**** BEGIN LOGGING AT Tue Nov 20 02:59:57 2007 Nov 20 13:21:07 ejka * r9584 /trunk/ (Config.in include/kernel-defaults.mk): add external kernel tree support Nov 20 13:23:03 ejka: what's the rmdir command in that changeset for? Nov 20 13:23:36 also: what specifically are you using this for? Nov 20 13:26:41 if $(LINUX_DIR) exists, you'll get symlink inside Nov 20 13:27:16 but rmdir only kills it if it's empty Nov 20 13:27:20 and how likely is that? Nov 20 13:27:28 sure, and will fail if it is not Nov 20 13:27:41 with make failing Nov 20 13:28:29 shouldn't it just get rid of the old tree? Nov 20 13:28:43 and I use it to build my kernel git master, which is frequently rebased agains several trees Nov 20 13:28:46 or even better, just redirect LINUX_DIR instead of doing a symlink Nov 20 13:29:12 ghm, haven't thought about it Nov 20 13:29:41 if you just point LINUX_DIR at the right tree, then there's no need to remove the old tree or check for its presence Nov 20 13:29:50 yeah Nov 20 13:56:06 re Nov 20 16:27:58 olli * r9585 /packages/net/bmon/ (Makefile patches/001-bugfixes.patch): Nov 20 16:27:58 [Packages] bmon: Nov 20 16:27:58 * Fix compilation Nov 20 16:27:58 * Cleanup Makefile Nov 20 16:27:58 * Refresh patch Nov 20 18:22:03 olli * r9586 /packages/net/subversion/Makefile: Nov 20 18:22:03 [Packages] subversion: Nov 20 18:22:03 * Fix compilation Nov 20 18:22:03 * Cleanup Makefile Nov 20 18:22:03 * Add libtool fixes Nov 20 18:22:03 * Update 1.4.4 > 1.4.5 Nov 20 19:46:48 Hmm. robocfg.c:58: error: field 'ifr' has incomplete type Nov 20 19:47:12 noz: rm -rf staging_dir/mipsel; make Nov 20 19:47:22 noz: after svn update or git-svn rebase Nov 20 19:47:25 OK. Ta. That was quick! Nov 20 19:48:23 I have some issues with the build system right now. I think we should be able to reconfigure (changing toolchain), and just make and have it "Just Work"(TM) Nov 20 19:49:18 At the moment, it doesn't build a non-existent toolchain because the stamp files under staging_dir/mipsel are stil present. Nov 20 19:50:23 What causes the 'ifr' problem (other than the obvious "staging_dir/mipsel" answer! :) Nov 20 19:55:11 libnl installed include files that broke various packages Nov 20 19:55:15 i moved them to a different place Nov 20 19:59:23 noz: just committed a change that moves some of the toolchain stamps out of staging_dir/mipsel Nov 20 19:59:34 they weren't supposed to be there anyway Nov 20 19:59:42 after that, switching the toolchain should work Nov 20 19:59:51 Timely. Ta Nov 20 21:56:18 At what point does the Build/InstallDev section of a Makefile get invoked? tcpdump is failing to build because the pcap headers aren't found. Nov 20 21:57:10 However, the Build/InstallDev section of the Makefile has a copy step for them, and the tcpdump package depends on libpcap Nov 20 21:57:16 it gets invoked at from the compile target Nov 20 21:57:27 even when the package itself is not selected Nov 20 21:58:24 and " $(CP) $(PKG_INSTALL_DIR)/usr/include/pcap{,-bpf,-namedb}.h \ Nov 20 21:58:24 $(1)/usr/include/ Nov 20 21:58:24 " causes them to be copied to staging_dir/mipsel/usr/include, yes? Nov 20 21:58:40 ye Nov 20 21:58:42 yes Nov 20 22:00:11 (or not in my case!) It might be a partial build issue, given that I've just deleted staging_dir/mipsel. I'll delete the whole of build_dir and retry Nov 20 22:07:11 i see the problem Nov 20 22:10:00 fixing Nov 20 22:43:42 nbd * r9588 /trunk/include/package.mk: fix build issues that happen when the staging dir is cleared and the build directories stay Nov 21 00:06:27 make[3]: Entering directory `/usr/local/Linksys-devel/toolchain/kernel-headers' Nov 21 00:06:27 ln -s /usr/local/Linksys-devel/build_dir/toolchain-mipsel_gcc4.1.2/linux-2.6.23.1 Nov 21 00:06:27 ln: creating symbolic link `./linux-2.6.23.1' to `/usr/local/Linksys-devel/build_dir/toolchain-mipsel_gcc4.1.2/linux-2.6.23.1': File exists Nov 21 00:07:24 Grrr. That should probably be 'ln -sf', but it's late, and I'm not 100%, so I thought I'd ask a 2nd opinion... Nov 21 00:14:01 erm.. the ln -s command only has one argument Nov 21 00:15:19 noz: run make oldconfig **** ENDING LOGGING AT Wed Nov 21 02:59:56 2007