**** BEGIN LOGGING AT Thu Aug 30 02:59:57 2007
Aug 30 08:01:46 <CIA-4> nico * r8541 /trunk/package/ifenslave/Makefile: use $(INSTALL_BIN) instead of $(CP)
Aug 30 10:35:25 <LoneStar> is this document http://forum.openwrt.org/viewtopic.php?id=8410 enough for me to spread a workspace for development?
Aug 30 10:36:11 <[florian]> I think so
Aug 30 10:40:05 <LoneStar> ok. and, what is whiterussian?
Aug 30 10:40:22 <[florian]> it is the previous firmware release name
Aug 30 10:40:40 <LoneStar> previous? then what is the current?
Aug 30 10:40:54 <[florian]> kamikaze
Aug 30 10:41:46 <LoneStar> then i replace all whiterussian with kamikaze, right?
Aug 30 10:43:46 <[florian]> yes
Aug 30 10:44:30 <LoneStar> ok, do it now
Aug 30 11:03:00 <LoneStar> Checked out revision 8541.
Aug 30 11:05:38 <LoneStar> [*] Linux 2.4 (NEW)
Aug 30 11:05:49 <LoneStar> does that mean the default kernel is 2.4?
Aug 30 11:07:58 <LoneStar> where can I select the target cpu?
Aug 30 11:20:29 <LoneStar> i'm going to check out https://svn.openwrt.org/openwrt/trunk, right?
Aug 30 11:39:35 <LoneStar> I may be doing stupid by downloading a large directory
Aug 30 11:40:34 <loswillios> huh? just checkout https://svn.openwrt.org/openwrt/trunk/ and run make defconfig package/symlinks menuconfig
Aug 30 11:42:02 <LoneStar> i've got more than 100MB files and not seen an end
Aug 30 11:43:52 <LoneStar> so, that's normal
Aug 30 11:44:04 <loswillios> 127mb for me. look for the files in trunk/dl/
Aug 30 11:44:30 <loswillios> sure. you need something to compile, don't you?
Aug 30 11:44:49 <LoneStar> cross compiler?
Aug 30 11:45:09 <LoneStar> I have one, a bit old version
Aug 30 11:45:28 <loswillios> openwrt buildroot covers it all
Aug 30 11:45:46 <LoneStar> bootstrap?
Aug 30 11:46:18 <LoneStar> 取出修订版 8541
Aug 30 11:46:26 <LoneStar> it finished！
Aug 30 11:51:38 <LoneStar> loswillios, should I do something to enable soft_float for gcc?
Aug 30 12:47:22 <CIA-4> florian * r8542 /packages/libs/uclibc++/Makefile: Upgrade uclibc++ to 0.2.2, it does not break uclibc++ dependent packages that are already compiling
Aug 30 14:15:32 <CIA-4> nico * r8543 /trunk/package/acx-mac80211/Makefile:
Aug 30 14:15:32 <CIA-4> acx-mac80211 fixes:
Aug 30 14:15:32 <CIA-4>  * make Build/Compile conditionnal (only if in-kernel mac80211 is available)
Aug 30 14:15:32 <CIA-4>  * add a release number
Aug 30 14:19:25 <CIA-4> nico * r8544 /trunk/package/bcm43xx-mac80211/Makefile:
Aug 30 14:19:25 <CIA-4> bcm43xx-mac80211 fixes:
Aug 30 14:19:25 <CIA-4>  * make build depends on in-kernel mac80211 support
Aug 30 14:19:25 <CIA-4>  * rework kconfig hack
Aug 30 14:39:49 <CIA-4> florian * r8545 /packages/libs/taglib/Makefile: taglib now uses uclibc++ and compiles again
Aug 30 14:48:37 <CIA-4> florian * r8546 /packages/net/tinyproxy/patches/100-fix_typo.patch: Fix tinyproxy patch
Aug 30 14:50:29 <thepeople_mobile> thanks florian
Aug 30 14:50:58 <[florian]> you are welcome
Aug 30 14:55:59 <sn9> [florian]: what was with the goto in the wdt driver, anyway?
Aug 30 14:58:26 <[florian]> sn9: what's the problem with it ? this is an usual programming technique to handle errors in drivers instead of having a complex state machine to ensure you were at step x) and should free resources y,z ...
Aug 30 14:58:30 <[florian]> sn9: this really needs no patch
Aug 30 14:59:01 <sn9> but there are no resources referenced in this case
Aug 30 14:59:13 <[florian]> sn9: I said resources, this can be anything
Aug 30 14:59:19 <[florian]> don't bitch about that
Aug 30 14:59:43 <sn9> ok...
Aug 30 15:00:16 <[florian]> this won't make the driver work or not
Aug 30 15:05:48 <CIA-4> florian * r8547 /packages/multimedia/uvc-streamer/Makefile: uvc-streamer uses v4l2 which is 2.6 specific
Aug 30 15:38:12 <CIA-4> nico * r8548 /trunk/package/rt2x00/ (Makefile src/Makefile):
Aug 30 15:38:12 <CIA-4> rt2x00 fixes (thanks to sn9):
Aug 30 15:38:12 <CIA-4>  * make build depends on in-kernel mac80211 support
Aug 30 15:38:12 <CIA-4>  * rework kconfig handling
Aug 30 15:38:12 <CIA-4>  * download firmware images into $(DL_DIR)
Aug 30 15:38:13 <CIA-4>  * tweak source Makefile
Aug 30 15:43:58 <sn9> {Nico}: that is not my change
Aug 30 15:49:53 <{Nico}> right, but it should do the same
Aug 30 15:50:43 <{Nico}> sn9: are you following the development of rt2x00?
Aug 30 15:51:35 <sn9> the development has progressed to where it requires 2.6.23
Aug 30 15:51:54 <sn9> that's why i picked a snapshot known to be stable with .22
Aug 30 15:55:46 <{Nico}> you're working on rdc-2.6, right?
Aug 30 15:56:05 <sn9> and "should" is the key word in "should do the same" -- i'll investigate and adjust
Aug 30 15:56:40 <sn9> rt2x00 is necessary for rdc-2.6 at this point, yes
Aug 30 15:57:29 <{Nico}> i have tried it on a few arches with various pci/usb combinations and package selections
Aug 30 15:57:34 <{Nico}> but not on rdc-2.6
Aug 30 16:01:09 <sn9> the way i made the changes, they were supposed to make integrating later snapshots easier
Aug 30 16:04:27 <CIA-4> nico * r8549 /trunk/target/linux/generic-2.6/patches-2.6.22/601-usb_serial_endpoint_size.patch: add high-speed usb serial fix from Junxion (closes: #2274)
Aug 30 16:08:39 <CIA-4> thepeople * r8550 /trunk/package/dnsmasq/files/dnsmasq.init: fix typo closes #2289
Aug 30 16:09:15 <{Nico}> sn9: only the main Makefile is touched
Aug 30 16:11:21 <{Nico}> and it handles the case where rt2x00-{lib,pci,usb) are selected but other packages are not
Aug 30 16:11:57 <{Nico}> doesn't really make senses but would break builds
Aug 30 16:12:55 <sn9> as far as i can see, it does not handle that
Aug 30 16:13:38 <{Nico}> try it
Aug 30 16:18:47 <sn9> any particular reason to remove the _compat.h inclusion, other than annoyance when something is misconfigured?
Aug 30 16:37:57 <{Nico}> nope, i will revert that
Aug 30 16:41:11 <CIA-4> nico * r8551 /trunk/package/rt2x00/ (Makefile src/Makefile):
Aug 30 16:41:11 <CIA-4> rt2x00 fixes after [8548]:
Aug 30 16:41:11 <CIA-4>  * fix creating defines from kconfig
Aug 30 16:41:11 <CIA-4>  * revert auto-inclusion of rt2x00_compat.h
Aug 30 16:53:16 <CIA-4> nbd * r8552 /trunk/package/ (ppp/files/pppoe.sh pptp/files/pptp.sh): fix ifname in network state for pppoe and pptp, add a workaround for isps sending wrong netmasks for pptp (thx, moonwatcher)
Aug 30 16:59:35 <CIA-4> nico * r8553 /trunk/package/acx-mac80211/patches/003-endianness_fixes.patch: acx-mac80211: another bigendian fix
Aug 30 17:17:58 <CIA-4> thepeople * r8554 /trunk/package/dnsmasq/files/dnsmasq.init: fix incorrect parameters
Aug 30 17:56:30 <sn9> {Nico}: warning -> error, and slight improvements to the rt2x00 build: http://homepage.mac.com/danielg4/rdc.patch
Aug 30 18:57:32 <CIA-4> olli * r8555 /packages/net/ssmtp/ (3 files in 2 dirs):
Aug 30 18:57:32 <CIA-4> ssmtp:
Aug 30 18:57:32 <CIA-4>  * Cleanup Makefile
Aug 30 18:57:32 <CIA-4>  * Refresh patches
Aug 30 19:17:29 <CIA-4> olli * r8556 /packages/net/ssmtp/Makefile: ssmtp: some more cleanups
Aug 30 19:42:01 <LoneStar> everytime I type make world, it starts from the begining
Aug 30 19:42:30 <LoneStar> i've got gdb built again and again
Aug 30 19:42:57 <sn9> that shouldn't happen unless something changed or it didn't finish
Aug 30 19:44:40 <LoneStar> :(
Aug 30 19:45:54 <LoneStar> the build was stopped for several times
Aug 30 19:46:23 <LoneStar> the last time it said autoconf is not high enough :)
Aug 30 19:46:34 <sn9> what distro?
Aug 30 19:46:46 <LoneStar> redhat AS3
Aug 30 19:46:58 <sn9> that's way too old
Aug 30 19:47:15 <LoneStar> yes, I know now
Aug 30 19:49:13 <LoneStar> it proceed after I upgraded autoconf
Aug 30 19:50:22 <LoneStar> now I nohup it, and hope it go further when I wake up next morning ;)
Aug 30 19:51:44 <LoneStar> oh, btw, I learned that AR7 is a kind of ADSL chip, labeled TNETDxxxx
Aug 30 19:52:08 <LoneStar> do you know if it is the same as TNETV?
Aug 30 20:01:22 <sn9> some similarities
Aug 30 20:27:49 <xaiki_> hey what's the status of the easygate bcm mtd driver ?
Aug 30 21:11:12 <CIA-4> nbd * r8557 /trunk/include/quilt.mk: allow the kernel to provide own series files for specific subdirs
Aug 30 21:12:33 <CIA-4> nbd * r8558 /trunk/ (7 files in 3 dirs): clean up recursive dependency handling, use timestamp.pl again, because it saves memory and execution time
Aug 30 22:19:20 <CIA-4> nico * r8559 /trunk/ (2 files in 2 dirs): build eeprom_93cx6 as a module, add a KernelPackage for it and for crc-itu-t
Aug 30 22:23:38 <sn9> {Nico}: how is that different from ecb or arc4?
Aug 30 22:24:11 <CIA-4> nico * r8560 /trunk/package/rt2x00/ (Makefile src/config src/rt2x00_compat.h):
Aug 30 22:24:11 <CIA-4> more rt2x00 fixes (thanks to sn9):
Aug 30 22:24:11 <CIA-4>  * pkg: add missing kmod dependencies and autoloads
Aug 30 22:24:11 <CIA-4>  * src: change #warning to #error when checking kconfig
Aug 30 22:24:11 <CIA-4>  * src: remove uneeded config file
Aug 30 22:24:53 <{Nico}> sn9: arc4 and ecb are included in the kmod-crypto package
Aug 30 22:25:25 <{Nico}> while eeprom-93cx6 and crc-itu-t were not included in any package
Aug 30 22:25:25 <sn9> oh, then that should be a dependency, too
Aug 30 22:25:39 <{Nico}> added them already
Aug 30 22:25:57 <{Nico}> and skipped firmware_class as it is buitin
Aug 30 22:26:48 <{Nico}> i will commit rdc-2.6 changes later as i don't have any build tree for it yet
Aug 30 22:28:08 <sn9> what about ifneq ($(CONFIG_PACKAGE_kmod-rt2x00-lib),) ?
Aug 30 22:28:31 <sn9> CONFIG_MAC80211 is already a dependency, too
Aug 30 22:29:22 <{Nico}> actually, it's more. it's used to check if we're on 2.6.22 or later
Aug 30 22:29:39 <{Nico}> and if not, we won't build anything
Aug 30 22:30:57 <sn9> IMHO, that belongs in kmod-mac80211
Aug 30 22:31:35 <sn9> that way, no modules that depend on it get built on .21
Aug 30 22:32:14 <{Nico}> it does not work like that anymore
Aug 30 22:32:39 <{Nico}> when you select a package, the kconfig gets modified to reflect your selection
Aug 30 22:33:00 <sn9> and?
Aug 30 22:33:15 <{Nico}> but on a 2.6.21 or earlier, there's symbol for it, so it does not appear in the kconfig
Aug 30 22:33:37 <{Nico}> s/there's/there's no/
Aug 30 22:34:07 <sn9> and?
Aug 30 22:35:14 <{Nico}> CONFIG_MAC80211 is undefined and ifneq ($(CONFIG_MAC80211),) is false
Aug 30 22:35:38 <sn9> all the more reason for the test to go in kmod-mac80211
Aug 30 22:36:41 <{Nico}> that won't solve the problem
Aug 30 22:37:25 <sn9> if kmod-mac80211 is made to require >=.22, neither rt2x00 nor anything else that requires mac80211 need concern itself with it
Aug 30 22:37:25 <{Nico}> because we can't prevent someone to select kmod-mac80211 in menuconfig
Aug 30 22:37:53 <sn9> yes, we can
Aug 30 22:38:03 <{Nico}> how?
Aug 30 22:38:24 <sn9> DEPENDS:=
Aug 30 22:39:26 <sn9> in kmod-mac80211
Aug 30 22:40:22 <{Nico}> sure, we'll talk about that later, i have to go...
Aug 30 22:40:36 <{Nico}> see you later
Aug 30 22:43:20 <CIA-4> nbd * r8561 /trunk/ (include/subdir.mk package/Makefile target/Makefile): fix a dependency bug
Aug 31 00:32:26 <Weedy_lappy> https://dev.openwrt.org/changeset/8552
Aug 31 00:32:33 <Weedy_lappy> FUCKING FINALY
Aug 31 00:32:39 * Weedy_lappy hugs nbd 
Aug 31 01:29:15 <ByronT> hello
Aug 31 01:29:22 <wrt> ByronT?
Aug 31 01:29:36 <ByronT> yes?
Aug 31 01:30:11 <ByronT> ah... bot
Aug 31 01:33:29 <ByronT> good gracious... you guys weren't kidding about initial setup time on a nslu2
**** ENDING LOGGING AT Fri Aug 31 02:59:56 2007