**** BEGIN LOGGING AT Sat Sep 25 02:59:57 2010 Sep 25 06:14:52 Weird. Two boxes both running the same OS with the same packages and same install. One built openssl-native the other didnt Sep 25 06:18:33 didn't means build error? Sep 25 06:20:03 Correct. Sep 25 06:23:38 so something has to be different Sep 25 06:23:45 what's the error? Sep 25 06:23:54 Agreed Sep 25 06:24:10 thats what i am looking at now.. Sep 25 06:24:20 i am getting one error while compliling cpp program .. arm-angstrom-linux-gnueabi/lib/libgcc_s.so, not found (try using -rpath or -rpath-link) Sep 25 06:24:24 One thing that pops to mind is that my fs is encrypted Sep 25 06:24:25 i am using gumstix Sep 25 06:24:35 for over Os Sep 25 06:25:21 ?? Sep 25 06:25:21 what does this /libgcc_s.so means Sep 25 06:25:25 anyone alive Sep 25 06:26:23 ?? Sep 25 06:27:00 ???????????????????????????/// Sep 25 06:28:06 Have you tried using rpath? Sep 25 06:28:22 libgcc_s.so is a linkable shared object Sep 25 06:28:59 how to rpatch?? Sep 25 06:29:05 how to rpath?? Sep 25 06:30:06 ?? Sep 25 06:30:10 I dont know. Have you tried google? gcc docs? Sep 25 06:30:31 yup Sep 25 06:32:09 ynezz: would you look at that! It works on my laptop now. No git pull or anything :| Sep 25 06:36:12 r0ck_: http://en.wikipedia.org/wiki/Rpath_%28linking%29 Sep 25 06:40:44 w|zzy: maybe it's the issue with the BB threads Sep 25 06:41:06 I'm seeying it sometimes also, just reruning the bitbake fixes it Sep 25 06:41:13 :) Sep 25 06:41:23 I ran it about 10 times before i had to go to work Sep 25 06:41:26 came back and pop it works! Sep 25 06:41:34 I even turned bb threads to 1 and -j1 Sep 25 06:42:02 hm, strange then Sep 25 06:43:29 Just knocked it up to 2 & -j 2 Sep 25 06:43:53 i would normally go -j 3 on this computer but i figured running a 2nd bb thread would run it hard Sep 25 06:49:54 Wow. BB doesnt handle threads very well Sep 25 06:57:35 gm Sep 25 06:58:04 hello eFfeM Sep 25 07:04:48 w|zzy: I see you found your way here :-) welcome aboard Sep 25 07:05:05 Few teething problems but i think i will settle in with OE Sep 25 07:05:10 Thanks Sep 25 07:06:00 feel free to ask questions if the docs do not help you, but note that weekends are somewhat less busy here than weekdays, and mornings are especially quiet Sep 25 07:06:55 What timezone are most people from? Sep 25 07:06:59 Because it is 1700 here! Sep 25 07:07:33 Europe I think Sep 25 07:07:44 it's 0907 here Sep 25 07:10:36 w|zzy: either europe or us and for the us people it is midnight Sep 25 07:11:17 where are u? e australia ? nw zealand? japan? Sep 25 07:11:42 Australia Sep 25 07:12:37 used to talking to people from eu & us Sep 25 07:31:33 What are the package sets? where would i find a list? Sep 25 08:01:01 w|zzy: was afk what do you mean with package set? Sep 25 08:01:31 for example "task-base" is a package set. Sep 25 08:02:03 ah it is a bundle of packages, I rarely use them, they live in recipes/tasks Sep 25 08:02:18 Ahh. There we go. Sep 25 08:02:48 Thankyou. Sep 25 08:03:10 I dont know why i am bothing trying to compile them on the laptop and why they are having so many issues. Sep 25 08:03:26 It worked flawlessly on my i7 running the same OS in a VM. Sep 25 08:04:07 best way to start is choose a distro (probably angstrom or minimal) and start building something small, i never build tasks directly they are used by images Sep 25 08:04:15 what issue do you have Sep 25 08:05:48 oe_runmake failed, makefile.org older than makefile Sep 25 08:05:51 few things like that. Sep 25 08:05:53 Only on the laptop Sep 25 08:06:20 I chose angstrom Sep 25 08:06:36 makefile.org older than makefile? never seen that Sep 25 08:06:42 I can just transfer the files that are compiled on my desktop to this laptop. Sep 25 08:06:49 time solved that issue. Sep 25 08:07:29 i just fixed solved the glibc issue by winding back bb threads to 1. Sep 25 08:07:35 I guess this lowly c2d isnt fast enough Sep 25 08:09:00 its just annoying when things dont work i feel compelled to work it out. Sep 25 08:09:22 i'm compiling on a core 2 duo if I recall correctly. D920 or so, 4 G of mem though Sep 25 08:09:49 03Frans Meulenbroeks  07org.openembedded.dev * r4144fd7afd 10openembedded.git/recipes/transmission/ (transmission.inc transmission_1.91.bb transmission_1.93.bb): Sep 25 08:09:49 transmission: moved to 1.93 Sep 25 08:09:49 also introduced INC_PR and added a few indents Sep 25 08:09:49 Signed-off-by: Frans Meulenbroeks Sep 25 08:10:05 I should be more correct.. Dual core centrino. Sep 25 08:10:19 It seems to be running ok since i pushed bb threads back. Sep 25 08:13:26 then maybe memory or cpu cycles is indeed the issue Sep 25 08:14:27 Yup. Not with the makefile issue but maybe with this one. Sep 25 08:33:52 I didnt realise TI was sending out http://elinux.org/OMAP_and_DaVinci_Software_for_Dummies Sep 25 08:36:56 s/was/is Sep 25 08:37:22 I wonder for what is the package revision with two numbers good for, ie. PR="r1.7" Sep 25 08:40:28 good for problems Sep 25 08:41:41 what kind of problems ? Sep 25 08:42:44 if you remove it later and there is DISTRO_PR bump.. you'll maybe miss that upgrade Sep 25 08:43:11 ah Sep 25 08:43:43 so it's not mandatory, but recommended Sep 25 08:44:06 or if there is distro with DISTRO_PR ie 9, then there is PR=r1.9 for them, then you add r1.7 and they have r1.7.9 which is still lower than original r1.9 Sep 25 08:45:37 just adding new recipes and I wonder what's correct Sep 25 08:45:50 better to start with "r0.0", right? Sep 25 08:46:33 or add the PR later as r1.0? r0 should be default anyway Sep 25 08:48:01 what recipe it is? Sep 25 08:48:28 why not just skip PR (keeping default r0) Sep 25 08:49:08 http://pastebin.com/nw06m6Ch Sep 25 08:49:18 yep, I'll skip that PR Sep 25 08:50:22 ericben: wrt " Sep 25 08:50:22 you can have a SOC which supports iwmmx and wish not using it so MACHINE_FEATURES seems sane here." Sep 25 08:50:29 does it make sense not to use it ? Sep 25 09:01:26 03Frans Meulenbroeks  07org.openembedded.dev * r35ab1c6a2a 10openembedded.git/recipes/transmission/transmission_2.04.bb: Sep 25 09:01:26 transmission: added 2.04, DP = "-1" until more tested Sep 25 09:01:26 Signed-off-by: Frans Meulenbroeks Sep 25 09:01:26 03Frans Meulenbroeks  07org.openembedded.dev * r9156b88e1a 10openembedded.git/recipes/libevent/libevent_1.4.14b-stable.bb: Sep 25 09:01:26 libevent: added 1.4.14b-stable; DP = "-1" until more tested Sep 25 09:01:27 Signed-off-by: Frans Meulenbroeks Sep 25 09:10:17 really should have set a locale Sep 25 09:12:31 oh, git fetcher hates me http://pastebin.org/1176982 Sep 25 09:12:57 I wonder what's causing the weird git error "fatal: Not a valid object name 1" Sep 25 09:13:48 ynezz: wrong SRCREV (or none) Sep 25 09:15:49 ah, nice descriptive error message "1" :) Sep 25 09:16:22 1 is apparently the default number Sep 25 09:16:57 I would understand it for svn, but for git :) Sep 25 09:17:30 but thanks, I had typo SRC_REV Sep 25 09:39:59 03Khem Raj  07master * r91a42c0ea9 10openembedded.git/recipes/binutils/binutils_cvs.bb: Sep 25 09:39:59 binutils_cvs.bb: Use FILESPATHPKG Sep 25 09:39:59 Signed-off-by: Khem Raj Sep 25 11:35:33 03Frans Meulenbroeks  07org.openembedded.dev * r07ecc255f3 10openembedded.git/recipes/mythtv/mythplugins_0.23+fixes.bb: Sep 25 11:35:33 mythwebplugins: fixed mythweb packaging Sep 25 11:35:33 Signed-off-by: Frans Meulenbroeks Sep 25 11:52:14 hi eFfeM : concerning iwmmx, I don't know in which state it's now but in the past we had many problems with that (toolchain + kernel). But that was the past :-) Sep 25 12:04:55 ah ok Sep 25 12:07:41 strange thing of the day : bitbake linux ... build fine but when it comes at "the following files were installed but not shipped in any package:" it displays only 2 lines / second which makes the packaging stage during very very long Sep 25 12:12:04 hi ericben I've that too Sep 25 12:15:50 hi GNUtoo|laptop : are you using bitbake master ? Sep 25 12:16:12 yes Sep 25 12:16:32 ok Sep 25 12:16:36 same thing here Sep 25 12:18:08 hm? Sep 25 12:18:12 whats up? Sep 25 12:18:23 kergoth told me he fixedmaster Sep 25 12:18:34 for this kind of problem ? Sep 25 12:25:23 thats why I am asking Sep 25 12:30:40 hi playya Sep 25 12:33:58 woglinde: do you know if the problem is bitbake or oe related ? Sep 25 12:36:53 when you dont show me the problem Sep 25 12:36:59 I cannt decide it Sep 25 12:37:16 strange thing of the day : bitbake linux ... build fine but when it comes at "the following files were installed but not shipped in any package:" it displays only 2 lines / second which makes the packaging stage during very very long Sep 25 12:37:28 this is the problem Sep 25 12:38:16 I was thinking when you said "kergoth told me he fixedmaster" you were talking about this Sep 25 12:39:10 I meant master bitbake Sep 25 12:39:21 but this looks like the normal oe stuff Sep 25 12:39:29 with meta-data Sep 25 12:39:53 yes but logging is really very slow now Sep 25 12:40:14 hm Sep 25 12:40:30 in fact I never noticed it was so slow before Sep 25 12:40:49 kernel is a good test for this as nearly all the files trigger this log Sep 25 12:41:06 gcc-cross Sep 25 12:44:20 ok here are the numbers : doèpackage was 292.78 seconds one week ago, it's now 1241 seconds on the same kernel ! Sep 25 12:45:22 hm Sep 25 12:45:31 ask kergoth about it Sep 25 12:45:31 same build machine Sep 25 12:45:37 I'll do Sep 25 12:45:50 GNUtoo|laptop: do you confirm you have the same thing ? Sep 25 12:48:15 it seems this problem is local as tinderbox shows do_package at ~250-300 s for linux Sep 25 12:49:11 ericben, if you mean the kenrel yes: Sep 25 12:49:19 NOTE: /kernel/arch/arm/mm/tlb-v4.S Sep 25 12:49:20 etc... Sep 25 12:49:45 yes there si the log but how long does your do_package take ? Sep 25 12:49:52 very very long Sep 25 12:49:58 because of this NOTE Sep 25 12:49:59 here 20 minutes ! Sep 25 12:50:02 maybe 15min Sep 25 12:50:03 or more Sep 25 12:50:16 was less than 5 before Sep 25 12:50:24 ok Sep 25 12:50:35 and is less than 5 for others on tinderbox Sep 25 12:50:51 ah ok Sep 25 12:50:52 strange Sep 25 12:50:54 I'm cleaning the hard drive as it was quite full Sep 25 12:50:56 I use bitbake master Sep 25 12:51:19 but I don't think this will change something Sep 25 12:51:23 no Sep 25 12:51:27 it's the NOTE stuff Sep 25 12:51:29 I think Sep 25 12:51:40 do the other have NOTE: Sep 25 12:51:59 the NOTE thing is normal Sep 25 12:52:05 ok Sep 25 12:52:07 hmmm Sep 25 12:52:09 it's in class/packages.bbclass Sep 25 12:52:37 ok Sep 25 12:56:44 basically it should be because theses files are not shipped Sep 25 12:57:17 but I mean , you could supress the warning rm-ing them or something like that Sep 25 12:57:19 right? Sep 25 13:00:21 in fact for some package this NOTE is interesting because you know immediatly if a file you expect to be in the package is missing Sep 25 13:03:15 so maybe we should prevent some files which wont be packaged to fall into the kernel's install Sep 25 13:07:47 indeed Sep 25 13:09:20 seems I have a problem on my machine. File access are ultra slow. Sep 25 13:20:51 do you know why in kernel.bbclass's install we copy arch/${ARCH}/* to ${D}/kernel/arch/${ARCH}/ ? Sep 25 13:34:11 ericben, I've noticed that also Sep 25 13:34:35 Crofton|work: at least .h could be usefull but all the .c .S seems unecessary Sep 25 13:35:02 maybe this was an easy way to handle the fact that include dir were splitted in several arch dir in the kernel ? Sep 25 14:59:02 GNUtoo|laptop: ericben, the note messages are because install apparently installs all the src to imiage which is not a good thing Sep 25 14:59:08 never got to digging into it Sep 25 14:59:35 ok Sep 25 15:02:05 in fact part of the sources are installed in ${D}/kernel Sep 25 15:03:07 and only used in sysroot_stage_all_append Sep 25 15:03:30 I made a quick test putting this kernel dir in ${D}/.. instead of ${D} Sep 25 15:03:44 and do_package is now 35 seconds ! Sep 25 15:49:54 he zecke Sep 25 15:56:03 anyone working on Qt recipes? Sep 25 15:58:00 better aks on the ml Sep 25 15:58:01 *g* Sep 25 15:58:56 cbrake, maybe? Sep 25 16:05:21 was more thinking if Jama is done Sep 25 16:06:51 ah Sep 25 16:11:05 03Holger Hans Peter Freyther  07master * r8c4a51b4d7 10openembedded.git/recipes/qt4/qt4.inc: qt4: Package the debug binaries into tools-dbg Sep 25 16:11:06 03Holger Hans Peter Freyther  07master * rd784cb863c 10openembedded.git/recipes/qt4/qt4.inc: qt4.inc: Package the xmlpatterns utility Sep 25 16:11:08 03Holger Hans Peter Freyther  07master * rfd24b00618 10openembedded.git/: Merge branch 'holger/qt4-4.7' Sep 25 16:11:10 03Holger Hans Peter Freyther  07master * rabfbe217fd 10openembedded.git/recipes/qt4/ (qt4-embedded.inc qt4-x11-free.inc): qt4: Bump the PR after packaging changes Sep 25 16:11:11 03Holger Hans Peter Freyther  07master * rfab7e6681c 10openembedded.git/recipes/qt4/qt4.inc: qt4.inc: Package the bearer plugins. Sep 25 16:11:12 03Holger Hans Peter Freyther  07master * r506a9b6c9f 10openembedded.git/recipes/qt4/qt4.inc: qt4.inc: Package qmlviewer into a new package. Sep 25 16:11:18 merge branch? Sep 25 16:11:22 forgotten rebase? Sep 25 16:13:13 woglinde: no, even passed --no-ff to the merge Sep 25 16:13:30 woglinde: this way I onlt need to bump PR on the last commit and git bisect should/could work. Sep 25 16:17:57 hmm the qml plugins are annoying to pack Sep 25 16:19:07 hm I wonder how pyqt will handle this Sep 25 16:21:48 pyQt and QML? Sep 25 16:22:16 ericben was afk; i was wondering why the kernel soruces are sintalled at all, what does sysroot_stage_all_append with it ? Sep 25 16:24:10 zecke why not? Sep 25 16:27:05 hm, retrieving src_uri from a wiki attachement; does not seem a good plan; will dig into this Sep 25 16:27:22 meanwhile anyone a suggestion on my RRECOMENDS and kernel modules email Sep 25 16:27:30 grabbing some food, biab Sep 25 16:32:06 woglinde: nothing against it, feels a bit odd... why you would want to combine python and JS Sep 25 16:34:17 hm no I meant running qml guis from python Sep 25 16:45:49 khem: yet? Sep 25 16:53:51 ouch Sep 25 16:53:57 I really don't know what I did Sep 25 16:54:05 but bitbake exits with 1 Sep 25 16:54:40 Iah ok Sep 25 16:54:54 maybe my last local modification Sep 25 16:55:41 ok works now Sep 25 16:58:58 hmm... bad patches to cover things up Sep 25 17:09:31 zecke: your Qt changes look useful Sep 25 17:09:59 zecke: hello Sep 25 17:10:13 hi khem and cbrake Sep 25 17:10:18 woglinde: hello Sep 25 17:10:40 I have linaro changes ported to OE and results looks damn good Sep 25 17:10:47 khem which? Sep 25 17:10:48 linaro gcc Sep 25 17:10:51 4.5 Sep 25 17:10:52 aeh? Sep 25 17:10:54 ah Sep 25 17:10:58 hrw made 4.4 Sep 25 17:11:24 woglinde: there are very good patches for armv7 Sep 25 17:11:41 I can see perf improve by atleast 20% on avg Sep 25 17:11:56 gcc test results look pretty good too Sep 25 17:12:24 khem: how is fp support? Sep 25 17:12:33 improved Sep 25 17:12:35 khem intressting Sep 25 17:12:42 but I am looking for some benchmarks Sep 25 17:12:51 do they forces patches upstream? Sep 25 17:12:52 zecke, there was an x that was in machine config: (related to ctrl+x of nano ) Sep 25 17:12:54 to see neon Sep 25 17:13:18 wow 20% perfs Sep 25 17:13:19 should I post these patches to ml or a branch before I commit them Sep 25 17:13:44 is there some interest in reviewing them or do all trust my judgement and testing hehe Sep 25 17:13:46 * GNUtoo|laptop wants to try mplayer on a 20% improved system Sep 25 17:13:56 GNUtoo|laptop: cool Sep 25 17:14:16 I bet I've to rebuild from scratch tough Sep 25 17:14:19 GNUtoo|laptop: I dont have time to do all benchmarking but if you can lend hand that will be fantastic Sep 25 17:14:43 ok but how am I supposed to do benchmarks Sep 25 17:14:46 GNUtoo|laptop: yes most of gcc changes specially to optimizers will require that Sep 25 17:14:59 GNUtoo|laptop: like mplayer you said Sep 25 17:15:03 ok Sep 25 17:15:14 I must find the good command line arguments then Sep 25 17:15:19 GNUtoo|laptop: btw. is your system armv7 Sep 25 17:15:35 nokia900 Sep 25 17:15:42 GNUtoo|laptop: may be you can dump frame/second info Sep 25 17:15:48 GNUtoo|laptop: perfect Sep 25 17:16:00 I've also a nexusone but it has no alsa Sep 25 17:16:07 ok Sep 25 17:16:09 GNUtoo how do you will recompile all n900 stuff? Sep 25 17:16:40 woglinde, ? Sep 25 17:16:54 I won't do it now tough Sep 25 17:17:05 I need to fix bootability of n900 first Sep 25 17:17:30 I am pretty confident that it will improve people's life :) Sep 25 17:18:10 gcc or the bootability? Sep 25 17:18:16 gcc Sep 25 17:18:22 basically my n900 is a bit slow with mplayer Sep 25 17:18:34 GNUtoo|laptop: are you using gcc 4.5 Sep 25 17:18:40 GNUtoo|laptop: if not try that Sep 25 17:18:43 I mean it can play some video like big buck bunny 480p,but not sintel 480p Sep 25 17:18:54 so I've 3 solutions: Sep 25 17:18:59 s/3/some Sep 25 17:19:07 *use a better gcc Sep 25 17:19:14 *try to use the DSPs Sep 25 17:19:29 *add alsa to nexusone and watch videos on nexusone(very bad idea) Sep 25 17:19:38 use the gstreamer dsp plugin Sep 25 17:19:44 ok Sep 25 17:19:46 but I wonder Sep 25 17:19:49 solution 1 is cheap Sep 25 17:19:49 what's faster Sep 25 17:19:58 between CPU and DSP Sep 25 17:20:01 you can still have other two on top Sep 25 17:20:01 indeed Sep 25 17:20:07 you dont loose anything with 1 Sep 25 17:20:14 with 3 I loose my mind Sep 25 17:20:40 with 2 it's ok as long as it's fast, which I doubt Sep 25 17:20:46 so let's go for 1 Sep 25 17:21:21 there is also the convert the video in non-x264 thing Sep 25 17:21:38 JaMa, what do you think? Sep 25 17:28:51 woglinde, for the dsp I eared there was a simplier gst plugin : gst-dsp or something like that Sep 25 17:30:17 http://github.com/felipec/gst-dsp Sep 25 17:30:21 what do you think of it? Sep 25 17:30:43 how do I use it Sep 25 17:30:48 I've the firmware from maemo Sep 25 17:30:55 I load the firmware Sep 25 17:30:59 I compile gst-dsp Sep 25 17:31:05 I add it to gstreamer path Sep 25 17:31:08 and then? Sep 25 17:34:44 03Khem Raj  07master * r190f46585f 10openembedded.git/recipes/quilt/ (quilt-0.48/no_rm_symlinks.patch quilt.inc): Sep 25 17:34:44 quilt-0.48: Check if file is a symlink before deleting. Sep 25 17:34:44 Signed-off-by: Khem Raj Sep 25 17:37:35 woglinde, hi do you know http://github.com/felipec/gst-dsp Sep 25 17:37:36 ? Sep 25 17:37:51 alright pushed the branch to repo Sep 25 17:38:23 khem/import-linaro-gcc-4.5 Sep 25 17:38:25 gnutto sure Sep 25 17:38:32 ask felpic on #beagle Sep 25 17:38:47 I would appreacite if people give it a try Sep 25 17:40:50 cbrake: hey, I will have to update the qmake specs right now, the 8 patches apply cover up and error with rpath-link Sep 25 17:41:15 khem: do you have experience with bfd? Sep 25 17:41:21 03Frans Meulenbroeks  07org.openembedded.dev * ra58ba9c0e4 10openembedded.git/recipes/lesstif/files/ (4 files): Sep 25 17:41:22 lesstif : moved unused files to obsolete dir Sep 25 17:41:22 Signed-off-by: Frans Meulenbroeks Sep 25 17:41:22 03Frans Meulenbroeks  07org.openembedded.dev * r84b4ba3060 10openembedded.git/recipes/lemon/files/snprintf.patch: Sep 25 17:41:22 lemon : moved unused files to obsolete dir Sep 25 17:41:22 Signed-off-by: Frans Meulenbroeks Sep 25 17:41:23 03Frans Meulenbroeks  07org.openembedded.dev * rc49791f38a 10openembedded.git/recipes/ltrace/files/no-usr-include.patch: Sep 25 17:41:23 ltrace : moved unused files to obsolete dir Sep 25 17:41:24 Signed-off-by: Frans Meulenbroeks Sep 25 17:41:24 03Frans Meulenbroeks  07org.openembedded.dev * r5a40c12ed4 10openembedded.git/recipes/libxslt/files/destdir.patch: Sep 25 17:41:25 libxslt : moved unused files to obsolete dir Sep 25 17:41:26 Signed-off-by: Frans Meulenbroeks Sep 25 17:41:26 03Frans Meulenbroeks  07org.openembedded.dev * rd607d8aa18 10openembedded.git/recipes/libqpe/libqpe-opie/save-windows-pos-dynamic.patch: Sep 25 17:41:27 libqpe : moved unused files to obsolete dir Sep 25 17:41:27 Signed-off-by: Frans Meulenbroeks Sep 25 17:41:28 03Frans Meulenbroeks  07org.openembedded.dev * r86369d52f1 10openembedded.git/recipes/libmimedir/files/ (4 files): Sep 25 17:47:55 khem: gcc-linaro, did they already add some features? Sep 25 17:48:13 zecke: lot of code improvements Sep 25 17:48:17 for armv7 Sep 25 17:48:20 mainly from csl Sep 25 17:49:45 zecke: do we have some dev space on openembedded.org Sep 25 17:49:58 like openembedded.org/~khem Sep 25 17:50:48 zecke: and I would like to forward my openembedded.org mails to another account can you help Sep 25 17:56:09 03Frans Meulenbroeks  07org.openembedded.dev * r2d97a1b871 10openembedded.git/recipes/perl/libintl-perl_1.11.bb: Sep 25 17:56:09 libintl-perl_1.11.bb: removed Sep 25 17:56:09 this one does not fetch any more (it is not available on cpan any more) Sep 25 17:56:09 also we still have 1.16 and 1.20 Sep 25 17:56:09 Signed-off-by: Frans Meulenbroeks Sep 25 17:56:49 quite some recipes do not seem to fetch any more Sep 25 17:58:01 khem: we have dev-space, i don't think I can give an account though. I have mine on oe.linuxtogo.org/~zecke Sep 25 17:58:02 eFfeM: sometimes sources move Sep 25 17:58:10 find new source and update them Sep 25 17:58:17 dont just move them to obsolete Sep 25 17:58:21 khem: for the email, I am sorry, I don't know anything about our mailsetup Sep 25 17:58:37 zecke: its fine, I have uclibc.org Sep 25 17:58:37 khem: (linaro) I went through the launchpad, do you know if they have anything from other than CSL? Sep 25 17:58:50 no I dont think so Sep 25 17:58:57 few backports Sep 25 17:58:58 khem, for libintl it is kinda useless to keep the old 1.11 version Sep 25 17:59:04 but was not planning to remove the others Sep 25 17:59:11 eFfeM: no it can be removed Sep 25 17:59:54 what is the problem removing libintl-perl_1.11 ? Sep 25 18:00:35 oops misread thought you wrote can not Sep 25 18:01:00 but will ask for others, most often they are older recipes where it is kinda pointless to update Sep 25 18:01:20 e.g. i don't want to invest time in finding a new source for linux-uml 2.4 Sep 25 18:01:44 eFfeM: you can leave it alone Sep 25 18:01:51 or ask people on ml Sep 25 18:02:03 which u did rightly Sep 25 18:02:35 i will (but as I do a lot of perl libs i found the libintl one being within my maintainer responsibility) Sep 25 18:02:50 issue is that ml often does not give any feedback Sep 25 18:03:12 will post the list once it is done Sep 25 18:03:21 coffee break Sep 25 18:03:38 still post to ml Sep 25 18:03:50 and you will have record Sep 25 18:04:09 and in computer science assumptions are often wrong :) Sep 25 18:06:28 khem: the g++ failures are due not finding libstdc++? Sep 25 18:06:35 yes :) Sep 25 18:07:04 that was because I did not populate_sysroot before running the test for g++ Sep 25 18:07:09 its rerunning as we speak Sep 25 18:07:12 :) Sep 25 18:07:28 khem: (binutils) I just ran into a bfd assertion in a very old binutils (angstrom's). How do you normally find patches for that? just use git log on the directory and search for patches in that area? Sep 25 18:07:39 khem: ah, do you use dejagnu with a remote setup? Sep 25 18:07:50 yes Sep 25 18:21:15 hi all Sep 25 18:22:31 anyone here worked with the sgx drivers from TI before? Sep 25 18:23:02 I see recipes that deal with 3_01_00_07 but the one I could get downloaded from the TI site is already 4_00_00_01 Sep 25 18:29:19 stefan_schmidt, try #beagle Sep 25 18:29:32 Crofton|work: ok :) Sep 25 18:38:58 hi stefan Sep 25 18:39:06 hi woglinde Sep 25 18:43:10 Crofton|work: I have posted gcc 4.5 improvements branch try it out if you will Sep 25 18:44:52 ok Sep 25 18:45:02 what will they "improve" Sep 25 18:45:20 I'm in the middle of something else, but once we decide we have good hw, I want to test them Sep 25 18:45:51 perf Sep 25 18:45:53 on armv7 Sep 25 18:45:59 significantly Sep 25 18:46:00 cool Sep 25 18:46:08 and also neon code gen hopefully Sep 25 18:47:16 this is very interesting Sep 25 18:47:22 for the linaro toolchain, this site gives details on tests & benchmarks once they fix the web site : http://ex.seabright.co.nz/helpers/buildlog Sep 25 18:48:10 khem: The world build stopped at 54964 of 63626 Sep 25 18:48:21 stefan_schmidt: hmmm Sep 25 18:48:27 what happens if you restart Sep 25 18:48:29 khem: I started over soem hours ago, this time with tinderbox and a log file Sep 25 18:48:47 eFfeM: sorry was also afk (on the roof :-) I think ionstalling arch/${ARCH} in the kernel is done in order to have all the includes which are here Sep 25 18:48:49 khem: it was able to get up some more recipes when I restarted, but not much Sep 25 18:49:11 stefan_schmidt: ok Sep 25 18:49:17 eFfeM: this may allow to build out of tree modules Sep 25 18:49:20 stefan_schmidt: so now we need your log Sep 25 18:49:24 so we can see what fails Sep 25 18:49:28 and try to fix those Sep 25 18:49:51 khem: new build is at 9000 now Sep 25 18:49:58 khem: will upload the log later Sep 25 18:50:10 stefan_schmidt: ok Sep 25 18:50:18 how long does it take on your machine Sep 25 18:50:31 my laptop it takes like 3 days Sep 25 18:51:13 I started the build yesterday and it was finished today. I missed the "time§ command. But this time I added it as well :) Sep 25 18:51:29 I would think between 1 and two days Sep 25 18:51:42 Sounds like a good weekend job for the machine :) Sep 25 18:52:47 khem: if we are getting world fixed we could think about adding it to the weekly testing builds Sep 25 18:52:58 ericben: hm, I'd say if the files are needed for out-of-tree modules, they should go to staging Sep 25 18:53:11 eFfeM: that's what is donea ctually Sep 25 18:53:51 but if they are staged i feel oe should not complain about them not being packaged Sep 25 18:54:18 stefan_schmidt: -k will help you to get a list of failing recipes Sep 25 18:54:23 in fact they are first put in ${D}/kernel by install, then staged by sysroot_stage_all_append Sep 25 18:54:48 but the remain in ${D}/kernel which makes packages complaining Sep 25 18:54:50 eFfeM: yup, doing that already: time bitbake -k world | tee world.log Sep 25 18:56:21 I simply copied in ${D}/../kernel and no more complain : is this an acceptable solution ? of maybe in fact we should not copy the files in ${D}/kernel in install stage but directly copy them from ${S} to stagging in sysroot_stage_all_append Sep 25 18:56:38 this way, there won't be the files in ${D} and package won't complain Sep 25 18:57:09 do you think this is a ringht fix ? Sep 25 18:57:13 ericben: i've not really an idea on this Sep 25 18:58:21 once my server gives back the hand (I had many tmp and the rm last since 4 hours ...) I'll test this fix and submit a patch to the ML Sep 25 19:08:09 ericben: install should look like sysroot Sep 25 19:08:19 so make it to install is better Sep 25 19:08:29 and then sysroot will take care of staging it Sep 25 19:08:43 btw : does anyone here have a good feedback of the better filesystem to use for a OE compile server ? Seems having the right FS for small sources & object file should improve the build times. Sep 25 19:08:51 khem: ok Sep 25 19:10:01 khem: but is it fine to install something at install stage not in ${D} ? Sep 25 19:10:33 btrfs Sep 25 19:12:24 khem: ok will try that next time. I though it was not yet ready for common usage Sep 25 19:12:37 it is not Sep 25 19:12:45 ok :) Sep 25 19:13:11 you can try reiser4 Sep 25 19:13:49 not sure it's worth spending time to bench this. Sep 25 19:14:01 yes Sep 25 19:14:32 ext4 is good enough I think Sep 25 19:14:46 for kernel install/staging : do you think copying what has to be stagged outside ${D} is acceptable ? Sep 25 19:16:00 hmmm prolly not a good idea Sep 25 19:16:03 why it that needed Sep 25 19:16:45 actual problem is that many sources & .h files end in $[D} and are not packaged leading to package's class colmplaining (through bb.note) and this brings 4 minutes of log ! Sep 25 19:17:04 only to say these files are not in a package Sep 25 19:17:22 because they are onyl used in sysroot_stage_all_append which will copy them into ${SYSROOT_DESTDIR}${STAGING_KERNEL_DIR} Sep 25 19:17:49 so my idea was to not install the files in ${D} but directly copy them from ${S} to ${SYSROOT_DESTDIR}${STAGING_KERNEL_DIR} Sep 25 19:18:02 so that package class won't complain Sep 25 19:18:09 all af this in is kernel.bbclass Sep 25 19:18:16 then delete them Sep 25 19:18:20 and also this makes only one copy instead of 2 Sep 25 19:18:40 oh i see Sep 25 19:18:46 they are needed in staging Sep 25 19:18:53 we can't deltet them else they won't be staged Sep 25 19:19:00 then they should be packaged into some -dev Sep 25 19:19:07 yes I think they are useful for out of tree module compilation Sep 25 19:19:15 ok kernel-dev Sep 25 19:19:21 is right place for them Sep 25 19:19:29 ok that could be a good fix Sep 25 19:20:12 then I'll go this way once I can test again Sep 25 19:20:25 ok Sep 25 19:20:28 good Sep 25 19:22:33 bye Sep 25 19:22:40 bye Sep 25 19:41:54 cbrake: I have an answer to your QtE build issue, qmake looks for rpath-link in a new variable Sep 25 20:16:04 zecke: great Sep 25 20:19:57 cbrake: I will need to build over night, the g++.conf needs to say QMAKE_LFGLAGS_RPATH instead of QMAKE_RPATH. Sep 25 20:20:21 cbrake: I will push it tomorrow morning, and then I need to package the qml plugins Sep 25 20:21:25 cbrake: for tomorrow I try to flash the CM-270 and then figure out if there is any benefit in using the iwmmx code that Qt has Sep 25 20:22:19 hm pyqt gives me soem headaches Sep 25 20:22:30 cbrake: holger/qt4-4.7 holds my current changes Sep 25 20:22:33 woglinde: what kind of? Sep 25 20:23:29 zecke determinig qt stuff for crosscompiling Sep 25 20:32:27 woglinde: like? Sep 25 20:33:18 hm rather it is to understand how mickeyl did compile the whole stuff Sep 25 20:39:55 hm why the heck he is patching configure.py and than dont use it Sep 25 20:39:56 hmm.. whats my next step to debug i2cdetect -yr 3 being empty Sep 25 20:40:14 should i get a usbee sx logic analyzer? Sep 25 20:43:47 woglinde: are you sure it is not called? Sep 25 20:44:06 will see later Sep 25 20:44:18 for what I now understand Sep 25 20:44:23 sip is called amnually Sep 25 20:44:43 rather than from configure.py Sep 25 20:45:20 next try with sip given the headerdir to the pyqt-sip-header Sep 25 20:47:00 khem: did you play with gcc plugins yet? Sep 25 20:47:14 only a bit Sep 25 20:47:15 yes Sep 25 20:47:23 wrote one printf parser Sep 25 20:47:53 but gplv3 is still a no no Sep 25 20:47:58 so had to abandon Sep 25 20:48:09 khem: for your work? Sep 25 20:48:13 es Sep 25 20:48:15 yes Sep 25 20:53:43 hm so now do_compile goes further Sep 25 21:07:14 good luck, I will sleep now Sep 25 21:07:26 gute nacht Sep 25 22:06:06 im probably just gonna kidnap an embedded device drivers consultant at gunpoint Sep 25 22:06:18 rofl Sep 25 22:06:28 in interest of expediency Sep 25 22:06:45 sometimes it is worth trading money for time Sep 25 22:07:13 yeah ive been attempting that Sep 25 22:16:09 these guys are in way too high of demand :) Sep 25 22:17:36 Didnt alan cox say device drivers were simple in linux? Sep 25 22:22:15 they are :) Sep 25 23:15:20 i2cdetect gives me an error because i have no I2C_FUNC_SMBUS_QUICK Sep 25 23:15:35 i2cdetect -r runs but is all blank Sep 25 23:16:38 i have 1 working i2c device, twl, sound chip, .. i tried sysfs/i2c/delete_device on it, to confirm i2cdetect -r would give nothing, but i couldnt actually delete it for some reason Sep 25 23:23:21 hmm heres real old patch http://linux.omap.com/pipermail/linux-omap-open-source/2006-October/008232.html Sep 25 23:23:32 but seems code is still same Sep 25 23:23:50 least for the files patched.. Sep 26 00:15:12 i dont quite get how icecc manage to do the cross compile thing distributed Sep 26 00:15:33 anybody knows more about this topic or knows a nice url? Sep 26 00:16:13 (if anybody uses distcc, id also like to know ;)) Sep 26 00:21:00 kusanagi: http://wiki.maemo.org/Icecream_Scratchbox_Howto Sep 26 00:25:18 oh, so u make a compiler eviroment and send them to the nodes so they are able to cross-compile Sep 26 00:25:38 am i getting it right? Sep 26 00:28:47 yes Sep 26 00:29:29 ty khem Sep 26 00:29:32 :D Sep 26 00:30:50 look into classes/icecc.bbclass too Sep 26 00:30:56 code is best documentation Sep 26 01:01:03 Hello, The OE opkg is same as openwrt, i mean is based on similars Makefiles? Sep 26 01:13:41 kristianpaul: ? Sep 26 01:13:44 opkg is opkg. google it Sep 26 01:13:54 oe and openwrt aren't the only projects that use/build it Sep 26 01:15:49 kergoth_: oh sorry Sep 26 01:16:04 http://code.google.com/p/opkg/ Sep 26 01:16:31 nice wikipedia save look in g00gl3 :) Sep 26 01:16:35 save me* Sep 26 01:17:37 20:14 < kergoth_> oe and openwrt aren't the only projects that use/build it Sep 26 01:17:47 who else uses? Sep 26 01:18:00 which projest* Sep 26 01:25:07 kristianpaul:http://www.opkg.org/ Sep 26 01:25:28 or http://tinyurl.com/2an7auf :) Sep 26 01:26:01 wow Sep 26 01:26:10 yeah tangogps ! Sep 26 01:27:00 ahh you hide it with tinyurl.! Sep 26 02:08:43 03Chris Larson  07master * r0754ca0195 10openembedded.git/classes/utility-tasks.bbclass: (log message trimmed) Sep 26 02:08:43 utility-tasks.bbclass: drop mrproper task Sep 26 02:08:43 This task is misleading and potentially quite harmful, as it wipes the entire Sep 26 02:08:43 DL_DIR, yet tasks are supposed to be recipe specific. We have the proper Sep 26 02:08:43 distclean task to wipe sources for a recipe + run do_clean, and the user can Sep 26 02:08:44 always wipe all of DL_DIR manually if necessary. Sep 26 02:08:45 Signed-off-by: Chris Larson Sep 26 02:24:10 kergoth_: what do u think of my patch for removing SCM dirs Sep 26 02:24:51 I thought of using export functionality of OS but it would slow down the process Sep 26 02:25:02 as export is order of checkout or fetch Sep 26 02:30:09 yeah, good point. seems sane enough, will have to test it out Sep 26 02:43:31 kergoth_: I cleaned my dldir Sep 26 02:43:36 and did a x11-image Sep 26 02:43:39 worked well Sep 26 02:43:42 booted well Sep 26 02:43:59 and now bitbake world is running **** ENDING LOGGING AT Sun Sep 26 02:59:57 2010