| 00:14.10 | *** join/#asterisk-dev superscrat (~asanders@173-21-89-217.client.mchsi.com) |
| 00:58.05 | *** join/#asterisk-dev mjordan (mjordan@nat/digium/x-ljqivywiivvazxbt) |
| 00:58.05 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 00:58.29 | coreyfarrell | mjordan: I can reopen the ticket if you'd like |
| 00:58.43 | mjordan | coreyfarrell: I e-mailed Kevin and asked him to take a look tomorrow |
| 00:58.49 | mjordan | since he was the one who worked the original test failures |
| 00:58.56 | mjordan | I'd like to at least understand where that got left |
| 00:59.44 | coreyfarrell | cool.. I do know that the tests are currently passing because they are flagged to expect failure. |
| 01:06.05 | mjordan | yeah... so something is wrong :-) |
| 01:11.18 | coreyfarrell | a couple months ago you mentioned something about breaking apart the testsuite, have you given that any more thought? |
| 01:31.48 | mjordan | sorry, shifting back and forth a lot |
| 01:32.00 | mjordan | honestly, with your patch that added a custom folder, I think there's less value there |
| 01:32.08 | coreyfarrell | it's cool |
| 01:32.10 | mjordan | That doesn't mean that breaking the tests out wouldn't be a good idea |
| 01:32.29 | mjordan | But the biggest reason I had for doing that - personally - was to allow people who are using the test suite to avoid merge conflicts |
| 01:32.41 | mjordan | pulling the tests out would just make it smaller, which is good |
| 01:32.50 | mjordan | but it would introduce more versioning issues |
| 01:33.11 | coreyfarrell | actually I was thinking a better approach would be to try pulling generic parts out of the testsuite and making them python libraries |
| 01:35.11 | mjordan | hm. That's not a bad idea. I had thought about going the opposite direction, but that might be easier. |
| 01:35.41 | mjordan | Plus we could actually structure lib/python/asterisk as a 'normal' python package, which would let us use normal Python unit tests to test the library infrastructure |
| 01:35.53 | coreyfarrell | like one thing I was thinking could be possibly useful in production is the code to spawn a sandboxed (but not chrooted or vm) asterisk instance. |
| 01:36.02 | mjordan | you might still run into some versioning things - but tests already have the ability to note which version of a Python library they need |
| 01:36.28 | mjordan | coreyfarrell: that would be taking the asterisk.py code in an interesting direction :-) |
| 01:37.28 | coreyfarrell | I could see having a utility running the sandbox spawner that would keep 8 instances running behind kamailio, being able to do zero downtime upgrades |
| 01:37.42 | coreyfarrell | obviously talking a beefy machine for 8 instances, you get the idea though |
| 01:39.05 | coreyfarrell | another thought (this would be harder) is to create a separate program to spawn the instances, allow it to be run as a daemon. this way a single process would control creation of instances, so we could (in theory) run tests in parallel |
| 01:42.30 | mjordan | hm |
| 01:42.37 | mjordan | I'd almost write that as a separate utility/project |
| 01:42.54 | mjordan | you *might* be able to extract that out of asterisk.py |
| 01:43.04 | mjordan | but there's some stuff in there that might not make much sense |
| 01:43.18 | mjordan | really, it's mostly just manipulating asterisk.conf and creating folders/symlinks :-) |
| 01:44.15 | coreyfarrell | yup.. the idea of being able to run a central 'asterisk spawn manager' daemon is that multiple copies of runtests.py can spawn instances or run 2 tests at a time |
| 01:45.23 | mjordan | hm |
| 01:45.26 | mjordan | that would be interesting |
| 01:46.15 | mjordan | in practice, I've found that going beyond two asterisk instances on the build agents gets "funky" - not because things don't work, but because processes get starved and the test timing gets wonky. But that's on our crappy HW. On a real system, you could do some fun things with that |
| 01:47.22 | coreyfarrell | so if we tagged the test-config.yaml with the number of asterisk instances, the central daemon would be able to queue tests, limit the number of concurrent instances. |
| 01:49.37 | coreyfarrell | so even on your crappy HW you'd be able to set the instance limit to 2, so it would only run 2 tests at the same time if they both needed one instance, any test with 2 or more instances would run serially |
| 01:50.16 | mjordan | ah. What I meant in this case is that some tests *require* 3 instances |
| 01:50.44 | mjordan | and there's some things you'd have to do with SIPp definitions in test-config.yaml even in those cases where you have a single Asterisk instance in the test |
| 01:50.49 | mjordan | but it could be doable |
| 01:51.13 | coreyfarrell | oh right to have it not talk to 127.0.0.1 |
| 01:52.14 | mjordan | yeah |
| 01:52.37 | mjordan | still, the pluggable module managing the SIPp instance 'knows' about its test object |
| 01:52.44 | mjordan | and the test object 'knows' about its asterisk instances |
| 01:52.55 | mjordan | really, what you need to tell the SIPp pluggable module is "talk to the first instance", or the second, etc. |
| 01:53.22 | mjordan | the pluggable module could then request the IP address of that instance from the test_object's asterisk.py wrapped asterisk instances |
| 01:53.32 | mjordan | so it's doable to pull that out and make it all dynamic |
| 01:53.35 | coreyfarrell | I was thinking the other day that a SIPP generator library would be nice. I DESPISE XML.. so something where you could run a bunch of python calls from an import to generate a sipp scenario |
| 01:53.41 | mjordan | heh |
| 01:53.48 | mjordan | I'm not that anti-XML. |
| 01:53.52 | mjordan | It's like violence. |
| 01:53.58 | mjordan | If it hasn't solved your problem yet, USE MORE |
| 01:54.00 | coreyfarrell | sometimes it's warrented |
| 01:54.23 | mjordan | sooo.... |
| 01:54.28 | mjordan | looks to see if something still exists |
| 01:54.47 | mjordan | ooo |
| 01:54.47 | mjordan | it does |
| 01:54.48 | mjordan | http://svn.asterisk.org/svn/sniff2sipp/trunk/ |
| 01:54.52 | coreyfarrell | hmm, does python even support 'here' documents? |
| 01:55.12 | coreyfarrell | like cat > file << EOF .. |
| 01:55.12 | mjordan | 'here' documents? |
| 01:55.15 | mjordan | ah |
| 01:55.21 | mjordan | dunno |
| 01:55.29 | mjordan | most XML readers want either a file or a string |
| 01:55.45 | coreyfarrell | IRC won't let me type the command |
| 01:55.47 | mjordan | if one supported a stream, that could work |
| 01:56.01 | coreyfarrell | try again: cat > /tmp/scenario.xml << EOF |
| 01:56.27 | coreyfarrell | not sure why it wouldn't let me type that without the 'try again: ' |
| 01:57.03 | mjordan | How would you pass that into a python interpreter? |
| 01:58.42 | coreyfarrell | ahh, the print '''multiline string...''' is how you do it in python |
| 02:00.07 | coreyfarrell | sorry was just thinking out loud about a sipp generator... if python code would be easier to read/write to generate the sipp scenarios |
| 02:00.55 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 02:02.07 | mjordan | I think if it skipped the boilerplate parts, than probably yes |
| 02:05.26 | coreyfarrell | I was thinking something similar to the way you'd do an HTTP request.. sipp.invite(from='', to='', ...optional settings follow) |
| 02:07.02 | coreyfarrell | where sipp would be an object created with an output file... you'd just call functions that use templates with default variables to generate the file.. |
| 02:11.37 | coreyfarrell | before I keep forgetting to ask, anything I can do to help with https://github.com/matt-jordan/twisted-ari/ ? |
| 02:15.42 | mjordan | heh |
| 02:15.47 | mjordan | so, the issue I ran into there was with SwaggerPy |
| 02:15.57 | mjordan | SwaggerPy was originally written to eventually support both sync and async uses |
| 02:16.20 | mjordan | unfortunately, it really ties itself hard into using requests, which is synchronous and would play havoc with twisted |
| 02:16.48 | mjordan | I worked on refactoring that, but it ended up getting kind of messy, and I started worrying that I was about to rewrite all of swaggerpy and make it something of a mess |
| 02:17.10 | mjordan | really, I should just put a swagger code generator that's twisted compatible behind twisted-ari, as opposed to making it two libraries |
| 02:17.44 | mjordan | Since a swagger schema reader for twisted is kind of specific. Even SwaggerPy eventually just assumes it is talking to Asterisk, so it's really not that reusable |
| 02:18.00 | mjordan | but that's where that kind of got left off |
| 02:19.10 | coreyfarrell | damn I really need to learn this swagger thing.. I'm thinking of creating a rest interface for my queue manager, it's a bit tough since I don't really even understand it for asterisk (a perfectly good example in front of me) |
| 02:19.19 | mjordan | it's actually quite awesome |
| 02:19.29 | mjordan | it's a schema for REST APIs |
| 02:19.42 | mjordan | I have my issues with it, but it does provide more good things than bad things |
| 02:20.05 | mjordan | the really nice thing is once you have a schema, generating client bindings - assuming your language has a library for swagger - is really easy. node.js is the prime user of it |
| 02:20.17 | mjordan | it also lets you build easy validators for your REST API |
| 02:20.36 | mjordan | so your initial "does this thing match the spec" sanity checking gets taken care of |
| 02:20.46 | mjordan | plus documentation, which is nice |
| 02:21.11 | mjordan | my only real issue with it is they made HUGE changes between versions, which really hurts compatibility |
| 02:21.27 | coreyfarrell | I'm also considering using node.js for my server.. maybe that's a better option than python? |
| 02:22.26 | coreyfarrell | a big part of why python is high on the list is that the learning time will be dual purpose (my program + the asterisk testsuite), and will make it easier for me to create tests in the testsuite for my software |
| 02:22.45 | gtjoseph | node.js +1000 |
| 02:22.52 | coreyfarrell | it does seem like node.js is in better shape for this sort of thing |
| 02:22.59 | coreyfarrell | hi gtjoseph! |
| 02:23.14 | gtjoseph | hey |
| 02:24.38 | coreyfarrell | one thing I'm still fuzzy on with nodejs is the "right" way to create a daemon with it (go from centos initscript to running as non-root). |
| 02:25.59 | gtjoseph | with a systemd service file it's easy. |
| 02:26.03 | mjordan | I'd use node.js. |
| 02:26.09 | mjordan | And I suck at JavaScript :-) |
| 02:26.19 | mjordan | But for writing async servers, it's kind of the best. |
| 02:26.28 | mjordan | Today. |
| 02:26.30 | mjordan | At least. |
| 02:26.54 | gtjoseph | until someone invents the next big thing |
| 02:27.09 | gtjoseph | that should have been Next Big Thing |
| 02:27.26 | coreyfarrell | gtjoseph: yeah so I'm still on centos 6 so no systemd |
| 02:28.01 | gtjoseph | you could just so a su - nobody /usr/bin/node in rc.local |
| 02:28.26 | gtjoseph | although i'd bet there are some goot init scripts out there |
| 02:29.04 | coreyfarrell | ah I just found a gist |
| 02:32.43 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 02:42.01 | *** part/#asterisk-dev mjordan (mjordan@nat/digium/x-ljqivywiivvazxbt) |
| 03:03.47 | *** join/#asterisk-dev mjordan (~mjordan@75.76.55.191) |
| 03:03.47 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 05:38.45 | *** join/#asterisk-dev Demon_VoIP (~demon@eoks.ivanovo.ru) |
| 06:44.22 | *** join/#asterisk-dev oej (~oej@kosmos2.upsys.se) |
| 06:44.22 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 07:20.11 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 07:20.11 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 07:27.13 | *** join/#asterisk-dev pchero_work (~pchero@109.70.54.56) |
| 07:34.58 | *** join/#asterisk-dev oej (~oej@kosmos2.upsys.se) |
| 07:34.59 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 08:22.24 | *** join/#asterisk-dev spditner_ (~simon@69-196-132-178.dsl.teksavvy.com) |
| 08:58.05 | *** join/#asterisk-dev oej (~oej@2a03:f0c0:1001:1000:c416:d5ec:fdf8:5f52) |
| 08:58.05 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 11:16.35 | *** join/#asterisk-dev oej (~oej@2a03:f0c0:1001:1000:c416:d5ec:fdf8:5f52) |
| 11:16.35 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 12:52.05 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@216-191-106-163.dedicated.allstream.net) |
| 12:54.12 | wanda_ | hello =) it's in 1H and you need't to move http://barcamp.org/w/page/97007235/BarCampMumble-VoIP-OpenSource biiises o/ |
| 12:55.02 | wanda_ | the next message will be to share an abstract about the meeting :*D |
| 13:11.10 | *** join/#asterisk-dev mjordan (mjordan@nat/digium/x-wryeibvxrtznuhlc) |
| 13:11.10 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 13:11.56 | *** join/#asterisk-dev oej (~oej@2a03:f0c0:1001:1000:c416:d5ec:fdf8:5f52) |
| 13:11.56 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 13:27.37 | *** join/#asterisk-dev leedm777 (leedm777@nat/digium/x-mbpomtgdiixdupfb) |
| 13:29.21 | *** join/#asterisk-dev superscrat (~asanders@173-21-89-217.client.mchsi.com) |
| 13:39.55 | *** join/#asterisk-dev newtonr (RustyNewto@nat/digium/x-swkeeyzulqbttczp) |
| 13:39.55 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 13:56.34 | *** join/#asterisk-dev leedm777 (leedm777@nat/digium/x-ycbwtlxquydfhspa) |
| 13:57.54 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-kxlxuzjrcyfezseu) |
| 14:23.55 | *** join/#asterisk-dev sgalarneau (Adium@nat/digium/x-otncgdctiktyngxv) |
| 14:33.50 | *** join/#asterisk-dev putnopvut (putnopvut@asterisk/master-of-queues/mmichelson) |
| 14:33.50 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 14:34.27 | *** join/#asterisk-dev gtjoseph (~gtj0@unaffiliated/gtj) |
| 14:39.55 | mjordan | oh, how I hate parsing version strings. |
| 14:40.32 | file | mjordan, but it's easy! *cough* |
| 14:57.03 | *** join/#asterisk-dev rmudgett (rmudgett@nat/digium/x-ellidwlrtiudnaoi) |
| 14:57.17 | *** join/#asterisk-dev sgriepentrog (sgriepentr@nat/digium/x-fdnrrjxpdfcxezjc) |
| 15:33.45 | *** join/#asterisk-dev w9sh (~sph@50-79-224-193-static.hfc.comcastbusiness.net) |
| 16:22.57 | *** part/#asterisk-dev w9sh (~sph@50-79-224-193-static.hfc.comcastbusiness.net) |
| 16:27.13 | coreyfarrell | kharwell: thanks for looking at those rest_api tests.. when you do your fix please verify that Asterisk isn't killed to shutdown |
| 16:30.19 | kharwell | coreyfarrell: np, it is something I missed when fixing the original issue. I'll double check as well that Asterisk isn't killed. |
| 16:33.37 | coreyfarrell | kharwell: great I've commented on ASTERISK-25167 that you are working on those 3 tests. when you have a fix please reference ASTERISK-25167 in your commit msg but don't close it. |
| 16:33.54 | *** join/#asterisk-dev superscrat (~asanders@173-21-89-217.client.mchsi.com) |
| 16:33.59 | kharwell | coreyfarrell: will do |
| 16:48.18 | *** join/#asterisk-dev spditner_ (~simon@69-165-160-168.dsl.teksavvy.com) |
| 17:08.03 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@216-191-106-163.dedicated.allstream.net) |
| 18:30.25 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 18:30.25 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 19:05.31 | *** join/#asterisk-dev w9sh (~sph@50-79-224-193-static.hfc.comcastbusiness.net) |
| 19:10.58 | *** join/#asterisk-dev mjordan (mjordan@nat/digium/x-mihoxqvculxhhjyc) |
| 19:10.58 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 19:14.48 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 19:14.48 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 19:21.25 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 19:21.25 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 19:32.24 | *** join/#asterisk-dev oej (~oej@h87-96-134-129.dynamic.se.alltele.net) |
| 19:32.24 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 19:33.02 | *** join/#asterisk-dev mjordan (mjordan@nat/digium/x-mrdspkxageymlgbv) |
| 19:33.02 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 19:44.57 | *** join/#asterisk-dev superscrat (~asanders@173-21-89-217.client.mchsi.com) |
| 20:25.50 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-179-40-172.cust.bezeqint.net) |
| 20:25.50 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 20:50.09 | *** join/#asterisk-dev [TK]D-Fender (~joe@64.235.216.2) |
| 21:38.55 | *** join/#asterisk-dev leedm777 (~leedm777@50-81-28-152.client.mchsi.com) |
| 21:40.58 | *** join/#asterisk-dev superscrat (~asanders@173-21-89-217.client.mchsi.com) |
| 21:56.45 | *** join/#asterisk-dev Corydon76 (~quassel@c-69-137-93-158.hsd1.tn.comcast.net) |
| 21:56.45 | *** mode/#asterisk-dev [+o Corydon76] by ChanServ |
| 22:28.54 | *** part/#asterisk-dev mjordan (mjordan@nat/digium/x-mrdspkxageymlgbv) |
| 23:12.19 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-kxlxuzjrcyfezseu) |
| 23:36.09 | *** join/#asterisk-dev w9sh (~sph@50-79-224-193-static.hfc.comcastbusiness.net) |
| 23:36.42 | *** part/#asterisk-dev w9sh (~sph@50-79-224-193-static.hfc.comcastbusiness.net) |