**** BEGIN LOGGING AT Tue Jun 23 02:59:58 2015 Jun 23 07:37:25 GOOD MORNING Jun 23 08:13:30 hi mckoan Jun 23 08:19:18 mornin Jun 23 08:28:30 meh I am having huge issues getting this c++ program compiled. For some reason it is unable to find the include path that I believe is correctly spesified. If I take the --sysroot path and concatinate that with my -I path and concatenate that with my #include directive the file does indeed exist. Jun 23 08:28:43 But g++ says it does not exist Jun 23 08:28:56 How can I debug this further? Jun 23 08:30:22 I am aware that include paths are normally a headache, but I know how it works and it should be working. Jun 23 08:31:36 Those paths that I concatenate come from the actual cmd line that bitbake prints on the failure. Jun 23 08:34:17 mmm wait let me use "" instead of <> Jun 23 08:38:10 no effect Jun 23 08:55:33 g++ -v says: ignoring nonexistent directory "/usr/include/opendnp3" Jun 23 08:55:58 thats a lie Jun 23 08:56:59 ./home/developer/w/mongoose/setup-scripts/build/tmp-angstrom_v2014_12-glibc/sysroots/mongoose-wb/usr/include/opendnp3 Jun 23 09:15:10 for some reason "/home/developer/w/mongoose/setup-scripts/build/tmp-angstrom_v2014_12-glibc/sysroots/mongoose-wb/usr/include/opendnp3" is not the same as "usr/include/opendnp3" Jun 23 09:16:48 it is like the --sysroot is not used? Jun 23 09:17:18 I have tried -I ${STAGING_DIR}${includedir}/opendnp3 Jun 23 09:17:21 wont work Jun 23 09:23:34 is there currently a way to do bbappends on a per image basis; imagine you want support for making FAT partitions in development image (via busybox bbappend), but you don't want support for creating FAT partitions in the production image Jun 23 09:24:44 is there a way to specify that the bbappend should only be used in the case building the development image? Jun 23 09:32:44 instead of inserting the behaviour into a bbappend, why not use a "include" directive and then use anonymous python to include different behaviour depending on the build type Jun 23 09:35:12 http://hastebin.com/qitibopeja.mel Jun 23 09:35:56 actually, just use anonymous python inside your bbappend to either do the work or not Jun 23 09:36:21 forget what I said about include directive Jun 23 09:53:11 koen: I found the problem. gcc documentation states: --sysroot=dir Jun 23 09:53:11 Use dir as the logical root directory for headers and libraries. For example, if the compiler normally searches for headers in /usr/include and libraries in /usr/lib, it instead searches dir/usr/include and dir/usr/lib. Jun 23 09:53:26 That is not true Jun 23 09:53:49 I have to manually do this -> CXXFLAGS_prepend = "-I${STAGING_DIR_HOST}${includedir}/opendnp3 " Jun 23 09:54:03 instead of this -> CXXFLAGS_prepend = "-I${includedir}/opendnp3 " Jun 23 15:30:09 I'm trying to _append to do_kernel_compilemodules, is this possible as it doens't seem to be working Jun 23 15:57:06 ok, I've got it; my function was the problem (suprise suprise) Jun 23 16:47:00 how does BitBake know source code has changed? I was under the impression it only checked for file names and not the content itself, but I changed a line in a patch and BitBake detected the change Jun 23 16:48:11 checksums, probably Jun 23 16:48:40 neverpanic: do you know where in the source that's done? Jun 23 16:48:59 Unfortunately not for patchfiles; I'm only familiar with the fetching part Jun 23 16:49:37 i can't seem to find it **** ENDING LOGGING AT Wed Jun 24 02:59:58 2015