**** BEGIN LOGGING AT Fri Nov 28 02:59:58 2014 Nov 28 05:25:54 build #756 of ar7 is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/756 Nov 28 09:03:53 build #731 of xburst is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/731 Nov 28 10:04:09 build #645 of iop32x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/645 Nov 28 10:41:24 anyone about Nov 28 11:43:01 cyrus r43421 trunk/package/ utils/busybox/files/sysntpd utils/busybox/Makefile utils/busybox/files/ntpd-hotplug * ntp: add hotplug class for ntp-events Nov 28 14:04:28 jow_laptop, ping ! Nov 28 14:04:47 nighty-_: pong Nov 28 14:05:44 jow_laptop, question : how can I can use temporary variable ( not to be stored in uci variables / files ) for fields ? Nov 28 14:06:04 you need to provide more context Nov 28 14:06:07 jow_laptop, because I need to concatenate values before storing them into variables Nov 28 14:06:15 ok Nov 28 14:06:18 ah ok Nov 28 14:06:18 let me show you Nov 28 14:06:28 assuming you want to concat three inputs Nov 28 14:06:55 set all fields remove and write callbacks to the same function Nov 28 14:06:57 http://pastebin.com/9gUE5zQg Nov 28 14:07:28 I can want concatenate bgscan="learn:30:-45:300:/etc/wpa_supplicant/network1.bgscan" Nov 28 14:07:35 I can want concatenate as bgscan="learn:30:-45:300:/etc/wpa_supplicant/network1.bgscan" Nov 28 14:07:49 or bgscan="simple:30:-45:300" Nov 28 14:08:21 where these value are short_interval and sst and long intervals Nov 28 14:09:03 jow_laptop, can you show short example ? Nov 28 14:12:39 nighty-_: yeah - wait Nov 28 14:16:57 nighty-_: http://pastebin.com/ARbNmgvh - something like that Nov 28 14:17:20 nighty-_: that approach will calculate the final value multiple times though so you need to optimize it Nov 28 14:17:46 nighty-_: I suggest to attach the callback only to bgscan_mode, thats a ListValue so it cannot be empty Nov 28 14:17:58 nighty-_: assign an empty function to all others Nov 28 14:18:30 ok let me look at this and understand it Nov 28 14:18:36 the idea is simply Nov 28 14:18:45 you override the default .write() callback Nov 28 14:19:03 but instead of writing only the own value into uci, you fetch the values of all related fields too Nov 28 14:19:18 concat a custom value and write that into uci Nov 28 14:20:44 jow_laptop, I am not sure I see where the concatanation occurs here Nov 28 14:20:51 line 9 Nov 28 14:21:08 oh Nov 28 14:21:09 ok Nov 28 14:21:13 syntax sugar for string.format("%s:%d:%d:%d", mode or "simple", sinv or 1, ssth or -45, linv or 300) Nov 28 14:21:18 sorry I was looking the definitions Nov 28 14:23:41 jow_laptop, thanks I will study this :) Nov 28 14:24:18 jow_laptop, what is self.map:set(section, "bgscan", value) for ? Nov 28 14:24:47 it will triger the uci set wireless.$section.bgscan=foo Nov 28 14:26:58 I see Nov 28 14:27:03 last question Nov 28 14:27:15 how I can set default for a multivalue ? Nov 28 14:27:26 defaults Nov 28 14:27:35 in that example the scan_freq Nov 28 14:31:32 jow_laptop, Nov 28 14:33:33 as far as I can see there is no default support for checkboxes Nov 28 14:33:33 uhmmmm somehow I can error running that code Nov 28 14:36:05 http://pastebin.com/v67GZjXL Nov 28 14:36:16 where did I screw up ? Nov 28 14:38:28 morning guys Nov 28 14:44:15 nighty-_: maybe the foo = bar = baz assignment is invalid Nov 28 14:44:31 looks like it Nov 28 14:44:43 nighty-_: replace it with foo = baz \n bar = baz Nov 28 14:44:50 But I thought it was some trick that you used Nov 28 14:45:04 so I was trying to see where I had made a mistake Nov 28 14:47:30 jow_laptop: can i ask you what do you think about that? https://dev.openwrt.org/ticket/18396#ticket Nov 28 14:54:06 oh hey jow_laptop did know you were in here Nov 28 14:54:19 I just got a response from MikroTik about new rb951g Nov 28 14:54:32 apprantely they've changed bootloader and it now expects the kernel to be in a different location Nov 28 14:54:48 I've pasted their response in comments here: https://dev.openwrt.org/ticket/18434#comment:1 Nov 28 14:55:55 dlorhh: I'm the wrong guy to ask Nov 28 15:00:54 jow_laptop, concatanation does not appear to work Nov 28 15:02:21 thats a bit unsepcific Nov 28 15:03:54 DomeMaster: well, the most important info is missing, i.e. which is the new offset Nov 28 15:04:03 jow_laptop, yeah sorry :) Nov 28 15:04:19 I know KanjiMonster, I've asked back to get as much info as possible Nov 28 15:04:19 jow_laptop, let me show you the code again Nov 28 15:04:28 also I'm suspicious, because I'm getting ARP replies Nov 28 15:04:33 that means the kernel must be loaded right? Nov 28 15:07:28 jow_laptop, http://pastebin.com/vK9sXA3n Nov 28 15:12:10 http://pastebin.com/NHLSBaYQ Nov 28 15:12:28 KanjiMonster: do you agree? Nov 28 15:21:52 DomeMaster: well, it might also be the bootloader after trying to verifying the kernel, failing it, and then switch to a recovery mode Nov 28 15:22:46 KanjiMonster: just got response from suppourt, they were indeed assuming that the kernel was on flash. So hopefully they can resolve this issue but it looks like there is another issue as well namely the offset thing Nov 28 15:22:52 I'll keep you updated on my progress Nov 28 15:27:36 nighty-_: make sure the form model is not cached by removing /tmp/luci-modulecache/ Nov 28 15:28:37 ok Nov 28 15:28:53 the code you pasted should not reslt in the uci you quoted Nov 28 15:29:00 jow_laptop, how do I get the name of the interface from within Luci ? Nov 28 15:29:13 which interface do you refer to? Nov 28 15:31:50 this code goes into wifi.lua Nov 28 15:32:03 at line roughly 550 Nov 28 15:32:05 nighty-_: also you need to declare "local value" outside of the if/else condition Nov 28 15:32:17 ah ok Nov 28 15:32:37 like local value; if ... then value = ... else value = ... fi Nov 28 15:33:03 worked :) Nov 28 15:33:06 thanks Nov 28 15:33:43 so now I need to have the interface name ( like wlan0 wlan0-1 or whatever it is ) Nov 28 15:33:49 I'd aslo suggest to make a 3rd mode bgscan_mode = none Nov 28 15:34:04 make the other bgscan options depend on mode simple and mode learn Nov 28 15:34:34 I see Nov 28 15:34:36 and do a self.map:del(section, "bgscan") when mode == "none" Nov 28 15:36:03 ah yes Nov 28 15:36:21 I would rather called it disabled then rather than none Nov 28 15:36:24 ? Nov 28 15:37:17 sure Nov 28 15:37:21 it was just a suggestion Nov 28 15:37:25 as for the wireless ifname Nov 28 15:37:36 you can get it using wnet:ifname() Nov 28 15:37:42 howver that is not stable Nov 28 15:37:56 it might shift from wlan0 to wlan0-1 if another network is added on the radio etc. Nov 28 15:38:18 I suggest using wnet:id() Nov 28 15:38:28 it will give you an identifier like "radio0.network1" Nov 28 15:38:45 that stay the same unless the /etc/config/wireless gets manually reordered by a user Nov 28 15:39:14 the problem is that all of this needs to end up in wpa_supplicant.conf Nov 28 15:39:30 and it does not understand radio0.network1 Nov 28 15:39:32 yeah but the learn file must not strictly carry the name of the wlan iface Nov 28 15:39:37 where do you put the dtb files in the src Nov 28 15:39:42 ok Nov 28 15:39:45 not fidiing them Nov 28 15:39:53 as far as I understood it the learn file just needs to be unique per radio Nov 28 15:40:00 and not change Nov 28 15:40:11 ok then good enough I guess Nov 28 15:40:14 its a kind of cache file Nov 28 15:40:25 (btw not sure if you really want to put that into /etc by default Nov 28 15:40:36 wnet:id() will return a string ? Nov 28 15:40:39 yes Nov 28 15:42:05 root@jj:~# lua -e 'local m = require "luci.model.network"; m.init(); local wnet = m:get_wifinet("wlan0"); print(wnet:id())' Nov 28 15:42:08 radio0.network1 Nov 28 15:42:10 root@jj:~# Nov 28 15:42:34 "wnet" is already provided by wifi.lua Nov 28 15:42:45 it refers to the wireless network currently edited Nov 28 15:45:56 nighty-_: eventually you also need to program hooks for cfgvalue(), this is needed to break the "bgscan" value back up into pieces that are fed to the indivudual fields Nov 28 15:46:16 ok Nov 28 15:46:29 nighty-_: unfortunately I need to run now, maybe I'll be back here tonight, otherwise feel free to ask me about that tomorrow Nov 28 15:46:39 ok Nov 28 15:47:00 the cfgvalue I do not understand why it is needed Nov 28 15:47:08 but you can explain this later Nov 28 15:47:16 cfgvalue() reads the value from uci and feeds it to the field Nov 28 15:47:23 usually one uci option is tied to one field Nov 28 15:47:33 now you tie 4 or 5 fields to one uci option Nov 28 15:47:46 however we only covered the path user input -> uci Nov 28 15:47:54 whats missing is uci -> html form defaults Nov 28 15:48:00 I see Nov 28 15:48:09 not really sure how to do that Nov 28 15:48:26 I assumed that Nov 28 15:48:29 I can tell you later Nov 28 15:48:45 bbl Nov 28 15:50:04 ok Nov 28 16:47:06 where do the dtb files get put Nov 28 16:47:15 in the src Nov 28 16:48:13 arch/arm/boot/*.dtb Nov 28 16:50:25 s/arm/yourarch/g Nov 28 16:50:38 the arch dir in the kernel src or ? Nov 28 16:51:27 build_dir/target-mips_mips32_uClibc-0.9.33.2/linux-atheros/linux-3.10.49/arch/arm/boot/dts/ sort of thing Nov 28 16:51:38 (ignore that I used the paths for two conflicting soc types) Nov 28 16:52:26 ok because I have no build_dir yet Nov 28 16:52:34 we are just working on patching Nov 28 16:52:40 and getting things in place Nov 28 16:53:51 did you mean dts files? Nov 28 16:55:13 dts gets patched and pu tin kernel Nov 28 16:55:18 correct Nov 28 16:56:24 this is so much a pain becuase the docs are all over the place on how to add a board Nov 28 16:57:05 dtb is for boot loader Nov 28 16:57:36 no, dtb is compiled dts. Nov 28 16:57:37 dts is for the seems so much is missing in the 3.14.18 kernel Nov 28 16:57:43 ok Nov 28 16:57:53 which is why I gave you the path out in the build dir :) Nov 28 16:58:14 dts= src dtb=binary Nov 28 16:58:20 ok Nov 28 16:58:24 get it now Nov 28 16:58:57 well we are using head at the min patchng it Nov 28 16:59:10 and there is no build dir Nov 28 16:59:21 yes, because you are presumably patching the dts files. Nov 28 16:59:23 that's fine Nov 28 16:59:33 ok Nov 28 17:00:18 so I have to patch kernel next Nov 28 17:00:28 and then work on the boot loade Nov 28 17:00:36 boot lodaer Nov 28 17:00:38 consider the dts file as just another file in the kernel source Nov 28 17:00:43 this is getting painfull Nov 28 17:01:14 you use the same tools for working on a kernel patch, it's just a .dts file instead of .[ch] Nov 28 17:01:25 ok Nov 28 17:03:42 well am new to openwrt and tying to port a new board the lamobo-r1 routerboard Nov 28 17:04:35 they are currently listed as the bananapi-r1 but they are looking to just use the lamobo name now Nov 28 17:04:59 but I wounder should use the 3.14 or the 3.18 kernel Nov 28 17:05:14 but I have to patch both Nov 28 17:05:19 either way Nov 28 17:45:10 uhmmm Nov 28 17:45:46 jow_laptop, I think there is a strange behaviour in Luci .default Nov 28 17:46:06 jow_laptop, let me paste the code Nov 28 17:47:24 http://pastebin.com/mkY1mpR5 Nov 28 17:48:08 when you change the value to none the fields dependent on bgscan_mode == simple or learn disappear Nov 28 17:49:04 but when you try to save and apply the change is shown in uci configuration but on the luci web interface comes back to "Simple" etc... Nov 28 17:49:33 I think it is because my fields are split and they can be read from file again Nov 28 17:49:53 as you said earlier about the UCI -> LUCI cfgvalues() Nov 28 17:55:26 luka r43422 trunk/target/ (11 files in 2 dirs) * x86: add support for 3.14 Nov 28 18:13:06 build #675 of au1000 is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/675 Nov 28 19:15:00 rmilecki r43423 trunk/package/system/procd/files/nand.sh * procd: fix typo in nand.sh s/TRX/TAR/ Nov 28 19:47:42 florian r43424 packages/utils/open-plc-utils * open-plc-utils: moved to github Nov 28 19:48:04 jow_laptop, http://pastebin.com/n7sPryDn Nov 28 19:48:07 latest code **** ENDING LOGGING AT Sat Nov 29 02:59:59 2014