**** BEGIN LOGGING AT Tue May 26 02:59:58 2015 May 26 08:22:25 morning all May 26 09:42:15 bluelightning: hello May 26 09:42:22 hi Noor May 26 09:42:45 bluelightning: need little help regarding your patch on devtools :) May 26 09:42:54 Noor: sure, what's up? May 26 09:42:55 https://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=fbfc06a969200e582a059c9943e6fd17aca70e30 May 26 09:43:12 ah right, the recent one May 26 09:43:16 bluelightning: can you just guide me or point me how to use this functionality May 26 09:43:38 we want to change the say defconfig May 26 09:43:53 what will be its command to devtool May 26 09:44:04 w.r.t your patch May 26 09:44:15 Noor: it's not really for inputs to the build like a defconfig, more for files that end up on the target system May 26 09:44:35 aaahhhh May 26 09:44:42 I got a hint to that May 26 09:44:51 but just wanna make sure May 26 09:46:03 hmm, I see I missed adding a description to the help text, that's very poor of me May 26 12:32:41 Noor: there's a reason i said, in our bug, that we'd want to add a new sub-command if we want that type of functionality.. May 26 12:32:44 * kergoth yawns May 26 12:33:35 kergoth: yeah May 26 12:34:46 kergoth: I just was making sure that I am getting correct understanding ..... I was thinking that its something related to configs of the packages on the target ...... just wanna confirm that May 26 13:16:23 is there some howto on implementing your own packaging format in bitbake? May 26 13:21:31 jaeckel: not that I know of... it's not something that people sit down to do very often May 26 13:21:49 jaeckel: you'd probably need to look at how the other packaging formats have been implemented May 26 13:22:24 yeah, that's what I'm currently doing and it seems it's not that easy to extend May 26 13:22:28 sadly I know there are a bunch of places outside of the packaging classes that have specific knowledge of the implementation of various packaging systems, which I know is not ideal May 26 13:26:54 is there a reason why it's not possible to use tar as rootfs class? May 26 13:28:02 jaeckel: yes, because we'd basically have to implement a package manager on top of tar (with dependency handling, etc.) in order for that to work May 26 13:28:22 and in the end that wouldn't really buy us much May 26 13:29:11 well it would give the possibility to have images without all this stuff... May 26 13:29:45 this stuff as in package manager, dependency handling, etc. May 26 13:30:43 e.g. I don't want to have all this in my resulting image, what would you propose to use as rootfs type? May 26 13:31:05 jaeckel: but which package manager you choose to assemble the package is separate from whether or not you have package management in the final image May 26 13:31:15 assemble the image, I meant May 26 13:31:32 uhm May 26 13:31:42 package management being in the final image is controlled by whether or not "package-management" is in IMAGE_FEATURES May 26 13:32:52 which variable defined the package management to use to assemble an image? IMAGE_PKGTYPE? May 26 13:33:00 ...defines... May 26 13:34:21 IMAGE_PKGTYPE, but in normal usage that is not set by the user, it's set by the package class (the package class being selected by PACKAGE_CLASSES) May 26 14:49:13 bluelightning: how would you propose to implement a new output image format? it's basically a gpg signed tarball of the rootfs that I want to create, so I thought about extending image_types.bbclass, but that also doesn't seem to be that easy... May 26 14:49:43 jaeckel: that should be pretty easy, and unlike packaging formats is commonly extended May 26 14:50:00 jaeckel: just set IMAGE_CMD_ to the command to create the image file May 26 14:50:23 jaeckel: then add to IMAGE_FSTYPES May 26 14:50:55 in fact since this is just a signed version of another type perhaps you could use the compression format stuff instead May 26 14:51:08 that was my approach May 26 14:51:30 so I created a image_types.bbappend in my own layer, is that correct? May 26 14:52:03 ...in the classes folder... May 26 14:53:26 you can't use the bbappend mechanism to extend classes I'm afraid, so that won't work May 26 14:53:36 grah :D May 26 14:54:11 you can add your own class and add that to IMAGE_CLASSES though May 26 15:12:26 cool, that worked May 26 15:33:46 thx May 26 18:37:30 Quick question. Let's say I want to build my own version of GRUB, but I still want to build the in-tree version of GRUB. I copy the .bb file from oe-core/meta/recipes-bsp/grub to the corresponding location in my own layer and rename it to, say, grub-bshelton. May 26 18:38:05 Assuming I still want to use the patches that oe-core has for GRUB, is there a way I can reference these from my new recipe rather than copying all of them into my layer? May 26 18:38:11 Or is there a better way to handle this scenario? May 26 18:48:32 bbappend? May 26 18:59:42 Can I bbappend a recipe and have it build under a different name, as well as in the original (unmodified) form? May 26 19:01:54 nope May 26 19:02:02 you could create a recipe and include/require the other one May 26 19:02:18 using a relative path from the root of its layer. e.g. require recipes-devtools/foo/bar.bb May 26 19:02:28 of course, getting to its file:// files is slightly uglier May 26 19:02:58 Can I do the same thing with FILESEXTRAPATHS to get to its files? May 26 19:03:05 if I did copy the recipe? May 26 19:04:33 not sure why you'd want to copy a recipe you can include, but yes, though again, locating the files would be ugly. either you'd have to hardcode reference to ${COREBASE} in FILESPATH, or you'd have to use inline python to search BBPATH May 26 19:06:04 I couldn't do, e.g., FILESEXTRAPATHS_prepend := recipes-bsp/grub/files May 26 19:06:05 ? May 26 19:09:58 that's correct, FILESPATH doesn't seach BBPATH, it requires absolute paths May 26 19:11:17 ${@':'.join('%s/recipes-bsp/grub/files' % p for p in '${BBPATH}'.split(':'))} May 26 19:11:20 or similar, may be of use May 26 20:25:13 what do I need to do beyond making sure "en_US.UTF-8" is in GLIBC_GENERATE_LOCALES and "en-us" is in IMAGE_LINGUAS to get en_US.UTF-8 in my images? May 26 20:25:23 at the moment, I just get regular en_US **** ENDING LOGGING AT Wed May 27 02:59:59 2015