**** BEGIN LOGGING AT Thu Mar 30 03:00:02 2017
Mar 30 13:20:49 <heeen> how can I find where a distro feature is added
Mar 30 13:24:02 <rburton> bitbake -e | less
Mar 30 13:24:07 <rburton> search for DISTRO_FEATURES
Mar 30 13:24:21 <rburton> scroll up and look at the history of the variable
Mar 30 13:28:58 <heeen> got it, thanks
Mar 30 13:29:22 <heeen> so this is weird, this build has x11 headers
Mar 30 13:29:28 <heeen> but x11 is not in the distro features
Mar 30 13:29:54 <Crofton|work> |   arm-oe-linux-gnueabi-objcopy  -j .text -j .secure_text -j .secure_data -j .rodata -j .hash -j .data -j .got -j .got.plt -j .u_boot_list -j .rel.dyn -j .efi_runtime -j .efi_runtime_rel  -O binary spl/u-boot-spl spl/u-boot-spl-nodtb.bin
Mar 30 13:29:54 <Crofton|work> | /bin/sh: bc: command not found
Mar 30 13:29:54 <Crofton|work> | /home/balister/src/oe-core/build/tmp-glibc/work/sulfur-oe-linux-gnueabi/u-boot/1_2017.01-r0/git/scripts/Makefile.spl:201: recipe for target 'spl/u-boot-spl-pad.bin' failed
Mar 30 13:29:55 <Crofton|work> weird
Mar 30 13:30:43 <heeen> and if I try to build libx11 it says it is skipped because it is not in DISTRO_FEATURES
Mar 30 13:38:35 <heeen> I can find libX11 even
Mar 30 13:51:15 <rburton> heeen: then either you've some crazy to let x11 built (ie distro feature manipulation in the distro config) or your libx11 was built before the distro features were changed.
Mar 30 15:47:55 <darknighte> hey folks, I'm trying to build a single package with a different package format (long story), but cannot see a way to do so without setting PACKAGE_CLASSES.  Anyone know a better way to handle this?
Mar 30 15:49:17 <darknighte> ultimately, I'd like to be able to issue a bitbake foo-image and have bb build/package for the default, which is .ipk in this case, and bitbake odd-package to generate the odd man out.
Mar 30 15:51:48 <kergoth> can you not just 'inherit package_rpm' in the recipe in question?
Mar 30 15:52:01 <kergoth> package_classes just gets added to inherit, after all
Mar 30 15:52:58 <darknighte> makes sense.  I'll try that out.
Mar 30 15:53:40 <kergoth> of course, it'd almost certainly still emit the other format too, but it should emit the other as well
Mar 30 15:53:47 <Crofton|work> Marex, this make any sense to you?
Mar 30 15:53:48 <Crofton|work> |   arm-oe-linux-gnueabi-objcopy  -j .text -j .secure_text -j .secure_data -j .rodata -j .hash -j .data -j .got -j .got.plt -j .u_boot_list -j .rel.dyn -j .efi_runtime -j .efi_runtime_rel  -O binary spl/u-boot-spl spl/u-boot-spl-nodtb.bin
Mar 30 15:53:48 <Crofton|work> | ERROR: oe_runmake failed
Mar 30 15:53:48 <Crofton|work> | /bin/sh: bc: command not found
Mar 30 15:53:51 <kergoth> that is, it'd write ipk and rpm both
Mar 30 15:54:08 <kergoth> Crofton|work: missing DEPENDS on bc-native? iirc we used to have to do that in kernel builds
Mar 30 15:54:14 * Crofton|work waits for Marex to say rude things about Xilinx
Mar 30 15:54:45 <Crofton|work> oh right, I just can't bitbake it to fix sysroot nay more :)
Mar 30 15:55:01 <kergoth> indeed
Mar 30 15:55:31 * kergoth thinks do_populate_sysroot is rather misleading nowadays, name wise
Mar 30 15:58:01 <Crofton|work> yep
Mar 30 15:58:10 <Crofton|work> thanks kergoth for reminding me what I already knew
Mar 30 15:58:17 <kergoth> np :)
Mar 30 21:41:10 <Marex> Crofton|work: altera still didn't give me a plushy FPGA sleeping pillow :(
Mar 30 21:41:34 <Crofton|work> lol
Mar 30 21:41:44 <Marex> Crofton|work: so what's the problem ?
Mar 30 21:41:47 <Crofton|work> Did Xilinx?
Mar 30 21:42:08 <Crofton|work> u-boot recipe needs to DEPEND on bc-native
Mar 30 21:42:30 <Crofton|work> seems like needed for xilinx/spl
Mar 30 21:42:33 <Crofton|work> at some point
Mar 30 21:42:41 <Crofton|work> I can send patch
Mar 30 21:42:53 <Crofton|work> can you think of other tools used in only some machines?
Mar 30 21:43:08 <Crofton|work> nrossi, awake?
Mar 30 21:44:03 <nrossi> Crofton|work: i am, whats up?
Mar 30 21:45:47 <Crofton|work> xilnx/spl needs bc to build
Mar 30 21:46:03 <Crofton|work> but not in DEPENDS
Mar 30 21:46:16 <Crofton|work> making sure I ahve my facts in line
Mar 30 21:47:02 <Marex> Crofton|work: why does it need bc ?
Mar 30 21:47:11 <Crofton|work> dunno
Mar 30 21:47:13 <nrossi> Crofton|work: not sure tbh lets see
Mar 30 21:47:34 <Crofton|work> jsut failed because it couldn;t find it, so I added bc-native to DEPEND :)
Mar 30 21:48:54 <Crofton|work> [balister@thuvia u-boot (master)]$ grep DEPEND *
Mar 30 21:48:54 <Crofton|work> grep: files: Is a directory
Mar 30 21:48:54 <Crofton|work> u-boot_2017.01.bb:DEPENDS += "dtc-native"
Mar 30 21:48:54 <Crofton|work> u-boot-fw-utils_2017.01.bb:DEPENDS = "mtd-utils"
Mar 30 21:48:54 <Crofton|work> u-boot-mkimage_2017.01.bb:DEPENDS = "openssl"
Mar 30 21:49:06 <nrossi> looks like its used in spl nodtb padding at least
Mar 30 21:49:09 <Crofton|work> the += looks funny
Mar 30 21:49:16 <Crofton|work> xilinx only
Mar 30 21:49:28 <nrossi> not xilinx specific
Mar 30 21:49:38 <Crofton|work> although I'm not sure why should do arch machine only depends :)
Mar 30 21:49:45 <Crofton|work> ah OK
Mar 30 21:50:09 <Crofton|work> ok, so I should send in patch
Mar 30 21:50:12 <nrossi> http://git.denx.de/?p=u-boot.git;a=blob;f=scripts/Makefile.spl;h=5370648e853d2d2e9d211069f2dc5d3e12f0f765;hb=HEAD#l210
Mar 30 21:50:20 <Crofton|work> I wonder about dropping the + from the DEPDNS line
Mar 30 23:15:49 <Marex> Crofton|work: bitbake -e , analyze the DEPENDS, the += is probably some legacy b/s
**** ENDING LOGGING AT Fri Mar 31 03:00:01 2017