| 00:29.06 | *** join/#asterisk-dev snuff-work (~snuff-wor@210.9.148.102) |
| 00:29.06 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 05:57.33 | *** join/#asterisk-dev tsearle (~tsearle@37.19.10.33) |
| 05:57.33 | *** mode/#asterisk-dev [+o tsearle] by ChanServ |
| 06:23.23 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:2c4f:caa0:951:5a4e) |
| 07:03.05 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 08:36.11 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:2c4f:caa0:951:5a4e) |
| 11:38.20 | *** join/#asterisk-dev sahmed (~sahmed@37.19.10.33) |
| 11:53.38 | *** join/#asterisk-dev jkroon (~jkroon@165.16.203.126) |
| 12:29.54 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 12:30.56 | seanbright | kinda hard to fix an issue when you can't reproduce it |
| 12:31.50 | file | WebRTC in a nutshell. |
| 12:32.11 | seanbright | srtp error 7 seems to indicate that the ssrc changed after negotiation |
| 12:32.19 | seanbright | that's about as far as i have gotten |
| 12:32.30 | seanbright | shrugs |
| 12:35.59 | seanbright | file: good weekend? |
| 12:36.19 | file | 'twas acceptable! my dog hurt herself somehow and I discovered it last night though, so today is "fun" |
| 12:36.39 | seanbright | yeah i saw that on the twitter |
| 12:36.56 | seanbright | sick pets make me sad |
| 12:36.58 | file | I be sleeeeeeepy |
| 12:37.00 | seanbright | i bet |
| 12:37.47 | file | and YOU? I saw you back/forth on that WebRTC issue |
| 12:38.10 | seanbright | well, i was hoping it was just the reporters were linking against the wrong libsrtp |
| 12:38.18 | seanbright | unfortunately that is not the case |
| 12:38.45 | seanbright | created a debian 8 VM on AWS and spent the better part of yesterday trying to make it fail and i couldn't :( |
| 12:39.08 | seanbright | other than that, pretty good weekend |
| 12:39.35 | file | Amazon thanks you for your business. |
| 12:39.45 | seanbright | yeah |
| 12:39.49 | seanbright | t2.micro for the win |
| 12:42.41 | seanbright | so SRTP is SRTP is SRTP, right? DTLS vs. SDES is just how the keying material is exchanged, correct? |
| 12:46.21 | file | yes |
| 12:46.29 | seanbright | ok |
| 12:48.01 | file | and SSRC in WebRTC generally doesn't change unless you're doing clever things |
| 12:48.31 | seanbright | hmmm, clever |
| 12:48.52 | file | yeah, changing streams - but then you have to do a re-negotiation |
| 12:50.34 | seanbright | and the renegotiation requires a re-invite? |
| 12:50.42 | file | yes |
| 12:50.47 | seanbright | gotcha |
| 12:50.55 | seanbright | so user error |
| 12:50.58 | seanbright | closed as wontfix |
| 12:50.59 | seanbright | heh |
| 12:51.02 | file | ha |
| 13:18.58 | *** join/#asterisk-dev cresl1n (uid299068@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 13:18.58 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 14:11.46 | *** join/#asterisk-dev elguero (~miguel323@74-95-21-41-Connecticut.hfc.comcastbusiness.net) |
| 14:24.08 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-fnkdsuqttedaluuh) |
| 14:24.09 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 14:38.00 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-yoayfdebilufvurd) |
| 14:38.00 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 14:55.29 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-bdcvgrhaaydblvdm) |
| 14:55.30 | *** mode/#asterisk-dev [+o rmudgett] by ChanServ |
| 14:58.26 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:2c4f:caa0:951:5a4e) |
| 15:26.28 | *** join/#asterisk-dev AsteriskRoss (259d3426@gateway/web/freenode/ip.37.157.52.38) |
| 15:26.38 | AsteriskRoss | Hi, the git change 10152 fails to compile "res_srtp.c:604:2: error: format â%sâ expects argument of type âchar *â, but argument 6 has type âintâ [-Werror=format=]" |
| 15:29.29 | seanbright | ugh |
| 15:29.33 | seanbright | what version of libsrtp? |
| 15:30.15 | seanbright | AsteriskRoss: ^^ |
| 15:31.10 | AsteriskRoss | 1.4.4 on CentOS 7 |
| 15:32.02 | AsteriskRoss | I've worked around it, just wanted you to know |
| 15:32.08 | gtjoseph | on centos7 you need to compile your own libsrtp |
| 15:33.00 | gtjoseph | https://wiki.asterisk.org/wiki/display/AST/libsrtp |
| 15:33.08 | AsteriskRoss | Its from epel-release |
| 15:33.19 | gtjoseph | i know but it's old |
| 15:33.30 | gtjoseph | 1.5.4 is the minimum version we support |
| 15:34.00 | AsteriskRoss | Ok cool |
| 15:34.50 | seanbright | well, we still need to check for the function |
| 15:35.06 | seanbright | you could have libsrtp 1.4.4 installed for some other program |
| 15:35.12 | seanbright | shouldn't cause asterisk to fail |
| 15:35.15 | seanbright | i'm patching now |
| 15:35.26 | gtjoseph | yep |
| 15:35.39 | seanbright | or we could make sure they have at least 1.5.4... |
| 15:35.58 | gtjoseph | are you going to do a configure.ac thing? |
| 15:36.15 | AsteriskRoss | Is version 2.X still not recommended? |
| 15:36.15 | gtjoseph | if so, you could warn that it's not 1.5.4 |
| 15:36.26 | gtjoseph | we haven't really tested 2.0 much at all |
| 15:36.46 | AsteriskRoss | 1.6? |
| 15:37.23 | gtjoseph | should be ok but getting 1.5.4 exactly is the best option. |
| 15:37.31 | seanbright | 1.5.4 <= x < 2.0 |
| 15:37.35 | gtjoseph | that's the latest distributed by Fedora. |
| 15:37.50 | seanbright | although 2.x is fine in my installs |
| 15:38.21 | AsteriskRoss | Ok I'll build the recommended 1.5.4, as always thank you for your help |
| 15:38.59 | seanbright | sorry folks. will have it fixed in a moment. |
| 15:39.11 | gtjoseph | np |
| 16:06.41 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-lqewvwjsrlvawatf) |
| 16:13.58 | coreyfarrell | given the known issues with 1.4.4 should we have configure.ac reject it? if so should that restriction be 16 / 16.0 / master only? |
| 16:15.34 | gtjoseph | i think we should warn in 13/16 and reject in 16/master |
| 16:15.41 | gtjoseph | oops, meant 13/15 |
| 16:16.56 | file | I'd be down with that |
| 16:34.34 | seanbright | well, the good news is that neither 1.4.0 nor 1.4.4 include a version number in their header files |
| 16:37.42 | seanbright | 1.5.0 has srtp_get_version() and srtp_get_version_string() but you would need to link it to get it. so that probably screws cross-compiles. |
| 16:41.27 | seanbright | so https://gerrit.asterisk.org/#/c/asterisk/+/10163/ would get us halfway there. if HAVE_SRTP_GET_VERSION is not defined, we know we're < 1.5.0 |
| 16:42.34 | gtjoseph | how about pkg-config --modversion |
| 16:42.57 | seanbright | we don't currently use pkg-config to find libsrtp, but i guess we could change that |
| 16:48.36 | file | falls over |
| 16:48.47 | seanbright | props file back up |
| 16:51.40 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00a99595745b635fef.ipv6.abo.wanadoo.fr) |
| 16:52.29 | seanbright | huh... so jssip sends an UPDATE instead of an INVITE when putting a call on hold |
| 16:56.26 | *** join/#asterisk-dev dakudos (~dakudos@c-73-243-106-97.hsd1.co.comcast.net) |
| 17:46.05 | *** join/#asterisk-dev dakudos (~dakudos@c-73-243-106-97.hsd1.co.comcast.net) |
| 17:46.41 | coreyfarrell | which branches shoudl I cherry-pick https://gerrit.asterisk.org/10065 (remove compile jansson from install_prereq)? should it go to 13 / 15? if not then should it go into 16.0? |
| 17:47.16 | gtjoseph | which branches is --with jansson-bundled the default? |
| 17:47.33 | gtjoseph | and no on 16.0 |
| 17:48.07 | coreyfarrell | jansson-bundled is never default. but jansson < 2.11 is rejected by 16/16.0/master. |
| 17:48.34 | gtjoseph | hmmm |
| 17:48.39 | coreyfarrell | jansson-bundled is available in all branches |
| 17:48.47 | file | hum |
| 17:48.54 | coreyfarrell | 13/15 will warn about 2.10 but not reject. |
| 17:49.04 | gtjoseph | my vote would be to remove it in 16 and master |
| 17:49.56 | file | gtjoseph: I thought the same but then it seemed a bit odd to do it in 16.1 |
| 17:49.58 | coreyfarrell | so it'll be in 16.0.0 but not 16.1.0 |
| 17:50.20 | gtjoseph | file yeah it's only an install_prereq change though |
| 17:50.28 | file | gtjoseph: yeahhhhh |
| 17:50.38 | file | I'm not that bothered |
| 17:50.42 | gtjoseph | and it's not critical to 16.0 |
| 17:50.45 | coreyfarrell | that's why I was thinking that if it's OK to change in 16.1.0 then it's ok to change to 13.next and 15.next. |
| 17:51.04 | gtjoseph | hmmm. true |
| 17:51.15 | gtjoseph | @cresl1n: ^^ thoughts? |
| 17:51.18 | coreyfarrell | but whatever you guys want it's not a big deal IMO |
| 17:52.34 | rmudgett | There is currently a memory corruption problem when using < 2.11 that has been reported. Pointing the reporters to use 2.11 fixes the problem. |
| 17:53.03 | gtjoseph | that doesn't address the cherry-pick issue though |
| 17:53.23 | rmudgett | That sort of says we are now really dependent upon using 2.11 in all branches |
| 17:53.25 | coreyfarrell | the current install_prereq script builds 2.11 |
| 17:58.04 | rmudgett | True but we still allow use of < 2.11 if it was already installed. |
| 18:01.00 | coreyfarrell | yes, the question now is if './contrib/script/install_prereq install-unpackaged' should continue to install jansson in 13/15. if not then users would need to have install the system jansson-devel or use --with-jansson-bundled. |
| 18:02.29 | file | I don't think it should be touched in 13/15 |
| 18:02.49 | file | it's a change in behavior that may cause ripples there, I'm thinking scripts/whatever that do builds |
| 18:04.59 | coreyfarrell | that seems reasonable which is why I was thinking the same logic should apply to 16.1.0. if we don't want to put it into 16.0 (which I understand) maybe I should do it in master only? |
| 18:08.28 | file | I think master is safest |
| 18:08.53 | rmudgett | Isn't 16 requiring 2.11? |
| 18:10.04 | coreyfarrell | rmudgett: yes |
| 18:15.31 | seanbright | i like that i have to try something "30-100" times to reproduce an error |
| 18:15.42 | seanbright | (one that i have never encounteted in thousands of calls) |
| 18:16.51 | rmudgett | Sounds like you are trying too hard if 1000's don't fail |
| 18:19.46 | seanbright | i would also like chrome to let me mark interfaces as "will never ever work so stop trying them" |
| 18:19.58 | seanbright | these are my dreams |
| 18:20.26 | file | If you pass an empty stun server that'll go away |
| 18:20.38 | file | And if the server has a public host candidate it should still work |
| 18:20.56 | seanbright | will try |
| 18:21.22 | seanbright | "sendto() has failed twice returning a transient error" |
| 18:21.30 | seanbright | is it work figuring out chromium's patch process to fix that? |
| 18:21.32 | seanbright | heh |
| 18:21.35 | seanbright | worth* |
| 18:23.38 | file | dunnnnnnnno |
| 18:30.13 | rmudgett | Do there is no donut |
| 18:36.52 | jpsharp | the donut is a lie along with the cake? |
| 18:38.25 | rmudgett | Remember, there is no donut. |
| 18:39.45 | file | I like trains. |
| 18:59.06 | *** join/#asterisk-dev dakudos (~dakudos@c-73-243-106-97.hsd1.co.comcast.net) |
| 19:12.41 | *** join/#asterisk-dev jkroon (~jkroon@165.16.204.171) |
| 19:25.22 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@LFbn-1-5359-81.w90-105.abo.wanadoo.fr) |
| 20:19.40 | seanbright | what is the use-case for ao2_bump() being NULL tolerant? |
| 20:44.23 | rmudgett | convenience |
| 20:44.33 | rmudgett | it corresponds to ao2_cleanup |
| 20:44.49 | rmudgett | ao2_cleanup was required for use with RAII_VAR |
| 20:45.15 | seanbright | why would you want to increase the refcount of a NULL pointer? |
| 20:45.22 | seanbright | and how is that not a logic error? |
| 20:45.27 | rmudgett | so you don't crash |
| 20:46.12 | rmudgett | otherwise you would have to check if not null and then ao2_ref |
| 20:46.33 | rmudgett | many cases we need to bump the ref if the pointer is not null |
| 20:49.06 | seanbright | ok |
| 20:49.24 | seanbright | that seems fragile |
| 20:50.38 | rmudgett | fragile? Where we could optionally be passes an ao2 object that we need to get a ref if we were passed an object? |
| 20:50.46 | rmudgett | *passed |
| 20:51.27 | seanbright | i can't parse that sentence |
| 20:51.46 | seanbright | even when the spelling fix |
| 20:52.40 | seanbright | the NULL check is still happening, just under the covers |
| 20:53.37 | seanbright | i don't know that eliding an explicit NULL check really makes that much of a difference from a readability standpoint |
| 20:53.38 | rmudgett | That's all ao2_bump is. It is a macro that increases the ref if the passed pointer is not null. Then returns the passed pointer |
| 20:53.51 | seanbright | yes, i have seen the implementation. |
| 20:55.08 | seanbright | my expectation when reading: ao2_bump(whatever); is that when that statement executes, whatever's refcount will have been incremented. |
| 20:55.45 | seanbright | i guess that is what documentation is for. |
| 20:55.53 | seanbright | but it's still dumb |
| 21:00.47 | tm1000 | seanbright: no astricon this year for you? :( |
| 21:00.56 | seanbright | tm1000: negative |
| 21:01.27 | seanbright | if astridevcon has a conference bridge of some kind i can 'stop by' |
| 21:01.54 | file | it will, it will |
| 21:03.07 | tm1000 | it has for three years or something |
| 21:03.16 | seanbright | well times change, tm1000 |
| 21:03.51 | file | impossible! |
| 21:30.15 | jameswf | an option https://usercontent.irccloud-cdn.com/file/if56xOnC/image.png |
| 21:30.41 | file | make that a use case for Zulu! |
| 21:30.44 | file | :P |
| 21:34.03 | jameswf | Improvement/Holograms |
| 21:37.08 | rmudgett | gtjoseph: Shouldn't https://gerrit.asterisk.org/#/c/asterisk/+/10132/ be part of the fix for https://issues.asterisk.org/jira/browse/ASTERISK-27121 |
| 22:36.25 | seanbright | gtjoseph: you around? |
| 23:02.43 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-fnkdsuqttedaluuh) |
| 23:26.37 | *** join/#asterisk-dev gtjoseph (~gtjoseph@unaffiliated/gtj) |
| 23:26.37 | *** mode/#asterisk-dev [+o gtjoseph] by ChanServ |
