| 00:16.16 | *** join/#asterisk-dev otherwiseguy (~otherwise@72.21.225.66) |
| 01:52.42 | *** join/#asterisk-dev mjordan (~mjordan@75.76.55.191) |
| 01:52.42 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 01:52.53 | mjordan | waves |
| 01:55.58 | *** join/#asterisk-dev sruffell (~sruffell@asterisk/the-kernel-guy/sruffell) |
| 01:55.58 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 02:10.34 | coreyfarrell | mjordan: glad that fixed the crash.. any leaks? |
| 02:10.45 | mjordan | nothing major |
| 02:10.49 | mjordan | have a patch I'm working on right now |
| 02:10.55 | mjordan | but things are moving a bit better :-) |
| 02:11.00 | coreyfarrell | anyone looking at /channels/features/ - test_features_channel_interval ? |
| 02:11.01 | mjordan | local channel tests all pass with no major leaks |
| 02:11.12 | mjordan | hm |
| 02:11.14 | mjordan | nope! |
| 02:11.26 | coreyfarrell | that's the current segfault on bamboo :) |
| 02:11.53 | mjordan | I'll take a look; that should have been run when I ran all the tests locally |
| 02:12.18 | coreyfarrell | and it ran for you? |
| 02:12.31 | mjordan | I'll check again |
| 02:12.38 | mjordan | it's using a similar pattern to CDR/CEL |
| 02:13.45 | coreyfarrell | fun |
| 02:16.01 | mjordan | just checked it in; I'm working on tweaking up the bamboo plan |
| 02:17.05 | sgriepentrog | mjordan: pong |
| 02:18.06 | mjordan | heh |
| 02:18.12 | mjordan | I can't even remember now :-) |
| 02:18.26 | mjordan | Ah! |
| 02:18.35 | sgriepentrog | np. Probably wanted to tell me about the ref thing you found? |
| 02:18.43 | mjordan | coreyfarrell fixed a nasty double de-ref of a format in sip_new - so if you're getting crashes, I'd update |
| 02:18.44 | mjordan | ha |
| 02:18.47 | mjordan | I didn't find it ;-) |
| 02:18.51 | mjordan | I did cause it |
| 02:18.58 | sgriepentrog | Right. Yeah, updated, that helped. |
| 02:19.32 | coreyfarrell | sgriepentrog: if you are running into leaks with chan_sip feel free to email ref's to me. |
| 02:19.35 | sgriepentrog | I had briefly attempted to augment the ref log with backtraces, but that was a bust. |
| 02:20.23 | coreyfarrell | I think I tried that once. lots of overhead for backtraces :) |
| 02:20.25 | sgriepentrog | Actually, that's what I'm working on, doing some of the more obvious just load/unload right now. |
| 02:20.46 | sgriepentrog | It started working... then blew up in the backtrace functions. |
| 02:21.49 | sgriepentrog | I may still be able to use it by hard coding what I'm looking for though... so it's not running constantly. |
| 02:22.57 | mjordan | ooo |
| 02:23.03 | sgriepentrog | ? |
| 02:23.05 | mjordan | masquerades get ast_format_none |
| 02:23.07 | mjordan | interesting |
| 02:23.18 | coreyfarrell | you might have better luck if you restrict it to a single ao2_alloc. like add an option for refs backtrace |
| 02:26.37 | sgriepentrog | exactly. Trigger it based on the alloc file/line and then backtrace everything that touches that object. |
| 02:26.47 | sgriepentrog | WIll do if I get the need for it. |
| 03:24.54 | coreyfarrell | mjordan: ping |
| 03:25.47 | coreyfarrell | on r3753 what was causing the assert? |
| 06:41.59 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 06:42.00 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 12:08.29 | *** join/#asterisk-dev sgriepentrog (~sgriepent@107.220.72.78) |
| 12:08.29 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 12:08.29 | *** join/#asterisk-dev jrose_atDigium (~jrose@nat/digium/x-ejvuuweptorhqtbt) |
| 12:08.29 | *** join/#asterisk-dev HeN (uid3747@gateway/web/irccloud.com/x-pmvnauczqyaanlcv) |
| 12:08.29 | *** join/#asterisk-dev coreyfarrell (~cfarrell@2001:4830:117f:2:d15d:4aa2:a953:122f) |
| 12:08.29 | *** join/#asterisk-dev luckman212 (~luckman21@unaffiliated/luckman212) |
| 12:08.29 | *** join/#asterisk-dev ChanServ (ChanServ@services.) |
| 12:08.29 | *** join/#asterisk-dev seanbright (~sean@asterisk/contributor-and-bug-marshal/seanbright) |
| 12:08.29 | *** join/#asterisk-dev angler (~angler@pdpc/sponsor/digium/angler) |
| 12:08.30 | *** join/#asterisk-dev leedm777 (~leedm777@nat/digium/x-hncdnfiaearlyvfm) |
| 12:08.30 | *** join/#asterisk-dev ivan` (~ivan@unaffiliated/ivan/x-000001) |
| 12:08.30 | *** join/#asterisk-dev wdoekes (~walter@wjd.osso.nl) |
| 12:08.30 | *** join/#asterisk-dev tm1000 (tm1000@2600:3c00::f03c:91ff:fe6e:6894) |
| 12:08.30 | *** join/#asterisk-dev felipe_ (~felipe@unaffiliated/felipe) |
| 12:08.30 | *** join/#asterisk-dev Gugge (gugge@194.150.109.208) |
| 12:08.30 | *** join/#asterisk-dev Qwell (north@pdpc/sponsor/digium/Qwell) |
| 12:08.30 | *** mode/#asterisk-dev [+oooo malcolmd ChanServ seanbright Qwell] by sinisalo.freenode.net |
| 12:08.30 | *** join/#asterisk-dev Stefan27 (~Stefan27@212.247.4.149) |
| 12:08.30 | *** join/#asterisk-dev zerohalo (~zerohalo@2601:6:f80:224:acbf:d145:ab95:9423) |
| 12:08.30 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 12:08.30 | *** join/#asterisk-dev TazzNZ (~TazzNZ@office-nat1.ytechops.co.nz) |
| 12:08.30 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-vhpgydvexexfnifc) |
| 12:08.30 | *** mode/#asterisk-dev [+o pabelanger] by sinisalo.freenode.net |
| 12:14.23 | *** join/#asterisk-dev w9sh (~sph@50-79-224-193-static.hfc.comcastbusiness.net) |
| 12:33.18 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 12:33.18 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 13:04.47 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 13:04.47 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 13:22.41 | *** join/#asterisk-dev WIMPy (~wimpy@e183095026.adsl.alicedsl.de) |
| 13:22.50 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 13:22.50 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 14:38.25 | *** join/#asterisk-dev sgriepentrog (~sgriepent@50.247.43.97) |
| 15:02.25 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@64.235.216.2) |
| 15:16.40 | *** join/#asterisk-dev sgriepentrog (~sgriepent@50.247.43.97) |
| 15:25.41 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@64.235.216.2) |
| 16:05.26 | opticron | so the crash bug I encountered when placing a pjsip->pjsip call occurs when calling from the pjsip client on my cell phone to the hardphone next to me, but not when calling to another pjsip client on another mobile device I have sitting here |
| 16:05.30 | opticron | wat |
| 16:48.25 | opticron | the format on that ast_rtp was never non-null...which means that that ast_rtp never received an audio packet |
| 16:48.40 | opticron | wonders if the SDP got messed up somewhere |
| 18:55.13 | opticron | https://reviewboard.asterisk.org/r/3761/ |
| 18:55.22 | opticron | the first part of the fix preventing the crash |
| 18:55.40 | opticron | the phone that doesn't send media also isn't getting the right codec offers in its sdp |
| 18:56.34 | *** join/#asterisk-dev mjordan (~mjordan@75.76.55.191) |
| 18:56.34 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 18:57.10 | mjordan | r3761: yeah... |
| 18:57.27 | mjordan | How does the RTP engine get into that state? |
| 19:03.35 | opticron | that's the state it starts in |
| 19:03.47 | opticron | if it never receives an audio packet, that's the state it stays in |
| 19:04.31 | opticron | I could also set the format using the format on outgoing audio frames (which would solve the case I'm experiencing) but wouldn't solve the case of never receiving or sending an audio packet |
| 19:10.34 | mjordan | opticron: how about setting it to ast_format_none on start up? |
| 19:11.05 | mjordan | functionally, it will do what your patch does, but it does so in a fashion that's less 'hard codey' |
| 19:11.28 | mjordan | plus, we can make ast_format_none less "invalid", from the perspective that it could be non-audio, etc. |
| 19:11.42 | mjordan | you can also assert on that format if it feels appropriate |
| 19:39.48 | opticron | hmmm...is ast_format_none slated to be the new init state for all formats? |
| 19:40.50 | opticron | the codec issues with pjsip in media formats branch are due the difference in behavior between ast_format_cap_append in trunk and ast_format_append_from_cap in media_formats |
| 19:41.46 | opticron | the one in trunk only appends ast_format_cap_iscompatible codecs while the one in media_formats doesn't use ast_format_cap_iscompatible_format when appending |
| 19:42.48 | opticron | makes me wonder how many places blindly replaced one with another during the swap |
| 19:57.06 | opticron | answer: mostly in chan_sip and chan_pjsip with one in ooh323 and one in a unit test |
| 20:01.39 | mjordan | well |
| 20:01.48 | mjordan | ast_format_none is the default format for channels |
| 20:01.53 | mjordan | it is a place holder for a NULL format |
| 20:02.03 | mjordan | so if you need a 'null' format without it crashing things, use that |
| 20:02.12 | opticron | I'll work on a patch for that later,then |
| 20:02.34 | mjordan | as far as ast_format_cap_iscompatible... dunno. I'm not sure what all implications there are. |
| 20:02.54 | mjordan | I'm not sure the replacement was blind however |
| 20:09.01 | mjordan | kerbloooey |
| 20:13.55 | coreyfarrell | mjordan: did you have a chance to look at r3756? |
| 20:45.47 | mjordan | looks |
| 20:46.30 | mjordan | ooo |
| 20:46.31 | mjordan | I like |
| 20:49.13 | mjordan | first run of channel drivers: https://bamboo.asterisk.org/bamboo/browse/ATB-MRCD-3 |
| 20:49.27 | mjordan | a number of those are IAX2 failures and masquerade induced crashes |
| 20:53.55 | *** join/#asterisk-dev bulkorok (~Benjamin@gw1.pinguin.ag) |
| 21:09.14 | *** part/#asterisk-dev bulkorok (~Benjamin@gw1.pinguin.ag) |
| 21:13.43 | coreyfarrell | nice |
| 21:40.45 | coreyfarrell | ugh testsuite takes forever to run. worst part is that tests/pbx/dialplan_reload and one of the fastagi tests lock up on my system |
| 21:40.56 | mjordan | huh |
| 21:40.58 | mjordan | that's odd |
| 21:41.13 | mjordan | well, dialplan_reload is a bit wonky in that it does like 50 reloads |
| 21:41.27 | mjordan | You can always run tag groups, or exclude those tests from your top most test-config.yaml |
| 21:42.05 | coreyfarrell | would be nice if I could say "run all tests that require chan_sip" |
| 21:42.36 | mjordan | ./runtests.py -g sip |
| 21:42.42 | mjordan | (might be SIP) |
| 21:42.49 | coreyfarrell | at some point after the freeze I might look at improving some of the testsuite cli options |
| 21:43.22 | coreyfarrell | hmm yeah well I don't want to miss anything |
| 21:43.52 | coreyfarrell | once this testsuite completes I plan to post a review to finish the chan_sip upgrade to ao2 |
| 21:44.39 | coreyfarrell | that is sip_mwi_subscription and sip_registry |
| 21:44.47 | coreyfarrell | err sip_subscription_mwi |
| 21:44.47 | mjordan | cool |
| 21:44.51 | mjordan | getting closer! |
| 21:45.08 | coreyfarrell | I also have a patch for res_smdi |
| 21:47.23 | coreyfarrell | so the only remaining reference to astobj.h will be in main/netsock.c (and not the header) |
| 21:48.22 | mjordan | \o/ |
| 21:48.31 | mjordan | and netsock is only used by chan_iax |
| 21:48.35 | mjordan | IIRC |
| 21:49.32 | coreyfarrell | yes that's right |
| 22:31.20 | *** join/#asterisk-dev otherwiseguy (~otherwise@23-228-175-118.mci.googlefiber.net) |