**** BEGIN LOGGING AT Sat Dec 08 02:59:59 2012
Dec 08 10:16:20 <russell--> did packages feed get purged?  /me just building for mikrotik and noticed much i've become accustomed to missing
Dec 08 10:24:24 <russell--> like, perl for example... it's in feeds, but not showing up in make menuconfig
Dec 08 10:39:08 <russell--> HEAD CONFIG_ALL=y (with packages.git feed), .config has 2108 lines, r33395's .config was 5669 lines
Dec 08 10:48:55 <russell--> weirdness, did it harder, and it worked
Dec 08 10:49:42 <russell--> scripts/feed install -a helps (though i thought i'd done that before ...)
Dec 08 13:42:20 <DonkeyHotei> nbd: https://lists.openwrt.org/pipermail/openwrt-devel/2012-December/017805.html
Dec 08 13:42:58 <nbd> i'm following the thread, yes
Dec 08 13:44:51 <DonkeyHotei> it looks like he's basically suggesting rewriting /etc/init.d/boot for all devices
Dec 08 13:45:15 <DonkeyHotei> just to accomodate the geos
Dec 08 13:45:25 <DonkeyHotei> s/m/mm/
Dec 08 13:51:42 <DonkeyHotei> new reply posted
Dec 08 13:57:12 <DonkeyHotei> nbd: well?
Dec 08 13:57:49 <nbd> i don't know how you read that into what he's writing
Dec 08 13:58:50 <DonkeyHotei> "That needs fixing then. Don't just do things wrongly to work around a bug elsewhere — *fix* the bug."
Dec 08 13:59:26 <DonkeyHotei> the only bug that could be is loading modules before starting hotplug
Dec 08 14:00:08 <DonkeyHotei> what device other than the geos would that help, and what could it break everywhere else?
Dec 08 14:01:04 <DonkeyHotei> nbd: that's how
Dec 08 14:01:48 <nbd> if you say so ...
Dec 08 14:01:49 <nbd> ;)
Dec 08 14:02:42 <DonkeyHotei> but seriously, is /etc/init.d/boot really supposed to be doing things the way it is?
Dec 08 14:04:48 <nbd> yes
Dec 08 14:05:19 <nbd> leave it alone unless you know what you're doing, which i assume you don't ;)
Dec 08 14:05:33 <DonkeyHotei> exactly
Dec 08 14:06:10 <DonkeyHotei> so, it has to be worked around
Dec 08 14:06:46 <nbd> as i wrote in the email, this issue has already wasted way more time than it deserves
Dec 08 14:07:05 <nbd> the simple temporary solution it to just make an init script that figures out the mac address and then sets it for the atm device
Dec 08 14:07:11 <nbd> without hooking into hotplug
Dec 08 14:07:35 <nbd> works, doesn't have many corner cases, won't require changing /etc/init.d/boot
Dec 08 14:12:31 <DonkeyHotei> nbd: well, that's what http://patchwork.openwrt.org/patch/2987/ does, and that's what dwmw2_gone thought could be improved upon. blogic doesn't like it either, because it apparently conflicts with some as yet unknown changes he's making for the 3.7 kernel
Dec 08 14:14:18 <nbd> i think the hotplug event thing may not be working because of an atm driver issue
Dec 08 14:14:24 <nbd> because it's not registered via platform device
Dec 08 14:14:37 <nbd> so maybe it's not an issue with hotplug2, udevtrigger or /etc/init.d/boot at all
Dec 08 14:15:20 <DonkeyHotei> hotplug responds to sysfs node creation AFAICT
Dec 08 14:15:45 <nbd> yes, and udevtrigger generates events for devices that are already present at start time
Dec 08 14:16:11 <DonkeyHotei> but only for platform devices?
Dec 08 14:16:28 <DonkeyHotei> maybe that's the real bug
Dec 08 14:16:34 <nbd> maybe only for devices that are properly registered
Dec 08 14:16:45 <nbd> i.e. have any bus attached to them
Dec 08 14:17:08 <nbd> which the lantiq ones don't, because they're registering the ATM device from the module_init path, which is stupid
Dec 08 14:17:33 <DonkeyHotei> well, they don't sit on a bus
Dec 08 14:17:47 <nbd> either way, i'm going to stop caring now
Dec 08 14:18:02 <nbd> the driver will eventually be changed to be registered via device tree
Dec 08 14:18:09 <nbd> when john's 3.7 work hits the tree
Dec 08 14:18:32 <DonkeyHotei> and until then?
Dec 08 14:19:46 <nbd> john is the maintainer of the target. if he says the patch conflicts with his work, then it'll have to wait
Dec 08 14:20:24 <DonkeyHotei> ok.
Dec 08 14:54:13 <blogic> omg
Dec 08 14:54:28 <blogic> nbd: i stopped caring a few days ago
Dec 08 14:54:40 <blogic> there is no poin tin discussing
Dec 08 14:59:13 <DonkeyHotei> well, if dwmw2_gone stopped caring too, hopefully, that now makes it unanimous
Dec 08 14:59:24 <DonkeyHotei> i look forward to seeing 3.7
Dec 08 14:59:36 <DonkeyHotei> that is all i can say about it now
Dec 08 15:00:05 <blogic> stop it please
Dec 08 15:00:20 <blogic> you are starting to become a great annoyance
Dec 08 15:02:15 <OpenWrt-Commits> juhosg r34578 trunk/ (10 files in 10 dirs) * kernel: update linux 3.6 to 3.6.9
Dec 08 15:02:44 * DonkeyHotei sighs
Dec 08 17:06:39 <OpenWrt-Commits> florian r34579 packages/libs/ libqmi/Makefile libqmi * libqmi: helper library and utils to talk to QMI enabled modems
Dec 08 17:06:42 <OpenWrt-Commits> florian r34580 packages/libs/ libjpeg-turbo libjpeg-turbo/Makefile * libjpeg-turbo: add package
Dec 08 17:06:43 <OpenWrt-Commits> florian r34581 packages/admin/zabbix/Makefile * zabbix: update to 2.0.3
Dec 08 17:06:49 <OpenWrt-Commits> florian r34582 packages/utils/ (6 files in 3 dirs) * nut: add package NUT 2.6.5
Dec 08 17:06:51 <OpenWrt-Commits> florian r34583 packages/libs/ libdbi-drivers libdbi-drivers/Makefile * libs: add libdbi-drivers
Dec 08 22:32:29 <dwmw2_gone> hm, is there a way to list all options in a given config section, from the shell?
Dec 08 22:32:42 <dwmw2_gone> # cat /etc/config/ups
Dec 08 22:32:42 <dwmw2_gone> config driver 'openups'
Dec 08 22:32:42 <dwmw2_gone>         option driver usbhid-ups
Dec 08 22:32:42 <dwmw2_gone> 	option port usb
Dec 08 22:33:03 <dwmw2_gone> what shell function (in an initscript) will call my own function with $1=driver $2=usbhid-ups
Dec 08 22:33:11 <dwmw2_gone> and then with $1=port $2=usb etc.
Dec 08 22:33:26 <dwmw2_gone> I can only find things that'll "return all the driver= options" and "return all the port= options" etc.
Dec 08 22:33:40 <dwmw2_gone> I can't quite work out config_list_foreach()
Dec 08 22:33:57 <dwmw2_gone> or list()
Dec 08 22:34:19 <dwmw2_gone> they seem to require that something_LENGTH be set. And it isn't.
Dec 08 22:39:02 <dwmw2_gone> aha, perhaps I have to use option_cb()
Dec 09 00:16:52 <russell--> orbit2 doesn't build for me as of HEAD (as of last night, for mikrotik)
Dec 09 00:17:49 <russell--> r34577
**** ENDING LOGGING AT Sun Dec 09 02:59:58 2012