**** BEGIN LOGGING AT Sun Jul 13 02:59:59 2014 Jul 13 14:31:00 I'm trying to slim down the size of my image. Currently I have both eglibc-binary-localedata-en-us and eglibc-binary-localedata-en-gb installed. Jul 13 14:32:00 But it is not clear to me wether I need both, just one, or none of them. Anyone knows what they actually contain ? They are relatively big packages Jul 13 14:32:58 I'm guessing they contain language-specific libc message strings Jul 13 15:55:59 hey everybody. I want to compile a openembedded -distro image, which fails during build (do_install) of apr-1.5.1. paste can be seen here : http://bpaste.net/show/459878/. Anybody got any ideas? Perhaps problem is known and solved already ? apr-1.5.3 is available, dont know why 1.5.1 is being autoselected for that image Jul 13 16:02:39 I just reakized theres a patch adressing exactly that failing test thats build at runtime Jul 13 16:02:51 s/rekized/realized Jul 13 22:43:39 Changes in ROOTFS_POSTPROCESS_COMMAND wont trigger rebuilds Jul 13 22:45:11 Is there a vardep missing ? Jul 13 23:18:09 kroon: iirc there was patch for that on ML Jul 13 23:19:03 JaMa, ah, thanks, will check Jul 13 23:23:37 adding ROOFS_POSTPROCESS_COMMAND to do_rootfs[vardeps] would make sense Jul 13 23:25:05 but my guess is most people add a function, and append the function to ROOTFS_POSTPROCESS_COMMAND Jul 13 23:25:31 so changes in the actual function wont get detected, i think Jul 13 23:26:16 but i guess i can add my own rootfs[vardeps] entries in the image recipe Jul 13 23:26:17 that's true. bitbake can do shell dependency analysis, but it won't know to do so with ROOTFS_POSTPROCESS_COMMAND unless it was flagged as a function Jul 13 23:27:01 i expect if you were to set ROOTFS_POSTPROCESS_COMMAND[func] = "1" and also add it to the do_rootfs vardeps, then changes to the function should cause do_rootfs to rerun Jul 13 23:27:04 might be worth a try Jul 13 23:27:29 kergoth, ah, sounds good, lemme try Jul 13 23:34:32 kergoth, indeed, seems to work Jul 13 23:34:41 cool, glad to hear it Jul 13 23:35:07 func is the internal flag that bitbake sets when you use foo () { } syntax to define a shell function Jul 13 23:35:23 so by setting it manually, we get it to treat it the same way, and parse the shell for deps Jul 13 23:36:24 maybe one of us should respond on the ml that it might be a good idea to do this Jul 13 23:50:06 kergoth, I responded on the ml Jul 13 23:55:33 hmm, that trick seems pretty rarely used though Jul 14 00:44:36 kroon: true, but it's also pretty rare to have variables that contain shell code but which aren't shell functions :) **** ENDING LOGGING AT Mon Jul 14 02:59:59 2014