| 00:39.46 | *** join/#asterisk-dev snuff-work (~snuffy@210.8.167.13) |
| 00:39.46 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 06:21.15 | *** join/#asterisk-dev hehol (~hehol@2001:1438:1009:200:5582:6440:404c:d87) |
| 07:00.09 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 07:00.09 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 07:53.28 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 07:53.28 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 07:58.18 | *** join/#asterisk-dev danjenkins_ (~dan@cpc11-folk2-2-0-cust171.1-2.cable.virginm.net) |
| 08:06.23 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 08:06.23 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 08:30.47 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 08:30.47 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 08:32.14 | *** join/#asterisk-dev stevePearPear (~stevePear@cm244.epsilon47.maxonline.com.sg) |
| 08:34.09 | *** join/#asterisk-dev paolo_ (~smuxi@ip-62-143-36-139.hsi01.unitymediagroup.de) |
| 08:41.15 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 08:41.15 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 08:51.20 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 08:51.20 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 10:35.17 | *** join/#asterisk-dev SirLouen (~pickupdog@95.63.247.9) |
| 10:35.23 | SirLouen | hello |
| 10:36.32 | SirLouen | i've been trying to edit the code behind playback stream in channels without success (file.c and channel.c) |
| 10:37.13 | SirLouen | for some reason, if i launch the playback function it does it sequentially, first the caller and after the callee |
| 10:37.50 | SirLouen | i would like to know if anyone has an idea to tweak this somehow in order to get both channels having the playback at the same time |
| 11:15.50 | *** join/#asterisk-dev hexanol (~bibi@modemcable094.94-70-69.static.videotron.ca) |
| 11:22.22 | paolo_ | i've got WARNINGS. "chan_pjsip_write: Can't send 10 type frames with PJSIP". the calle doesn't set ptime in SDP. So ptime is 20. but it seems like chan_pjsip assumes ptime 10. |
| 11:22.50 | file | SirLouen, what are you trying to do and when? |
| 11:22.59 | file | paolo_, that's not what that means |
| 11:23.22 | file | type 10 is AST_FRAME_CNG |
| 11:23.26 | SirLouen | file i'm trying to inyect the same audio file during a Dial for example |
| 11:23.37 | file | or, a comfort noise frame |
| 11:23.53 | file | SirLouen, like after both sides have answered? |
| 11:23.57 | SirLouen | yes |
| 11:24.22 | file | then you either need to use audiohooks to mix the audio in to the streams going to both sides, or to replace the media going to both sides |
| 11:24.38 | SirLouen | replace the media going to both sides? |
| 11:25.17 | SirLouen | i'm only able to code somehow that while the caller gets the audio the callee gets a silence |
| 11:25.24 | paolo_ | file: i am making my first steps with chan_pjsip. And i trying to get rid of warnings :) |
| 11:25.25 | SirLouen | and after, it goes reverse |
| 11:25.40 | file | that's because the act of playing a sound file to a channel is blocking |
| 11:25.56 | SirLouen | yes |
| 11:25.57 | file | you have to use audiohooks and a separate thread to inject media asynchronously |
| 11:26.14 | file | OR... |
| 11:26.55 | file | you could go really low level and use the file API to get a frame and intelligently send it to both sides while ignoring what would normally go |
| 11:27.41 | SirLouen | i see... |
| 11:28.27 | SirLouen | what i've seen is that ast_playstream is mutually blocking each other |
| 11:29.10 | SirLouen | actually ast_playstream calls ast_readaudio_callback |
| 11:29.36 | SirLouen | that is the func that actually performs the frame processing |
| 11:30.25 | file | yup, it's written to do that |
| 11:30.43 | file | you can also do this externally... by originating a call where one side goes to chanspy and the other a playback |
| 11:30.59 | SirLouen | yes i was thinking this |
| 11:31.06 | SirLouen | chanspy as the audiohook originator |
| 11:31.10 | file | that's easiest, and what many people do |
| 11:31.48 | SirLouen | i'm looking to russell's app_jack |
| 11:31.57 | SirLouen | looks cool |
| 11:32.36 | SirLouen | but i wanted to try out a playstream coding solution |
| 11:35.52 | *** join/#asterisk-dev Tim_Toady (~fuzzy@snf-33276.vm.okeanos.grnet.gr) |
| 11:39.36 | SirLouen | file do you have an example of the callfile to inyect in chanspy? |
| 11:39.44 | file | no |
| 11:40.23 | file | but one side goes to a Chanspy with barge-in... the other a playback |
| 11:41.33 | russellb | "Asterisk Cookbook" has an example |
| 11:41.39 | russellb | if you can find the text |
| 11:46.31 | SirLouen | ok i'm going to check it out i found "Triggering Audio Playback into a Call Using DTMF" |
| 11:46.39 | SirLouen | but i think could be appliable |
| 11:46.53 | *** join/#asterisk-dev vassilux (~vassilux@LMontsouris-656-01-80-156.w82-127.abo.wanadoo.fr) |
| 11:48.18 | vassilux | hi , is there a possibility to call volume function from func_volume.c from my C module ? |
| 11:52.45 | SirLouen | russellb the problem i see is that you inyect the audio only with the dtmf but what happens if you want to force it for example 10 seconds in call? |
| 12:02.48 | *** join/#asterisk-dev dms1 (~dms1@2601:0:5480:e4::1008) |
| 12:08.21 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 12:08.21 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 12:09.55 | paolo_ | one error after another :-/ res_pjsip/config_transport.c:433 transport_tls_cipher_handler: Cipher 'ALL' is unsupported. Isn't it working yet ? |
| 12:10.35 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@216-191-106-163.dedicated.allstream.net) |
| 12:13.42 | file | no, although it could be added |
| 12:14.44 | file | files a bug about it |
| 12:18.24 | file | ASTERISK-24199 |
| 12:18.30 | file | oh, no bot |
| 12:18.36 | file | paolo_, https://issues.asterisk.org/jira/browse/ASTERISK-24199 |
| 12:28.01 | paolo_ | file: That was fast :) |
| 12:29.42 | *** join/#asterisk-dev oej (~Adium@h87-96-134-129.dynamic.se.alltele.net) |
| 12:29.43 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 12:33.46 | *** join/#asterisk-dev protocoldoug (~quassel@unaffiliated/protocoldoug) |
| 12:53.48 | *** join/#asterisk-dev otherwiseguy (~otherwise@23-228-175-118.mci.googlefiber.net) |
| 13:04.12 | *** join/#asterisk-dev jameswf (uid27319@gateway/web/irccloud.com/x-otpnrrdmdkmvmxzk) |
| 13:05.56 | *** join/#asterisk-dev otherwiseguy (~otherwise@23-228-175-118.mci.googlefiber.net) |
| 13:10.38 | *** join/#asterisk-dev oron (~oron_lapt@local.xorcom.com) |
| 13:37.33 | *** join/#asterisk-dev dms1 (~dms1@c-50-168-154-89.hsd1.sc.comcast.net) |
| 13:39.06 | *** join/#asterisk-dev sruffell (~sruffell@asterisk/the-kernel-guy/sruffell) |
| 13:39.06 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 13:39.26 | SirLouen | i cant figure out how to replace the dtmf with another event like 10 sec inside the call playback for chanspy idea :( |
| 13:40.06 | SirLouen | because with local chans if i use wait and dialed channel bridges, wait execution stops |
| 13:40.35 | *** join/#asterisk-dev mjordan (~mjordan@nat/digium/x-xqmvcenaqrhpojwc) |
| 13:40.35 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 13:46.36 | *** join/#asterisk-dev rmudgett (~rmudgett@nat/digium/x-ahwyvpcmvqzstszi) |
| 14:01.21 | *** join/#asterisk-dev kharwell (~kharwell@nat/digium/x-sjhculwujyuhmryq) |
| 14:15.56 | *** join/#asterisk-dev putnopvut (~putnopvut@asterisk/master-of-queues/mmichelson) |
| 14:15.56 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 14:20.06 | *** join/#asterisk-dev newtonr (~newtonr@nat/digium/x-afhwpcusyvehdrgz) |
| 14:20.06 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 14:24.03 | *** join/#asterisk-dev otherwiseguy (~otherwise@69.76.221.33) |
| 14:28.18 | wdoekes | mjordan: isn't the whole per-logger-verbosity-level a source of slowness? making the regression go away by making it a tiny bit slower sounds like one of the better alternatives (the other good alternative being undoing the problem that caused the regression) |
| 14:30.27 | mjordan | wdoekes: yeah... if I could go back in time, I'd probably ask for a different way to do that feature. Or just not do that feature. |
| 14:30.51 | mjordan | wdoekes: As it is, I think where we're at today is functional, even if it does incur a bit of extra processing. My hesitancy to put it in 11/12 was probably not warranted |
| 14:36.21 | mjordan | putnopvut: good morning! |
| 14:36.44 | mjordan | putnopvut: If you missed the bamboo spam, it appears as if RLS is a tad explodey in the unit tests/most subscription tests |
| 14:38.52 | rmudgett | I did improve the verbosity level per logger better by not generating a verbose 10 message if nothing cared about it. However, if something did care then the other logger channels would need to filter it if they didn't care. |
| 14:39.22 | rmudgett | s/logger better/logger performance/ |
| 14:42.26 | putnopvut | mjordan: YAY!!! |
| 14:42.51 | mjordan | OTOH: tests. Doing their thing by dying spectacularly. |
| 14:45.58 | rmudgett | Dropping like flies? |
| 14:51.29 | mjordan | rmudgett: yes, if flies exploded |
| 14:51.58 | rmudgett | Fly bombs |
| 14:52.41 | rmudgett | runs away |
| 14:52.55 | opticron | potassium iodide crystals? |
| 14:55.25 | opticron | maybe it was nitrogen triiodide |
| 14:57.30 | *** join/#asterisk-dev gtjoseph (~gtj0@unaffiliated/gtj) |
| 15:08.34 | *** join/#asterisk-dev protocoldoug (~quassel@unaffiliated/protocoldoug) |
| 15:10.59 | file | the RLS unit test failures are because the resource_list object uses the config file by default, but the test is trying to create a resource list within sorcery - which is not possible with config file |
| 15:11.21 | file | so sorcery needs to be changed to make that non-crash, and the test needs an explicit sorcery config to use something like astdb |
| 15:11.23 | file | or memory. |
| 15:13.47 | putnopvut | Yeah, the unit tests currently assume that you have a [res_pjsip_pubsub] category in your sorcery.conf with resource_list=memory in it. |
| 15:14.25 | putnopvut | Is there a way, programmatically, to tell sorcery to use the memory backend for the test? |
| 15:14.35 | file | not currently |
| 15:15.00 | file | it would have to be extended to allow you to build an explicit list yourself |
| 15:16.18 | putnopvut | One option is to have the unit test check if you have a sorcery.conf and if you have the appropriate section and option in it. If not, don't run the unit test. It would be better if we weren't dependent on the config though. |
| 15:23.43 | mjordan | I think for now that's acceptable. Definitely worth opening a JIRA issue for that however. |
| 15:24.53 | mjordan | at the same time, sorcery_wizard_create probably shouldn't blow up if it is asked to create something that it can't support |
| 15:25.00 | file | aye |
| 15:25.16 | mjordan | I mean, it is effective. It didn't create the thing that it couldn't support. |
| 15:25.35 | mjordan | marks this as a "Feature" and walks away |
| 15:25.59 | file | mic. dropped. |
| 15:37.03 | gtjoseph | hmmm... chan_ooh323 in 12 won't build this morning. Cast errors related to ast_ends_with and ast_strndup. Anyone else? |
| 16:08.13 | *** join/#asterisk-dev sgriepentrog (~sgriepent@nat/digium/x-kuxuictpmdgvmhgz) |
| 16:45.04 | *** join/#asterisk-dev ipengineer (~zconkle@static-71-252-134-63.dllstx.fios.verizon.net) |
| 16:45.32 | mjordan | gtjoseph: hm. what version of gcc are you using? |
| 16:56.14 | mjordan | opticron: how's the bridge test failures coming? |
| 16:59.35 | opticron | I have a solution for it, but I'm still finishing the implementation |
| 17:02.36 | mjordan | gtjoseph: compiled okay for me using gcc 4.6.3 |
| 17:26.54 | mjordan | rmudgett: #3 turns a core attended transfer into a 3-way bridge, right? |
| 17:28.28 | rmudgett | mjordan: Default is *3 |
| 17:30.03 | mjordan | did we document that anywhere? |
| 17:32.09 | *** join/#asterisk-dev lirakis (~lirakis@50-198-204-238-static.hfc.comcastbusiness.net) |
| 17:32.12 | lirakis | hey guys |
| 17:32.50 | lirakis | im looking into an issue where queue's that who's agents have a domain with an SRV record, but not an A record are all getting marked as "invalid" on log in |
| 17:33.07 | lirakis | im digging through the code path to see where the device state gets set to invalid |
| 17:33.35 | lirakis | and ... as far as i can tell, chan_sip.c ast_device_state() is the only place that sets "invalid" |
| 17:33.44 | lirakis | if it cant parse the "number" or the tech |
| 17:34.21 | lirakis | oh wait |
| 17:34.21 | lirakis | ... |
| 17:34.42 | lirakis | in devicestate.c ... sip_devicestate() res is defaulted to invalid |
| 17:35.43 | lirakis | so guess its ... failing if ((p = find_peer(host, NULL, 1))) { ... to |
| 17:36.06 | lirakis | and trying to do the else { hp = ast_gethostbyname(host, &ahp); |
| 17:36.15 | lirakis | and that ... is some how failing to do an srv |
| 17:36.20 | lirakis | so hp is not set |
| 17:36.35 | lirakis | or .. rather is set to some thing that would not evaluate true when if(hp) is checked |
| 17:36.40 | lirakis | then it falls through and returns res |
| 17:36.51 | lirakis | ... soo... what is going on in ast_gethostbyname(host, &ahp); |
| 17:36.57 | lirakis | that it doesnt do SRV |
| 17:38.58 | lirakis | any thoughts would be appreciated |
| 17:39.36 | mjordan | I can't really follow. However, chan_sip does not support SRV records in a meaningful way. |
| 17:39.50 | mjordan | dnsmgr may help you. |
| 17:40.00 | lirakis | mjordan, basically at this point... ive tracked this down to ast_gethostbyname() |
| 17:40.02 | mjordan | It provides (very limited) SRV support. But only of the primary record. |
| 17:40.07 | lirakis | the host does not have an A record |
| 17:40.11 | lirakis | only SRV |
| 17:40.34 | putnopvut | The RLS unit tests and the subscription-related testsuite tests should all now pass properly on the next go-around. |
| 17:40.38 | mjordan | \o/ |
| 17:40.44 | opticron | mjordan, https://reviewboard.asterisk.org/r/3903/ for the rest-api bridge test fixes |
| 17:40.53 | lirakis | so im guessing it fails some how ... and the previous app_queue calls to ast_device_state result in getting the default "invalid" state |
| 17:40.56 | putnopvut | Well, acutally the unit tests will not be run, but they'll also not fail or crash :) |
| 17:41.40 | lirakis | mjordan, this is outside of chan_sip really at this point |
| 17:41.51 | lirakis | im looking in utils.c |
| 17:41.59 | lirakis | ast_gethostbyname() |
| 17:42.04 | mjordan | lirakis: yes and no. ast_gethostbyname is doing its job. |
| 17:42.11 | mjordan | its job is just not the one you want. |
| 17:42.38 | lirakis | well the issue is the queue is trying to "qualify" agents |
| 17:42.40 | mjordan | the problem is, you want full DNS support with SRV records. chan_sip doesn't support that without dnsmgr. Even with dnsmgr, its support is limited. |
| 17:42.51 | lirakis | and it thinks that not having an A record means they are invalid |
| 17:43.07 | mjordan | You could try enabling dnsmgr, which will attempt to look at the SRV records and remember the first entry |
| 17:43.24 | lirakis | mjordan, hmm ok |
| 17:43.30 | rmudgett | mjordan: |
| 17:43.30 | rmudgett | <configOption name="atxferthreeway" default="*3"> |
| 17:43.30 | rmudgett | <synopsis>Digits to dial to change an attended transfer into a three-way call</synopsis> |
| 17:43.30 | rmudgett | <description> |
| 17:43.30 | rmudgett | <para>This option is only available to the transferrer during an attended |
| 17:43.30 | rmudgett | transfer operation. Pressing this DTMF sequence will result in the transferrer, |
| 17:43.31 | rmudgett | the transferees, and the transfer target all being in a single bridge together.</para> |
| 17:43.31 | rmudgett | </description> |
| 17:43.32 | rmudgett | </configOption> |
| 17:43.34 | mjordan | Or you could continue down the path you are on, but you are unlikely to receive much encouragement with mucking with ast_gethostbyname. It's job is to act as a moped, not a race car. |
| 17:43.35 | lirakis | mjordan, thanks for that |
| 17:43.39 | mjordan | rmudgett: GOOD LORD PASTEBIN :-P |
| 17:43.57 | file | mjordan, vroom |
| 17:45.00 | putnopvut | mjordan: file: based on discussions we were having earlier, has there been an issue opened for the sorcery crash or for improving sorcery (when testing) to allow for temporarily changing the sorcery backend of a particular object type? |
| 17:45.19 | file | putnopvut, no |
| 17:48.30 | putnopvut | Okie-dokie. |
| 17:49.07 | *** join/#asterisk-dev protocoldoug (~quassel@unaffiliated/protocoldoug) |
| 17:53.01 | lirakis | mjordan, sorry got a phone call |
| 17:53.16 | lirakis | mjordan, yeah i actually DONT want ast_gethostbyname to be a racecar |
| 17:53.25 | lirakis | i would rather the queue NOT qualify agents by dns at all |
| 17:53.32 | lirakis | since we use an outbound proxy |
| 17:53.51 | lirakis | i want the proxy to be doing the NAPTR/SRV/A record stuff |
| 17:54.15 | lirakis | mjordan, so ... if there is alteratively a way to disable dns qualification for the queue - that would be fine for me as well |
| 17:55.37 | putnopvut | Sorcery issue created: https://issues.asterisk.org/jira/browse/ASTERISK-24203 |
| 18:03.59 | mjordan | opticron: review feedback provided |
| 18:05.32 | mjordan | lirakis: you could try file's crazy "ignore the world with an outbound proxy" patch. It got rejected due to causing ... things ... but it would most likely (maybe?) cause chan_sip to not perform a host lookup https://reviewboard.asterisk.org/r/2851/ |
| 18:05.48 | file | I'd go with "maybe" |
| 18:06.05 | lirakis | mjordan, thats fine and actually i think im just gonna change the default state to some thing "valid" |
| 18:06.11 | mjordan | shakes the magic 8 ball |
| 18:06.15 | lirakis | mjordan, b/c we use an outbound proxy for EVERYTHING |
| 18:06.27 | lirakis | mjordan, and the box is ONLY used for queues specifically |
| 18:19.40 | opticron | mjordan, review updated |
| 18:20.52 | pabelanger | lirakis, are your registering SIP to asterisk or your proxy? |
| 18:21.04 | lirakis | we register to our proxy servers |
| 18:21.11 | lirakis | asterisk is simply an application server |
| 18:21.15 | pabelanger | right |
| 18:21.21 | pabelanger | so, we do the same thing basically |
| 18:21.31 | lirakis | and the application server has a proxy infront of it as well |
| 18:21.44 | lirakis | so all signalling goes in and out a proxy to/from the app server |
| 18:21.46 | pabelanger | we then use a local channel to dial the SIP subscribers from the proxy |
| 18:22.06 | lirakis | we are entirely proxy based for non-applications |
| 18:22.12 | pabelanger | I also tested the patch mjordan linked above, but found it didn't work is all situations. |
| 18:22.35 | lirakis | i think im just going to change the default res to some thing valud |
| 18:22.37 | lirakis | *valid |
| 18:22.43 | pabelanger | so, for now, we hack a lot of device_state function within the local channel, then use some magic look up to do DNS via a python script |
| 18:22.44 | lirakis | b/c we dont care if asterisk cant resolve the host |
| 18:23.06 | lirakis | that way if it detects some other device state like .. "on a call" it will still be set for the agent |
| 18:23.27 | lirakis | pabelanger, yikes |
| 18:23.29 | lirakis | lol |
| 18:23.53 | lirakis | yeah i dont want to do any of that since our proxies are very good and do proper rfc3263 lookup |
| 18:24.25 | pabelanger | for the most part we just do DIAL(SIP/1231@foo.com) and then use the global outboundproxy to pass the invite to the proxy. Then to proxy does the dns loop up on the domain foo.com |
| 18:24.43 | pabelanger | asterisk will not do DNS lookups in that case |
| 18:31.30 | mjordan | opticron: putnopvut: I *think* once opticron's patch goes in, we are good from the perspective of the test failures |
| 18:33.59 | putnopvut | yay |
| 18:35.21 | mjordan | file: sgriepentrog: jrose_atDigium: rmudgett: putnopvut: opticron: can you guys think of anything we've missed? |
| 18:35.51 | jrose_atDigium | Nothing in particular from me. |
| 18:36.06 | putnopvut | Can't think of anyhthng |
| 18:36.07 | rmudgett | I was just looking at opticron's patch. I've got a comment for the review. |
| 18:37.52 | file | I lack anything. |
| 18:48.41 | *** join/#asterisk-dev protocoldoug (~quassel@unaffiliated/protocoldoug) |
| 19:00.00 | mjordan | crosses his fingers |
| 19:00.08 | mjordan | hold on to your butts |
| 19:02.06 | opticron | latches on to the door frame |
| 19:03.04 | *** join/#asterisk-dev anthm (~anthm@freeswitch/developer/anthm) |
| 19:25.00 | *** join/#asterisk-dev danjenkins_ (~dan@cpc11-folk2-2-0-cust171.1-2.cable.virginm.net) |
| 19:27.13 | pabelanger | you said butts |
| 19:31.08 | *** join/#asterisk-dev otherwiseguy (~otherwise@75.87.131.147) |
| 19:49.33 | mjordan | https://www.youtube.com/watch?v=-W6as8oVcuM |
| 19:49.43 | mjordan | dear god it may have worked the first time through |
| 19:49.58 | mjordan | that's a ... sign |
| 19:50.59 | leifmadsen | heh heh heh |
| 19:52.12 | mjordan | lucky 13! |
| 19:53.19 | leifmadsen | neat! |
| 19:55.11 | *** join/#asterisk-dev protocoldoug (~quassel@unaffiliated/protocoldoug) |
| 20:16.14 | *** join/#asterisk-dev oron (~oron_lapt@line25-214.adsl.actcom.co.il) |
| 20:42.14 | *** part/#asterisk-dev lirakis (~lirakis@50-198-204-238-static.hfc.comcastbusiness.net) |
| 20:53.07 | *** join/#asterisk-dev [TK]D-Fender (~chatzilla@64.235.216.2) |
| 21:14.34 | *** join/#asterisk-dev tzafrir (~tzafrir@bzq-218-28-58.cablep.bezeqint.net) |
| 21:14.35 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 21:23.33 | *** join/#asterisk-dev Jacke (jacke@valhalla.bofh.org) |
| 21:47.04 | putnopvut | I'm trying to write an ARI app using ari-py. As a sort of baseline "hello world" type thing, I'm trying to print a channel name when I get a StasisStart event on a channel. There are a couple of odd things happening. |
| 21:48.01 | putnopvut | First, in my callback that gets called on StasisStart, I'd expect to be given a channel, based on examples in the ari-py docs. However, I'm given a dictionary instead. I can get a channel by grabbing value at key 'channel' in the dictionary though. So that's worked through. |
| 21:48.45 | putnopvut | But then trying to get the channel name is proving to be a bit troublesome as well. Looking at the docs, it appears that a channel has a "name" field defined for it. However, both channel.name and channel.get('name') throw exceptions. |
| 21:49.22 | putnopvut | So, if I'm not supposed to be doing channel.name or channel.get('name'), is there a way to get the name of the channel? Or is that just not possible? |
| 21:50.52 | sgriepentrog | As debugging, maybe try print dir(channel) to see what's there? |
| 21:52.02 | putnopvut | The problem is that the objects are set up a bit cleverly in ari-py. Instead of having direct methods, they have __getattr__ defined to look in the swagger model to find what's available. |
| 21:52.39 | putnopvut | This is what dir(channel) gives: http://ur1.ca/hye88 |
| 21:53.10 | putnopvut | I can look into that 'json' member a bit more, I suppose. |
| 21:53.23 | sgriepentrog | Ah. Looks like you don't get the name of the channel, just the id. |
| 21:53.56 | putnopvut | Aha, I can do channel.json.get('name') it appears. |
| 21:54.02 | putnopvut | Thanks for the dir() tip! |
| 21:54.05 | sgriepentrog | np |
| 22:01.44 | mjordan | putnopvut: yeah, it's a bit clever. Really, what it gets back is just the JSON blob for the event. It then pulls out the ID for the object (depending on the object, that could be id or some other field) and creates an attribute for it |
| 22:01.49 | mjordan | most things, however, are left in the JSON blob. |
| 22:02.00 | mjordan | it's a nice 'starter' library |
| 22:03.55 | putnopvut | mjordan: the thing that was throwing me off was your script that tested the call forwarding enhancements. In that script, your dial_cb (called on event Dial), grabs channel.get('peer'), channel.get('caller'), and channel.get('forwarded'), which is where I got the idea that channel.get('name') would do what I wanted. |
| 22:04.26 | mjordan | ah |
| 22:04.34 | mjordan | yes, in that case, what you get back is a dictionary of channels |
| 22:04.42 | mjordan | as the event corresponding to a dial contains multiple channels |
| 22:04.49 | mjordan | in StasisStart, however ... not so much. |
| 22:05.03 | putnopvut | Okay, thanks |
| 22:05.56 | *** part/#asterisk-dev sgriepentrog (~sgriepent@nat/digium/x-kuxuictpmdgvmhgz) |
| 22:06.01 | putnopvut | I like python a lot, but sometimes the duck typing can be bothersome |
| 22:06.24 | *** join/#asterisk-dev sgriepentrog (~sgriepent@nat/digium/x-kuxuictpmdgvmhgz) |
| 22:07.05 | mjordan | yeah... in this case, it'd be nice if we had a layer that let you subscribe to a specific channel, and it constructed what was delivered so it was your channel's updates. That would eventually result in needing to track the channels/etc. in the library itself, unfortunately. |
| 22:12.01 | *** topic/#asterisk-dev by mjordan -> Asterisk Development Discussion -=- http://www.asterisk.org/developers -=- Tier 2 and 3.14159265 support is in #asterisk -=- Asterisk 13.0.0-beta1 now released. Do the (beta) release dance! -=- Follow on Twitter at @AsteriskDev |
| 22:19.02 | *** join/#asterisk-dev WIMPy (~wimpy@e183095026.adsl.alicedsl.de) |
| 23:48.48 | *** part/#asterisk-dev kharwell (~kharwell@nat/digium/x-sjhculwujyuhmryq) |