**** BEGIN LOGGING AT Sun Dec 09 02:59:58 2012
Dec 09 12:34:12 <OpenWrt-Commits> cyrus r34584 trunk/package/network/ipv6/6relayd/Makefile
Dec 09 12:34:12 <OpenWrt-Commits> 6relayd: Always send NA with hop-limit 255
Dec 09 12:34:12 <OpenWrt-Commits> * thanks risa2000 for investigating
Dec 09 13:33:12 <OpenWrt-Commits> cyrus r34585 trunk/package/network/ipv6/6distributed/Makefile * 6distributed: Correctly handle duplicate prefixes
Dec 09 13:55:05 <OpenWrt-Commits> cyrus r34586 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 09 13:55:05 <OpenWrt-Commits> ipv6-support: Next iteration
Dec 09 13:55:05 <OpenWrt-Commits> * Make ula_prefix a global option
Dec 09 13:55:05 <OpenWrt-Commits> * Add preliminary support for NAT (Linux >= 3.7, ip6tables >= 1.4.17)
Dec 09 13:55:07 <OpenWrt-Commits> * Minor fixes for prefix fallback
Dec 09 15:13:49 <mesx> hi folks, sorry for buging you with this, but no one else knows... https://forum.openwrt.org/viewtopic.php?id=40937
Dec 09 15:15:24 <DonkeyHotei> you operate your own dslam?
Dec 09 15:15:31 <mesx> no
Dec 09 15:16:08 <DonkeyHotei> "my Huawei DSLAM"
Dec 09 15:16:53 <mesx> on my line, (on the other end is Huawei DSLM)
Dec 09 15:17:03 <mesx> which can be seen from wireshark
Dec 09 15:17:09 <mesx> img i posted
Dec 09 15:17:37 <DonkeyHotei> the different place you tested is annex b and your place is annex a?
Dec 09 15:17:43 <mesx> yes
Dec 09 15:17:48 <blogic> is it yozur dslam ?
Dec 09 15:18:06 <blogic> i am about to go offline
Dec 09 15:18:13 <mesx> no, its my operators dslam, sry for not clearing that
Dec 09 15:18:15 <blogic> but lets try and chat later or tomorrow
Dec 09 15:18:19 <mesx> ok
Dec 09 15:18:22 <DonkeyHotei> did you build separate annex-b and annex-a builds of owrt?
Dec 09 15:18:23 <mesx> i will be here
Dec 09 15:18:29 <mesx> no
Dec 09 15:18:48 <mesx> i just changed in /etc/config/network annex
Dec 09 15:18:52 <blogic> ah
Dec 09 15:18:58 <blogic> make sure its the right annex
Dec 09 15:19:03 <mesx> how? :)
Dec 09 15:19:08 <blogic> wlel
Dec 09 15:19:11 <blogic> is it a or b ?
Dec 09 15:19:16 <blogic> i mean your line
Dec 09 15:19:20 <DonkeyHotei> annex b and annex a use different dsp firmwares, selectable as different packages in menuconfig
Dec 09 15:19:22 <mesx> at my place is A
Dec 09 15:19:29 <mesx> ah so
Dec 09 15:19:30 <blogic> and you made a A firmware ?
Dec 09 15:19:39 <mesx> didnt check for that
Dec 09 15:19:43 <mesx> so i dont konw
Dec 09 15:19:47 <blogic> ls /lib/firmware/
Dec 09 15:19:52 <mesx> sec
Dec 09 15:20:04 <DonkeyHotei> sounds like you made a b-only build
Dec 09 15:20:28 <mesx> so im in root of my aa git
Dec 09 15:20:28 <mazilo> Did anyone here manage to get a Seagate GoFLEX Net from Fry's Electronics for $9.99 (http://www.frys.com/product/6483422)?
Dec 09 15:20:32 <mesx> where to look?
Dec 09 15:21:21 <DonkeyHotei> () makes links not clickable in xchat/hexchat; i wish people would use <>
Dec 09 15:21:33 <blogic> in the device
Dec 09 15:21:57 <blogic> there will be a /lib/firmware/ltq-dsl-fw-b-danube.bin
Dec 09 15:21:58 <blogic> or
Dec 09 15:22:04 <mesx> I'm not at home currently :/
Dec 09 15:22:05 <blogic> /lib/firmware/ltq-dsl-fw-a-danube.bin
Dec 09 15:22:28 <mesx> ok, what confip option to choose when building a-dsl-fw
Dec 09 15:22:39 <blogic> ok
Dec 09 15:22:42 <blogic> then lets chat tomorrow
Dec 09 15:22:47 <blogic> maybe the xtu bits are broken
Dec 09 15:22:51 <mesx> ok, ty
Dec 09 15:23:21 <DonkeyHotei> mazilo: all locations within 100 mi of me show "unavailable"
Dec 09 15:24:28 <mazilo> DonkeyHotei: Sorry about that. Same here. But, I managed to get some units from friends who live near by Fry's Electronics stores.
Dec 09 15:24:41 <mazilo> I think this is a really good deal.
Dec 09 15:25:42 <mazilo> It is a Marvell Kirkwood platform and should work right out of the box with OpenWRT.
Dec 09 15:26:57 <DonkeyHotei> mazilo: prior to april, there was a fry's directly across the interstate from me. i can still get to one without too much trouble, but i no longer often do so due to less money available to spend and no car
Dec 09 15:27:46 <DonkeyHotei> had i known of this deal, i might have gone for it
Dec 09 15:28:47 <DonkeyHotei> but whenever fry's has deals this good on anything, it's never more than 2 days before they're permanently sold out
Dec 09 15:29:04 <mazilo> IC.
Dec 09 15:29:24 <mazilo> BTW, J&R (http://www.jr.com/product/productListing.jsp?N=4294205859) has both Pogoplug P21 and Pogoplug Mogile (POGOV4A101 ) onsale for $19.99 (FREE S/H). The former one is based on an OXNAS chipset with a close source driver.
Dec 09 15:29:29 <DonkeyHotei> at least around here
Dec 09 15:29:38 <DonkeyHotei> techies know a good deal
Dec 09 15:30:00 <mazilo> I believe the POGOV4A101 should work with OpenWRT right out the box.
Dec 09 15:30:43 <mazilo> These all devices I mentioned come with ONLY 128/128 MB RAM/NAND.
Dec 09 15:31:27 <DonkeyHotei> i'm still hoping to find an F9 referrer willing to share the referral credit, but that would mean i need to get this sx762 working, and it's not cooperating
Dec 09 15:31:57 <mazilo> DonkeyHotei: What is F9?
Dec 09 15:32:05 <DonkeyHotei> future-nine
Dec 09 15:32:17 <mazilo> DonkeyHotei: Oh IC.
Dec 09 15:33:30 <DonkeyHotei> near as i can figure, the device is missing some kind of pull-up for serial
Dec 09 15:35:03 <mazilo> DonkeyHotei: R U talking about PogoPlug P21 or POGOV4A101?
Dec 09 16:00:48 <mesx> blogic: I just checked my make config file and kmod--ltq-dsl-firmawre-a-danube is included
Dec 09 16:42:26 <DonkeyHotei> mesx: both a and b?
Dec 09 17:13:10 <mesx> yes
Dec 09 17:31:51 <DonkeyHotei> then i should hope selecting it in /etc/config/network would do it, but i could be wrong; you should wait for blogic
Dec 09 17:33:43 <mesx> yeah, I'm dealing with this problem for a month, so I waiting day or two isnt a problem :) ty anyway
Dec 09 17:37:43 <DonkeyHotei> i'm having my own problems with an sx762
Dec 09 17:38:45 <DonkeyHotei> [Sun 2012-12-09 07:33:30 AM PST] <DonkeyHotei> near as i can figure, the device is missing some kind of pull-up for serial
Dec 09 17:39:07 <DonkeyHotei> i don't want to do anything to the device just blind
Dec 09 17:39:59 <DonkeyHotei> mesx: do you have serial on yours?
Dec 09 17:43:34 <mesx> yeah serial works fine on mine, in fact, i got 3 sx763 routers and serial works on all of them
Dec 09 17:43:43 <DonkeyHotei> odd
Dec 09 17:44:11 <DonkeyHotei> you got the Tx line from the point in the pic?
Dec 09 17:44:53 <mesx> DonkeyHotei: http://wiki.openwrt.org/_detail/toh/gigaset/serialv2_sx76x_danube.jpg?id=toh%3Agigaset%3Asx76x
Dec 09 17:46:11 <DonkeyHotei> that's the pic i meant, yes
Dec 09 18:39:51 <luka> mesx: kirkwood should work out of the box yes
Dec 09 18:44:10 <mesx> luka: it doesnt :/
Dec 09 18:44:42 <luka> give me some logs, output, etc...
Dec 09 18:45:01 <luka> which board...
Dec 09 18:47:11 <mesx> did you see my forum post? https://forum.openwrt.org/viewtopic.php?id=40937
Dec 09 18:47:57 <mesx> i can give you logread and ps output
Dec 09 18:49:58 <mesx> logread https://gist.github.com/4246456
Dec 09 18:50:45 <mesx> ps output, with some additonal pppd_options https://gist.github.com/4246455
Dec 09 18:52:18 <DonkeyHotei> luka: mazilo was the one talking about kirkwood; mesx has a lantiq
Dec 09 20:32:22 <OpenWrt-Commits> swalker r34587 packages/lang/python-crypto/ patches/001-no-host-paths.patch patches * [packages] python-crypto: remove host include path (#12592)
Dec 09 22:17:12 <swalker> updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/index.html
Dec 09 22:19:26 <luka> swalker: Sorry, you are not authorized to view this page.
Dec 09 22:20:05 <luka> swalker: i get that error
Dec 09 22:20:13 <swalker> luka: should be fixed, lame web interface
Dec 09 22:20:27 <luka> swalker: works, thx
Dec 09 22:21:05 <heilterminalcons> you are right e.g. for memcached i dont have access
Dec 09 22:21:14 <heilterminalcons> how to fix that?
Dec 09 22:22:41 <swalker> heilterminalcons: nag Kaloz/thepeople until you stop caring afaik
Dec 09 22:30:22 <heilterminalcons> tried it but no luck
Dec 09 22:30:32 <heilterminalcons> i will keep on trying
Dec 09 22:45:41 <dwmw2_gone> can anyone help me with luci-statistics stuff?
Dec 09 22:46:08 <dwmw2_gone> trying to work out how to make it render a graph where the ds isn't "value".
Dec 09 22:48:52 <jow_laptop> sources = { "something_not_value" }
Dec 09 22:49:01 <jow_laptop> see e.g. load.lua
Dec 09 22:49:28 <jow_laptop> implicit default is always  sources = { "value" }
Dec 09 22:49:35 <dwmw2_gone> thanks
Dec 09 22:50:38 <dwmw2_gone> 16846 write(2, "No DS called 'value' in '/tmp/rrd/geos.infradead.org/nut-openups/percent-charge.rrd'", 84) = 84
Dec 09 22:51:14 <dwmw2_gone> http://david.woodhou.se/nut.lua
Dec 09 22:51:15 <jow_laptop> can you pastebin an rrdtool info output of it?
Dec 09 22:51:26 <dwmw2_gone> it's "percent"
Dec 09 22:52:07 <jow_laptop> ah, sorry, it should have been  sources = { charge = { "percent" } }
Dec 09 22:52:12 <dwmw2_gone> thanks
Dec 09 22:52:25 <jow_laptop> since there can be multiple sources with multiple valuetypes each
Dec 09 22:53:13 <dwmw2_gone> no change
Dec 09 22:53:25 <dwmw2_gone> er
Dec 09 22:53:26 <dwmw2_gone> percent
Dec 09 22:53:28 <dwmw2_gone> not charge
Dec 09 22:53:35 <dwmw2_gone> 		        sources = { percent = { "percent" } },
Dec 09 22:53:43 <dwmw2_gone> yay
Dec 09 22:54:32 <dwmw2_gone> There is 20th century charset stuff going on here
Dec 09 22:55:00 <jow_laptop> I have to remember the rrdtool structure every time I touch the graph definitions
Dec 09 22:55:06 <dwmw2_gone> My °C temperature ends up as °C in the image
Dec 09 22:55:09 <jow_laptop> I always forget it
Dec 09 22:55:20 <jow_laptop> yes, rrdtool graph only handles iso-8859-1
Dec 09 22:55:36 <jow_laptop> so make sure you encode your source file accordingly
Dec 09 22:56:13 <dwmw2_gone> welcome back to the 20th century ☹
Dec 09 22:56:26 <jow_laptop> its a limitation of the underlying libgd
Dec 09 22:56:39 <jow_laptop> rrdtool 1.2 would handle it properly but it is *huge*
Dec 09 22:56:45 <jow_laptop> with ttf support etc.
Dec 09 23:01:44 <dwmw2_gone> do I need to put the 'instances' in?
Dec 09 23:01:51 <dwmw2_gone> or can I just let it show everything that's there?
Dec 09 23:02:07 <jow_laptop> you need to specify either instances or sources
Dec 09 23:02:39 <jow_laptop> there is no code currently which inspects the rrd structure, so it does not auto-discover instaces
Dec 09 23:05:24 <dwmw2_gone> can I fix the scale?
Dec 09 23:05:33 <dwmw2_gone> so the percentage UPS charge always goes to 100%?
Dec 09 23:05:47 <dwmw2_gone> and can I stop it from having an average?
Dec 09 23:05:53 <dwmw2_gone> That a bit pointless for time left and charge fields
Dec 09 23:05:59 <dwmw2_gone> and probably temperature
Dec 09 23:08:26 <jow_laptop> not with the current code
Dec 09 23:08:46 <dwmw2_gone> ok, thanks
Dec 09 23:08:50 <dwmw2_gone> this'll do for now :)
Dec 09 23:10:46 <jow_laptop> actually there's an option to turn off min and max, but avg always stays
Dec 09 23:11:18 <jow_laptop> as for the scale, thats fully implicit atm, rrdtool graph does it
Dec 09 23:18:08 <dwmw2_gone> my voltage graph appears to go up to 40V
Dec 09 23:18:11 <dwmw2_gone> but I can't smell smoke
Dec 09 23:20:26 <dwmw2_gone> http://david.woodhou.se/nut.png
Dec 09 23:20:30 <dwmw2_gone> that image doesn't even make sense
Dec 09 23:20:48 <dwmw2_gone> battery voltage average 12.9V... yet a flat horizontal line at about 25?
Dec 09 23:20:56 <dwmw2_gone> output voltage looks saner.
Dec 09 23:21:19 <jow_laptop> they're stacked
Dec 09 23:21:48 <jow_laptop> in the data = { ... } table add  overlay = true
Dec 09 23:22:53 <jow_laptop> if you want to see the individual lines, add  noarea = true  to the per-source options
Dec 09 23:23:21 <jow_laptop> that will prevent it from filling the are underneath the line with color
Dec 09 23:23:25 <jow_laptop> *the area
Dec 09 23:24:31 <dwmw2_gone> 		data = {
Dec 09 23:24:31 <dwmw2_gone> 			noarea = true,
Dec 09 23:24:31 <dwmw2_gone> 			overlay = true,
Dec 09 23:24:31 <dwmw2_gone> 			instances = {
Dec 09 23:24:32 <dwmw2_gone> like that?
Dec 09 23:25:13 <dwmw2_gone> sorry, per-source options for noarea
Dec 09 23:26:11 <jow_laptop> yes
Dec 09 23:26:41 <dwmw2_gone> overlay doesn't work there
Dec 09 23:27:45 <dwmw2_gone> 				voltage_output  = { color = "00e000", title = "Output voltage", noarea=true, overlay=true },
Dec 09 23:27:46 <dwmw2_gone> 				voltage_battery = { color = "0000ff", title = "Battery voltage", noarea=true, overlay=true },
Dec 09 23:27:46 <dwmw2_gone> 				voltage_input   = { color = "ffb000", title = "Input voltage", noarea=true, overlay=true }
Dec 09 23:27:47 <dwmw2_gone> that works
Dec 09 23:28:17 <jow_laptop> yes, was about to say that
Dec 09 23:28:18 <jow_laptop> sorry
Dec 09 23:32:06 <dwmw2_gone> hah. My UPS reports battery current of 655.28A when it's *discharging*
Dec 09 23:32:23 <dwmw2_gone> that fucks *that* graph, but other than that I think it's all working
Dec 09 23:32:49 * dwmw2_gone goes to tidy up the nut BuildDev bits, the collectd-mod-nut build, and the new luci-statistics support 
Dec 09 23:33:35 <jow_laptop> I can probably hack the rrdtool.lua abstraction to allow more stuff, e.g. value filters to cap data above or below certain thresholds
Dec 09 23:33:46 <jow_laptop> just need to read into the cdef language again
Dec 09 23:33:56 <jow_laptop> its been over five years or so since I wrote that
Dec 09 23:34:14 <dwmw2_gone> I'll get its firmware fixed
Dec 09 23:34:52 <jow_laptop> line thickness is another attribute I'd like to expose
Dec 09 23:35:04 <jow_laptop> its currently always 1 for area and 2 for noarea
Dec 09 23:35:23 <jow_laptop> but when overlaying lines that tend to report the same values it would be helpful if they could have different thinkess
Dec 09 23:35:32 <jow_laptop> so that all are visible
Dec 09 23:36:29 <jow_laptop> problem with the old rrdtool 1.0.x is that it does not support alpha transparency so intersections of areas are not visiable, the topmost one just masks everything below
Dec 09 23:38:07 <jow_laptop> anyhow, I suppose there're bigger issues to deal with
Dec 09 23:52:35 <dwmw2_gone> I'm happy just to have UPS graphs from nut for now :)
Dec 09 23:54:40 <dwmw2_gone> http://david.woodhou.se/luci-stats-ups.png
Dec 09 23:56:20 <jow_laptop> now I see what you mean with avg is useless
Dec 09 23:56:32 <jow_laptop> forgot about the fact that last is printed too
Dec 09 23:56:47 <dwmw2_gone> Last is OK
Dec 09 23:56:52 <dwmw2_gone> means you don't have to interpret the graph
Dec 09 23:57:03 <jow_laptop> right
Dec 09 23:57:07 <dwmw2_gone> you get accurate value for "current" reading.
Dec 09 23:57:13 <dwmw2_gone> which admittedly may be a few minutes ago. But the latest
Dec 09 23:57:21 <dwmw2_gone> with 3 signficant figures or whatever
Dec 09 23:59:28 * dwmw2_gone goes back to the beginning of the evening's hacking session, and looks at the Build/InstallDev target
Dec 09 23:59:43 <dwmw2_gone> can we use 'make DESTDIR=$(1) install' for that?
Dec 10 00:00:25 <jow_laptop> I don't see why not
Dec 10 00:00:39 <dwmw2_gone> it'll install some bits (like executables) that we don't need in the buildroot
Dec 10 00:00:47 <dwmw2_gone>  and I suppose they'll get packaged into the SDK too?
Dec 10 00:00:54 <jow_laptop> for the target buidl we have a variable  PKG_INSTALL:=1  which essentially means the install phase is invoked with make ... install DESTDIR=$(1)
Dec 10 00:01:39 <dwmw2_gone> this is already set for nut
Dec 10 00:01:57 <jow_laptop> ok, then it is not propagated to InstallDev apparently
Dec 10 00:02:02 <dwmw2_gone> but that's for install, not InstallDev? $(1) is different here, right?
Dec 10 00:03:44 <jow_laptop> well as far as I understand it, InstallDev has no default implementation
Dec 10 00:03:49 <dwmw2_gone> makes sense
Dec 10 00:03:59 <jow_laptop> and the macro contents are simply expanded after the package install phase has been run
Dec 10 00:04:53 <jow_laptop> so if PKG_INSTALL is set, there should've already been an make install into $(PKG_INSTALL_DIR) at the time the InstallDev commands are executed
Dec 10 00:05:18 <jow_laptop> so you should be able to blindly copy e.g. $(PKG_INSTALL_DIR)/usr/include into $(1)/usr/include
Dec 10 00:05:23 <dwmw2_gone> is that the same directory?
Dec 10 00:05:49 <dwmw2_gone>  /home/dwmw2/git/openwrt/staging_dir/target-i386_uClibc-0.9.33.2/usr/include is where I had to put the headers
Dec 10 00:06:05 <jow_laptop> no, PKG_INSTALL_DIR is a temporary installation prefix where Install draws it files from
Dec 10 00:06:08 <dwmw2_gone> that *isn't* the directory that gets used for the rootfs, or the package build, is it?
Dec 10 00:06:17 <dwmw2_gone> ah, right.
Dec 10 00:06:21 <jow_laptop> we cannot use upstream "make install" to e.g. create packages because we want to only copy certain files
Dec 10 00:06:31 <jow_laptop> no man pages, headers, ...
Dec 10 00:06:39 <dwmw2_gone> right
Dec 10 00:07:01 <dwmw2_gone> ok, I think I can make this one work then; thanks
Dec 10 00:22:36 <dwmw2_gone> jow_laptop: http://git.infradead.org/users/dwmw2/openwrt-packages.git — do those three look sane ?
Dec 10 00:26:18 <jow_laptop> minus the debug echos in InstallDev it looks sane
Dec 10 00:26:52 <dwmw2_gone> oops
Dec 10 00:26:53 <dwmw2_gone> thanks
Dec 10 00:27:48 <jow_laptop> there's also a convention to call uci configs files/foo.config and init scripts files/foo.init in the package source
Dec 10 00:28:05 <dwmw2_gone> ok
Dec 10 00:28:17 <jow_laptop> so maybe move files/ups to files/ups.init and adjust the make rules accordingly
Dec 10 00:29:03 <jow_laptop> is upsdd dropping privileges?
Dec 10 00:29:30 <dwmw2_gone> not yet.
Dec 10 00:29:33 <dwmw2_gone> I'll sort that out later
Dec 10 00:29:34 <jow_laptop> just wondering if it is allowed to write to /var/run... ok
Dec 10 00:30:29 <dwmw2_gone> I'll also recompile it to have /var/run/ups as its sysconfdir, isntead of using symlinks
Dec 10 00:30:36 <dwmw2_gone> after I've made *all* the conf files get autogenerated
Dec 10 00:30:47 <dwmw2_gone> pushed those changes (untested yet)
Dec 10 00:30:53 <stintel> interesting, dwmw2_gone thanks for your work on this :)
Dec 10 00:33:49 <dwmw2_gone> hth
Dec 10 00:36:17 <jow_laptop> $(CP) is defined as cp -fpR so you can copy the entire include/ dir in InstallDev
Dec 10 00:36:22 <jow_laptop> in case they ever add subdirs
Dec 10 00:36:38 <dwmw2_gone> ok
Dec 10 00:36:52 <dwmw2_gone> does that work if it already exists?
Dec 10 00:37:01 <jow_laptop> then you can also spare the $(INSTALL_DIR) before and copy directly to $(1)/usr/
Dec 10 00:37:13 <jow_laptop> I suppose it does (-f)
Dec 10 00:37:15 <dwmw2_gone> $(1)/usr/include might already exist
Dec 10 00:37:26 <dwmw2_gone> doesn't that create $(1)/usr/include/include/ ? :)
Dec 10 00:38:31 <jow_laptop> it merges the dirs
Dec 10 00:38:36 <jow_laptop> just tested
Dec 10 00:44:36 <thepeople> swalker: even I have to go through Kaloz, I was just trying to keep tack of who was wanting access :-)
**** ENDING LOGGING AT Mon Dec 10 02:59:58 2012