| 01:34.08 | *** join/#asterisk-dev Laureano (bef573b3@gateway/web/freenode/ip.190.245.115.179) |
| 01:40.13 | Laureano | Hello, when a blind xfer is initiated, the flag AST_SOFTHANGUP_ASYNCGOTO is set on the channel to break the normal frame flow so an async goto can be. Should this flag be removed after the flag? Or it will stay set on the channel for the remainder of the call? Blind transfers from queues seems to be conflicting with it's behaviour. |
| 01:40.22 | Laureano | *can be done |
| 01:40.35 | Laureano | *after the xfer |
| 02:05.51 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 02:53.29 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-cwbeqmljjrdgkpnd) |
| 04:29.02 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 05:21.45 | *** join/#asterisk-dev babak (uid19622@gateway/web/irccloud.com/x-oxqlpxokqqsxqpqb) |
| 05:30.27 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 06:13.13 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 06:13.13 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 06:31.23 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 06:36.38 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 06:36.38 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 06:53.33 | *** join/#asterisk-dev pchero_work (~pchero@109.70.54.56) |
| 07:23.56 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 07:23.57 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 07:31.49 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 07:34.41 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 07:34.41 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 08:32.47 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 09:28.40 | *** join/#asterisk-dev Qwell (north@asterisk/developer/Qwell) |
| 09:28.40 | *** mode/#asterisk-dev [+o Qwell] by ChanServ |
| 09:33.45 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 10:02.05 | *** join/#asterisk-dev snuff-home (~brad@180-150-6-49.NBN.cranbourne.aussiebb.net) |
| 10:02.05 | *** mode/#asterisk-dev [+o snuff-home] by ChanServ |
| 10:28.55 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 10:28.55 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 10:34.12 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 11:35.09 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 11:56.46 | *** join/#asterisk-dev snuff-home (~brad@180-150-6-49.NBN.cranbourne.aussiebb.net) |
| 11:56.46 | *** mode/#asterisk-dev [+o snuff-home] by ChanServ |
| 12:01.46 | *** join/#asterisk-dev snuff-home (~brad@180-150-6-49.NBN.cranbourne.aussiebb.net) |
| 12:01.46 | *** mode/#asterisk-dev [+o snuff-home] by ChanServ |
| 12:17.17 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@216-191-106-163.dedicated.allstream.net) |
| 12:24.07 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 12:24.07 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 12:56.16 | *** join/#asterisk-dev leedm777 (leedm777@nat/digium/x-gydzyanhlizxivrs) |
| 13:17.08 | *** join/#asterisk-dev leedm777 (leedm777@nat/digium/x-bhwxmxthhdiplitp) |
| 13:26.32 | *** join/#asterisk-dev Laureano (bef573b3@gateway/web/freenode/ip.190.245.115.179) |
| 13:31.06 | Laureano | Morning, I'm working on 23730 to understand under what conditions this issue is reproduced. When a blind xfer is initiated, the flag AST_SOFTHANGUP_ASYNCGOTO is set on the channel to break the normal frame flow so an async goto can be done. Should this flag be removed after the xfer? Or it will stay set on the channel for the remainder of the call? Blind transfers from queues seems to be conflicting with it's behaviour. |
| 13:31.25 | file | it should be cleared |
| 13:31.47 | *** join/#asterisk-dev leedm777 (leedm777@nat/digium/x-wwoheytxifbmwmii) |
| 13:34.45 | Laureano | Thanks file. I will look for the code that should clear it then. app_queue checks if the channel has any softhangup flag set, my guess is to detect if the channel is still up, but it seems that under some conditions the flag is cleared after that check is done, thus leading to a features' blind xfer not being logged as such in queue_log. |
| 14:13.19 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-bejgiqavinznkaac) |
| 14:15.33 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 14:15.33 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 14:15.40 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-kthtpujkjildnigr) |
| 14:27.36 | *** join/#asterisk-dev gtjoseph (~gtj0@unaffiliated/gtj) |
| 14:36.39 | *** join/#asterisk-dev Synthase_ (uid63346@gateway/web/irccloud.com/x-zpfwspmkgjlgvaqu) |
| 14:46.04 | *** join/#asterisk-dev vjfromgt (vjfromgt@68.132.66.209) |
| 14:47.09 | vjfromgt | i am looking for a developer for chan_mobile - add android sms - am i in the right place? my email is vjfromgt@gmail.com |
| 16:30.57 | *** join/#asterisk-dev Tim_Toady (~fuzzy@snf-33276.vm.okeanos.grnet.gr) |
| 16:36.31 | *** join/#asterisk-dev superscrat (asanders@nat/digium/x-zbvwspliohmhooyz) |
| 17:16.16 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 17:16.16 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 18:30.29 | *** join/#asterisk-dev leedm777 (leedm777@nat/digium/x-mjwtqdocoiqtgplp) |
| 18:39.33 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 18:39.34 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 19:39.14 | *** join/#asterisk-dev roramirez (~root@boxtub.com) |
| 19:39.36 | roramirez | Hi guys! |
| 19:43.07 | roramirez | what branch support new features?, i searching info in wiki but can not find. |
| 19:47.14 | file | Asterisk master absolutely, 13 depending on things |
| 19:58.42 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-179-40-172.cust.bezeqint.net) |
| 19:58.42 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 20:13.45 | roramirez | ok file . My questions is related with Gerrit patch #1447 for create a issue on jira. |
| 20:14.46 | leifmadsen | where is my mjordan?! |
| 20:15.20 | file | leifmadsen, classified |
| 20:15.24 | file | roramirez, you can file it against master |
| 20:15.29 | leifmadsen | question for the room: if you were given a spec file that would build out a set of Asterisk RPMs, and the command (mock) to do that, would you expect the dependencies to be easily available for you from a hosted platform, or would you want those spec files as well so you could buidl them local? |
| 20:16.04 | leifmadsen | I'm trying to determine if I ask Digium (or whoever) to host the set of "dep" RPMs for building, and just provide a nice simple mock.cfg that would point at that repo for the dep building |
| 20:16.22 | file | knows nothing |
| 20:17.00 | roramirez | thanks file |
| 20:21.04 | *** join/#asterisk-dev pchero (~pchero@109.70.54.56) |
| 21:19.30 | danjenkins | hey so I'm going to ask the question I tried to ask the other day but everyone was dead after astricon |
| 21:20.29 | danjenkins | if the browsers dont need tls certs to be able to do DTLS stuff, why do we have to pass in TLD certs into asterisk for WebRTC ? |
| 21:20.35 | danjenkins | TLS* |
| 21:21.22 | danjenkins | or maybe i should ask the question during the working day |
| 21:21.41 | file | because the browser generate them |
| 21:39.42 | DivideBy0 | during those presentations, I was wondering how the browsers handle identity with those certs if it's P2P |
| 21:40.08 | nbjoerg | it's s3curity |
| 21:40.37 | DivideBy0 | s3curity = security theater? |
| 21:41.08 | nbjoerg | to a degree, ydes |
| 21:41.55 | DivideBy0 | based on what I saw in the presentations/panels, it seems relatively trivial for a mitm. I haven't sat down to read the specs yet though |
| 21:46.00 | danjenkins | so yes @file why can't asterisk generate them |
| 21:46.46 | nbjoerg | you don't want to create certificates dynamically if you don't have to |
| 21:47.01 | nbjoerg | it's quite expensive in terms of cpu load |
| 21:47.28 | nbjoerg | s/certificates/private keys/ |
| 21:56.54 | danjenkins | how can the browsers do it so "quickly" |
| 21:57.21 | danjenkins | i know chrome is moving to the elipical one (cant remember the name) |
| 21:58.03 | nbjoerg | danjenkins: ecdsa is cheaper, but a browser generally doesn't have to care as much about cpu load as a server |
| 21:58.19 | danjenkins | thats the one |
| 21:59.00 | danjenkins | ok :) I was just curious, it's something Torrey asked me the other day |
| 21:59.44 | danjenkins | he also said that hecause his certs were so long that he said in firefox he got IP fragmentation; after he took his certs out and used self signed certs, the problem disappeared |
| 21:59.58 | *** join/#asterisk-dev [TK]D-Fender (~joe@64.235.216.2) |
| 22:00.17 | danjenkins | i think it was firefox |
| 22:01.52 | leedm777 | danjenkins: I saw the same thing using chan_respoke. See https://github.com/respoke/chan_respoke#dtls-certificate |
| 22:02.47 | danjenkins | interesting. leedm777 do you consider this a bug with the browser/asterisk and using self signed certs is just a plaster over the top |
| 22:02.54 | danjenkins | this is way above what I understand |
| 22:03.20 | leedm777 | Honestly, I haven't dug into it too deeply. |
| 22:03.41 | leedm777 | On the surface, self-signed certs seems like a Bad Thing⢠|
| 22:04.07 | leedm777 | But I don't know if there are other aspects of the negotiation that compensate for the lack of a CA |
| 22:04.47 | DivideBy0 | it looks like there's an identity provider proposed as a draft, but no real spec yet |
| 22:04.54 | danjenkins | can you remember if it was firefox or non-browser specific? |
| 22:05.31 | leedm777 | non-browser specific. |
| 22:05.36 | danjenkins | ah ok |
| 22:05.50 | leedm777 | the packets coming out of Asterisk weren't making it to the browser, because fragementation. |
| 22:07.31 | leedm777 | The browsers had other tricks to reduce packet size, in an attempt to avoid fragementation-induced packet loss |
| 22:08.06 | leedm777 | The details escape me, though. |
| 22:09.04 | danjenkins | just wondering if this should be made into a bug on chrome bug tracker for example |
| 22:09.23 | danjenkins | oh you say it wasnt making it out of asterisk |
| 22:09.40 | leedm777 | it was getting lost in the network between Asterisk and the browser. |
| 22:10.48 | leedm777 | so it all of that depends on your network configuration. if Asterisk and the browser are on the same network, it almost always works. |
| 22:11.58 | danjenkins | ah interesting |
| 22:12.20 | danjenkins | so that sounds like it should be documented on the asterisk wiki |
| 22:12.44 | danjenkins | this is where someone points and says.... it is |
| 22:13.11 | danjenkins | and basically people should stop trying to use, you know, their proper TLS certs and just generate self signed ones :D |
| 22:41.39 | *** join/#asterisk-dev babak (uid19622@gateway/web/irccloud.com/x-xgfcfrqgzmotaxgw) |
| 22:50.18 | coreyfarrell | leifmadsen: I think just having the rpm spec use dep names from EPEL would be good enough. |
| 22:52.19 | nbjoerg | danjenkins: what stops you from creation a self signed cert in advance? |
| 22:52.33 | nbjoerg | danjenkins: it also gives you full control over what algorithm and key size is used |
| 22:53.50 | nbjoerg | I'm also a bit puzzled, certs should only be exchanged on initial connection setup and even a 4k RSA key is less than half of what should be safe for any modern network |
| 22:54.12 | nbjoerg | [given that ipv6 requires 1200+ Bytes as minimum MTU] |
| 22:55.23 | nbjoerg | and seriously, if you use 4kbit RSA keys on the internet, you are likely doing something wrong... |
| 23:02.19 | *** part/#asterisk-dev coreyfarrell (~coreyfarr@67-8-56-126.res.bhn.net) |
| 23:12.24 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-bejgiqavinznkaac) |
| 23:18.44 | danjenkins | oh nothing stops you from doing it, it's just an extra step to getting asterisk running and im guessing a lot of people would try and use their proper certs now that you need it for getUserMedia in Chrome for example |
| 23:18.58 | danjenkins | yes it's an interesting dilemma |
| 23:19.14 | danjenkins | if thats the case should it go in as a bug in asterisk? |
| 23:20.18 | nbjoerg | it would be nice to see a proper analysis first for why they got pushed into fragmentation |
| 23:20.59 | nbjoerg | that said, if you have a non-trivial cert chain, yes, could happen |