**** BEGIN LOGGING AT Tue Nov 11 02:59:57 2008 Nov 11 03:53:49 hmm Nov 11 03:54:12 my configure script looks for a header in the directory where the sources are, in my local directory Nov 11 03:54:24 and it seems I get a sanity error because of that Nov 11 03:54:28 not sure though if that is really it Nov 11 03:54:43 ERROR: This autoconf log indicates errors, it looked at host includes. Rerun configure task after fixing this. Nov 11 03:55:01 Path was '/opt//openembedded/build-angstrom-i386//work/i386-angstrom-linux-uclibc/mediatomb-static-0.11.0+svn20081111-r3-linux-uclibc-i386/mediatomb' Nov 11 03:55:21 weird, heh? Nov 11 03:56:13 or is the path that this message is printing - the configure location? Nov 11 03:57:08 that path is the working directory where the sources were unpacked for the build Nov 11 03:57:42 ooh ok, so the error does not actually say what host includes are being looked for - i.e. what triggers it Nov 11 03:57:50 bitbake seems halfbaked so far, but i'm using it on a non-linux host. giving it a shot on lunix now to see if i have better luck that way Nov 11 03:58:09 I think I found it, it called a -config script from the system Nov 11 03:59:12 well, if you go to the dir one up from that path, you'll find there is a dir called temp in addition to mediatomb. in temp youll find a log.do_configure.XXXXX which will show you waht hapened when configure ran and what the error is Nov 11 03:59:46 actually, configure runs through, it seems that bitbake analyzes the config.log file after the configure run has finished Nov 11 03:59:54 but indeed I could see the path in the log Nov 11 03:59:54 thx Nov 11 04:00:03 you can run configure manually in the dir shown and play with the options until it works, then go edit the corresponding .bb file to give configure right switches when run from bitnake Nov 11 04:00:07 would have been helpful if bb would print the path it did not like Nov 11 04:00:37 I know, I know, thx, I got it working now Nov 11 04:00:41 I was only confused about the error message Nov 11 04:00:52 I thought it was telling me the path it did not like :) Nov 11 04:00:56 which was not the case :> Nov 11 04:01:28 nah, it just tells the working path and then you get to go digging Nov 11 04:01:46 well, too bad :) Nov 11 04:01:57 i quickly started working with teo terminal windows. one tio rerun bitbake and another to move through its tree and fix up the bits that are wrong Nov 11 04:01:59 I should have a closer look at this Nov 11 04:02:18 should not be too hard to add that one extra message, printing the path it did not like Nov 11 04:02:55 I usually have a shitload of terms open :) recipes, logs, devshells, etc Nov 11 04:04:22 are you fairly familiar with bitbake internals? Nov 11 04:04:34 not too much, but I know python Nov 11 04:04:40 or just goingto go diving in to make that error more useful Nov 11 04:04:56 I am sort of familiar with bbclasses Nov 11 04:05:04 I guess it may be somewhere in insane.bbclass Nov 11 04:06:42 if you have any ideas regarding the question i asked earlier, feel free to chime in Nov 11 04:06:54 let me see the backlog Nov 11 04:07:06 I was coding, so was not following the channel Nov 11 04:08:11 hmm, sorry Nov 11 04:08:13 don't know Nov 11 04:08:26 you want to edit the script and pass to a further stage? Nov 11 04:08:54 which you can't because it recreates the script and reruns the unfinished stage that fails with the script that it creates? or something like that? Nov 11 04:09:05 exactly Nov 11 04:09:33 cant figure out what generates the script Nov 11 04:09:37 I understand the problem, but have no idea how to work around that Nov 11 04:09:51 why don't you fix the .bb recipe so that stage works out? Nov 11 04:10:00 especially if you already know what you have to change? Nov 11 04:10:02 if i knew that, i could either flip it off for a quick run with the edited script or change the metadata from which the script is built Nov 11 04:10:51 i don't see anything in the bb recipe about cp but the script contains a line with cp -t and the -t only exists on linux yet bitbake is suppossed to work on other host platforms Nov 11 04:11:16 maybe that cp comes from one of the bbclasses that your recipe is inheriting? Nov 11 04:11:25 I'd search for that in the classes directory Nov 11 04:11:37 actually, I am pretty sure that this is the case Nov 11 04:12:32 i'm not too familiar with the classes. is there an easy way to find out the entire inheritance chain of classes Nov 11 04:12:51 i'm finding that oe has decent docs but bitbake is barely documented Nov 11 04:15:00 not really, can't say much about overall inheritance Nov 11 04:15:58 hmm, my grep shows mostly -dPR or stuf flike that Nov 11 04:15:59 but no -t Nov 11 04:16:16 in relation to cp that is Nov 11 04:18:21 autoconf-native-2.61-r1, run.do_stage.XXXXX line 760: cp -fpPR -t "$to" "$from"/* Nov 11 04:18:50 can be rewritten as cp -fpPR "$from"/* "$to" i believe Nov 11 04:19:30 check autotools.bbclass Nov 11 04:19:38 I have cp -fpPR "$from"/* "$to" Nov 11 04:19:42 but I did not update recently Nov 11 04:19:49 so either someone added that line Nov 11 04:19:55 or I don't know :) Nov 11 04:27:02 crap, almost 6 a.m. Nov 11 04:27:04 time to leave Nov 11 04:27:15 nite Nov 11 04:34:24 I am sure this has been asked, I am trying to clone the git repository behind a corp. firewall. It doesn't appear the http protocal is enabled. Is this possible or is there a snapsnot I download somewhere. Nov 11 04:39:03 thanks Nov 11 04:50:13 I thought we had one Nov 11 04:50:17 but I could be wrong Nov 11 04:50:40 if you have issues with git, likely you will have trouble with other ports used to fetch source though Nov 11 04:52:26 well, I noticed the welcome mesg to this #irc channel mentioned the ro repo at http://repo.or.cz/r/openembedded.git Nov 11 04:53:03 I just try it via http protocal (at home) and it works. I will try this at work tomorrow Nov 11 04:53:09 what machine are you going to build for? Nov 11 04:53:52 I am playing. but the end goal with be cross-builds for x86, x86_64, and ia64 Nov 11 05:29:26 ah, found it i believe. line 175 of autotools.bbclass has the offending cp template. fixed that and reruning the bitbake Nov 11 05:33:20 yep, looks like it working Nov 11 05:34:06 Jin|away: thanks for mentioning the bbclasses. with some digging i found and fixed the offending line and all is going well from here Nov 11 05:34:51 also i should be able to patch up other stuff from here on out more easily Nov 11 05:35:35 if all goes well, i'll have a working oe on mac os x Nov 11 05:36:41 Avengence: bitbake is documented -> http://bitbake.berlios.de/manual/ Nov 11 05:36:56 kergoth__: read that manual. it doesnt get into what i needed to know Nov 11 05:37:26 ahh, so when something doesnt include what you happen to need, that makes it 'barely' documented Nov 11 05:37:27 i see Nov 11 05:37:36 i was looing for something more like dev docs than user docs. something that would roughly explain how it all comes together Nov 11 05:39:21 usage isnt even fully explained as far as various options, much less waht happens under the hood during a build Nov 11 05:41:00 "how it all comes together"? i fail to see what's missing here. you want documentation of implementation details? read the code Nov 11 05:41:35 * kergoth__ shrugs Nov 11 05:43:19 yeah, that is kinda want i want without having to actually read all the code. some paragraphs on the overall structure with some diagrams or flow charts would go a long way Nov 11 05:43:53 surem the code explains it all, but do i really have the time to read and comprehend all the code in every project i pick up for a brief time Nov 11 08:25:28 morning Nov 11 08:25:58 morning Nov 11 10:01:20 hi blindman Nov 11 10:01:22 hi lrg Nov 11 10:01:46 hey pb__ Nov 11 10:04:56 Hi Nov 11 10:05:37 I have a problem while building an x11-image The Gtk+ 2. 14.2 build fails Nov 11 10:06:12 I tried on two differents systems ( One Ubuntu64 and one with Standard Ubuntu 32 , both 8.04 version ) Nov 11 10:06:19 The error is : Nov 11 10:06:43 /home/drakou/OE/tmp/cross/armv5te/lib/gcc/arm-angstrom-linux-gnueabi/4.2.4/../../../../arm-angstrom-linux-gnueabi/bin/ld:.libs/libgdk_pixbuf-2.0.ver:2: ignoring invalid character `\001' in script Nov 11 10:06:43 | /home/drakou/OE/tmp/cross/armv5te/lib/gcc/arm-angstrom-linux-gnueabi/4.2.4/../../../../arm-angstrom-linux-gnueabi/bin/ld:.libs/libgdk_pixbuf-2.0.ver:2: syntax error in VERSION script Nov 11 10:16:00 trying to compile ncurses ; the process hangs at Running tic to install /tmp/testfs//usr/share/terminfo ... Nov 11 10:16:34 it remains here for a long period of time and the filesystem size does not increase at all Nov 11 10:26:55 Did anyone there managed to build gtk-2.14.2 ? Or build a full x11-image without problems ? I am new to OE, I am trying to start using it but I am really blocked with that problem. Nov 11 10:27:37 One mont ago I tried to use it and I get exactly the same error... So I gave up ,( I didn't have time to solve the problem ) Nov 11 10:31:34 drakou: maybe you should try a different version of gtk+ Nov 11 10:32:14 otherwise, I think you will probably need to debug this problem for yourself. Nov 11 10:32:27 Yes, but, How can I select the Gtk version ? I am trying to build the angstrom2008 distro ... I am trying first a standard configuration, and it's not working Nov 11 10:32:44 Nobody uses gtk+ ? it's impossible ... Nov 11 10:32:45 drakou: the easiest way to experiment is to build different versions by hand: bitbake -b Nov 11 10:32:59 ok Nov 11 10:33:15 plenty of people use gtk, but maybe not the particular version you are building Nov 11 10:34:07 openmoko seems to be using gtk+-fastscaling_2.10.14, and that version does build ok for me Nov 11 10:35:04 yes ... I guess, but it's the standard Angstrom 2008 distro that I am using... Maybe I shouldn't. I need a system for a digital photo/multimedia frame Nov 11 10:35:37 What distro do u use ? Nov 11 10:46:29 openmoko is its own DISTRO. I use "minimal" for some other things but I think I have a set of custom preferred versions for those. Nov 11 10:47:04 ERROR: Nothing PROVIDES 'openembedded/packages/gtk+/gtk+-fastscaling_2.10.14.bb' Nov 11 10:47:44 The prefered version number are defined in the distro or in the image ??? If want to define a new set of prefered versions, do I have to define a new distro ? Nov 11 10:54:56 OK -> Prefered versions are in the distro :) But I still don't undestrand why the 2.14.2 version of Gtk+ is not building ... I'll try other versions Nov 11 10:55:53 drakou: I also got this gtk error about a month or so ago. It builds for me now with recent OE. Nov 11 10:57:35 I got OE from repos yesterday, so it's a recent one :) Nov 11 10:58:12 The problem is when creating the libgdk_pixbuf-2.0.ver file Nov 11 11:04:09 I have exactly the same error with gtk + 2.12.11 Nov 11 11:05:14 drakou: I got the same error with all version I tried too. It may be something earlier in the build Nov 11 11:06:54 I really would like to use OE but I have no time to spend more than a week solving that problem ... What is our build computer distro ? ( Maybe there's a problem with the Ubuntu 8.04 I am using ) Nov 11 11:11:57 florian: good morning Nov 11 11:12:08 good morning Nov 11 11:12:12 pb__: hey Nov 11 11:12:17 florian: moin Nov 11 11:12:44 pb__: talking with woglinde he suggested me to ask you about this: http://rafb.net/p/Os3YTs32.html Nov 11 11:19:56 pb__: pls help on this too (oebug 4784) http://rafb.net/p/Flsgou13.html Nov 11 11:28:14 hi ant_work Nov 11 11:29:07 I don't think I can immediately help you with either of those things. Nov 11 11:29:22 the first one suggests that gcc's shlibs data has gotten clobbered, or written out wrong, but I don't know why that would happen. Nov 11 11:29:30 no idea on the latter Nov 11 11:35:09 Hello; How are people doing? Nov 11 11:35:27 I'm still sick :) Nov 11 11:35:30 * otavio goes to check his mail backlog :-) Nov 11 11:35:31 otavio just fine, you? Nov 11 11:35:38 zecke: really? flu? Nov 11 11:35:46 dcordes: doing fine too. Nov 11 11:35:53 otavio: not that bad, just a bad cold... coughing badly Nov 11 11:35:58 birds flu? Nov 11 11:36:17 zecke: oh right; i wish you get better ... Nov 11 11:36:36 zecke: i can bet that you program more and better when not sick ;-) Nov 11 11:37:33 otavio you think it's a conspiracy from moko? Nov 11 11:38:26 I found where is the problem with GTK+ (any version ) not building . I compared the logs between OE and a normal build on my computer . I will need a little of help to know why it's happening Nov 11 11:39:02 With A clean build, on my computer the command line is Nov 11 11:39:03 /usr/bin/nm -B .libs/gdk-pixbuf.o .libs/gdk-pixbuf-animation.o .libs/gdk-pixbuf-data.o .libs/gdk-pixbuf-io.o .libs/gdk-pixbuf-loader.o .libs/gdk-pixbuf-scale.o .libs/gdk-pixbuf-simple-anim.o .libs/gdk-pixbuf-scaled-anim.o .libs/gdk-pixbuf-util.o .libs/gdk-pixdata.o .libs/gdk-pixbuf-enum-types.o pixops/.libs/libpixops.a | sed -n -e 's/^.*[ ]\([ABCDGIRSTW][ABCDGIRSTW]*\)[ ][ ]*\([_A-Za-z][_A-Za-z0-9]*\)$/\1 \2 \2/p' | /bin/sed 's/.* //' | sort | uniq > . Nov 11 11:39:03 libs/libgdk_pixbuf-2.0.exp Nov 11 11:39:15 And with OE : Nov 11 11:39:18 | arm-angstrom-linux-gnueabi-libtool: link: /home/drakou/OE/tmp/cross/armv5te/bin/arm-angstrom-linux-gnueabi-nm -B .libs/gdk-pixbuf.o .libs/gdk-pixbuf-animation.o .libs/gdk-pixbuf-data.o .libs/gdk-pixbuf-io.o .libs/gdk-pixbuf-loader.o .libs/gdk-pixbuf-scale.o .libs/gdk-pixbuf-simple-anim.o .libs/gdk-pixbuf-util.o .libs/gdk-pixdata.o .libs/gdk-pixbuf-enum-types.o pixops/.libs/libpixops.a | sed -n -e 's/^.*[ ]\([ABCDGIRSTW][ABCDGIRSTW]*\)[ ][ ]*\([_A-Za Nov 11 11:39:18 -z][_A-Za-z0-9]*\)$/  /p' | /bin/sed 's/.* //' | sort | uniq > .libs/libgdk_pixbuf-2.0.exp Nov 11 11:40:15 Instead of writing \1 \2 \2, there is the caracters with ascii code 1 ,2 and 2 ... so this is the buggy line Nov 11 11:40:43 How could I solve that ?? Nov 11 11:40:46 hi zecke Nov 11 11:40:52 get well soon :-} Nov 11 11:41:43 drakou: that's very odd. it looks like the oe version of the command is being passed through "echo -e" or some such filter. Nov 11 11:41:55 what does the makefile itself look like? Nov 11 11:42:28 I can't find that command neither in the Makefile or in Makefile.am ... I don't know where this is generated Nov 11 11:42:53 oh, maybe libtool is generatingit Nov 11 11:43:10 you'll have to find the corresponding libtool call, then inspect the libtool script to see what's up Nov 11 11:44:25 dcordes: can be. I hope zecke can find their dare and survive ;-) Nov 11 11:44:27 zecke: :P Nov 11 11:45:10 I've fixed the qemu-native build failure locally; any objection in pushing it? Nov 11 11:45:34 the problem must be in libtool I guess ... maybe the macro used is in aclocal