**** BEGIN LOGGING AT Fri Apr 18 02:59:59 2014
Apr 18 04:37:55 <adri> nbd: ping. :P
Apr 18 05:36:05 <DonkeyHotei> adri: pong
Apr 18 05:47:29 <builder_> build #582 of orion is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/582
Apr 18 06:36:30 <builder_> build #575 of cobalt is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/575
Apr 18 06:47:27 <builder_> build #31 of pxa is complete: Failure [failed shell compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/pxa/builds/31
Apr 18 07:06:08 <adri> DonkeyHotei: the post is at http://adrianchadd.blogspot.com/
Apr 18 07:49:41 <builder_> build #31 of omap is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/31
Apr 18 09:37:59 <Pteridium> vampo: have you had success with vlans with the ARV7519PW or the ar9 based Zyxel?
Apr 18 09:38:00 <Pteridium> the routers that have the RTL8306
Apr 18 09:41:10 <vampo> Pteridium: I have not done any extended testing but at least port based vlans should work
Apr 18 09:41:33 <vampo> not sure about full 802.1q
Apr 18 09:41:48 <vampo> what problems do you see?
Apr 18 09:42:50 <Pteridium> i haven't success with the arv4518, maybe the reset gpio is wrong or bad defined
Apr 18 09:43:29 <Pteridium> so no port based vlans for now
Apr 18 09:44:23 <vampo> I don't have that board but maybe I can help if you gave me some details of the setup
Apr 18 09:45:00 <Pteridium> which details do you need? a complete bootlog?
Apr 18 09:45:26 <Pteridium> wit a lot of debug, of course
Apr 18 09:45:32 <vampo> ifconfig and swconfig dev eth0 show
Apr 18 09:45:37 <Pteridium> ok
Apr 18 09:58:38 <Pteridium> vampo: excuse me the delay. http://pastebin.com/xVuddqT8
Apr 18 09:59:30 <Pteridium> the gpio 13 is the switch reset signal AFAIK
Apr 18 10:00:25 <Pteridium> at least it was defined as 13 in AA
Apr 18 10:01:28 <builder_> build #33 of x86_64 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/x86_64/builds/33
Apr 18 10:02:27 <vampo> ok, first things first, the vlan config doesn't look good
Apr 18 10:02:42 <vampo> you have untagged ports on multiple vlans
Apr 18 10:02:49 <vampo> start with:
Apr 18 10:03:08 <vampo> swconfig dev eth0 vlan 0 set ports ' '
Apr 18 10:03:27 <vampo> and repeat for vlans 2, 3, 4 and 5
Apr 18 10:04:16 <vampo> until you see only vlan1 in the list
Apr 18 10:09:39 <Pteridium> ok, give me some minutes to test
Apr 18 10:09:52 <Pteridium> and thanks for your advices
Apr 18 10:10:00 <vampo> Pteridium: sure :)
Apr 18 10:10:16 <Pteridium> and this is the dts: http://pastebin.com/06fGhsbU
Apr 18 10:10:37 <vampo> these changes will be lost on reboot. you should transfer them to the config
Apr 18 10:10:54 <Pteridium_tests> ok
Apr 18 10:11:15 <vampo> the current driver tries to initialise port isolation and creates these vlans on every reset
Apr 18 10:35:35 <Pteridium> vampo: no luck
Apr 18 10:35:37 <Pteridium> http://pastebin.com/xyCbsBaV
Apr 18 10:36:46 <Pteridium> tried with eth0 and switch0
Apr 18 10:39:07 <Pteridium> maybe my fault: in the dts i've defined gpios = <&gpio 13 1>, while you defined gpios = <&gpio 19 0>
Apr 18 10:40:27 <vampo> does network work with this dts without vlans? The gpio is for reset/power the switch
Apr 18 10:42:01 <Pteridium> yes, with option vlan '0' and option ifname 'eth0' works fine
Apr 18 10:42:52 <Pteridium> 'eth0' instead 'eth0.1'
Apr 18 10:43:12 <vampo> ok, so the gpio is correct
Apr 18 10:43:36 <Pteridium> sure?
Apr 18 10:43:50 <Pteridium> sure it is correct?*
Apr 18 10:44:07 <vampo> well, at least it's not powering off the switch :)
Apr 18 10:44:17 <Pteridium> ok :)
Apr 18 10:45:00 <vampo> if you toggle that gpio, eth should stop even without vlans
Apr 18 10:45:13 <vampo> and at least then you know if you have the right one
Apr 18 10:46:07 <Pteridium> ok, give me a moment to make the tests.
Apr 18 11:45:02 <Pteridium_> vampo: you are right. setting gpio 13 to value 0 rtl8306 works, while with 1 it is reseted or powered off
Apr 18 11:46:50 <Pteridium> but only setting option ifname 'eth0'
Apr 18 11:47:01 <Pteridium> with option ifname 'eth0.1' nothing
Apr 18 11:48:05 <vampo> Pteridium: good, so we have confirmed that 13 is the correct GPIO
Apr 18 11:48:44 <Pteridium> yes, thanks to you
Apr 18 11:48:47 <vampo> now, before we try vlans, let's check if the switch - cpu link recovers after a reset
Apr 18 11:49:04 <Pteridium> checked and recovers
Apr 18 11:49:13 <vampo> good :)
Apr 18 11:49:41 <vampo> reset of the switch only, right? not the whole board
Apr 18 11:50:00 <Pteridium> yes, only the switch
Apr 18 11:50:02 <vampo> as in flipping the gpio, connectivity does return
Apr 18 11:50:08 <Pteridium> yes
Apr 18 11:50:47 <vampo> cool, so it isn't the initial vlan setup that kills the connection, it is the vlans themselves
Apr 18 11:51:07 <vampo> i.e. the cpu - switch still works
Apr 18 11:51:30 <Pteridium> ah
Apr 18 11:51:40 <Pteridium> but...
Apr 18 11:52:19 <vampo> can you check if any packets come out of the switch when vlans are configured? and if they do, are they tagged?
Apr 18 11:53:14 <Pteridium> how can i check it? with wireshark?
Apr 18 11:53:24 <vampo> to test, set up vlans, set up eth0.1 and start a ping from the board to the connected PC
Apr 18 11:53:35 <Pteridium> ok
Apr 18 11:54:12 <vampo> yes, on the pc check if wireshark can see anything coming out and check for the vlan headers
Apr 18 11:54:40 <Pteridium> i'll do as you say, but now is time to lunch
Apr 18 11:55:00 <vampo> ;) enjoy
Apr 18 11:55:27 <Pteridium> thanks for your help
Apr 18 11:55:39 <Pteridium> see you in a while
Apr 18 13:23:09 <jow_laptop> zajec__: I took a look at the nested form thing
Apr 18 13:23:42 <jow_laptop> zajec__: however there are a number of logical things that need to be sorted out
Apr 18 14:17:05 <jow_laptop> karlp: just read your mail regarding local modifications
Apr 18 14:17:16 <jow_laptop> karlp: there is a feature for that (CONFIG_SOURCE_TREE_OVERRIDE)
Apr 18 14:17:59 <jow_laptop> karlp: you enable source tree override in menuconfig, then you symlink the .git directory of a local clone of the package source to git-src within the openwrt package folder
Apr 18 14:18:46 <jow_laptop> karlp: e.g. I do for netifd,  ln -s /home/jow/devel/netifd.git/.git package/network/config/netifd/git-src
Apr 18 14:19:22 <jow_laptop> karlp: then a "make package/netifd/{clean,compile}" will pick the sources from my local clone regardless of whats defined in the Makefile
Apr 18 14:20:40 <zajec__> jow_laptop: yeah, i've relaized that
Apr 18 14:20:50 <zajec__> jow_laptop: all these classes seem a bit overcomplicated to me :(
Apr 18 14:21:22 <jow_laptop> zajec__: there is no way to make it significantly easier, that structure evolved over th course of six years
Apr 18 14:21:42 <jow_laptop> of course you can trim down the abstraction dramaticly, but then your views will get extremely complicated
Apr 18 14:21:58 <jow_laptop> because you end up recoding the crud logic every time
Apr 18 14:22:47 <jow_laptop> I played a bit around with nesting entire forms and its quite complicated
Apr 18 14:23:07 <jow_laptop> the rendering part is easy, a form will render itself into arbritary parent containers
Apr 18 14:23:22 <jow_laptop> but the event delegation is a nightmare
Apr 18 14:23:46 <jow_laptop> for example a nested form shall have no own save/reset etc. buttons but react on the actions of the parent form
Apr 18 14:24:16 <zajec__> right
Apr 18 14:25:26 <jow_laptop> structure wise you can already declare your wifi-device + related wifi-iface form layout, even without nesting support
Apr 18 14:26:52 <jow_laptop> ah no, you can't... a few things are missing
Apr 18 14:27:19 <jow_laptop> I'll try another approach... nested sections
Apr 18 14:49:03 <jow_laptop> zajec__: the subsection approach looks more promising and fits more logically into the existing structure
Apr 18 14:51:34 <zajec__> ok...
Apr 18 14:51:42 <zajec__> but i still wonder if this could be simplified
Apr 18 14:53:00 <zajec__> for example i find it confusing *Section handle the way data is displayed
Apr 18 14:53:17 <zajec__> i was wondering if we could have something like a layouts
Apr 18 14:53:48 <zajec__> like... DefaultLayout, TabbedLayout, TableLayout?
Apr 18 14:54:18 <zajec__> well, I guess I need to understand code better before going deeper
Apr 18 14:55:32 <zajec__> jow_laptop: [ot] do you know who could commit/push my [PATCH][packages] ntfs-3g: bump to 2014.2.15
Apr 18 14:58:58 <OpenWrt-Commits> nbd r40527 trunk/include/package.mk * build: reimplement package build directory override
Apr 18 15:02:53 <zajec__> jow_laptop: is "taboption" used for anything else than TypedSection?
Apr 18 15:02:59 <jow_laptop> zajec__: and what would the introduction of layouts simplify?
Apr 18 15:03:23 <jow_laptop> taboptions are supported by TypedSections, NamedSections, SingleSections
Apr 18 15:03:29 <jow_laptop> basically by all except TableSections
Apr 18 15:03:37 <zajec__> oh
Apr 18 15:03:43 <jow_laptop> because tabbing within a grid view of sections does not really make sense
Apr 18 15:04:11 <jow_laptop> but I could implement it
Apr 18 15:04:31 <jow_laptop> tabbing in a table section could devide the table rows
Apr 18 15:04:34 <jow_laptop> erm table cols
Apr 18 15:05:14 <jow_laptop> so yeah, out of lazyness I coded TableSections to ignore the tabbing, it currently simply renders the fields fro mthe first tab
Apr 18 15:06:37 <zajec__> jow_laptop: honestly, I don't know how to proceed
Apr 18 15:06:50 <zajec__> I just get irritated every time I try to understand some part :(
Apr 18 15:08:32 <jow_laptop> don't worry
Apr 18 15:08:47 <jow_laptop> I'm still wokring on the subsection support, then we can look over it together if you like
Apr 18 15:11:10 <zajec__> well, maybe it just needs documenting
Apr 18 15:11:14 <zajec__> maybe it would be easier then
Apr 18 15:11:48 <jow_laptop> crtainly :)
Apr 18 15:12:38 <zajec__> Right now I look at L.cbi.Map and... WTH is finish? WTH is section? WTH is formvalue? and so on...
Apr 18 15:17:01 <karlp> jow_laptop: thanks for those suggestions, I'll try them out at work next week, and try and write some edits to the "creating packages" wiki page with how to do this
Apr 18 15:17:16 <karlp> I'd asked here when it was removed what an alternative workflow was, but hadn't heard anything,
Apr 18 15:17:31 <karlp> if this way handles make clean properly (whch it sounds like it does) then it's better anyway :)
Apr 18 15:18:07 <zajec__> jow_laptop: btw. what I wanted some layout classes? to for example avoid duplicating tabs implementation in Map and TypedSection
Apr 18 15:20:39 <jow_laptop> zajec__: reason for the two distinct implementations is that the underlying object model fundamentally differs
Apr 18 15:54:04 <DonkeyHotei> adri: SIOCSIFHWADDR never reaches the ath9k code
Apr 18 15:57:52 <zajec__> jow_laptop: what are fields? why ever tab has fields instead of options directly?
Apr 18 16:29:20 <builder_> build #49 of sunxi is complete: Failure [failed shell]  Build details are at http://buildbot.openwrt.org:8010/builders/sunxi/builds/49
Apr 18 17:17:22 <builder_> build #32 of mpc83xx is complete: Failure [failed shell compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/32
Apr 18 17:28:37 <OpenWrt-Commits> nbd r40528 trunk/package/base-files/files/etc/rc.button/wps * base-files: remove duplicate wps button handler (#15621)
Apr 18 17:40:11 <builder_> build #538 of ppc44x is complete: Failure [failed shell_10]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/538
Apr 18 17:40:16 <builder_> build #538 of rb532 is complete: Failure [failed shell_10]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/538
Apr 18 18:05:36 <builder_> build #31 of adm8668 is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/adm8668/builds/31
Apr 18 18:07:51 <builder_> build #507 of sibyte is complete: Failure [failed shell_10 shell_15]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/507
Apr 18 19:08:03 <builder_> build #571 of lantiq is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/571
Apr 18 19:37:54 <jow_laptop> zajec__: fields are options
Apr 18 19:39:27 <zajec__> jow_laptop: what do u think about my p910nd module?
Apr 18 19:40:29 <jow_laptop> it looks okay
Apr 18 19:40:52 <zajec__> will you apply & push it? ;)
Apr 18 19:41:58 <jow_laptop> soon
Apr 18 19:42:13 <zajec__> alrighty :)
Apr 18 19:42:29 <jow_laptop> I want to cleanup some internal cbi functions as well
Apr 18 19:43:07 <jow_laptop> for example sections() in *Section classes should be called section_ids()
Apr 18 19:43:42 <jow_laptop> and I think about renaming section() to addSection(), option() to addOption() etc.
Apr 18 19:43:52 <zajec__> ack!
Apr 18 19:43:55 <jow_laptop> the current names are more or less transcriptions of cbi.lua
Apr 18 19:44:04 <zajec__> functions like section() are confusing
Apr 18 19:44:38 <jow_laptop> I already started renaming internal members
Apr 18 19:44:45 <zajec__> i also find it a bit confusing to use "package" for what uci usually calls "config"
Apr 18 19:44:56 <jow_laptop> uci on the cli calls it package too
Apr 18 19:45:00 <zajec__> err
Apr 18 19:45:02 <zajec__> no?
Apr 18 19:45:03 <jow_laptop> as does the uci library
Apr 18 19:45:22 <jow_laptop> but yeah, its switching back and forth here and there
Apr 18 19:45:29 <zajec__> $ uci help
Apr 18 19:45:31 <zajec__>         show       [<config>[.<section>[.<option>]]]
Apr 18 19:45:32 <zajec__>         get        <config>.<section>[.<option>]
Apr 18 19:46:19 <jow_laptop> uci export network | head
Apr 18 19:46:44 <zajec__> oh...
Apr 18 19:46:50 <zajec__> so ever "uci" mixes two names
Apr 18 19:46:55 <jow_laptop> yes
Apr 18 19:47:10 <jow_laptop> however the uci C api uniformly refers to the toplevel namespace as "package"
Apr 18 19:47:23 <zajec__> jow_laptop: if I may complain, i don't like _load_cb, as it goes to deeply into *Section internal IMO
Apr 18 19:47:24 <jow_laptop> afair a "config" (-file) can contain multiple packages
Apr 18 19:47:46 <zajec__> jow_laptop: i don't thnk _load_cb should touch fields directly
Apr 18 19:48:17 <jow_laptop> ack
Apr 18 19:48:33 <jow_laptop> it should call each sections load() which in turn calls load() on its fields
Apr 18 19:48:44 <zajec__> makes sense
Apr 18 19:49:04 <jow_laptop> so that the section internal arrangement is opague to the map
Apr 18 19:49:30 <zajec__> exactly
Apr 18 19:49:37 <jow_laptop> especially with the current subsection nesting, the load_cb would fail anyway
Apr 18 19:49:47 <zajec__> and then section could even think about embedding subsections ;)
Apr 18 19:49:58 <zajec__> exactly :P
Apr 18 20:08:15 <builder_> build #31 of cns3xxx is complete: Failure [failed shell_15]  Build details are at http://buildbot.openwrt.org:8010/builders/cns3xxx/builds/31
Apr 18 20:35:53 <adri> DonkeyHotei: interesting
Apr 18 20:36:00 <adri> DonkeyHotei: I wonder what the right way to do it then is
Apr 18 21:11:32 <builder_> build #517 of avr32 is complete: Failure [failed shell compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/517
Apr 18 21:26:20 <DonkeyHotei> adri: i'm open to ideas
Apr 18 21:26:34 <DonkeyHotei> adri: however, i did suggest sysfs
Apr 18 21:38:56 <adri> yeah
Apr 18 21:39:02 <adri> Well, I dunno if sysfs is going to get accepted in
Apr 18 21:42:22 <DonkeyHotei> as i said i'm open to ideas
Apr 18 22:05:05 <Hauke> we made an official statement on the Linksys WRT1900AC support: https://forum.openwrt.org/viewtopic.php?pid=230686
Apr 18 23:47:30 <builder_> build #445 of iop32x is complete: Failure [failed shell_10]  Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/445
Apr 18 23:52:31 <builder_> build #576 of cobalt is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/576
**** ENDING LOGGING AT Sat Apr 19 02:59:58 2014