**** BEGIN LOGGING AT Thu Sep 24 02:59:58 2015 Sep 24 05:11:47 build #91 of ramips.rt305x is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ramips.rt305x/builds/91 Sep 24 05:18:21 hm.... ar71xx is broken Sep 24 05:19:46 that's not good Sep 24 05:22:02 can somebody apply http://patchwork.ozlabs.org/patch/519204/ - it should fix buildbot arm64 issue http://buildbot.openwrt.org:8010/builders/arm64/builds/92/steps/compile_4/logs/stdio Sep 24 05:22:50 and its the same solution like other CAN symbols that are not appearing in generic kernel config (like CONFIG_CAN_TI_HECC=n in can.mk) Sep 24 05:50:58 build #93 of arm64 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/arm64/builds/93 Sep 24 06:56:53 build #90 of ar71xx.mikrotik is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.mikrotik/builds/90 Sep 24 07:40:19 * russell-- just found an old alix3 that's running r12616 Sep 24 07:40:50 Linux clumsy 2.6.25.16 #2 SMP Thu Sep 18 01:59:57 PDT 2008 i586 unknown Sep 24 08:38:42 nbd r47030 trunk/target/linux/arm64/config-default * arm64: disable ext2/ext3 filesystem support (ext4 is enough) Sep 24 08:38:48 nbd r47031 trunk/target/linux/ uml/config/x86_64 uml/config/i386 * uml: disable ext2/ext3 filesystem support (ext4 is enough) Sep 24 08:38:53 nbd r47032 trunk/target/linux/imx6/config-3.18 * imx6: disable ext2/ext3 filesystem support (ext4 is enough) Sep 24 08:38:59 nbd r47033 trunk/package/network/services/dropbear/files/dropbear.init * dropbear: add respawn param in case dropbear crashes Sep 24 08:39:09 nbd r47034 trunk/target/linux/ar71xx/image/Makefile * ar71xx: fix bogus hardcoded kernel image size for Seama images (fixes #20585) Sep 24 08:42:33 nbd r47035 trunk/package/kernel/linux/modules/can.mk * kernel: add missing CAN related config symbol Sep 24 09:10:05 nbd r47036 trunk/target/linux/ar71xx/ (7 files in 7 dirs) * ar71xx: Add WeIO Board Support Sep 24 09:10:10 nbd r47037 trunk/package/libs/libnl/Makefile * libnl: split libnl into smaller libraries Sep 24 09:20:05 rmilecki r47038 trunk/target/linux/oxnas/base-files/lib/upgrade/platform.sh * oxnas: drop some unused sysupgrade code Sep 24 09:20:17 rmilecki r47039 trunk/target/linux/oxnas/base-files/lib/upgrade/platform.sh * oxnas: add platform_pre_upgrade for sysupgrade Sep 24 11:20:30 rmilecki r47040 trunk/target/linux/ generic/patches-4.1/406-mtd-old-rootfs-squashfs-splitter.patch generic/config-4.1 generic/patches-4.1/430-mtd-add-myloader-partition-parser.patch * kernel: remove old "rootfs" splitter Sep 24 11:39:41 build #92 of avr32 is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/92 Sep 24 11:58:55 build #101 of brcm63xx is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/101 Sep 24 13:51:41 hi! I'm cross-compiling something, and it fails because it cannot find "libffi.so.6". I look and there's only "libffi.so.6.0.1" in the toolchain target dir. whose mistake is that? shouldn't there be a "libffi.so.6" also created? Sep 24 13:52:24 bah, scratch that, I'm blind. sorry! Sep 24 14:08:38 yay, I'm running common lisp on my router! :) I'll try to set up a package and create a pull request Sep 24 14:23:58 build #95 of brcm63xx.smp is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx.smp/builds/95 Sep 24 14:27:42 build #89 of mpc83xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/89 Sep 24 15:03:27 nbd r47041 trunk/package/kernel/mac80211/ patches/335-mac80211-fix-handling-of-PS-filtering-with-fast-xmit.patch patches/334-mac80211-fix-tx-sequence-number-assignment-with-soft.patch * mac80211: fix a few issues with fast-xmit support Sep 24 15:03:34 nbd r47042 trunk/package/kernel/mac80211/ patches/543-ath9k_entropy_from_adc.patch patches/530-ath9k_extra_leds.patch patches/336-ath9k-declare-required-extra-tx-headroom.patch patches/501-ath9k_ahb_init.patch * ath9k: fix handling of tx headroom for padding (fixes #20556) Sep 24 15:31:22 when I'm adding a package called 'ecl' (embedded common lisp), shall I make the name longer? e.g. common-lisp-ecl? Sep 24 15:41:03 nbd r47043 trunk/package/kernel/mt76/Makefile * mt76: update to the latest version, adds performance improvements Sep 24 15:44:57 nbd r47044 branches/ (7 files) * ath9k: fix handling of tx headroom for padding Sep 24 15:45:08 I could also use some guidance on whether I can assume that host gcc accepts -m32 Sep 24 15:46:35 i don't think you can Sep 24 15:46:45 you can't even assume that the host compiler is gcc Sep 24 15:47:06 on my system it isn't Sep 24 15:50:54 hrm, then it's concept failed, right? ecl host for now needs to be 32 bits to cross compile to 32 bit target... Sep 24 15:52:14 why does it need that? Sep 24 15:53:21 nbd: because it needs to be fixed. the host ecl compiler generates C code that is then compiled by the cross compiling gcc to yield the target binary, and it's not properly written with cross-compilation in mind... Sep 24 15:54:07 build #89 of ep93xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/89 Sep 24 15:54:42 and that can't be fixed? Sep 24 15:55:22 nbd: it can, but it's not a trivial task, and I'm not the author. Sep 24 15:55:44 hm Sep 24 15:56:49 the host compiler needs to assemble the heap of the target that will be used to bootstrap the target environment. it's not trivial at all, and needs good infrastructure, otherwise you'll spend your life debugging it Sep 24 15:57:09 I've seen a bit of how it happens on SBCL and it can be tricky Sep 24 16:08:28 nbd r47045 trunk/target/linux/generic/patches-3.18/ 092-03-spi-Only-idle-the-message-pump-in-the-worker-kthread.patch 092-02-spi-Pump-transfers-inside-calling-context-for-spi_sy.patch 092-01-spi-Check-to-see-if-the-device-is-processing-a-messa.patch Sep 24 16:08:28 kernel: backport some SPI layer improvements from 4.1 to 3.18, significantly improves flash speed on ramips Sep 24 16:08:53 nbd r47046 branches/ chaos_calmer/target/linux/generic/patches-3.18/092-01-spi-Check-to-see-if-the-device-is-processing-a-messa.patch chaos_calmer/target/linux/generic/patches-3.18/092-03-spi-Only-idle-the-message-pump-in-the-worker-kthread.patch chaos_calmer/target/linux/generic/patches-3.18/092-02-spi-Pump-transfers-inside-calling-context-for-spi_sy.patch Sep 24 16:08:53 OpenWrt Sep 24 16:08:53 kernel: backport some SPI layer improvements from 4.1 to 3.18, significantly improves flash speed on ramips Sep 24 16:15:26 when I work on the packaging, what should I touch to make menuconfig recompile? Sep 24 16:16:09 * attila_lendvai is staring at http://wiki.openwrt.org/doc/devel/packages but cannot find anything Sep 24 16:29:18 attila_lendvai: you want to recompile the kconfig interprepter? Sep 24 16:30:33 jow_laptop: ah, it's an interpreter? ok, less assuming from my side: I've created a new dir under feeds/packages/lang/ecl and I don't see the new package when I run make menuconfig. Sep 24 16:31:41 thats the wrong directory Sep 24 16:31:52 put it directly under package/ while developing it Sep 24 16:33:59 that worked, thanks! Sep 24 16:49:44 do I need to write Package/ecl/install, listing every single file, or can I just say somehow to package everything the package created to go under /usr/ ? Sep 24 16:52:57 you can use PKG_INSTALL:=1, this will call upstreams "make install" with prefix set to $(PKG_INSTALL_DIR) Sep 24 16:53:12 in your /install section you can then copy from $(PKG_INSTALL_DIR)/usr to $(1)/usr Sep 24 16:53:39 but usually you want to skip stuff like /usr/share/doc, /usr/share/man etc. Sep 24 17:00:17 hi Sep 24 17:00:53 i'd like to drop ownership of some packages, should i just delete them? Sep 24 17:01:09 tripolar: which ones? Sep 24 17:01:44 jow_laptop rtorrent dovecot Sep 24 17:02:09 rtorrent is a mess and i don't use it anymore, dovecot is nice but i don't use it anymore too Sep 24 17:02:21 we could move them into the abandonned repo Sep 24 17:02:27 yeah would be nice Sep 24 17:02:47 libtorrent too it belongs to rtorrent Sep 24 17:03:03 can you move it? Sep 24 17:03:48 will see if I can do Sep 24 17:11:59 if I touch the makefile then the host dependencies are also rebuilt? that hurts, it takes minutes... using make package/ecl/compile Sep 24 17:19:28 I need to call the ./configure script twice. for the first time it errors with a message to edit a file, and the second time it works fine. any ideas how to do that? $(call Build/Configure/Default,) stops the build when it fails the first time... Sep 24 17:19:50 better fix the root cause Sep 24 17:19:58 don't hardcode workarounds into the build recipes Sep 24 17:21:07 but this is "normal". I mean that file needs changing based on whether the target is big/little endian, etc... hrm, maybe I should just add the file, and copy it there before calling ./configure Sep 24 17:21:40 yep Sep 24 17:21:45 jow_laptop did it work? Sep 24 17:22:01 tripolar: I can't atm, not having all required keys here Sep 24 17:22:10 tripolar: will do when at home Sep 24 17:28:54 "Can't locate Thread/Queue.pm" that's not in the build depnds check yet is it? Sep 24 17:29:33 karlp: uhm ... context? Sep 24 17:31:05 trunk make, on a fresh linux desktop, failed to build autotools Sep 24 17:31:50 autconf or automake? Sep 24 17:32:00 automake Sep 24 17:32:02 jow_laptop thanks Sep 24 17:33:34 karlp: nasty - it depends on how the distro builds perl Sep 24 17:33:52 karlp: redhat stuff apparently still didn't gave up on perl threads Sep 24 17:34:49 eh, it's no biggy, I've installed the package and continued, just knew you were fiddling with this stuff recently. Sep 24 17:35:03 karlp: yeah I guess we can add a prereq check like perl -MThread::Queue -e 1 Sep 24 17:35:53 I've copied the perl makefile, and I'm wondering if there's a way to avoid recompiling host tools when I issue make package/ecl/compile... I tried to save the file without updating the modification time, but the host still gets recompiled (and it takes minutes) Sep 24 17:36:20 * attila_lendvai means he used the makefile of the perl package as a template, but it's probably irrelevant Sep 24 17:37:08 karlp: can you test this? http://pastebin.com/jytDspWV Sep 24 17:37:37 attila_lendvai: it likely has a BUILD_DEPENDS := foo/host somewhere Sep 24 17:37:48 attila_lendvai: you could temporarely disable that Sep 24 17:38:03 jow_laptop: ah, right! I'm slow... thanks! Sep 24 17:44:35 I need to make sure that the path contains the host bin/ when I call ./configure. any ideas how? shall I just stop using "$(call Build/Configure/Default,)"? will I lose something? Sep 24 17:44:55 ah, no, I'll lose a lot Sep 24 17:45:26 jow_laptop: yep, that patch works, I removed perl-Thread-Queue, distclean in that new tree I was starting up and then make menuconfig failed for me on the prereq check Sep 24 18:23:53 jow_laptop: and maybe you have an idea on how to avoid rebuilding when testing the packaging? :) Sep 24 18:53:50 attila_lendvai: rebuilding what? aren't you trying to test the packaging? don't you want the package rebuilt? Sep 24 18:54:09 make package/yourthing/{clean,compile} V=s might be useful if that's what you're looking for? Sep 24 18:54:10 jow_laptop: thanks for all the help, I have a v0.1 working! Sep 24 18:54:28 karlp: I wanted to avoid recompiling, but managed to fix it in a few iterations Sep 24 18:54:51 my problem was that compile took minutes, it's a big thing Sep 24 18:54:56 * attila_lendvai needs to run Sep 24 19:32:51 build #93 of ar7 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/93 Sep 24 22:16:06 rmilecki: build for wgt634u failed. logs/target/linux/compile.txt says: scripts/Makefile.build:403: recipe for target 'crypto/ocf' failed Sep 24 22:33:28 CC [M] crypto/ocf/ubsec_ssb/ubsec_ssb.o Sep 24 22:33:28 crypto/ocf/ubsec_ssb/ubsec_ssb.c:181:5: error: 'SSB_DEVTABLE_END' undeclared here (not in a function) Sep 24 22:33:31 SSB_DEVTABLE_END Sep 24 22:33:34 ^ Sep 24 22:58:57 * russell-- just replaced it with "{ 0, }," as in previous kernels Sep 24 23:29:26 looks like there was a recent (february) release of ocf-linux after being kind of stagnant for a long time Sep 25 01:46:52 sweet, found another bcm96368MVWG.dts board Sep 25 02:25:27 build #95 of xburst is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/95 Sep 25 02:38:42 build #93 of mxs is complete: Failure [failed shell_14] Build details are at http://buildbot.openwrt.org:8010/builders/mxs/builds/93 **** ENDING LOGGING AT Fri Sep 25 02:59:58 2015