| 00:20.03 | *** join/#asterisk-dev infobot (ibot@208.53.50.136) |
| 00:20.03 | *** 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 |
| 01:49.14 | *** join/#asterisk-dev sh_smith (~sh_smith@cpe-76-174-26-91.socal.res.rr.com) |
| 07:05.57 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 07:19.54 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:4559:7079:edb9:3261) |
| 08:10.36 | *** join/#asterisk-dev jkroon (~jkroon@105.4.4.196) |
| 12:50.27 | *** join/#asterisk-dev scgm11_ (~scgm11@r186-52-168-22.dialup.adsl.anteldata.net.uy) |
| 13:06.09 | *** join/#asterisk-dev scgm11_ (~scgm11@r186-52-168-22.dialup.adsl.anteldata.net.uy) |
| 13:28.01 | *** join/#asterisk-dev AsteriskRoss (259d3426@gateway/web/freenode/ip.37.157.52.38) |
| 13:49.02 | *** join/#asterisk-dev cresl1n (uid299068@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 13:49.02 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 13:56.50 | gtjoseph | coreyfarrell: rough guess... how long does a full testsuite run with REF_DEBUG take? |
| 14:00.12 | *** join/#asterisk-dev CELYA_ (~Thunderbi@80.12.34.118) |
| 14:04.18 | gtjoseph | and other than enabling REF_DEBUG in menuselect are there other things that need to be done? |
| 14:05.02 | *** join/#asterisk-dev CELYA_ (~Thunderbi@80.82.229.178) |
| 14:06.45 | gtjoseph | and what output should be captured? |
| 14:17.04 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-xpaefegcoeobywgn) |
| 14:17.05 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 14:43.57 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-nefqhlgxyelbovel) |
| 14:43.57 | *** mode/#asterisk-dev [+o rmudgett] by ChanServ |
| 14:45.05 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-eghtnxdqwzqktkys) |
| 14:45.05 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 14:55.26 | coreyfarrell | gtjoseph: I think it used to take like 6 hours but I also did some patches that should make it run faster. enabling REF_DEBUG is all you have to do. |
| 14:55.58 | coreyfarrell | if instance of asterisk which leaks will cause that test to fail and the output of refcounter.py is included in the logs/ artifacts. |
| 14:56.56 | coreyfarrell | BTW this takes a lot of disk space. |
| 14:58.36 | gtjoseph | what's "a lot"? |
| 15:16.32 | coreyfarrell | I don't remember. I'm running a compile now so I can run one test to see how much space the refs log adds to one test. |
| 15:19.29 | coreyfarrell | at one point I did some startup optimizations which significantly reduced the number of calls to ao2_ref so the refs logs might be smaller than I'm remembering. |
| 15:23.23 | coreyfarrell | gtjoseph: is it intentionally that jenkins emails me instead of the codereview list? |
| 15:23.35 | gtjoseph | hmmm. no. |
| 15:23.48 | gtjoseph | checks |
| 15:23.52 | file | it goes to both |
| 15:24.16 | gtjoseph | meaning both jenkins and gerrit are emailing? |
| 15:24.44 | file | actually first question |
| 15:24.47 | file | coreyfarrell: what email are you referring to? |
| 15:25.42 | coreyfarrell | file: the one that says: Build Started https://jenkins2.asterisk.org/job/asterisk-unittests/job/master/25/ ' |
| 15:26.10 | coreyfarrell | looks like it is from gerrit but it's sent to me directly |
| 15:26.41 | file | ah |
| 15:26.55 | coreyfarrell | I noticed that email doesn't go to the mailing list because my email rule doesn't catch it :( |
| 15:33.16 | coreyfarrell | gtjoseph: I just ran tests/manager/originate, the refs log was 3.8mb (8x the size of the full log). so if you're running 1000 tests figure that's at least 4gb added to the tmp folder where asterisk instances are created, plus copies of those logs for all failed tests. |
| 15:33.49 | gtjoseph | hmmm. ok thanks. |
| 16:00.33 | coreyfarrell | I have one major item on my pre-16 wish-list. can we make it require jansson >= 2.11 (current latest)? to do so would require an option '--with-jansson-bundled' as that version of jansson is unavailable on many distro's. |
| 16:01.14 | gtjoseph | i don't have a problem with that personally. |
| 16:02.58 | file | my only concern would be the ripples from the distros |
| 16:03.54 | gtjoseph | what's the jansson version on fedora27-28 and ubuntu17-18? |
| 16:04.19 | gtjoseph | fedora27 is already 2.11 |
| 16:05.28 | coreyfarrell | EPEL6 also has 2.11. Redhat decided to put jansson into RHEL7 so they're still on 2.10. AFAIK they've ignored the request to update RHEL7. |
| 16:06.21 | coreyfarrell | I'm not sure about ubuntu (or how to find out) |
| 16:06.54 | coreyfarrell | file: what do you mean ripples from the distros? |
| 16:07.00 | file | 18.04 is 2.11 |
| 16:07.10 | file | coreyfarrell: packagers |
| 16:07.27 | file | if 2.11 is widespread then I think it'd be fine |
| 16:09.45 | coreyfarrell | gtjoseph: do the jenkins agents have 2.11? the jenkins page doesn't say what distro's are running anymore, what is the current list? |
| 16:10.01 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00bdebef4ef10dc1c1.ipv6.abo.wanadoo.fr) |
| 16:10.10 | gtjoseph | centos7 fedora27, fedora28 ubuntu16, ubuntu18 |
| 16:13.45 | coreyfarrell | centos7 does not have jansson 2.11 so for that you'd either have to enable bundled or update jansson 2.11 (it's backwards compatible). |
| 16:16.43 | gtjoseph | was there an actual api change in 2.11 or was it just the threading stuff behind the scenes? |
| 16:17.54 | coreyfarrell | 2.11 has API additions, no breaking changes. you can safely replace the system package on RHEL7 if that's the question. |
| 16:18.48 | gtjoseph | no it was... what happens if we require 2.11 and someone tries to build with an earlier version? |
| 16:18.57 | gtjoseph | does it still compile but may be unstable? |
| 16:21.09 | coreyfarrell | if we require 2.11 then ./configure would reject 2.10. |
| 16:21.34 | gtjoseph | oops. certified/13.21 needs backports for gcc fixes. compile fails on fedora 28. |
| 16:21.52 | gtjoseph | ok but what of the check weren't there |
| 16:22.43 | gtjoseph | in other words, do we HAVE to make it required or can we do like we do with srtp and say "at your own risk" on earlier ver4sions. |
| 16:22.53 | file | or put another way: why do you want it required? |
| 16:23.04 | gtjoseph | :) |
| 16:23.52 | gtjoseph | if asterisk will still compile and run on earlier versions, then the packagers won't care and we can still do a bundled 2.11 if needed. |
| 16:24.39 | coreyfarrell | new features I'm interested in: json_vsprintf https://jansson.readthedocs.io/en/2.11/apiref.html#c.json_vsprintf and "o*" of json_pack. |
| 16:25.09 | gtjoseph | ah, ok i get it now |
| 16:25.22 | coreyfarrell | we could create a replacement stub for json_vsprintf that would work on <=2.10 but the json_pack feature would be awesome and really can't be done outside of jansson |
| 16:26.48 | gtjoseph | can we test for the "o*" in ./configure and use it it it's there? or is it too much hassle to use a workaround if it's not there? |
| 16:27.09 | rmudgett | There was also an ASTERISK memory corruption issue that was "fixed" by using libjansson 2.11 |
| 16:27.24 | gtjoseph | yep |
| 16:27.25 | rmudgett | *upgrading to |
| 16:27.40 | gtjoseph | but that falls in the "use older versions at your own risk" category |
| 16:28.02 | gtjoseph | oops. i gotta run |
| 16:28.37 | coreyfarrell | gtjoseph: the whole benefit of "o*" is that it simplifies our code. having conditional code to work around it not existing would do the opposite. |
| 16:55.24 | seanbright | we should just roll our own json lib |
| 16:55.26 | seanbright | problem solved |
| 17:18.28 | *** join/#asterisk-dev scgm11_ (~scgm11@r186-52-168-22.dialup.adsl.anteldata.net.uy) |
| 17:49.33 | *** join/#asterisk-dev dakudos (~dakudos@c-73-243-106-97.hsd1.co.comcast.net) |
| 18:01.11 | *** join/#asterisk-dev malcolmd (malcolmd@pdpc/sponsor/digium/malcolmd) |
| 18:01.11 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 18:35.08 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@LFbn-1-5358-11.w90-105.abo.wanadoo.fr) |
| 19:06.42 | *** join/#asterisk-dev kharwell (~kharwell@user-24-214-15-130.knology.net) |
| 19:06.42 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 19:29.43 | *** join/#asterisk-dev joshelson (~joshelson@67.133.173.134) |
| 19:47.25 | *** join/#asterisk-dev scgm11_ (~scgm11@r186-52-168-22.dialup.adsl.anteldata.net.uy) |
| 20:00.34 | coreyfarrell | gtjoseph: any objection to putting the pjproject md5 file into the asterisk repo? |
| 20:00.49 | gtjoseph | for what purpose? |
| 20:02.14 | coreyfarrell | it'll be slightly more secure since the pjproject md5 will be verified as part of the asterisk tarball but also jansson doesn't have md5 checksum so I'll be able to share more of the download/verify code if the md5 was already in third-party. |
| 20:03.19 | gtjoseph | yeah ok |
| 20:03.22 | gtjoseph | go for it |
| 20:47.28 | jpsharp | If I've got a "Switch => DUNDi/blah" in a dialplan context, how often does that make a dundi query? every time something in that context is called? |
| 21:34.17 | *** join/#asterisk-dev joshelson (~joshelson@67.133.173.134) |
| 22:01.38 | rmudgett | jpsharp: Looks like switches are checked after normal extens in the context. After switches then context includes are checked. |
| 22:03.06 | jpsharp | Hmm. Okay. |
| 22:10.57 | jpsharp | Does it matter about order in the context configs? |
| 22:14.32 | coreyfarrell | rmudgett: I'm addressing your findings, you think that module.conf errors should return -1, failure to load modules -2? do you know if the behavior of exit(2) vs exit(1) is documented anywhere? |
| 22:15.32 | rmudgett | jpsharp: The switches are searched in the order listed in the context. |
| 22:16.14 | rmudgett | coreyfarrell: I looked for documentation about -2 which thus causes asterisk to exit with code 2. I could not find any documentation. |
| 22:17.02 | rmudgett | I was noting that when you attempt to load a module and it fails to load we were exiting with a -2 |
| 22:17.47 | rmudgett | -2 in load_resource_list() |
| 22:19.02 | jpsharp | rmudgett: I'm seeing a lot of dundi traffic via the two switch statments I have, even though other extension patterns should match them. |
| 22:19.06 | jpsharp | But I'll dig a little further. |
| 22:19.27 | jpsharp | so much dundi traffic that it's actually causing weird locks and SIP timeouts. |
| 23:11.56 | *** part/#asterisk-dev kharwell (~kharwell@user-24-214-15-130.knology.net) |