**** BEGIN LOGGING AT Thu Mar 01 02:59:59 2012 Mar 01 04:06:37 evenin' Mar 01 07:44:44 how can I set the root password in openembedded? Right now the root password is empty, which kind of sucks:P Mar 01 07:46:45 git grep for _ZAP_ iirc Mar 01 07:48:01 CONFIG_IR_STREAMZAP? only things that comes up Mar 01 07:53:32 _zap_ :) Mar 01 07:53:45 iirc, it was function Mar 01 07:54:06 something like zap_root_password Mar 01 07:54:16 or root_password_zap Mar 01 07:58:54 zap_root_passwd it was)) thanks Mar 01 07:59:54 I just have to overwrite that function? Mar 01 08:17:43 good morning Mar 01 08:21:16 good morning Mar 01 08:21:22 morning Mar 01 08:24:15 yey, it worked Jay7, nice to have it password protected:D Mar 01 08:32:44 mertsas_: great :) Mar 01 08:43:27 Next I'm struggling with is adding a user. I have added USERADD_PACKAGES = "${PN}" USERADD_PARAMS = "-m -G video,audio -s /bin/sh username", but username is not created as a user. What more than this and inherit useradd do I need? Mar 01 09:39:43 how can I perform a custom action in do_populate_sysroot() an then contine with the standard behaviour? Mar 01 09:40:20 I have a recipe without do_populate_sysroot, if I implement it will I loose the default behaviour of do_populate_sysroot? Mar 01 09:44:02 mckoan: yes, but you could do do_populate_sysroot_prepend() instead Mar 01 09:44:44 bluelightning: aha, you're right! Mar 01 09:49:34 if you're referring to OE-Core btw you'd be well advised not to try to put stuff in the sysroot manually since it will not play nicely with sstate Mar 01 09:50:21 bluelightning: not using oe-core yet Mar 01 09:50:29 right, fair enough Mar 01 10:03:46 hi Mar 01 10:05:20 I need to know one thing ... if there a two recipe which are not dependent on each other ... is there a way to make sure that one recipe is not built before the other one .... like A recipe always configure built befiore B is staged Mar 01 10:08:19 Noor: possibly... what's the reason you want to do that specifically? Mar 01 10:10:00 bluelightning: I have a situation where I need gettext version 0.14 and gnutls version 2.10 .... in my setup when gettext for target is built before gnutls .... gnutls fails to configure Mar 01 10:10:44 I am trying to resolve the issue but for the time being I want to make sure that gettext for target is stagged after gnutls configure is executed Mar 01 10:12:06 in gnutls Makefile @GETTEXT_MACRO_VERSION@ is not replaced by the gettext version :( Mar 01 10:12:32 you could make do_populate_sysroot in gettext depend on gnutls do_configure using the depends varflag (git grep "\[depends\]" for some examples) Mar 01 10:12:37 that's a huge, huge hack though Mar 01 10:13:43 Noor: Will you be able to rebuild gnutls (after gettext have been staged)? Mar 01 10:13:47 bluelightning: in this way when we will run gettext it will build gnutls as well? Mar 01 10:14:17 esbenh: I cant .... that is the problem Mar 01 10:15:46 If I build gettext then built gnutls I get this "gettext infrastructure mismatch: using a Makefile.in.in from gettext version 0.17 but the autoconf macros are from gettext version @GETTEXT_MACRO_VERSION@" Mar 01 10:16:29 As there is no support for per-recipe staging, I guess you have to hunt down the reason for gnutls failing when gettext is staged, and fix it. Mar 01 10:17:41 esbenh: yes I am doing that but to make the ball rolling on over server and for continious builds I want to give my team a work arround Mar 01 10:17:55 how can I make systemd start udhcpc at startup Mar 01 10:17:57 but the final fix will be in gnutls Mar 01 10:52:33 is there any autoconf expert here Mar 01 10:53:06 Noor: I would not call myself exert but I did quite a few things with autoconf Mar 01 10:54:52 Jin^eLD: Need to know that in one Makefuile.in I am getting $(top_srcdir)/m4/gettext.m4 in am__aclocal_m4_deps and when configuration changes then gettext.m4 does not come in am__aclocal_m4_deps .... can you give me pointer what I can look to investigate it Mar 01 10:56:13 Makefile.in is regenerated when autoreconf command is executed Mar 01 10:56:28 Noor: well I assume Makefile.in is generated from Makefile.am (which will be the case if the project uses automake too) Mar 01 10:56:34 so your source to look for problems will be Makefile.am Mar 01 10:56:48 the .in is already in an inhuman form :) Mar 01 10:57:50 so your gettext.m4 is in the m4 directory and you also have ACLOCAL_AMFLAGS = -I m4 Mar 01 10:57:55 in the Makefile.am? Mar 01 10:58:40 is there any output files which I can look that what has changed which caused gettext.m4 not included there Mar 01 10:59:24 Jin^eLD: my Makefile.am says ACLOCAL_AMFLAGS = -I m4 -I gl/m4 Mar 01 10:59:26 I am not sure, config.log will only have autoconf stuff, not sure here automake logs anything (if it does at all) Mar 01 11:00:10 so just that I understand correctly: the problem is that at some point gettext.m4 is getting lost for you? Mar 01 11:01:27 yeah ... and this happend when gettext version 0.14 is there in sysroot :) Mar 01 11:01:59 Jin^eLD: just notisec a folder name autom4te.cache which has automake cache and traces output but need to understand it Mar 01 11:02:26 notice Mar 01 11:02:48 I have to admit that most of my problems so far were somewhere around configure, so I did not have to debug automake yet Mar 01 11:03:04 I did notice however that sometimes stuff that I put into m4 is not taken Mar 01 11:03:40 but I never came to investigating it, simply dumped what I needed into $(top_srcdir)/acinclude.m4 Mar 01 11:05:27 hhhmmm no problem Jin^eLD .... I look into it thanks Mar 01 11:35:55 hi Mar 01 11:36:29 im using new bitbake version 1.10.2 Mar 01 11:36:57 with latest openembedded Mar 01 11:37:52 during parsing its giving parsing error for various files saying "ERROR: EOL while scanning string literal" Mar 01 11:40:33 the line on which error is coming is "${@oe_filter_out('(-L\S+|-l\S+)', '-Wl,-O1 ${TARGET_LINK_HASH_STYLE}', d)}" Mar 01 11:43:19 its also showing error in line "DISTRO_UPDATE_ALTERNATIVES ?= "${@base_conditional("ONLINE_PACKAGE_MANAGEMENT", "none", "", "${PREFERRED_PROVIDER_virtual/update-alternatives}", d)}"" Mar 01 12:06:08 hi, all! Mar 01 12:06:27 is it possible to have multiple kernels in single build? Mar 01 12:14:00 hii Mar 01 12:14:09 can anyone help me Mar 01 12:14:15 with these parsing errors Mar 01 12:14:19 i am new to OE Mar 01 12:15:08 anuj_: it looks like extra " in the end Mar 01 12:15:26 ah maybe you've added it Mar 01 12:15:29 yes Mar 01 12:16:59 ERROR: EOL while scanning string literal (, line 1) while parsing /openembedded/recipes/uclibc/uclibc-initial_0.9.31.bb Mar 01 12:17:16 NOTE: :EOL while scanning string literal (, line 1) while evaluating: ${@oe_filter_out('(-L\S+|-l\S+)', '-Wl,-O1 ${TARGET_LINK_HASH_STYLE}', d)} Mar 01 12:18:30 hmm that's not oe-core.. for oe-classic you need older bitbake Mar 01 12:19:00 1.12 or older Mar 01 12:23:57 k i will check with older version of bitbake Mar 01 12:26:42 hey guies do we have an android ndk in OE Mar 01 12:33:31 with bitbake 1.8.12 we are getting some different parsing errors Mar 01 12:33:32 http://pastebin.com/ZM9iAHnb Mar 01 12:35:51 1.8.12 is maybe too old :) Mar 01 12:36:13 k Mar 01 12:36:20 just use 1.12 Mar 01 12:36:38 1.12 is higer than 1.10.2 Mar 01 12:36:39 ?? Mar 01 12:36:54 yes Mar 01 12:37:20 it's 1.12.0 (1.12 is name of bitbake branch) Mar 01 12:38:36 k Mar 01 12:40:38 i have a problem with a SRC_URI. I would like to write a recipes for xine-lib. But the download URL has a whitespace. I tried to use "%20" bit i didn't work. Mar 01 12:40:51 Here is the URL http://sourceforge.net/projects/xine/files/xine-lib%201.2/1.2.1/xine-lib%201.2-1.2.1.tar.bz2 Mar 01 12:43:04 http://downloads.sourceforge.net/project/xine/xine-lib/1.2.1/xine-lib-1.2.1.tar.bz2 Mar 01 12:43:12 there is no space afaik Mar 01 12:43:42 and use SOURCEFORGE_MIRROR variable Mar 01 12:47:39 JaMa: with 1.12.0 we are getting same parsing errors Mar 01 12:49:31 http://pastebin.com/hFuu8W6z Mar 01 12:52:17 we are using classic OE http://cgit.openembedded.org/openembedded/ Mar 01 12:54:05 decembersoul you are aware that there are already recipes for xine? Mar 01 12:54:06 I have problems building gcc for device with oe-core/meta-openembedded of a few days ago. http://pastebin.com/Qh18pLSa Mar 01 12:54:18 I mean gcc for target Mar 01 12:54:40 slapin same as for host Mar 01 12:54:55 patching the same source from diffrent recipes Mar 01 12:55:02 should be fixed now Mar 01 12:56:10 woglinde: It seems it doesn't, I build from scratch and see this. I do bitbake gcc -c cleanall and do build anagi and get this. Mar 01 12:56:24 stateclean Mar 01 12:56:26 again Mar 01 12:56:35 hm Mar 01 12:56:42 than look up the recipes Mar 01 12:56:48 maybee it suffer the same Mar 01 12:56:52 as the cross stuff Mar 01 12:56:56 some stray patch Mar 01 12:56:59 woglinde: cleanall cleans all including source. Mar 01 12:57:35 btw. http://sourceforge.net/projects/xine/files/xine-lib/1.2.1/xine-lib-1.2.1.tar.bz2 has no space Mar 01 12:57:46 woglinde: so stray it soesn't apply? Mar 01 12:57:49 ups Mar 01 12:57:50 jama Mar 01 12:58:08 slapin stray that it will be applied from two recipe at the same time Mar 01 12:58:17 slapin look at the patch for gcc-cross Mar 01 12:58:32 can anyone help me with this Mar 01 12:58:33 ?? Mar 01 13:00:29 slapin -> http://lists.linuxtogo.org/pipermail/openembedded-core/2012-February/018296.html Mar 01 13:00:48 anuj_ what exactly is the error? Mar 01 13:01:35 be aware classic is end of life only really big errors will be fixed Mar 01 13:03:02 woglinde: please consider applying this http://pastebin.com/J41yKBm7 or else it is badly broken, since I have no time at the moment for proper bug report/mailing of the patch Mar 01 13:03:18 slapin me neither Mar 01 13:03:34 so I hope somebody can Mar 01 13:03:40 woglinde /openembedded/recipes/uclibc/uclibc_0.9.30.3.bb: EOL while scanning string literal (, line 1) ERROR: Command execution failed: Exited with 1 Mar 01 13:03:53 hm I thought the patch is already in Mar 01 13:04:50 slapin: sending it with git send-email doesn't take more time then pasting it to pastebin Mar 01 13:04:55 how can I get some damn commit access somewhere in *contrib repos? than I think pull requests be of less pain at my side.... Mar 01 13:05:15 JaMa I though anderes already did Mar 01 13:05:21 ups andreas Mar 01 13:05:25 * JaMa too, but I'm not using 4.5 Mar 01 13:05:39 JaMa: I don't have normal scriptable mail access in this whole dump around here Mar 01 13:05:40 woglinde: where is the xine recipes? Not at oe-core? Mar 01 13:05:53 but he removed it only from gcc-runtime http://git.openembedded.org/meta-openembedded/commit/?id=11ab6e41b1131ea489e69c868ba2767201d39f93 Mar 01 13:06:02 lol Mar 01 13:06:06 mailing patches from here is so much pain you can't imagine it Mar 01 13:06:36 why he or richard did not search for similarities Mar 01 13:06:40 this suckz Mar 01 13:09:18 Ok, think I'm just lazy, but please apply the damn thing. I'm past 2 days of planned release damnit... Mar 01 13:10:57 but that of course is my problem :( Mar 01 13:12:57 and you get paid for whining about your mail setup? :) Mar 01 13:13:38 jama he could pay you for sending the patch in Mar 01 13:13:52 ok i found libxine 1.1.19 at the classic oe tree. Is there a new one for oe-core? Mar 01 13:13:53 so now going to lunch for real Mar 01 13:13:59 decembersoul no Mar 01 13:14:08 but you can hope it over Mar 01 13:14:23 i will try. Thanks Mar 01 13:14:24 and it would go either to multimedia or meta-oe Mar 01 13:14:27 but not core Mar 01 13:14:44 so be aware when sending the patches Mar 01 13:14:47 and remove PR Mar 01 13:15:10 yeah right.. nobody paid me anything for OE work and still I'm not whining :) Mar 01 13:15:33 jama you are making something wrong *me runs* Mar 01 13:15:57 :) Mar 01 13:16:35 * JaMa will start whining :) Mar 01 13:16:47 JaMa, woglinde: I'm paid for having public transport coordinates/points on some glamour blonde laptop screen, that's all. Mar 01 13:17:16 so I'm not paid for whining either Mar 01 13:17:22 * JaMa is paid for Java apps runing on windows :) Mar 01 13:17:25 but this might work, too. Mar 01 13:19:04 well, I really into donation for OE devs and mailed patches to OE earlier, but this time I'm out of infrastructure... Mar 01 13:19:23 hi wondering is there any web server packages inside openembeddeD? Mar 01 13:19:48 kay_: lighttpd Mar 01 13:19:53 kay_: apache too Mar 01 13:20:16 the apache works like normal apache? Mar 01 13:22:03 Anybody to hire me? I can whine properly. Please gimme commit access to meta-openembedded-contrib? I promise I will send patches then... Mar 01 13:25:56 you can send patches from any git repo too Mar 01 13:26:25 and oe has mirrors on github Mar 01 13:33:48 * woglinde wonders why jama is helping with meta-java Mar 01 13:34:13 * woglinde wonders too what you stops to apply the patch locally Mar 01 13:35:58 then it will rot and nobody will benefit from it Mar 01 13:36:16 nobody said it should rot Mar 01 13:36:24 but for your deadline its okay Mar 01 13:36:40 better thank me I pointed you at the right spot Mar 01 13:36:49 deadline is the whole life thing Mar 01 13:36:52 otherwise you could have called it a day Mar 01 13:37:22 woglinde: thank you very much Mar 01 13:37:27 hi when i try to bitbake apache Mar 01 13:37:32 it gives me nothing provide apache Mar 01 13:37:38 any idea why? Mar 01 13:37:47 hm we should make a recipe for ngix Mar 01 13:37:57 nobody needs apache Mar 01 13:39:00 nginx even Mar 01 13:39:22 woglinde: apache is cool when you build for some cool heavily-loaded embedded box with lots of resources Mar 01 13:39:37 woglinde: lighttpd is not needed either Mar 01 13:40:18 this whole web interface idea is wrong way I think Mar 01 13:40:42 like for sheevaplug? Mar 01 13:40:43 but it is common for most computer users Mar 01 13:40:51 denisATeukrea o.O Mar 01 13:40:59 think about it again Mar 01 13:41:16 why is apache bad? Mar 01 13:41:30 I didnt say its bad Mar 01 13:41:30 I mean if you want to run a webapp Mar 01 13:41:42 but there are now more solutions which are better Mar 01 13:41:56 denisATeukrea: yeah Mar 01 13:42:13 its like php or mysql Mar 01 13:42:21 sure you can use it Mar 01 13:42:25 but the price is high Mar 01 13:42:30 * denisATeukrea doesn't like php a lot but it's the standard Mar 01 13:42:42 standard? Mar 01 13:42:45 it seems everybody runs python web servers these days. or some tiny web server + tomcat Mar 01 13:42:53 ah ok Mar 01 13:43:05 I tought that hosting providers provided mainly php Mar 01 13:43:12 that is to say I prefer python Mar 01 13:43:53 I dont care about hostings providers Mar 01 13:44:17 ok but some people do when writting applications Mar 01 13:44:21 well, I host several of sheevaplugs with tomcat for some indie-web-development, works perfectly. Mar 01 13:44:28 do we have tomcat in OE? Mar 01 13:44:35 no Mar 01 13:44:56 but we have java Mar 01 13:45:14 and thanks to me for oe-core too Mar 01 13:47:03 well, is there some "Donate" button somewhere on OE sites? I think I could spend $10 from time to time, because thanks to OE I can be lazy most of time at work and get a good salary. Mar 01 13:47:29 slapin it is Mar 01 13:47:38 configure: error: cannot compute sizeof (off_t) for lighttpd Mar 01 13:47:52 any idea how to solve it? Mar 01 13:48:03 woglinde: any ideas on where? Mar 01 13:48:07 kay update the recipe to a newer version? Mar 01 13:48:14 ok Mar 01 13:48:23 slapin I thought in wiki we have paypal Mar 01 13:48:44 hms Mar 01 13:48:50 need to talk to florian Mar 01 13:50:08 hey woglinde how u update a reciepe? Mar 01 13:51:41 kay raising the version number? Mar 01 13:52:02 apache seems to give the same problem too Mar 01 13:52:07 is apache or apache2? Mar 01 13:53:13 the recipe i am using is the newest already Mar 01 13:55:42 let me see Mar 01 13:55:52 by the way how can I clean packaged staging stuff with bitbake? -c cleanstate and -c stateclean doesn't work :( Mar 01 13:57:43 there are two 's' Mar 01 13:57:50 it's shared state so sstate Mar 01 13:57:53 -> cleansstate Mar 01 14:06:00 JaMa: thanks alot! Mar 01 14:06:14 today is my day here it seems Mar 01 14:54:17 morning all Mar 01 14:54:50 hi pb Mar 01 14:57:29 hm lighttp in meta-oe builds fine Mar 01 15:10:17 Hi there, I noticed that the SD card emulation of qemu doesn't like unaligned image sizes. Mar 01 15:12:38 In certain cases qemu rounds down (hw/sd.c:sd_reset) which then leads to a fail in the guest because of the attempt to access beyond end of device Mar 01 15:14:51 With pure OE-core this doesn't happen because the image is not given to qemu as an SD card usually Mar 01 15:15:14 qemuarm for example uses the old versatile which has an IDE Mar 01 15:16:06 In my case, I've created a meta layer that is using the vexpress model and the only way to specify a rootfs is via SD card Mar 01 15:16:52 The code that calculates the image size is at meta/classes/image_types.bbclass: runimagecmd() Mar 01 15:19:03 I've created the following patch to have the image size aligned: http://pastie.org/private/zhtetjzqsqbwsntbssaduw Mar 01 15:31:25 kenws: seems cut off at the right Mar 01 15:31:56 kenws: I'm guessing this ought to be optional though Mar 01 15:35:03 bluelightning: ok, thanks for your comment. how would you do this optionally? invent a new variable to test for? Mar 01 15:35:19 kenws: I think so yes Mar 01 15:36:03 I wonder, is it likely that people would want an alignment other than 4MB? Mar 01 15:36:31 i.e. is the variable a yes/no or a value Mar 01 15:36:48 It rather seems to me that it is a bug in qemu that it rounds down. Mar 01 15:37:32 pb_: this is what I thought at first. I talked to peter maydell about this and it seems that this is just the correct way to emulate an SD card device Mar 01 15:39:48 why can't it just round up instead of down? that seems a bt strange. Mar 01 16:03:34 pb_: My knowledge about SD cards and qemu is limited. What Peter states is: the reason qemu rounds is that sd cards have write-protect groups that are pretty large and you can't have an sd card with a handful of bytes hanging off the end of the wp group Mar 01 16:04:57 pb_: I think 2MiB alignment would be fine. Mar 01 16:07:08 I suspect it can't round up because the file isn't large enough so qemu can only round it down to something reasonable Mar 01 16:11:33 bluelightning: pb_: I guess I'll bring this topic up on the oe-core ml or do you have any other suggestions? Mar 01 16:14:22 kenws: might be worth doing so yes Mar 01 16:14:59 ok Mar 01 16:14:59 Is there a way to extend/overwrite bbclasses from a meta layer? Mar 01 16:17:04 no, I'm afraid not Mar 01 16:17:51 well, not to extend... if you copy the class into your layer and ensure BBPATH is set up in the correct order via your layer.conf then it can provide its own overlayed version Mar 01 16:18:47 bluelightning: I see, but this creates quite some redundancy Mar 01 16:19:03 the ideal situation is to avoid having to do this by getting the functionality into OE-Core Mar 01 16:30:45 bluelightning: I think having a variable that specifies the alignment would be nice. In case OE-core doesn't want any alignment it could be set to 1. Mar 01 16:31:26 kenws: indeed, I don't think that should be too controversial particularly if the qemu behaviour is unlikely to change Mar 01 16:34:30 Sounds, good. I'll provide a patch. Where would you place place a variable such as IMAGE_ROOTFS_ALIGNMENT? Right into meta/conf/bitbake.conf or are there better places? Mar 01 16:35:53 kenws: well, you could just do a ?= assignment in image_types.bbclass for the default value Mar 01 16:36:09 that's probably the preferred way Mar 01 16:36:45 ah, true. thanks! Mar 01 16:38:38 hm.. what is classes/image-swab.bbclass? Mar 01 16:38:51 I see there PARALLEL_MAKE_pn-* Mar 01 16:39:01 core or oe? Mar 01 16:39:07 core Mar 01 16:39:31 I'm trying to collect list of recipes with disabled P_M Mar 01 16:39:48 should I take image-swab.bbclass items into account? Mar 01 16:40:24 dont know Mar 01 16:40:49 same here :) Mar 01 17:08:22 Jay7: it's for enabling swabber, which analyses host usage during a build Mar 01 17:08:43 bluelightning: ah, ok Mar 01 17:09:09 well.. let's run some builds now Mar 01 18:25:55 dpkg fetch failed Mar 01 18:26:12 hm Mar 01 18:26:14 ERROR: No recipes available for: /devel/arm/git/setup-scripts/sources/meta-openembedded/meta-oe/recipes-graphics/xorg-lib/pixman_0.24.2.bbappend Mar 01 18:26:18 | File: '/home/oe/sources/dpkg_1.15.8.7.tar.bz2' has md5 checksum 4960bf4cfcae5ff15be988ab408a87ad when d1731d4147c1ea3b537a4d094519a6dc was expected Mar 01 18:26:20 | File: '/home/oe/sources/dpkg_1.15.8.7.tar.bz2' has sha256 checksum 7e6a7a892110926146024d34d851fd51929c0a4546696848230a64dd5bb4636e when 1ec1376471b04717a4497e5d7a27cd545248c92116898ce0c53ced8ea94267b5 was expected Mar 01 18:26:25 checksum mismatch Mar 01 18:27:19 aha lol Mar 01 18:27:23 core has now .4 Mar 01 18:27:24 not .2 Mar 01 18:28:42 hehe.. both checksum are wrong.. Mar 01 18:28:56 hm Mar 01 18:29:21 ah Mar 01 18:29:28 it's html file Mar 01 18:30:20 seems wrong url in recipe Mar 01 18:31:38 moved? Mar 01 18:31:46 sounds likely Mar 01 18:31:57 no.. looks like shapshot.debian.org was reorganized Mar 01 18:35:03 well.. ftp.debian.org have no such file Mar 01 18:35:22 seems dpkg recipe should be upgraded Mar 01 18:38:27 * Jay7 pings RP Mar 01 18:43:35 now do_patch is failed Mar 01 18:43:44 noman.patch does not apply Mar 01 18:44:02 well.. I'll remove it from testing list Mar 01 20:43:59 hm.. I have good enough results of some first PARALLEL_MAKE-testing runs Mar 01 20:44:15 ~4-5 packages confirmed Mar 01 20:44:34 others need more testing Mar 01 20:47:45 khem ping Mar 01 21:26:30 woglinde: yo Mar 01 21:28:22 khem sorry no problem anymore Mar 01 21:28:34 ah time heals :) Mar 01 21:28:53 Jay7: Pretty good work Mar 01 21:29:03 no was only about va_list mangling note Mar 01 21:29:10 for arm eabi since 4.4 Mar 01 21:29:14 guess there is no fix Mar 01 21:29:18 for the note Mar 01 21:29:37 k Mar 01 21:29:50 or is there one? Mar 01 21:29:58 I did not found something Mar 01 21:30:12 va_list is mangling what ? Mar 01 21:30:40 -> note: the mangling of 'va_list' has changed in GCC 4.4 Mar 01 21:31:51 I hoped you know something about it Mar 01 21:32:00 oh yes I do Mar 01 21:32:13 I was not understanding context from your message Mar 01 21:32:19 but now I know waht you mean Mar 01 21:33:21 its and ABI change Mar 01 21:33:28 so where are you seeing it Mar 01 21:35:30 by now all system headers should be clean of it Mar 01 21:37:04 jni.h in openjdk for instance Mar 01 21:37:28 ok so you want to fix it or avoid it ? I can provide both Mar 01 21:37:47 -Wno-psabi should silence the warning Mar 01 21:39:05 I want the correct solution Mar 01 21:39:39 ok then provide the preprocessed output of the file Mar 01 21:40:36 what is your host machine gcc ? Mar 01 21:41:00 hm oh Mar 01 21:41:09 rm_work now kicks in Mar 01 21:41:19 http://pastebin.com/sagTDNg6 Mar 01 21:41:23 QA issue Mar 01 21:42:47 Jay7: yeah it warns since you are linking with libs from both usr/lib and /lib Mar 01 21:43:30 somehow we have this esoteric view that / and /usr can be separately mounted Mar 01 21:43:52 and /usr may not be yet available Mar 01 21:44:11 when programs under / execute Mar 01 21:44:21 hence the warning Mar 01 21:44:41 we have been movin stuff into / Mar 01 21:44:53 and I am skeptic of that change Mar 01 21:45:33 khem: now /usr should be mounted by initramfs on every modern distribution Mar 01 21:45:59 yes but some folks seems to be not that modern yet Mar 01 21:46:21 and this QA ldd output is ok? libck-connector.so.0 => not found Mar 01 21:47:32 whole output http://fpaste.org/cEyj Mar 01 21:47:54 currently confirmed recipes are: bind dpkg slang tcp-wrappers js pth Mar 01 21:48:31 ynezz: I think thats a problem since its not staged it seems Mar 01 21:48:52 ynezz: do u have libck-connector.so.0 in sysroot ? Mar 01 21:49:50 'find build/tmp-angstrom_2010_x-eglibc/sysroots -name libck-connector.so.0' returns nothing Mar 01 21:53:57 * Jay7 started another set of builds and gone sleep Mar 01 22:01:53 ynezz: interesting Mar 01 22:02:01 so how did the dep came in I wonder Mar 01 22:07:09 http://fpaste.org/Yspd Mar 01 22:07:25 it does exist, but not in the sysroot Mar 01 22:46:51 Hello. Mar 01 22:47:07 I'm suffering problems while trying to rebuild my OE distro. Mar 01 22:47:29 In this moment one of source URLs stopped responding and I can not build my image. Mar 01 22:48:12 Is there any posibility to change bitbake behavior to continue building with older sources if current ones download fails? Mar 01 23:18:31 otwieracz: that happens a lot with TI packages... Mar 01 23:19:35 our workaround is 1) download the source tarball by hand and put it in oe/downloads, and 2) create the md5sum file and update the recipe Mar 01 23:20:10 unless you can just replace the SRC_URI with a a valid one Mar 01 23:20:21 s/a// Mar 01 23:36:06 mr_science for ti packages you only need if the license do not allow a direct download Mar 01 23:36:11 good nite **** ENDING LOGGING AT Fri Mar 02 02:59:58 2012