**** BEGIN LOGGING AT Fri Apr 27 03:00:00 2012 Apr 27 04:39:06 build #4 of ppc40x is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/4 Apr 27 05:30:10 juhosg * r31487 /packages/multimedia/minidlna/Makefile: Apr 27 05:30:10 packages/minidlna: fix config file installation Apr 27 05:30:10 Reported-by: Gabor Varga Apr 27 06:30:58 build #4 of orion is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/4 Apr 27 06:31:45 juhosg * r31488 /trunk/target/linux/generic/config-3.3: linux/3.3: yet another missing symbol Apr 27 07:02:25 build #4 of uml is complete: Failure [failed compile_10] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/4 Apr 27 08:12:08 build #4 of cobalt is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/4 Apr 27 08:50:20 build #5 of brcm47xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/5 Apr 27 08:50:21 build #4 of brcm63xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/4 Apr 27 09:36:37 can I used curl configure OpenWrt? Apr 27 09:37:11 like curl http://192.168.0.1/cgi-bin/luci/admin/network/wireless/: for get wireless list Apr 27 09:37:44 curl http://192.168.0.1/cgi-bin/luci/admin/network/wireless/ssid=ssid;passwd=passwd for connec to one of them? Apr 27 09:37:50 is that hard? Apr 27 09:38:40 if you've enabled the json-rpc module, you can do things like that Apr 27 09:38:47 http://luci.subsignal.org/trac/wiki/Documentation/JsonRpcHowTo Apr 27 09:42:51 * xiangfu reading http://luci.subsignal.org/trac/wiki/Documentation/JsonRpcHowTo Apr 27 09:45:41 karlp, install this package: luci-mod-rpc ? Apr 27 09:48:17 try to install luci-mod-rpc now. Apr 27 09:54:22 what's the difference between a module and an application? Apr 27 09:54:51 I have an application I want to add luci support for, that should be an application right? Apr 27 09:55:17 karlp, I think so. Apr 27 09:55:21 the luci docs only talk about "how to write a module" Apr 27 09:56:13 karlp, I finish install luci-mod-rpc. can you give me more advice? I tried: 'http://192.168.42.1/cgi-bin/luci/rpc/sys" but it give error. Apr 27 09:56:42 karlp, I think you needs write two part. 1. the application Makefile 2. add a luci mod Apr 27 09:57:07 karlp, for example. how can I get the 'dmesg' output? Apr 27 10:05:39 karlp: module vs. application is merely a convention Apr 27 10:06:09 karlp: simply speaking a module is a thing that covers many tasks (like the admin area) and an application is a thing that covers only one application (like samba) Apr 27 10:07:12 for apps that manage the config file for a third party app, is it common to have the luci pages write and store into a uci file in /e/c, and on "apply" parse that file and generate an appropriate config file for the other app and restart it? Apr 27 10:07:41 or do some people try and hve the lua code behind pages directly parse the form and construct and store into the apps' native config format? Apr 27 10:08:07 I was thinking that having a stored uci config would get some shared benefits from ucitrack and so on Apr 27 10:08:13 and being able to use the cbi models? Apr 27 10:08:18 yes Apr 27 10:09:03 writing native non uci configs is way more complex as you have to reimplement most of what cbi already does for uci files Apr 27 10:09:12 yeah, that was what I was feeling Apr 27 10:09:22 seemed like a lot of extra work just to not store parts ofa config file twice Apr 27 10:09:24 or you need to write some kind of uci-to-native wrapper Apr 27 10:09:43 e.g. ini style configs can be mapped quite well to uci Apr 27 10:10:15 however in most cases its way simpler to just generate whatever needed format from uci Apr 27 10:10:47 so the cbi model has an onapply hook to not just save into uci, but also to do the generation Apr 27 10:11:14 or is it safe to put that in the init script to always parse out the uci file into the native file? Apr 27 10:11:33 so that anyone making a change via vi on the device and running the init script doesn't get any surprises? Apr 27 10:11:46 usually the init script does the conversion Apr 27 10:12:01 since we always support non-gui operation Apr 27 10:12:13 so it makes no sense to move the conversion logic into luci Apr 27 10:12:31 yeah, good points. was just trying to go through how this would all work in my head. Apr 27 10:13:02 there are also various approaches Apr 27 10:13:12 sometimes we map the complete subset of native options to uci Apr 27 10:13:17 like done for dropbear or openvpn Apr 27 10:13:29 sometimes we only map a subset which gets filtered through templates, like samba Apr 27 10:13:39 yeah, the samba one is what I'm looking at now. Apr 27 10:13:47 it looks like a pretty good match for me. Apr 27 10:25:43 where does the luci-samba application get into the make menuconfig options? there's no Package definition in the makefile at http://luci.subsignal.org/trac/browser/luci/trunk/applications/luci-samba/Makefile ? Apr 27 10:26:29 openwrt package definitions are all in .../contrib/package/luci/Makefile Apr 27 10:27:27 but there's nothing stopping me / wrong with putting the luci stuff into a sub pacakge in the applications general package makefile right? Apr 27 10:28:06 no Apr 27 10:36:42 you mentioned earlier that I need a Build/Prepare step like: http://www.pastebay.net/672787 Apr 27 10:36:59 but that doesn't have a way of specifying the package part, it seems to be for the entire package? Apr 27 10:37:15 is that why the luci support parts are often broken out into a separate makefile? Apr 27 11:00:00 yes Apr 27 11:38:57 in a typedsection, it just shows, "This section contains no values yet" Apr 27 11:39:06 if I add .addremove=true, I can make multiple copies of it, Apr 27 11:39:21 isn't there a way to have it show the options, even if there isn't anything in the underlying uci file yet? Apr 27 11:39:50 yeah but its complex Apr 27 11:40:15 so it's better to just ship something with the defaults in the config file :) Apr 27 11:40:16 you have to fake an empty section Apr 27 11:41:18 in the end I did it like this in various places: http://luci.subsignal.org/trac/changeset/7500 Apr 27 11:41:49 it will present a "Setup" button instead of the "This section contains no values yet" message and create a new section with defaults when pressed Apr 27 11:42:47 that dictionary of key/values correspond to optionnames? (start/limit/leasetime etc) Apr 27 11:42:55 yes Apr 27 11:43:47 ok, that should work. I have some basic config that can be shipped out in the default uci file, but they can add multiple whole sections for bridges, each of which have a set of options, Apr 27 11:43:58 looks like I want that sort of "add whole new section with defaults" Apr 27 11:44:06 not the cbi add section thing. Apr 27 11:44:11 more experimentation! Apr 27 11:44:13 here's another example, this time for the ntp server setup: http://luci.subsignal.org/trac/changeset/7915 Apr 27 11:47:23 for a while I simply used to create a uci section on the load of the page Apr 27 11:47:41 but that is irritating for the users since merely visting the form will result in unsaved changes Apr 27 11:47:56 without any further input Apr 27 11:48:16 huh, somehow now I've got a duplicate initial section. but saving the file only results in one section in the file on disk Apr 27 11:48:35 paste your model somewhere Apr 27 11:50:10 http://pastebin.com/zrKX7vrH Apr 27 11:53:30 oh, doh, this was a permissions thing Apr 27 11:53:49 killing and rerunning "make runhttpd" wsa saving changes to apply across runs Apr 27 11:55:30 ok Apr 27 12:05:36 jow * r31489 /trunk/toolchain/gcc/patches/ (2 files in 2 dirs): [toolchain] gcc: fix getenv() patches for 4.4.7 and 4.5-linaro Apr 27 12:19:02 if I'm running make runhttpd, is that meant to automatically reload models for me? Apr 27 12:19:31 yes Apr 27 12:20:33 in the dhcp example, it's always the config dhcp dhcp section, Apr 27 12:20:52 I want to have multiple "config bridge ????" sections, Apr 27 12:21:16 each with a set of values for each one, can I just let uci give them made up names? Apr 27 12:21:27 and keep the id for each one as an option within that section? Apr 27 12:21:33 probably Apr 27 12:21:59 otherwise they have to enter the name in a box somewhere before pressing the button to create a new section, Apr 27 12:55:22 heh, now *other* people are hassling for my patches to be merged :) Apr 27 12:56:32 jow_laptop: all working pretty beautifully thanks :) Apr 27 12:56:38 whether is far too nice to review patches today :P Apr 27 12:56:45 * jow_laptop is about to go to the beach... Apr 27 12:56:57 woo Apr 27 12:56:58 good times! Apr 27 13:00:13 hmm, even if the luci pages are in a separate package, the init script changes and default uci config should be in the parent package right? so that you can still properly use uci from teh command line Apr 27 13:00:21 yeah, that must be it. Apr 27 13:00:23 yes Apr 27 13:00:57 karlp: I've been doing luci stuff (for proto-openconnect and proto-l2tp) as patches against the luci repo itself, so those packages are generated as subpackages from the luci build Apr 27 13:01:13 rather than doing a completely separate source package. Apr 27 13:01:42 I'm not sure I'll ever make a complete enough config parser, Apr 27 13:02:05 it's why I was originally looking at trying to put the luci as an option inside the main package's makefile, Apr 27 13:02:32 I'm not quite sure where the right home is for it if I don't do full support for the entire config Apr 27 13:02:50 dwmw2_gone: I plan to rework it soon so that lucis build facilities get installed into openwrt for other packages to use Apr 27 13:03:06 dwmw2_gone: similar how python installs python.mk for python extension packages Apr 27 13:03:17 in staging_dir Apr 27 13:03:18 nice Apr 27 13:03:31 dwmw2_gone: then I'd also need to get patches into both luci for the app and packages for teh init script changes, Apr 27 13:03:58 karlp: yes. I have three git trees with outstanding patches. openwrt.git, openwrt-packages.git and luci.git :) Apr 27 13:04:31 yeah, and then there's the upstream package itself and then work with our configs to be installed after the standards. Apr 27 13:04:39 it gets a bit ugly sometimes :) Apr 27 13:05:23 well, time to write the uci->native converter and poke the init scripts a bit, see if this all works. Apr 27 13:06:34 off to the beach, bbl Apr 27 13:09:47 have fun! Apr 27 13:38:13 build #5 of at91 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/5 Apr 27 16:17:26 hmm, recompiling trunk on wr1043nd after a while, configured radvd, wide-dhcp6c, i get ipv6 address on pppoe-wan and br-lan, dhcp6c and dhcp6s are up but no ipv6 address on my wireless associated android ICS tablet Apr 27 16:29:43 hmm, something seems broken, cant figure out what Apr 27 16:34:06 ip -6 neigh shows nothin Apr 27 16:34:19 maybe some kernel 3.3.3 quirk with ipv6 Apr 27 16:35:36 got it Apr 27 16:35:41 no default ipv6 route Apr 27 16:44:45 florian * r31490 /trunk/package/kernel/modules/pcmcia.mk: [package] fix kmod-pcmcia-serial for 3.3+ kernels Apr 27 16:44:49 florian * r31491 /trunk/target/linux/brcm63xx/ (16 files in 2 dirs): [brcm63xx] switch to gpio-keys-polled Apr 27 16:49:18 florian * r31492 /trunk/package/base-files/ (Makefile files/lib/firstboot/05_firstboot_skip): [package] base-files: fix typo in 05_firstboot_skip script (#11359) Apr 27 17:21:11 nbd: ping pls Apr 27 17:24:04 bleh, ash vs /bin/sh differences. Apr 27 17:24:08 stupid shells cripting Apr 27 17:24:53 mooo! Apr 27 17:25:05 jow_laptop/nbd: ping-pong Apr 27 17:25:35 are you aware of any problems with futex on MIPs? This may well be eglibc-specific, I don't know. Apr 27 17:29:01 could be kernel too, who knows Apr 27 17:36:29 meh, this aint right, rebuilded trunk, configured radvd, wide-dhcp6c and dhcp6s, ppp with ipv6 exactly as a working older r30k dir-825 Apr 27 17:36:40 default ipv6 route aint coming up no more Apr 27 17:36:48 i can only add it manually Apr 27 17:42:03 cant figure out netifd, maybe something wrong in /lib/netifd.. ? Apr 27 18:14:18 build #3 of avr32 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/3 Apr 27 18:16:13 dape: well add debugging to /lib/netifd/ppp-up Apr 27 18:16:21 its invoked for both v4 and v6 Apr 27 18:16:40 see if $LLREMOTE is set on one invocation Apr 27 18:39:05 has 6in4 been tested with netifd? firewall's reflection.hotplug was missing its wanip using https://wiki.openwrt.org/doc/uci/network#static.ipv6-in-ipv4.tunnel Apr 27 18:45:31 dape: is IPv6 working at all on the ppp device? Apr 27 18:55:33 jow_laptop/nbd: ping Apr 27 19:32:54 jow_laptop: im not sure my skills allow me to figure out debugging, i hope you mean putting logger lines in the script? Apr 27 19:33:28 dwmw2_gone: yep, ipv6 enabled, i can manually add defaultipv6 route to fe80::1 and everything works great but wont come up automagically after ppp auths Apr 27 19:34:31 build #4 of x86 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/4 Apr 27 19:37:07 jow_laptop: maybe a line like logger -t LLREMOTE "LLREMOTE is $LLREMOTE" in ppp-up ? Apr 27 19:37:56 hmm, got it, seems that $LLREMOTE is.. empty? Apr 27 19:49:24 juhosg * r31493 /packages/net/nginx/Makefile: Apr 27 19:49:25 packages: Upgrade nginx to 1.0.15 Apr 27 19:49:25 Changelog: http://nginx.org/en/CHANGES-1.0 Apr 27 19:49:25 This upgrade includes a fix for a security vulnerability, CVE-2012-2089. Apr 27 19:49:25 Signed-off-by: Mark Mentovai Apr 27 19:49:27 juhosg * r31494 /packages/libs/slang2/Makefile: Apr 27 19:49:27 packages: Update s-lang library to 2.2.4 Apr 27 19:49:27 See ftp://space.mit.edu/pub/davis/slang/v2.2/changes.txt for changes Apr 27 19:49:27 Signed-Off-By: Jeroen van Bemmel Apr 27 19:51:55 hehe, nginx 1.2.0 went stable too recently Apr 27 20:17:53 what's the syntax of a dhcp CNAME entry? it's not in the wiki... Apr 27 21:18:22 build #4 of ppc44x is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/4 Apr 27 22:00:13 build #3 of sibyte is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/3 Apr 27 23:01:42 build #3 of ar7 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/3 Apr 27 23:07:23 build #4 of rb532 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/4 Apr 27 23:18:26 build #4 of ixp4xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/4 Apr 27 23:38:13 jow_laptop: ping Apr 28 00:41:24 How to find led gpio number? Apr 28 00:47:38 leds-gpio: probe of leds-gpio failed with error -16 Apr 28 00:47:51 How to fix this? Apr 28 01:03:54 what's the platform? Apr 28 01:05:02 build #3 of kirkwood is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/3 Apr 28 01:08:19 philipp64|laptop: arthoes ar9342 Apr 28 01:12:47 I have send a patch to add support for tl-wr1041n to ML. Apr 28 01:13:27 The only defect is the leds. Apr 28 01:50:09 build #6 of s3c24xx is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/6 Apr 28 01:53:57 whats this arthoes platform :P Apr 28 02:01:11 build #4 of au1000 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/4 Apr 28 02:30:39 build #4 of ar71xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/4 **** ENDING LOGGING AT Sat Apr 28 03:00:00 2012