**** BEGIN LOGGING AT Wed Jan 07 02:59:57 2009
Jan 07 03:22:12 <RoundSparrow> Does setting CONFIG_ATH9K_DEBUG=y in the trunk/.config file work?  Is there a proper place to set this that does work?
Jan 07 04:25:26 <RoundSparrow> I see the ath9k AP mode 5 weeks ago had support for N mode added, but how do you pick it? https://dev.openwrt.org/changeset/13497/trunk/package
Jan 07 04:26:08 <CIA-30> nbd * r13911 /trunk/package/ (22 files in 3 dirs): nuke the old ath9k package, rename kmod-ath9k-new in compat-wireless to kmod-ath9k
Jan 07 04:26:17 <Bartman007> hehe.
Jan 07 04:26:19 <CIA-30> nbd * r13912 /trunk/ (5 files in 5 dirs): make broadcom-wl robust against implicit kernel config changes (e.g. netfilter), upgrade it from minor version .2 to .3 (adds fixed bssid support for ad-hoc)
Jan 07 04:26:24 <Bartman007> RoundSparrow: he told you.
Jan 07 04:27:09 <nbd> xMff: ^^ please make sure that luci will be able to set the config option for r13912 :)
Jan 07 04:27:14 <nbd> i will merge the change to 8.09 as well
Jan 07 04:27:28 <nbd> this should make some freifunk people happy
Jan 07 04:27:30 <RoundSparrow> Bartman007: I don't follow?
Jan 07 04:28:10 <Bartman007> RoundSparrow: looks like that patch just nuked ath9k and replaced it with ath9k-new, unless you were already speaking about ath9k-new
Jan 07 04:28:17 <Bartman007> s/patch/commit/
Jan 07 04:28:56 <nbd> i nuked it because there's no point in testing an unmaintained old version
Jan 07 04:29:15 <CIA-30> nbd * r13913 /branches/8.09/ (5 files in 5 dirs): merge r13912 to 8.09
Jan 07 04:29:33 <Bartman007> RoundSparrow: to select 11g vs 11n, use the hwmode option
Jan 07 04:31:05 <RoundSparrow> the docs don't cover anything newer than g https://svn.openwrt.org/openwrt/trunk/docs/wireless.tex
Jan 07 04:31:35 <RoundSparrow> like is "11gn" or "11bgn" valid values?
Jan 07 04:31:53 <Bartman007> there are several other madwifi specific options it does not cover
Jan 07 04:33:10 <Bartman007> RoundSparrow: btw. I mapped out the wnr2000 gpios, I'll add the code later tonight, but being able to run a writable filesystem from flash will require recompiling uboot because of the checksum madness that netgear does.
Jan 07 04:34:19 <Bartman007> I don't understand *why* they statically compile the checksum requirement into the bootloader when they could make it a configurable option, aside from fearing modification
Jan 07 04:37:32 <RoundSparrow> Bartman007: cool, good to hear.  Yha, that's no fun
Jan 07 04:38:59 <CIA-30> nbd * r13914 /trunk/package/base-files/files/etc/profile: fix $HOME in /etc/profile (#4424)
Jan 07 04:39:01 <Bartman007> especially since I don't seen any obvious jtag port. I see some points that may be jtag, but it's just a guess, and won't be fun to deal with
Jan 07 04:39:44 <CIA-30> nbd * r13915 /branches/8.09/package/base-files/files/etc/profile: merge r13914 to 8.09
Jan 07 04:40:34 <RoundSparrow> Bartman007:  It looks like iw is the program to control the rate params?  http://wireless.kernel.org/en/users/Documentation/iw
Jan 07 04:46:55 <Bartman007> RoundSparrow: possibly, I haven't really messed with mac80211 much
Jan 07 04:47:54 <RoundSparrow> ath9k uses it, doesn't it?
Jan 07 04:49:39 <Bartman007> I was naively hoping that hostapd would handle it ;)
Jan 07 04:51:21 <Bartman007> I planned to investigate once I got devices working
Jan 07 04:51:47 <RoundSparrow> hostapd says: unknown hw_mode 'n'
Jan 07 04:51:59 <Bartman007> RoundSparrow: 11n is proper syntax IIRC.
Jan 07 04:52:04 <RoundSparrow> (hand edited config file changed g to n)
Jan 07 04:52:19 <RoundSparrow> in that file it was just hw_mode=g
Jan 07 04:52:47 <RoundSparrow> I killed the hostapd running and start interactive with "hostapd -dd  /tmp/run/hostapd-wlan0.conf"
Jan 07 04:52:49 <Bartman007> oh, oops. it was late when I tweaked that script, I forgot
Jan 07 04:53:57 <RoundSparrow> "Line 3: unknown hw_mode '11n'"
Jan 07 04:54:56 <Bartman007> yeah, g is the correct syntax. I'm looking at a config and I guess 11g and 11a determine the frequencies it uses?
Jan 07 04:55:18 <Bartman007> I see a seperate section of 11n configuration down below... http://hostap.epitest.fi/gitweb/gitweb.cgi?p=hostap.git;a=blob_plain;f=hostapd/hostapd.conf
Jan 07 04:55:26 <Bartman007> search for ieee80211n
Jan 07 04:56:57 <RoundSparrow> fyi, openwrt has hostapd v0.6.6 and looks like 0.6.7 was released
Jan 07 04:58:35 <RoundSparrow> hum, Line 4: unknown configuration item 'ieee80211n'
Jan 07 04:59:51 <RoundSparrow> changelog says we need 0.6.7 for these params? http://hostap.epitest.fi/gitweb/gitweb.cgi?p=hostap.git;a=blob_plain;f=hostapd/ChangeLog
Jan 07 05:00:41 <RoundSparrow> 0.6.7 released just today.  "* added IEEE 802.11n HT capability configuration (ht_capab) "
Jan 07 05:01:07 <Bartman007> let me make a quick patch
Jan 07 05:03:20 <RoundSparrow> This is useful reference, bottom of page: http://wireless.erley.org/git.html
Jan 07 05:05:49 <RoundSparrow> Bartman007:  iw is a package, but opkg gave me 0.9.5 - looks lik someone did a patch to update to 0.9.7, but maybe not added to trunk? https://dev.openwrt.org/ticket/4373
Jan 07 05:07:02 <RoundSparrow> "iw list" doesn't show anything higher than 54, so I'm guessing something in mac80211 or ath9k driver is still an issue.  But I really don't understand this stuff ;)
Jan 07 05:09:09 <Bartman007> I don't claim to either, but I figured I'll have to learn if I decide to add 11n support to the mac80211 UCI configs.
Jan 07 05:24:16 <kupesoft> anyone wanna add my patch to /etc/profile? ;p
Jan 07 05:24:35 <kupesoft> I noticed it was touched today
Jan 07 05:26:16 <Bartman007> kupesoft: which what where?
Jan 07 05:26:33 <kupesoft> sec, let me dig it up
Jan 07 05:26:47 <kupesoft> I mailed it out in November, it adds xterm support
Jan 07 05:26:50 <kupesoft> debian-style
Jan 07 05:27:01 <kupesoft> guess it just slipped under the radar
Jan 07 05:27:04 <kupesoft> xterm fancy title bar
Jan 07 05:27:05 <kupesoft> http://lists.openwrt.org/pipermail/openwrt-devel/2008-November/003406.html
Jan 07 05:28:02 <kupesoft> silly, but its useful locally
Jan 07 05:28:06 <kupesoft> (it's
Jan 07 05:31:20 <Bartman007> but it doesn't work for me! xterm-color :P
Jan 07 05:32:03 <kupesoft> Look at the top of the title bar
Jan 07 05:32:25 <Bartman007> xterm != xterm-color :
Jan 07 05:32:29 <Bartman007> :P
Jan 07 05:32:32 <kupesoft> ah
Jan 07 05:32:33 <kupesoft> i'll fix that
Jan 07 05:32:41 <kupesoft> and it'll bump the email ;p
Jan 07 05:32:47 <Bartman007> it was a joke
Jan 07 05:33:22 <Bartman007> I can't remember, what is it that makes PS1 appear in the titlebar?
Jan 07 05:34:56 <RoundSparrow> Bartman007:  hostapd has a CONFIG_IEEE80211N var.  I get lost in Makefile land, but not sure if openwrt is setting that
Jan 07 05:36:00 <kupesoft> Bartman007: \[\e]0; and \a\]
Jan 07 05:41:13 <frogonwheels> anybody know what would prevent a package with an empty /install from being built?
Jan 07 05:42:13 <Bartman007> mmmm, stuffed bellpepper.
Jan 07 05:49:13 <kupesoft> Bartman007: [ $(expr index "$TERM" xterm) -gt 0 ] && export PS1="\[\e]0;\u@\
Jan 07 05:49:13 <kupesoft> h: \w\a\]$PS1"
Jan 07 05:49:35 <kupesoft> Is that the best way to search for a substring with shell scripting, I haven't done shell scripting since first or second year uni ;p
Jan 07 05:52:59 <kupesoft> echo "$TERM" | grep -q xterm && ...
Jan 07 05:53:02 <kupesoft> so many options
Jan 07 07:38:48 <Bartman007> RoundSparrow: only compile tested, no uci config updates yet. I have to run, have fun testing.
Jan 07 07:38:57 <Bartman007> http://agb.openwrt.org/ondeck/hostapd-0.6.7.patch
Jan 07 09:41:34 <CIA-30> juhosg * r13916 /trunk/package/ath9k/: [package] ath9k: remove empty dirs
Jan 07 09:42:30 <CIA-30> juhosg * r13917 /trunk/target/linux/brcm47xx/ (files-2.6.23/ patches-2.6.23/): [brcm47xx] remove empty dirs
Jan 07 09:45:10 <CIA-30> juhosg * r13918 /trunk/target/linux/generic-2.6/ (5 files in 5 dirs): [kernel] generic-2.6: remove empty dirs
Jan 07 09:47:45 <CIA-30> juhosg * r13919 /trunk/target/linux/generic-2.6/ (4 files in 4 dirs): [kernel] generic-2.6: remove more empty dirs
Jan 07 11:07:54 <CIA-30> agb * r13920 /trunk/package/kernel/modules/other.mk: [package] kernel: added GPIO driver for CS5535/CS5536 companion chip. Thanks Markus Wigge
Jan 07 11:11:15 <CIA-30> mirko * r13921 /packages/lang/python/files/python-package.mk: prevent packaging precompiled python files (*.pyc, *.pyo)
Jan 07 13:41:12 <CIA-30> mirko * r13922 /packages/lang/pyserial/ (. Makefile): added python package <pyserial>
Jan 07 14:04:04 <CIA-30> mirko * r13923 /packages/lang/pyyaml/ (. Makefile): added python package <PyYaml>
Jan 07 15:03:13 <CIA-30> mirko * r13924 /packages/phone/paroli/Makefile: changed git repository
Jan 07 17:48:16 <MateIn4> I finally created a usr8200 patch to add LED support and fix the PHYs. Sent it to mailing list.
Jan 07 17:53:39 <Bartman007> MateIn4: hmm, I don't see it yet
Jan 07 17:55:41 <MateIn4> I guess it takes a while...
Jan 07 17:58:39 <larsc> MateIn4: theres no file attached to your mail
Jan 07 18:00:30 <MateIn4> ok, that was stupid, retrying...
Jan 07 18:17:39 <CIA-30> juhosg * r13925 /trunk/package/mac80211/ (23 files in 2 dirs): [package] mac80211: update compat-wireless to 2009-01-07, and refresh ath9k patches
Jan 07 18:21:56 <CIA-30> juhosg * r13926 /trunk/package/mac80211/patches/ (3 files): [package] mac80211: oops, add missing patches
Jan 07 18:29:18 <CIA-30> lars * r13927 /packages/ (4 files in 4 dirs): Force symlink creation.
Jan 07 18:42:38 <roundsparrow> juhosg:  I want to make sure you caught the discussion last night about  iw, hostapd, and hostapd being compiled with N support. http://forum.openwrt.org/viewtopic.php?pid=79159#p79159
Jan 07 18:44:53 <Bartman007> roundsparrow: did you see my patch for hostapd?
Jan 07 18:45:12 <roundsparrow> yep, but it didn't go into Trunk, right?
Jan 07 18:45:32 <Bartman007> right, I haven't tested it and it doesn't have any UCI 11n support yet
Jan 07 18:47:40 <roundsparrow> I put a link to your patch in that thread
Jan 07 18:48:55 <Bartman007> roundsparrow: thanks
Jan 07 18:49:06 <roundsparrow> juhosg said he is atively working on performance tuning, the high cpu in softirq is probably more critical than the N mode at the moment ;)  The router is CPU bound.
Jan 07 18:50:10 <MateIn4> Does anyone know how to add a new LED trigger to LuCI?
Jan 07 18:54:20 <xMff> in System -> LED Config ?
Jan 07 18:56:11 <MateIn4> xMff: yes, but I want to add a new trigger condition (USB insertion/removal)
Jan 07 18:59:11 <xMff> MateIn4: take a look at /usr/lib/lua/luci/model/cbi/admin_system/leds.lua
Jan 07 19:03:53 <MateIn4> xMff: I've been looking at that in the source tree but it's not clear how to add new event types. These would be hotplug events.
Jan 07 19:04:48 <xMff> MateIn4: CyrusFF wrote that led trigger part but as far as I know it relies on uci in /etc/config/system which implements the actual handling
Jan 07 19:05:32 <MateIn4> I'll take a look there
Jan 07 20:08:28 <sw> ejka: are you still working on cpmac.c - it needs some fixes for kernel 2.6.28?
Jan 07 20:18:28 <johnrw> xMff: ping
Jan 07 21:09:48 <dbro> I'm trying to compile and install some example programs that come with libsamplerate, which has a Makefile in feeds/packages/libs/libsamplerate. What command should I use? I've tried "make package/feeds/packages/libs/libsamplerate-clean V=99" and I get this: "No rule to make target ..."
Jan 07 21:11:11 <xMff> look at feeds/packages/libs/libsamplerate/Makefile how the actual packages are called (define Package/foobar)
Jan 07 21:11:16 <xMff> johnrw: pong
Jan 07 21:11:39 <dbro> it says this: "define Package/libsamplerate"
Jan 07 21:11:58 <Bartman007> dbro: works here
Jan 07 21:12:42 <dbro> hmm... any suggestions then?
Jan 07 21:13:53 <dbro> it shows up in the menus when I run "make menuconfig"
Jan 07 21:15:32 <Bartman007> verify that the path was typed correctly? (no I don't have any suggestions)
Jan 07 21:16:22 <dbro> hey this seems to work: " make package/feeds/packages/libsamplerate-clean V=99" -- I removed the substring "/libs"
Jan 07 21:16:36 <dbro> chalk it up to the bozo syndrome
Jan 07 21:17:07 <Bartman007> oh yes, the feeds script doesn't install the hierarchy
Jan 07 21:17:20 <dbro> ok - that's tripped me up before.
Jan 07 21:22:27 <dbro> btw, the Makefile is set up for version 0.1.3 of libsamplerate - do you want me to do a quick diff to update the version number and md5sum for version 0.1.4 ?
Jan 07 21:35:38 <frogonwheels> Is there an example in a makefile somewhere of a virtual package - one that is there just for it's dependencies?
Jan 07 21:36:24 <frogonwheels> oh xMff - did you understand my comments about the asterisk dialplan config?
Jan 07 21:36:45 <xMff> frogonwheels: I think yes
Jan 07 21:36:59 <xMff> currently playing with the openwrt package
Jan 07 21:38:38 <frogonwheels> xMff: I should send you an update - playing with luci interface on the virtualbox revealed some bugs
Jan 07 21:39:12 <xMff> so it actually does something for you? I never tested it
Jan 07 21:39:31 <frogonwheels> xMff: yeah - many bits work fine.
Jan 07 21:39:39 <frogonwheels> xMff: there are some validation issues as mentioned..
Jan 07 21:39:51 <xMff> cool
Jan 07 21:39:57 <frogonwheels> xMff: (like you can't actually select multiple 'includes' in the dialplan)
Jan 07 21:40:21 <frogonwheels> xMff: and some of those sections that need to be turned inside out don't work particularly well..
Jan 07 21:40:37 <xMff> okay, so I rewrite that stuff :)
Jan 07 21:40:55 <frogonwheels> xMff: but I'm guessing they were pretty confusing :)
Jan 07 21:41:19 <frogonwheels> xMff: I've also put lots more 'list' support in from the original one you had.
Jan 07 21:49:35 <frogonwheels> Bartman007: about the virtual package thing - is there an example?   I tried just having empty /install sections - but they don't seem to be being built :(
Jan 07 21:50:23 <xMff> frogonwheels: try to install some directory... like /tmp
Jan 07 21:50:30 <xMff> frogonwheels: maybe that's enough
Jan 07 21:50:36 <Bartman007> frogonwheels: what about python? I think the virt package python-core can be satisfied via python or python-mini, something like that.
Jan 07 21:52:41 <frogonwheels> ok thanks.
Jan 07 21:54:02 <frogonwheels> that's a PROVIDES which is not quite what I want.
Jan 07 22:00:47 <Bartman007> ah.
Jan 07 22:37:51 <CIA-30> lars * r13928 /packages/lang/pygobject/ (Makefile patches/010-binary-constants-module.patch): Update pygobject 2.16.0
Jan 08 00:05:11 <dbro> hello, I'm getting an error when compiling a package - it can't find libsndfile, but it seems to me like the right files are there in ./staging_dir/mipsel/usr/lib/ , and I'm setting the LDFLAGS in the Makefile - is there something else I can try?
Jan 08 00:05:55 <xMff> maybe the application makefile overrides your ldflags again
Jan 08 00:06:06 <xMff> with some static definition
Jan 08 00:06:11 <dbro> could be... how can I check?
Jan 08 00:06:30 <xMff> read the Makefile and search for stuff like LDFLAGS=...
Jan 08 00:06:48 <dbro> it's a very complicated Makefile- one of those automatically generated ones
Jan 08 00:07:56 <xMff> dbro: ah, then maybe look for a Makefile.in
Jan 08 00:08:11 <xMff> dbro: is it generated by a configure script?
Jan 08 00:08:18 <dbro> yes
Jan 08 00:08:34 <dbro> I cant see any mention of LDFLAGS in those files. CFLAGS are there, and look good
Jan 08 00:08:36 <xMff> dbro: have you set PKG_CONFIGURE_ARGS in the openwrt makefile?
Jan 08 00:08:54 <dbro> no - what's the syntax for that?
Jan 08 00:09:06 <xMff> dbro: maybe you need to pass sth. like "--with-libsndfile=/foo/bar" to configure
Jan 08 00:09:39 <dbro> ok, I'll try to find something like that
Jan 08 00:10:45 <xMff> search for TARGET_CONFIGURE_ARGS and PKG_CONFIGURE_OPTS
Jan 08 00:10:51 <xMff> plenty of packages use them
Jan 08 00:10:54 <dbro> ok thanks
Jan 08 01:03:58 <CIA-30> nico * r13929 /trunk/package/ (3 files in 3 dirs): fix kernel modules compile (as in [13724])
Jan 08 01:13:35 <dbro> one hour later, and I havent been able to make any progress with getting the compile process to find libsndfile. I didn't learn much from the TARGET_CONFIGURE_ARGS stuff- can you help me with the syntax?
Jan 08 01:13:54 <dbro> there seem to be OPTS, ARGS, and VARS
Jan 08 01:14:09 <dbro> for both TARGET and PKG
Jan 08 01:15:12 <dbro> TARGET strikes me as the right one to work with
Jan 08 01:21:24 <CIA-30> nico * r13930 /trunk/package/mac80211/Makefile: let all mac80211 subpackages inherit some defaults
Jan 08 01:49:15 <CIA-30> nico * r13931 /trunk/ (94 files in 23 dirs):
Jan 08 01:49:15 <CIA-30> [massive] add support for alternative C libraries (currently only glibc/eglibc)
Jan 08 01:49:15 <CIA-30> other (related) changes:
Jan 08 01:49:15 <CIA-30> - kernel headers are now installed using "make headers_install" on 2.6
Jan 08 01:49:15 <CIA-30> - target names now contain an openwrt "vendor" tag (e.g. mips-openwrt-linux-gnu)
Jan 08 01:49:16 <CIA-30> - build directory names now contain gcc/libc name/version
Jan 08 01:49:20 <CIA-30> - default cpu for x86 is now i486 (required to build glibc/eglibc)
Jan 08 02:07:41 * Bartman007 shakes his fist at {Nico} and kicks off test builds
Jan 08 02:07:50 <Bartman007> :)
Jan 08 02:11:59 <{Nico}> yeah!
Jan 08 02:12:47 <{Nico}> so now, we have 3 x more builds to tests :)
Jan 08 02:18:46 <CIA-30> nico * r13932 /trunk/package/iptables/Makefile: temp fix for iptables build failure after [13931]
Jan 08 02:23:17 <roundsparrow> doh, start over ;)
Jan 08 02:24:54 <roundsparrow> Stupid question who doesn't get these 6 levels of nested makefiles.  If I want to set CONFIG_ATH9K_DEBUG=y for a package, where?  I've tried the root .config and it didn't seem to work
Jan 08 02:28:21 <{Nico}> roundsparrow: you can add it to the MAKE_OPTS in package/mac80211/Makefile
**** ENDING LOGGING AT Thu Jan 08 02:59:57 2009