**** BEGIN LOGGING AT Sun Feb 21 02:59:57 2010 Feb 21 03:20:01 build #14 of mpc52xx is complete: Exception [exception failed slave lost shell_19 compile_15] Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/14 Feb 21 03:29:05 looks like the wordpress trac has such an auto has-patch tagger if their trac code can be found Feb 21 06:57:33 build #16 of ep93xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/16 Feb 21 07:45:23 builder status Feb 21 07:45:35 builder : help Feb 21 07:45:41 builder: help Feb 21 07:45:41 Get help on what? (try 'help ', or 'commands' for a command list) Feb 21 07:45:50 builder: comands Feb 21 07:45:57 builder: commands Feb 21 07:45:57 buildbot commands: commands, dance, destroy, excited, force, hello, help, join, last, leave, list, notify, source, status, stop, version, watch Feb 21 07:46:13 builder: status brcm63xx-trunk Feb 21 07:46:13 no such builder 'brcm63xx-trunk' Feb 21 07:46:19 builder: list Feb 21 07:46:20 try 'list builders' Feb 21 07:46:29 builder: list builders Feb 21 07:46:29 Configured builders: adm5120 ar7 ar71xx atheros au1000 avr32 brcm47xx brcm63xx brcm_2_4 cobalt ep93xx etrax gemini goldfish ifxmips iop32x ixp4xx kirkwood mpc52xx octeon orion ppc40x ppc44x ps3 pxcab ramips rb532 rdc s3c24xx sibyte ubicom32 uml x86 xburst Feb 21 07:46:42 builder: status brcm63xx Feb 21 07:46:43 brcm63xx: idle, last build 6h18m27s ago: build successful Feb 21 08:34:09 build #14 of gemini is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/14 Feb 21 10:18:01 ping rtz2 Feb 21 11:19:39 <[florian]> cshore: ping? Feb 21 12:47:45 florian * r19788 /trunk/target/linux/cobalt/ (3 files in 2 dirs): [cobalt] add missing kernel configuration and patches for 2.6.32 Feb 21 12:52:49 is there some way to set some kernel options when selecting a profile, or does this require a subtarget? Feb 21 13:09:31 hi, how do I enable kernel debugging when building an image? Feb 21 13:52:57 sKAApGIF: global build options -> build krnel wich symbol table information Feb 21 14:14:19 florian * r19789 /trunk/package/mac80211/patches/560-ath9k_fix_eap_handling.patch: mac80211: fix an eap frame handling bug in ath9k - fixes client mode rekeying issues Feb 21 14:14:29 florian * r19790 /trunk/package/mac80211/patches/570-ap_vlan_injection.patch: mac80211: fix lookup for vlan'd stations for injected frames - fixes wpa rekeying with wds ap Feb 21 14:14:34 florian * r19791 /trunk/package/mac80211/patches/570-ap_vlan_injection.patch: mac80211: make the vlan injection patch more specific Feb 21 14:14:38 <[florian]> oh crap Feb 21 14:14:44 <[florian]> got trapped by git svn rebase Feb 21 14:16:51 lars * r19792 /packages/net/freeswitch/Makefile: Feb 21 14:16:51 [packages] freeswitch: Don't look for libgnutls-config. The package does not Feb 21 14:16:51 depend on gnutls and it will only find the hosts libgnutls-config anyway. Feb 21 14:27:18 florian * r19793 /trunk/target/linux/brcm63xx/files-2.6.30/arch/mips/bcm63xx/boards/board_bcm963xx.c: [brcm63xx] fix gpio-leds registration for boards having less than 8 LEDs (#6528) Feb 21 14:28:06 florian * r19794 /trunk/target/linux/au1000/ (Makefile au1500/config-2.6.32): [au1000] switch to 2.6.32, resync kernel config, enable gpio sysfs support Feb 21 14:59:50 {Nico}: ping Feb 21 15:01:37 lars * r19795 /packages/Xorg/xorg/xserver/xorg-server/files/xorg-driver.mk: Feb 21 15:01:37 [packages] xserver-xorg: Fix setting verion suffix to CONFIGURED_STAMP in Feb 21 15:01:38 xorg-driver.mk Feb 21 15:31:57 lars * r19796 /packages/libs/speex/Makefile: Feb 21 15:31:57 [packages] speex: Provide ogg prefix path. Otherwise it will start looking for Feb 21 15:31:57 libogg in host paths. Feb 21 15:43:10 florian * r19797 /trunk/package/mac80211/patches/ (560-ath9k_fix_eap_handling.patch 570-ap_vlan_injection.patch): [package] fix accidently commited ancient mac80211 patches in r19789-19791 (#6716) Feb 21 16:48:19 thepeople * r19798 /trunk/target/linux/avr32/config-default: add missing symbol Feb 21 20:00:22 xMff: wouldn't it be a good idea to add a "recommanded devices" page to the wiki and link it on the frontpage? Feb 21 20:00:39 xMff: with say 5 devices from different categories on it? Feb 21 20:00:57 xMff: there are at least 2 people a day asking that question Feb 21 20:02:46 <_trine> rtz2, but would that ever be up to date Feb 21 20:03:13 <_trine> it would be a constantly changing situation Feb 21 20:03:36 _trine: it doesn't change that much Feb 21 20:03:52 _trine: if a device works well now, it will still do so in a year Feb 21 20:05:08 rtz2: yes, if someone does it and documents all of the candidates well enough (install, recovery, upgrade procedures ...) Feb 21 20:05:54 well, I don't which router is supported well and which isn't :/ Feb 21 20:06:05 <_trine> rtz2, that's true but they wound not be the latest equipment,, as that would be constantly changing,, you could finish up recommending a wrt54g Feb 21 20:06:33 <_trine> it would need to be regularly updated Feb 21 20:06:47 _trine: you don't need the latest equipment Feb 21 20:06:52 all information needs to be regulary updated :) Feb 21 20:06:54 _trine: only something that works Feb 21 20:07:26 <_trine> rtz2, fo me,, I want the newest if I'm looking to buy Feb 21 20:07:32 people having that question usually don't get an answere at all at the moment Feb 21 20:07:54 <_trine> rtz2, you do have a good point though Feb 21 20:08:08 _trine: well, the differences aren't that big between a 2 year old model and a new one Feb 21 20:08:08 <_trine> there should be something Feb 21 20:08:19 I don't think it has to be the newest always, just something that is actually 95% - 100% supported Feb 21 20:08:21 maybe n support, but if you can do without that ... Feb 21 20:08:26 right Feb 21 20:08:57 <_trine> if you are going to buy a new router it should be the newest I think Feb 21 20:08:57 and it should be out of "read this forum thread, check that ticket and recompile this revision" phase Feb 21 20:09:14 yes Feb 21 20:09:25 but I have no clue, for what this applays Feb 21 20:10:23 I think it might still take some month until stuff like the wrt160nl are truly stable Feb 21 20:10:38 the ubn gear works well now, like ns2, ns5, picostation, bullet Feb 21 20:10:54 but those are not "basement routers" Feb 21 20:10:59 they're more for outdoor stuff Feb 21 20:11:20 xMff: doesn't matter that much Feb 21 20:11:38 xMff: this would be one category Feb 21 20:12:01 xMff: maybe something really cheap with only g support Feb 21 20:12:09 and something with usb Feb 21 20:12:12 all brcm-2,4 devices ;) Feb 21 20:12:26 usb ... well the msot good devices I know are eol'd Feb 21 20:12:52 <_trine> maybe there should be a section with new routers that are very likely to be stable in the medium future Feb 21 20:13:44 <_trine> the wrt160nl and the wndr3700 and the dir 825 might fit that category Feb 21 20:13:51 but such a thing would still be very useful, even if it's not completly up-to-date Feb 21 20:14:27 hauke * r19799 /trunk/ (5 files in 2 dirs): Feb 21 20:14:27 ext4: add Support for large (2TB+) block devices and files Feb 21 20:14:27 CONFIG_LBDAF is needed for mounting ext4 filesystems with huge_file enabled. Feb 21 20:14:27 This is the default option. Feb 21 20:14:27 This closes #6714. Feb 21 20:15:08 http://www.amazon.de/TP-Link-300Mbps-Ultimate-Wireless-Netzwerk/dp/B002YETVTQ/ref=sr_1_1?ie=UTF8&s=ce-de&qid=1265902525&sr=8-1 Feb 21 20:15:21 this one has the potential to bwecome a wrt54 killer Feb 21 20:15:31 in terms of price, features, and support Feb 21 20:16:27 <_trine> I was looking at that last week Feb 21 20:16:37 <_trine> someone else mentioned it Feb 21 20:18:51 the TL-WR1043ND is the big brother of the one linked above Feb 21 20:23:05 sweet, gigabit Feb 21 20:25:38 <_trine> often manufacturers in their advertisements don't mention the processor speed Feb 21 20:26:38 hauke * r19800 /trunk/package/kernel/modules/fs.mk: ext4: fix for commit r19799 Feb 21 20:27:31 build #14 of au1000 is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/au1000/builds/14 Feb 21 20:27:31 tripolar * r19801 /packages/net/ntpd/Makefile: add new ntpd-openssl package to get a ntpd that is linked against openssl Feb 21 20:38:39 tripolar * r19802 /packages/net/ntpd/Makefile: fix typo Feb 21 20:55:02 lars * r19803 /packages/admin/osiris/patches/010-dont-add-host-include-paths.patch: [packages] osiris: Don't add host include paths Feb 21 21:02:42 lars * r19804 /packages/Xorg/lib/imlib2/Makefile: Feb 21 21:02:42 [packages] imlib2: Pass Xorg include and libarary path. Otherwise it will look Feb 21 21:02:42 for it for them in the host environment. Feb 21 21:37:31 lars * r19805 /packages/net/kismet/ (3 files in 2 dirs): [packages] kismet: Don't add host include paths Feb 21 21:42:11 tripolar * r19806 /packages/net/ntpd/Makefile: use ntpd instead of because == ntp Feb 21 22:05:16 build #18 of octeon is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/18 Feb 21 22:05:36 lars * r19807 /packages/net/snort/patches/010-dont-add-host-paths.patch: [packages] snort: Don't add host include and libraray paths Feb 21 22:56:03 thepeople * r19808 /trunk/target/linux/generic-2.6/config-2.6.30: fix missing symbol Feb 21 23:05:06 build #17 of ep93xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/17 Feb 21 23:26:42 thepeople * r19809 /trunk/target/linux/generic-2.6/config-2.6.32: fix missing symbol Feb 21 23:39:40 ping nbd Feb 21 23:39:51 pong Feb 21 23:40:25 I'm looking at the busybox blkid and it is missing some functionality I was using Feb 21 23:40:50 the busybox blkid only shows uuid, nothing else Feb 21 23:41:00 what do you need? Feb 21 23:41:22 I was also using blkid to get the filesystem type (for the auto fsck) Feb 21 23:41:34 in addition I was using it to find a particular uuid Feb 21 23:41:49 though I could do that with a shell function and the busybox blkid Feb 21 23:42:14 and the busybox blkid doesn't show the filesystem label (for mounting via label) Feb 21 23:42:35 the autofsck is probably not that big a deal... I could drop that Feb 21 23:42:49 is it hard to add the missing functionality to busybox? Feb 21 23:42:50 and only do fsck for defined in fstab filesystems Feb 21 23:43:09 nbd: no idea...I'll have to look at it Feb 21 23:45:19 nbd: it's certainly more involved that what I was hoping for...I don't know the busybox code at all Feb 21 23:51:50 apparently blkid has support for labels Feb 21 23:52:02 maybe you're missing a config option somewhere Feb 21 23:52:07 hmm....ok Feb 21 23:53:10 Usage: blkid Feb 21 23:53:11 Print UUIDs of all filesystems. Feb 21 23:53:18 is all that the blkid --help shows Feb 21 23:54:50 ok, I just tested....it shows the label if present Feb 21 23:55:00 I just didn't have a label Feb 21 23:55:15 or maybe you need busybox 1.16 ;) Feb 21 23:55:26 :) Feb 21 23:55:26 the bb help tests aren ot always correct, sometimes they're incomplete, sometimes they're wrong, ... Feb 21 23:55:30 *texts Feb 21 23:56:33 I'm inclined to just ditch the auto determine fs type and fsck for automount, and only do fsck for defined fstypes in /etc/config/fstab Feb 21 23:57:01 maybe add it back later Feb 21 23:57:23 can't you just throw all available fsck utils at the fs until one succeeds? Feb 21 23:57:59 xMff: I think that would be dangerous...especially with the fix option Feb 21 23:58:18 well use dry-run then Feb 21 23:59:30 xMff: I'm rather nervous about trying to determine it that way....it's not vital features. Fsck if you know the fstype ahead of time still works without doing that Feb 22 00:00:30 xMff: so I think the fsck of devices *not* defined in /etc/config/fstab should just be shelved for now Feb 22 00:02:43 xMff: I'd hate to hose someone's filesystem by running the wrong fsck, because it thought it could repair an fs that was actually of a different type Feb 22 00:03:13 k Feb 22 00:03:36 xMff: we still have normal fsck, so the important feature is there Feb 22 00:04:09 xMff: and if I ever add the fstype detection to busybox we can add it back Feb 22 00:05:05 xMff: it should probably be fairly easy to do, but right now I've not got the time to go off on that sort of tangent Feb 22 00:06:05 xMff: I don't need the feature myself, it was just a nice idea and came for free with the other blkid Feb 22 00:06:21 I got your point Feb 22 00:06:25 sorry Feb 22 00:06:35 sometimes I talk a lot Feb 22 00:07:17 I *will* beat your over the head with my point...'cause I'm sometimes a nob that way Feb 22 00:08:01 my brother and sister used to complain about that Feb 22 00:09:11 ;) Feb 22 00:09:37 anyway, I'm just about ready with the new version of block-mount then Feb 22 00:11:27 I'ved also added the is_rootfs option to the fstab in this....so extroot will just call the mount by section function with the parameter for find_rootfs = 1, in which case only the rootfs wll be mounted Feb 22 00:15:36 cool Feb 22 00:17:20 with that and your idea about the modules, no extroot config file is necessary; just fstab Feb 22 00:18:43 oh, what do you think of ede's idea of making extroot a base system type thing....I don't think it's that core myself, but I'll do it if you want Feb 22 00:26:59 as long as we can users to flip a switch in menuconfig and can get rid of the dozens of shaky scripts, it doesn't matter were you put it :) I think the Utilities category would be a good choice Feb 22 00:27:40 that's where I have it now Feb 22 00:27:49 but it the packages feed Feb 22 00:27:50 +tell Feb 22 00:27:58 somehow I miss a lot of words lately Feb 22 00:28:02 me too Feb 22 00:28:08 AOL +1 Feb 22 00:28:30 it=in Feb 22 00:29:55 oh, the scripts have been much simplified in terms of the number and no more plugins in fstab anymore Feb 22 00:30:30 * xMff is tempted to start rewriting the network scripts in lua... Feb 22 00:31:18 using the shell for that getting to you? Feb 22 00:32:29 it is starting to approach the limits of ash, the biggest issue is the lack of stuff like arrays or sophisticated string functions Feb 22 00:33:02 imho Feb 22 00:33:06 xMff: yeah, once you start trying to do complex stuff shell isn't so good Feb 22 00:33:18 pppd is quirky in particular Feb 22 00:33:21 and pptp too Feb 22 00:33:53 xMff: I think some of the quirkiness in pptp comes from how the rest of the scripts are written Feb 22 00:34:04 xMff: that reminds me: I have modifed the /lib/network/ppp script so that it works properly with pppoe+pptp Feb 22 00:34:13 right now they're both trying to be ppp0 Feb 22 00:34:24 then there's a new ticket stating that the wan requests dhcp initially even if it is set to static, I suspect it is some race conditions where ifaces are brought up when the config is still from /rom and not jffs2 Feb 22 00:34:43 or something similar Feb 22 00:35:11 xMff: hmm...iface hotplug shouldn't be happening utnil /etc/init.d/boot Feb 22 00:35:13 xMff: when does it load stuff from /rom? Feb 22 00:35:34 rtz2: before it pivots Feb 22 00:35:48 thepeople: as we discussed earlier, it should be splitted from the base iface config imho, this would ease stuff a lot since ppp.sh & friends don't have to deal with the ipproto stuff Feb 22 00:35:52 cshore: well, it pivots pretty early Feb 22 00:36:07 cshore: before it loads the network stuff afaik Feb 22 00:36:13 rtz2: yes Feb 22 00:36:22 rtz2: as long as the overlay is not there, stuff is originating from rom Feb 22 00:36:37 and some platforms have hardcoded default network configs Feb 22 00:36:41 xMff: the overlay is added before init Feb 22 00:36:47 xMff: yep, nbd and blogic were talking about a rewrite of the scripts some time ago, I wonder how that is going Feb 22 00:36:58 cshore: k, then the race is somewhere else Feb 22 00:37:01 xMff: well, the overlay is there before init is started, or did I get this wrong? Feb 22 00:37:17 rtz2: no, you're right Feb 22 00:37:30 rtz2: still, lot of the network setup happens asyncronously, detached from the init scripts Feb 22 00:37:33 so, how can it load stuff from there? Feb 22 00:37:51 xMff: /etc/init.d/network runs later than the hotplug coldplug of interfaces though Feb 22 00:38:16 xMff: but doesn't it get kicked of from init scripts? Feb 22 00:38:19 I have seen cases where init is finished before the network is even brought up Feb 22 00:38:32 on fast systems Feb 22 00:38:49 which messes all sorts of things up Feb 22 00:38:54 right Feb 22 00:39:05 * thepeople has encountered that Feb 22 00:39:13 xMff: that's ok, but how can the network start before init? Feb 22 00:39:40 remindes me http://wiki.openwrt.org/inbox/techref/taskinit_discussion should probably be advertised for discussion somewhere Feb 22 00:39:53 iirc the overlay jffs2 stuff needs some time to settle, mount won't block Feb 22 00:40:20 or maybe not, however, there are races all over the places Feb 22 00:40:47 I wonder if it would be possible to do init in parrellel with dependancies Feb 22 00:41:15 thepeople: xMff and I talked about that a bit....that's what the link I posted is for discussing Feb 22 00:41:16 the danger in doing parallel init is to bomb out low memory devices on boot Feb 22 00:41:17 lars * r19810 /packages/net/quagga/patches/150-no-cross-fs-link.patch: [packages] quagga: Fix compile Feb 22 00:41:28 thepeople: i'm already working on something Feb 22 00:42:11 thepeople: not so much parallel init as asynchronous Feb 22 00:42:20 xMff: yep, although if it bombs on start there is a good likely hood of it bombing later Feb 22 00:42:38 I didn't have much time last week, but I will send something to the mailinglist this week Feb 22 00:42:49 cshore: true Feb 22 00:42:53 rtz2: I meant to ask about that Feb 22 00:43:26 it does make progress, but not as fast as planned :/ Feb 22 00:43:39 rtz2: isn't that always the way? Feb 22 00:43:48 right ... Feb 22 00:44:39 thepeople: the system will be able to track dependencies and check for certain conditions, so it should avoid the races Feb 22 00:45:08 rtz2: that would be great Feb 22 00:46:00 rtz: as long as you can add new tasks dynamically for hotplug events Feb 22 00:46:51 thepeople: regarding the network config rewrite, I think both blogic and nbd are busy with more essential stuff so maybe we should give them a head start by starting to work on it :) Feb 22 00:47:31 xMff: I agree, just want to make sure we don't duplicate the wheel :-) Feb 22 00:47:59 cshore: not in the first version, but I have already some ideas Feb 22 00:48:36 xMff: do you want to use lua for it? Feb 22 00:48:53 rtz2: base stuff in C and lua for scripting hooks Feb 22 00:49:43 it is trivial to embed within a C host application so it would be an ideal choice Feb 22 00:49:54 xMff: so it would become mandatory and I could also use it? Feb 22 00:49:54 I would prefer to see it in shell still to prevent yet another dependancy Feb 22 00:50:40 or c and shell Feb 22 00:50:44 thepeople: one could make it optional by just dlopen()'ing liblua at runtime if it exists Feb 22 00:50:48 http://openwrt.pastebin.ca/1805251 has my modified ppp.sh that deals with having pptp + pppoe at the same time Feb 22 00:51:06 xMff: but then what happens if it isn't? Feb 22 00:51:34 thepeople: users loose the ability to write lua scripts Feb 22 00:52:27 xMff: but what would likely happen is that essensial bootup scripts will get written in lua making it another manditory depenency Feb 22 00:53:09 thepeople: I actually don't think that's a bad thing....shell it rather limited Feb 22 00:53:43 thepeople: it's already used for ucitrigger which is enabled by default Feb 22 00:54:32 rtz2: I am aware, I don't think that is a good thing either, but you can remove ucitrigger with no negative consiquenses Feb 22 00:54:45 cshore: yes it is quite limited Feb 22 00:56:00 ash is quite a pain at times Feb 22 01:01:59 problem with shell code as scripting language for some hostapp is the limited interaction Feb 22 01:02:18 some signals, environment variables and exit codes Feb 22 01:02:42 maybe piping stdout/stdin Feb 22 01:05:47 xMff: for booting up the system what else are you looking for? Feb 22 01:05:52 I've put the script in #6720 Feb 22 01:05:53 anyway, the goal is to leave only the high level stuff to scripts, not stuff like ifconfig, route and whatever Feb 22 01:06:30 thepeople: something that could work closer to the kernel Feb 22 01:06:47 at least for the whole firewall and network business Feb 22 01:07:05 cshore: it is easier if you submit diffs to see what changed Feb 22 01:07:14 and to apply it Feb 22 01:07:34 it is far easier to iterate through interfaces or routing tables in C or some other embededded langauge then doing ash / awk / grep in complicated shell scripts Feb 22 01:08:01 and screen-scraping output from the cli tools Feb 22 01:11:25 I don't think that most of the networking scripts are overly complicated, I think most of what makes them hard to read is the modularity of them, function calls all over to different scripts Feb 22 01:14:08 the scripts itself are not that complicated but they're just not robust enough imho - keeping state vars in sync, keep firewall updated across ip changes, keep pppd running Feb 22 01:14:51 I think that is the fault of scripts, not the fact that they are in shell Feb 22 01:15:59 I think there are assumptsions in there that have been sticky issues Feb 22 01:16:01 there is some stuff that is very hard to do in shell Feb 22 01:17:58 yes there is, but I don't think that means we should discount it Feb 22 01:18:53 I did not say we should discout it Feb 22 01:19:37 thepeople: it's more a question of maintainability and readability Feb 22 01:19:44 but if a C program needs like 0.02 seconds to establish the same firewall ruleset the current ash scripts need like 5 seconds for then you know that there are limits Feb 22 01:20:38 thepeople: parsing out some sed/awk/grep magic is way more complicated then looking at c or lua code Feb 22 01:21:17 rtz2: I concur Feb 22 01:21:53 hey, for init.d scripts do you need a postinst to enable them? Feb 22 01:22:25 it was a design decision to leave init scripts of opkg installed packages disabled Feb 22 01:22:34 rtz2: is it worth the tradeoff in space for lua, also I don't think the complexity comes from sed/awk/grep most of time Feb 22 01:23:05 if they are built in they are enabled by default, installed after they are left disabled Feb 22 01:23:22 I do think that the enabled by default when built in should be changed tho Feb 22 01:23:23 ok, thanks: that's what I needed to know Feb 22 01:23:53 thepeople: maybe not Feb 22 01:23:59 I just wanted to make sure I didn't need to write a section since I'm adding a init.d file Feb 22 01:24:09 rtz2: I also have a biased viewpoint having worked alot with the scripts in openwrt as well as ash in general Feb 22 01:24:10 thepeople: I admit, I'm not sure about that Feb 22 01:24:49 thepeople: well, we can see how it goes Feb 22 01:25:54 xMff: there are a number of things that can be done faster in C, but that doesn't make it the best chose all the time Feb 22 01:26:16 thepeople: same goes for shell Feb 22 01:26:43 agreed Feb 22 01:27:40 shell scripts are nice for one-shot operations for example setup stuff at boot Feb 22 01:30:06 Ok, I've added diff/patch to #6720 Feb 22 01:30:19 cshore: thanks Feb 22 01:31:49 xMff: you know, leobaillard probably hates you after this comment ;) Feb 22 01:32:16 rtz2: who is leobaillard? Feb 22 01:32:28 cshore: he's in #openwrt Feb 22 01:33:08 cshore: he seems rather clueless about the config thing and xmff writes this: Feb 22 01:33:09 leobaillard: well that should be straightforward, you need to renumber the port in the vlan settings, get rid of the wan interface, define a new one for your management access and tweak the firewall config Feb 22 01:33:12 :D Feb 22 01:33:22 it does has it's irony Feb 22 02:19:29 hi Feb 22 02:19:49 where is defined which .h files are taken when uclibc is created? Feb 22 02:26:34 jow * r19811 /trunk/ (3 files in 3 dirs): [package] ppp: enable multilink support (#4837) **** ENDING LOGGING AT Mon Feb 22 02:59:58 2010