**** BEGIN LOGGING AT Sat Jun 30 02:59:57 2007 Jun 30 08:40:54 rwhitby * r7785 /trunk/target/linux/ixp4xx-2.6/config/default: Enable USB EHCI hub options to bring ixp4xx kernel usb support in line with known good nslu2-linux kernel usb hub settings. Jun 30 10:59:34 [mbm], can you read my answers in priv msg? Jun 30 11:00:07 The first WIP-Day is now about to begin - wellcome to all who are present and all who will show-up during the following 24 hours :0) (Now I just hope my UTC calibration is correct, with me being on UTC+1)) Jun 30 11:00:41 :) Jun 30 11:00:44 <[mbm]> forum2006: yeah, I see your msg Jun 30 11:01:03 huh, strange i'm not registred at freenode at all Jun 30 11:02:03 there's a feature: /quote ns set unfiltered on Jun 30 11:02:11 allows registered people to chat with unregistered people Jun 30 11:05:12 nbd * r7786 /trunk/package/ (5 files in 3 dirs): replace the hotplug2 fork handling hack with a better solution implemented in upstream (thanks, iSteve) Jun 30 11:05:24 <[mbm]> nbd: yeah, I'v got the filtering crap off.. annoys me Jun 30 11:05:30 me too Jun 30 11:05:42 i've never received any query spam Jun 30 11:05:56 one subject which could be good to discuss at some point in time would be: how to procede with an update/restructuring of the existing documentation? Jun 30 11:06:16 glp_home: are you talking about the wiki or the documentation in svn? Jun 30 11:06:30 I'm talking about the wiki Jun 30 11:07:10 ok, i think the primary source of documentation should be what we have in svn Jun 30 11:07:18 of course it's not nearly finished and still needs a lot of work Jun 30 11:07:45 but the wiki is too unstructured to be reliable. and people occasionally tend to put in misinformation as well Jun 30 11:08:31 ok - then I think it is needed to add this information to the wiki - just to help people to remember that the svn exists Jun 30 11:08:58 we should start autobuilding the docs in svn and put them on downloads.openwrt.org Jun 30 11:09:05 <[mbm]> need to figure out some sort of quality assurance for the documentation Jun 30 11:09:07 at the moment people are often pasting links to a build on my servef Jun 30 11:09:08 server Jun 30 11:09:22 yes Jun 30 11:09:31 <[mbm]> nbd: yeah, should get the buildbot to just toss up the latest docs Jun 30 11:10:25 but maybe it would be an idea to ask members if they would like to help with re-viewing the wiki Jun 30 11:11:47 a lot of people seem to find old or wrong information - but it doesn't really get picked-up anywhere + few then actually correct it by themselves :/ Jun 30 11:12:51 but the buildbot doing the svn documentation would be really nice :) Jun 30 11:14:43 <[mbm]> figure there will be two levels of documentation Jun 30 11:15:10 <[mbm]> the official manuals which will be short and will be considered the absolute truth Jun 30 11:15:15 <[mbm]> and the community stuff Jun 30 11:17:17 yes sounds good Jun 30 11:18:04 <[mbm]> as soon as the documentation becomes too big people avoid it and look elsewhere for simple solutions Jun 30 11:21:43 yeah Jun 30 11:29:52 I'll add a comment about the svn documentation at he top of the wiki Jun 30 11:30:24 <[mbm]> well, right now the svn documentation is pretty bare and aimed mostly at other developers Jun 30 11:30:39 <[mbm]> which is why it wasn't mentioned in the 7.06 release notes Jun 30 11:31:03 <[mbm]> probably need to start a documentation team Jun 30 11:31:39 i know - so i'll write soemthing like '...the most up to date development documentation is found in the svn ...' ? Jun 30 11:32:03 and a docu-team would be a good activity:) Jun 30 11:32:04 <[mbm]> wouldn't even botehr plugging it yet Jun 30 11:32:18 <[mbm]> wait until there's something useful there Jun 30 11:32:25 oki Jun 30 11:33:24 <[mbm]> probably need to post a help-wanted or something to find someone to head the documentation team Jun 30 11:34:06 <[mbm]> needs to be someone who's been around the community long enough Jun 30 11:36:14 <[mbm]> should put an infobot in #openwrt too; most questions can be answered by canned responses Jun 30 11:43:45 [mbm], svn password does not work Jun 30 11:45:04 trying with this svn command: Jun 30 11:45:06 svn --username=olli -q checkout https://svn.openwrt.org/openwrt/packages/ packages/ Jun 30 11:45:15 <[mbm]> forum2006: looks right on this end Jun 30 11:46:18 hmm, 403 forbidden Jun 30 11:47:48 svn: PROPFIND of '/openwrt/packages': 403 Forbidden (https://svn.openwrt.org) Jun 30 11:48:11 <[mbm]> forum2006: did you give me the wrong password hash? Jun 30 11:48:55 [mbm], i don't think so. i tried both passwords and trac password works fine Jun 30 12:01:15 olli * r7787 /packages/utils/mutt/Makefile: mutt: update to 1.5.16 (#1990) Jun 30 12:02:44 :) Jun 30 12:05:54 :) Jun 30 12:09:19 nbd * r7788 /trunk/package/kernel/modules/usb.mk: fix the path to the yealink module (#1944) Jun 30 12:09:57 so, the packages must compile at least on two platforms, right? Jun 30 12:11:43 nbd * r7789 /trunk/package/kernel/modules/usb.mk: cleanup Jun 30 12:21:06 nbd * r7790 /trunk/tools/Makefile: add update target to tools/ Jun 30 12:21:55 nbd * r7791 /trunk/ (3 files in 3 dirs): add portability fixes from #1720 Jun 30 12:23:24 [mbm]: such an help-wanted add could be placed in the forum --> show an interest by adding a post in the thread and the developer group will get back? Jun 30 12:24:39 <[mbm]> falling asleep here; I'll check back in several hours Jun 30 12:31:07 nbd * r7792 /trunk/target/linux/generic-2.4/base-files/: nuke old base-files directory Jun 30 12:32:43 [mbm]: have a good sleep :) Jun 30 12:41:32 olli * r7793 /packages/net/maradns/ (Makefile patches/001-cross_compile_fix.patch): Update maradns: 1.2.12.04 > 1.2.12.06 (#1976) Jun 30 12:55:42 could someone please confirm this https://dev.openwrt.org/ticket/1973 ? Jun 30 13:08:58 nbd, trying to update elinks but i got: http://pastebin.ca/597168 , should i leave the package or is there an easy fix? Jun 30 13:17:00 something is wrong with r7793: http://pastebin.ca/597188 Jun 30 13:17:45 this is a fresh checkout and I'm trying to build brcm47xx with gcc-4.2.0 and binutils-2.17 Jun 30 13:21:19 fixing that... Jun 30 13:22:19 puh, lucky it was not my fault :) Jun 30 13:23:19 nbd * r7794 /trunk/tools/ipkg-utils/patches/140-portability.patch: fix accidentally committed broken patch Jun 30 13:23:24 loswillios: fixed Jun 30 13:23:31 thanks Jun 30 13:26:29 nbd, is it difficult to fix this elinks thingy Jun 30 13:26:30 ? Jun 30 13:28:21 no idea Jun 30 13:28:28 busy with other things right now Jun 30 13:33:21 nbd * r7795 /trunk/package/wlcompat/ (17 files in 4 dirs): update wlcompat for new ioctl interface (fixes #1357) Jun 30 13:48:51 does the toolchain not support QUILT or is - make toolchain/gcc-{clean,prepare} QUILT=1 V=99 (something wrong with the syntax) Jun 30 13:49:40 it supports that Jun 30 13:50:04 do you get an error? Jun 30 13:50:35 I get it not even doing gcc Jun 30 13:50:51 are you using the latest rev? Jun 30 13:50:54 trying to do uClibc Jun 30 13:50:59 ah Jun 30 13:51:04 yeah, there's an issue with dependency checking Jun 30 13:52:03 as a workaround, you can do make -C toolchain/gcc clean prepare QUILT=1 TOPDIR=$PWD V=99 Jun 30 13:52:27 thanks Jun 30 14:06:04 olli * r7796 /packages/net/miau/Makefile: Update miau: 0.6.3 > 0.6.4 Jun 30 14:20:23 nbd: Do you think it would be much effort for you to allow choice of uclibc (like now it is with kernel) ? (possibly I could update the patches but I have no chance to add the hooks to that) - also someway to allow - me to set - http://paste.debian.net/31704 Jun 30 14:31:40 h3sp4wn: after 7.07, i will create a compatibility library to make the old binaries on broadcom work with the new uclibc installed Jun 30 14:31:48 h3sp4wn: then we can upgrade for all targets Jun 30 14:55:03 nbd: Is that to be released tommorow ? Jun 30 14:56:51 is there a way set a var which is also defined in include/site/mipsel-linux-uclibc ? just want to override one for one package. if i put ac_cv_sizeof_off_t=4 in the configurepart of the Makefile it doesn't work :| Jun 30 14:57:27 WIP-Day --> i jsut submitted a basic kamikaze installation/configuration howto Jun 30 14:57:27 jsut = just Jun 30 14:57:58 but I was wondering about the default generated network file Jun 30 14:59:40 h3sp4wn: not tomorrow, but soon Jun 30 15:00:03 glp_home: what about it? Jun 30 15:00:09 using the image i downloaded this morning generates a network file missing the two options 'gateway' and 'dns' - which you need to add manually to make the LAN work properly Jun 30 15:01:32 nbd: ok, will it still uses 2.6.19.2 (or 2.6.19.7 ?) - Is there any reason the microcode loading patch cannot be added (or do I still need to ask Kaloz - he said to me the next release would have the new net driver) Jun 30 15:02:02 i've seen several people having problems with getting kamikaze to work - and i was just wondering if why those options shouldn't be generated by default? Jun 30 15:02:35 h3sp4wn: it will probably still use the old driver Jun 30 15:03:55 juhosg * r7797 /trunk/target/linux/ (27 files in 3 dirs): [adm5120] move 2.6.22 specific patches into the right place Jun 30 15:05:52 glp_home: they don't have to be generated by default if they're empty Jun 30 15:06:35 juhosg * r7798 /trunk/target/linux/adm5120-2.6/patches/: [adm5120] restore old patches for 2.6.21.5 kernel Jun 30 15:21:32 juhosg * r7799 /trunk/target/linux/adm5120-2.6/patches/: Oops! I've copied r7557 instead of r7757, sorry Jun 30 15:23:03 juhosg * r7800 /trunk/target/linux/adm5120-2.6/patches/: [adm5120] restore right patches for 2.6.21.5 Jun 30 15:31:22 olli * r7801 /packages/net/weechat/ (Makefile patches/001-cross_compiling.patch): Update weechat: 0.2.4 > 0.2.5 Jun 30 16:11:56 some packages i don't get... Jun 30 16:12:35 I submitted the 2000st ticket :P Jun 30 16:14:08 thepeople, another one who can't compile miniupnpd Jun 30 16:14:27 <_trine> guess its you who now has to show his bum Jun 30 16:15:15 loswillios, so, miniupnpd rc6 compiles again? Jun 30 16:15:32 nope Jun 30 16:16:25 maybe it's because of gcc-4.2.0 Jun 30 16:19:01 miniupnpd rc6 with gcc 4.2.1 compiles fine Jun 30 16:19:23 4.1.2 Jun 30 16:19:34 ? Jun 30 16:19:53 yeah, 4.1.2 Jun 30 16:20:00 I'll write an email to the author then Jun 30 16:21:28 let's test it with .22-rc6 and gcc 4.2.0 Jun 30 16:22:11 .22-rc6 and gcc 4.2.0 it's still broken Jun 30 16:22:37 I forgot to mention it's .22-rc6 Jun 30 16:26:47 loswillios, i'll update it to miniupnpd rc6 even if it doesn't fix the .22 issue... Jun 30 16:35:36 forum2006: yes, good idea Jun 30 16:40:32 olli * r7802 /packages/net/miniupnpd/ (3 files in 2 dirs): Update miniupnpd: 1.0-RC5 > 1.0-RC6 Jun 30 16:42:21 loswillios, done Jun 30 16:44:42 To add the a version of hostapd with the devicescape driver - best to add a new template or just add it to the existing one Jun 30 16:59:51 forum2006: I see you have commit again Jun 30 17:00:21 thepeople, yep. [mbm] was friendly. Jun 30 17:00:53 thepeople, but only to /packages. which is enough for now. Jun 30 17:09:35 :-) Jun 30 17:09:45 thepeople: maybe it's 2.6.22-rc6 related b/c changelog of miniupnpd-RC6 says: "Redirection descriptions are now stored in the linux/netfilter flavour of the daemon." Jun 30 17:10:10 and the build fails because it can't find linux/netfilter/bla.h Jun 30 17:35:37 nbd, i have packaged cfdisk (from util-linux). the problem is that it depends on libncurses but ncurses is not in /trunk/package Jun 30 17:50:57 if i add a package does svn automatically set the propset? Jun 30 17:53:41 let's see what happens Jun 30 17:59:35 olli * r7803 /packages/libs/libsndfile/ (. Makefile): Add libsndfile (#1959) Jun 30 18:27:18 nbd * r7804 /trunk/include/site/ (7 files): fix the value of ac_cv_sizeof_off_t in the sitefiles (reported by crazy_imp) Jun 30 19:28:45 nbd * r7805 /trunk/package/dnsmasq/files/dnsmasq.init: include the network state in the dnsmasq init script. this makes it possible to run a dhcp server on dynamically assigned interfaces without hardcoding interface names Jun 30 19:30:35 nbd * r7806 /trunk/package/iptables/files/firewall.init: make the firewall script run after the network script again (required for working with dynamically assigned interfaces), include the network state Jun 30 19:47:21 nbd * r7807 /trunk/package/madwifi/patches/124-node_unref_r2484.patch: add madwifi changeset 2484 (node refcounting patch) - i will resync with madwifi svn after the 7.07 release Jun 30 19:50:42 nbd * r7808 /branches/whiterussian/openwrt/package/qos-scripts/files/usr/lib/qos/generate.sh: add missing function (https://dev.openwrt.org/ticket/1745#comment:3) Jun 30 20:00:26 nbd * r7809 /trunk/package/base-files/files/etc/init.d/boot: fix a warning (#1800) Jun 30 21:09:00 nbd * r7810 /trunk/ (4 files in 2 dirs): move parts of the failsafe code to the generic base-files - will be used on atheros as well as soon as i've added support for it Jun 30 21:09:16 messing around with updating hostapd to the latest version (needed for mac80211) - cannot understand why - http://paste.debian.net/31734 doesn't work - also how is the actual mac80211 package being used (I see nothing in the bcm43xx-mac80211 Makefile that makes it use that not the in kernel version) Jun 30 21:10:28 h3sp4wn: we will use the mac80211 package outside of the kernel tree so people don't have to rebuild the kernel tree whenever we add mac80211 updates from wireless-dev Jun 30 21:11:24 nbd: I know that is the theory - in that case should the in tree mac80211 be disabled ? Jun 30 21:11:34 yeah Jun 30 21:11:41 h3sp4wn: Yes Jun 30 21:12:16 Sorry if there are slight crapnesses in the packaging of those two Jun 30 21:12:31 I started by building in kernel, then moved them out to packages. Jun 30 21:13:35 h3sp4wn: The way it uses the out-of-tree version is by installing the ipkg, which will overwrite the module in /lib/modules Jun 30 21:13:59 noz: but the headers are different in wireless-git to 2.6.22 Jun 30 21:14:17 (hence with rt2x00 I cannot get parts of it to be build) Jun 30 21:14:21 You are right, we should definitely disable the in-kernel version, but we use the same config name, so it gets built Jun 30 21:14:44 I've got a patch, in generic-2.6, that should fix the headers. Jun 30 21:15:31 could be something stupid I did though Jun 30 21:16:48 You should have a target/linux/generic-2.6/patches-2.6.22/210-mac80211_include_wireless_dev.patch Jun 30 21:21:02 I have that - I will take a closer look Jun 30 21:42:54 nbd * r7811 /trunk/include/kernel.mk: fix kernel config feature autodetection Jun 30 21:49:18 ndb: i've just tried you patch (Ticket #1973). sry, doesn't work Jun 30 21:49:30 same kernel panic Jun 30 21:49:40 nbd * r7812 /trunk/include/kernel-build.mk: merge the kernel config parts at dump time, so that overrides that disable certain features work Jun 30 21:50:33 i can't understand why but when i try to associate with osx with a wrong wep key, that wds code is executed Jun 30 21:50:51 ok, i'll check the code again Jun 30 21:51:05 can i do something? Jun 30 21:51:15 (usefull) :) Jun 30 21:51:41 yes, test a new patch when i got one :) Jun 30 21:51:46 sure Jun 30 21:51:58 can u reproduce the kernel panic? Jun 30 21:52:28 only Kevin` confirmed to me that behaviour Jun 30 21:52:36 but only in irc Jun 30 21:58:29 i haven't tested yet Jun 30 22:19:12 Cdtdaddy: i'll use your fix for now Jun 30 22:19:32 Cdtdaddy: i really don't get why the node table is NULL, but the additional check probably won't hurt Jun 30 22:19:49 Cdtdaddy: please test the new http://nbd.name/309-micfail_detect.patch Jun 30 22:20:33 ok Jun 30 22:22:41 compiling... Jun 30 22:24:09 it is possible that ieee80211_node_leave is called 2 times? Jun 30 22:24:24 i have no deep understand of the code.. Jun 30 22:24:40 it's possible, but that would be a bug in the refcounting system Jun 30 22:29:10 ndb: ok the patch works Jun 30 22:29:18 ok. i'll commit it Jun 30 22:29:25 thank u Jun 30 22:30:35 nbd * r7813 /trunk/package/madwifi/patches/309-micfail_detect.patch: fix a crash related to wep and mic handling (#1973) Jun 30 23:00:00 WIP-Day is now about to pass into its second half :) Jun 30 23:02:08 glp_home: huh? is it 2 days long? :) Jun 30 23:04:20 crazy_imp: never been good at neither math nor figuring out what date it is - so i'm just kind'o stumbling along :) Jun 30 23:04:31 :-) Jun 30 23:04:34 hehe :D Jun 30 23:04:47 glp_home: so it is on british summer time ? Jun 30 23:08:41 h3sp4wn: summertime is a mess, already the concept makes you think that the weather is nice and sunny, that you will spend a lot of time on a beach with one of those drinks with an umbrella in one hand and your other hand wiping the sweat off your brow. Jun 30 23:09:13 and the reality is rain, rain, rain, rain Jun 30 23:10:19 noz * r7814 /trunk/target/linux/brcm47xx-2.6/files/include/asm-mips/mach-bcm947xx/gpio.h: Bring mach-bcm947xx/gpio.h into line with new SSB code, as per #1640 Jul 01 00:04:18 nbd: does the stuff you've added from #1984 also work with gcc 3.4.6 and binutils 2.16 ? Jul 01 00:14:04 crazy_imp: should. why are you asking? Jul 01 00:16:02 because i had the same problem and it vanished after i changed gcc&binutils in the toolchain Jul 01 00:17:41 nbd: (and i tried r7765, and you added that stuff in r7756) Jul 01 00:19:19 nbd: same problem like #1613 **** ENDING LOGGING AT Sun Jul 01 02:59:56 2007