| 00:18.42 | *** join/#asterisk-dev infobot (ibot@208.53.50.136) |
| 00:18.43 | *** topic/#asterisk-dev is Asterisk Development Discussion -=- http://www.asterisk.org/developers -=- Tier 2 and 3.14159265 support is in #asterisk -=- Check out our blog! blogs.asterisk.org -=- Follow on Twitter at @AsteriskDev |
| 05:22.14 | *** join/#asterisk-dev Worldexe (~Worldexe@95-107-33-134.dsl.orel.ru) |
| 06:47.18 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:f1c7:3396:dff1:6173) |
| 07:10.59 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 08:08.05 | *** join/#asterisk-dev wdoekes (~walter@wjd.osso.nl) |
| 08:08.06 | *** mode/#asterisk-dev [+o wdoekes] by ChanServ |
| 09:05.46 | wdoekes | good morning. can someone tag the current 13 branch with 13-dev or something? now it looks like 13.15.0-rc1 while we're at 13.22. confused me for a while |
| 09:05.51 | wdoekes | $ git log -1 --format=oneline && git describe |
| 09:05.51 | wdoekes | f4ddc56b9a83664f2b8728a9276cf74e2c5e8844 (HEAD -> 13, origin/13) Merge "res_pjsip_sdp_rtp: include ice in ANSWER only if offered" into 13 |
| 09:05.54 | wdoekes | 13.15.0-rc1-1649-gf4ddc56b9a |
| 12:17.42 | gtjoseph | coreyfarrell: i need you asap. are you around? |
| 13:17.04 | gtjoseph | coreyfarrell: never mind. false alarm :) |
| 14:04.22 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-qrqefbgcleobiymb) |
| 14:04.23 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 14:09.11 | *** join/#asterisk-dev sahmed (~sahmed@4.7.148.242) |
| 14:40.55 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-xmtiamkkbzdhdwnm) |
| 14:40.55 | *** mode/#asterisk-dev [+o rmudgett] by ChanServ |
| 14:46.16 | *** join/#asterisk-dev tsearle (~tsearle@37.19.10.33) |
| 14:46.16 | *** mode/#asterisk-dev [+o tsearle] by ChanServ |
| 14:51.20 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-mebfjajdhwcocjzs) |
| 14:51.20 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 15:17.57 | rmudgett | tsearle: ping |
| 15:18.04 | tsearle | hello |
| 15:18.31 | rmudgett | I don't understand what you meant in your comment on the deadlock review |
| 15:19.09 | tsearle | so pjsip has a timer to retransmit unreliable 183 message every 60 seconds |
| 15:19.20 | rmudgett | Why are we retransmitting the 183 response periodically? |
| 15:19.31 | rmudgett | Are you using 100rel? |
| 15:19.36 | tsearle | no |
| 15:19.54 | tsearle | but I think the rfc for early media does say you should periodically re-transmit 183 |
| 15:20.07 | tsearle | in the non-100rel case |
| 15:20.18 | gtjoseph | what casues the timer to stop? |
| 15:20.33 | tsearle | stop as in cancelling it? |
| 15:20.53 | gtjoseph | yeah |
| 15:21.19 | tsearle | probably when you send pjsip a new response to send |
| 15:22.16 | tsearle | I have't fully digged that part of code, but I can tell you that it does try to re-xmit every 60 seconds |
| 15:22.38 | tsearle | and in this case the other guy is also re-xmitting every 60 seconds too |
| 15:22.55 | tsearle | so asterisk is trying to both forward a 183 and internally re-xmit a 183 at the same time |
| 15:23.02 | rmudgett | Which is the problem because we have two threads working on the same tdata at the same time. |
| 15:23.12 | tsearle | exactly |
| 15:24.01 | tsearle | today I've been working on moving the txdata clone function out to sip_transport.h (same place where rx_data is clone method is defined) |
| 15:24.35 | tsearle | and trying to clone it every time tdata is getting saved in session / transaction objects |
| 15:24.43 | rmudgett | We need two tdata's to truly resolve. However, that is going to involve making sure that the life of tdatas are managed. |
| 15:24.45 | tsearle | so far it's still deadlockign |
| 15:24.58 | tsearle | checked that already |
| 15:25.15 | tsearle | tdatas are being reference counted |
| 15:25.30 | tsearle | so cloning shouldn't cause a memory leak on that front |
| 15:26.13 | tsearle | there is only one exception to the proper reference counting bit |
| 15:26.20 | rmudgett | reference counts just keep a data structure around until the count reaches zero. They don't provide reentrancy protection. |
| 15:26.21 | tsearle | and that's when it's saving an ack |
| 15:26.27 | tsearle | and there is a nice "this is a bug" comment there |
| 15:27.09 | tsearle | but making sure that the session and transaction layers are having seperate txdata objects (both reference counted) would |
| 15:27.20 | tsearle | which is what you tasked me to try to do, heh |
| 15:27.44 | file | This sounds like something that should be brought up with Teluu before going down the rabbit hole further |
| 15:28.09 | gtjoseph | yeah exactly |
| 15:28.36 | tsearle | load tests of my original patch indicated that it resolved my deadlock issue (and didn't create new one) |
| 15:28.57 | gtjoseph | tsearle: if you can write an email fully describing the issue we can push it with Teluu. |
| 15:30.03 | tsearle | the description in the jira isn't clear enough? |
| 15:30.16 | gtjoseph | i didn't look. :) |
| 15:31.39 | rmudgett | The description could use a SIP scenario description of what is happening in the transaction. |
| 15:31.54 | tsearle | I can include a pcap |
| 15:32.27 | tsearle | or sipp files |
| 15:35.10 | coreyfarrell | jenkins not currently working? logs say it ran ./configure but then makeopts does not exist. this is not happening to me locally. https://jenkins2.asterisk.org/job/asterisk-unittests/job/master/68/console |
| 15:40.48 | tsearle | @gtjoseph added a little ascii callflow into the jira and more explanation |
| 15:40.56 | gtjoseph | great thanks |
| 15:41.15 | tsearle | let me know if it's sufficint |
| 15:41.21 | tsearle | and when you do, then I can go home :-P |
| 15:41.58 | tsearle | and I'll stop working on the clone patch until I get an update from you guys |
| 15:44.51 | rmudgett | With that call flow I now understand what you meant by the other side sending a 183. |
| 15:45.27 | tsearle | cool |
| 15:45.37 | tsearle | I have your permission to call it a day then? |
| 15:48.14 | gtjoseph | you are excused :) |
| 15:49.03 | tsearle | :D |
| 15:49.09 | tsearle | take care all! |
| 15:52.29 | seanbright | it's not ascii art without {noformat}, just sayin' |
| 16:38.11 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00c8bd9bf6268f6c0c.ipv6.abo.wanadoo.fr) |
| 17:36.01 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@2a01cb0583d31d00c8bd9bf6268f6c0c.ipv6.abo.wanadoo.fr) |
| 17:44.00 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d003548710b3a2fddd0.ipv6.abo.wanadoo.fr) |
| 17:49.47 | coreyfarrell | looks like the testsuite didn't get 16 or 16.0 branches. I'm surprised jenkins isn't failing the checks which need those branches. |
| 20:40.48 | gtjoseph | coreyfarrell: no gates have run yet so there was no need for the testsuite. it was on my checklist though and i just created the branches now. |
| 20:41.52 | coreyfarrell | oh right testsuite is not used by the asterisk check. |
| 20:42.32 | gtjoseph | :) |
| 20:43.09 | gtjoseph | have you been looking at the new jenkins stuff in tests/CI? |
| 20:44.06 | coreyfarrell | I glanced but to be honest I don't know where to start. can I use that stuff from a stand-alone docker image (or is that on a todo list)? |
| 20:46.14 | gtjoseph | the shell scripts you can use but the jenkinsfiles obviously need jenkins |
| 20:46.35 | gtjoseph | i'm still working on the docs |
| 20:49.10 | coreyfarrell | cool I'll probably start playing with it once the first draft of those docs are published |
| 20:49.42 | gtjoseph | yeah i tried to make it so there are no hidden things you need to actually use the stuff |
| 20:49.56 | gtjoseph | so i'd like feedback. docs next week. |
| 21:23.51 | coreyfarrell | gtjoseph: I'm running buildAsterisk.sh on a container right now, from what I can tell TESTSUITE_DIR must be declared for runTestsuite.sh, otherwise no args/variables needed for build, install or unit tests? |
| 21:25.20 | gtjoseph | you can specify any env vars on the command line with --env-var which will become ENV_VAR in the script |
| 21:25.37 | gtjoseph | so start it with --testsuite-dir=<dir> |
| 21:27.08 | coreyfarrell | oh neat. so I need ./tests/CI/runTestsuite.sh --testsuite-dir=/usr/src/testsuite, otherwise everything should run on defaults? |
| 22:29.52 | sahmed | Hi, could anyone please tell me what is done by ao2_bump? |
| 22:33.54 | kharwell | it increases the reference count of an ao2 object by one and returns the object. It also checks for NULL on the given object. |
| 22:34.00 | kharwell | before doing so |
| 22:35.07 | sahmed | Got it, thanks @<kharwell> |
| 22:58.12 | AsteriskRoss | Has something changed today with opus? Asterisk isn't allowing custom opus settings in codec.conf |
| 23:12.14 | kharwell | AsteriskRoss: yes a new version was put out, but I believe the only change was with the documentation. Are you seeing any kinds of warnings/errors when loading codec_opus (or at asterisk startup)? |
| 23:15.43 | AsteriskRoss | @kharwell It looks like its loading but not correctly, I'll see if I can spot an error when loading |
| 23:17.07 | AsteriskRoss | Ok there aren't any messages displayed but its not loading the options from the codec.conf |
| 23:23.14 | AsteriskRoss | In a version of asterisk pulled from GIT yesterday it works without issue |
| 23:24.12 | kharwell | AsteriskRoss: after looking back at the release history and git log it appears another change may have gone into this latest release besides the documentation one. |
| 23:24.38 | kharwell | I'd say go ahead and file an issue on the asterisk issue tracker for it |
| 23:24.57 | file | include the configuration |
| 23:26.24 | kharwell | actually I am wrong! I was looking in master and not the 15 branch for opus. The only change should be the documentation change |
| 23:26.57 | AsteriskRoss | I'm using 13 |
| 23:27.00 | kharwell | when I load 15 I do receive this warning though: "config_options.c:1058 xmldoc_update_config_type: Cannot update type 'opus' in module 'codec_opus' because it has no existing documentation!" |
| 23:27.21 | gtjoseph | yes, those are a result of today's build |
| 23:27.39 | gtjoseph | pull the earlier version and install it manually |
| 23:28.13 | gtjoseph | the 1.3.0 version still should work |
| 23:28.40 | gtjoseph | i'll be fixing the issue wuth 1.8.1 first thing in the AM. |
| 23:28.46 | gtjoseph | sorry, 1.3.1 |
| 23:28.49 | file | oh, I understand what happened... yeah... |
| 23:29.13 | *** join/#asterisk-dev sh_smith (~sh_smith@cpe-76-174-26-91.socal.res.rr.com) |
| 23:29.14 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-qrqefbgcleobiymb) |
| 23:29.30 | gtjoseph | "xmldoc dump" wants the docs without the xml and docs top level elements. |
| 23:29.52 | gtjoseph | but everything else wants the xml and docs top level elements |
| 23:30.18 | gtjoseph | i took the elements out when i saw that i couldn't generate the wiki docs for the rest of asterisk |
| 23:50.20 | AsteriskRoss | Just to clarify to manually install i need to copy: |
| 23:50.28 | AsteriskRoss | codec_opus-13.0_1.3.0-x86_64/codec_opus.so >> /usr/lib64/asterisk/modules/ |
| 23:50.44 | AsteriskRoss | codec_opus-13.0_1.3.0-x86_64/format_ogg_opus.so >> /usr/lib64/asterisk/modules/ |
| 23:51.02 | AsteriskRoss | odec_opus-13.0_1.3.0-x86_64/codec_opus_config-en_US.xml >> /var/lib/asterisk/documentation/thirdparty/ |
| 23:51.12 | AsteriskRoss | Is that it? |
| 23:56.28 | AsteriskRoss | Ok its all working now! |