**** BEGIN LOGGING AT Mon Jul 11 02:59:56 2011
Jul 11 04:26:27 <dirtyfreebooter> anybody around? current trunk for ar71xx, the final gcc hits a gcc_assert() on even a stub program like main(){ return 0;}
Jul 11 04:28:14 <dirtyfreebooter> i've tried that simple no-op program and i get an internal compiler error (from the gcc_assert()): http://pastebin.com/SaN1HfPP
Jul 11 07:32:13 <loswillios> jow_laptop, nbd: there's still no way to make packages depend on uClibc-0.9.32 or is there? (pulseaudio, backfire compatibility)
Jul 11 07:32:48 <sekon_> Hello i a trying compile backfire
Jul 11 07:33:01 <sekon_> lzma kernel image seems to be too big
Jul 11 07:33:10 <sekon_> any tips on reducing its size
Jul 11 07:39:02 <blogic> which target ?
Jul 11 07:58:25 <sekon_> AR2315
Jul 11 08:00:51 <KanjiMonster> loswillios: should be possible with @CONFIG_<whatever_selects_uclibc_32> iirc
Jul 11 08:09:12 <loswillios> KanjiMonster: there used to be something like @CONFIG_UCLIBC_0_9_32 but it's gone. I think nbd wanted to look into it
Jul 11 08:11:28 <KanjiMonster> loswillios: hm... as a workaround you could depend on @!CONFIG_UCLIBC_0_9_30&&!CONFIG_UCLIBC_0_9_31 or similar; should have the same effect
Jul 11 08:51:53 <loswillios> KanjiMonster: nope, that doesn't work anymore
Jul 11 08:53:19 <_trine> can someone explain this to me
Jul 11 08:53:21 <_trine> http://dpaste.com/566289/
Jul 11 08:53:34 <_trine> I was just reading the wiki
Jul 11 08:53:47 <_trine> and decided to try the crpto command
Jul 11 08:54:06 <_trine> but I backed out when I didnt understand what it would do
Jul 11 08:54:30 <KanjiMonster> loswillios: dud you check that these are actual config symbols? I didn't; the appropriate ones are UCLIBC_VERSION_0_9_30_1 to UCLIBC_VERSION_0_9_30_3 ;)
Jul 11 08:57:33 <KanjiMonster> loswillios: but I think the menuconfig has some problems with these types of dependencies; I once tried to make two packages exclude each other and it didn't work reliably
Jul 11 09:01:47 <loswillios> KanjiMonster: ah, right. The CONFIG_ vanished. I'll commit this then: http://pastebin.com/ttm1diir
Jul 11 09:05:18 <KanjiMonster> loswillios: ah, I forgot that you don't add the CONFIG_ part - I can't test this, don't have a backfire checked out; but at a glance it looks like it should work ;)
Jul 11 09:06:09 <loswillios> yeah it does, thanks
Jul 11 09:08:18 <CIA-89> loswillios * r27589 /packages/sound/pulseaudio/Makefile: [packages] pulseaudio: not supported in backfire release, requires newer uClibc
Jul 11 10:23:37 <jow_laptop> loswillios: for what does it need .32 exactly?
Jul 11 10:23:46 <jow_laptop> whats the issue with .30 ?
Jul 11 10:54:02 <loswillios> jow_laptop: pulseaudio needs NPTL
Jul 11 11:09:10 <jow_laptop> ok
Jul 11 11:09:12 <KanjiMonster> loswillios: ah, then depending on .32 would have been wrong anyway (since iirc (e)glibc also has NPTL)
Jul 11 11:12:05 <loswillios> true
Jul 11 11:13:03 <jow_laptop> don't worry too much about it
Jul 11 11:13:15 <jow_laptop> I'll split off the pacvkages branch for backfire anyway
Jul 11 11:13:21 <loswillios> there was a time when I depended on the newer stuff (e/glibc,uclibc-git) but now it makes sense to just list the old-stuff for some time. progress is great :)
Jul 11 11:13:28 <jow_laptop> we can just mark PA broken there or remove it completely
Jul 11 11:13:51 <loswillios> oh, packages will be branched?
Jul 11 11:13:54 <jow_laptop> yes
Jul 11 11:14:00 <jow_laptop> along with rc5
Jul 11 11:14:09 <KanjiMonster> the clean, overengineered solution would be to add a new LIBC_HAS_NPTL bool, let uclibc .32 and (e)glibc select it, and then just depend on that ;)
Jul 11 11:15:11 <loswillios> fortunately, every libc in trunk does have NPTL :)
Jul 11 11:15:40 <KanjiMonster> yeah
Jul 11 11:18:08 <loswillios> jow_laptop: btw, did you succeed with a localized openwrt? I remember reading about you trying an utf-8 build or somethign
Jul 11 11:30:43 <jow_laptop> loswillios: yeah, it boiled down to then eed to build uclibc with locale support
Jul 11 11:30:56 <jow_laptop> without that the multibyte functions are no-ops
Jul 11 11:31:13 <jow_laptop> and iconv which utilizes them fails
Jul 11 11:42:08 <loswillios> ok
Jul 11 11:42:23 <Kaloz> KanjiMonster: nah, we should do a hidden feature for arches with nptl.. or maybe the oposite, we should check how many has it
Jul 11 11:45:05 <loswillios> jow_laptop: that means that stuff like bash/weechat can depend on libnursesw and have utf8 support?
Jul 11 11:45:49 <jow_laptop> no
Jul 11 11:45:56 <jow_laptop> as we still have no locale support in uclibc
Jul 11 11:46:26 <jow_laptop> and the utf8 awareness in uclibc is detemrined by the presence of at least one utf-8 on the hostsystem (!) during build
Jul 11 11:46:28 <loswillios> even when compiled with UCLIBC_HAS_LOCALE?
Jul 11 11:46:33 <jow_laptop> so its not really determinisitc either
Jul 11 11:46:45 <jow_laptop> *at least one utf-8 locale
Jul 11 11:47:19 <loswillios> ah, hm
Jul 11 11:47:21 <jow_laptop> you need  UCLIBC_HAS_LOCALE and at least one *.UTF-8 locale on your buildbox
Jul 11 11:47:25 <jow_laptop> e.g. de_DE.UTF-8
Jul 11 11:49:28 <loswillios> I'll try it, thanks
Jul 11 11:50:23 <KanjiMonster> Kaloz: afaik in trunk all of them have it (unless uclibc doesn't support if for some targets or it's disabled for them), so it would be always true
Jul 11 11:51:55 <Kaloz> KanjiMonster: no
Jul 11 11:52:06 <Kaloz> KanjiMonster: not all arch has nptl support
Jul 11 11:52:12 <Kaloz> s/has/have/
Jul 11 11:52:46 <loswillios> jow_laptop: is there something about this on the wiki yet?
Jul 11 11:53:42 <KanjiMonster> Kaloz: okay, then it does make sense to have it
Jul 11 11:57:42 <jow_laptop> loswillios: no
Jul 11 13:36:48 <jow_laptop> is there any freeswitch maintainer atm ?
Jul 11 13:38:11 <jow_laptop> it currently faiils due to some stupid -Werror in some zeromq plugin
Jul 11 13:38:45 <jow_laptop> I would like to push backfire and do not feel to stab around in the shitpile the FS buildsystem is
Jul 11 13:39:13 <jow_laptop> so if it does not get fixed we'll have to release it without
Jul 11 14:12:59 <mazilo> jow_laptop: I will go to fix it.
Jul 11 14:26:17 <jow_laptop> mazilo: thank you. see http://openwrt.wwsnet.net/ar7/logs/package/feeds/packages/freeswitch/compile.txt
Jul 11 14:26:26 <jow_laptop> or alternatively any other target
Jul 11 14:27:21 <mazilo> jow_laptop: Right now, I am trying to compile it locally and see if zmq will crash.
Jul 11 14:27:47 <jow_laptop> hmm
Jul 11 14:27:55 <jow_laptop> actually it looks like my warning fix did help
Jul 11 14:28:09 <jow_laptop> it craps out due to a missing posix_fadvice now
Jul 11 14:29:01 <mazilo> jow_laptop: I have some more new modules to add to freeswitch and hopefully will update soon within a couple of days once I iron out some quirks.
Jul 11 14:29:24 <jow_laptop> http://pastebin.com/4BDXagCc
Jul 11 14:29:38 <jow_laptop> its important that you compile backfire
Jul 11 14:29:48 <jow_laptop> FS on trunk probably works, didn't check
Jul 11 14:30:33 <mazilo> jow_laptop: ATM, I can only work on trunk due limited HDD space, unfortunately. :(
Jul 11 14:31:04 <mazilo> jow_laptop: I will update the Makefile with the libzmq_cv_cxx_werror_flag="" \
Jul 11 14:31:32 <jow_laptop> mazilo: well trunk does not help me
Jul 11 14:31:42 <jow_laptop> I try to build backfire, not trunk
Jul 11 14:31:49 <jow_laptop> its a huge difference
Jul 11 14:32:07 <mazilo> jow_laptop: IC.
Jul 11 14:32:55 <mazilo> jow_laptop: Can you please commit the Makefile with your working copy?
Jul 11 14:36:22 <jow_laptop> will do later
Jul 11 14:47:52 <loswillios> KanjiMonster: you added 3.0 support right? I'm having problems with cryptsetup, it fails with "Could not determine kernel version used", apparently because of "check_kernel" in build_dir/target-arm_v5te_uClibc-0.9.32_eabi/cryptsetup-1.3.1/lib/libdevmapper.c:178
Jul 11 14:53:01 <blogic> loswillios: what does that part of the code say ?
Jul 11 14:53:34 <blogic> http://code.google.com/p/cryptsetup/source/browse/trunk/lib/libdevmapper.c#174
Jul 11 14:54:28 <blogic> is the node created in /dev ?
Jul 11 14:55:52 <loswillios> blogic: no, it doesn't get this far. this is the output: http://pastebin.com/JYLa0wam~
Jul 11 14:56:02 <loswillios> args, http://pastebin.com/JYLa0wam
Jul 11 14:56:29 <loswillios> yes, dm_mod is loaded
Jul 11 14:57:24 <loswillios> I'm not sure where the 'Could not determine kernel version used.'-message is coming from. At least I can't find it in cryptsetup
Jul 11 14:58:36 <jow_laptop> probably from insmod
Jul 11 14:58:41 <jow_laptop> or mdadm ?
Jul 11 14:58:48 <jow_laptop> or some other realted userspace util
Jul 11 15:01:27 <loswillios> seems to be libdevmapper https://bugzilla.redhat.com/show_bug.cgi?id=710646#c5
Jul 11 15:17:28 <CIA-89> loswillios * r27590 /packages/utils/lvm2/Makefile: [packages] lvm2: update to 2.02.86, fixes linux-3.0 compatibility
Jul 11 15:28:10 <jow_laptop> grmbl
Jul 11 15:28:16 * jow_laptop restarts the rc5 builds
Jul 11 15:30:37 <loswillios> oops
Jul 11 15:31:58 <jow_laptop> :P
Jul 11 16:57:01 <CIA-89> kaloz * r27591 /trunk/package/kernel/modules/crypto.mk: fixup orion compile
Jul 11 17:15:37 <mazilo> jow_laptop: Now, the freeswitch compilation has started to get all kind of wierd errors as shown here (http://pastebin.com/xPqm0d6q) starting on line #6 where it refereced to files on a host '/usr/include/" directory
Jul 11 18:38:12 <luka12345|wiik> blogic: ping
Jul 11 18:47:34 <blogic> luka12345|wiik: hi
Jul 11 18:48:09 <luka12345|wiik> i just wanted to tell you that adm6996i and adm6996m are the same
Jul 11 18:48:25 <luka12345|wiik> at least regarding registers
Jul 11 18:48:59 <luka12345|wiik> only thing is that on my board with adm6996m port numbering start with 0 and end with 3
Jul 11 18:49:20 <luka12345|wiik> and on siemens port numbering starts with 1 and ends with 4
Jul 11 18:49:32 <luka12345|wiik> and switch is working
Jul 11 18:49:59 <luka12345|wiik> but im going to look more into the source and maybe add new functionality
Jul 11 18:50:32 <luka12345|wiik> i wanted to ask you do you want me to send you up to date patch to be applied to sx763
Jul 11 18:50:39 <luka12345|wiik> regarding mac and stuff...
Jul 11 18:53:38 <blogic> sure
Jul 11 18:53:45 <blogic> please do will make it easier ;)
Jul 11 18:53:55 <blogic> i am running low on time, vacation starts in 8 days ;)
Jul 11 18:53:59 <luka12345|wiik> np, i'll do it tonight
Jul 11 18:54:23 <luka12345|wiik> and also can you tell me why in dmesg i get this:
Jul 11 18:54:24 <luka12345|wiik> WARNING: FW version 12.1.0 too old.  Minimum required FW version is 12.4.0
Jul 11 18:54:54 <blogic> its the voice firmware
Jul 11 18:55:01 <blogic> drivers are newer than firmware
Jul 11 18:55:07 <blogic> i iwll try to get an update
Jul 11 18:55:58 <luka12345|wiik> great :)
Jul 11 18:56:28 <luka12345|wiik> i'll send my patch to the list when it's done...
Jul 11 18:59:54 <luka12345|wiik> blogic: one more thing, is the uboot in the current trunk ready for flashing? except for the lost packages issue...
Jul 11 19:00:34 <blogic> no
Jul 11 19:00:41 <blogic> the lost package leads to not working uboot
Jul 11 19:01:17 <blogic> as in ethernet is not working properly i was unable to transfer an image
Jul 11 19:01:41 <luka12345|wiik> can image be sent via serial?
Jul 11 19:02:48 <blogic> hmm
Jul 11 19:02:55 <blogic> you need to enable the command for it
Jul 11 19:03:42 <luka12345|wiik> ok...
Jul 11 19:04:02 <blogic> package/uboot-lantiq/files/board/infineon/easy50712/config.mk:CONFIG_BOOTSTRAP_TEXT_BASE = 0xb0000000
Jul 11 19:04:13 <blogic> if you modify that to say 0xb00a0000
Jul 11 19:04:34 <blogic> you can use the old uboot to flash the new to 0xb00a0000 and then do go 0xb00a0000 to test it
Jul 11 19:04:51 <blogic> and then play with it until serial load works and then revert the 0xb00a0000 and play with it
Jul 11 19:05:28 <luka12345|wiik> i thought you said that chainloading this uboot wont work
Jul 11 19:06:38 <blogic> it will
Jul 11 19:06:46 <blogic> you need to chnage that define for it
Jul 11 19:07:05 <luka12345|wiik> ok, thx
Jul 11 19:07:19 <blogic> also
Jul 11 19:07:45 <blogic> to start httpd you need to use the button on the pcb next to the switch and not the one next to the usb
Jul 11 19:07:51 <blogic> i only found 1 button to use ;)
Jul 11 19:08:43 <luka12345|wiik> that is more than fine with me :)
Jul 11 19:09:18 * luka12345|wiik  is going to do some patching :)
Jul 11 19:13:20 <blogic> hehe
Jul 11 19:13:39 <blogic> and i found the gpio for the switch... let me find the piece of paper where i wrote it down
**** ENDING LOGGING AT Mon Jul 11 20:00:49 2011

**** BEGIN LOGGING AT Mon Jul 11 20:02:14 2011
Jul 11 20:21:26 <Chocky> is there a RTC in the Buffalo G300NH?
Jul 11 20:39:50 <CIA-89> kaloz * r27592 /trunk/target/linux/orion/ (Makefile files/arch/arm/mach-orion5x/dt2-setup.c): [orion]: upgrade to 3.0-rc6 and optimize for armv5te again
Jul 11 20:39:51 <Chocky> no one knows!
Jul 11 20:57:22 <jow_laptop> Chocky: no
Jul 11 20:58:07 <Chocky> aw.
Jul 11 20:58:38 <Chocky> so, what is better. rdate or ntpclient.  and when does the latter run?
Jul 11 21:00:05 <jow_laptop> when wan is brought up
Jul 11 21:00:25 <Chocky> what if I want it on the LAN?
Jul 11 21:03:02 <jow_laptop> then you need to edit /etc/config/ntpclient and change wan to lan
Jul 11 21:03:20 <Chocky> oh yes
Jul 11 21:03:24 <Chocky> :o
Jul 11 21:19:02 <luka12345|wiik> blogic: ping
Jul 11 21:23:38 <CIA-89> jow * r27593 /packages_10.03.1/: [packages] branch off packages for 10.03.1
Jul 11 21:29:36 <CIA-89> jow * r27594 / (branches/packages_10.03.1/ packages_10.03.1/): [packages] move 10.03.1 branch to the right directory
Jul 11 21:31:57 <CIA-89> jow * r27595 /branches/backfire/feeds.conf.default: [backfire] switch packages feed to the 10.03.1 branch
Jul 11 22:07:06 <Chocky> so, why doesn't setting timezone in /etc/config/system have any obvious effect?
Jul 11 22:09:49 <jow_laptop> because you didn't reboot?
Jul 11 22:10:04 <jow_laptop> or you didn't specify a valid tz string
Jul 11 22:46:33 <Chocky> I rebooted, for sure.
Jul 11 22:46:42 <Chocky> and I'm pretty sure it's valid, but I can triple check :p
Jul 11 22:47:10 <Chocky> in fact, the value was put into the config before I even booted the first time
Jul 11 22:47:29 <Chocky>   option 'timezone' 'America/Los Angeles'
Jul 11 22:49:22 <Chocky> jow_laptop: ping
Jul 11 22:53:52 <jow_laptop> it has nss
Jul 11 22:53:56 <jow_laptop> it has no effect
Jul 11 22:54:08 <jow_laptop> the timezone variable is just used as hint for the gui
Jul 11 22:54:16 <Chocky> ugh
Jul 11 22:54:30 <jow_laptop> sorry
Jul 11 22:54:35 <jow_laptop> option timezone is correct
Jul 11 22:54:39 <jow_laptop> but it needs a tz string
Jul 11 22:54:45 <jow_laptop> no zonename
Jul 11 22:54:49 <jow_laptop> http://wiki.openwrt.org/doc/uci/system#time.zones
Jul 11 22:54:50 <jow_laptop> rtfm
Jul 11 22:54:53 <jow_laptop> ;)
Jul 11 22:54:58 * Chocky rererertfm
Jul 11 22:55:22 <Chocky> yes, is it not valid?
Jul 11 22:55:23 <jow_laptop> but to be fair the page is not clear on that
Jul 11 22:55:28 <Chocky> oh
Jul 11 22:55:46 <jow_laptop> America/Los Angeles is wrong
Jul 11 22:55:47 <Chocky> I need the second columnd
Jul 11 22:55:52 <jow_laptop> correct
Jul 11 22:56:42 <Chocky> I'm using glibc, does that help?
Jul 11 22:56:55 <jow_laptop> no
Jul 11 22:56:59 * Chocky sobs
Jul 11 22:57:00 <Chocky> ok
Jul 11 22:57:22 <jow_laptop> glibc respects the same tz strings
Jul 11 22:57:37 <jow_laptop> symbolic names will not work as we ship no zoneinfo/tzdata
Jul 11 22:57:45 <jow_laptop> its just too big for most targets
Jul 11 22:58:00 <Chocky> how big is tzdata?
Jul 11 22:58:20 <jow_laptop> on my ubuntu its 6.4M
Jul 11 22:58:28 * Chocky slaps
Jul 11 22:59:21 <Chocky> be a lot smaller in a squashfs, but still
Jul 11 22:59:39 <Chocky> surely there's an easy fix for mapping of those names.  that's only a few K
Jul 11 23:00:05 <jow_laptop> still a waste of space imo
Jul 11 23:00:19 <Chocky> few hundred bytes compressed?
Jul 11 23:01:01 <Chocky> my filesystem is 35M :p
Jul 11 23:01:49 <jow_laptop> ~16K
Jul 11 23:01:54 <Chocky> ok, well.  I'm sure there's been some discussion on this.  I'll leave it for now
Jul 11 23:01:57 <jow_laptop> only for  Foo/bar -> TZ String
Jul 11 23:02:31 <Chocky> for our case, we'll have to put it somewhere, so the point is moot.
Jul 11 23:02:42 <Chocky> probably PHP for now
Jul 11 23:02:51 <jow_laptop> plus maybe another 7K for zone code to offset mapping
Jul 11 23:03:07 <Chocky> one line grep :p
Jul 11 23:03:07 <jow_laptop> php... shudder
Jul 11 23:03:16 <Chocky> aw.
Jul 11 23:03:33 <Chocky> PHP is perhaps not my choice either, but my job is to make it all work
Jul 11 23:05:06 <Chocky> moving on
Jul 11 23:05:14 <Chocky> why is ath9k not spotting any interfaces?
Jul 11 23:05:25 <jow_laptop> hm?
Jul 11 23:05:44 <Chocky> it should bring up wlan0, ath0 or something, right?
Jul 11 23:05:54 <jow_laptop> no it shouldn't if wifi is disabled
Jul 11 23:06:03 <Chocky> good point
Jul 11 23:07:48 <Chocky> so, um:   Mon Jul 11 23:03:04 UTC 2011
Jul 11 23:08:11 <jow_laptop> the system value is written to /etc/TZ on boot
Jul 11 23:08:19 <jow_laptop> works for uclibc
Jul 11 23:08:22 <Chocky> crap
Jul 11 23:08:32 <jow_laptop> maybe glibc needs another way
Jul 11 23:08:39 <jow_laptop> environment variable or whatever
Jul 11 23:11:31 * Chocky pokes again at glibc patches he has in trac
Jul 11 23:13:51 <Chocky> right
Jul 11 23:14:05 <Chocky> you need to dump that /etc/TZ into TZ environment var
Jul 11 23:15:46 * luka12345|wiik has done enough patching for today
Jul 11 23:16:00 <Chocky> time for sewing?
Jul 11 23:39:05 <Chocky> um
Jul 11 23:39:33 <Chocky> where would I set export TZ=... to make sense?  not /etc/profile, surely
Jul 11 23:40:10 <jow_laptop> I don't know
Jul 11 23:40:24 <Chocky> where are system-wide vars set in openwrt?
Jul 11 23:40:33 <jow_laptop> intuitively I'd have tried /etc/profile
Jul 11 23:40:38 <jow_laptop> in /etc/profile
Jul 11 23:40:41 <Chocky> i.e, /etc/eniviroment
Jul 11 23:40:48 <Chocky> but that's only for login shells, right?
Jul 11 23:40:58 <jow_laptop> probably, yes
Jul 11 23:42:31 * Chocky tries
Jul 11 23:42:42 <Chocky> since that doesn't exist in Wrt
Jul 11 23:43:01 <Chocky> no joy
Jul 11 23:43:05 <Chocky> ugh
Jul 11 23:43:22 <Chocky> leave in profile for now
Jul 11 23:43:42 <jow_laptop> so profile works ?
Jul 11 23:44:06 <Chocky> well, I'm sure it'll appear to work great if you login and do "time", but it's not clear that'll work eg. for php under a web server
**** ENDING LOGGING AT Tue Jul 12 02:59:57 2011