**** BEGIN LOGGING AT Mon Aug 08 02:59:58 2011 Aug 08 05:04:15 urh! eglibc.org is down :( Aug 08 06:27:17 does anybody know if there's some eglibc svn mirror? Aug 08 07:44:43 hello, anyone disabling CCACHE during build? How do yo do it w/out editing bitbake.conf? Aug 08 09:12:48 hello Aug 08 09:13:30 I had buided a x11-image for mini6410 .... but when I try to play a wmv file I m getting following messages.. Aug 08 09:16:06 http://pastebin.mozilla.org/1292924 Aug 08 09:16:25 can anyone help whats going wrong.... Aug 08 09:16:37 also video is playing too slow... Aug 08 10:25:32 morning Aug 08 10:57:05 hi all Aug 08 10:58:09 i have a minor question ... today I noticed that we done have native recipe for subversion and svn is used in some recipes .... is there any specific reason that we did not have native recipe of subversion Aug 08 11:00:21 or should I go ahead and submit subversion's native version? Aug 08 11:05:47 done -> don't Aug 08 11:47:17 Noor, do they use bblcassextend? Aug 08 11:48:36 Crofton|work: Nopse Aug 08 11:50:03 could be because subversion is a PITA to compile Aug 08 11:59:26 XorA: hello Aug 08 12:32:13 hey AnssiN_ Aug 08 12:32:16 hay ant_work Aug 08 18:19:43 Hey guys, I'm having some trouble setting up a cross-compilation toolchain Aug 08 18:20:11 when building binutils, I get "WARNING: `makeinfo' is missing on your system. Aug 08 18:20:11 ", except I already have the texinfo package Aug 08 18:20:17 I'm on Ubuntu 10.04 Aug 08 18:20:27 Any suggestions? Aug 08 18:26:39 check temp/log and temp/run files in workdir to see where its getting makeinfo from **** BEGIN LOGGING AT Mon Aug 08 18:30:38 2011 Aug 08 18:30:58 GeorgeH: are you getting that warning when compiling binutils Aug 08 18:31:11 yup Aug 08 18:31:16 and will do, thanks joelagnel Aug 08 18:32:33 GeorgeH: which version of binutils are you compiling ? there is a patch already applied to newer binutils to fix it Aug 08 18:34:59 I pulled this from sourceware's CVS earlier today Aug 08 18:37:02 last version given in NEWS (binutils/binutils) is 2.21 Aug 08 18:39:48 ok then it should have the bug fix already Aug 08 18:40:01 so it really is missing on your build system Aug 08 18:44:23 okay, so inside binutils/bfd/config.log, I see MAKEINFO='/usr/arm/src/binutils/missing makeinfo --split-size=5000000' Aug 08 18:47:26 GeorgeH: hmm /usr/arm/src/ ? Aug 08 18:47:29 whats that Aug 08 18:47:47 get rid of that Aug 08 18:48:25 /usr/arm/src/ is where I have binutils... is that horribly wrong? Aug 08 18:51:26 GeorgeH: are you using oe ? Aug 08 18:51:35 os something else Aug 08 18:52:12 GeorgeH: and you should have texinfo installed Aug 08 18:53:19 I do, and yes, but don't I need to get binutils compiled first? Aug 08 18:53:50 i appear to have makeinfo installed at /usr/bin/makeinfo Aug 08 18:57:22 well, I think I solved that problem. Edited the binutils makefile and set MAKEINFO to '/usr/bin/makeinfo Aug 08 18:57:22 '... solved that error Aug 08 18:57:34 RP__: around? Aug 08 18:57:40 but before I go any further, am I doing this completely wrong? Aug 08 18:58:11 Sorry for being pretty inept, I'm not very familiar with linux toolchains yet Aug 08 19:02:49 hello, i'm following the instructions on http://gumstix.org/create-a-bootable-microsd-card.html but i can't find any file called MLO. does anyone know where this would be? Aug 08 19:03:48 do i need to run bitbake x-load in order to see it? Aug 08 19:11:54 How does one report a bug? Aug 08 19:13:21 Do I report it to openembedded-devel@lists.openembedded.org or openembedded-issues@lists.openembedded.org? Aug 08 19:14:30 hlzr: openembedded-devel Aug 08 19:14:56 khem: thank you Aug 08 19:15:28 and if you explain your problem well then chances are it will get good answers Aug 08 19:19:07 My issue is, when I try to build an image, GNU MPFR breaks when it tries to apply a patch. The output of the failed build is at . Can anyone assist me in fixing this bug? Aug 08 19:23:14 hlzr: your checkout appears corrup Aug 08 19:23:16 or at least that patch Aug 08 19:23:59 galak: hi Aug 08 19:24:54 Tartarus: how can I check if that is the case? Aug 08 19:33:45 RP__: hi, wanted to see if we can figure out how to deal with this --with-cpu for glibc Aug 08 19:35:17 galak: set PACKAGE_ARCH = TUNE_ARCH conditionally for eglibc Aug 08 19:35:44 galak: Can you clearly explain what the problem is? Aug 08 19:36:02 galak: since with this option eglibc becomes machine/sub-arch specific Aug 08 19:36:04 galak: I'm obviously misunderstanding whats needed somewhere :/ Aug 08 19:36:22 galak: Under what conditions should that option be passed in? Aug 08 19:36:47 RP__: I think that option is always to be used for ppce5500 e.g. Aug 08 19:37:20 RP__: in general eglibc behaves differently when --with-cpu is specified Aug 08 19:37:35 so may be it should be in the recipe to detect with-cpu and make it machine specific Aug 08 19:37:47 or tune files can do it too Aug 08 19:37:55 with pn-eglibc override Aug 08 19:39:42 * RP__ suspects the answer is to pass it in if TUNE_FEATURES has ppce5500 in it Aug 08 19:40:00 if that is the case, it will be packaged as tune specific anyway Aug 08 19:40:01 RP__: for ppc we need to specify --with-cpu to eglibc to get proper builds cfg for the platform Aug 08 19:40:23 galak: Is a built without that going to work on platform? Aug 08 19:40:46 no correctly, it would work for 99% of things in some cases Aug 08 19:40:57 for example mathlib sqrt function will be broken Aug 08 19:41:11 RP__: glibc becomes cpu specific in short Aug 08 19:41:14 galak: So the default tune configs in the default arch file won't work? Aug 08 19:41:27 RP__: correct Aug 08 19:41:36 for ppc it should since we use tune-arch in triplet Aug 08 19:41:39 RP__: you have to utilize a tune-ppcXXXX Aug 08 19:41:46 galak: ok, so the file should do AVAILTUNES = "ppce5500", not with += Aug 08 19:41:48 ppc603e-angstrom-linux Aug 08 19:41:57 khem, I started from scratch - I think I have everything setup correctly now, but I'm not entirely sure what to set BBFILES to inside local.conf. Should it point to the distro or machine folders inside my openembedded folder, or should it point to an empty folder that I will populate as I go on? Aug 08 19:42:13 galak: this means that the available tune list is resticted only to things that work Aug 08 19:42:29 then you can do the GLIBC_EXTRA_OECONF unconditionally Aug 08 19:42:58 oh, shoot, openembedded/recipes. I apologize for the stupid question Aug 08 19:44:15 RP__: I think I still need GLIBC_EXTRA_OECONF to be something like: Aug 08 19:44:18 GLIBC_EXTRA_OECONF += "${@bb.utils.contains("TUNE_FEATURES", "ppc603e", "--with-cpu=603e", "", d)}" Aug 08 19:44:25 because of nativesdk builds Aug 08 19:44:43 galak: hmm, yes, right Aug 08 19:44:52 galak: the above will certainly work and be fine Aug 08 19:45:04 galak: We need to get nativesdk using the tune system really Aug 08 19:45:12 but one step at a time Aug 08 19:45:32 * RP__ doesn't fancy including ia32 arch in a ppc built Aug 08 19:45:36 build Aug 08 19:47:38 RP__: ok, I'll work that up for the tune files that need it Aug 08 19:47:56 RP__: did my query about 32-bit vs 64-bit tune file make sense? Aug 08 19:48:15 galak: Can you also check whether all the AVAILTUNE values make sense on a given machine? Aug 08 19:48:34 meaning? Aug 08 19:54:13 galak: meaning as I mentioned above - that the tunes listed there should be ones that will work on a given platform Aug 08 19:54:41 it sounds like tune-powerpc and tune-powerpc-nf won't work on the 5500 for example Aug 08 19:54:44 ok, will take a look Aug 08 19:55:08 RP__: so the following still has issues with nativesdk: Aug 08 19:55:09 GLIBC_EXTRA_OECONF += "${@bb.utils.contains("TUNE_FEATURES", "ppc603e", "--with-cpu=603e", "", d)}" Aug 08 19:58:44 thoughts on why? Aug 08 19:59:01 galak: nativesdk doesn't select a different tune :/ Aug 08 19:59:19 galak: It should probably reset GLIBC_EXTRA_OECONF itself Aug 08 19:59:25 how do I distinguish things for nativesdk vs not? Aug 08 20:00:59 brb Aug 08 20:05:25 RP__: are you suggesting the following: Aug 08 20:05:33 GLIBC_EXTRA_OECONF_virtclass-nativesdk = "" Aug 08 20:05:38 in meta/recipes-core/eglibc/eglibc.inc Aug 08 20:06:35 galak: or GLIBC_EXTRA_OECONF = "" in nativesdk.bbclass Aug 08 20:08:33 should nativesdk.bbclass just set TUNE_FEATURES = ""? Aug 08 20:31:18 galak: I have a suspicion that would cause other issues Aug 08 20:39:47 RP__: ok Aug 09 00:41:32 hm, is there a way to make bitbake a little more verbose about which specific recipe it's executing? I'm trying to find out which damn recipe is trying to clone from git with the wrong credentials Aug 09 00:45:24 tzanger - try -D, but pipe output to a file, it gets quite verbose Aug 09 00:45:32 Jefro: perfect, danke Aug 09 00:47:40 well I guess I can't say thanks yet, but I'm trying it :-) Aug 09 00:53:26 :) good luck with it - sometimes those things are hard to track down Aug 09 00:54:44 yeah Aug 09 00:54:46 I see the recipe now Aug 09 00:55:14 but I'm curious if there's a better way... when using svn+ssh to grab code, is there already an environment variable that svn or ssh listens to to get the username from? Aug 09 00:55:48 it's defaulting to my login name, which is not a bad default, but I am wondering if there's already a way to specify it without modifying the command line Aug 09 00:56:26 you can set the username for an ssh login in .ssh/config Aug 09 01:00:13 true, but that affects all ssh logins everywhere Aug 09 01:00:48 no, you can set the login for a specific server Aug 09 01:01:40 oh? Aug 09 01:01:46 I'm gonna read me some man pages, that sounds like the solution Aug 09 01:02:56 Forgive my lack of experience with OE, but is the deploy/ipk directories in tmp, suitable to be used directly to host opkg feeds? Aug 09 01:03:46 tzanger: Aug 09 01:03:48 Host foofoofoo Aug 09 01:03:50 Hostname foo.bar.com Aug 09 01:03:52 User FooBar Aug 09 01:06:23 incandescant: danke **** ENDING LOGGING AT Tue Aug 09 02:59:56 2011