**** BEGIN LOGGING AT Fri Mar 27 02:59:57 2009 Mar 27 03:16:41 Who here handles Atheros development? Mar 27 05:28:41 Zombie: nbd for the most part Mar 27 05:30:41 I have some hardware I want analyized Mar 27 05:30:58 Zombie: what hardware? Mar 27 05:32:17 Its a LevelOne WBR 6000 Mar 27 05:33:05 It has an Atheros AR7130 Mar 27 05:33:07 oh, an ap81 board. that would be juhosg's domain Mar 27 05:33:39 oh, it has an actual ar7130 on the board? I guess it's an earlier design w/ minipci Mar 27 05:34:03 I'm willing to ship this hardware under the condition it is eventually returned. Mar 27 05:35:24 you would have to work out those details with him, and round-trip shipping likely may cost as much as the device itself (he's in hungary) Mar 27 05:36:30 juhosg is still the one to talk to, he's supreme master of the ar71xx target :) Mar 27 05:37:56 Yeah... Mar 27 05:38:05 Thats what RoundSparrow told me. Mar 27 13:49:14 juhosg: you see any way to recover the Planex without a serial port? I saw no sign of a tftp or other recovery like on the Trendnet or Netgear Mar 27 13:56:44 hi Mar 27 13:56:50 outlaw45: what? Mar 27 13:57:00 anyone could assist with some compiling problems? Mar 27 13:57:11 trying to compile rxtx for openwrt kamikaze but make fails on including sys/io.h, anyone has an idea? Mar 27 14:01:32 trying to compile rxtx for openwrt kamikaze but make fails on including sys/io.h, anyone has an idea? Mar 27 14:06:00 <[Fate]> lo Mar 27 14:06:09 hi Mar 27 14:06:15 trying to compile rxtx for openwrt kamikaze but make fails on including sys/io.h, anyone has an idea? Mar 27 14:06:16 outlaw45: what? Mar 27 14:11:44 <[Fate]> mh shouldn't "libpthread.so.0" by default be available in an openwrt build? Mar 27 14:12:21 if no application needs it, why should it be included :-) Mar 27 14:12:32 I believe you have to install the pthread package Mar 27 14:12:44 I'm trying to compile rxtx for openwrt kamikaze but make fails on including sys/io.h, anyone has an idea? Mar 27 14:13:11 outlaw45: I think you already made clear what your problem is Mar 27 14:13:13 <[Fate]> I'm seen no such package in "make menuconfig" Mar 27 14:13:20 <[Fate]> I've Mar 27 14:13:28 <[Fate]> thought it's part of uClibc probably Mar 27 14:13:51 CyrusFF: I know the problem, just dont understand why io.h isn't included in the toolchain... Mar 27 14:14:04 <[Fate]> try verifying that statement Mar 27 14:14:20 <[Fate]> look at how the package is built, with which include paths Mar 27 14:14:30 Fate: I needed the phtread package for jamvm and found it in the openwrt repository Mar 27 14:14:40 it works at the moment... :) Mar 27 14:15:35 <[Fate]> gna Mar 27 14:15:36 <[Fate]> found it Mar 27 14:15:38 <[Fate]> part of busybox config Mar 27 14:15:44 <[Fate]> nah Mar 27 14:15:47 <[Fate]> base system that is Mar 27 14:15:51 :) Mar 27 14:15:56 <[Fate]> blindness is a virtue Mar 27 14:16:05 <[Fate]> but I've another bug with openwrt DEPENDS Mar 27 14:16:18 haha, it's no problem if you learn from your mistakes :) Mar 27 14:16:19 <[Fate]> if I add the required libs as DEPENDS in my package Makefile Mar 27 14:16:31 <[Fate]> they are built as packages (I can see them in bin/packages/) Mar 27 14:16:36 <[Fate]> but they don't appear in the final image for some reason Mar 27 14:17:07 <[Fate]> they do, if I don't use dependencies and select them manually Mar 27 14:18:15 don't know that.... Mar 27 14:19:13 I'm using this channel to get answers to my own question :) Mar 27 14:19:38 [Fate]: could you paste your DEPENDS-line Mar 27 14:20:07 <[Fate]> DEPENDS:=+libpcap +libexpat Mar 27 14:21:32 and if you select the package with this dependencies as bultin (*) what does show up beneath libpcap and libexpat? Mar 27 14:23:48 <[Fate]> "--" Mar 27 14:24:05 ok as it should Mar 27 14:25:06 are there any warnings regarding dependencies when you run make or make menuconfig? Mar 27 14:25:20 otherwise I have no idea what could cause this Mar 27 14:33:27 I'm trying to compile rxtx for openwrt kamikaze but make fails on including sys/io.h, anyone has an idea? Mar 27 14:38:46 <[Fate]> make[3]: *** No rule to make target `package/preconfig'. Stop. Mar 27 14:39:00 <[Fate]> I see this immediately before kernel building Mar 27 14:39:04 <[Fate]> but I don't think it's related Mar 27 18:32:29 hi folks Mar 27 18:32:35 aroscha? Mar 27 18:33:04 wondering where to specify a different athX_hal.tgz file Mar 27 18:42:06 i can only select the madwifi stable default branch Mar 27 18:43:14 it's the only supported version Mar 27 18:57:42 hmmm :( Mar 27 18:57:42 i thought i saw that on the fon target Mar 27 18:57:42 that you can chose a different hal Mar 27 19:01:00 what is actually the current policy on startup scripts for apps? is there a common policy like "all external packages should come with an /etc/init.d/* and /etc/rc.d/* startup script" ? Mar 27 19:01:11 or is that always left to the integrator Mar 27 19:07:06 all startup scripts should have '#!/bin/sh /etc/rc.common' as the first line Mar 27 19:08:02 then /etc/init.d/foo enable/disable work, and various other functions are defined. Mar 27 19:08:37 also need to define $START so the rc.d link is started in the right order Mar 27 20:49:27 bartman007: ok, thx. But is the policy to have a startup script for every app out there, or is writing the startupscript in case it does not exist left to the user? Mar 27 20:49:33 like i have iperf for example Mar 27 20:49:38 i like to run it as a server Mar 27 20:49:46 but this might not be useful for everybody Mar 27 20:50:17 does it make sense to have a /etc/config/iperf file which has an option "runAsServer" $bool ? Mar 27 20:50:21 and a startup script Mar 27 20:50:22 ? Mar 27 20:50:45 if we do that for every app, then just checking the startup scripts for every installed app is already a long process Mar 27 20:50:46 no? Mar 27 20:50:50 what do you think? Mar 27 21:15:02 nbd * r15055 /trunk/package/ (3 files in 3 dirs): wifi: fix hostapd + autochannel Mar 27 21:30:15 grml..... the olsrd script by alina does not work :(( Mar 27 21:30:20 i think we should replace it Mar 27 21:32:29 it actually isn't far off though Mar 27 21:32:42 it is mismatching interfaces Mar 27 21:32:52 with a little tweak it got close for me Mar 27 21:33:29 russell: you got it working? Mar 27 21:33:46 working ish Mar 27 21:33:52 hehe Mar 27 21:34:00 i just made my own small one now Mar 27 21:34:02 hold on, i'm plugging in my development device Mar 27 21:34:07 ok Mar 27 21:34:12 can you mail me the diff? Mar 27 21:34:47 i can, but it's a bit hacky Mar 27 21:34:56 i just bypass a check Mar 27 21:35:20 and it needs a corresponding /etc/config/olsrd Mar 27 21:37:49 aroscha: http://openwrt.pastebin.com/f708d10d2 Mar 27 21:38:22 look in /var/run/olsrd.conf to see what comes out Mar 27 21:40:21 sorry, /var/etc/olsrd.conf Mar 27 21:42:21 my test config looks like this: http://openwrt.pastebin.com/f71ce257 Mar 27 21:43:16 but, i may be misunderstanding how alina wanted the interfaces to be specified Mar 28 01:41:34 aroscha_: including the #! line I mentioned allows the use to control which services run by running `/etc/init.d/foo enable/disable' Mar 28 01:42:19 if a package in included in an image it will be enabled by default; if installed via opkg it is disabled by default Mar 28 01:45:51 and yes ideally we will have a startup script for every app that benefits from it. deciding on whether to create a UCI config is a matter left to your discretion. UCI makes writing LUCI/X-Wrt modules for the application trivial, and ensures that the config wil be saved when using sysupgrade (though I have a few other ideas about how to achieve that automatically) Mar 28 02:19:50 Bartman007: that should be written down somewhere! ;-) Mar 28 02:20:49 along with a whole philosophy of /etc/config/foo document Mar 28 02:27:45 russell_: agreed. I have lots of work to do once we launch the new wiki **** ENDING LOGGING AT Sat Mar 28 02:59:57 2009