**** BEGIN LOGGING AT Mon Aug 02 02:59:57 2010 Aug 02 04:00:26 anyone about here? Aug 02 04:00:55 I'm writing a recipe that I want to run ./configure (some options) && make && make install Aug 02 04:01:16 but using "inherit autotools" causes autoconf to be run as well Aug 02 04:01:41 and autoconf is failing because "error: possibly undefined macro: AC_BAKEFILE" Aug 02 04:02:09 bakefiles don't seem to be well supported with openembedded (not that I'm familiar with them) Aug 02 04:02:29 so I think my best option is to cause autoconf to not be run Aug 02 04:02:33 how do I go about that? Aug 02 04:17:52 can you write your own do_configure() func in your recipe? (I'm a newb, so not sure) Aug 02 04:56:08 I'm a tool Aug 02 04:56:30 why google when vi classes/autotools.bbclass is around Aug 02 06:13:18 porcoesphino: if your package does require autotools just leave out the inherit autotools and write your own do_configure and do_install Aug 02 06:13:45 there are plenty of examples around there Aug 02 06:42:46 cheers eFfeM Aug 02 06:42:54 hey, I'm having an issue... Aug 02 06:43:12 it looks like the last include from CXX_FLAGS is getting dropped Aug 02 06:43:16 but I cant tell why Aug 02 06:43:45 I mean CPP_FLAGS Aug 02 06:43:54 a work around is to dup the -I/dir into CXX_FLAGS Aug 02 06:44:12 and it works... but it seems messy and I can't track down why it's being dropped Aug 02 06:44:16 anyone have any ideas? Aug 02 06:46:39 good morning Aug 02 06:56:46 porcoesphino: seems strange, no idea Aug 02 06:56:49 mckoan: hi Aug 02 07:00:45 Here's what I did. Bitbake failed during make. I copied the command that had been executed into an editor. Found the rule to make that file in the Makefile in the 'work' directory, copied that line to the text editor. Then hunted all of the variables down that it was using. CPPFLAGS was -isystem/dir1 -I/dir2 but then when you looked at the log from the make file only isystem/dir1 was being used. Aug 02 07:01:07 adding -I/dir2 to CXXFLAGS is the work around.. Aug 02 07:01:16 Any issues with that logic? Aug 02 07:01:29 if not.. then CPPFLAGS is being overriden somewhere? Aug 02 07:02:09 morning Aug 02 07:11:56 hrw: gm Aug 02 07:35:01 Ah, I forgot to say cheers eFfeM Aug 02 07:45:41 porcoesphino: don't really have an idea off-hand Aug 02 07:48:03 eFfeM_work: the same thing is happening with LDFLAGS if that changes anything Aug 02 07:48:15 well... in that case it's the last -L/dir Aug 02 07:49:05 i fear i have little knowledge in that area Aug 02 07:49:20 never bumped into it Aug 02 07:53:19 hi Aug 02 07:53:32 some stable maintainer here? Aug 02 08:02:22 ~seen hrw Aug 02 08:02:25 hrw is currently on #maemo #oe. Has said a total of 3 messages. Is idling for 59m 50s, last said: 'morning'. Aug 02 08:02:38 ping hrw Aug 02 08:03:12 woglinde: pöng Aug 02 08:04:05 hrw could you please ack http://patchwork.openembedded.org/patch/2478/ and http://patchwork.openembedded.org/patch/2479/ Aug 02 08:07:06 woglinde: feel free to add my Ack Aug 02 08:08:11 thanks Aug 02 08:58:27 for anyone who searches the above issue I was having. It looks like it was due to recreating the functions I wanted from autoconf.class instead of inheriting and then overidding do_configure with oe_runconf. I'm still not sure what exactly it was that was overriding/removing the last flags from the variables though Aug 02 09:00:29 porcoesphino whats your problem? Aug 02 09:01:36 hey guys Aug 02 09:01:58 hi xora Aug 02 09:02:38 hi Aug 02 09:02:55 hi jama Aug 02 09:03:02 wb from vacation Aug 02 09:03:35 morning Aug 02 09:03:45 * JaMa|Wrk already looking forward to longer vacation in September :) (3 weeks in US) Aug 02 09:04:18 ieehks zs Aug 02 09:04:21 ups US Aug 02 09:04:50 hey JaMa|Wrk Aug 02 09:05:01 ah, it's from a bit back Aug 02 09:05:10 I dont like travel to countrys which forces me to write down "That I am not a terrorist" Aug 02 09:05:21 I altered my recipe and it disappeared. That was just for anyone who searches the archives.. Aug 02 09:05:45 * JaMa|Wrk wants to see those parks and doesn't care about writing "I am not a terrorist" Aug 02 09:06:21 basically my last -I flag was being removed from CPPFLAGS and my last -L from LDFLAGS Aug 02 09:06:50 porcoesphino pastebin the recipe Aug 02 09:07:07 porcoesphino then its easier to understand your problem Aug 02 09:07:16 * XorA prefers to write I am not, rather than I am :-) Aug 02 09:09:56 woglinde: it's a tad pointless now, but I'm curious so sure :) Aug 02 09:10:19 JaMa|Wrk: which parks are you going to? Aug 02 09:10:46 shading = bash? Aug 02 09:12:15 XorA: Yosemite, Sequoia & Kings Canyon National Parks, Grand Cannyon, Hoover Dam and also SFO, LA, meteorcrater, Albuquerque, ... Aug 02 09:12:36 here's the current working recipe: http://pastebin.com/KWt2N737 Aug 02 09:13:14 JaMa|Wrk, is going to my places. Aug 02 09:13:20 here's the one that had the random error where the last -I flag was being removed from CPPFLAGS and my last -L from LDFLAGS: http://pastebin.com/pGLtpw2e Aug 02 09:14:27 actually... those directories shouldn't be like that.. Aug 02 09:14:44 I'm still in the process of testing the resultant package Aug 02 09:14:51 JaMa|Wrk: nice tour Aug 02 09:17:26 porcoesphino why you need to override do_configure? Aug 02 09:17:49 and you do_install twice Aug 02 09:17:51 JaMa|Wrk: nice tour. don't forget bryce and zion (better than grand canyon imho) Aug 02 09:17:52 +have Aug 02 09:18:39 btw wrt "i am not a terrorist". ofc a terrorist would never ever lie on that question. Aug 02 09:20:04 porcoesphino dont use hardcoded paths like /home/OE/angstrom-dev/sysroots/armv7a-angstrom-linux-gnueabi/usr/include/mysql Aug 02 09:20:13 there are VARIABLES for Aug 02 09:20:25 JaMa|Wrk: have a look at http://www.dse.nl/~meulenbr/holiday2000/ Aug 02 09:20:39 woglinde: ha. I was doing install twice.. still that's on odd effect if that's what was causing it.. Aug 02 09:20:49 effem 2000 over ten years Aug 02 09:21:15 woglinde: yeah, it this works fine I'll go back to making sure the directories aren't hard coded Aug 02 09:21:39 porcoesphino again why do you need to override do_configure? Aug 02 09:21:50 woglinde: was there 10 years ago Aug 02 09:22:09 effem yes Aug 02 09:22:12 before 2001 Aug 02 09:22:20 btw nicest park imho is yellowstone (but quite crowded), been here a few years back Aug 02 09:22:29 canadian rockies was also nice Aug 02 09:22:36 woglinde: ah sorry. because autoconf requires something that I don't have installed on the machine and ./configure is all I need to execute Aug 02 09:23:36 ie. I want ./configure && make && make install instead of automake && autoconf && ./configure && make && make install Aug 02 09:23:51 although I could just fix the mixing .m4 file... Aug 02 09:23:57 *missing Aug 02 09:25:44 eFfeM_work: added to our wiki, thanks! Aug 02 09:25:49 hrw: FYI: an important mdbus bug has been fixed and i release 2.0.2 Aug 02 09:25:57 released, that is Aug 02 09:26:33 mickey|office: thx - will try to remember to push new ver Aug 02 09:41:17 JaMa|Wrk: btw we liked the north rim of the grand canyon better than the south rim (but both bryce and zion have better hiking opportunities (at least in my opinion) Aug 02 11:25:53 Iam having Mistral, TMDSEVM3530 board. for that board what recepie image need to be selected in oe tree using bitbake Aug 02 11:26:13 what you want Aug 02 11:26:19 will omap3evm-demo-image work for Mistral, TMDSEVM3530 ? Aug 02 11:26:33 maybee Aug 02 11:26:34 i want a gui image on top of that Mistral, TMDSEVM3530 Aug 02 11:26:50 is there a machine file for mistral? Aug 02 11:27:16 no i got for overo ... omap3evm-demo-image-overo.tar.bz2 Aug 02 11:27:30 how to make it for Mistral ? Aug 02 11:29:04 i am using gumstix oe tree.. will i get images for Mistral board. (Or) Any any known oe tree having compatible support for Mistral Aug 02 11:53:58 jo gnutoo Aug 02 11:54:05 hi Aug 02 11:54:17 I've no music in wesnoth,I'll look if sdl mixer has ogg support Aug 02 11:54:21 it has mad Aug 02 11:54:31 s#ogg#ogg/vorbis# Aug 02 11:54:52 why you need sdl-mixer? Aug 02 11:55:03 what should that change? Aug 02 11:56:58 I'm not shure Aug 02 11:57:08 I've already sdl-mixer Aug 02 11:57:17 but I need ogg support Aug 02 11:57:20 it has mad support Aug 02 11:59:05 hmmm Aug 02 11:59:08 EXTRA_OECONF = "--disable-music-mp3 --enable-music-ogg --enable-music-ogg-tremor --enable-music-mp3-mad-gpl" Aug 02 11:59:20 but... Aug 02 11:59:20 http://pastebin.com/4RXhVKxF Aug 02 11:59:21 maybe it's dlopened Aug 02 11:59:21 I'll look Aug 02 11:59:43 ah tremor not installed Aug 02 12:01:47 woglinde, how's things going? Aug 02 12:01:58 slowly Aug 02 12:02:39 ah? Aug 02 12:02:48 I'll ask on #wesnoth-dev Aug 02 12:10:12 btw I'll have an oe question later about new staging and python Aug 02 12:20:19 woglinde, by slowly you mean that work is slow? Aug 02 12:21:16 yes Aug 02 12:22:56 ok Aug 02 12:34:16 hmmm PR bump didn't change anything Aug 02 12:34:31 same ldd thing Aug 02 12:34:32 rebooting Aug 02 14:26:16 Hi Aug 02 14:26:40 hi Aug 02 14:27:08 Which variable is responcible to hold the value of CROSS compiler flag -m64 ? Aug 02 14:29:41 While doing bitbake for the QEMU recipe. It is automatically configured for the 32 bit architecture. How one can configure for 64 bit architecture Aug 02 14:30:08 hm Aug 02 14:30:26 qemu-native is builded with the host compiler toolchain Aug 02 14:33:58 hello Aug 02 14:34:38 hi ant Aug 02 14:36:01 woglinde: in the (hypotetical) free time, give a look at dietlibc failure / segfaults on do_compile, pls Aug 02 14:36:35 hm ieehks Aug 02 14:36:50 dev is a friend of you, isn't ? Aug 02 14:37:03 eh..frieund even Aug 02 14:37:15 let say Freund ; Aug 02 14:37:36 new releases perhaps? Aug 02 14:38:03 sorry Aug 02 14:38:07 no time Aug 02 14:38:13 until october for it Aug 02 14:38:25 I thought kexeclib was better anyway Aug 02 14:38:48 you see, with latest klibc and latest kexec tools I'm back to 100 kb binary size Aug 02 14:38:56 ugh Aug 02 14:39:00 why this? Aug 02 14:39:08 disbale some options maybee Aug 02 14:39:11 pre-latest kexec-tools was 140 kb !!! Aug 02 14:39:22 I mean 2.01 vs. 2.02 Aug 02 14:39:28 pls help ;) Aug 02 14:39:29 you maybee could try uclibc with a stripped down featureset Aug 02 14:39:37 there is a strange --zlib option Aug 02 14:40:16 I'd like someone check the recipe Aug 02 14:40:29 and finally, I should rename it Aug 02 14:40:38 it seems 'static' is taboo Aug 02 14:41:25 pb_ once suggested to create a -dev package for the klibc-static binaries iirc Aug 02 14:41:38 -dev? Aug 02 14:41:43 iirc ... Aug 02 14:42:00 is like your recipe kexec-tools-dietlibc Aug 02 14:42:05 this is static per se Aug 02 14:42:28 yes dielibc builds all stuff static Aug 02 14:42:37 but kexec-tools-dietlibc-static seems wrong Aug 02 14:42:39 there is some shared lib options Aug 02 14:42:45 but that isnt widely used Aug 02 14:42:51 ah Aug 02 14:43:01 in klibc we force static too Aug 02 14:43:10 now defaults is shared since some time iirc Aug 02 14:43:13 :/ Aug 02 14:50:26 LDFLAGS = "-Wl,--warn-common -m64 -g -L${STAGING_LIBDIR} -Wl,-rpath-link,${STAGING_LIBDIR} -Wl,-O1" is this correct ? Aug 02 14:51:12 depends what you are building? Aug 02 14:51:29 qemu or qemu-native? Aug 02 14:51:46 woglinde: qemu Aug 02 14:52:14 for what machine you are building? Aug 02 14:52:42 woglinde: for x86 64 bit target machine Aug 02 14:52:51 hm ah Aug 02 14:52:53 ;( Aug 02 14:53:02 that is n't well tested Aug 02 14:53:37 is there noc configure option? Aug 02 14:53:48 woglinde: may be but I should not hit configure error Aug 02 14:54:08 I meant configure option to enabled 64 bit build Aug 02 14:54:23 Error: zlib check failed Make sure to have the zlib libs and headers installed. Aug 02 14:54:36 hm Aug 02 14:55:48 ./configure --cpu=x86_64 I am hitting the same error while doing manual compilation Aug 02 14:56:06 sorry compilation- configure Aug 02 14:56:06 hm Aug 02 14:56:08 DEPENDS = "zlib" Aug 02 14:56:12 is in Aug 02 14:56:15 yes Aug 02 14:57:16 hn Aug 02 14:57:25 where is the qemu-native recipe gone Aug 02 14:57:49 ah virtclass Aug 02 14:58:24 EXTRA_OECONF_append_virtclass-native = " --extra-cflags="-I${STAGING_INCDIR_NATIVE}"" Aug 02 14:59:28 anybody faced "Error: zlib check failed" while baking qemu recipe for x86 64 bit architecture? Aug 02 14:59:57 nope Aug 02 15:00:11 zlib check is suppose to be for target or host? Aug 02 15:00:11 not many people building qemu for target Aug 02 15:00:16 both Aug 02 15:00:21 humm Aug 02 15:00:46 ok in that case my host machine is x86 32 bit machine Aug 02 15:00:59 hoi stefan_schmidt Aug 02 15:01:12 hi woglinde Aug 02 15:01:27 all openjdk stuff now merged Aug 02 15:01:36 woglinde: yeah, just seen it Aug 02 15:01:42 Test build is on the way Aug 02 15:01:59 I am now looking at the jamvm problem Aug 02 15:02:00 woglinde: btw, I removed the empty dirs git-svn left over from llvm and ecj Aug 02 15:02:05 woglinde: great! Aug 02 15:02:52 enable libffi dont works Aug 02 15:03:25 stefan I commited somethin wrong? Aug 02 15:03:31 with git-svn Aug 02 15:03:37 woglinde: nothing wrong Aug 02 15:03:41 okay Aug 02 15:03:51 woglinde: it just seems as it keeps the dirs empty Aug 02 15:04:01 woglinde: or did you only remove the files and not the dir? Aug 02 15:04:21 hm here is all removed Aug 02 15:04:21 IIRC I had the same issue with git-svn myself Aug 02 15:04:57 woglinde: i think its a known limitation of git-svn. Not a problem. I just removed the dirs to make it clearer in svn Aug 02 15:05:21 okay Aug 02 15:19:28 hello stefan_schmidt Aug 02 15:19:37 anything new wrt dfu-util? Aug 02 15:20:01 ant_work: What new? Aug 02 15:20:02 morning Aug 02 15:20:09 hi kergoth Aug 02 15:20:24 ant_work: only some small fixes after 0.1, nothing that would bring a 0.2 on the table Aug 02 15:20:29 anything to test wrt forcing HID profile in CSR dongles? Last time upload failed :/ Aug 02 15:20:39 ok Aug 02 15:20:56 ant_work: no work for this from my side Aug 02 15:21:03 I have yet to test the release. I'll do Aug 02 15:21:30 I built some snapshot in may iirc Aug 02 15:36:57 hmm woglinde ..your commits are not appearing in cgit... Aug 02 15:37:15 7 hours ago Aug 02 15:37:28 is some server down? Aug 02 15:38:27 hmmm Aug 02 15:39:15 * kergoth wonders how much work it'd be to use the ast in bitbake to try to write back out a recipe after making a change to it, for things like khem's removal of do_stage () { autotools_do_stage } rather than using sed for it Aug 02 15:39:52 ant? Aug 02 15:40:04 ant yes linuxtogo has some problems Aug 02 15:40:11 hm look in stable Aug 02 15:47:59 ant_work its up Aug 02 15:51:58 woglinde, what time did you push those changes that ant_work was talking about? Aug 02 15:56:59 ka6sox they are the stable branch Aug 02 15:57:02 no panic Aug 02 15:57:53 its in git but not appearing to cgit? Aug 02 15:58:17 it is there Aug 02 15:58:26 ant didnt looked up in the stable branch Aug 02 15:58:50 http://cgit.openembedded.org/cgit.cgi/openembedded/log/?h=stable/2009 Aug 02 15:59:04 oh...whew. Aug 02 15:59:58 * ka6sox has trap set on words "server,git,tinderbox" Aug 02 16:01:27 *g* Aug 02 16:18:00 * ant_work hides Aug 02 16:18:11 ant stand up Aug 02 16:18:13 no problem Aug 02 16:18:23 false alert Aug 02 16:18:41 sorry ka6sox :P Aug 02 16:19:15 woglinde: it's so rare to have patches for stable that I didn't checked... Aug 02 16:19:38 yes Aug 02 16:19:51 ant_work...tis okay...about 13hrs ago I put new HD controller firmware in. Aug 02 16:24:50 I've read, before going to sleep....I just wondered about that :) Aug 02 16:25:01 is a LSI? Aug 02 16:27:48 hmmm my glib-2.24.1 build is not behaving Aug 02 16:28:10 configure: error: Could not find a glib-genmarshal in your PATH Aug 02 16:28:28 hi khem Aug 02 16:30:23 ant_work: hello Aug 02 16:33:25 hmm I guess glib depends on glib-native Aug 02 16:36:08 bbl, bye Aug 02 16:46:48 if you try to startx on a tftp/nfs mounted kernel, and it attempts to run a file in the build directory of OE which is not inside the rootfs of the system, what did I setup wrong? Aug 02 16:48:07 embedded root = /tftpboot/oe/gnome/ it attempts to access ~/oe/ODT/build/tmp/sysroots/i686-linux/usr/bin/mcookie when I startx obviously failing Aug 02 17:01:48 hello, does anyone have that fixed ncurses recipe? Aug 02 17:02:22 why does my alsa-utils is running my host ncurses5-config and no the OE one? Aug 02 17:02:33 not* Aug 02 17:13:42 hi Aug 02 17:20:04 marcompile: it might need http://patchwork.openembedded.org/patch/2248/ but I am not sure whether it required for my ncurses 5.7 recipe only or for the 5.4 one too Aug 02 17:20:04 Hello, is anyone else besides me interested in why does alsa-utils is running my hosts ncurses5-config and not the OE one? Aug 02 17:20:45 ensc|w, the problem is that alsa-utils is running my hosts ncurses5-config and that says it uses -ltinfo (that is not included in ncurses 5.4) Aug 02 17:21:24 It may work if I compile a ncurses 5.7 (with -ltinfo) on OE, but there is still a bug somewhere that alsa-utils is running my hosts ncurses5-config Aug 02 17:21:41 ensc|w, thanks for the info, anyway Aug 02 17:22:00 marcompile: do you have the correct ncurses5-config in the $PATH Aug 02 17:22:01 ? Aug 02 17:22:34 (somewhere at tmp/sysroots/x86_64-linux/usr/bin/armv5te-linux-gnueabi or so) Aug 02 17:22:59 let me check Aug 02 17:23:11 or only ncurses-config? Aug 02 17:24:28 i have that directory in my PATH but there is no ncurses* there Aug 02 17:25:39 old 5.4 recipe does not inherit binconfig :( Aug 02 17:26:22 so there it's expected that there is no ncurse5-config... Aug 02 17:29:21 so do you think I need to inherit binconfig on the ncurses 5.4 recipe or there is a good ncurses recipe somewhere? Aug 02 17:30:11 try my 5.7 from the maillist and give feedback... ;) Aug 02 17:30:52 is that pre-./configured ? Aug 02 17:30:55 gah some stuff still uses binconfig :-( Aug 02 17:31:02 I would not touch the 5.4 one; ncurses stuff is very porcelainish Aug 02 17:36:52 ensc|w, what ncurses recipe is not on git? Aug 02 17:37:39 ensc|w, that* Aug 02 17:37:49 marcompile: http://patchwork.openembedded.org/patch/2507/ Aug 02 17:38:54 andyj: seems something used native tools during build. if you figure out which package is asking for it that will help Aug 02 17:39:07 ensc|w: your patches look sane Aug 02 17:39:18 I will test them today or tomorrow Aug 02 17:39:46 khem: ok, thx Aug 02 17:40:46 khem: well my startx script has the path into my build system, not sure which package makes that I would think xserver/xorg? Aug 02 17:44:17 hehe yeah, that bug is everywhere it seems Aug 02 17:44:33 the link to mcookie Aug 02 17:44:49 khem: I built the angstrom-gnome-image-big if that is any help Aug 02 17:45:22 andyj: seems to be in every build i do Aug 02 17:46:18 buZz: so do you know how to go about fixing it after the fact? I just need to get this OS up and running so I can test some things on my dev board Aug 02 17:48:10 or even if you can point me in the right direction, right now I'm chasing down Xorg -configure spitting out that I am missing output drivers Aug 02 17:48:13 yeah it freaking easy Aug 02 17:48:20 just edit the startx file ;) Aug 02 17:48:35 find the line pointing to your build env, change to local filesystem, done Aug 02 17:48:42 lol well if I toast the mcookie check is that going to wreak havoc on the fs? Aug 02 17:48:43 the last part is probably right already Aug 02 17:48:54 dont do that, you need mcookie Aug 02 17:49:10 well there is no mcookie in my rootfs for the system only on my build system Aug 02 17:49:36 not even in /usr/bin ? Aug 02 17:50:10 it is on my image .. Aug 02 17:50:11 nope not there Aug 02 17:50:19 weird, should just be there Aug 02 17:50:33 i built angstrom-gnome-image some days ago, it had it in /usr/bin Aug 02 17:50:37 ya I think something went terribly wrong in my build and hope I don't have to do another 6 hour bitbake :) Aug 02 17:51:02 when running bitbake -b ncurses_5.7.bb i get ERROR: Parsing error data_fn virtual:native:/stuff/build/recipes/ncurses/ncurses_5.7.bb and fn /stuff/build/recipes/ncurses/ncurses_5.7.bb don't match Aug 02 17:51:17 andyj: bitbake saves a lot of work Aug 02 17:51:30 6 hours is pretty short for angstrom-gnome-image-big from nothing O_o Aug 02 17:51:40 buZz: ls /usr/bin | grep mcookie returns nothing Aug 02 17:51:52 ya I love the OE build setup Aug 02 17:52:28 well it might have taken longer than 6 hours, I left it running when I left work one day last week :) I also had a base-image already built Aug 02 17:53:51 ensc|w, any ideas? Aug 02 17:56:14 i'm setting the PREFERRED_VERSION_eglibc to 2.12 in my machine file, but a build continues to pull in eglibc 2.9. Has anyone seen this before? Aug 02 17:56:31 still pretty fast Aug 02 17:56:34 buZz: do you think I should just git pull and try bitbaking the image again? Aug 02 17:56:42 my first angstrom gnome build was about 2 days :D Aug 02 17:56:50 (on a p4 xeon with 8gb ram) Aug 02 17:57:02 andyj: well, it should really really be in your image already Aug 02 17:57:10 heh I'm running this from a VM with 2gb and only 2 of my four cores shared to it Aug 02 17:57:11 so just bitbake again, should be done a lot quicker Aug 02 17:57:16 :) Aug 02 17:57:36 so don't toast my tmp or do? Aug 02 17:57:45 this proc only has hyperthreading, but i have a second processor and heatsink waiting for me to mount them in the system with tie-wraps ;) Aug 02 17:57:54 no, should just work straight up Aug 02 17:58:03 buZz: thanks Aug 02 17:59:46 marcompile: which bitbake version? does it work with other recipess having BBCLASSEXTENDS="native"? Aug 02 18:00:38 bitbake version 1.8.18 Aug 02 18:01:29 do you think I need to upgrade to git's bitbake? Aug 02 18:01:56 marcompile: to use the -b option with BBCLASSEXTENDS: yes Aug 02 18:02:24 oic Aug 02 18:02:25 thanks Aug 02 18:06:21 WARNING, WARNING, WARNING: This is a Bitbake from the Unstable/Development 1.9 Branch. This software contains gaping security holes and is dangerous to use! Aug 02 18:06:26 lol Aug 02 18:06:43 *fear* Aug 02 18:09:59 marcompile: if it's in caps and repeated three times... it must be serious! Aug 02 18:10:15 it is Aug 02 18:12:58 well, this bitbreak 1.9 does not work at all Aug 02 18:13:08 well, not for me Aug 02 18:16:16 rphillips: which distro ? Aug 02 18:16:58 angstrom Aug 02 18:17:09 rphillips: you should use a proper override like PREFERRED_VERSION_eglibc_local = "2.12" and PREFERRED_VERSION_eglibc-initial_local = "2.12" Aug 02 18:17:31 ah, thank you Aug 02 18:17:38 Distro's lock versions for a reason Aug 02 18:18:53 yeah, it makes sense Aug 02 18:19:23 ensc|w, sorry to bother you again but bb 1.9 does not work for me, what do you suggest me to do to compile your new ncurses 5.7 recipe? Aug 02 18:36:01 ensc|w, if I just bitbake ncurses, it builds the old one Aug 02 18:42:24 marcompile: it might have default_preferences set to -1 Aug 02 18:46:36 what should I set default_preferences to? Aug 02 18:47:50 if I bitbake ncurses-5.7, I get ERROR: Nothing PROVIDES 'ncurses-5.7' Aug 02 18:55:36 remove anything like DEFAULT_PREFRENCES from the recipe Aug 02 18:55:53 and also make sure that your distro does not choose old one Aug 02 18:56:03 hmm better is to overrride it in your local.conf Aug 02 18:57:52 PREFERRED_VERSION_ncurses_local = "5.7" Aug 02 19:33:31 hrm. Aug 02 19:34:46 khem, thanks Aug 02 19:34:53 its working nao Aug 02 19:36:21 03Michael 'Mickey' Lauer  07org.openembedded.dev * ra93982c1a4 10openembedded.git/recipes/libgee/ (libgee.inc libgee_0.5.2.bb libgee_git.bb): libgee: refactor into common inc file and add 0.5.2 release Aug 02 19:42:35 hmmm relocate bbclass stuff is borked on mac Aug 02 19:45:40 does anyone actually get to build something with OE on mac? Aug 02 19:45:52 it still has lots of issues Aug 02 19:46:31 a few annoyances went away with the siteinfo change, since it doesn't try to get at darwin info, but there's plenty remaining. mktemp syntax, gnu specific sed issues, etc Aug 02 20:11:55 in busybox how can i edit it so its not built with everything? Aug 02 20:12:11 its 60m and i need it down to 1.7m Aug 02 20:30:46 when sending a series of patches against a single .bb file, is there a policy of (not) bumping ${PR} ? I'm thinking of adding a level of dot for intermediary patches Aug 02 20:41:27 yann, I think if you changed something you just bump the revision by some value Aug 02 20:41:57 ok, so bumping by 0.1 would be ok then Aug 02 20:45:29 is there any feature that adds a package to RDEPENDS depending on FILES_${PN}- ? I'm hit by a strange behaviour looking like just that Aug 02 21:03:17 khem: ping Aug 02 21:05:11 yann, well, there's code that adds rdepends based on what shared libraries your binaries are linked against Aug 02 21:06:40 that would fail Aug 02 21:06:47 kergoth: well, git-remote-http hardly depends on any lib from the rest of git, there is no such lib Aug 02 21:06:56 if one builds with other libc? Aug 02 21:07:09 there are a few other magic rdepends bits, i.e. perl modules, but that's about it Aug 02 21:08:07 ant___: hey Aug 02 21:08:14 hi khem Aug 02 21:08:26 I'm doing second arch testing v5 patches Aug 02 21:08:34 kergoth: hm, I have split all of perl in another package, so that one is out as well Aug 02 21:08:38 ant___: thx Aug 02 21:08:39 I see some strange paths.. Aug 02 21:08:39 http://paste.debian.net/82065/ Aug 02 21:09:18 ant___: like what Aug 02 21:09:24 kergoth: I'm at a loss about selecting one of oe and bitbake for the bugreport - I guess oe will fit... Aug 02 21:11:31 khem, http://pastebin.com/mCYxfXbR Aug 02 21:12:02 klcc is creating /bin Aug 02 21:12:14 hm Aug 02 21:13:55 ant___: klcc is a cross stuff is it ? Aug 02 21:14:16 http://pastebin.com/BcjUTa9g Aug 02 21:14:22 this is about angstrom- Aug 02 21:14:33 klcc is a crosscompiler, yes Aug 02 21:14:50 well, is like a wrapper around gcc Aug 02 21:16:25 and ld Aug 02 21:18:23 the second paste looks ok Aug 02 21:18:27 there is no problem there Aug 02 21:18:37 yes, angstrom links are ok Aug 02 21:18:58 g77->gfortran seems broken Aug 02 21:19:00 let me see Aug 02 21:19:12 so what is the problem Aug 02 21:19:37 no evident problems, just stale symlink Aug 02 21:19:51 and too many /bin in sysroots Aug 02 21:20:03 but it is building Aug 02 21:23:54 about gfortran ? Aug 02 21:23:58 yes I will fix that Aug 02 21:24:14 well we have to get used to the bins Aug 02 21:24:31 atleast its clear with the preceding dirs Aug 02 21:25:51 about klibc legacy staging (klcc), now the line is: Aug 02 21:25:52 cp '${STAGING_DIR_TARGET}/bin/klcc' '${TOOLCHAIN_PATH}/bin/${TARGET_PREFIX}klcc' Aug 02 21:26:13 how would it translate after all patches are submitted? I'm a bit confused... Aug 02 21:26:37 khem, the /bin comes from there Aug 02 21:27:04 didnt I fix it in one of the patches ? Aug 02 21:27:22 hm..this is in my version, with your patches Aug 02 21:27:29 v5 Aug 02 21:27:35 ok thats correct Aug 02 21:27:49 so the klcc should be there along with gcc and all Aug 02 21:27:58 I suppose Aug 02 21:28:16 is fine then Aug 02 21:29:00 yes, but how can we INSTALL it without legacy staging? Aug 02 21:29:49 add same to do_install_append Aug 02 21:30:23 same export ? Aug 02 21:30:33 export INST=${STAGING_DIR_TARGET} Aug 02 21:31:36 ok, I'll try once armv4 builds ends Aug 02 21:32:22 ant___: I can provide a patch Aug 02 21:32:32 tell me where does klcc existi in your build tree Aug 02 21:34:28 khem, /oe/build/tmp/sysroots/armv5te-angstrom-linux-gnueabi/bin Aug 02 21:34:53 thats staged Aug 02 21:35:01 I want to know where is it in you build Aug 02 21:35:18 search in your klibc-x.x.x dir Aug 02 21:35:25 in build tre Aug 02 21:35:42 moment, rm_work here Aug 02 21:36:57 eh Aug 02 21:39:32 /oe/build/tmp/work/collie-angstrom-linux-gnueabi/klibc-1.5.18-r17.2/klibc-1.5.18/klcc Aug 02 21:42:57 khem: built now for armv4 in /oe/build/tmp/work/collie-angstrom-linux-gnueabi/klibc-1.5.18-r17.2/klibc-1.5.18/klcc/klcc Aug 02 21:43:35 and where is ls Aug 02 21:44:11 khem: I don't follow you Aug 02 21:44:19 looking anywayù Aug 02 21:45:56 there should be ls binary somewhere in the klib build tree Aug 02 21:47:56 anyway try this patch and see if it works http://pastebin.com/VtWwzcZb Aug 02 21:49:09 /oe/build/tmp/work/collie-angstrom-linux-gnueabi/klibc-1.5.18-r17.2/klibc-1.5.18/usr/utils Aug 02 21:49:14 there are utils Aug 02 21:49:37 cat, ls, ... Aug 02 21:50:24 khem: thx, I'll see what happens. Aug 02 21:50:39 still NOTE: Running task 3287 of 5022 Aug 02 21:50:46 half an hour Aug 02 21:51:01 * ant___ heads down for a beer Aug 02 21:54:42 ant___: beer - what a good idea :} Aug 02 21:56:32 hm, does not help the compilation, though :> Aug 02 22:05:41 yann: whats the issue Aug 02 22:06:10 khem: never mind, a typo in an envvar :) Aug 02 22:06:28 finally, it appears that the beer *did* help somehow :) Aug 02 22:09:52 hm, no info about building locale packages in the oe usermanual:( Aug 02 22:10:33 neither in the wiki Aug 02 22:15:52 write one if you can I will be happy Aug 02 22:16:30 khem: unfortunately I don't really see how to solve my problem - an app with l10n files in a custom dir... Aug 02 22:17:01 khem: build seems freezed NOTE: Running task 3687 of 5022 (ID: 696, /oe/openembedded/recipes/e2fsprogs/e2fsprogs_1.41.9.bb, do_populate_sysroot) Aug 02 22:17:01 NOTE: package e2fsprogs-1.41.9-r24: task do_populate_sysroot: Started Aug 02 22:17:50 ant___: check whats happening with top Aug 02 22:17:57 ah, top reveals qemu-arm Aug 02 22:18:50 http://pastebin.com/9zm1Z1HA Aug 02 22:19:13 seems stopped after first locale Aug 02 22:19:32 611 andrea 20 0 156m 53m 1552 R 100 0.7 0:08.92 qemu-arm Aug 02 22:21:30 no, other locales generated before Aug 02 22:21:32 anyone with a hint about generating -locale- package from an arbitrary dir ? Aug 02 22:24:11 ant___: thats strange Aug 02 22:24:26 ant___: qemu-arm means its doing some locale stuff I guess Aug 02 22:25:06 khem, small indices Aug 02 22:25:16 /oe/build/tmp/work/armv4-angstrom-linux-gnueabi/e2fsprogs-1.41.9-r24/temp ─ Aug 02 22:25:22 ant___: kill it and start again Aug 02 22:25:38 log.sysroot_stage_all.436 -> Aug 02 22:25:40 Found la files:. Aug 02 22:25:59 Found la files: Aug 02 22:26:00 it could be stuck in some race Aug 02 22:26:34 :/ I was so happy first build went through w/out races.. Aug 02 22:27:14 -j5 4 BB threads Aug 02 22:27:40 I have the log fwiw Aug 02 22:29:22 btw this ubuntu could freeze my host yesterday :/ Aug 02 22:29:35 perhaps is some buildhost bug Aug 02 22:30:41 probably Firefox Aug 02 22:41:08 hm, FILES_${PN}-locale from bitbake.conf was promising, but finally not - I probably missed something Aug 02 23:37:28 do I have to subscribe to oe-dev to be allowed to post ? Aug 03 01:14:48 * kergoth finally gets around to resurrecting his IMAGE_FEATURES patch Aug 03 01:16:36 * kergoth has way, way too many branches with proofs of concepts and things that never got merged.. starting to work through them Aug 03 01:58:02 kergoth: how about the per package dbg file feature Aug 03 02:08:44 khem, it's already in, has been for ages, afaicr, i expect no one knows about it though :) INHERIT += "package_dbg" should work Aug 03 02:09:20 * kergoth just watched bitbake dump like 500 "eeek, this shouldn't happen" errors :( **** ENDING LOGGING AT Tue Aug 03 02:59:56 2010