**** BEGIN LOGGING AT Thu Dec 06 02:59:59 2012
Dec 06 12:30:02 <RealOpty> having trouble getting  kernel video modules to build. http://paste.ubuntu.com/1414564/ <- my config. using a fresh trunk.
Dec 06 13:50:08 <nbd> RealOpty: please use ./scripts/diffconfig.sh to trim your config for pasting
Dec 06 13:53:07 <RealOpty> nbd, I forgots about that script. here you go :) http://paste.ubuntu.com/1414731/
Dec 06 14:11:52 <OpenWrt-Commits> nbd r34531 trunk/package/mac80211/ (34 files in 2 dirs)
Dec 06 14:11:52 <OpenWrt-Commits> mac80211: clean up patches, replace patching config.mk with a generated sed expression
Dec 06 14:11:56 <OpenWrt-Commits> nbd r34532 trunk/target/linux/ generic/config-3.7 generic/config-3.6 * kernel: add missing config symbols
Dec 06 14:11:59 <OpenWrt-Commits> nbd r34533 trunk/package/kernel/modules/video.mk * kernel: fix video-core build on 3.6+
Dec 06 14:12:03 <nbd> RealOpty: fixed now
Dec 06 14:14:17 <OpenWrt-Commits> hcg r34534 packages/lang/perl/perlbase.mk * perl: Install /usr/bin/prove so test harness works
Dec 06 14:17:31 <RealOpty> nbd, appreciate it. :)
Dec 06 14:48:28 <WillieNL> i noticed glibc has been removed from trunk any reason why ?
Dec 06 14:49:11 <WillieNL> last i had tested was 2.14 noticed someone tryed/failed with 2.16
Dec 06 14:50:33 <[florian]> we decided there was no point in keeping glibc since eglibc is there
Dec 06 14:50:41 <WillieNL> argh again that :S
Dec 06 14:50:52 <WillieNL> <- head to desk
Dec 06 14:51:08 <WillieNL> where did i spend my time for to debug it till dependency trouble was fixed ?
Dec 06 14:51:22 <WillieNL> also it was inteded for builds on cf cards not nand
Dec 06 14:51:55 <WillieNL> since a couple apps like collectd and other graphing related stuff around had trouble with eglibc and uclibc
Dec 06 14:52:32 <[florian]> eglibc presents no runtime nor build time difference to glibc
Dec 06 14:52:43 <WillieNL> update time starts for me and notice i can not push dependencys to current versions but back to where i started must push it back into the svn code
Dec 06 14:53:12 <WillieNL> [florian]: don't get me wrong but i had my reasons togo glibc as faster less debug consuming solution
Dec 06 14:53:41 <[florian]> I don't get you wrong, we just cannot afford maintaining glibc considering that eglibc is a much saner alternative
Dec 06 14:53:54 <WillieNL> instaid of patch all userland applications better fix the lacking toolchain features then use arm optimized solutions for x86 builds
Dec 06 14:54:07 <WillieNL> well why was the ancient glibc theire so long ?
Dec 06 14:54:07 <[florian]> what are you talking about?
Dec 06 14:54:19 <[florian]> because no one cared about actually removing it
Dec 06 14:54:27 * WillieNL runs wrt on centrino dothan boxes
Dec 06 14:54:51 <WillieNL> i want max performance with less trouble seeking why do i miss threading support and other crap
Dec 06 14:55:27 <WillieNL> never asked the devs to fix my stuff but this is like been pushed out the trunk
Dec 06 14:55:39 <cyrusff> WillieNL: debian, ubuntu and probably others use eglibc as there main c-library (and dropped glibc some time ago)
Dec 06 14:56:18 <WillieNL> cyrusff: well possible i dunno about debian but i had my trouble wich the devs must remember and not just me wheire a few using it
Dec 06 14:56:42 <WillieNL> now i wanna pickup the task and give it an other push and notice it is removed
Dec 06 14:57:09 <WillieNL> also i do not expect the devs to debug it every month in a auto build
Dec 06 14:58:21 <WillieNL> while it was ok for that dunno about glibc 2.16 have they looked at dependency versions or blindly upgrade 2.14 to 2.16 and look new build wich mosth likely will fail during build or when try to run the produced binarys on missing librarys or features in some lib
Dec 06 14:59:10 <[florian]> sorry to say this, but I cannot make much sense of what you state here
Dec 06 14:59:44 <WillieNL> short story with glibc in svn wich still can do it my self i know resulted in using openwrt as non embedded i686/x64 distro on modern more powerfull machines with 533 or 800mhz fsb and pcie network adapters
Dec 06 15:00:49 <WillieNL> [florian]: as dev you must know just update glibc and nothing else in your toolchain will fail
Dec 06 15:00:56 <cyrusff> WillieNL: why don't you just use eglibc then. if the big non-embedded distros use it for x86 and amd64 PCs / Laptops / Servers why do you think it won't work for you?
Dec 06 15:01:27 <WillieNL> cyrusff: have said this a lot times even today eglibc did not render in less trouble on x86/i386 builds
Dec 06 15:01:29 <[florian]> there is pratically no difference between glibc and eglibc, besides the fact that eglibc is more configurable at build time
Dec 06 15:01:36 <[florian]> and that it does receive proper fixing for non x86 architectures
Dec 06 15:01:41 <WillieNL> like apache will be a horror
Dec 06 15:02:26 <WillieNL> [florian]: why was collectd then so much drama if even the upstream dev refuse to support the ancient version supported/used by wrt back then
Dec 06 15:02:42 <jow_laptop> drama?
Dec 06 15:02:57 <WillieNL> don't say that it worked even prebuild versions had the memleaks xD
Dec 06 15:03:14 <jow_laptop> never had memleaks on the standard uclibc version
Dec 06 15:03:14 <[florian]> is openwrt responsible for bad upstream software?
Dec 06 15:03:17 <WillieNL> and yes with glibc 2.14 ntp had some memleak i know
Dec 06 15:03:41 <[florian]> if you do care about the embedded case and fast aspects of your system
Dec 06 15:03:50 <[florian]> you really should not use any glibc
Dec 06 15:03:53 <[florian]> but rather uclibc or musl
Dec 06 15:03:57 <[florian]> which have a smaller footprint
Dec 06 15:04:11 <WillieNL> [florian]: well if can with threading support wich was a problem for eglibc and uclibc also not important for mosth embedded users
Dec 06 15:04:39 <[florian]> uclibc support nptl now, so it's really in par with glibc
Dec 06 15:04:43 <WillieNL> i have 4G sd wich copy to a 1G ramdisk
Dec 06 15:04:56 <[florian]> most uclibc developpers work on i386 and x86_64, so uclibc gets proper fixing on these architectures anyway
Dec 06 15:04:57 <WillieNL> size is not a topic
Dec 06 15:05:11 <[florian]> it's not just abous size, it's about the memory footprint and the speed of execution
Dec 06 15:05:20 <WillieNL> well ok
Dec 06 15:05:28 <[florian]> glibc has a whole lof of heuristic to allocate memory this or that way
Dec 06 15:05:40 <WillieNL> still then build some apps with threading suppor was impossible
Dec 06 15:05:43 <WillieNL> how is that today ?
Dec 06 15:05:48 <[florian]> everything builds and works fine
Dec 06 15:06:16 <WillieNL> heh well that was what people said back then before in did run into 12 weeks seeking true mailing lists and stuff
Dec 06 15:06:38 <WillieNL> why something simple as graph creation and stuff was not working well
Dec 06 15:06:50 <WillieNL> if that did work i may never did look at the threading part and stuff
Dec 06 15:07:04 <WillieNL> also i must go befor stores close
Dec 06 15:07:07 <WillieNL> bbl
Dec 06 15:18:34 <dwmw2_gone> /home/dwmw2/git/openwrt/build_dir/target-i386_uClibc-0.9.33.2/linux-x86_geos/linux-3.6.8/drivers/usb/host/pci-quirks.c:471: undefined reference to `uhci_check_and_reset_hc'
Dec 06 15:35:08 <dwmw2_gone> nbd: that's yours, I think?
Dec 06 15:36:05 <dwmw2_gone> perhaps you want quirk_usb_handoff_uhci() to be ifdeffed out too?
Dec 06 15:48:25 <OpenWrt-Commits> florian r34535 packages/net/radsecproxy/Makefile * radsecproxy: update to 1.6.2
Dec 06 15:55:32 <dwmw2_gone> [florian]?
Dec 06 15:55:43 <dwmw2_gone> your feedback on the nut package (in August)
Dec 06 15:56:00 <blogic> dwmw2_gone: yes, [florian] is nuts
Dec 06 15:56:03 <blogic> .-)
Dec 06 15:56:10 <dwmw2_gone> - make the driver macro self-contained, such that it also does the call to
Dec 06 15:56:10 <dwmw2_gone> BuildPackage
Dec 06 15:56:10 <dwmw2_gone> - rename this macro to BuildDriver for instance and get rid of the foreach
Dec 06 15:56:10 <dwmw2_gone> since it can be error prone with respect to escaping make variables
Dec 06 15:56:22 <dwmw2_gone> I'm trying this, but it doesn't work and my brain is hurting when I try to debug
Dec 06 15:56:59 <nbd> dwmw2_gone: i'll take care of it
Dec 06 15:57:02 <nbd> thanks
Dec 06 15:57:14 <dwmw2_gone> I have updated to build 2.6.5, and added the openups driver
Dec 06 15:57:34 <dwmw2_gone> nbd: you're talking about the uhci build failure, right?
Dec 06 15:57:41 <nbd> yes
Dec 06 16:00:37 <heilterminalcons> How do you think about having opendpi in OpenWrt?
Dec 06 16:02:34 <dwmw2_gone> Makefile:173: *** Package/nut-driver-tripplite_usb is missing the TITLE field.  Stop.
Dec 06 16:11:17 <[florian]> dwmw2_gone: this commonly happens when you make a typo in the make macro you invoke (happened to me)
Dec 06 16:11:39 <[florian]> and also for some reason I think you need to split the actually macro creating the define/Package/foo
Dec 06 16:11:45 <[florian]> from the one doing Package/foo/install
Dec 06 16:12:16 <[florian]> dwmw2_gone: coreutils and linux-atm should be good examples
Dec 06 16:16:32 <dwmw2_gone> thanks
Dec 06 16:16:55 <dwmw2_gone> if you have to split the macro and invoke it twice, does that mean you need a list of the subpackages in a variable, and to use foreach?
Dec 06 16:17:03 * dwmw2_gone should read the examples before asking questions
Dec 06 16:17:07 * dwmw2_gone goes to do so
Dec 06 16:17:54 <[florian]> that is correct, yes
Dec 06 16:18:08 <[florian]> also makes the addition/deletion a couple of liners
Dec 06 16:18:16 <dwmw2_gone> I thought the use of foreach was one of the things you objected to?
Dec 06 16:18:53 <[florian]> let me see my own answer again
Dec 06 16:18:56 <dwmw2_gone> I was wanting to define a list of the drivers anyway. Makes it easier to keep in sync with upstream if you can lift the actual text from its Makefile.am
Dec 06 16:19:07 <dwmw2_gone> Subject: 	Re: [OpenWrt-Devel] [PATCH] New package NUT 2.6.3 (RESENT)
Dec 06 16:19:07 <dwmw2_gone> Date: 	Mon, 13 Aug 2012 15:06:26 +0200 (13/08/12 14:06:26)
Dec 06 16:20:27 <[florian]> actually using foreach like he did was just fine
Dec 06 16:20:34 <dwmw2_gone> ok
Dec 06 16:20:39 <[florian]> no idea why I wrote that
Dec 06 16:20:49 <dwmw2_gone> but I think I'm going to start with a list of drivers anyway.
Dec 06 16:21:01 <dwmw2_gone> and make it look up the description
Dec 06 16:21:25 <dwmw2_gone> define Driver/description/tripplite_usb
Dec 06 16:21:25 <dwmw2_gone> 	Driver for older Tripp Lite USB UPSes (not PDC HID)
Dec 06 16:21:25 <dwmw2_gone> endef
Dec 06 16:21:59 <[florian]> you can use some nice tricks in your global variable containing the list drivers, like:
Dec 06 16:22:08 <[florian]> DRIVERS="foo:This is description for foo"
Dec 06 16:22:14 <dwmw2_gone> ah but that defeats the object
Dec 06 16:22:32 <dwmw2_gone> I want to *copy* the text from drivers/Makefile.am in the nut package itself
Dec 06 16:22:42 <dwmw2_gone> SERIAL_DRIVERLIST = bcmxcp belkin belkinunv bestfcom    \
Dec 06 16:22:42 <dwmw2_gone>  bestfortress bestuferrups bestups dummy-ups etapro everups     \
Dec 06 16:22:49 <dwmw2_gone> nice and easy to stay in sync when they add new drivers then
Dec 06 16:23:06 <[florian]> right, but that requires you having access to this list before hand
Dec 06 16:23:19 <dwmw2_gone> manual process when updating to a new version
Dec 06 16:23:22 <[florian]> if you want this fully auto-generated of course
Dec 06 16:23:26 <[florian]> ok, then this is just fine
Dec 06 16:23:32 <dwmw2_gone> still easier than staring at what's in the package makefile right now, and comparing
Dec 06 16:23:57 <dwmw2_gone> I'll probably even use $(filter-out ) to disable some, if I have good reason to do so
Dec 06 18:18:03 <OpenWrt-Commits> jow r34536 trunk/package/network/ services/ppp/files/ppp.sh services/ppp/Makefile * ppp: unconditionally start pppd with nodefaultroute and usepeerdns (#12580)
Dec 06 18:18:21 <OpenWrt-Commits> jow r34537 packages/net/xl2tpd/files/l2tp.sh
Dec 06 18:18:22 <OpenWrt-Commits> xl2tpd: unconditionally configure daemon with nodefaultroute and usepeerdns (#12580)
Dec 06 18:22:47 <OpenWrt-Commits> jow r34538 branches/attitude_adjustment/ package/ppp/Makefile package/ppp/files/ppp.sh * AA: ppp: backport r34536
Dec 06 18:23:33 <OpenWrt-Commits> jow r34539 branches/packages_12.09/net/xl2tpd/files/l2tp.sh * AA: xl2tpd: backport r34537
Dec 06 20:53:10 <OpenWrt-Commits> cyrus r34540 trunk/package/network/ipv6/6distributed/Makefile * 6distributed: Add compatibility with 6rd and 6to4
Dec 06 20:53:16 <OpenWrt-Commits> cyrus r34541 trunk/package/network/ipv6/6relayd/Makefile * 6relayd: More RFC-compatibility and size reduction
Dec 06 20:53:21 <OpenWrt-Commits> cyrus r34542 trunk/package/network/ ipv6/ipv6-support/Makefile ipv6/ipv6-support/files/network6.config ipv6/ipv6-support/files/support.sh ipv6/ipv6-support/files/dhcpv6.sh
Dec 06 20:53:21 <OpenWrt-Commits> ipv6-support: next iteration
Dec 06 20:53:21 <OpenWrt-Commits> * Introduce mode 6rd
Dec 06 20:53:21 <OpenWrt-Commits> * Introduce mode 6to4
Dec 06 20:53:22 <OpenWrt-Commits> * Fix announcement of DNS-Servers received from DHCPv6
Dec 06 20:53:25 <OpenWrt-Commits> * Fix some corner-cases in relaying behaviour
Dec 06 21:32:43 <tonikasch> bye
Dec 06 22:39:49 <OpenWrt-Commits> florian r34543 trunk/target/linux/ adm8668/files/arch/mips/adm8668/setup.c adm8668/files/arch/mips/adm8668/platform.c adm8668/files/arch/mips/adm8668/Makefile * [adm8668] move setup-related functions to their own file
Dec 06 22:39:58 <OpenWrt-Commits> florian r34544 trunk/target/linux/ adm8668/files/arch/mips/adm8668/platform.c adm8668/files/arch/mips/adm8668/time.c adm8668/files/arch/mips/adm8668/Makefile * [adm8668] move time related functions to their own file
Dec 06 22:40:03 <OpenWrt-Commits> florian r34545 trunk/target/linux/adm8668/files/arch/mips/adm8668/platform.c * [adm8668] use platform_add_devices
Dec 06 22:40:07 <OpenWrt-Commits> florian r34546 trunk/target/linux/adm8668/files/arch/mips/adm8668/irq.c * [adm8668] simplify IRQ controller code
Dec 06 22:40:14 <OpenWrt-Commits> florian r34547 trunk/target/linux/ (5 files in 3 dirs) * [adm8668] implement early_printk without an early_console
Dec 06 22:40:23 <OpenWrt-Commits> florian r34548 trunk/target/linux/ (14 files in 6 dirs) * [adm8668] get rid of the specific UART driver and use AMBA PL010
Dec 06 22:40:28 <OpenWrt-Commits> florian r34549 trunk/target/linux/ adm8668/files/arch/mips/adm8668/early_printk.c adm8668/files/arch/mips/include/asm/mach-adm8668/adm8668.h * [adm8668] get rid of the UART defines
Dec 06 22:40:33 <OpenWrt-Commits> florian r34550 trunk/target/linux/ adm8668/files/arch/mips/adm8668/time.c adm8668/files/arch/mips/adm8668/clock.c adm8668/files/arch/mips/adm8668/Makefile adm8668/files/arch/mips/include/asm/mach-adm8668/adm8668.h * [adm8668] provide a system clock to be used by the MIPS timer
Dec 06 22:40:37 <OpenWrt-Commits> florian r34551 trunk/target/linux/adm8668/files/arch/mips/adm8668/setup.c * [adm8668] remove frequency from /proc/cpuinfo
Dec 06 22:40:42 <OpenWrt-Commits> florian r34552 trunk/target/linux/ adm8668/files/arch/mips/adm8668/irq.c adm8668/files/arch/mips/include/asm/mach-adm8668/irq.h adm8668/files/arch/mips/include/asm/mach-adm8668/adm8668.h * [adm8668] cleanup and make interrupt code more robust
Dec 06 22:40:52 <OpenWrt-Commits> florian r34553 trunk/target/linux/ (7 files in 5 dirs) * [adm8668] cleanup PCI controller code
Dec 06 22:40:57 <OpenWrt-Commits> florian r34554 trunk/target/linux/ adm8668/files/arch/mips/adm8668/platform.c adm8668/patches-3.3/003-adm8668_nor_map.patch adm8668/files/drivers/mtd/maps/adm8668.c adm8668/config-3.3 * [adm8668] replace the custom mtd driver with a partition parser
Dec 06 22:41:02 <OpenWrt-Commits> florian r34555 trunk/target/linux/ generic/config-3.6 generic/config-3.3 * [kernel] add CONFIG_GPIO_PL061 config symbol
Dec 06 22:41:06 <OpenWrt-Commits> florian r34556 trunk/target/linux/ adm8668/files/arch/mips/adm8668/clock.c adm8668/files/arch/mips/include/asm/mach-adm8668/adm8668.h * [adm8668] move clock frequencies into clock driver
Dec 06 22:41:11 <OpenWrt-Commits> florian r34557 trunk/target/linux/ adm8668/patches-3.3/005-tulip_platform.patch adm8668/patches-3.3/004-tulip_pci_split.patch adm8668/config-3.3 adm8668/files/arch/mips/adm8668/Makefile * [adm8668] patch the tulip driver to accept a platform variant
Dec 06 22:41:16 <OpenWrt-Commits> florian r34558 trunk/target/linux/adm8668/files/arch/mips/adm8668/platform.c * [adm8668] use the tulip platform driver
Dec 06 22:41:21 <OpenWrt-Commits> florian r34559 trunk/target/linux/ adm8668/files/arch/mips/adm8668/net_intr.c adm8668/files/arch/mips/adm8668/net_core.c adm8668/files/arch/mips/adm8668/net.h * [adm8668] remove the "old" tulip forked driver
Dec 06 22:41:25 <OpenWrt-Commits> florian r34560 trunk/package/mac80211/Makefile * mac80211: add support for adm8211 wireless driver
Dec 06 22:41:30 <OpenWrt-Commits> florian r34561 trunk/target/linux/ adm8668/files/arch/mips/adm8668/platform.c adm8668/files/arch/mips/include/asm/mach-adm8668/adm8668.h * [adm8668] add support for on-board EHCI controller
Dec 06 22:41:34 <OpenWrt-Commits> florian r34562 trunk/target/linux/ adm8668/Makefile adm8668/config-3.3 * [adm8668] enable and advertise USB support
Dec 06 22:41:39 <OpenWrt-Commits> florian r34563 trunk/target/linux/adm8668/files/arch/mips/adm8668/setup.c * [adm8668] do not reset WAN and LAN blocks
Dec 06 22:41:45 <OpenWrt-Commits> florian r34564 trunk/target/linux/ adm8668/patches-3.3/001-adm8668_arch.patch adm8668/files/arch/mips/adm8668/gpio.c adm8668/files/arch/mips/include/asm/mach-adm8668/gpio.h adm8668/config-3.3 * [adm8668] add a gpiolib driver
Dec 06 22:41:50 <OpenWrt-Commits> florian r34565 trunk/target/linux/ adm8668/files/arch/mips/adm8668/proc.c adm8668/files/arch/mips/adm8668/Makefile adm8668/files/arch/mips/include/asm/mach-adm8668/adm8668.h * [adm8668] remove the now deprecated /proc/adm8668 interface
Dec 06 22:41:55 <OpenWrt-Commits> florian r34566 trunk/target/linux/ adm8668/files/arch/mips/adm8668/irq.c adm8668/files/arch/mips/adm8668/platform.c adm8668/files/arch/mips/include/asm/mach-adm8668/adm8668.h * [adm8668] rename interrupt number constants
**** ENDING LOGGING AT Fri Dec 07 02:59:59 2012