**** BEGIN LOGGING AT Tue Jul 05 02:59:57 2011 Jul 05 03:02:45 philipp64|laptop: pong Jul 05 03:25:29 build #60 of at91 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/60 Jul 05 03:39:14 build #58 of ubicom32 is complete: Failure [failed compile_3] Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/58 Jul 05 03:43:14 nbd: is it worth doing a version bump for ocf for kernels prior to 2.6.38? Jul 05 03:56:38 blogic: err, wrong person... is it worth doing a version bump for ocf for kernels prior to 2.6.38? Jul 05 08:26:16 build #47 of ixp4xx is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/47 Jul 05 09:08:35 obsy * r27452 /packages/net/transmission/Makefile: [packages] transmission: add dependency to librt Jul 05 10:01:27 nbd: feel free to use http://inuyasha.ath.cx/~jonas/0001-linux-generic-Update-Overlayfs-to-v10-for-2.5.38-39.patch - 2.6.37 takes a bit more work since it's missing some of the rcu security things, and I don't feel confident enough to backport it without compromising security Jul 05 11:56:07 jow * r27453 /packages/libs/glib2/patches/002-missing-gthread-include.patch: [packages] glib2: add missing gthread.h include, observed during host compilation of glib2 on CentOS 5.6 (#8794) Jul 05 11:59:17 kaloz * r27454 /trunk/target/linux/generic/config-3.0: [target/generic/3.0]: add missing symbols Jul 05 12:44:03 florian * r27455 /packages/multimedia/imagemagick/Makefile: [package] imagemagick: add tools (#8430) Jul 05 12:44:05 florian * r27456 /packages/multimedia/imagemagick/Makefile: [package] imagemagick: add tiff plugin (#8430) Jul 05 12:44:37 florian * r27457 /trunk/target/linux/generic/ (6 files in 6 dirs): [kernel] sched_connmark depends on NF_CONNTRACK_MARK (#8976, #9423) Jul 05 12:44:40 florian * r27458 /trunk/package/kernel/modules/netsupport.mk: [package] add kmod-sched-netem (#8839) Jul 05 12:44:43 florian * r27459 /trunk/package/e2fsprogs/Makefile: [package] e2fsprogs: fix blkid dependencies (#9377) Jul 05 12:44:49 florian * r27460 /trunk/package/iptables/Makefile: [package] iptables: add iptables-mod-led (#9400) Jul 05 12:55:08 build #62 of brcm63xx is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/62 Jul 05 14:50:44 build #60 of s3c24xx is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/60 Jul 05 16:24:27 ping nbd Jul 05 16:27:26 https://dev.openwrt.org/changeset/27436/trunk needs CONFIG_PACKAGE_kmod-mac80211 to be enabled in order to get compiled Jul 05 16:27:43 how do we enable a package without making it a dependency Jul 05 17:11:28 jow * r27461 /packages/net/wing/patches/010-libc-compat.patch: [packages] wing: the host build fails with glibc 2.5 - add compat macros, observed on CentOS 5.6 Jul 05 17:11:51 has anyone tried the patch on #9638 yet Jul 05 17:11:52 ? Jul 05 17:13:50 it looks like Debian wheezy needs it too (but I haven't tested yet to see if it works) Jul 05 17:18:34 more to the point, how have other host builds been solved for libm etc not in /lib or /usr/'lib (I see /usr/lib32/libm.so and /usr/lib/x86_64-linux-gnu/libm.so for debian wheezy, for example) Jul 05 17:43:39 jow * r27462 /packages/net/wing/patches/ (010-libc-compat.patch 011-perl-compat.patch): [packages] wing: fix illegal modification of constant value triggered in at least perl 5.8.8, observed on CentOS 5.6 Jul 05 17:54:22 cshore: no,you can do a simple regression check and add it Jul 05 17:55:09 ok Jul 05 18:00:11 jow_laptop: ping Jul 05 18:01:30 I'm seeing a case where compiling might be picking up header files from the host... I'm wondering why when we compile we don't force the cross-compiler to leave our /usr/include (etc) out of the sysinc path... Jul 05 18:05:53 because it wasn't possible in the past? Jul 05 18:06:44 philipp64|laptop: speaking of which, how would you? Jul 05 18:07:16 from gcc 4.3 and later, you can use --sysinc=% etc. Jul 05 18:27:19 zandbelt * r27463 /packages/net/asterisk-1.8.x/Makefile: [packages] asterisk-1.8.x: upgrade to 1.8.4.4; put keydir in static storage Jul 05 18:29:13 zandbelt: Thanks. Jul 05 18:41:32 mazilo: I heard yate isn't compiling these days, is that true? Jul 05 18:41:48 cshore: Yup, it is broken. Jul 05 18:42:01 damn, ok, I'll put that on my todo list Jul 05 18:42:09 cshore: Thanks. Jul 05 18:44:12 I see no one reported a bug though Jul 05 19:33:39 I just finished building complete backfire branch for x86 and yate was compiled okay Jul 05 19:34:13 it also built okay on xburst in backfire Jul 05 19:34:45 so if its broken it must be either trunk specific or pebkac :) Jul 05 20:31:26 gah Jul 05 20:31:50 trying to bring up openwrt on my G300NH Jul 05 20:32:03 the serial port wiring is super dodgy, but did manage to flash openwrt Jul 05 20:32:16 but I can see ethernet is not coming up. Jul 05 20:32:26 trying to add a USB serial port. Jul 05 20:58:46 wrong serial driver. sigh Jul 05 21:19:16 mooh Jul 05 21:23:48 kaloz * r27464 /trunk/target/linux/generic/config-3.0: [generic/3.0]: add some symbols Jul 05 21:25:15 Chocky: there are two variants around Jul 05 21:25:33 Chocky: one hw revision uses the rtl8366s, the other a rtl8366rb Jul 05 21:25:49 if the switch doesn ot work with the g300nh image, try the g301nh one isntead Jul 05 21:34:35 hmmmm....libext2fs uses a different version of libcom_err than e2fsprogs which is creating problems Jul 05 21:35:02 ok Jul 05 21:36:26 wait. waits building com_err if it's not e2fsprogs? something's fishy here Jul 05 21:37:27 aha, kerberos is Jul 05 21:37:39 ok, for now I drop it Jul 05 21:37:46 (personally, not from archive) Jul 05 21:45:19 xmff: hooray. thanks Jul 05 21:45:48 Chocky: np Jul 05 21:46:13 -1 one point to buffalo. but +10 for them going OSS stuff. they used to be really unfriendly. Jul 05 21:59:06 cshore: well, we could move libcom_err 2.x and 3.x to different directories Jul 05 21:59:15 like libnet 1.0 and libnet 1.1 Jul 05 21:59:36 or patch one of both users to cope with the newer/older version Jul 05 22:00:32 xmff: you know, I'm kinda glad I've given so much IRC help in the past. Means I don't need to feel bad when I get instant answers. Also means I get to abuse those who are obviously guessing ;-) Jul 05 22:00:33 I think different dirs probably makes sense.... Jul 05 22:01:29 let me add it to my todo (atm I'm just trying to get a build done) Jul 05 22:17:15 are sem_destroy, sem_wait, sem_post, and sem_init from pthreads? Jul 05 22:17:35 most likely Jul 05 22:18:21 xmff: buffalo etc al is mips, not mipsel? Jul 05 22:18:40 hmm...because pthreads in the -L (.../toolchain/lib), but those aren't found by the linker for blkid Jul 05 22:18:50 Chocky: mips (be) Jul 05 22:18:55 ta Jul 05 22:25:02 jow * r27465 /packages/libs/apr/Makefile: [packages] apr: ensure rlimit support, fixes httpd mod_cgi link failures, upstream httpd bug 51371 Jul 05 22:35:52 what the correct syntax for -rpath-link if you need more than one -rpath-link dir? (is presently -Wl,-rpath-link,$(top_builddir)/lib ; I have to figure out how to get the staging dir toolchain /lib into there) Jul 05 22:36:44 you can repeat them Jul 05 22:37:16 the -Wl,...? Jul 05 22:37:25 or just -rpath Jul 05 22:38:04 -Wl,-rpath-link=... -Wl,-rpath-link= Jul 05 22:38:13 thanks Jul 05 22:41:38 what the short name for $(STAGING_DIR)/toolchain.../lib ? Jul 05 22:42:47 hm, I fear you're doing something very wrong here Jul 05 22:43:24 what gcc version is that? Jul 05 22:44:00 you mean host or openwrt? Jul 05 22:44:12 openwrt Jul 05 22:45:07 4.5.x with Linary enhancements Jul 05 22:45:40 ok Jul 05 22:45:45 and whats the actual error? Jul 05 22:47:17 http://pastebin.com/DRggRqeQ Jul 05 22:47:56 failure is the link of blkid Jul 05 22:50:39 cshore: then either -lrt or -lpthread is missing Jul 05 22:50:48 it should be not needed to add rpath flags Jul 05 22:51:02 in blkid or in libext2fs? Jul 05 22:52:01 blkid Jul 05 22:52:19 it looks like -lpthread is commented out for dynamic libraries in the Makefiles (upstream) Jul 05 22:52:30 well Jul 05 22:52:38 add it back :) Jul 05 22:52:46 and it will probably work as intended Jul 05 23:14:37 cshore * r27466 /trunk/package/e2fsprogs/patches/ (. 100_add_missing_libpthread_for_blkid): [package] e2fsprogs: Added libpthread back to blkid link, otherwise blkid fails to link) Jul 05 23:15:01 oh crap, an extra file got in there Jul 05 23:15:18 and the 100* seems to lack an extension Jul 05 23:20:33 oh, there is no extra file, just missing extension Jul 05 23:22:27 ah, that was Kaloz's commit, not part of mine Jul 05 23:23:19 cshore * r27467 /trunk/package/e2fsprogs/patches/ (2 files): [package] e2fsprogs: Fixed filename of patch (needed .patch extension) Jul 06 00:07:42 hm Jul 06 00:08:13 my .config has 2 entries for kmod-input-gpio-keys-polled. One is set, one is not Jul 06 00:09:33 now they are both set Jul 06 00:24:42 * Chocky pokes Jul 06 00:25:11 the image builder is building stuff for all the ar71xx platforms. I only want mine. especially since it fails for some of them, since the target image is too big Jul 06 00:26:58 * Chocky pokes xmff Jul 06 00:27:09 ? Jul 06 00:27:31 how can I stop the image builder generating ROMs for everything? Jul 06 00:28:39 http://wgaugsburg.de/trac/browser/kamikaze/contrib/imagebuilder/trunk/patches/trunk/200-IB-Makefile.patch Jul 06 00:28:44 see line 167, in the last hunk Jul 06 00:28:50 * Chocky looks Jul 06 00:29:45 but profile is the list of packages, not the target hardware, right? Jul 06 00:29:57 I already have a custom target file. Jul 06 00:30:54 ar71xx also uses for image building Jul 06 00:34:37 then USER_PROFILE= what? Jul 06 00:35:01 nothing Jul 06 00:35:06 just modify the line as shown Jul 06 00:39:02 I think that's only going to work if my profile entry exactly matches the target hardware name, right? Jul 06 00:42:03 yes, seems so Jul 06 00:42:38 and perhaps that's a reasonable assumption, but I had a custom package list under an entirely different name. and the default one for the g301 etc doesn't work, since ath9k isn't build by default. I think. Jul 06 01:37:24 MOO! **** ENDING LOGGING AT Wed Jul 06 02:59:57 2011