**** BEGIN LOGGING AT Fri Dec 16 02:59:56 2011
Dec 16 04:20:52 <philipp64|laptop> blogic: ping
Dec 16 08:37:59 <blogic> philipp64|laptop: yp
Dec 16 09:12:23 <builder_> build #81 of rdc is complete: Failure [failed compile_7]  Build details are at http://buildbot.openwrt.org:8010/builders/rdc/builds/81
Dec 16 10:01:26 <acoul> http://paste.pocoo.org/show/521634/
Dec 16 10:52:50 <karlp> on http://luci.subsignal.org/trac/wiki/Documentation/LuCI-0.10#Hooks it says, "Modified configurations got committed and the CBI is about to restart associated services"
Dec 16 10:52:59 <karlp> how can I define what hte "associated service" is?
Dec 16 10:53:22 <karlp> or should I just add an on_commit hook to call an /etc/init.d script myself?
Dec 16 11:00:46 <karlp> I can see three different ways some luci stuff calls init.d scripts, os.execute, luci.util.exec, and luci.sys.call... Is there any noticeable difference between them?
Dec 16 11:18:46 <karlp> I'm using a o = section:option(DynamicList, ....) and then o.datatype="range(1,247)"
Dec 16 11:19:04 <karlp> which works well for the first entry in the list, it gets highlighted in red if the number is outside the range,
Dec 16 11:19:14 <karlp> but not for subsequent numbers in the list.
Dec 16 11:19:26 <karlp> is range meant to work on lists like that?
Dec 16 14:30:33 <jow_laptop> yes
Dec 16 14:30:37 <jow_laptop> karlp: ^
Dec 16 15:05:51 <dape-at-work> uhm, do we really need CONFIG_TMPFS_XATTR ?
Dec 16 15:07:31 <KanjiMonster> iirc yes, this is needed for the overlay fs when booting a fresh image while waiting for the jffs2 to be formatted
Dec 16 15:09:03 <dape-at-work> aight, thanks
Dec 16 15:09:36 <dape-at-work> KanjiMonster:  you dont mind if i test without it on my serialized wr740 :P
Dec 16 15:11:06 <acoul> http://paste.pocoo.org/show/521729/
Dec 16 15:11:07 <KanjiMonster> dape-at-work: I wouldn't even mind if you shot with a tazer at it; it's your hardware, you are free to do with it what you want ;)
Dec 16 15:13:34 <dape-at-work> meh, i only have a pepper spray..
Dec 16 15:13:36 <fullstop> You know, there's a limited number of flash erase/write cycles.  Every time you flash, there's one less.
Dec 16 15:13:39 <karlp> jow_laptop: by "yes" does that mean I should file a bug for it? or does it work for you? the bottom of the page shows: luci 0.10+svn809
Dec 16 15:13:48 <karlp> duh, 8090
Dec 16 15:14:01 <fullstop> My father used to say that about switches to keep us from playing with them.
Dec 16 15:16:13 <jow_laptop> karlp: I will not fix it most likely, its too complex
Dec 16 15:16:31 <karlp> that's fine, just wondering.
Dec 16 15:16:41 <karlp> the server side validation works too.
Dec 16 15:17:30 <jow_laptop> os.execute and luci.sys.call are identical
Dec 16 15:17:38 <karlp> do you want me to file a bug on subsignal that can be wontfixed for tracking?
Dec 16 15:17:44 <jow_laptop> luci.util.exec gathers the output into a string buffer
Dec 16 15:17:51 <jow_laptop> no need to file a bug
Dec 16 15:17:53 <karlp> ahh, good to know.
Dec 16 15:18:43 <karlp> http://luci.subsignal.org/trac/wiki/Documentation/LuCI-0.10#Datatypes links to the source file for the list of what the datatypes can be,
Dec 16 15:18:56 <karlp> could you possibly add a comment to the code maybe?
Dec 16 15:19:08 <jow_laptop> for what?
Dec 16 15:19:29 <karlp> that it doesn't work on dynamic lists?
Dec 16 15:19:45 <jow_laptop> but it does work
Dec 16 15:20:28 <karlp> I thought you just agreed that it doesn't, and that you wouldn't fix it?
Dec 16 15:20:44 <jow_laptop> well the validation catches the error
Dec 16 15:20:51 <jow_laptop> so its works as intended
Dec 16 15:21:04 <karlp> well, only if I add an option.validate() as well.
Dec 16 15:21:13 <jow_laptop> huh?
Dec 16 15:21:14 <karlp> if I only have option.datatype="range(1,247)"
Dec 16 15:21:25 <karlp> then the red highlighting only works for the first entry in a dynamic list
Dec 16 15:22:17 <karlp> gah, now it works for the first two, but not the third
Dec 16 15:23:10 <jow_laptop> works as intended for me
Dec 16 15:23:22 <jow_laptop> highlights every field with a bad value
Dec 16 15:23:46 <karlp> ok, great then :)
Dec 16 15:24:14 <karlp> I'll put it down to some bad cahching or something on my side then.
Dec 16 15:24:39 <karlp> oh, one other thing, what's the difference between save, and save and apply?
Dec 16 15:24:51 <karlp> save doesn't seem to update the /etc/config/blah file,
Dec 16 15:24:55 <karlp> save and apply does though.
Dec 16 15:25:11 <karlp> I've got a hook on "on_commit" which does the ini.d restart, which works nicely.
Dec 16 15:25:20 <karlp> can I just get ride of the save button somehow?
Dec 16 15:25:26 <jow_laptop> save is uci save
Dec 16 15:25:37 <jow_laptop> save & apply is uci save, uci commit + init script
Dec 16 15:25:59 <karlp> is the/an init script supposed to be restarted automatically?
Dec 16 15:26:18 <karlp> there's a note that seems to suggest that, but I haven't been able to see how it's linked,
Dec 16 15:26:28 <jow_laptop> yes
Dec 16 15:26:33 <karlp> or do you just mean if the module has an on_commit hook to call the init script?
Dec 16 15:26:54 <karlp> I was grepping the source for init.d and only found os.execute and friends.
Dec 16 15:27:34 <jow_laptop> you need an entry in /etc/config/ucitrack
Dec 16 15:28:55 <karlp> is there any documentation for that file?
Dec 16 15:29:04 <jow_laptop>  /bin/luci-reload
Dec 16 15:29:12 <karlp> is that config "luci CBI name", option init "init.d script name" ?
Dec 16 15:29:32 <jow_laptop> its config 'configname', option init 'init.d name'
Dec 16 15:31:19 <karlp> configname is the first name in the Map() call?
Dec 16 15:31:29 <jow_laptop> yes
Dec 16 15:31:38 <jow_laptop> which corresponds to the /etc/config/foo you operate on
Dec 16 15:31:45 <karlp> right.
Dec 16 15:32:04 <karlp> I'll give that a go, though the on_commit hook and exec is also working so far.
Dec 16 15:32:15 <karlp> that ucitrack looks like I can restarted dependent services though, which is cool.
Dec 16 15:33:13 <jow_laptop> problem with on_commit hooks is that they break the "queue saves, batch commit later" paradigm since they're only invoked when hitting save & apply on the corresponding model (e.g. not through the changes menu)
Dec 16 15:33:45 <karlp> yeah.  so, is there a way for a package to automatically add entries to ucitrack?
Dec 16 15:33:57 <jow_laptop> yes, I usually do that from within postinstall
Dec 16 15:33:58 <karlp> I've currently got a "first boot" script that does a bunch of uci cli calls,
Dec 16 15:34:13 <karlp> hmm, I should revisit postinstall.
Dec 16 15:34:39 <jow_laptop> http://luci.subsignal.org/trac/browser/luci/trunk/applications/luci-radvd/root/etc/uci-defaults/luci-radvd
Dec 16 15:34:51 <jow_laptop> http://luci.subsignal.org/trac/browser/luci/trunk/applications/luci-radvd/ipkg/postinst
Dec 16 15:36:21 <jow_laptop> basically I ship a uci-defaults script with each package and a postinstall snipped which invokes uci-defaults in the opkg install case
Dec 16 15:36:28 <jow_laptop> to avoid the need for a reboot to apply settings
Dec 16 15:36:59 <jow_laptop> not super elegant but works ok
Dec 16 15:37:28 <jow_laptop> at some point we really need to add a generic "delayed postinstall" facility to the buildroot
Dec 16 15:40:23 <jow_laptop> brb, I need to get myself some more coffee
Dec 16 15:58:51 <Chocks> jow_laptop: ~100 more flashes each on both AG300N and G300NH v1.  again, with same size image.  go figure.
Dec 16 20:20:15 <philipp64|laptop> hmm....  just did a build of libupnp and linuxigd, but it looks like the libupnp header files aren't being installed where linuxigd expects them (i.e. in staging_dir's usr/include) ... http://fpaste.org/CSGe/
Dec 16 20:50:50 <mazilo> What is the variable points to OpenWRT root directory?
Dec 16 20:54:47 <mazilo> Is "$(TOPDIR)" the right variable to use in any OpenWRT Makefile package to point to the OpenWRT root directory?
Dec 16 20:57:17 <jow_laptop> yes
Dec 16 20:59:22 <mazilo> jow_laptop; Thanks.
Dec 16 21:36:36 <philipp64|laptop> hmm...  rebuilding from scratch fixes that... must be a missing dependency.
Dec 16 22:15:05 <danitool> is there any plan to support ath9k onboard wlan chips in the bcm6358 boards?
Dec 16 22:16:15 <danitool> currently openwrt doesn't detect the ath calibration data partition, and wireless doesn't work spite having all modules loaded
Dec 16 22:43:38 <russell--> anybody have the uart pinout for the airrouter?
Dec 16 23:26:57 <stintel> nbd: I have kernel oops with ath9k it, trunk r29557 see http://pastebin.com/UvKBevgi
Dec 16 23:27:54 <stintel> something obvious or do I create a ticket ?
Dec 16 23:39:19 <nbd> stintel: make target/linux/clean world
Dec 16 23:40:26 <stintel> nbd: this is how I compiled: make package/kernel/clean target/linux/clean package/busybox/clean world V=99
Dec 16 23:42:26 <nbd> hm
Dec 16 23:44:04 <stintel> there are two devices in my rspro, one of them works
Dec 16 23:44:10 <stintel> the other doesn't
Dec 16 23:44:24 <stintel> there is only phy0 not phy1
Dec 16 23:44:56 <nbd> oh, right
Dec 16 23:45:03 <nbd> i see the bug now
Dec 16 23:45:28 <stintel> some extra info helps sometimes :)
Dec 16 23:47:01 <CIA-79> nbd * r29558 /trunk/package/mac80211/patches/581-ath9k_extra_platform_leds.patch: ath9k: fix a crash in led init
Dec 16 23:47:06 <nbd> stintel: there, should be fixed now
Dec 16 23:47:22 <stintel> nbd: thanks :)
Dec 16 23:47:25 <stintel> I'll test soon
Dec 16 23:47:56 <stintel> waiting for my x86 build to finish
Dec 17 00:02:24 <Memphiz> jow_laptop: anything you need from me for https://dev.openwrt.org/ticket/10562 ^^?
Dec 17 00:35:02 <stintel> heh, forgot to do git pull
Dec 17 00:56:26 <philipp64|laptop> anyone know if the AR5BXB92 works over plain-old USB? my Alix doesn't do PCI-e 1x over the mini-PCIe slot.
**** ENDING LOGGING AT Sat Dec 17 02:59:57 2011