**** BEGIN LOGGING AT Tue Jun 09 02:59:57 2020
Jun 09 04:01:22 <owrt-snap-builds> build #358 of lantiq/xrx200 is complete: Failure [failed defconfig dltar]  Build details are at http://buildbot.openwrt.org/master/images/builders/lantiq%2Fxrx200/builds/358  blamelist: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Jun 09 04:02:40 <owrt-snap-builds> build #401 of ramips/rt288x is complete: Failure [failed kmodconfig]  Build details are at http://buildbot.openwrt.org/master/images/builders/ramips%2Frt288x/builds/401  blamelist: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Jun 09 04:02:44 <owrt-snap-builds> build #344 of malta/be is complete: Failure [failed defconfig dltar]  Build details are at http://buildbot.openwrt.org/master/images/builders/malta%2Fbe/builds/344  blamelist: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Jun 09 04:34:27 <dengqf6> dwmw2_gone: you have to use vlan filtering
Jun 09 06:57:16 <ldir> don't see how I can be blamed for tar download issues or permissions on /tmp - hmmm
Jun 09 07:23:09 <dwmw2_gone> dengqf6: shouldn't it do that for me automatically if I just add a wan.1 interface to the br-lan bridge?
Jun 09 07:23:41 <dwmw2_gone> I typed 'wan.1' into the interface list for br-lan, now it shows up as 'Software VLAN: "wan.1" in the list.
Jun 09 07:23:51 <dwmw2_gone> in luci
Jun 09 07:23:54 <dwmw2_gone> but doesn't actually *do* it.
Jun 09 07:24:05 <dwmw2_gone> I thought DSA was supposed to make things transparent?
Jun 09 07:34:27 <jow> dwmw2_gone: seems it was overpromised and underdelivered :)
Jun 09 07:35:32 <dwmw2_gone> root@garden:~# ip link add wan name wan.1 type vlan id 1
Jun 09 07:35:32 <dwmw2_gone> ip: RTNETLINK answers: Invalid argument
Jun 09 07:38:53 <dwmw2_gone> that happens even when I plug in the USB Ethernet and do it on eth1. Not a DSA problem.
Jun 09 07:43:16 <dwmw2_gone> although it also happens on my 19.07 Lantiq box where it clearly *is* using VLAN on the vdsl
Jun 09 07:47:40 <dwmw2_gone> aha, the implicit one does work on eth1, the external Ethernet. I just add 'eth1.202' to the br-lan bridge in the interface dropdown, and it creates it
Jun 09 07:50:16 <dwmw2_gone> just not on wan
Jun 09 07:54:15 <ldir> wanders about an init script and comes across '  . "${IPKG_INSTROOT}/lib/functions/network.sh   '  - I've not seen $IPKG_INSTROOT before
Jun 09 07:55:14 <ldir> then reads https://openwrt.org/docs/guide-developer/procd-init-scripts?s[]=ipkg&s[]=instroot and is still confused
Jun 09 07:56:41 <ldir> is it me or should init scripts just simply do (at the top) [ "$IPKG_INSTROOT" ] && exit 0
Jun 09 07:56:54 <jow> it's not clearly definefd
Jun 09 07:57:05 <jow> IPKG_INSTROOT is used for two purposes actually
Jun 09 07:57:20 <jow> most scripts use it to detect a host build environment
Jun 09 07:57:36 <jow> however afair it is also set when the opkg --dest flag is used
Jun 09 07:59:09 <ldir> stumbling across https://github.com/openwrt/packages/pull/11353/commits/39f2e30b41dbfbb7967348d4dc8ce79840a08455 is what started my 'thought' process...and then the rabbit hole happened (as usual)
Jun 09 08:02:24 <ldir> git grep "INSTROOT.*functions" reveals a) all sorts of formatting inconsistencies  b) not many do it - leading to c) why/wth? :-)
Jun 09 08:07:14 <ldir> in packages at least - in base there's only 1... in odhcpd
Jun 09 08:09:40 <ldir> stuff gets cargo culted (often by me!) so it's good to get it right :-)
Jun 09 08:13:19 <jow> I was mistaken, IPKG_INSTROOT is only set when installing packages in the build host environment
Jun 09 08:13:39 <jow> it makes no sense to use it in init scripts
Jun 09 08:14:06 <jow> my guess is that `. "${IPKG_INSTROOT}/lib/functions/network.sh` was done to fix an error in the host build phase
Jun 09 08:14:15 <jow> *image build phase
Jun 09 08:14:33 <jow> the proper fix is moving the sourcing into the init script procedures
Jun 09 08:25:11 <dwmw2_gone> nick[m]: morning.
Jun 09 08:25:31 <nick[m]> hi
Jun 09 08:25:49 <dwmw2_gone> thanks for DAWN. I've been looking for something like this.,
Jun 09 08:27:25 <dwmw2_gone> if I can avoid getting distracted by Real Work (or that Banana Pi R2 which I really ought to get supported in master before we do a release) I'm probably going to play with luci-app-dawn and see if I can make it look more like the 'Associated Stations' on the main overview page. Which gives me host information as well as actual speeds.
Jun 09 08:28:48 <dwmw2_gone> e.g. http://david.woodhou.se/stations.png
Jun 09 08:29:28 <dwmw2_gone> aside from the host information which obviously is out of scope, do we have all the information needed for the rest of that?
Jun 09 08:29:50 <dwmw2_gone> I'll have to get host information locally the same way the overview page does.
Jun 09 08:42:35 <nick[m]> Since that is information parsed from libiwinfo (I'm not totaly sure) that should be easy to do
Jun 09 08:42:39 <ynezz> dwmw2_gone: FYI https://github.com/blogic/utopology and u80211d should probably provide that in the future
Jun 09 08:42:43 <nick[m]> but we have to think about about rate
Jun 09 08:43:20 <ynezz> dwmw2_gone: it's wip, maybe John could provide more details in order to not waste/duplicate effort? :)
Jun 09 08:45:04 <nick[m]> update rate*
Jun 09 08:45:47 <nick[m]> what is this utopolgy project doing?
Jun 09 08:49:46 <nick[m]> dwmw2_gone: I'm just adding the ifname and the hostname
Jun 09 08:49:51 <dwmw2_gone> ta
Jun 09 08:49:54 <dwmw2_gone> that'll help
Jun 09 08:50:15 <dwmw2_gone> I think I'm going to do a generic MAC address lookup too, like wireshark has to print vendor names
Jun 09 08:50:26 <dwmw2_gone> a few places could use that
Jun 09 08:51:51 <ldir> jow: sorry - got diverted into re-arranging the living room rug, table, plants etc.
Jun 09 08:57:40 <nick[m]> "I think I'm going to do a generic MAC address lookup too, like wireshark has to print vendor names" I would suggest doing that in luci :)
Jun 09 08:57:41 <nick[m]> but that would be awesome
Jun 09 09:01:24 <dwmw2_gone> nick[m]: right. I'd do it somewhere generically, and luci-app-dawn (and others) could use it.
Jun 09 09:02:17 <dwmw2_gone> I found it really useful in wireshark the other day when trying to make sense of a tcpdump from a DSA root port, which didn't look like proper 802.1Q so wireshark couldn't display packets sanely, but I could find the one from the right vendor quite easily.
Jun 09 09:03:48 <jow> dwmw2_gone: there currently is this which you might be able to reuse: https://github.com/openwrt/luci/blob/master/applications/luci-app-nlbwmon/htdocs/luci-static/resources/view/nlbw/display.js#L50 + https://github.com/openwrt/luci/blob/master/applications/luci-app-nlbwmon/htdocs/luci-static/resources/view/nlbw/display.js#L156
Jun 09 09:04:28 <jow> repo + generation script here: https://github.com/jow-/oui-database/
Jun 09 09:04:44 <jow> ideally it should be packaged as .ipk so that users can opt into installing it locally
Jun 09 09:05:03 <dwmw2_gone> shiny; thanks
Jun 09 09:06:17 <dwmw2_gone> a perl script to process a file in another repository which in turn was generated by a perl script...
Jun 09 09:06:28 <dwmw2_gone> no, in turn was generated by a python script
Jun 09 09:06:40 <jow> :P
Jun 09 09:07:50 <jow> I needed something compact which was sorted so that a simple binary searhc can be done
Jun 09 09:08:11 <jow> a binary format would probably be slightly smaller still but thats hard to work with in JS
Jun 09 09:09:21 <dwmw2_gone> [
Jun 09 09:09:22 <dwmw2_gone>    "0",
Jun 09 09:09:22 <dwmw2_gone>    24,
Jun 09 09:09:22 <dwmw2_gone>    "Officially Xerox, but 0:0:0:0:0:0 is more common",
Jun 09 09:09:22 <dwmw2_gone>    "1000000",
Jun 09 09:09:24 <dwmw2_gone>    24,
Jun 09 09:09:24 <dwmw2_gone>    "Xerox Corporation",
Jun 09 09:09:27 * dwmw2_gone blinks
Jun 09 09:09:30 <dwmw2_gone> I was expecting more []
Jun 09 09:09:37 <jow> its a flat array
Jun 09 09:09:48 <dwmw2_gone> [
Jun 09 09:09:48 <dwmw2_gone>    ["0",
Jun 09 09:09:48 <dwmw2_gone>    24,
Jun 09 09:09:48 <dwmw2_gone>    "Officially Xerox, but 0:0:0:0:0:0 is more common",],][  "1000000",
Jun 09 09:09:48 <dwmw2_gone>    24,
Jun 09 09:09:48 <dwmw2_gone>    "Xerox Corporation",
Jun 09 09:09:51 <dwmw2_gone> ]
Jun 09 09:10:05 <dwmw2_gone> or something more competently edited in hexchat on my side :)
Jun 09 09:10:22 <jow> three items are one record
Jun 09 09:10:25 <dwmw2_gone> yeah
Jun 09 09:15:02 <ldir> jow: is https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=commit;h=3150474988223155f82186ad883e9ea522264a18 fair/reasonable?
Jun 09 09:15:04 <dwmw2_gone> I'd just be doing it in lua though, wouldn't I? Or perhaps a ubus call
Jun 09 09:15:39 <jow> dwmw2_gone: an ubus call makes sense - you could implement it as rpcd plugin
Jun 09 09:16:00 <jow> dwmw2_gone: and it should probably support looking up many MAC addresses at once
Jun 09 09:16:12 <jow> to avoid too many roundtrips from the ui
Jun 09 09:16:57 <jow> ldir: yep, looks sensible
Jun 09 09:18:35 <ldir> I'll send it in for that nice Mr Dedecker to Ack.
Jun 09 09:20:19 <ldir> I still think that 1st/2nd line "[ "$IPKG_INSTROOT" ] && exit 0" is neater/consumes less space than multiple source directives in each function.
Jun 09 09:22:09 <jow> that will break init script activation in the image build phase
Jun 09 09:22:21 <jow> keep in mind that init scripts actually are executed at build time
Jun 09 09:22:53 <jow> so they must not do target specific things (like sourcing /lib/functions) or bail out (like prematurely calling exit 0)
Jun 09 09:23:05 <jow> otherwise /etc/init.d/XXX enable will fail
Jun 09 09:24:51 <ldir> I sorta get that - or I've seen the warning at https://openwrt.org/docs/guide-developer/procd-init-scripts?s[]=ipkg&s[]=instroot#procd_init_script_parameters at least
Jun 09 09:25:46 <ldir> ah, yes I get you - just for 'enable/disable'. hmmm
Jun 09 09:27:10 <jow> I guess all that could be made more robust by intelligently parsing out START= and STOP= from the script text and placing the corresponding .../etc/rc.d/ symlinks manually
Jun 09 09:28:05 <jow> this way the init script must not actually be run to get activated
Jun 09 09:28:42 <ldir> A guard such as ' if IPKG_INSTROOT set and (parameter != start & parameter != stop) exit 0'
Jun 09 09:36:38 <nick[m]> dwmw2_gone: I mentioned u here. https://github.com/openwrt/luci/pull/4146
Jun 09 09:37:12 <nick[m]> is that okay? can u have a look at this how u want that to be displayed?
Jun 09 09:37:33 <dwmw2_gone> will do; thanks
Jun 09 09:38:38 <jow> nick[m]: any reason why you define a separate MAX_INTERFACE_NAME ?
Jun 09 09:39:24 <jow> nick[m]: I think it can never exceed IFNAMSIZ which is 16 bytes iirc
Jun 09 09:40:19 <jow> nick[m]: and regarding the hostname, not sure if you should prefer /proc/sys/kernel/hostname over whatever is set in uci atm
Jun 09 09:41:27 <nick[m]> jow: I think I can remove MAX_INTERFACE_NAME and search for the equivalent in limits.h. Not sure why I added this a long time ago.
Jun 09 09:42:35 <nick[m]> jow: what is the problem with getting hostname from uci?
Jun 09 09:43:14 <jow> it is not strictly a problem, more a philosophical question
Jun 09 09:43:54 <jow> my reasoning was that the intent is to report the current state (effective hostname) not the desired state (what is written in a config but not necessarily applied yet)
Jun 09 09:44:10 <nick[m]> in general I have to tidy up the code and the printf statements and stuff like this...
Jun 09 09:44:22 <jow> in most cases there won't be a difference
Jun 09 09:45:05 <nick[m]> jow: hmmm. actually I favor uci because I'm doing everything with uci. I understand your point. I'm not sure.
Jun 09 09:50:50 <andy2244> Hey how do i add a lib depends that needs kernel 5.x ?
Jun 09 10:00:07 * ldir suddenly gets what jow is on about and feels very stupid
Jun 09 10:01:16 <ldir> insufficient sleep = shouldn't attempt to re-write code
Jun 09 10:13:40 <dwmw2_gone> nick[m]: I see garbage in 'iface' from remote systems which I haven't updated
Jun 09 10:13:48 <dwmw2_gone> 	"Lumleys6": {
Jun 09 10:13:48 <dwmw2_gone> 		"1A:62:2C:5D:94:6C": {
Jun 09 10:13:53 <dwmw2_gone> ...
Jun 09 10:13:54 <dwmw2_gone> 			"neighbor_report": "2c5d946cef090000510b070603000b00",
Jun 09 10:13:55 <dwmw2_gone> 			"iface": "�k�w���",
Jun 09 10:13:55 <dwmw2_gone> 			"hostname": ""
Jun 09 10:14:16 <dwmw2_gone> that's 'ubus call dawn get_network' output
Jun 09 10:18:17 <nick[m]> dwmw2_gone: I have to "\0" the char buffers, or memset them, if devices won't send me a hostname
Jun 09 10:18:32 <nick[m]> did not thought about "backwards compatbility"
Jun 09 10:18:35 <dwmw2_gone> s/,.*//
Jun 09 10:18:42 <dwmw2_gone> You are receiving data from the network. Don't trust it.
Jun 09 10:18:58 <dwmw2_gone> This isn't strictly a backwards compatibility thing.
Jun 09 10:19:24 <dwmw2_gone> You were listening to port 1026 on a public IP address and accepting data :)
Jun 09 10:19:28 <dwmw2_gone> I switched to multicast now :)
Jun 09 10:22:14 <nick[m]> dwmw2_gone: yeah that security stuff is really another thing I have to look at. I made some basic symmetric encryption, but I hope if people deploy that in a public network, they have some control channel, or some seperated vlan (yeah I know vlans are not for security....)
Jun 09 10:22:16 <dwmw2_gone> Before switching, I started looking at an access control setup, where we only accept incoming connections from the local network. Or from hosts we've discovered and connected to.
Jun 09 10:23:25 <dwmw2_gone> I was just pondering setting up vlans. I only have 64 public Legacy IP addresses at home, and it's getting a bit silly when all the TVs and even my watch have one each.
Jun 09 10:23:32 <dwmw2_gone> Some of those really *can* be stuck behind NAT.
Jun 09 10:23:42 <dwmw2_gone> but only if I can get OpenWRT to actually do VLAN
Jun 09 10:25:26 <ynezz> s/OpenWrt/DSA/
Jun 09 10:25:38 <dwmw2_gone> true
Jun 09 10:26:22 <dwmw2_gone> I just heckled at the 'convert lantiq xrx200 to DSA' PR on that topic :)
Jun 09 10:26:26 <dwmw2_gone> since it *does* work there currently.
Jun 09 10:26:38 <dwmw2_gone> just not on *all* the other APs.
Jun 09 10:28:06 <ynezz> yeah, just rebasing onto 5.4.45 and noticed few upstream fixes for DSA
Jun 09 10:29:13 <ynezz> mt7530 related fixes
Jun 09 10:29:35 <dwmw2_gone> the 'enable learning for the cpu port' one already bit me :)
Jun 09 10:29:42 <dwmw2_gone> but we ingested that already.
Jun 09 10:29:50 <dwmw2_gone> was fixed in master by the time I'd worked it out
Jun 09 10:36:37 <dwmw2_gone> What is the policy for version upgrades in packages in a stable (specifically 19.07) release?
Jun 09 10:37:13 <dwmw2_gone> Should I leave Domoticz on v4.9700 even if a newer version does fix problems?
Jun 09 10:38:39 <ynezz> depends on the number of changes (likely some of them can introduce regression) so backporting just the fix should be preferred
Jun 09 10:39:13 <ynezz> but since you're the maintainer of the package, you decide I would say, but I don't know all the rules in that package feeds repo :]
Jun 09 10:39:58 <dwmw2_gone> I know of three users. Myself (and I've already updated). stintel (who doesn't use it any more), and the user in #12259 who is having problems and I'm giving the new version to test.
Jun 09 10:50:03 <ynezz> and I've thought, that home automation is current new craze :)
Jun 09 10:50:31 <dengqf6> dwmw2_gone: nope
Jun 09 10:52:12 <dengqf6> bridge vlan add vid 1 dev wan tagged
Jun 09 10:53:07 <dengqf6> dwmw2_gobe: ^run this after adding wan to br-lan and enabling filtering
Jun 09 10:53:25 <dengqf6> ip link set br-lan type bridge vlan_filtering 1
Jun 09 10:54:13 <dwmw2_gone> ip: either "dev" is duplicate, or "type" is garbage
Jun 09 10:54:23 <stintel> ynezz: the combination home automation and openwrt isn't very common
Jun 09 10:54:46 <dwmw2_gone> I don't really want to encourage the use of domoticz.db on internal flash :)
Jun 09 10:54:50 <stintel> most people I know using open source home automation stuff are running it in docker
Jun 09 10:54:50 <dwmw2_gone> I use external storage for mine
Jun 09 10:55:43 <stintel> mine used to be on the SSD in my APU@
Jun 09 10:55:46 <stintel> APU2*
Jun 09 10:56:07 <dengqf6> dwmw2_gone: you probably need ip-full
Jun 09 10:56:21 <dwmw2_gone> dengqf6: so if I want three different vlans on the physical network, bridge vlan add vid [123] dev wan tagged.
Jun 09 10:56:33 <dwmw2_gone> then add wan.1 to (e.g.) br-lan1, wan.2 to br-lan2, etc. ?
Jun 09 10:57:07 <dengqf6> dwmw2_gone: what is your previous swconfig settings?
Jun 09 10:57:34 <dwmw2_gone> this is a new AP
Jun 09 10:57:54 <dwmw2_gone> the main router is the lantiq box with swconfig, and just adding 'eth0.1.1' to the br-lan device list seemed to do it all magically.
Jun 09 10:58:03 <dwmw2_gone> I can add 'eth0.1.2' to br-lan6, etc.
Jun 09 10:58:29 <dengqf6> dwmw2_gone: q-in-q?
Jun 09 10:58:35 <dwmw2_gone> yes
Jun 09 10:58:46 <dwmw2_gone> 'eth0.1' on the XRX200 is the LAN ports.
Jun 09 10:59:06 <dwmw2_gone> eth0.1.1 is sending 802.1q (vid 1) out the LAN port on the wire.
Jun 09 10:59:21 <dwmw2_gone> so not q-in-q except *internally* to the box.
Jun 09 10:59:50 <dengqf6> No other swconfig settings?
Jun 09 11:00:17 <dwmw2_gone> on the lantiq, just the defaults,
Jun 09 11:00:29 <dengqf6> so you don't even need vlan filtering
Jun 09 11:00:31 <dwmw2_gone> vlan1 to ports 0 1 2 4 6t
Jun 09 11:00:42 <dwmw2_gone> vlan2 to ports 5 6t (which is its 'wan' port)
Jun 09 11:00:55 <dwmw2_gone> so eth0.1 is the lan ports, eth0.2 is the wan port on the back of hte box.
Jun 09 11:01:10 <dwmw2_gone> hence if I want to *send* 802.1q, I use eth0.1.1 etc.
Jun 09 11:01:13 <dengqf6> Create a new interface, in physical setting: set in to "@lan.1"
Jun 09 11:01:15 <dwmw2_gone> which all Just Works
Jun 09 11:01:35 <dwmw2_gone> that's what worked on the lantiq with swconfig not DSA
Jun 09 11:01:38 <rmilecki> noltari: nice work!
Jun 09 11:01:48 <dwmw2_gone> didn't work on the RT-AC85P. Although I can try again now I have ip-full installed.
Jun 09 11:01:53 <nick[m]> dwmw2_gone: the uninitalized string buffers should be gone in https://github.com/berlin-open-wireless-lab/DAWN/commit/77619e9b18c1f4370ee1fd16c1d383d3005602c5
Jun 09 11:01:53 <dengqf6> as long as you don't bridge untagged and tagged ports together
Jun 09 11:02:19 <dengqf6> dwmw2_gone: Have you tried the "@lan.1" interface?
Jun 09 11:02:29 <dwmw2_gone> I tried wan.1 not @wan.1
Jun 09 11:02:47 <dengqf6> Is it bridged to br-lan?
Jun 09 11:03:33 <dwmw2_gone> That's *all* I've done. In the br-lan settings, where I can type arbitrary text into the 'Interface' list, added 'wan.1'
Jun 09 11:03:38 <dwmw2_gone> and now '@wan.1'
Jun 09 11:03:45 <dwmw2_gone> that worked on the Lantiq box
Jun 09 11:03:48 <dwmw2_gone> not the one using DSA
Jun 09 11:04:52 <dengqf6> dwmw2_gone: bridge all lan and wan to br-lan
Jun 09 11:05:04 <dwmw2_gone> that I already had.
Jun 09 11:05:15 <dengqf6> enable vlan filtering then set wan to tagged vlan 1
Jun 09 11:05:32 <dengqf6> do NOT use wan.1 anymore
Jun 09 11:08:02 <dwmw2_gone> Ok, I removed wan.1 from br-lan through luci, applied.
Jun 09 11:08:09 <dwmw2_gone> then
Jun 09 11:08:09 <dwmw2_gone> root@garden:~# ip link set br-lan type bridge vlan_filtering 1
Jun 09 11:08:09 <dwmw2_gone> root@garden:~# bridge vlan add vid 1 dev wan tagged
Jun 09 11:08:27 <dwmw2_gone> I can still talk to it. Which, since I'm still sending untagged packets through its WAN port, probably means it isn't working?
Jun 09 11:08:36 <dwmw2_gone> I'd have to change the *other* end too, and I haven't yet
Jun 09 11:10:03 <dwmw2_gone> nick[m]: thanks. Will update and make luci-app-dawn use it. If it works for me are you OK with me pushing the package update?
Jun 09 11:10:20 <dwmw2_gone> (for dawn itself; we'll do more review for my luci-app-dawn changes)
Jun 09 11:10:55 <nick[m]> yeah! actually I hoped for it because then I don't have to do an extra PR ;)
Jun 09 11:11:05 <dwmw2_gone> ok, thanks
Jun 09 11:13:09 <dengqf6> add wan to br-lan
Jun 09 11:13:20 <dwmw2_gone> it already is
Jun 09 11:14:01 <dwmw2_gone> I'm connected through it (with no vlan yet).
Jun 09 11:14:34 <dwmw2_gone> I can move that wire. But if I have wan and wan.1 both in br-lan that ought to work and I can reconfigure the other end *afterwards* to switch from untagged to vid1
Jun 09 11:21:34 <dwmw2_gone> nick[m]: +        strncpy(hostname, ptr.o->v.string, HOST_NAME_MAX);
Jun 09 11:22:18 <dwmw2_gone> that doesn't necessarily nul-terminate it
Jun 09 11:22:26 <dwmw2_gone> welcome to the "I hate strncpy" club...
Jun 09 11:22:37 <dwmw2_gone> while we're at it can we truncate at the first . ?
Jun 09 11:22:39 <dwmw2_gone> if there is one
Jun 09 11:24:31 <nick[m]> dwmw2_gone: strlcpy?
Jun 09 11:27:17 <jow> snprintf(hostname, HOST_NAME_MAX, "%s", ptr.o->v.string)
Jun 09 11:27:32 <jow> or strlcpy() if it happens to be available
Jun 09 11:27:39 <jow> it is on musl but not glibc iirc
Jun 09 11:28:37 <nick[m]> hmmm snprintf sonuds good
Jun 09 11:39:38 <plntyk> I readded and slightly update talks and slides from old wiki https://openwrt.org/about/talks - somebody with admin access might want to add a link from about page as its not linked to any content atm
Jun 09 12:13:46 <ynezz> plntyk: nice, you're better discussing this on forum, main communication channel for docs https://forum.openwrt.org/c/documentation/
Jun 09 12:14:39 <dwmw2_gone> nick[m]: hm, I seem to still have garbage in iface
Jun 09 12:15:13 <dwmw2_gone> oh, because I didn't include that patch :)
Jun 09 13:06:21 <dwmw2_gone> nick[m]: I had added separate Hostname and Interface columns, in https://github.com/dwmw2/luci/commits/dawn
Jun 09 13:06:43 <dwmw2_gone> But now seeing what you did, I've tweaked it slightly to put it all in the 'AP' column
Jun 09 13:06:50 <dwmw2_gone> https://github.com/dwmw2/luci/commits/dawn2
Jun 09 13:06:54 <dwmw2_gone> take whichever you prefer
Jun 09 13:07:11 <dwmw2_gone> It adds the ($hostname $iface) in parens the AP, *if* they exist.
Jun 09 13:07:42 <dwmw2_gone> I think I prefer my original though, although it takes more horizontal space.
Jun 09 13:08:04 <dwmw2_gone> for the new separate columns
Jun 09 13:18:12 <dwmw2_gone> Especially as I plan to screw with the MAC address display, as discussed.
Jun 09 14:31:27 <rr123> why ca
Jun 09 14:31:59 <rr123> why is luci not using the default bootstrap css instead of creating everything manual in cascade.css
Jun 09 14:38:36 <jow> rr123: because it existed before the bootstrap theme was added
Jun 09 14:39:04 <jow> to use the default bootstrap css you need to change all the luci markup and all the other themes to use the same css calsses and structure
Jun 09 14:41:45 <rr123> jow: make sense, so it's for historical reasons then. new to luci and all those internal css etc are intimidating to me, the vue-based oui looks simpler but you have to learn vue, would like to have a pure-bootstrap theme so others can hack luci easier
Jun 09 14:42:29 <rr123> by the way openwrt-2020 looks modern and works well on cell phones too
Jun 09 14:43:52 <rr123> one more question you mentioned luci will go 100% js client does it mean, the lua-template <% %> will be gone in the future? e.g. no server side rendering at all.
Jun 09 14:46:23 <RobertP> I did two changes to openWRT and I'd like to commit these changes for review - is this described somewhere?
Jun 09 14:46:46 <jow> rr123: correct
Jun 09 14:47:04 <RobertP> One is to add OpenVPN out of the box plugins (auth-pam & root-iface-down) second one is to add SYNPROXY support (kernel modules & iptables)
Jun 09 14:49:53 <rr123> that might make luci one day REST-api ready and let the client side to handle the rest, at the moment luci is stil tightly coupled with the system. Thanks for the pointers.
Jun 09 14:50:20 <ldir> RobertP: https://openwrt.org/submitting-patches
Jun 09 14:50:44 <rr123> the core developers really should co-author an openwrt book :)
Jun 09 14:51:11 <RobertP> @ldir: thanks
Jun 09 16:10:36 <dwmw2_gone> OK...
Jun 09 16:10:43 <dwmw2_gone> # ubus -v call macvendor get_multi '{"macs":{"one":"","two":""}}}'
Jun 09 16:10:43 <dwmw2_gone> {
Jun 09 16:10:43 <dwmw2_gone> 	"one": "unknownone",
Jun 09 16:10:43 <dwmw2_gone> 	"two": "unknowntwo"
Jun 09 16:10:46 <dwmw2_gone> }
Jun 09 16:10:50 <dwmw2_gone> I'll make it return actual data later.
Jun 09 16:11:07 <karlp> are you including an oui db now?
Jun 09 16:11:43 <dwmw2_gone> right now it just says "unknown" but that's the plan
Jun 09 16:12:23 <dwmw2_gone> first I want to work out why http://david.woodhou.se/dawn-show-mac-vendor.patch fails when I make the actual call
Jun 09 16:12:42 <karlp> is this for a ui to help people turn macs into station names? do things not normally provide names to dhcp already? I normally see things like "wif's iphone" and so on?
Jun 09 16:13:22 <dwmw2_gone> no, this is MAC to hardware vendor like Wireshark does.
Jun 09 16:13:39 <karlp> right, obviously, but what's the end use?
Jun 09 16:14:01 <dwmw2_gone> when displaying MAC addresses in dawn etc. I'd like to display the vendor with them.
Jun 09 16:14:04 <karlp> because oui vendor can lead to interestingly misleading names.
Jun 09 16:14:19 <dwmw2_gone> this is only under the MAC itself.
Jun 09 16:15:22 <dwmw2_gone> so if you look at http://david.woodhou.se/dawn-with-nicknames.png where I have 'lantiq.infradead.org wlanX' in parens in the 'MAC' column. I've done that differently and have separate Hostname and Interface columns
Jun 09 16:15:43 <dwmw2_gone> In parens in the MAC column instead I've have the OUI vendor name
Jun 09 16:18:15 <dwmw2_gone> but mostly right now I just want to know why it doesn't bloody work :)
Jun 09 16:21:13 <dwmw2_gone> oh, maybe acl on the macvendor service
Jun 09 16:31:05 <dwmw2_gone>         "luci-app-dawn": {
Jun 09 16:31:05 <dwmw2_gone>                 "description": "Grant UCI access for luci-app-dawn",
Jun 09 16:31:05 <dwmw2_gone>                 "read": {
Jun 09 16:31:05 <dwmw2_gone>                         "uci": [ "dawn" ],
Jun 09 16:31:07 <dwmw2_gone>                         "ubus": { "macvendor": [ "get_multi" ] }
Jun 09 16:31:07 <dwmw2_gone>                 },
Jun 09 16:31:10 <dwmw2_gone> doesn't fix it
Jun 09 16:31:58 <jow> dwmw2_gone: did you do an /etc/init.d/rpcd reload?
Jun 09 16:32:03 <dwmw2_gone> restart
Jun 09 16:33:09 <jow> do you get a permission denied error?
Jun 09 16:34:09 <dwmw2_gone> that would be useful
Jun 09 16:34:21 <karlp> dawn is only useful if you have multiple openwrt APs? or is it useful for a single device situation as well?
Jun 09 16:34:45 <dwmw2_gone> I get http://david.woodhou.se/dawn-denied.txt
Jun 09 16:35:09 <dwmw2_gone> karlp: for a single device setup, the 'Associated Stations' on the main luci overview is a little bit nicer IMHO
Jun 09 16:35:15 <jow> dwmw2_gone: ah ok, that looks like a dawn specific problem
Jun 09 16:35:21 <dwmw2_gone> dawn is nice because it shows you that for *multiple* APs
Jun 09 16:35:55 <dwmw2_gone> jow: that goes away if I comment out the single line from david.woodhou.se/dawn-show-mac-vendor.patch that does the ubus call
Jun 09 16:36:37 <dwmw2_gone> I've explicitly set macnames={} after the call (throwing away the results) too, FWIW. Just in case
Jun 09 16:39:18 <dwmw2_gone> Oh, I fail lua 101
Jun 09 16:39:23 <dwmw2_gone> 		local macargs = {}
Jun 09 16:39:24 <dwmw2_gone> 		macargs['macs'] = macnames
Jun 09 16:39:24 <dwmw2_gone> 		macnames = utl.ubus("macvendor", "get_multi", macargs)
Jun 09 16:39:25 <dwmw2_gone> that works
Jun 09 16:39:25 <jow> dwmw2_gone: ah, you triggered and error in the error handler :P
Jun 09 16:39:49 <jow> template.lua:76 is actually printing an error message that the template could not be evaluated
Jun 09 16:39:50 <dwmw2_gone> I can't just do macnames = utl.ubus("macvendor", "get_multi", { "macs" : macnames } )
Jun 09 16:39:55 <dwmw2_gone> haha
Jun 09 16:40:17 <dwmw2_gone> OK, It's now printing my bogus "unknown" results from the non-existent OUI database. Will fix that now :)
Jun 09 16:40:27 <jow> and since the dawn template is a string iirc, the filename "name" value is nil
Jun 09 16:40:41 <jow> will fix that
Jun 09 16:45:47 <dwmw2_gone> thanks
Jun 09 16:49:29 <karlp> dwmw2_gone: you could do utl.ubus("macvendor", "get_multi", { macs=macnames } ) to get your table made inline
Jun 09 16:50:38 <dwmw2_gone> better; thanks.
Jun 09 16:54:22 <karlp> (unless the key is an illegal identifer, macargs.macs = macnames woudl have also worked...
Jun 09 16:55:02 <karlp> (which always traps me when I try and do that in python)
Jun 09 16:56:11 <dwmw2_gone> what will utl.ubus return when it fails?
Jun 09 16:56:24 <dwmw2_gone> I made it call something that doesn't exist
Jun 09 16:56:26 <dwmw2_gone> 		macnames = utl.ubus("macvedndor", "get_multi", { macs = macnames })
Jun 09 16:56:26 <dwmw2_gone> 		if macnames is nil then
Jun 09 16:56:26 <dwmw2_gone> 		   macnames = {}
Jun 09 16:56:27 <dwmw2_gone> 		end
Jun 09 16:56:36 <dwmw2_gone> I get the same unhelpful failure
Jun 09 16:56:43 <dwmw2_gone> *unconditionally* setting macnames to {} afterwards works OK
Jun 09 16:57:07 <dwmw2_gone> but whatever it is, it isn't nil.
Jun 09 16:57:28 <karlp> been a while, but doesn't it do classic multireturn?
Jun 09 16:57:37 <karlp> status, data = call()
Jun 09 16:57:54 <karlp> if status then data = useful stuff else data = error description
Jun 09 16:58:15 <dwmw2_gone> ah, no I fail lua 101 again
Jun 09 16:58:19 <dwmw2_gone> if macnames == nil then
Jun 09 16:58:20 <dwmw2_gone> not
Jun 09 16:58:25 <dwmw2_gone> if macnames is nil then
Jun 09 16:58:29 <karlp> just do "if macnames" then
Jun 09 16:58:58 <karlp> are you running uhttpd in the foreground so you can see stderr?
Jun 09 16:59:02 <karlp> I find that super helpful :)
Jun 09 16:59:12 <dwmw2_gone> I wasn't.
Jun 09 16:59:48 <karlp> I have a script aht symlinks all my dev work and then prints a uhttpd command line to run :)
Jun 09 16:59:51 <karlp>  /usr/sbin/uhttpd -f -h /www -x /cgi-bin -u /ubus -A 1 -n 3 -R -p 0.0.0.0:80
Jun 09 17:00:16 <karlp> (because I never remeber the right invocation of xargs to grab the current running params of uhttpd and rerun them)
Jun 09 19:15:47 <anonanoob> Hi devs, trying to dl a file getting 403 forbidden, it's been verified on $openwrt by another party as well; can you please let me know if this will be posted?  Starting from https://openwrt.org/toh/views/toh_fwdownload?dataflt%5BBrand*%7E%5D=asus&dataflt%5BModel*%7E%5D=RT-AC58 selecting
Jun 09 19:15:48 <anonanoob> https://downloads.openwrt.org/snapshots/targets/ipq40xx/generic/openwrt-ipq40xx-generic-asus_rt-ac58u-initramfs-fit-uImage.itb 403's.
Jun 09 19:57:19 <DeX77> similar issue like anonanoob but with https://downloads.openwrt.org/snapshots/targets/ipq806x/generic/
Jun 09 19:57:29 <DeX77> sha256sum doesn't match with file
Jun 09 20:01:24 <DeX77> interesting: using some mirror directly yields correct sha256sum
Jun 09 20:24:17 <anonanoob> flash/ram here https://openwrt.org/toh/views/toh_admin_low-flash-ram
Jun 09 22:06:03 <f00b4r0> xback: pling?
Jun 09 23:11:41 <owrt-snap-builds> build #402 of ramips/rt288x is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org/master/images/builders/ramips%2Frt288x/builds/402
Jun 10 00:01:44 <ianc> I've just compiled a build off master - it's about two weeks newer than my previous build. procd is in a crashloop at boot. Enabling debugging, it's telling me that the ubus instance is failing, with an error every second or so that "Instance ubus::instance1 exit with error code 65280 after 0 seconds". Anyone know where I can look in the code to
Jun 10 00:01:45 <ianc> find the meaning of error code 65280?
Jun 10 00:03:17 <ianc> this is before the init process is complete
Jun 10 00:04:04 <ianc> the error is produce by service/instance.c
Jun 10 00:04:18 <ianc> error message that is
Jun 10 00:05:03 <owrt-snap-builds> build #359 of lantiq/xrx200 is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org/master/images/builders/lantiq%2Fxrx200/builds/359
Jun 10 00:18:43 <owrt-snap-builds> build #345 of malta/be is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org/master/images/builders/malta%2Fbe/builds/345
Jun 10 01:49:38 <lmore377> I'm trying to port openwrt to a new device (Askey RAC2V1K).
Jun 10 01:53:57 <lmore377> It has an IPQ8065 processor so I decided to use the Netgear R7500 as a base. I already managed to dump the original device tree and got it booting (got leds working and wifi sorta works if I use the firmware from the original rom) but now I'm trying to get the switch (QCA8337) working but I can't figure out where to define how the switch is
Jun 10 01:53:58 <lmore377> connected. Can I get any help with this?
Jun 10 02:03:47 <lmore377> Also I had to change some stuff in the qcom-ipq8064.dtsi file to get the router to not crash because of the regulators. How can I put those in the dts file I'm making instead?
**** ENDING LOGGING AT Wed Jun 10 02:59:57 2020