| 02:24.04 | gtjoseph | file: you around for an outbound_registration question? |
| 02:24.21 | gtjoseph | related to ransport reload |
| 06:08.45 | *** join/#asterisk-dev goddva (~goddva@77.40.154.242) |
| 06:52.55 | *** join/#asterisk-dev zerohalo (~zerohalo@c-174-62-142-83.hsd1.ma.comcast.net) |
| 06:57.47 | *** join/#asterisk-dev bof22 (~Thunderbi@185.13.183.107) |
| 07:30.31 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 08:15.56 | *** join/#asterisk-dev pchero_work (~pchero@109.70.54.56) |
| 08:17.22 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 08:17.22 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 11:42.09 | file | I wasn't then, but now I exist |
| 12:50.56 | *** join/#asterisk-dev oej (~oej@130.28.1.50) |
| 12:50.56 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 13:25.40 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@216-191-106-163.dedicated.allstream.net) |
| 13:40.10 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-kjxbquxwzjlfmpdn) |
| 13:40.10 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 14:07.12 | *** join/#asterisk-dev putnopvut (~putnopvut@asterisk/master-of-queues/mmichelson) |
| 14:07.13 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 14:28.02 | *** join/#asterisk-dev malcolmd (malcolmd@pdpc/sponsor/digium/malcolmd) |
| 14:28.02 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 14:40.09 | *** join/#asterisk-dev zerohalo (~zerohalo@c-174-62-142-83.hsd1.ma.comcast.net) |
| 14:56.46 | *** join/#asterisk-dev WIMPy (~wimpy@e183095026.adsl.alicedsl.de) |
| 15:04.20 | *** join/#asterisk-dev oej (~oej@195.149.146.45) |
| 15:04.20 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 15:13.21 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-rjmwdvmwwvrpgzao) |
| 15:33.53 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-hbvehucqnjqovfuv) |
| 15:33.54 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 15:42.23 | *** join/#asterisk-dev oej (~oej@195.149.146.45) |
| 15:42.23 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 16:16.09 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-heilxxkjicjxjpfu) |
| 16:16.21 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-lrbvsoklbonuhpbq) |
| 16:16.46 | seanbright | putnopvut: i'm not following your question |
| 16:17.03 | seanbright | in my use case there are only host candidates |
| 16:17.18 | seanbright | no srflx or relay |
| 16:17.33 | putnopvut | So if there are no srflx candidates, then how do you not end up sending duplicate hosts? |
| 16:18.11 | putnopvut | Let's say your machine has two IPs, a public (1.2.3.4) and private (10.0.0.1). I would suspect that we would by default advertise 1.2.3.4 and 10.0.0.1 as host candidates |
| 16:18.29 | putnopvut | If you replace the 10.0.0.1 with 1.2.3.4, then wouldn't we end up advertising 1.2.3.4 twice as a result? |
| 16:18.33 | seanbright | correct, but that is not a common setup as far as i am concerned |
| 16:18.41 | putnopvut | What is your setup then? |
| 16:18.47 | seanbright | asterisk has a single private IP |
| 16:18.57 | seanbright | router has a static NAT configured |
| 16:19.24 | putnopvut | Right, then the public address of the router would be a srflx candidate, not a host candidate. |
| 16:19.49 | seanbright | ... |
| 16:19.59 | seanbright | now i'm confused |
| 16:20.06 | seanbright | externaddr in sip.conf |
| 16:20.11 | seanbright | what does it do? |
| 16:20.21 | seanbright | it allows asterisk to "pretend" to be on the edge of the network |
| 16:20.36 | seanbright | that is all my patch aims to do, but for ice negoitation |
| 16:20.48 | seanbright | so asterisk doesn't have a public IP assigned to it, but it is pretending that it does |
| 16:21.19 | seanbright | because the one-to-one nat is effectively the same thing as having the public IP assigned directly to the asterisk box |
| 16:21.23 | putnopvut | That is correct. The thing with ICE is that the protocol differentiates between an address on the host itself vs. an address on the edge of a NAT network. This falls into the weighting algorithms. |
| 16:21.49 | putnopvut | With SIP externaddr, SIP doesn't care about where the actual address is. We just say one. ICE actually does care. |
| 16:22.09 | seanbright | but... i don't care that ICE cares |
| 16:22.10 | seanbright | heh |
| 16:22.29 | putnopvut | Like I said before, I doubt that misadvertising the type of address would actually cause a problem. I just figured it didn't hurt to ask in case someone said "Actually, there is one..." |
| 16:22.40 | seanbright | i fail to see how the externally visible behavior is any different |
| 16:22.53 | file | I would say this is a fuzzy area. Because it's a 1 to 1 despite the IP not being on the server but on the edge, it's associated with that server despite not being host. It shoulddddddddd all work fine. |
| 16:24.17 | seanbright | putnopvut: in your mind you would rather see the host candidates removed and synthetic srflx candidates added? |
| 16:24.25 | putnopvut | Yeah, I agree about the fuzziness here. Despite the fact that the address is at the edge, if it's 1-to-1, it's practically a host candidate anyway |
| 16:24.47 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-bwnmkssevvesikqr) |
| 16:24.47 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 16:24.56 | putnopvut | It's what the protocol rigidity side of my brain says to do, but in practice, I think it may not matter. |
| 16:25.37 | putnopvut | I'll go ahead and mark on the review that it's probably fine the way it is, and I'll give the actual content another look. |
| 16:25.59 | seanbright | i'm happy to add a big red warning in the configuration |
| 16:26.20 | seanbright | you don't want both the mapping and stun and/or turn enabled at the same time |
| 16:26.57 | putnopvut | That sounds like a good idea. |
| 16:27.08 | putnopvut | Not sure how you'll make it red though ;) |
| 16:27.30 | seanbright | metaphorically speaking |
| 16:27.39 | file | red hilighter on your monitor! |
| 17:47.23 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 17:47.23 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 18:24.41 | *** join/#asterisk-dev DivideBy0 (~DivideBy0@unaffiliated/divideby0x0) |
| 18:50.18 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-snuanyyodepldurj) |
| 18:50.18 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 18:51.45 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 18:51.45 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 18:59.18 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 18:59.18 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 19:41.11 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-179-40-172.cust.bezeqint.net) |
| 19:41.11 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 19:58.31 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-hltukllbnqtfkxmp) |
| 19:58.31 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 21:17.35 | gtjoseph | rmudgett: did you actually try the makefile patch? |
| 21:18.48 | rmudgett | I did a distclean, removed the '-' on the include line, and did another distclean |
| 21:19.20 | rmudgett | Since makeopts was no longer present I only got a file not found message but it still tried to do the distclean |
| 21:19.34 | gtjoseph | it's not supposed to prevent a distclean. |
| 21:19.45 | gtjoseph | it's supposed to prevent an uninstall |
| 21:20.19 | rmudgett | I would have still tried to do the uninstall |
| 21:20.40 | gtjoseph | have yo tried it? |
| 21:20.48 | rmudgett | Removing the '-' from the include simply makes the include not optional |
| 21:21.00 | gtjoseph | which makes the rule to make makeopts fail |
| 21:21.15 | coreyfarrell | rmudgett: it sounds like you just removed the '-', and didn't use gtjoseph's patch? |
| 21:21.17 | rmudgett | I did not actually do the patch because I didn't want my root directory deleted |
| 21:21.38 | rmudgett | Right that is what I did. I just removed the - |
| 21:22.06 | gtjoseph | there's a rule to make makeopts that's spits out the error and exits... |
| 21:22.08 | rmudgett | It still did the distclean regardless of makeopts not being present |
| 21:22.30 | gtjoseph | it's not supposed to prevent distclean |
| 21:23.00 | rmudgett | I substituted distclean for uninstall because I did not want it to delete my root directory |
| 21:23.15 | gtjoseph | well, maybe you should try it as a non-root user |
| 21:23.34 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-rrvdxaniwiknxfuc) |
| 21:23.37 | rmudgett | I didn't even want to do that |
| 21:23.56 | gtjoseph | then maybe a -1 was a little premature. :) |
| 21:24.04 | gtjoseph | since I DID test it. |
| 21:26.19 | coreyfarrell | gtjoseph: he only removed the '-' before include, he didn't use your ifeq conditional so 'include makeopts' was unconditional. |
| 21:26.43 | gtjoseph | yeah |
| 21:27.11 | gtjoseph | i'm triple checking. :) |
| 21:28.25 | gtjoseph | well, it works for me |
| 21:28.48 | gtjoseph | make uninstall goes no firther than the makeopts missing error |
| 21:29.25 | rmudgett | You don't get the makeopts target run to "regenerate" makeopts? |
| 21:29.46 | gtjoseph | The configure script must be executed before running 'make'. |
| 21:29.53 | rmudgett | yes |
| 21:30.00 | gtjoseph | and it stops |
| 21:30.13 | rmudgett | because that target has an exit if it is made |
| 21:30.25 | gtjoseph | yes |
| 21:30.34 | gtjoseph | and because the include triggered it |
| 21:31.15 | gtjoseph | a non-optional include will trigger that rule if makeopts doesn't exist |
| 21:31.33 | rmudgett | I was just reading about GNU make's include files. If the include file is out of date (or doesn't exist) but a rule is present to create the makefile then it builds the missing makefile and restarts make. |
| 21:32.04 | rmudgett | yse |
| 21:32.04 | rmudgett | s/yse/yes/ |
| 21:32.04 | gtjoseph | exactly. since that rule doesn't actually create the file, it'll always fail |
| 21:32.16 | coreyfarrell | rmudgett: what version of 'make' do you have? I'm seeing the same results as gtjoseph, removing the '-' before include prevents 'make distclean' from doing anything more than print the "run ./configure" error. |
| 21:32.54 | coreyfarrell | then it stops with -- Makefile:335: recipe for target 'makeopts' failed |
| 21:38.09 | coreyfarrell | I was just looking at _uninstall, maybe we should change the offending line to say: rm -f "$(DESTDIR)$(ASTMODDIR)/"*.so |
| 21:40.37 | gtjoseph | originally i had a check that $(DESTDIR)$(ASTMODDIR) was set as the first rule but because there are other places (like sounds) that use those variables, I opted just to prevent running any target |
| 21:40.49 | gtjoseph | xcept the cleans |
| 21:41.44 | rmudgett | Do you get a bunch of "/bin/sh: 0: Illegal option -" lines if a distclean was run earlier? |
| 21:42.09 | gtjoseph | not me |
| 21:43.00 | gtjoseph | with the full patch, I get a good distclean |
| 21:43.23 | gtjoseph | with just removing the '-', i get the configure error and an exit |
| 21:45.20 | rmudgett | This was on the v11 branch. Haven't tried other branches |
| 21:45.36 | gtjoseph | oh, interesting. |
| 21:49.58 | gtjoseph | ok, well I do get all the distclena errors on stock 11 |
| 21:51.08 | coreyfarrell | maybe those errors are from the static pjproject? |
| 21:51.17 | gtjoseph | rmudgett: well someing ELSE is busted in the 11 makefile |
| 21:51.26 | gtjoseph | nope, on 11 |
| 21:51.32 | gtjoseph | stock even |
| 21:51.49 | gtjoseph | the patch DOES prevent the uninstall |
| 21:52.11 | gtjoseph | but there are other pre-existing issue with the 11 makefile i guess. |
| 21:59.50 | coreyfarrell | interesting, I get those /bin/sh errors even with 'make -n distclean' |
| 22:00.03 | gtjoseph | yeah, me too |
| 22:00.35 | gtjoseph | i'm looking for $(shell) |
| 22:04.53 | coreyfarrell | gtjoseph: it's doc/core-en_US.xml and doc/full-en_US.xml |
| 22:05.11 | gtjoseph | really? |
| 22:05.28 | coreyfarrell | fixed in master with commit 5bf283d4 |
| 22:05.49 | gtjoseph | fixed in master? |
| 22:05.57 | gtjoseph | it's not a problem there as far as I know |
| 22:06.16 | coreyfarrell | the /bin/sh error was fixed by that commit in master. |
| 22:06.20 | gtjoseph | i commentred out the block that starts woth MOD_SUBDIRS at like 247 and it works |
| 22:06.35 | *** join/#asterisk-dev [TK]D-Fender (~joe@64.235.216.2) |
| 22:10.47 | gtjoseph | well, it doesnt cherry-pick so i'm not messing with it. |
| 22:15.56 | rmudgett | gtjoseph: I'm removing my -1's on the makefile reviews |
| 22:16.06 | gtjoseph | thanks. :) |