**** BEGIN LOGGING AT Fri Jun 20 03:00:00 2014 Jun 20 04:37:37 <5EXAA8SU8> build #552 of kirkwood is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/552 Jun 20 04:56:45 <5EXAA8SU8> build #95 of pxa is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/pxa/builds/95 Jun 20 06:59:45 <5EXAA8SU8> build #206 of imx6 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/imx6/builds/206 Jun 20 07:13:55 _FaulP3lz: shhhhhh Jun 20 07:48:08 <5EXAA8SU8> build #630 of ramips is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/630 Jun 20 07:51:59 <5EXAA8SU8> build #96 of x86_64 is complete: Failure [failed compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/x86_64/builds/96 Jun 20 07:55:49 * russell-- just submitted a ticket about my asiarf awm003 memory detection problem: https://dev.openwrt.org/ticket/16844 Jun 20 08:00:51 <5EXAA8SU8> build #571 of xburst is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/571 Jun 20 08:09:03 i'm still trying to find out why a mpd 0.18.11 works with "-ggdb3" but not without on Openwrt Jun 20 08:23:20 define "works" and "not works"? Jun 20 08:30:42 playing any file (tested with mp3,m4a,flac) resulst in "instant" sigsegv Jun 20 08:31:14 I already tracked it down but the code is really simple at that stage Jun 20 08:34:06 http://paste.debian.net/105466/ :/ Jun 20 08:34:43 not understanding that much of gdb / stack debugging things Jun 20 08:40:24 plntyk: backtrace Jun 20 08:43:48 <5EXAA8SU8> build #95 of adm8668 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/adm8668/builds/95 Jun 20 08:44:13 http://paste.debian.net/105406/ - there are backtraces of 2 files - mp3+m4a Jun 20 08:46:40 oh, so you cranked up debugging and the problem goes away? Jun 20 08:46:54 maybe a misoptimization by gcc Jun 20 08:47:04 removed null pointer check or something Jun 20 08:49:09 yep ... if I remove "-Os" from CPU_FLAGS in include/target.mk I get playback :/ Jun 20 08:49:17 just tested Jun 20 08:49:48 if you know where the fault is happening, you could sprinkle in some printfs to check stuff Jun 20 08:50:13 interestingly there is not much in AudioFormat::GetSampleSize() that can crash Jun 20 08:50:23 its essentially an inlined switch/case Jun 20 08:52:15 yep thats what is strange Jun 20 08:53:15 can you pastebin a listing of the *.o files in mpd's build_dir ? Jun 20 08:53:29 and rebuild with -Os again to get a crashing executable Jun 20 08:53:48 have you considered taking up an instrument, live music is better anyway. Jun 20 08:56:11 i learned playing electric bass a while ago Jun 20 08:56:14 :) Jun 20 08:56:24 nice! Jun 20 08:56:39 * russell_ considers all the options Jun 20 09:29:46 <5EXAA8SU8> build #548 of ar71xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/548 Jun 20 10:00:05 jow r41282 trunk/package/network/config/netifd/Makefile * netifd: fix handling of "lookup main" for ip rules (#15315) Jun 20 11:35:49 jow_laptop: is netifd's source url correct? fatal: reference is not a tree: 26505641d8037a8e0a32945251653e4a116b342e Jun 20 11:43:12 oh crap, I pushed into the wrong repo Jun 20 11:43:17 will fix in a second Jun 20 12:30:25 jow r41283 trunk/package/network/config/netifd/Makefile * Revert "netifd: fix handling of "lookup main" for ip rules (#15315)" Jun 20 13:28:39 can not file a ticket at luci, keep saying my captcha input is wrong Jun 20 13:28:49 tmp/.config-package.in:23750:error: recursive dependency detected! Jun 20 13:28:56 tmp/.config-package.in:23750: symbol PACKAGE_luci-app-pbx is selected by PACKAGE_luci-app-pbx-voicemail Jun 20 13:29:07 quite a few circular dependencies on luci now Jun 20 13:29:19 http://luci.subsignal.org/trac/newticket Jun 20 13:35:08 this has been fixed a while ago Jun 20 13:35:14 needs updates in both openwrt and luci Jun 20 13:41:47 jow_laptop: i did updated luci Jun 20 13:42:00 not openwrt apparently Jun 20 13:42:03 let me pull trunk Jun 20 13:42:25 yeah i had to stick with some commit-id on trunk for a while to share SDK with others Jun 20 13:42:57 blogic r41284 trunk/package/base-files/files/lib/upgrade/common.sh * base-files: regenerate the /lib64 symlink when switching to a ramdisk Jun 20 13:43:01 blogic r41285 trunk/target/linux/octeon/image/Makefile * octeon: build a sysupgrade image for erlite Jun 20 13:43:05 blogic r41286 trunk/target/linux/octeon/base-files/ lib/upgrade lib/upgrade/platform.sh * octeon: add sysupgrade support for the erlite Jun 20 13:43:45 jow_laptop: is it at all possible to use external-kernel-git(3.0.x) to build openwrt's toolchain? i had a toolchain issue for more than one weeek. Jun 20 13:43:58 a package built with buildroot's toolchain(arm-hardfp) worked fine Jun 20 13:44:02 iirc yes but that otpion is not really maintained Jun 20 13:44:25 but built with openwrt's arm-eglibc with pretty much the same ./configure flags on gcc can run that program, but no output Jun 20 13:44:59 the only difference is that in buildroot i can build the toolchain with any kernel headers Jun 20 13:45:17 openwrt's toolchain is tightly coupled with all its magics Jun 20 13:48:20 I also removed $(call toolchain_test,CONFIG_SOFT_FLOAT,softfloat) from toolchain/wrapper/Makefile since this check mandates I must use SOFT_FLOAT for arm/eglibc Jun 20 13:48:40 ? Jun 20 13:48:51 but turn that config on/off made no difference, though it seems by removing it i can build a armhf toolchain and the whole rootfs Jun 20 13:48:56 it is supposed to check CONFIG_SOFT_FLOATCONFIG_SOFT_FLOAT from .config Jun 20 13:49:34 jow_laptop: yeah, i don't know where it went wrong, it simply disallows me to turn SOFT_FLoAT off then compile arm/eglibc Jun 20 13:50:30 try to turn that off, build with glibc, and arm will fail Jun 20 13:50:37 i'm testing with imx6 Jun 20 13:51:20 you can't turn it off in menuconfig? Jun 20 13:51:58 i can, but it won't finish the build Jun 20 13:52:18 saying something like i need that on Jun 20 13:52:53 that happened since gateworks turned on the CPU_SUBTYPE from vfpv3 to neon I believe Jun 20 13:53:05 s/turned on/switch/ Jun 20 13:53:29 jow_laptop: do you know any estimation date on the BB-rc? Jun 20 13:53:38 nope Jun 20 13:53:59 even an pre-rc tag will help, as I new know if my git pull is done in the middle of something Jun 20 13:54:11 s/new/never/ Jun 20 13:55:50 what about make some 'rc' tag every 3 months or 6 months, not a release, but a reference point for us users Jun 20 13:56:36 something like buildroot does, it works stable and great, but it needs something like uci/luci Jun 20 13:56:36 <5EXAA8SU8> build #96 of mpc83xx is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/96 Jun 20 14:11:29 its probably worth considering, but we're in the middle of establishing some proper processes at all, then we can think about stuff like that Jun 20 14:12:27 that's great. you mean the github move? Jun 20 14:13:03 i really think openwrt needs some tag to stage the development, bleeding edge for years is too hard for us who wants to use it Jun 20 14:13:18 not only that, also how to build releases, how to build package feeds independently of the core, so that we can provide upgrade repositories and stuff like that Jun 20 14:13:42 how can I help? I used to donate some hardwares :) Jun 20 14:14:41 and then there's always the clash of interests... like openwrt being quite router centric while buildroot is more of a general purpose embedded target Jun 20 14:15:03 openwrt should look beyond router Jun 20 14:15:20 IoT gateway is a hot word for it, people are using it for non-routers Jun 20 14:15:39 it could be used for so many things Jun 20 14:16:28 I rises and falls with the availability of contributors/develoeprs maintaining such features Jun 20 14:16:49 oepnwrts buildroot for example is only developed as far as required to build the router firmware, simply speaking Jun 20 14:17:05 there is no dedicated group of people doing only toolchain/build system work Jun 20 14:17:20 need attract Yorin to openwrt to take care of toolchain :) Jun 20 14:17:46 he is on buildroot irc 24x7 Jun 20 14:17:50 I've seen that BR has a few quite nice features regarding toolchain wrapping, cross toolchain support etc. Jun 20 14:18:21 what about leverage crosstool-ng and keey it outside of openwrt? Jun 20 14:18:22 but from simply looking at it I already know that its a huge task to keep up with that, it would consume most of the resources Jun 20 14:19:29 make openwrt works with external toolchains will mitigate the pain Jun 20 14:19:50 currently i can use openwrt's own toolchain as external toolchain Jun 20 14:19:59 but that's it, no other external toolchain can work Jun 20 14:20:31 for example, the one with --sysroot(e.g. buildroot, crosstool) do not work with openwrt as external ones Jun 20 14:21:54 also,glibc deserves more love :) Jun 20 14:22:57 thats one of the thing I meant, for the vast majority of openwrts traditional targets glibc is no viable choice Jun 20 14:23:06 so there's no real interest in maintaining it Jun 20 14:23:33 somewhone who uses it in production, maybe even multiple targets, say arm and x86 needs to keep up with it Jun 20 14:24:03 as for the --sysroot - I know, it would make things much simpler Jun 20 14:24:16 unfortunately a lot of proprietary toolchains do not support it yet Jun 20 14:24:46 or at least didn't when I wrote the toolchain wrapping stuff years ago Jun 20 14:26:52 I'd be interested how the ct-ng toolchain fails though Jun 20 14:28:09 could not find --libc.so.6 is the current error Jun 20 14:28:36 the SYSROOT definition used in openwrt disagrees with external toolchain's sysroot Jun 20 14:28:43 so it could not find libc.so Jun 20 14:28:50 tried to hack that with no success yesterday Jun 20 14:28:57 ok Jun 20 14:29:15 it's hard-coded to STAGING-DIR style 'sysroot' Jun 20 14:29:38 maybe we should extend scripts/ext-toolchain.sh to simply use --sysroot if available, else fall back to the legecy -Wl,-rpath-link + indjecting -I and -L Jun 20 14:30:40 YES Jun 20 14:31:15 I also remember that multilib toolchains where giving me headaches Jun 20 14:31:42 that's another issue indeed Jun 20 14:31:48 at the moment i'm stuck with sysroot mismatches Jun 20 14:32:42 ok Jun 20 14:33:08 i'm happy to test any external-toolchain changes Jun 20 16:59:04 <5EXAA8SU8> build #94 of omap is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/94 Jun 20 17:40:46 <5EXAA8SU8> build #601 of ppc44x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/601 Jun 20 17:41:40 <5EXAA8SU8> build #570 of sibyte is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/570 Jun 20 17:41:59 <5EXAA8SU8> build #601 of rb532 is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/601 Jun 20 17:53:14 cyrus r41287 trunk/package/network/config/netifd/Makefile * netifd: various routing table handling fixes Jun 20 18:36:18 <5EXAA8SU8> build #581 of avr32 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/581 Jun 20 18:47:37 <5EXAA8SU8> build #612 of uml is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/612 Jun 20 18:49:43 <5EXAA8SU8> build #508 of iop32x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/508 Jun 20 19:10:33 <5EXAA8SU8> build #641 of cobalt is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/641 Jun 20 19:14:40 <5EXAA8SU8> build #647 of orion is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/647 Jun 20 19:14:50 kaloz r41288 trunk/target/linux/mvebu/files/arch/arm/boot/dts/armada-xp-mamba.dts * [mvebu]: define the power led for the WRT1900AC Jun 20 19:21:41 nbd: https://git.kernel.org/cgit/linux/kernel/git/linville/wireless-next.git/commit/?id=418378fed0506b5ec0b43c03bc11929ec1f88073 Jun 20 19:21:58 nbd: after next compat-wireless update we'll be able to reduce b43 size for mips74k :) Jun 20 20:00:38 <5EXAA8SU8> build #536 of au1000 is complete: Failure [failed shell compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/536 Jun 20 20:44:37 kaloz r41289 trunk/package/kernel/linux/modules/hwmon.mk * [kernel]: package the gpio-fan module Jun 20 20:46:08 kaloz r41290 trunk/package/kernel/linux/modules/other.mk * [kernel]: make thermal framework registered sensors also register as hwmons Jun 20 21:27:00 <5EXAA8SU8> build #496 of mcs814x is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mcs814x/builds/496 Jun 20 22:36:58 kaloz r41291 trunk/target/linux/mvebu/base-files/lib/mvebu.sh * [mvebu]: make script executable, add WRT1900AC detection Jun 20 23:07:12 kaloz r41292 trunk/target/linux/mvebu/files/arch/arm/boot/dts/armada-xp-mamba.dts * [mvebu] update mamba dts Jun 21 02:26:34 I would like to create a package that uses uhttpd as part of a larger system. is there a way for me to cleanly add a section in /etc/config/uhttpd as part of installing my package? anyone have thoughts on that method vs starting /usr/sbin/uhttpd myself? **** ENDING LOGGING AT Sat Jun 21 03:00:00 2014