?qt=grep&q=terminfo Aug 23 07:22:59 ok, fine with me Aug 23 07:23:07 s/RSUGGESTS/RRECOMMENDS/ Aug 23 07:23:54 ok, that means khem's patch looks fine to me Aug 23 07:28:35 03Martin Jansa  07org.openembedded.dev * rc14b64632b 10openembedded.git/ (7 files in 2 dirs): Aug 23 07:28:35 xorg: upgrade recipe versions for 2010-08-23 Aug 23 07:28:35 Signed-off-by: Martin Jansa Aug 23 07:31:11 JaMa: as you know much about xorg: I've been peeking at freenx (triggered by ka6sox4; this might be your lucky day ka6sox4). Aug 23 07:31:27 woohoo! Aug 23 07:32:23 freenx comes with its own version of X, I've managed to build things but install seems to want to put stuff in /usr/include/X11R6 etc, could not really find how to easily change that, JaMa do you have an idea ? Aug 23 07:33:06 ka6sox4: that was quick :-) expected you to be afk (guess it it around 12.30 am where you live) Aug 23 07:33:51 ya, but having a discussion about why you can't just bring a kernel module that wasnt' built with the toolchain of the rest of the build and expect it to work. Aug 23 07:34:24 ka6sox4: did you want the client or the server (or both), freenx-server-0.7.3 builds fine for me (but I guess the recipe has some missing dependencies) nx-X11_3.4.0 builds fine but does not install Aug 23 07:34:30 eFfeM: well I never used freenx, so don't know about freenx build.. Aug 23 07:34:38 server. Aug 23 07:34:56 ka6sox4: server seems no problem Aug 23 07:35:28 eFfeM: but looking into it's configure.ac/Makefile.am should probably show where you get X11R6 and if there is configure option for that Aug 23 07:35:40 kewl...monday I'll build the 3.4.0 source and see if it works against it. Aug 23 07:35:47 JaMa: it is a std xorg tree with an Imakefile etc Aug 23 07:36:13 JaMa it contains the xorg tree Aug 23 07:36:35 let me push a few things Aug 23 07:38:48 oops, seems I spoke too early wrt server, need to peek into it again Aug 23 07:41:50 k Aug 23 07:42:18 I"m not going to push anything with the 3.4.0 sources till it builds and works for me. Aug 23 07:44:03 server = 0.7.3; client is 3.4.0 server is still the stuff from berlios Aug 23 07:45:58 no, I"m getting it from nomachine Aug 23 07:46:05 03Frans Meulenbroeks  07org.openembedded.dev * r373526c176 10openembedded.git/recipes/freenx/nx-client_3.4.0.bb: (log message trimmed) Aug 23 07:46:05 nx-client: new recipe Aug 23 07:46:05 first version; still needs to get install fixed, but pushing Aug 23 07:46:05 this so others can peek into it. Aug 23 07:46:05 There are probably also missing dependencies Aug 23 07:46:06 and it is untested. Aug 23 07:46:06 Feel free to improve Aug 23 07:46:06 from their released stuff Aug 23 07:46:30 I'll pull that on monday Aug 23 07:46:34 er tommorrow. Aug 23 07:46:38 ka6sox4: do they also have server? not on the page you gave me (or I just mixed up server and client and pushed the server stuff as client Aug 23 07:46:58 its different pieces. Aug 23 07:47:00 the stuff above seems to be server stuff indeed Aug 23 07:47:07 yep. Aug 23 07:47:10 ka6sox4: yeah saw that, doc and structure sucks Aug 23 07:47:27 its a mess..but we are sorting it Aug 23 07:47:57 ka6sox4: hwat I just pushed builds nx-X11-3.4.0-4 Aug 23 07:48:35 good...I hope all the other modules do as well... Aug 23 07:48:42 JaMa: if you have time and interest feel free to check out the recipe I just pushed (which may be nx-server instead of nx-client) Aug 23 07:49:00 i have a few of the modules compiling, some have some issues Aug 23 07:49:42 eFfeM what target? Aug 23 07:50:00 beagleboard Aug 23 07:50:41 nxssh, nxspool and nxesd have issues with configure, nxsensor has an issue with gkrellm, nxservice seems to have an issue with install Aug 23 07:50:49 uh looks ugly with that comment :) Aug 23 07:51:15 JaMa: true but commit often, commit early :-) Aug 23 07:51:51 ka6sox4: nxbuilder is just php, no idea where to install it but should be easy to fix (do you have a dir?) Aug 23 07:52:00 nxscripts ditto Aug 23 07:52:10 rest builds for me Aug 23 07:52:30 still sussing that out...its kind of archaic Aug 23 07:52:30 (well ok, didn't look at nx-X11-compat at all) Aug 23 07:53:02 * XorA|gone is holding out hopes of neatX/qtNX overtaking NX Aug 23 08:11:34 XorA: don't know qtNX (actually was mostly unaware of freeNX before ka6sox4 brought it up, if needed I used vnc, but most of the time I just do an ssh into the system and if needed start up an X app with display set to my workstation Aug 23 08:12:48 eFfeM some things I like is sound which vnc doesn't do Aug 23 08:15:29 eFfeM: neatX is google re-implementation of freenx Aug 23 08:17:33 is it Qt Based? Aug 23 08:25:16 ka6sox4: qtNX is Qt-based NX client Aug 23 08:26:21 okay now I have something else to look into tommorrow...with neatX Aug 23 08:27:17 hi eFfeM Aug 23 08:28:16 can u me what is the diff between bitbake virutal/something and bitbake something Aug 23 08:29:19 XorA, interesting...re-implemented in python. Aug 23 08:33:50 Noor: "virtual/something" can be provided by few recipes Aug 23 08:34:03 hrw, you just beat me Aug 23 08:49:28 khem, Tartarus my spe image with minimal builds fine Aug 23 08:49:48 going to office, back later Aug 23 08:50:27 hrw: and then how specific package is selected to build Aug 23 08:50:47 like bitbake virtual/kernel builds the kernel Aug 23 08:51:07 as u said this will be provided by different recipes Aug 23 08:51:19 how correct recipe is selected then Aug 23 09:16:43 http://www.kernel.org/pub/software/scm/git/docs/git-notes.html Aug 23 09:16:47 can be handy Aug 23 09:19:33 good morning Aug 23 09:22:18 isnt there special handling to virtual in bitbake? Aug 23 09:23:33 we have to set PREFERED_PRIVIDER_virtual_kernel to particular package for example "linux" Aug 23 09:23:57 cant we use PREFERED_PRIVIDER_kernel Aug 23 09:24:53 what benefit get by using virtual word? Aug 23 09:28:16 hi, Aug 23 09:40:23 for generating rootfs for beagleboard, we change the local.conf. that is MACHINE = "beagleboard". Aug 23 09:40:29 But if i am using "user define board", then what i must change in local.conf and other. that is MACHINE = ??? Aug 23 09:44:11 florian: good morning Aug 23 09:47:24 akumar: you will have to write a board description fpr your custom board. this isn#t too hard, usually you can base on a similar one Aug 23 09:48:58 florian: for this any link, which can help me. Aug 23 09:51:25 akheron: http://wiki.openembedded.net/index.php/Adding_a_new_Machine Aug 23 09:55:34 florian: thanks, let me see. Aug 23 10:00:09 03Stefan Schmidt  07org.openembedded.dev * r040ba60bd0 10openembedded.git/recipes/ibrdtn/ (7 files in 2 dirs): ibrdtn: Remove outdated versions and patches Aug 23 10:00:16 03Stefan Schmidt  07org.openembedded.dev * r12c96bfc67 10openembedded.git/recipes/ibrdtn/ (6 files in 2 dirs): ibrdtn: Add latest release and improve .inc usage Aug 23 10:08:05 re Aug 23 10:09:29 eFfeM_work: wb Aug 23 10:46:49 noor, wrt the uicmoc4 changes: these should be applied together and therefore should be in one patch Aug 23 10:47:23 and ideally you should introduce an INC_PR variable Aug 23 10:54:39 Hello everyone. First time for me in this channel so litte intro: I am ahappy owner of a Neo Freerunner trying to dev some apps with OE. I came to OE because of SHR for my Freerunner. Aug 23 10:57:59 Now my question: does someone already tried to build the recipe automoc-native ? because I think there is an issue as it does not build. I am trying to understand what is wrong (apparently a macro not defined Q_WS_QWS) but I would not be agaainst some help :) Aug 23 10:59:12 I think that Qt should define it but it is not the case.. so Aug 23 11:07:42 khem, pulled head, removed TMPDIR, no pstage outside TMPDIR; full rebuild; seemed initially to work but then it turned that I had kicked the offending package (libusb) from my image; when trying to build libusb it still gives me the hidden symbol/nonrepresentable secion error Aug 23 11:09:07 and verified that my compiler and lib are rebuild Aug 23 11:14:20 btw the build also barfs at libc.so not being found, the odd thing is that the file is there ./ppce500v2-oe-linux-gnuspe/usr/lib/libc.so; however libc.so.6 is in ./ppce500v2-oe-linux-gnuspe/lib/libc.so.6 Aug 23 11:21:43 OK Aug 23 11:21:50 eFfeM_work: Aug 23 11:22:10 eFfeM_work: i'll resubmit the patch Aug 23 11:36:34 Noor, please also update patchwork Aug 23 11:37:09 patchwork? Aug 23 11:37:40 http://wiki.openembedded.org/index.php/Patchwork Aug 23 11:38:02 http://patchwork.openembedded.org/project/openembedded/list/ Aug 23 11:38:18 if you are registered you can set your patches to superseded or so Aug 23 11:38:30 this picks up the patches and stores them so the do no t get lost Aug 23 11:38:31 OK Aug 23 11:52:46 eFfeM_work: if certain package is not using any PR and PR is coming from inc file just like uicmoc4-native Aug 23 11:53:17 Noor: better to convert those to use INC_PR first Aug 23 11:53:19 if we change PR to INC_PR in inc file do we have to set it in package recipe as well Aug 23 11:53:39 yes, set INC_PR to highest PR in any recipe including that .inc Aug 23 11:53:41 Noor: make introducing INC_PR a separate patch Aug 23 11:53:55 and then put PR="${INC_PR}" to every recipe including that .inc Aug 23 11:54:00 and then put PR="${INC_PR}.0" to every recipe including that .inc Aug 23 11:54:20 and if you modify the inc file later on, increment INC_PR, if you modify the recipe increase the .0 in PR Aug 23 11:54:46 and to be more exact "set INC_PR to highest PR + 1" Aug 23 11:57:09 if we dont use INC_PR In package then package recipe PR will be 0 right? Aug 23 11:58:29 if there is no .inc file there is no INC_PR and PR will start with "r0", if there is INC_PR then PR is as JaMa said before Aug 23 11:58:49 and 0 or r0 is the default for PR (I think it is r0) Aug 23 11:59:12 eFfeM_work: should it work If I resubmit the previous patch as it is and then add another patch to introduce INC_PR? Aug 23 11:59:28 or u need first INC_PR then new style staging patch Aug 23 11:59:28 better do it the other way around Aug 23 11:59:32 yes Aug 23 12:00:13 tell me one more thing. its related to git . Can we revert chnges of only one file? Aug 23 12:00:32 I used git reset --hard Aug 23 12:00:41 it cleared all my changes Aug 23 12:01:28 git reset --soft does not reset Aug 23 12:01:41 git checkout file gives you the original file from git Aug 23 12:10:30 Noor: actually wanted to say that git reset --soft does not remove files, but that is probalby not what you wanted Aug 23 12:11:09 I want to revert my say uicmoc4-native.inc Aug 23 12:11:22 how will I do that Aug 23 12:12:11 if i do git reset hard it discards changes in all file and bring me to the particular file whoes hash i give in reset command Aug 23 12:12:18 git checkout uicmoc4-native.inc will revert to the version that is in the archive Aug 23 12:12:29 OK Aug 23 12:12:45 and if Ihave commit it to my local repository Aug 23 12:12:57 will it bring from server or my checked in file Aug 23 12:13:05 if you want to revert the commit and keep the chagned files do Aug 23 12:13:14 git reset --soft HEAD^ Aug 23 12:13:34 or use a hash, HEAD^ is revert one commit Aug 23 12:13:55 OK Aug 23 12:14:00 I'll try that Aug 23 12:53:02 anyone tried building cmake-native with nucrses-5.7? Aug 23 12:56:32 yes Aug 23 12:56:37 works for me on 2 buildhosts Aug 23 12:56:53 hm; get this: NOTE: package module-init-tools-cross-3.2.2-r5: task do_package: Started Aug 23 12:56:53 NOTE: Couldn't find shared library provider for libc.so.6 Aug 23 13:00:31 khem, btw wrt my hidden symbol problem: that was with gcc 4.4.4, I've reverted to gcc 4.3.3 by saying PREFERRED_GCC_VERSION = "4.3.3" in my local.conf; rm tmpdir; after that (and some patience) libusb builds Aug 23 13:00:36 either with gcc-4.5? Aug 23 13:02:12 echo "this is tekst" | cut -f3 -d ' ' - how to get rid of \n at the end of 'tekst'? Aug 23 13:02:25 http://pastebin.com/NbpFxuxc Aug 23 13:02:42 is what I see Aug 23 13:02:44 hrw: echo -n Aug 23 13:03:28 eFfeM_work: what if it is not 'echo' but other tool which gives me tekst? Aug 23 13:03:53 hrw: you can use sed inbetween or tr Aug 23 13:04:09 dpkg-deb -I libc6-armel-cross_2.12.1-0ubuntu1hrw1_all.deb |grep Section|cut -d' ' -f3 - gives "devel\n" Aug 23 13:05:22 tr -d \n ;) Aug 23 13:05:37 yup Aug 23 13:10:40 hrw: head -c -1 also works Aug 23 13:20:09 RP__: ping Aug 23 13:20:38 hi zecke Aug 23 13:20:58 pb_: ping as well. :) Aug 23 13:46:57 ok, cmake native is trying to link againts host libs Aug 23 14:43:58 I tried to build the Angstrom image on the latest code and it failed with the error below Aug 23 14:43:59 ERROR: Task 4232 (/beagleboard/angstrom/angstrom-setup-scripts/sources/openembedded/recipes/intltool/intltool_0.40.3.bb, do_configure) failed with 256 Aug 23 14:44:09 does anyone have an idea? Aug 23 14:59:44 ~pastebin Aug 23 15:00:08 ani: Please copy/paste all of the do_configure log (the location should be a few lines before what you pasted in your terminal) there, once the bot speaks up Aug 23 15:02:27 ERROR: TaskFailed event exception, aborting ERROR: Build of /beagleboard/angstrom/angstrom-setup-scripts/sources/openembedded/recipes/intltool/intltool_0.40.3.bb do_configure failed ERROR: Task 4232 (/beagleboard/angstrom/angstrom-setup-scripts/sources/openembedded/recipes/intltool/intltool_0.40.3.bb, do_configure) failed with 256 ERROR: Function do_configure failed NOTE: Task failed: ('function do_configure failed', '/beagleboard/angstrom/a Aug 23 15:03:52 m-linux-gnueabi/intltool-0.40.3-r4.0/temp/log.do_configure.20733') ERROR: Logfile of failure stored in: /beagleboard/angstrom/angstrom-setup-scripts/build/tmp-angstrom_2008_1/work/armv7a-angstrom-linux-gnueabi/intltool-0.40.3-r4.0/temp/log.do_configure.20733 Log data follows: | automake (GNU automake) 1.11.1 | Copyright (C) 2009 Free Software Foundation, Inc. | License GPLv2+: GNU GPL version 2 or later Aug 23 15:04:15 | There is NO WARRANTY, to the extent permitted by law. | | Written by Tom Tromey | and Alexandre Duret-Lutz . | AUTOV is 1.11 | NOTE: Executing autoreconf --verbose --install --force --exclude=autopoint -I /beagleboard/angstrom/angstrom-setup-scripts/build/tmp-angstrom_2008_1/work/armv7a-angstrom-linux-gnueabi/intltool-0.40.3-r4.0/intltool-0.40.3/ -I/beagleboard/angstrom/angstrom-setup-scripts/build/ Aug 23 15:05:28 ani: Please put the whole of the logfile into http://pastebin.com and tell us the url Aug 23 15:05:59 e/aclocal | autoreconf: Entering directory `.' | autoreconf: configure.in: not using Gettext | autoreconf: running: aclocal -I /beagleboard/angstrom/angstrom-setup-scripts/build/tmp-angstrom_2008_1/work/armv7a-angstrom-linux-gnueabi/intltool-0.40.3-r4.0/intltool-0.40.3/ -I /beagleboard/angstrom/angstrom-setup-scripts/build/tmp-angstrom_2008_1/sysroots/armv7a-angstrom-linux-gnueabi/usr/share/aclocal-1.11 -I /beagleboard/angstrom/angstrom-setu Aug 23 15:06:39 I guess the log is too long. Can I email it to an ID? Aug 23 15:06:51 ani: Please use http://pastebin.com Aug 23 15:14:14 http://pastebin.com/ET20P2Wr Aug 23 15:16:16 03Khem Raj  07org.openembedded.dev * rde2e963701 10openembedded.git/recipes/linux/linux-efikamx_2.6.31.bb: Aug 23 15:16:16 linux-efikamx_2.6.31.bb: Add recipe for linux kernel used for Efika MX Aug 23 15:16:16 Signed-off-by: Khem Raj Aug 23 15:16:16 Acked-by: Frans Meulenbroeks Aug 23 15:16:16 03Khem Raj  07org.openembedded.dev * r8fdc993235 10openembedded.git/conf/machine/efikamx.conf: Aug 23 15:16:16 efikamx.conf: Add config file for efika MX Aug 23 15:16:17 Signed-off-by: Khem Raj Aug 23 15:16:17 Acked-by: Frans Meulenbroeks Aug 23 15:17:13 03Khem Raj  07org.openembedded.dev * r41569c5f4e 10openembedded.git/recipes/ (172 files in 96 dirs): Aug 23 15:17:13 ncurses_5.7.bb,recipes: Make DEFAULT_PREFERENCE Aug 23 15:17:13 * Select 5.7 if none pinned. Aug 23 15:17:13 * Bump PR for recipes which have ncurses dependency. Aug 23 15:17:13 Signed-off-by: Khem Raj Aug 23 15:17:14 Acked-by: Martin Jansa Aug 23 15:17:35 03Khem Raj  07org.openembedded.dev * r231757d9f3 10openembedded.git/recipes/u-boot/u-boot-efikamx_1.1.8.bb: Aug 23 15:17:35 u-boot-efikamx_1.1.8.bb: Efika MX bootloader Aug 23 15:17:35 Signed-off-by: Khem Raj Aug 23 15:17:35 Acked-by: Frans Meulenbroeks Aug 23 15:37:51 gm Aug 23 16:35:09 03Martin Jansa  07org.openembedded.dev * r6504340970 10openembedded.git/recipes/xserver-common/ (2 files in 2 dirs): Aug 23 16:35:09 xserver-common: update 89xTs_Calibrate.xinput_calibrator Aug 23 16:35:09 * after 5c3846f9bbf1dbb4a9ee1d916107747c2f9dc909 we already know that user won't be forced to provide new calibration, so no need to request Display,CPU resources Aug 23 16:35:09 Signed-off-by: Martin Jansa Aug 23 16:35:10 03Martin Jansa  07org.openembedded.dev * r45e4104cf9 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Aug 23 16:35:10 EFL: bump SRCREV a bit more again, fix for missing cursor in text fields Aug 23 16:35:10 Signed-off-by: Martin Jansa Aug 23 16:35:11 03Martin Jansa  07org.openembedded.dev * r003d072660 10openembedded.git/recipes/openmoko-3rdparty/mcnavi_0.2.10.bb: Aug 23 16:35:11 mcnavi: bump PR after EFL .so names change Aug 23 16:35:11 Signed-off-by: Martin Jansa Aug 23 16:35:12 03Martin Jansa  07org.openembedded.dev * r22e3772f79 10openembedded.git/ (conf/distro/shr.conf recipes/e17/e-wm_svn.bb): Aug 23 16:35:12 e-wm: introduce E17_DEFAULT_CONFIG to remove e-wm-config-default from RRECOMMENDS if needed Aug 23 16:35:13 Signed-off-by: Martin Jansa Aug 23 16:35:15 03Martin Jansa  07org.openembedded.dev * rdf080ed906 10openembedded.git/ (9 files in 6 dirs): (log message trimmed) Aug 23 16:35:16 e-wm: use illume2 instead of illume1, fix illume-keyboard paths and locales packaging Aug 23 16:35:31 * illume2 and illume1 shouldn't be compiled both anymore Aug 23 16:35:31 http://trac.enlightenment.org/e/changeset/50884 Aug 23 16:35:31 keep only illume2 as it's preferred by upstream Aug 23 16:35:31 * separate illume2 (illume-home) to new package e-wm-config-illume2 and Aug 23 16:35:31 make it default Aug 23 17:11:34 likewise: howdy Aug 23 17:12:19 likewise: I have added support for Genisi efika-MX to OE and I know you own efika-MX :) would you try it out ? Aug 23 17:15:34 khem: ah, I *just* wrote you an email response. Aug 23 17:15:42 khem: I have a MX for you, if you want one. Aug 23 17:15:56 khem: not the newest board revision, but Ubuntu runs fine. Aug 23 17:32:33 I had to hack the cmake-native recipe to get it to build w/ gcc-4.5 and F13 build machine Aug 23 17:33:23 I am going to push the patch and hope it is a NOP on machines that already work :) Aug 23 17:36:15 Crofton: how can test this easiest? Aug 23 17:36:24 just build cmake-native will do? Aug 23 17:37:11 yeah Aug 23 17:37:15 it failed for me Aug 23 17:37:31 I had to add ltinfo to the list of libs to link ccmake against Aug 23 17:37:48 also, it is using host libraries, which may be the reason for the hack Aug 23 17:38:18 cmake makes my eyes bleed, so I stopped trying to fix the host libs once I had a ahck Aug 23 17:54:57 03Chris Larson  07org.openembedded.dev * r9d0d70da60 10openembedded.git/classes/patch.bbclass: Aug 23 17:54:57 patch: allow importing patches into other dirs Aug 23 17:54:57 Use the 'patchdir' flag. If relative, it's assumed to be relative to ${S}. Aug 23 17:54:57 It creates a patchset class per patch directory, so for quilt this results in Aug 23 17:54:57 multiple .pc/patches directories to manage each of the patchsets. Aug 23 17:54:57 Signed-off-by: Chris Larson Aug 23 18:01:56 ever since i saw their stupid ass filename, "CMakeLists.txt", I've avoided getting in depth with cmake Aug 23 18:01:59 just bugged me Aug 23 18:02:01 heh Aug 23 18:03:48 kergoth, maybe that's because of windows users Aug 23 18:04:26 maybe it's because of file association, if it's not .txt no text editor would open it Aug 23 18:05:49 if so, that's a really stupid reason. it's not that hard to create an installer or even a script to install the association Aug 23 18:06:48 03Koen Kooi  07org.openembedded.dev * r4522417763 10openembedded.git/recipes/base-files/ (base-files/beagleboard/profile base-files_3.0.14.bb): Aug 23 18:06:48 Revert "base-files: beagleboard: profile: use nano" Aug 23 18:06:48 This is a distro setting, not a machine setting Aug 23 18:06:48 This reverts commit c323496195aaec26055da823a6c0166c7836c6b2. Aug 23 18:07:04 03Koen Kooi  07org.openembedded.dev * r60844bb680 10openembedded.git/conf/machine/beagleboard.conf: Aug 23 18:07:04 Revert "beagleboard.conf: use qt4-x11-free-gles" Aug 23 18:07:04 This is a distro setting, not a machine setting Aug 23 18:07:04 This reverts commit f7635a8467bb23901c2cfb01fe2371d277be9af7. Aug 23 18:07:04 kergoth, you are lucky Aug 23 18:07:43 someone created a "testing-testing" branch? nice naming there Aug 23 18:08:22 rofl Aug 23 18:08:44 always best to test the test to make sure it works. Aug 23 18:29:01 re Aug 23 18:32:52 ls Aug 23 18:32:52 likewise: I have a MX genisi donated me one :) Aug 23 18:34:07 khem, any idea about the ppc/spe problem I encountered with 4.4.4 ? Aug 23 18:38:55 eFfeM: as I said yesterday Aug 23 18:39:22 if the problem still persists try to use 4.5 Aug 23 18:39:29 03Martin Jansa  07org.openembedded.dev * rc0916bda7c 10openembedded.git/ (conf/distro/shr.conf recipes/e17/e-wm_svn.bb): Aug 23 18:39:29 Revert "e-wm: introduce E17_DEFAULT_CONFIG to remove e-wm-config-default from RRECOMMENDS if needed" Aug 23 18:39:29 * This reverts commit 22e3772f799948bf4a7806e74d83948ee2b2e1f4. Aug 23 18:39:29 * e-wm-config-default is still needed to show 1st time wizard (selecting Aug 23 18:39:29 which config to use) Aug 23 18:39:30 * sorry I should have tried after removing ~/.e (I thought I did) Aug 23 18:39:31 khem, I moved to dev head, rm-ed tmpdir, but still the problem remains Aug 23 18:39:35 03Martin Jansa  07org.openembedded.dev * r423e9ed161 10openembedded.git/recipes/xserver-common/ (2 files in 2 dirs): Aug 23 18:39:35 xserver-common: fix patch Aug 23 18:39:35 * never ever modify patch by hand after testing it, sorry for noise Aug 23 18:40:01 eFfeM: which package is this showing up in Aug 23 18:40:31 libusb Aug 23 18:40:51 more specific libusb_0.1.12.bb Aug 23 18:43:00 4.3.3 has no problems with it, didn't try others Aug 23 18:44:08 wrt 4.5: not too sure about the maturity of it, so not too eager to make that step Aug 23 18:55:11 hi :) Aug 23 18:55:30 is there a script or utility which could perform git pulls and rebuild updated packages? Aug 23 18:57:22 git pull && bitbake console-image Aug 23 18:57:30 or whatever image you want to build Aug 23 19:06:50 eFfeM, what you call freenx-client is really the server Aug 23 19:07:26 eFfeM: and this will rebuild packages which have newer versions? Aug 23 19:09:28 Techqua69 Aug 23 19:09:35 hrmph Aug 23 19:09:38 * kergoth mutters Aug 23 19:09:47 filip, yes, bitbake will always build the current selected versions of things Aug 23 19:09:52 given your and the distro's preferences Aug 23 19:10:54 kergoth: even if console-image was built yesterday? and for example there was a version bump in coreutils, then console-image will be rebuilt? Aug 23 19:11:02 yes. Aug 23 19:11:10 bitbake doesn't care that it's arleady built console-utils Aug 23 19:11:17 it goes by the name, version, and revision Aug 23 19:11:19 not just name Aug 23 19:11:30 er, coreutils Aug 23 19:12:34 okay, thanks for the clarification :) Aug 23 19:13:59 ka6sox-work: feared that, but could not find any code at all that builds something with the name nxclient Aug 23 19:14:07 filip: yes Aug 23 19:15:08 I'll let you rename it Aug 23 19:15:13 building it now Aug 23 19:15:26 ok, can you tell me what I need to get a client ? Aug 23 19:19:08 not gottten that far but I will do that when I am finished with this. Aug 23 19:20:03 well, I've downloaded all of the source packages, there is no one with nxclient in the name and a grep on the expanded sources did not lead to a makefile or so that delivers nxclient Aug 23 19:22:24 03Philip Balister  07org.openembedded.dev * r70b3cf1ad3 10openembedded.git/recipes/tasks/task-sdk-gnuradio-native.bb: task-sdk-gnuradio-native.bb : Update name of python dev package. Aug 23 19:22:35 03Philip Balister  07org.openembedded.dev * re5bed7acf5 10openembedded.git/recipes/python/python-numpy_1.4.1.bb: python-numpy_1.4.1.bb : Add version that builds with gcc-4.5. Aug 23 19:22:36 03Philip Balister  07org.openembedded.dev * r5ea3047995 10openembedded.git/recipes/cmake/ (cmake-2.8.2/tinfo.patch cmake-native_2.8.2.bb): Aug 23 19:22:36 cmake-native_2.8.2.bb : Force ccmake to link against -ltinfo. Aug 23 19:22:36 I needed this to build cmake-native for angstrom-2010 on a F13 machine. Aug 23 19:22:36 Hopefully, this does not break other builds. Aug 23 19:23:40 as you can see, I am a little worried about the cmake change :) Aug 23 19:33:33 eFfeM, the most active one is http://sourceforge.net/projects/opennx/ Aug 23 19:34:23 use ming to build it Aug 23 19:34:25 ka6sox-work: ah ok, expected this to come from nomahcine too Aug 23 19:34:56 eFfeM: whats your TARGET_ARCH Aug 23 19:35:08 wrt ppc issue Aug 23 19:35:53 khem, this is based upon calamari machine (e500v2) Aug 23 19:36:19 I am interested in TARGET_ARCH setting Aug 23 19:36:28 actually it is different hw, and I just copied calamari.conf to my own; didn't make a change Aug 23 19:36:29 one sec Aug 23 19:36:37 eFfeM, on this page:http://www.nomachine.com/sources.php they do have client programmes and libraries. Aug 23 19:36:43 need to vpn to work to find it Aug 23 19:36:47 but I don't know how complete they are. Aug 23 19:36:59 calamari has TARGET_ARCH = "powerpc" Aug 23 19:37:34 ka6sox-work: thre is no nxclient program there, the sourceforge page you mention says nxclient is proprietary Aug 23 19:37:43 khem, correct Aug 23 19:37:44 ka6sox-work, ming like the gnash test suite? Aug 23 19:37:47 hi btw Aug 23 19:37:55 hi Aug 23 19:38:09 eFfeM: ok. Aug 23 19:38:20 GNUtoo|laptop, dunno Aug 23 19:38:22 ah must be related to the new comit about nx-client Aug 23 19:38:22 ok Aug 23 19:38:30 lemme try to reproduce it with calamari and 4.4.4 Aug 23 19:38:38 eFfeM, that was our thought too... Aug 23 19:38:56 khem, thanks! Aug 23 19:39:15 i've just build minimal with no frills Aug 23 19:39:17 eFfeM: btw. do you have you gcc-cross build logs Aug 23 19:39:23 yes Aug 23 19:39:28 I am interested in do_compile Aug 23 19:39:37 ok, need to copy it over, one sec Aug 23 19:41:49 eFfeM, uses wxWidgets and needs the nxssh lib as well as the nxesd for sound Aug 23 19:43:37 ka6sox-work: might look into it later not sure if I get to it today Aug 23 19:45:00 btw any shell wizard here? how do I set an env var in a shell while loop in a way that I can access it outside the loop ? Aug 23 19:45:16 khem: http://filebin.ca/mpqygp/log.do_compile.27078 Aug 23 19:45:18 2.78 MB Aug 23 19:45:22 VARIABLE=boo Aug 23 19:45:42 no wizardry required Aug 23 19:46:40 I'm using OpenNX. works fine Aug 23 19:47:07 XorA: thought that too, but i failed miserably Aug 23 19:47:10 ls | while read $i; do xx=1; done; echo $xx Aug 23 19:47:19 no return value Aug 23 19:47:22 I'm using Google's open-source NX server, and OpenNX open-source client (for Mac) Aug 23 19:47:45 likewise: ah but ka6sox-work wants an openembedded version Aug 23 19:48:00 and if you are not on intel hw you need cto compile yourself Aug 23 19:48:07 eFfeM: ok your gcc seems to be compiled correctly as I expected Aug 23 19:48:25 eFfeM: what was error message once again for my reference Aug 23 19:48:37 khem, ok, that makes me partly happy, but also unhappy Aug 23 19:48:45 khem, let me check, one sec Aug 23 19:50:42 eFfeM: ls | { while read $i; do xx=1; done; echo $xx ;} Aug 23 19:50:56 eFfeM: your command isnt quite doing what you think Aug 23 19:51:26 XorA: this is a simplfied example, I want to set a var in the while Aug 23 19:51:43 ls is piped to while, which terminates on the done, the echo $xx executes totally seperately Aug 23 19:52:42 khem here is the pastie of the last lines of libusb compile Aug 23 19:52:53 XorA: aarg, yes ofc Aug 23 19:54:46 XorA; thanks, ( ) fixed it Aug 23 19:54:47 ls | (while read $i; do xx=1; done; echo $xx) Aug 23 20:00:07 03Martin Jansa  07org.openembedded.dev * r4872124db6 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Aug 23 20:00:07 linux-openmoko-2.6.32: upgrade to 2.6.32.20 and add patch for WSOD with 242 timming Aug 23 20:00:07 Signed-off-by: Martin Jansa Aug 23 20:00:08 03Martin Jansa  07org.openembedded.dev * r5ff016d6f4 10openembedded.git/recipes/shr/shr-e-gadgets_git.bb: Aug 23 20:00:09 shr-e-gadgets: add edje_convert call before it's fixed in git repo Aug 23 20:00:09 Signed-off-by: Martin Jansa Aug 23 20:07:59 03Frans Meulenbroeks  07org.openembedded.dev * rd014258b95 10openembedded.git/recipes/htmldoc/ (3 files): Aug 23 20:07:59 Revert "htmldoc[-gui]: remove old versions" Aug 23 20:07:59 This reverts commit e5287904a9dddc8ddc0f486d8526bc8cfda427b7. Aug 23 20:07:59 Reverted request of the recipe maintainer Aug 23 20:07:59 This version is the latest official version, the 1.9 version Aug 23 20:08:00 is a development snapshot Aug 23 20:08:00 Signed-off-by: Frans Meulenbroeks Aug 23 20:21:15 Jay7: I've also seen that error in libusb Aug 23 20:21:32 that error = link error: libc.so, needed by libgcc_s.so, not found Aug 23 20:53:52 03Frans Meulenbroeks  07org.openembedded.dev * rc566d33ddc 10openembedded.git/recipes/freenx/nx-client_3.4.0.bb: Aug 23 20:53:52 nx-client: renamed to nx-server Aug 23 20:53:52 actually this was the server component. Aug 23 20:53:52 Apparenty nomachine has not made client sources available Aug 23 20:53:52 still has the issues mentioned in the original commit msg Aug 23 20:53:52 Signed-off-by: Frans Meulenbroeks Aug 23 20:59:47 eFfeM where is the link Aug 23 21:03:31 khem, link to what ? Aug 23 21:03:56 aarg forgot to paste the link Aug 23 21:04:51 khem: http://www.pastie.org/1110706 Aug 23 21:04:56 sorry, thought I'd sent before Aug 23 21:05:18 this one also has the libc.so warning Aug 23 21:06:29 gone for now, but you can still address me at eFfeM_work and I'll see it tomorrow morning Aug 23 21:06:48 thanks alot for looking into this Aug 23 21:07:12 * eFfeM is calling it a day, cya all Aug 23 21:21:38 hello, gays! I have a problem building rootfs jffs2 with OE and I don't know how to debug or how to see what exactly wrong Aug 23 21:22:19 mkfs.jffs2 -x lzo --root=... --eraseblock=0x20000 -n Aug 23 21:22:26 ...temp/run.do_rootfs.23382: line 468: 29172 Aborted Aug 23 21:22:32 ERROR: Function do_rootfs failed Aug 23 21:22:39 NOTE: package base-image-1.0-r0: task do_rootfs: Failed Aug 23 21:30:14 03Chris Larson  07org.openembedded.dev * r5678affd4e 10openembedded.git/classes/base.bbclass: Aug 23 21:30:14 base: use a variable for the modules to import/inject Aug 23 21:30:14 Signed-off-by: Chris Larson Aug 23 21:31:55 dv: a good place to start in debugging builds is to try running the failing command by hand Aug 23 21:35:10 dear Wormo, mkfs.jffs2 gets signal "SIGABRT". that's all. why it goes wrong - i don't know Aug 23 21:37:22 dv: what if you run your mkfs.jffs2 without arguments, does it show help message or still aborts? Aug 23 21:37:39 yes, it shows help Aug 23 21:39:39 check dmesg for an OOM Aug 23 21:40:37 no oom. and I have 4 Gb or RAM in my pc Aug 23 21:41:11 at least half of 4fb is free Aug 23 21:48:32 selinux report? Aug 23 21:50:35 * kergoth mutters Aug 23 21:50:59 no selinux setup at this pc! Aug 23 21:51:30 dv: what distro are you using as build machine? Aug 23 21:51:46 Mandriva PP Aug 23 21:51:54 03Chris Larson  07org.openembedded.dev * r0a11eb7146 10openembedded.git/classes/base.bbclass: Aug 23 21:51:54 Revert "base: use a variable for the modules to import/inject" Aug 23 21:51:54 This reverts commit 5678affd4e1b6453d6802b8cfb7823fec6f47269. Aug 23 21:52:32 ah Mandrake is a trip into bizzaro land for me :-)( Aug 23 21:52:35 hrm Aug 23 21:53:04 anyway, this mkfs.jffs2 is NOT the part of Mandriva. It was build specially for OE: ./tmp/staging/i686-linux/usr/bin/mkfs.jffs2 Aug 23 21:53:16 dv: yes, I was just curious :) Aug 23 21:54:04 other strange thing, It creates an output! but ends with SIGABRT and no other tasks executed after Aug 23 21:55:59 dv: tried running it under gdb? Aug 23 21:56:17 under strace. nothing! Aug 23 21:56:30 dv: also does it abort if you use different compressor or smaller set of files Aug 23 21:57:54 without "-x lzo" SIGABRT raise too Aug 23 21:58:47 dv: did OE produce any other native tools that work right? Aug 23 21:59:26 dv: also could still run in gdb rather than strace Aug 23 21:59:59 at least "cat" works! Aug 23 22:02:16 Starting program: /home/dv/workspace/oe_sam/tmp/staging/i686-linux/usr/bin/mkfs.jffs2 -x lzo --root=/home/dv/workspace/oe_sam/tmp/rootfs --faketime --output=/home/dv/workspace/oe_sam/tmp/deploy/glibc/images/at91sam9m10g45ek/Angstrom-base-image-glibc-ipk-2009.X-stable-at91sam9m10g45ek.rootfs.jffs2 --pad --little-endian --eraseblock=0x20000 -n Aug 23 22:02:36 Program received signal SIGABRT, Aborted. Aug 23 22:02:38 0xffffe424 in __kernel_vsyscall () Aug 23 22:02:48 did you say selinux was on, dv? Aug 23 22:03:06 said it was not Aug 23 22:03:09 no, selinux is switched off Aug 23 22:03:12 k Aug 23 22:03:26 and i'm calling programs as usual user Aug 23 22:04:53 Hello all, I am running ubuntu and did bitbake task-sdk-host. When I was trying to install the .ipk packages in my machine, I've got a error status Error: Wrong architecture 'i686-armv7a-sdk' Aug 23 22:06:54 The output of an SDK comes from bitbake meta-toolchain (or one of the meta-toolchain based recipes) Aug 23 22:06:57 And those outputs are tarballs Aug 23 22:11:27 I searched the build/tmp/deploy and I didn't see any .tar file except the .tar of my filesystem images Aug 23 22:11:41 yes Aug 23 22:11:47 you need to bitbake meta-toolchain Aug 23 22:11:50 not just task-sdk-host Aug 23 22:13:14 ahh ok.. got it.. thanks Aug 23 22:55:32 dv: it seems like you've got a problem with your native toolchain built by oe, because there aren't any assert() statements in mkfs.jffs2 source Aug 24 00:36:07 cbrake: you could note the combinations that are known to build in the message attached to an annotated tag Aug 24 00:36:14 cbrake: either instead of or in addition to the wiki page Aug 24 00:42:36 03Chris Larson  07org.openembedded.dev * r744016016a 10openembedded.git/classes/base.bbclass: Aug 24 00:42:36 base: use a variable for the modules to import/inject Aug 24 00:42:36 Signed-off-by: Chris Larson Aug 24 01:01:38 can someone point me to an example of two packages generated from the same source tarball? Aug 24 01:04:59 hollisb: how about an example with lots of packages - perl Aug 24 01:05:31 I'll look, but I was hoping for something as simple as possible (just slightly complex :) Aug 24 01:07:17 ok how about xscreensaver then Aug 24 01:07:45 less complicated than perl Aug 24 01:14:08 hollisb: every recipe automatically generates multiple packages if the files exist. Aug 24 01:14:20 hollisb: read conf/bitbake.conf, the PACKAGES and FILES_ variables Aug 24 01:14:26 ok, thanks Aug 24 01:20:17 I have a problem compiling mtd-utils 1.3.1, http://pastebin.ca/1923853 if someone has an idea what the root cause is. Aug 24 01:52:54 SRC_URI = "file:///home/balister/src/git/linux-omap-2.6-export \ Aug 24 01:53:14 this should take the kernel source in the export directory and copy it in to work? Aug 24 01:53:20 I'm getting an empty directory **** ENDING LOGGING AT Tue Aug 24 02:59:57 2010