| 01:57.42 | *** join/#asterisk-dev dandann00dle (uid120491@gateway/web/irccloud.com/x-djdmmwawwovodhlx) |
| 03:35.39 | *** join/#asterisk-dev leifmadsen (~leifmadse@asterisk/documenteur-extraordinaire/blitzrage) |
| 03:35.40 | *** mode/#asterisk-dev [+o leifmadsen] by ChanServ |
| 04:28.16 | *** join/#asterisk-dev tuxd00d (~tuxd00d@ip68-106-11-138.ph.ph.cox.net) |
| 06:43.52 | *** join/#asterisk-dev pchero_work (~pchero@109.70.54.56) |
| 06:44.05 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 07:24.54 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 07:24.54 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 07:37.28 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 07:37.43 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 07:47.26 | *** join/#asterisk-dev jkroon (~jkroon@uls-154-73-35-201.wall.uls.co.za) |
| 08:13.56 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 08:47.30 | *** join/#asterisk-dev jkroon_ (~jkroon@165.16.204.36) |
| 08:50.45 | *** join/#asterisk-dev hajekd (~hajekd@212.24.153.186) |
| 08:51.00 | *** join/#asterisk-dev hajekd (~hajekd@212.24.153.186) |
| 08:52.19 | *** join/#asterisk-dev hajekd (~hajekd@212.24.153.186) |
| 08:57.47 | *** join/#asterisk-dev jkroon_ (~jkroon@uls-154-73-35-201.wall.uls.co.za) |
| 09:27.41 | *** join/#asterisk-dev CELYA_ (~Thunderbi@LStLambert-657-1-90-73.w80-13.abo.wanadoo.fr) |
| 10:55.53 | coreyfarrell | just in case nobody noticed gerrit is down |
| 11:09.48 | stefan27 | does that mean you developers get a free day in the park? |
| 11:11.13 | coreyfarrell | it means I can't download a patch that I need for testing :( |
| 11:21.05 | coreyfarrell | file / gtjoseph: if you get a chance can you kick gerrit? |
| 11:22.31 | file | I can try |
| 11:23.19 | file | starting, so it says |
| 11:24.40 | file | back up |
| 11:30.09 | coreyfarrell | thanks! |
| 11:39.27 | *** join/#asterisk-dev pc-m (~pcm@modemcable094.94-70-69.static.videotron.ca) |
| 11:56.45 | *** join/#asterisk-dev hajekd (~hajekd@212.24.153.186) |
| 12:00.06 | gtjoseph | what was the issue with gerrit? |
| 12:01.53 | file | gtjoseph: VM host got restarted |
| 12:02.07 | gtjoseph | oh |
| 12:10.27 | gtjoseph | Oh man. I just realized Tom Petty died on Monday. |
| 12:21.56 | coreyfarrell | does anyone know why res_pjsip_session.so was made separate from res_pjsip.so? |
| 12:24.39 | coreyfarrell | it looks like they call each other during module load, I'm confused about how it can even work. load_module for res_pjsip_session calls ast_sip_register_service (res_pjsip) but load_module for res_pjsip calls ast_res_pjsip_init_message_ip_updater which calls ast_sip_session_register_supplement (res_pjsip_session). |
| 12:32.54 | gtjoseph | Ir probably should have been a source file in res/pjsip. |
| 12:33.06 | gtjoseph | s/Ir/It |
| 12:33.55 | file | There didn't used to be a dependency like that |
| 12:36.28 | coreyfarrell | hmm sorry I was working with an older revision of master, looks like pjsip_message_ip_updater.c has been removed. I'll have to see what I get once my rebuild finishes and I run it again. |
| 12:39.28 | file | It was renamed |
| 12:39.35 | file | More stuff gets added to it |
| 12:42.40 | gtjoseph | Maybe we should make all res_pjsip* modules just be source files on res/pjsip, then ifdef them in and out based on menuselect. |
| 12:45.51 | coreyfarrell | I think some of it makes sense as separate modules, but some doesn't. for example: why would you load res_pjsip without res_pjsip_session? |
| 12:46.46 | gtjoseph | true |
| 12:47.04 | coreyfarrell | for some of the other stuff if it were part of res_pjsip you'd probably need a run-time option to be able to enable/disable. |
| 12:47.41 | gtjoseph | yeah easy enough |
| 12:49.16 | coreyfarrell | that actually would be better than #ifdef because packagers aren't going to disable any of the options. |
| 12:49.32 | coreyfarrell | so you'd need an alternative to 'noload' |
| 12:50.15 | gtjoseph | we could add "feature" load/noload options to pjsip.conf/system |
| 12:50.35 | gtjoseph | or maybe enable/dosable to avoid confusion with modules |
| 12:55.10 | coreyfarrell | that would create a big diff between 15 and 16 though.. I assume you can't just merge everything together in released versions. |
| 12:56.43 | gtjoseph | hmmm. good question. |
| 12:57.03 | gtjoseph | we wouldn't be changing the api |
| 12:57.43 | coreyfarrell | modules.conf: require=res_pjsip_session.so |
| 12:57.52 | gtjoseph | and external modules shouldn't be linking directly to or dlopening the modeules |
| 12:58.49 | coreyfarrell | I think we'd have to leave an empty module for anything merged into res_pjsip. |
| 13:00.36 | gtjoseph | we could. I'd guess it'd piss off the packagers if we removed a bunch of modules. |
| 13:01.56 | coreyfarrell | that I'm not so worried about, the packagers would get an error before anything is deployed. more worried about end users who might have used the require directive of modules.conf. |
| 13:02.31 | file | Yeah that's not happening, every time we do that people have problems due to custom modules.conf |
| 13:04.04 | file | >_> |
| 13:05.24 | gtjoseph | what's "not happening"? the whole concept? |
| 13:05.26 | coreyfarrell | file: so what do you think of the code from res_pjsip_session.c to res_pjsip/pjsip_session.c, then creating an empty module for res_pjsip_session? |
| 13:06.08 | file | I'd still be iffy on that, it's a huge thing to do |
| 13:07.17 | gtjoseph | would only be 15 and master |
| 13:07.59 | file | makes merging more difficult |
| 13:08.11 | gtjoseph | now that IS true |
| 13:08.38 | file | if it were to be done it'd need to either be all branches or none, or else we're in for a world of manual merging |
| 13:08.56 | gtjoseph | we could leave the source files there and just modify the Makefile to compile them directly to res_pjsip.so |
| 13:09.11 | coreyfarrell | ewwww |
| 13:10.47 | gtjoseph | coreyfarrell: you should propose it on the asterisk-dev list |
| 13:10.52 | coreyfarrell | also 15.0.0 has been released with res_pjsip_session.so, so we need to produce that binary (even if empty) |
| 13:13.51 | coreyfarrell | does res_pjsip/pjsip_message_filter.c need to be in res_pjsip.so? maybe would be lower impact for that to become part of res_pjsip_session? |
| 13:15.12 | file | yes, or else the address information in messages will be incorrect |
| 13:15.30 | coreyfarrell | the only non-static symbols in that file are the init/cleanup functions, and they call functions from res_pjsip_session so they can't work without res_pjsip_session anyways. |
| 13:15.31 | gtjoseph | yeah the filter does multiple things at different points |
| 13:16.13 | coreyfarrell | I'm still confused about how it can successfully call ast_sip_session_register_supplement during startup of res_pjsip. |
| 13:16.46 | gtjoseph | possible because if the number of dlopen passes we make on modules? |
| 13:18.19 | gtjoseph | we _could_ move t he session supplement registration up to res_pjsip |
| 13:18.45 | file | that'd be a better option |
| 13:19.06 | gtjoseph | yeah it's a linked list. |
| 13:19.38 | gtjoseph | and nothing has to be initialized in res_pjsip_session to add a supp to the list |
| 13:20.55 | coreyfarrell | so move the list, reg, unreg to res_pjsip/pjsip_session.c with add_supplements converted to a non-static so it can be called from res_pjsip_session? |
| 13:21.21 | file | yes |
| 13:21.31 | file | I think that's fine |
| 13:21.35 | gtjoseph | yep |
| 13:21.35 | file | and a minimal change |
| 13:23.29 | coreyfarrell | hmm I'll give that a try. |
| 13:23.33 | gtjoseph | You'd probably need a new ast_sip_session_run_supplements() in res_pjsip/pjsip_supplements |
| 13:24.16 | gtjoseph | to use whereever res_pjsip_session iterated over t5he LL to run them |
| 13:26.10 | gtjoseph | Interesting. I'm getting this from gerrit on signin: "SignInFailure,SIGN_IN,Nonce+verification+failed." |
| 13:26.36 | file | time. |
| 13:26.52 | file | it is possibly out of sync, or crowd is out of sync |
| 13:27.31 | gtjoseph | yikes, the server is 3 minutes fast |
| 13:27.55 | coreyfarrell | do you think ast_sip_session_register_supplement / ast_sip_session_unregister_supplement should stay in res_pjsip_session.h or move to the main header? |
| 13:28.22 | gtjoseph | i'd say main header |
| 13:28.27 | file | res_pjsip.h should be included for anything regardless, so should be fine there |
| 13:31.27 | file | gtjoseph: gerrit fast? |
| 13:31.46 | gtjoseph | yeah and the ntp server isn't reachable from ot |
| 13:32.03 | gtjoseph | might have to do with the network issues yesterday |
| 13:32.06 | file | yes |
| 13:32.33 | gtjoseph | opening issue with IT |
| 13:34.51 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-enskycfesbiazptq) |
| 13:42.55 | *** join/#asterisk-dev newtonr (newtonr@nat/digium/x-xjvuzrtqlnqpizmb) |
| 13:42.55 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 13:46.12 | gtjoseph | i'm bouncing gerrit |
| 13:51.18 | gtjoseph | well, that didn't help |
| 13:52.57 | gtjoseph | file: were you able to log in recently? |
| 13:53.05 | file | I logged in this morning after I restarted it |
| 13:53.07 | file | still logged in |
| 13:53.22 | gtjoseph | dare i ask you to log out and try logging in again? |
| 13:54.02 | file | I can incognito it |
| 13:54.27 | file | no go |
| 13:54.43 | gtjoseph | hmmm. |
| 13:54.50 | gtjoseph | the time on the server is correct now |
| 13:55.32 | gtjoseph | oh i wonder if apache needs to be restarted |
| 13:56.36 | gtjoseph | nope |
| 13:56.38 | file | shouldn't need to be.. |
| 13:56.46 | file | crowd / oauth may be the one that is off now |
| 13:57.11 | gtjoseph | i'd exzpect issues with logging into either directly though and i'm nopt having an issue there |
| 13:57.56 | file | gerrit and jenkins are the only things that use the oauth component |
| 13:58.06 | gtjoseph | yeah true |
| 13:58.13 | file | Atlassian proper doesn't auth the same way |
| 13:58.58 | gtjoseph | jenkins was fine |
| 14:06.10 | *** join/#asterisk-dev dandann00dle (uid120491@gateway/web/irccloud.com/x-qthttcksqnxylcqo) |
| 14:22.01 | seanbright | the icalendar patch is not correct, but i can't log in to -1 it, just FYI |
| 14:22.13 | file | yeah gerrit is no good right now |
| 14:22.16 | file | I'll mark it -1 |
| 14:22.43 | gtjoseph | i've got an internal ticket open to IT to see what's going on |
| 14:27.31 | *** join/#asterisk-dev zerohalo (~zerohalo@2601:199:417f:7e6d:f4ce:d5c0:532b:a37c) |
| 14:32.12 | *** join/#asterisk-dev tuxd00d (~tuxd00d@ip68-106-11-138.ph.ph.cox.net) |
| 14:38.05 | *** join/#asterisk-dev tuxian (~tuxian@igilmour.plus.com) |
| 14:40.33 | file | gtjoseph: server is back, but I don't recall if Gerrit starts automatically |
| 14:40.52 | gtjoseph | pffft |
| 14:42.09 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-ocznyoavcfkfxhay) |
| 14:42.09 | *** mode/#asterisk-dev [+o rmudgett] by ChanServ |
| 14:42.55 | gtjoseph | i'm still getting the error |
| 14:45.11 | file | fun fun fun |
| 15:01.07 | gtjoseph | gerrit should be available again |
| 15:06.53 | *** join/#asterisk-dev leifmadsen (~leifmadse@asterisk/documenteur-extraordinaire/blitzrage) |
| 15:06.53 | *** mode/#asterisk-dev [+o leifmadsen] by ChanServ |
| 15:12.48 | file | gtjoseph: nifty |
| 15:13.21 | gtjoseph | time on yooden was 8 hours off :) |
| 15:13.50 | file | that's a bit much |
| 15:14.36 | *** join/#asterisk-dev hajekd (~hajekd@212.24.153.186) |
| 15:17.50 | seanbright | can we change the label in JIRA from "Reviewboard Link" to "Gerrit Link" or something? |
| 15:18.22 | file | maybe. |
| 15:19.00 | rmudgett | seanbright: The gerrit reviews tab lists the reviews that have to topic set to the issue. |
| 15:19.55 | rmudgett | s/have to/have the |
| 15:22.32 | file | I changed it to "Review Link" |
| 15:26.02 | *** join/#asterisk-dev kharwell (~kharwell@67.231.202.2) |
| 15:26.03 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 15:31.54 | seanbright | holy shit... |
| 15:32.05 | seanbright | err, let me hop over to a relevant channel |
| 15:39.32 | coreyfarrell | woohoo, I got one of the pjsip basic call tests down to 90 AO2 object leaks. |
| 15:40.44 | seanbright | i'm picking up what you're putting down |
| 15:40.53 | coreyfarrell | the other 7000 leaks were because res_pjsip / res_pjsip_session were being held open by circular loops. |
| 15:41.41 | gtjoseph | cool |
| 15:41.46 | seanbright | those leaks were during/after the call? |
| 15:41.49 | seanbright | or during shutdown |
| 15:41.50 | seanbright | ? |
| 15:44.37 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-179-40-170.cust.bezeqint.net) |
| 15:44.37 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 15:44.48 | coreyfarrell | I posted the refs.txt showing the 90 leaks to https://issues.asterisk.org/jira/browse/ASTERISK-27306 |
| 15:45.29 | coreyfarrell | I'm unsure the cause, they could be due to the call or the endpoint, or just due to shutdown. |
| 15:47.04 | seanbright | i would assume (i know, i know...) that if people are using pjsip in production and we were leaking 90 objects per call that we would hear about it every day |
| 15:51.22 | coreyfarrell | I've seen some pretty bad leaks in the past, I can't assume that others will catch it in production. |
| 15:55.03 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 15:55.21 | *** join/#asterisk-dev jkroon (~jkroon@vc-nat-gp-s-41-13-56-34.umts.vodacom.co.za) |
| 17:08.00 | *** join/#asterisk-dev CELYA_ (~Thunderbi@LStLambert-657-1-90-73.w80-13.abo.wanadoo.fr) |
| 17:08.41 | coreyfarrell | gtjoseph: it looks like JIRA's time is still 8 hours off. |
| 17:08.53 | coreyfarrell | or at least it was as of an hour ago. |
| 17:09.00 | gtjoseph | :) |
| 17:09.34 | gtjoseph | s/:)/:(/ |
| 17:12.39 | *** join/#asterisk-dev tsearle (~tsearle@67.231.202.2) |
| 17:12.39 | *** mode/#asterisk-dev [+o tsearle] by ChanServ |
| 17:18.20 | coreyfarrell | does the testsuite have anything that tests adding & removing endpoints for pjsip? currently in chan_sip I frequently creating/deleting static peers (type=friend), is that analogous to a pjsip endpoint? |
| 17:27.05 | file | Not really, and yes |
| 17:31.53 | coreyfarrell | hmm from what I can tell pjsip is leaking endpoints, I just don't know if it's a shutdown leak or if it would happen repeatedly when adding/removing. |
| 17:32.43 | coreyfarrell | the 90 leaked objects appear to be tied to the 4 endpoints used in the test. |
| 18:23.12 | *** join/#asterisk-dev kharwell (~kharwell@67.231.202.3) |
| 18:23.12 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 18:29.48 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 18:34.30 | coreyfarrell | pjsip_configuration.c calls load_all_endpoints after setting up sorcery, does anyone know what this is for? it causes leaks on startup, I removed the call and a couple tests I ran still worked fine. |
| 18:40.11 | file | looks like it was the result of one of Alexei's patches for improving realtime performance |
| 18:40.55 | gtjoseph | yeah it gets the cache populated iirc |
| 18:54.54 | *** join/#asterisk-dev kharwell (~kharwell@67.231.202.4) |
| 18:54.54 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 18:56.28 | coreyfarrell | ah thanks, looks like the leak might be somewhere in stasis. |
| 19:13.53 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d004dc7d3df4eb7617f.ipv6.abo.wanadoo.fr) |
| 19:47.17 | *** join/#asterisk-dev kharwell (~kharwell@67.231.202.2) |
| 19:47.17 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 19:49.03 | *** part/#asterisk-dev kharwell (~kharwell@67.231.202.2) |
| 20:17.46 | *** join/#asterisk-dev dandann00dle (uid120491@gateway/web/irccloud.com/x-xkcuzaiccgouzvlt) |
| 20:21.27 | *** join/#asterisk-dev ggo (~quassel@67.231.202.2) |
| 21:43.49 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 21:53.01 | *** join/#asterisk-dev newtonr (~newtonr@216.186.174.249) |
| 21:53.01 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 21:53.29 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 22:11.38 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 22:21.17 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 22:30.54 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 22:39.00 | *** join/#asterisk-dev snuff-work (~snuff-wor@210.9.148.102) |
| 22:39.00 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 22:49.04 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 22:58.38 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |
| 23:44.49 | *** join/#asterisk-dev hajekd (~hajekd@81.90.251.91) |