**** BEGIN LOGGING AT Fri Aug 14 02:59:58 2015 Aug 14 08:20:08 build #76 of cobalt is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/76 Aug 14 08:22:41 build #76 of cns21xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/cns21xx/builds/76 Aug 14 08:25:27 build #76 of orion is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/76 Aug 14 10:21:40 jogo r46592 trunk/ (5 files in 5 dirs) * image.mk: move KDIR_TMP definition to a central location Aug 14 12:55:37 is there a way to enable PUT/DELETE in uhttpd? or is that just nto a feature that's supported? Aug 14 12:56:23 hrm, code actually seems to make that pretty clear that it's GET/POST/HEAD only. Aug 14 13:04:23 jogo r46593 trunk/target/linux/bcm53xx/image/Makefile * bcm53xx: prefix netgear specific device variables with NETGEAR_ Aug 14 13:05:08 jogo r46594 trunk/target/linux/brcm47xx/image/Makefile * brcm47xx: prefix netgear specific device variables with NETGEAR_ Aug 14 13:05:28 jogo r46595 trunk/target/linux/brcm63xx/image/Makefile * brcm63xx: prefix netgear specific device variables with NETGEAR_ Aug 14 13:05:59 jogo r46596 trunk/ include/image.mk target/linux/brcm47xx/image/Makefile target/linux/brcm63xx/image/Makefile target/linux/bcm53xx/image/Makefile * image.mk: replace all Build/netgear-chk with a generic implementation Aug 14 13:06:08 jogo r46597 trunk/ target/linux/brcm2708/image/Makefile include/image.mk target/linux/brcm63xx/image/Makefile * image.mk: store the device name variable Aug 14 13:07:19 jogo r46598 trunk/ (74 files in 6 dirs) * kernel: update 4.1 to 4.1.5 Aug 14 13:30:03 hm... anybody an idea where those "depends on !(USE_UCLIBCXX) || BROKEN" entries come from when the make system generates its files+dependencies ? Aug 14 13:30:45 they are in tmp/.config-package.in and seem to relate to uclibc++ packages only but I cannot see where BROKEN comes into the mix Aug 14 13:35:40 Hello! I have added the source code for BCM63168 on the wiki with kernel 3.4, unfortunately I am no c / c++ developer so I am unable to contribute more than that. Aug 14 13:40:39 well the interest is limited since the platform does not have open DSL drivers (beside the wireless issue with many broadcom chipsets being originally close too) Aug 14 13:47:16 karlp: right, its simply not implemented Aug 14 13:48:31 karlp: we could consider adding stub support, like accepting it but respond with 405 Not supported unless the request is routed to a cgi / lua handler Aug 14 13:49:11 plntyk: on what version? trunk? Aug 14 13:49:26 yes trunk Aug 14 13:49:37 plntyk: got an example entry? Aug 14 13:50:43 mom Aug 14 13:54:03 oh wait Aug 14 13:54:49 hm.... seems a local change - I marked package/uclibc++ as BROKEN on a local tree since it fails to build with gcc5 anyway Aug 14 13:55:14 ok Aug 14 14:05:00 jow_laptop: I am just wondering if that single change introduced a "Warning: recursive dependency" when doing make/make menuconfig ... (libmysql.. has CXX_DEPENDS) and then it reports strongswan-mod-mysql - strongswan-full are recursive Aug 14 14:24:24 jow_laptop: right, yeah, I have a lua handler that was expecting it, but it's not a huge deal. Aug 14 15:11:03 jogo r46599 trunk/target/linux/ ipq806x/patches-4.1/165-arm-qcom-dts-Enable-NAND-node-on-IPQ8064-AP148-platform.patch ipq806x/patches-4.1/707-ARM-dts-qcom-add-mdio-nodes-to-ap148-db149.patch ipq806x/patches-4.1/708-ARM-dts-qcom-add-gmac-nodes-to-ipq806x-platforms.patch ipq806x/patches-4.1/157-ARM-DT-ipq8064-Add-ADM-device-node.patch * ipq806x: disable Aug 14 15:11:03 OpenWrtdma for spi on linux 4.1 as well Aug 14 15:11:28 jogo r46600 trunk/include/image.mk * image.mk: add a fit recipe Aug 14 15:11:34 jogo r46601 trunk/ include/image.mk target/linux/brcm63xx/image/Makefile * brcm63xx: move gzip build command to include/image.mk Aug 14 15:11:41 jogo r46602 trunk/target/linux/ipq806x/image/Makefile * ipq806x: switch to new image build system Aug 14 16:15:17 speaking of an api, should luci.http.context.request.message.headers be a table of all headers received? I get an empty table? Aug 14 16:16:46 or is that only for setting headers on output? Aug 14 16:36:31 karlp: its empty in cgi mode since that does not expose headers Aug 14 16:36:53 karlp: do you need access to custom headers? Aug 14 16:42:41 luka: https://github.com/openwrt/packages/pull/1596 Aug 14 16:55:57 jow_laptop: yeah, that was the plan, Aug 14 16:56:41 was goign to use X-HTTP-Method-Override to let post be put/delete Aug 14 16:58:08 what's "cgi mode" ? this is just in a normal registered dispatcher/controller Aug 14 17:08:01 karlp: it means luci runs as cgi program and not via embedded lua runtime Aug 14 17:11:12 when you say "that does not expose headers" do yo umean that the cgi mechanism in uhttpd? there's no standards reason for cgi to not have access to the headers is there? Aug 14 17:11:58 the standardized cgi protocol does not contain any possibility to hand over headers to a cgi program Aug 14 17:12:49 the only means of communication are environment variables (to expose a subset of standard http headers in a mangled form) and stdin (to feed http request message bodies to the cgi program) Aug 14 17:14:12 what we could implement is mapping arbritary headers, like X-Foo-Bar: baz -> setenv("HTTP_X_FOO_BAR", "baz") Aug 14 17:14:29 iirc php is doing that Aug 14 17:14:52 perl too if I understand some of this web reading I'm doing... Aug 14 17:15:27 mod_perl maybe, perl in cgi mode can only work with whatever the server is passing on (apache etc.) Aug 14 17:15:33 I always think, yeah, I'll just do this thing, it will be modern and sane and cool, and then I run into th epile of junk that is web programming :| Aug 14 17:16:05 CGI.pm has a .http() method that seems to give you headers., Aug 14 17:16:16 it prefixes them with HTTP_ and allcaps them. Aug 14 17:16:32 CGI.pm is "standard" right? ;) Aug 14 17:17:02 its just a wrapper around getenv Aug 14 17:17:52 https://tools.ietf.org/html/rfc3875#section-4.1.18 Aug 14 17:17:52 $cgi->http("X-Foo-Bar") -> return $ENV{"HTTP_X_FOO_BAR"} -> getenv("HTTP_X_FOO_BAR") Aug 14 17:18:34 "The server is not required to create meta-variables for all the header fields that it receives. Aug 14 17:18:37 " Aug 14 17:18:41 not required, no. Aug 14 17:18:43 but SHOULD :) Aug 14 17:20:19 * karlp shrugs, just tyring to work out best to work with this. Aug 14 17:22:49 so you're not using uhttpd-mod-lua ? Aug 14 17:22:55 just defaults? Aug 14 17:23:24 yeah Aug 14 17:23:35 hm Aug 14 17:23:38 source patch then Aug 14 17:23:41 I tried mod-lua a while ago, had some issue that I don't remember anymore. Aug 14 17:24:07 well, I'm defining the api, but with neither PUT/DELETE request, nor header support it leaves me down with adding post parameters, which are kinda gross. Aug 14 22:06:43 build #72 of mpc83xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/72 Aug 15 00:09:04 build #72 of ep93xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/72 **** ENDING LOGGING AT Sat Aug 15 02:59:58 2015