**** BEGIN LOGGING AT Thu Mar 14 02:59:58 2013 Mar 14 04:50:51 hi Mar 14 04:51:17 is it necessory to build native while building glib2 Mar 14 04:51:44 i.e., include BBCLASSEXTEND=native in bb file Mar 14 04:52:28 i'm getting do_compile error while building glib2 for arm Mar 14 04:52:40 gconvert.c:62:2: error: #error GNU libiconv in use but included iconv.h not from libiconv Mar 14 04:55:41 i think it's either glibc or a separate recipe Mar 14 04:56:05 make sure the recipe isn't pulled in Mar 14 05:13:11 well, both versions got pulled in Mar 14 05:13:48 you'll need to clean one or both Mar 14 05:14:08 then add something like PREFERRED_PROVIDER_virtual/libiconv = "libiconv" Mar 14 05:14:17 depending on which one you want Mar 14 05:17:22 somewhere in your recipes you have depends on one or more Mar 14 06:01:41 but eglibc is providing libiconv Mar 14 06:02:03 i want it to be used from eglibc Mar 14 06:02:21 so set it equal to that Mar 14 06:02:42 are you using samba? Mar 14 06:10:26 Good Moring Mar 14 07:11:01 i gave in extra_oeconf as --with-libtool=native but i'm buiding for arm, will it have any effect? Mar 14 07:11:26 sorry, --with-libiconv=native Mar 14 07:11:59 there were three options, no/gnu/native Mar 14 07:12:29 when i give gnu its throwing the same error Mar 14 07:12:50 i need iconv lib so i cant give no Mar 14 07:14:25 hi Mar 14 07:14:50 how to use if condition in .bb files Mar 14 07:14:58 ?? Mar 14 07:15:28 read the documentation for bitbake and oe? Mar 14 07:16:04 use case its working, i tried if in do configure but it was not working Mar 14 07:17:11 i didnt find if in documentation? Mar 14 07:17:35 can i use case outside Mar 14 07:17:47 i mean not in any function? Mar 14 07:23:30 @dm8tbr can provide me link to the documentation for bitbake and oe? Mar 14 07:23:49 are you google-impaired? Mar 14 07:27:43 Googling bitbake related stuff is much more difficult than many other things. I don't know why. Mar 14 07:28:33 Google for openembedded documentation and you get link to wiki page that says "under construction" and some links to source on github. Mar 14 07:30:30 Compare to "vaadin documentation" which gives book and javadoc or to "android documentation". Mar 14 07:31:48 http://www.openembedded.org/wiki/Documentation looks pretty useful to me and it was the first hit on google... Mar 14 07:57:45 Anyone working with the meta-qt5 layer on a 32-bit host? I get an "out of memory" from cc1plus that I can't explain. Mar 14 07:58:03 (and want to get someone else to build) Mar 14 07:59:03 and another Q: if I run 64-bit and build denzil/danny, can I compile SDKs for 32-bit machines? Mar 14 08:52:41 tasslehoff: that should be easy if you've install the multilib support Mar 14 09:06:56 nerdboy: great. I just remember there being some bugs there in the (oe-classic) past. Mar 14 09:26:28 hi all Mar 14 09:27:12 anybody has gone through insane.bb class??? in openembedded-core/meta/classes Mar 14 09:27:33 i need some clarification in function do_package_qa() Mar 14 09:28:24 i'm getting an error as "unsafe-references-in-binaries" Mar 14 09:29:32 should i set any flags to not to get this error?? Mar 14 09:45:12 good morning Mar 14 09:56:34 Hello guys Mar 14 09:57:26 when we set INCOMAPATIABLE_LICESNE .... does that mean those recipes which have that license won't build and other version of the recipe will be picked? Mar 14 12:23:40 Is this for mobile only? Mar 14 12:27:22 donnie what you mean? Mar 14 12:27:53 Well the site says it's like the ultimate linux and blah blah blah... I dig WebOS. if I can get it on PC. I'd be happy Mar 14 12:28:39 ? Mar 14 12:30:06 openembedded is useful for customize distributions/images which do not have diskmemory Mar 14 12:33:19 does anyone know which gst plugins through oe can give me aac support? it is not blindly obvious Mar 14 12:35:27 ant_work: pong Mar 14 12:42:10 hi ensc|w_ Mar 14 12:42:41 ant_work: hi, you pinged my yesterday Mar 14 12:42:47 I am looking for a solution wrt mlibc missing vfscanf() and basename() Mar 14 12:42:51 *klibc Mar 14 12:42:59 I see dietlibc has both Mar 14 12:45:05 a basename implementation should not be a problem... Mar 14 12:45:31 yes, I found a debian (refused) patch for that Mar 14 12:45:49 the point is vfscanf ..I don't think I need float in that code Mar 14 12:45:51 http://tinyurl.com/cvuxmjl Mar 14 12:45:57 (kexec-tools) Mar 14 12:46:34 are other *scanf() functions existing in klibc? Mar 14 12:46:41 yes Mar 14 12:48:53 and there is i.e. vfprintf Mar 14 12:49:46 I don't know if it's worth adding both to klibc or try torefactor that x86-only patch Mar 14 12:54:57 ant_work: vsscanf() in klibc is not written in a way which makes it easy to read from FILE *; dietlibc uses a generic implementation which allows custom getc/ungetc params Mar 14 12:56:46 you can copy and modify vsscanf() (-> lots of non trivial redundant code) or try to refactor it to allow getc/ungetc Mar 14 12:58:46 but for your case, I would avoid FILE* and read/mmap the edd file and give the buffer to vsscanf Mar 14 12:59:45 hm Mar 14 12:59:50 I see Mar 14 13:02:10 in other project compiling fine with klibc we did just FILE *f = fopen(...) and then while (fgets(line, .....) Mar 14 13:04:51 * ensc|w_ usually avoids FILE and uses fds directly Mar 14 15:32:33 how do I get mips-octeon-linux-gcc ? Mar 14 15:35:18 select the right machine? Mar 14 15:41:03 I'm reading on the wiki, but i'm lost. Mar 14 15:46:56 supaplex I don't believe the octeon support has been publically posted.. Mar 14 15:47:32 company I work for (Wind River) has a version, I wouldn't be surprised in Cavium/MontaVista have one as well.. Mar 14 15:48:01 I havn't gotten mail from the oe-core list in a while now.. does anyone know if the list is experiencing problems again? Mar 14 15:48:12 oh lovely... thanks for the heads up Mar 14 15:48:30 fray florian is here maybee he knows Mar 14 15:48:53 the problem w/ the octeon is that it requires a custom compiler.. and as far as I'm aware nobody has pushed that work back upstream (to gcc) Mar 14 15:48:53 plus then you have to get someone to write the cpu tunings for oe, and of course a BSP Mar 14 15:49:36 (AFAIK only Cavium can push the changes upstream for the compiler.. company I work for didn't write the code, so we can't contribute it.. [cavium wrote their mods for their SDK]) Mar 14 15:49:58 that makes sense Mar 14 15:50:07 tuning files are easy once the compiler support is there.. and then it's down to the BSP.. which is likely easy as well.. Mar 14 15:50:20 bsp? Mar 14 15:50:23 the best you can probably do is use the MIPS64 tuning today and then configure the kernel for cavium.. Mar 14 15:50:43 BSP -- board support package, this is the kernel configuration, kernel patches, bootloader, and device drivers Mar 14 15:51:01 well, I'll pen test binaries from uboot+tftp before I move to flash. Mar 14 15:51:05 even with all of that, you still need a secondry compiler and system for the simpleexec system to really make it useful.. Mar 14 15:52:10 he core in the Octeon is really just a MIPS64 core with a couple of additional instructions... if you don't use the instructions that is "fine", but it's a bit slower.. (Cavium also modified their ABI slightly to improve performance, which I suspect is why the gcc changes have never been sent upstream.. they 'broke' the ABI) Mar 14 15:52:49 interesting Mar 14 15:53:38 Octeon is a MIPS64 core + one or more "engines" the engines don't run linux, they run the "simpleexec" code.. Mar 14 15:54:25 * fray was involved in userspace/toolchain design for the octeon at WR about 3 years ago.. (I don't know much about the kernel side.. but the userspace side was 'almost' mips64) Mar 14 16:09:03 so far so good ... Mar 14 16:58:55 fray, just saw an email on oe-devel Mar 14 16:59:03 I think people must be doing real work :) Mar 14 17:01:52 hmm.. Mar 14 17:03:02 oe-core though... I haven't seen anything since about 8:45 last night.. (it's noon here) Mar 14 17:03:43 I saw something more recent Mar 14 17:03:48 and they are the same machine Mar 14 17:03:57 hmm Mar 14 17:04:09 no one else is complaining Mar 14 17:30:38 thanks for reminding of tune files... Mar 14 17:42:08 sent aarch64 includes to ml Mar 14 17:42:43 hrw, when did you send them? Mar 14 17:42:53 just now Mar 14 17:43:10 k Mar 14 17:43:28 fray, I have the email hrw is refering to Mar 14 17:43:39 did you receive it? Mar 14 17:44:37 I got it as well Mar 14 17:45:40 k Mar 14 18:05:01 who knows what Koen meant with 'nss-myhostname'? Mar 14 18:13:37 .00002 seconds via google http://en.wikipedia.org/wiki/NSS Mar 14 18:14:49 And I am sure if you look you can see.... http://en.wikipedia.org/wiki/Namespace_Specific_String#URN_Syntax Mar 14 18:15:00 MAn, this ADD generation today Mar 14 18:19:45 sure, but what does it mean when it comes to OE? Mar 14 18:28:17 bye Mar 14 19:19:18 Dude you are out there, Mar 14 19:20:04 I looked at days of hitory ..just wjere is this KOEN? Mar 14 19:20:22 what? Mar 14 19:20:50 if you're talking about koen kooi, try #angstrom, he doesn't hang out here Mar 14 19:21:38 then why the heck is hrw in here about question in reference to him? Mar 14 19:21:47 that is why I said he is "out there" Mar 14 19:22:25 [13:04:55] who knows what Koen meant with 'nss-myhostname'? Mar 14 19:22:45 I don't see what's confusing about that Mar 14 19:22:50 he was aksing if anyone else knew what koen meant Mar 14 19:22:54 he didn't ask koen himself Mar 14 19:22:59 i don't either Mar 14 19:23:02 because seems pretty straightforward to me Mar 14 19:23:15 And I siad it only took .0000000002 seconds to find on google Mar 14 19:23:33 then no, i don't see why you would be confused by his asking here. koen doesn't have to be here for him to ask what others thought. Mar 14 19:24:14 I am , because if the person is not on here, it's an incomplete conversation Mar 14 19:25:01 not really. he could (and probably was) have been referring to a mailing list thread or something else equally likely that folks in here would have seen. Mar 14 19:25:04 not that hard to grasp Mar 14 19:25:38 I stay logged in here 24/7 and I have never say him before Mar 14 19:25:47 I also don't see why you give a shit, honestly. do you really have nothing else to do than pick apart a question someone asks with a hostile attitude? Mar 14 19:26:02 plenty of us deal with koen on a regular basis. whether he's in here or not really doesn't matter Mar 14 19:26:06 nope Mar 14 19:26:10 [13:19:38] sure, but what does it mean when it comes to OE? Mar 14 19:26:16 and? Mar 14 19:27:05 just because your'e lacking the context to answer the question doesn't mean everyone else in here is missing the same Mar 14 19:27:16 and bitching and moaning about it doesn't help anyone Mar 14 19:27:23 I guess he does nto need to be in here then if he can't see it has NOTHING to do with OE, but part of the TCP/IP Mar 14 19:27:45 are you really that blind? Mar 14 19:27:54 nss-myhostname is a glibc module. oe includes glibc. Mar 14 19:27:57 https://fedoraproject.org/wiki/Features/nssmyhostname Mar 14 19:28:01 from *my* 2 seconds of googling Mar 14 19:28:17 clearly koen has an opinion about its use in oe Mar 14 19:29:06 * kergoth rolls eyes and goes back to real work Mar 14 19:30:13 out of curiosity where was a post about this? angstrom, oe-dev? Mar 14 19:40:21 fray: https://plus.google.com/u/0/110579591067038943673/posts/e3ezH2SACb8 Mar 14 19:43:31 huh.. didn't know anyone used google+ ;) Mar 14 19:44:36 interesting.. Mar 14 19:45:08 I definitely see the problem that nss-myhostname is trying to solve.. I just wonder if it's the right solution (based on the fedora link from Kergoth) Mar 14 19:46:08 the way I've always avoided that in the past is simply use 127.0.0.1 as the address for local system things that need name resolution, then ensure that is in the /etc/hosts Mar 14 19:46:27 "other stuff" comes on line later, and uses a different IP/name.. but then X11 and such are fine Mar 14 19:47:23 also is nice to restrict connections from non 127.0.0.1 addresses to things like X and dbus Mar 14 21:08:39 hello Mar 14 21:10:54 hello **** ENDING LOGGING AT Fri Mar 15 02:59:58 2013