**** BEGIN LOGGING AT Mon Feb 21 02:59:57 2011 Feb 21 06:15:02 lars * r25644 /trunk/target/linux/xburst/patches-2.6.37/300-battery-fixes.patch: [xburst] Refresh patches for 2.6.37.1 Feb 21 09:25:33 florian * r25645 /packages/net/asterisk-addons-1.4.x/: [packages] remove asterisk-addons-1.4.x (#8898) Feb 21 09:28:06 kaloz * r25646 /trunk/package/iptables/ (Makefile patches/005-imq.patch): [package/iptables]: as [25641] removed kernel IMQ support, remove it from iptables as well Feb 21 09:36:14 florian * r25647 /trunk/target/linux/ar7/base-files/etc/uci-defaults/ (. network): [ar7] add a default network config file, patch from Wipster Feb 21 10:18:25 what is the "Download Folder" option for? (it's in Advanced Configuration Options in menuconfig) Feb 21 10:19:15 Olipro: to set a common folder Feb 21 10:19:22 Olipro: to set a common dl folder Feb 21 10:19:33 i have several trees i dont want to have a dl/ for each Feb 21 10:19:45 so i can either symlink a generic dl/ or use the option Feb 21 10:19:46 ah ok Feb 21 10:20:05 what option do I want to set to give it a URL for fetching the package list? Feb 21 10:20:16 i.e. the URL the router will use when it's actually running OpenWRT Feb 21 10:22:18 also, is it safe to use ccache yet? Feb 21 10:27:19 never used opkg nor ccache Feb 21 10:27:20 sorry Feb 21 10:41:37 <[florian]> Olipro: it should be safe to use ccache now Feb 21 10:42:00 <[florian]> Olipro: some packages that are not using quotes when passing the TARGET_CC variable to makefiles might still fail; but it's easy to fix anyway Feb 21 11:43:54 kaloz * r25648 /trunk/toolchain/binutils/ (Config.in Makefile patches/2.19.1+cs/): [toolchain/binutils]: remove the CS variant - broken since it's addition Feb 21 12:09:38 <_trine> iax2 does not seem to be working in asterisk 1.8 Feb 21 13:09:30 acoul * r25649 /trunk/package/madwifi/patches/474_fix_ssid_scan_length.patch: package/madwifi: extend wlanconfig scan ssid field from 14 to 30 characters Feb 21 13:17:18 acoul * r25650 /packages/net/quagga/patches/000-upstream-bgp_lock_node.patch: net/quagga: add an upstream pending patch Feb 21 13:26:57 anyone else having issues compiling libnet? Feb 21 13:26:58 libs/libnet_resolve.o: relocation R_MIPS_26 against `socket' can not be used when making a shared object; recompile with -fPIC Feb 21 13:42:30 hmm Feb 21 13:42:36 hey that must be it! Feb 21 13:42:50 hm? Feb 21 13:42:51 * PsyMan looks at [florian] Feb 21 13:43:02 ar7 switch patch Feb 21 13:44:01 I think OpenWRT needs a trunk and a trunk_volatile :P Feb 21 13:44:25 commit to trunk volatile, if it doesn't break the build, move to trunk Feb 21 13:44:26 Olipro-: which libnet? there are two Feb 21 13:44:55 build_dir/target-mips_r2_uClibc-0.9.32/Libnet-1.0.2a Feb 21 13:47:55 cannot reproduce Feb 21 13:48:00 builds fine here Feb 21 13:48:10 also with -fPIC Feb 21 13:48:35 <[florian]> PsyMan: yes? Feb 21 13:49:31 that last ar7 commoti was the switch fixup? :p Feb 21 13:49:47 <[florian]> it should make marvell switches work Feb 21 13:49:49 hm, ccache is also buggered Feb 21 13:49:59 neat Feb 21 13:50:04 >_< Feb 21 13:50:09 now only thing remaining is acx from git Feb 21 13:50:21 and i can ditch that old f/w Feb 21 13:50:31 <[florian]> try it first ;) Feb 21 13:51:29 well, then i'll have to reflash Feb 21 13:51:44 and flashing once is time consumning by itself Feb 21 13:51:51 consuming Feb 21 13:52:07 and bothersome Feb 21 13:52:25 <[florian]> well, yes, but that's the price for new software :) Feb 21 13:53:04 so you more or less suggest to try and update the package after making sure the switch works fine Feb 21 13:53:32 well, when the acx package is updated Feb 21 13:53:42 <[florian]> acx is already updated in trunk Feb 21 13:53:48 <[florian]> are we talking about you flashing trunk? Feb 21 13:53:59 <[florian]> because if you are trying backfire, for sure the situation has not chnaged Feb 21 13:54:28 actually, the plan is to use a trunk snaapshot Feb 21 13:54:32 snapshot Feb 21 13:54:49 <[florian]> ok, then you might want to make sure that buildbot has built one up-to-date for you Feb 21 13:55:10 not yet Feb 21 13:55:17 probably toommorow Feb 21 13:55:21 tommorow Feb 21 13:56:01 kmod-acx-mac80211_2.6.32.29+20101030-1_ar7.ipk <- looks old :| Feb 21 13:56:24 that is. Feb 21 13:56:41 When executing 'quilt push 001-Makefile.am.patch', it complains with this warning message "warning: the `--unified-reject-files' option is obsolete; use `--reject-format=unified' instead". How do I suppress such a WARNING statement? Feb 21 13:56:47 <[florian]> are you guys saying that just because of the date, or have you checked acx-mac80211 git actually :) ? Feb 21 13:57:06 just because of date Feb 21 13:57:09 build #77 of ps3 is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/77 Feb 21 13:57:11 [florian]: i'm pretty sure there were a bunch of updates well after that.. Feb 21 13:57:14 it uses git to build? Feb 21 13:57:15 :p Feb 21 13:57:17 after it moved back to sf.net Feb 21 13:57:24 PsyMan: it uses the git checkout date as the revision Feb 21 13:57:58 ehh? Feb 21 13:58:04 [florian]: quite a lot of updates in shortlog :) Feb 21 13:58:08 http://acx100.git.sourceforge.net/git/gitweb.cgi?p=acx100/acx-mac80211 Feb 21 13:58:33 didn't get that :| Feb 21 13:59:13 PsyMan: the mac80211 version of the acx driver has no releases as such. it uses git as its version control repository, and the date in the package name is when the source was checked out of git Feb 21 13:59:16 do you mean that it uses a predefined revision/date or that it uses the latest one whenever it builds? Feb 21 13:59:24 <[florian]> I usually update that based on what the acx-mac80211 tells me, whether it's worth or not updating the snapshot Feb 21 13:59:59 things like the recalibration and sensitivity updates are likely well worth updating for Feb 21 14:00:11 indeed Feb 21 14:00:12 the spinlock check disabling not so mcuh Feb 21 14:00:12 :p Feb 21 14:00:14 much* Feb 21 14:00:30 also the improving of rate handling.. which was all in december :) Feb 21 14:01:17 <[florian]> EqUaTe: allright Feb 21 14:01:38 yay!? Feb 21 14:01:47 that said, the acx drivers are a royal pain anyway.. Feb 21 14:05:57 hmm Feb 21 14:06:00 reminds me Feb 21 14:06:14 last time i tried openwrt on that thing Feb 21 14:06:22 it would connect wirelessly Feb 21 14:06:37 but wouldn't let me connect to the router Feb 21 14:06:44 ping would fail too Feb 21 14:06:53 but that was backfire iirc Feb 21 14:06:59 yeah, i used to get that behaviour Feb 21 14:07:16 i rolled an update for acx-mac80211 and could get connectivity some of the time Feb 21 14:07:28 what device is it PsyMan ? Feb 21 14:07:43 dg834g Feb 21 14:07:54 don't know it.. similar to the linksys wag354g? Feb 21 14:08:25 lemme c Feb 21 14:08:43 yeah Feb 21 14:08:53 k Feb 21 14:08:59 so not the worlds best device then :) Feb 21 14:09:29 it was ok for back then Feb 21 14:09:38 even now, it's stable as rock Feb 21 14:09:49 they seem to be fairly good as adsl modems Feb 21 14:09:53 only issue is the dodgy software Feb 21 14:09:55 the wifi chip sucks though Feb 21 14:10:04 ? Feb 21 14:10:07 why? :p Feb 21 14:10:15 stock or non-stock, the wifi is poor Feb 21 14:10:27 stock, at least on the linksys, supports only as far as wpa1. Feb 21 14:10:35 yeah, same here Feb 21 14:10:38 anything that adds wpa2 has been unstable as all hell with wifi Feb 21 14:10:41 that's the issue Feb 21 14:10:55 that and upnp Feb 21 14:10:57 acx-mac80211 is the best bet, but it's taking its time. Feb 21 14:11:01 upnp i don't care about Feb 21 14:11:05 in fact i think it's a horrible idea :) Feb 21 14:11:21 that 10-10 revision works? Feb 21 14:11:34 or it has connection issues? Feb 21 14:12:05 no idea Feb 21 14:12:17 hmm Feb 21 14:12:24 so not tested there huh Feb 21 14:13:19 i'm sure some people claimed getting it to work using some january build Feb 21 14:13:30 we'll see i suppose Feb 21 14:14:26 i haven't done anything with the wag354g for a few months now Feb 21 14:14:44 it's been rather mad :( Feb 21 14:16:00 :\ Feb 21 14:17:02 work is finally starting to quiet down though Feb 21 15:30:36 build #67 of etrax is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/etrax/builds/67 Feb 21 16:03:56 I unsucessfully tried to compile asterisk-1.8.x for my Seagate DockStar. The compilation fails on asterisk-1.8.2.3/addons/app_mysql.c as shown on line# 4 (http://openwrt.pastebin.com/VGnn3ktc). The Makefile doesn't mention of needing mysql as one of dependencies. Feb 21 16:17:50 build #66 of ep93xx is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ep93xx/builds/66 Feb 21 16:32:38 anyone seen this before; Feb 21 16:32:44 warning: libcom_err.so.3, needed by ../lib/libext2fs.so, not found (try using -rpath or -rpath-link) Feb 21 16:32:45 ../lib/libext2fs.so: undefined reference to `error_message@com_err_3_MIT' Feb 21 16:32:45 ../lib/libext2fs.so: undefined reference to `com_err@com_err_3_MIT' Feb 21 16:38:40 build #77 of gemini is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/77 Feb 21 16:43:36 is there anyone in OpenWRT who focuses on maintaining the build system? Feb 21 16:47:05 build #14 of lantiq is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/lantiq/builds/14 Feb 21 17:10:47 build #88 of brcm47xx is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/88 Feb 21 17:30:38 build #65 of rdc is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/rdc/builds/65 Feb 21 17:34:12 build #66 of adm5120 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/66 Feb 21 17:35:50 the kernel config seems to be broken Feb 21 17:50:54 build #79 of pxcab is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/79 Feb 21 17:52:42 moin Feb 21 17:53:06 is there a (very rough) ETA for the 10.03.1 release? Feb 21 18:05:24 i know Feb 21 18:05:30 when it's done :) Feb 21 18:06:17 oh, i'm not asking for an exact date or anything Feb 21 18:06:42 i didn't give an exact date either Feb 21 18:06:44 more like an order of magnitude, like "this week/month/year" Feb 21 18:06:59 this decade :) Feb 21 18:07:02 or a list of bugs that are in need of fixing or whatever.. Feb 21 18:07:09 sometime between now and then Feb 21 18:08:17 my boss is hell-bent on using "just releases" even though the 10.03 branch from after the release carries my patches.. Feb 21 18:09:01 so i'm trying to filibuster the migration back to the 10.03 release for long enough that i don't have to do it... Feb 21 18:09:54 looks like i won't get around this particular boring work :-/ Feb 21 18:09:56 DonDiego: I think 1-2 months is reasonable Feb 21 18:10:24 but don't quote me on that Feb 21 18:11:34 ok, that gives me a rough idea.. Feb 21 18:12:08 DonDiego: the brnach itself is in good shape but the package repo has many issues Feb 21 18:12:29 we still need to backport a bunch of autotools fixups to backfire in order to solve that Feb 21 18:20:42 ok, thanks for the info Feb 21 19:30:12 packages/lib/jpeg when building on darwin host still fails because of bad ltconfig/ltmain.sh, it cannot detect darwin, and disables shared libraries from being built Feb 21 19:30:38 but detecting darwin is false, since its building for openwrt Feb 21 19:31:14 what is the proper way to specify the host to configure in an openwrt makefile? Feb 21 19:31:28 host = build host Feb 21 19:31:35 paste V=99 output and config.log please Feb 21 19:35:56 its probably doing an uname -m somewhere Feb 21 19:36:39 clean,prepare,configure: http://openwrt.pastebin.com/iiJWVzbt .. config.log: http://openwrt.pastebin.com/8B2WGRhP ... and since no shared libs are built, the pkg install fails when trying to copy the *.so files Feb 21 19:43:46 nevermind what i said about darwin, that is what it was doing before the recent libtool fixups Feb 21 19:44:15 but it seems to be failing on that reallocatable error Feb 21 19:44:58 jow * r25651 /packages/lang/php4/patches/110-openssl-1.0.0-compat.patch: [packages] php4: fix OpenSSL v1.0.0 fallout... Feb 21 19:45:27 I see the same error on linux, yet shared libs are built Feb 21 19:45:48 its some ancient ltconfig crap Feb 21 19:46:19 yea Feb 21 19:46:44 i've been working around by using the old hack of just linking ltconfig to true Feb 21 19:46:56 very good idea Feb 21 19:47:08 maybe I should incorperate that in autotools.mk Feb 21 19:47:16 since the ltconfig is old as hell Feb 21 19:48:03 the only problem with the old Makefile hack it hard coded ltconfig to /bin/true, but on darwin its /usr/bin/true ... just an FYI Feb 21 19:48:22 $$$$(which true) works nicely Feb 21 19:50:25 jow * r25652 /trunk/package/jshn/Makefile: [package] jshn: add build dependency on libubox, it needs the list.h header from it Feb 21 19:52:22 also with jpeg, when I don't specify LIBTOOL="$(STAGING_DIR)/to/path/libtool" to $(MAKE), I always get "make[3]: ./libtool: No such file or directory" Feb 21 19:52:35 trunk or backfire? Feb 21 19:52:39 trunk Feb 21 19:52:59 that is after i link ltconfig to true in the prepare target Feb 21 19:53:38 hm Feb 21 19:53:53 here is my simple makefile change: http://openwrt.pastebin.com/wfFx0eAi Feb 21 19:53:54 they ship no autofail templates, so autoreconf is not started Feb 21 19:54:52 man, every time i talk with you, i more and more hate autoconf and friends Feb 21 20:16:35 dirtyfreebooter: http://paste.openwrt.org/d292b72d8 Feb 21 20:24:27 xMff: works! thanks. Feb 21 20:36:38 xMff: on your router, does both your pppoe interface and lan have an inet IPv6 address, or only the lan? Feb 21 20:37:01 (by inet, I mean an address accessible from the internet) Feb 21 20:37:02 I don't have pppoe but both my wan and lan have an address, yes Feb 21 20:37:11 both public too Feb 21 20:37:18 yeah, on mine that doesn't work it seems Feb 21 20:37:38 I put the IP on the br-lan and the pppoe-wan2 doesn't have any inet IP Feb 21 20:37:40 and now everything works Feb 21 20:58:33 jow * r25653 /branches/backfire/include/autotools.mk: [backfire] autotools.mk: the old libtool fixup never fully worked, correct it Feb 21 21:00:41 I wonder how much bombs out now Feb 21 21:02:18 jow * r25654 /packages/libs/jpeg/Makefile: [packages] jpeg: package is not autoreconfigurable and ships broken libtool, fixup manually Feb 21 21:03:22 who is responsible for the libre / restund stuff? Feb 21 21:03:33 it still does not seem to work Feb 21 21:03:43 who maintains triggerhappy? it never compiled here Feb 21 21:10:28 thepeople: ping? ;) Feb 21 21:14:32 jow * r25655 /packages/net/l2tpv3tun/ (Makefile patches/100-nl_handle_alloc.patch): [packages] l2tpv3tun: fix libnl 2.0 fallout... Feb 21 23:59:04 hmmm, getting this with new qos-scripts --> Cannot find device "ifb0", maybe qos-scripts should depend on PACKAGE_kmod-ifb ? Feb 22 00:01:28 nbd: ^ Feb 22 00:04:03 I need to downgrade my local OpenWRT to r25010. If I do 'svn up -r 25010', then the packages under feeds won't get downgraded. What is the proper way to downgrade? Feb 22 00:08:13 maybe a new checkout with -r Feb 22 00:08:24 dunno for the feeds in fact Feb 22 00:10:10 stintel: oh, right Feb 22 00:10:58 nbd * r25656 /trunk/package/qos-scripts/Makefile: qos-scripts: add missing dependency Feb 22 00:14:46 Epsylon3: you can add the revision in feeds.conf, as in http://downloads.qi-hardware.com/software/images/NanoNote/Ben/2010-12-14/feeds.conf Feb 22 00:15:06 nbd: thanks Feb 22 00:25:58 :) thanks stintel Feb 22 00:28:02 Epsylon3: welcome Feb 22 00:28:10 anybody had a change to look at https://dev.openwrt.org/ticket/8590 ? Feb 22 00:28:49 i dont use qos here Feb 22 00:38:18 hmm, time for bed, nn all Feb 22 00:38:28 good night Feb 22 00:43:04 /bin/sh: can't load library 'librpc.so' <- looks like I need to clean my config Feb 22 00:59:04 yeah Feb 22 00:59:23 maybe a make dirclean as wel Feb 22 01:01:11 ah well, all I wanted was to see if it accepts the image, booting was optional ;) Feb 22 01:06:02 what is the good way to make image after "make package/kernel/compile" Feb 22 01:06:16 i just dont want to recompile all Feb 22 01:06:32 i'm doing some tests on platform setup Feb 22 01:10:23 Epsylon3: what image? Feb 22 01:10:25 kernel? Feb 22 01:10:34 no Feb 22 01:10:38 the full one Feb 22 01:10:46 squashfs Feb 22 01:10:50 well, that requires full compile Feb 22 01:11:09 erf ok.... Feb 22 01:11:10 which should be fairly fast, if no changes were made Feb 22 01:11:24 yea fast... Feb 22 01:14:41 maybe there is another way to "patch" kernel image only ? Feb 22 01:15:09 maybe with kexec Feb 22 01:15:50 sure Feb 22 01:16:16 also, make target/linux/install will rebuild the kernel image, only Feb 22 01:16:23 and modules Feb 22 01:20:58 but it won't include the updated kernel modules in the image Feb 22 01:22:58 xMff: there was a patch for l2tpv3tun sent in for exactly that issue, but it must have been overlooked. Feb 22 01:23:05 yea, modules are secured by some kernel crc ? Feb 22 01:23:18 i dont want to change modules here, only platform setup Feb 22 01:23:44 rmmm compiling since 20mn Feb 22 01:23:49 just for 2 lines Feb 22 01:24:06 kernel is compiled in 30 seconds (alone) Feb 22 01:24:43 maybe the package/cleanup Feb 22 01:27:44 25mn.... waiting Feb 22 01:28:08 really fast Feb 22 01:29:03 is your machine older than 10 years? Feb 22 01:29:12 Or do you have thousands of packages selected? Feb 22 01:29:21 yea Feb 22 01:29:23 i've Feb 22 01:29:49 its why i asked for a "fast" alternative Feb 22 01:29:59 <_trine> are there 1000's of packages ? Feb 22 01:30:00 to prevent make clean before Feb 22 01:30:16 if you only select the essential packages it speeds up things significantly. Feb 22 01:30:39 sure... but i dont want to change the .config Feb 22 01:30:49 else i will need 10 hours Feb 22 01:30:54 yeah, well. There's no free candy :) Feb 22 01:31:09 <_trine> mb__, how come you're not as grumpy as you once were :) Feb 22 01:31:25 huh? Feb 22 01:31:37 <_trine> you seem to be nice these days Feb 22 01:31:54 <_trine> a good sign if I may say so Feb 22 01:32:27 * mb__ scratches head Feb 22 01:32:38 <_trine> oh well nvm Feb 22 01:36:34 so well 35mn for a make -r world: build failed. Please re-run make with V=99 to see what's going on Feb 22 01:37:17 the weird thing is : package/kernel/compile works Feb 22 01:37:26 target/linux install fail Feb 22 01:41:31 I am getting: configure: error: Building GCC requires GMP 4.2+, MPFR 2.3.1+ and MPC 0.8.0+. This is a problem with my build environment right? (ubuntu) Feb 22 01:42:04 netprince_: which target? Feb 22 01:42:16 geode Feb 22 01:42:29 did you run make oldconfig or make menuconfig after updating? Feb 22 01:43:03 I did run make menuconfig to update all my target config files, I can check it again Feb 22 01:44:05 I'm working on r25562 Feb 22 01:44:24 the answer to my question was "make target/install" Feb 22 01:44:38 that compile linux kernel too Feb 22 01:44:45 [florian]: since I updated to the coreutils changes in SVN, I can no longer select coreutils-* via ~/.openwrt/defconfig ... Feb 22 01:45:03 nbd: Target Profile (PCEngines Alix) Feb 22 01:45:05 digging. Feb 22 01:47:01 Linux version 2.6.37.1 (root@epsybox) (gcc version 4.5.2 (Linaro GCC 4.5-2011.02-0) ) #3 Tue Feb 22 02:43:12 CET 2011 Feb 22 01:47:14 ok that works Feb 22 01:47:50 nbd: more info: http://openwrt.pastebin.com/E4LFibTe Feb 22 01:49:00 console=ttyS0,115200 root=/dev/mtdblock1 rw ip=192.168.1.1:192.168.1.2:::DB88FXX81:eth0:none Feb 22 01:49:18 what is this ip cmd line (i got that from uboot) Feb 22 01:49:27 192.168.1.1:192.168.1.2 Feb 22 01:49:31 there is 2 ips ? Feb 22 01:49:36 netprince_: you should update to latest Feb 22 01:49:45 nbd: ok, will try that Feb 22 02:47:04 build #79 of ppc44x is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ppc44x/builds/79 Feb 22 02:49:11 what's a good example of a LED platform driver for an x86 platform using GPIO? I need something to clone... Feb 22 02:57:56 wow hmm good question Feb 22 02:59:42 ls /lib/modules/2.6.35-25-generic/kernel/drivers/leds/ **** ENDING LOGGING AT Tue Feb 22 02:59:57 2011