**** BEGIN LOGGING AT Tue Jul 26 02:59:56 2011
Jul 26 03:42:34 <^Willie^> hmm segfaults using eglibc builds turns out permissions on half of the files are not right.
Jul 26 03:43:05 <^Willie^> a bit odd that all that not happend when using uclibc ..
Jul 26 03:44:31 <^Willie^> how can i fix that before the build ?
Jul 26 05:47:18 <^Willie^> hmm atleast rrdtool 1.3.8 is fully functional. incl graphs
Jul 26 08:05:57 <solca> hello, how do you guys resync the target kernel config file?
Jul 26 08:06:52 <solca> I always get asked in 'make oldconfig' but I have proper symbols in config-3.0
Jul 26 08:07:03 <solca> maybe they are not in the proper place
Jul 26 08:18:31 <KanjiMonster> solca: make kernel_oldconfig takes care of updating the target config
Jul 26 08:19:53 <KanjiMonster> if you then still get asked on a build, then these are config symbols made available through kmods and should go into generic
Jul 26 08:30:28 <solca> KanjiMonster: thx that fixed one but I still get another from it seems a new driver
Jul 26 08:30:38 <solca> EHCI support for AR7XXX/AR9XXX SoCs (USB_EHCI_ATH79)
Jul 26 08:30:56 <solca> EHCI support for AR7XXX/AR9XXX SoCs (USB_EHCI_ATH79) [Y/n/?] (NEW)
Jul 26 08:31:17 <solca> that should go into the generic config as N ?
Jul 26 08:31:19 <KanjiMonster> solca: hm, that shouldn't be available; are you building for ATH79?
Jul 26 08:31:32 <solca> nope, just for AR71xx
Jul 26 08:33:26 <KanjiMonster> solca: hm, no that should be in target/linux/ar71xx since its ar71xx/ath79-specific; and "n" I think (you have add the line manually to target/linux/...
Jul 26 08:35:27 <solca> yeah I though that too, will search why it pop ups in ar71xx, thx!
Jul 26 08:35:45 <^Willie^> KanjiMonster: libstdc++ is rarly used is it ?
Jul 26 08:36:26 <^Willie^> afther install still need to run manual chmod +x
Jul 26 08:36:45 <KanjiMonster> solca: probably because "depends on USB_EHCI_HCD && (SOC_AR71XX || SOC_AR724X || SOC_AR913X)" and the SOC_* are also used/defined by AR71XX
Jul 26 08:37:18 <KanjiMonster> ^Willie^: hm, never had to do that (on uClibc that is)
Jul 26 08:40:21 <solca> KanjiMonster: so it's a valid driver in ar71xx
Jul 26 08:41:33 <KanjiMonster> solca: yes and no - the problem is that in openwrt there are *two* targets for the same chips, AR71XX and ATH79, so you have some duplicates (the mainline kernel driver for ath79 and the openwrt driver for ar71xx)
Jul 26 08:42:37 <solca> it seems ATH79 is a new board type or something
Jul 26 08:42:46 <solca> in Linux 3.0
Jul 26 08:43:09 <solca> machine type
Jul 26 08:43:40 <loswillios> it's ar71xx partially merged upstream AFAIK
Jul 26 08:44:14 <solca> loswillios: it handles just reference boards AFAICT
Jul 26 08:45:06 <solca> KanjiMonster: so if I undef CONFIG_USB_EHCI_ATH79 in the target config I should be fine?
Jul 26 08:45:32 <^Willie^> KanjiMonster: uclibc is maintained yes on the other 2 libc's people do run into weird issues that can be expected to be fixed if was fixed for uclibc :(
Jul 26 08:45:52 <KanjiMonster> solca: yes
Jul 26 08:46:42 <KanjiMonster> and ATH79 is there since 2.6.38, but it's currently incomplete (e.g. no ethernet driver yet)
Jul 26 08:47:29 <solca> oh
Jul 26 08:47:56 <solca> I wonder why kernel devels don't pick openwrt patches as is
Jul 26 08:48:19 <solca> openwrt patches seems of very good code quality
Jul 26 08:48:56 <solca> maybe NIH syndrome :)
Jul 26 08:48:58 <KanjiMonster> because they don't conform fully to kernel code style, and often have quite a lot of workarounds that have to be cleaned up first
Jul 26 08:49:07 <KanjiMonster> also they have to be submitted by someone ;)
Jul 26 08:49:26 <solca> oh
Jul 26 08:49:44 <KanjiMonster> kernel devs don't go actively looking for stuff to add, they usually have their own hands full
Jul 26 08:51:23 <solca> right, at least MIPS boards are more or less in control, ARM is out of control IMHO
Jul 26 08:53:29 <solca> KanjiMonster: do you think it could be a good idea to make a patch to remove ATH79 so it doesn't conflict with AR71xx?
Jul 26 08:54:21 <KanjiMonster> solca: I'd rather see a patch to convert anything that's missing to use the existing ATH79 (since that needs to be done eventually anyway)
Jul 26 08:55:11 <loswillios> yup
Jul 26 08:55:37 <solca> but that would mean deprecate target/linux/ar71xx and create a new target/linux/ath79
Jul 26 09:01:33 <KanjiMonster> solca: no
Jul 26 09:02:04 <KanjiMonster> brcm47xx is actually BCM47XX, and brcm63xx is actually BCM63XX and they didn't get renamed ;)
Jul 26 09:03:52 <solca> good point, and the ar71xx_ namespace in all source code should be updated to ath79_ ?
Jul 26 09:04:57 <KanjiMonster> solca: yes
Jul 26 09:05:54 <solca> KanjiMonster: ok
Jul 26 09:06:44 <solca> KanjiMonster: I thought the kernel was compiled in target/linux/compile but it's getting compiled in target/linux/install, do you know why?
Jul 26 09:07:07 <KanjiMonster> short answer: no
Jul 26 09:07:21 <KanjiMonster> I'd have to look that up ;)
Jul 26 09:07:38 <solca> ok thx for all the help!!
Jul 26 10:26:47 <solca> nbd: finally finish reading the quilt tutorial you point me
Jul 26 10:26:59 <solca> great tool, glad is very good integrated into openwrt
Jul 26 10:27:42 <loswillios> yup, makes stuff a lot easier
Jul 26 10:48:01 <CIA-89> loswillios * r27778 /packages/sound/pulseaudio/Makefile: [packages] pulseaudio: disable avai
Jul 26 10:48:31 <loswillios> *avahi
Jul 26 10:54:21 <CIA-89> jow * r27779 /trunk/package/linux-atm/Makefile: [package] linux-atm: don't fail if an PREFIX env var is set
Jul 26 10:56:08 <CIA-89> jow * r27780 /branches/backfire/package/ (linux-atm/Makefile lua/Makefile): [backfire] merge r27779
Jul 26 10:57:20 <CIA-89> jow * r27781 /branches/backfire/package/lua/Makefile: [backfire] revert unrelated change in previous commit
Jul 26 11:20:09 <^Willie^> hmm just for a perl module -> i486-openwrt-linux-gnu-ld: unrecognised emulation mode: arch=i486 Supported emulations: elf_i386 i386linux
Jul 26 11:45:50 <builder_> build #62 of s3c24xx is complete: Failure [failed compile_10]  Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/62
Jul 26 12:48:06 <CIA-89> loswillios * r27782 /packages/sound/pulseaudio/Makefile: [packages] pulseaudio: remove optional gdbm dependency, bump PKG_RELEASE
Jul 26 13:21:23 <builder_> build #40 of mpc52xx is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/40
Jul 26 13:27:49 <gzanan> Does openwrt support RT3070 ar9217 right now?
Jul 26 14:31:28 <netprince> can anyone recommend a device which can forward packets at 500Mb (half a Gbps) minimum?
Jul 26 14:38:11 <WillieNL> wat if a dep of the package start his own configure/compiler process inside a build but using autoconf results in march=k8 ?
Jul 26 14:38:31 <WillieNL> host machine is amd64 indeed but from a i686 chroot.
Jul 26 14:38:47 <nbd> WillieNL: beat it up until it behaves ;)
Jul 26 14:38:59 <WillieNL> nbd: hmm 2.5 hours further..
Jul 26 14:39:11 <WillieNL> no clue, except rrdtool is not designed to crosscompile.
Jul 26 14:39:22 <nbd> there are no generic solutions for that kind of issue
Jul 26 14:39:36 <nbd> except: read the code, understand it, devise a fix
Jul 26 14:40:03 <WillieNL> rrdtool 1.3.8 works for me with graphs. but not with perl bindings yet and just drop the RRDs.pm (textfile) in the target was not enough. so i'm forced to fix something in there..
Jul 26 14:40:18 <WillieNL> nbd: lolz understand rrdtool ?
Jul 26 14:40:30 <nbd> no, understand the build system
Jul 26 14:40:34 <WillieNL> always a surprice where it fails next time or why it start including stuff we don't want
Jul 26 14:40:42 <WillieNL> hmm
Jul 26 14:41:20 <WillieNL> not shure if it can be build external outside the rrdtool tree as a seperate package linking against librrd and perl.
Jul 26 14:41:37 <WillieNL> not shure if dynamic linking will make that posible. but my rrd is dynamic linked.
Jul 26 14:43:07 <WillieNL> nbd: atleast i never had to deal with this kind of issues in 7 years gentoo :S
Jul 26 14:43:17 <WillieNL> also always avoided cross-compiling for this.
Jul 26 14:43:42 <nbd> well, lots of build systems look like they were designed by monkeys with typewriters
Jul 26 14:43:56 <nbd> and contain stupid stuff that breaks cross compiling in many interesting ways
Jul 26 14:44:06 <jow_laptop> AC_TRY_RUN() :D
Jul 26 14:45:21 <WillieNL> like hard define march=k8 in a makefile due a makefile.ac  file use cpu host /arch detection instaid of distro arch..
Jul 26 14:45:35 <WillieNL> something that will fail always if i686 on x68-64 like i do.
Jul 26 14:46:38 <WillieNL> jow_laptop: also the not luci included collectd modules are working  using collectd-web on this box atm for the extra modules.
Jul 26 14:46:55 <WillieNL> but wanna have collectd-web in my target :)
Jul 26 14:47:19 <WillieNL> perl:json and rest is done already
Jul 26 14:53:09 <blaise> jow_laptop: do you sleep?
Jul 26 14:53:47 <blogic> blaise: 9 minutes ago he was typing
Jul 26 14:54:06 <blogic> so unless he has a ueber skill of typing while sleeping, he probably is not sleeping
Jul 26 14:54:18 <blogic> although that would be a cool skill to have
Jul 26 14:54:22 <blaise> a lot of irc users can irc via sleep..
Jul 26 14:54:22 <blogic> ;)
Jul 26 14:54:41 <blogic> via or while ?
Jul 26 14:54:43 <WillieNL> nbd jow_laptop : this is wat rrdtool's makefile.ac and makefile.in produce -> http://pastebin.com/4WmYhYCv
Jul 26 14:54:49 <blaise> but, it was a question towards if he sleeps...
Jul 26 14:54:50 <blaise> lol
Jul 26 14:55:21 <WillieNL> this == rediculous ->> i486-openwrt-linux-gnu-ld  -L/usr/local/lib  RRDs.o -L../../src/.libs/ -Wl,--rpath -Wl,/usr/lib -lrrd  -o blib/arch/auto/RRDs/RRDs.so
Jul 26 14:55:32 <blogic> ah ok, the do you _ever_.. bit ewas missing
Jul 26 14:55:50 <WillieNL> looks like someone must rewrite those conftool scipts :(
Jul 26 14:56:15 <blaise> blogic: by the structure of the question, I believe it was implied..
Jul 26 14:56:17 <blaise> XD
Jul 26 14:56:24 <jow_laptop> WillieNL: no, someone has to nuke them from orbit, then kill its authors, and afterwards reimplement them sanely
Jul 26 14:57:02 <WillieNL> jow_laptop: if someone forked rrdtool 3 years ago there was no fire :)
Jul 26 14:57:23 <WillieNL> there is something like graphvis but dunno if that won't need rrdtool in the first place.
Jul 26 14:58:57 <jow_laptop> blaise: I linked a patch for the libotr thing
Jul 26 14:58:58 <WillieNL> jow_laptop: and wat about pull the bindings outside rrdtool and build that external as a new package ?
Jul 26 14:59:23 <jow_laptop> WillieNL: maybe possible
Jul 26 14:59:24 <WillieNL> no chance or word give it a try .
Jul 26 14:59:35 <WillieNL> jow_laptop: hmm ok sounds easyer to me..
Jul 26 14:59:41 <jow_laptop> perl is a nightmare to compile anyway
Jul 26 14:59:48 <jow_laptop> crosscompile rather
Jul 26 14:59:53 <WillieNL> eh compiled at once fine :D
Jul 26 14:59:59 <WillieNL> define some proper deps ;-)
Jul 26 15:00:05 <WillieNL> before install perl
Jul 26 15:00:07 <blaise> jow_laptop: awesome, I'm eager to see what the issue was.. :D
Jul 26 15:00:22 <jow_laptop> blaise: as expected, an autoconf fuckup
Jul 26 15:00:52 <jow_laptop> blaise: http://luci.subsignal.org/~jow/libotr.patch - the first part contains the patch
Jul 26 15:00:52 <WillieNL> jup automagic fuzz :p
Jul 26 15:01:31 <blaise> danke
Jul 26 15:02:24 <blaise> jow_laptop: I wasn't even aware there was a patch involved...
Jul 26 15:02:44 <jow_laptop> I found no way to fix it without patching
Jul 26 15:02:48 <blaise> ah
Jul 26 15:02:55 <jow_laptop> usually you can throw various magic variables and arguments at it
Jul 26 15:03:09 <jow_laptop> but not in this case
Jul 26 15:03:37 <blaise> so I need to remove my packages/libotr dir?
Jul 26 15:03:53 <jow_laptop> yes
Jul 26 15:04:16 <blaise> how will the patch work if there is no Makefile to patch?
Jul 26 15:04:18 <jow_laptop> rm -rf package/libotr; wget -qO- http://luci.subsignal.org/~jow/libotr.patch | patch -p0
Jul 26 15:04:38 <jow_laptop> make package/libotr/{clean,compile} V=99
Jul 26 15:06:50 <blaise> jow_laptop: you da man.
Jul 26 15:08:17 <blaise> jow_laptop: so, what is the possibility that this could be put into population?
Jul 26 15:08:26 <jow_laptop> don't see why not
Jul 26 15:08:42 <jow_laptop> but it makes no sense without a using program for it
Jul 26 15:09:33 <blaise> I've been working on porting irssi-otr plug-in as well
Jul 26 15:10:16 <blaise> which now that we have working libotr, there is a distinct possibility that my irssi-otr Makefile will work.. :D
Jul 26 15:11:37 <blaise> omg, autoconf/general.m4 is an angry beast, isn't it..
Jul 26 15:12:04 <jow_laptop> you have to nuke every .m4 file which contains a libtool copy
Jul 26 15:12:27 <jow_laptop> openwrt by default removes aclocal.m4 if it finds it (and PKG_FIXUP:=autoreconf is set)
Jul 26 15:12:33 <jow_laptop> but sometimes its elsewhere
Jul 26 15:12:48 <jow_laptop> e.g. in m4/program.m4 or sometimes even in acinclude.m4
Jul 26 15:13:47 <jow_laptop> you can do that by specifying  PKG_REMOVE_FILES:=foo/bar.m4 bla/blubb.m4
Jul 26 15:15:21 <blaise> hrmm, maybe I shouldn't have included iconv on the dependencies list
Jul 26 15:15:38 <jow_laptop> you shouldn't use that anyway
Jul 26 15:15:49 <jow_laptop> as openwrt has two versions of it which can be swapped in menuconfig
Jul 26 15:16:05 <jow_laptop> add "include $(INCLUDE_DIR)/nls.mk" after package.mk
Jul 26 15:16:17 <jow_laptop> and add $(ICONV_DEPENDS) to the DEPENDS:= line
Jul 26 15:20:02 <blaise> ah, aparently I didn't add iconv..
Jul 26 15:20:14 <blaise> but it needs python...
Jul 26 15:20:32 <blaise> so, I'm thinking python may have pulled iconv in..
Jul 26 15:20:43 <blaise> s/iconv/libiconv
Jul 26 15:29:07 <WillieNL> blaise: iconv and intl are mosthly not linked directly
Jul 26 15:29:23 <WillieNL> blaise: and specialy not added later in a working build ;-)
Jul 26 15:29:57 <WillieNL> also nls is disabled by default you need to enable that in the toolchain and target conf.
Jul 26 15:30:01 <WillieNL> else it fails.
Jul 26 15:30:49 <WillieNL> ahh sorry never mind
Jul 26 15:32:09 <WillieNL> blaise: in menuconfig if you goto iconv and select help you can view the "selected by" line.
Jul 26 15:46:14 <nbd> tripolar: hi
Jul 26 15:46:52 <tripolar> hi nbd, i wanted to ask for the patch but its wrong
Jul 26 15:47:00 <nbd> yeah
Jul 26 15:47:07 <nbd> NULL as a return value for u32 is not good
Jul 26 15:47:09 <tripolar> and i switched back to 2.6.39.3 as 3.0 has some performance gaps
Jul 26 15:47:09 <nbd> also the #else is wrong
Jul 26 15:47:30 <tripolar> the device becomes unrepsonsive when i copy files over nfs
Jul 26 15:47:33 <nbd> hm
Jul 26 15:47:41 <tripolar> maybe a sheduler problem
Jul 26 15:47:49 <nbd> hard to say
Jul 26 15:47:52 <nbd> does cpu usage go up?
Jul 26 15:47:58 <tripolar> the desktop also feels laggy compared to 2.6.39
Jul 26 15:48:01 <tripolar> yes
Jul 26 15:48:04 <tripolar> 100%
Jul 26 15:48:16 <tripolar> but with 2.6.39 everything is fine even with 100% load
Jul 26 15:48:16 <nbd> ok, maybe we need to so some profiling again
Jul 26 15:48:26 <tripolar> profiling?
Jul 26 15:48:39 <nbd> yeah, using oprofile you can figure out where the kernel spends most of its time
Jul 26 15:48:52 <tripolar> the performance also drops for nfs
Jul 26 15:49:12 <tripolar> from about 24mbit to 20mbit
Jul 26 15:49:29 <tripolar> and while copying a file i cant login into ssh
Jul 26 15:49:43 <nbd> copying it from the router or just passing traffic through the router?
Jul 26 15:49:49 <tripolar> from the router
Jul 26 15:49:52 <nbd> ok
Jul 26 15:49:56 <tripolar> with 2.6.39 it work fine
Jul 26 15:50:00 <tripolar> works
Jul 26 15:50:08 <tripolar> maybe 3.0.1 fixes this
Jul 26 15:50:25 <tripolar> nbd do you know where i can find the git repo for 3.0.1?
Jul 26 15:50:35 <tripolar> or is the repo released after the kernel is released?
Jul 26 15:51:47 <tripolar> or does grep push the changes right befor the release to http://git.kernel.org/?p=linux/kernel/git/stable/linux-3.0.y.git;a=summary
Jul 26 15:51:53 <tripolar> grep=greg
Jul 26 15:51:57 <nbd> not sure
Jul 26 15:52:54 <tripolar> even dnsmasq is unresponsive when i copy a file over nfs
Jul 26 15:57:28 <blaise> it's compiling sqlite, and stuff
Jul 26 15:57:37 <blaise> I can't imagine why..
Jul 26 15:57:38 <blaise> lol
Jul 26 15:58:31 <blaise> it just started on python-2.6.4
Jul 26 15:58:37 <blaise> finaly
Jul 26 15:59:42 <tripolar> blaise dont open a champagne befor you tried it on the router ;)
Jul 26 16:06:45 <tripolar> nbd another question - i wnt to by a new device for a prototype im developing : it should be possible to change transmission power, timing parameters of the MAC protocol (also including 802.11e/WMM) and it should support IEEE 802.11n - and of course i should be able to run openwrt on it
Jul 26 16:07:02 <tripolar> wnt=want by=buy
Jul 26 16:07:20 <tripolar> do you have some recommendations?
Jul 26 16:10:15 <nbd> tripolar: something with atheros 802.11n hw
Jul 26 16:10:28 <nbd> ar92xx or newer
Jul 26 16:10:43 <tripolar> okay - i will search for one
Jul 26 16:10:45 <tripolar> thanks
Jul 26 16:11:55 <tripolar> nbd and the n standard is completely supported by linux?
Jul 26 16:20:15 <blaise> nice
Jul 26 16:20:33 <nbd> tripolar: yes
Jul 26 16:20:35 <blaise> python didn't compile..
Jul 26 16:20:43 <blaise> I didn't honestly expect it to..
Jul 26 16:23:04 <blaise> although irssi-otr only requires python at build time.
Jul 26 16:23:32 <blaise> it doesn't need anything but libotr, glib, and irssi to run
Jul 26 16:24:45 <blaise> required to build irssi-otr is cmake (pkg-config), python, and wget..
Jul 26 16:25:10 <tripolar> nbd okay thanks for the info
Jul 26 16:26:55 <blaise> hrmmmm
Jul 26 16:28:44 <blaise> hrrrmmmmmmm....
Jul 26 16:31:00 <CIA-89> nbd * r27785 /branches/backfire/target/linux/ (10 files in 3 dirs): backport rtl8366 detection support from trunk
Jul 26 16:31:03 <nbd> git log
Jul 26 16:31:07 <nbd> oops
Jul 26 16:31:09 <nbd> -EWRONGWINDOW
Jul 26 16:44:12 <CIA-89> nbd * r27786 /trunk/target/linux/x86/alix2/target.mk:
Jul 26 16:44:12 <CIA-89> x86/alix2: fix typo in target makefile
Jul 26 16:44:12 <CIA-89> Cut and paste cruft from borrowing from Geos platform.
Jul 26 16:44:12 <CIA-89> Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
Jul 26 16:44:13 <CIA-89> nbd * r27787 /trunk/target/linux/x86/alix2/base-files/etc/config/network:
Jul 26 16:44:13 <CIA-89> x86/alix2: include WAN interface by default
Jul 26 16:44:13 <CIA-89> Include WAN interface by default, instead of putting both ports in the LAN bridge.
Jul 26 16:44:13 <CIA-89> Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
Jul 26 16:51:17 <philipp64|laptop> nbd: quick question... what would cause gpio_request() to fail? I'm trying out a modified experimental alix platform driver... and it's failing to acquire the S/W reset button.
Jul 26 16:51:49 <nbd> did you check if it's already requested somewhere else?
Jul 26 16:52:55 <larsc> check the return value it should tell you why it fails
Jul 26 16:58:03 <philipp64|laptop> ok... rebuilding...
Jul 26 17:02:02 <philipp64|laptop> if I don't change kernel version or gcc version, I don't need to rebuild the toolchain, right?
Jul 26 17:10:46 <Chocky> ok, kids
Jul 26 17:11:08 <Chocky> UI is complaining about libiwinfo not installed.  Fine, installed it, but no joy.  even rebooted.
Jul 26 17:21:10 <CIA-89> nbd * r27783 /trunk/package/mac80211/patches/300-pending_work.patch: mac80211: merge a few pending fixes for channel switch handling
Jul 26 17:29:22 <builder_> build #38 of octeon is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/38
Jul 26 17:29:23 <builder_> build #59 of atheros is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/59
Jul 26 17:29:25 <builder_> build #58 of cobalt is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/58
Jul 26 17:29:27 <builder_> build #57 of orion is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/57
Jul 26 17:40:55 <CIA-89> nbd * r27784 /branches/backfire/package/mac80211/patches/300-pending_work.patch: mac80211: merge a few pending fixes for channel switch handling (backport of r27783)
Jul 26 18:19:15 <tripolar> nbd what did you study or did you learn all this autodidactic?
Jul 26 18:19:51 <nbd> pretty much all autodidactic
Jul 26 18:26:00 <Chocky> well, I dunno the content of the conversation, but it's not like there are many options for going to school and learning openwrt, embedded Linux, kernel dev (there are some, I admit, but pretty superficial)
Jul 26 18:26:08 <RealOpty> Hauke, did you recover that board_data partition data on the netgear?
Jul 26 18:29:11 <Hauke> RealOpty: I haven't tried yet ;-)
Jul 26 18:31:16 * Chocky wiggles nbd
Jul 26 19:03:37 <builder_> build #57 of pxcab is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/57
Jul 26 19:03:38 <builder_> build #57 of ps3 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/ps3/builds/57
Jul 26 19:04:38 <jow_laptop> Chocky: lua -e 'require "iwinfo"'
Jul 26 19:04:42 <jow_laptop> look for errors
Jul 26 19:06:31 <Chocky> gotcha
Jul 26 19:06:46 <Chocky> and can someone put extern "C" wrappers in libudev.h, kthxbye
Jul 26 19:06:46 <jow_laptop> probably missing libnl-tiny
Jul 26 19:07:06 * Chocky spent 10 mins chasing linkage errors
Jul 26 19:07:08 <jow_laptop> can somebody kill C++ please? :P
Jul 26 19:07:14 <Chocky> aw
Jul 26 19:07:24 <Chocky> not really a fan, but I have to deal with a lot of C++ code
Jul 26 19:08:30 <Chocky> lua: error loading module 'iwinfo' from file '/usr/lib/lua/iwinfo.so':
Jul 26 19:08:30 <Chocky> 	/usr/lib/lua/iwinfo.so: undefined symbol: stat
Jul 26 19:08:30 <Chocky> stack traceback:
Jul 26 19:08:30 <Chocky> 	[C]: ?
Jul 26 19:08:30 <Chocky> 	[C]: in function 'require'
Jul 26 19:08:30 <Chocky> 	(command line):1: in main chunk
Jul 26 19:08:31 <Chocky> 	[C]: ?
Jul 26 19:08:38 <Chocky> fail
Jul 26 19:08:43 <Chocky> oh, wait a sec
Jul 26 19:10:57 <jow_laptop> stripped to death?
Jul 26 19:11:50 <Chocky> I have two routers ;-)
Jul 26 19:12:05 <Chocky> but same error
Jul 26 19:12:25 <Chocky> I don't think I turned on the crazy stripping options
Jul 26 19:12:38 <Chocky> I will include it in the build, and see what happens
Jul 26 19:12:47 <jow_laptop> never seen it missing stat
Jul 26 19:13:16 <Chocky> this is glibc, FWIW
Jul 26 19:13:29 <jow_laptop> ah, bloatlib
Jul 26 19:13:31 <jow_laptop> no idea
Jul 26 19:13:34 <Chocky> sigh
Jul 26 19:13:36 * Chocky slaps
Jul 26 19:13:40 <jow_laptop> maybe it has not stat...
Jul 26 19:13:48 <jow_laptop> because it maintainer hates it or so
Jul 26 19:14:01 <Chocky> possibly; ISTR, that stat is really stat64
Jul 26 19:14:12 <jow_laptop> or iwinfo needs some funny "I am gnu" defines
Jul 26 19:14:35 <Chocky> ok
Jul 26 19:14:46 <jow_laptop> or largefile source
Jul 26 19:14:54 * Chocky puts on autoslap
Jul 26 19:14:56 <Chocky> anyway
Jul 26 19:15:10 <Chocky> WiFi works in any case, but I assume that gives more info in UI
Jul 26 19:15:17 <jow_laptop> yeah
Jul 26 19:15:34 <Chocky> flashing takes ~5mins.  bah
Jul 26 19:15:48 <Chocky> 10.6MB
Jul 26 19:16:18 <Chocky> that's 30MB uncompressed.  we have a lotta crap
Jul 26 19:38:25 <Chocky> odd
Jul 26 19:40:20 * Chocky counter accuses openwrt devs of doing something uclibc specific
Jul 26 19:47:52 <Chocky> jow: any ideas?
Jul 26 19:49:04 <jow_laptop> don't see how calling stat() is uclibc specifc these days
Jul 26 19:49:32 <Chocky> I never said it was, I was just making a baseless accusation. just real tired of the pointless glibc bashing
Jul 26 19:49:51 <Chocky> I'm using glibc because when I stated, uclibc was clearly broken for what I was going, and glibc was not
Jul 26 19:50:03 <jow_laptop> then fix whatever breaks with glibc
Jul 26 19:50:07 <jow_laptop> I have zero interest in it
Jul 26 19:50:17 <Chocky> sigh
Jul 26 19:50:17 <jow_laptop> I don't have a test setup
Jul 26 19:50:22 <jow_laptop> I don't have an idea
Jul 26 19:50:22 <Chocky> I'm trying.  just asking for ideas
Jul 26 19:51:10 <jow_laptop> I tried iwinfo on my host system with x86/glibc and I tried it with uclibc on openwrt, it works in both cases
Jul 26 19:51:17 <Chocky> bully for you
Jul 26 19:51:44 <jow_laptop> so maybe see if stat() is protected with some ifdef stuff or whatever
Jul 26 19:51:47 <Chocky> apathy is allowed.  but can we refrain from bashing in future.  I'm aware of faults many many systems.
Jul 26 19:52:01 <Chocky> it seems that there's some __xstat stuff. due to 32 vs 64-bit stuff
Jul 26 19:52:08 <Chocky> which gets used with -O3
Jul 26 19:52:49 <Chocky> but even so. hm
Jul 26 20:02:38 <Chocky> ok.
Jul 26 20:02:57 <Chocky> compile with -O3 instead of -Os is a work around.  that suggest some kind of deeper issue, but I'll report a bug
Jul 26 20:03:23 * Chocky smooches jow
Jul 26 20:07:46 * Chocky boots wifi signal like crazy
Jul 26 20:14:39 <Chocky> https://dev.openwrt.org/ticket/9483
Jul 26 20:14:44 <Chocky> FWIW
Jul 26 20:19:35 <jow_laptop> bah apparently you need to link with libc
Jul 26 20:19:42 <jow_laptop> thats just stupid
Jul 26 20:44:48 <dirtyfreebooter> on latest trunk, dirclean'd, "make[3] -C feeds/packages/libs/avahi compile" shows up twice, one right after the other
Jul 26 20:45:34 <dirtyfreebooter> some strange dependency gone wrong? i wonder...
Jul 26 20:46:44 <dirtyfreebooter> the build still completed, and the install phase had it twice also.. "make[3] -C feeds/packages/libs/avahi install"
Jul 26 20:47:50 <dirtyfreebooter> ditto with ncurses, twice the compile output and twice in the install output
Jul 26 20:47:59 <dirtyfreebooter> no other packages in my build do that
Jul 26 21:00:06 <CIA-89> nbd * r27788 /trunk/package/mac80211/patches/ (6 files): ath9k: rework tx power handling - display the correct *current* tx power, and also fix tx power limits with ATH_USER_REGD set
Jul 26 21:01:31 <mhei> @jow: did you notice http://patchwork.ozo.com/patch/465 ? do you have still any objections?
Jul 26 21:05:37 <CIA-89> nbd * r27789 /branches/backfire/package/mac80211/patches/ (6 files): ath9k: rework tx power handling - display the correct *current* tx power, and also fix tx power limits with ATH_USER_REGD set (backport of r27788)
Jul 26 21:16:32 <Chocky> is it possible to turn off stuff in the UI easily.  there's waaaay too many options for our customers.
Jul 26 21:16:40 <Chocky> and some of the stuff it would be better for them to not poke at
Jul 26 21:18:16 <jow_laptop> mhei: its fine, just did not got around to integrate it yet
Jul 26 21:19:50 <mhei> jow: ok, thnx for your review
Jul 26 21:22:11 <jow_laptop> Chocky: not without touching code
Jul 26 21:23:09 <Chocky> that's fine, as long as it's relatively easy
Jul 26 21:23:13 <Chocky> I don't speak Lua, howeve
Jul 26 21:24:03 <jow_laptop> basically put a "--" (Lua comment) before the corresponding "node()" or "entry()" calls in /usr/lib/lua/luci/controller/*.lua
Jul 26 21:24:09 <Chocky> thanks
Jul 26 21:53:24 <CIA-89> jow * r27790 /packages/net/quagga/ (3 files in 3 dirs): [packages] quagga, quagga-old: use "cat" as default pager and "head -n" as pager for limited amount of lines (#9814)
Jul 26 21:54:45 <CIA-89> jow * r27791 /branches/packages_10.03.1/net/quagga/ (3 files in 3 dirs): [packages_10.03.1] merge r27790
Jul 26 21:59:39 <Chocky> jow: with the stripping options, etc.  how much do you think you'd save in space?
Jul 26 22:28:05 <CIA-89> jow * r27792 /trunk/package/firewall/ (Makefile files/lib/core_rule.sh): [package] firewall: prevent redundant rules if multiple ports and multiple icmp types are given in a rule block for both icmp and other protocols
Jul 26 22:28:39 <jow_laptop> Chocky: what stripping options?
Jul 26 22:29:02 <Chocky> in the advanced build, there's kernel stripping stuff, and also to strip unused symbols from libraries.
Jul 26 22:29:10 <Chocky> you can't add stuff later, but that's ok
Jul 26 22:29:17 <jow_laptop> ah, well maybe 10% orso
Jul 26 22:29:29 <Chocky> that's worthwhile
Jul 26 22:29:49 * Chocky tries
Jul 26 22:30:16 <Chocky> it won't be 10% of my image; I have a lot of non-binary stuff, but still
Jul 26 22:30:26 <Chocky> non-code that is
Jul 26 22:30:50 <WillieNL> jow_laptop: strip 10% ? hmm 30M tarred or 250M tarred data in packages folder..
Jul 26 22:31:06 <WillieNL> for x86 with libc stuff and nls
Jul 26 22:31:18 <WillieNL> using eglibc
Jul 26 22:31:56 <Chocky> I will have the answer in a second
Jul 26 22:33:09 <Chocky> hm
Jul 26 22:33:10 <WillieNL> jow_laptop: hmm how do you like to receive patches, via a ticket in the trac system?
Jul 26 22:33:18 <Chocky> perhaps I need to rebuild everything.  but the stripping should be after
Jul 26 22:33:25 <Chocky> it's the same size
Jul 26 22:33:37 <Chocky> maybe it's just padded to the same size
Jul 26 22:34:12 * Chocky strips opkg stuff too
Jul 26 22:35:09 <Chocky> hm, was anyway
Jul 26 22:35:34 * Chocky builds clean, and does something else for a bit
Jul 26 23:51:04 <CIA-89> nbd * r27793 /branches/backfire/toolchain/binutils/patches/2.19.1/310-backport_fPIE_mips_fix.patch: binutils: backport a fix for broken relocation entries on mips with -fPIE, fixes dbus-server crash with sstrip (backport of r25188, fixes #9810)
Jul 26 23:58:41 <CIA-89> jow * r27794 /branches/backfire/package/ (5 files in 3 dirs): [backfire] rc5->rc6
**** ENDING LOGGING AT Wed Jul 27 02:59:57 2011