| 00:25.17 | *** join/#asterisk-dev leifmadsen (~Leif@asterisk/documenteur-extraordinaire/blitzrage) |
| 00:25.17 | *** mode/#asterisk-dev [+o leifmadsen] by ChanServ |
| 01:15.59 | *** join/#asterisk-dev sruffell (~Adium@asterisk/the-kernel-guy/sruffell) |
| 01:16.00 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 01:49.27 | *** join/#asterisk-dev elguero (~miguel323@2001:470:1f06:12c4::2) |
| 04:39.43 | *** join/#asterisk-dev mickecarlsson (~Micke@h10n3c1o1101.bredband.skanova.com) |
| 05:29.01 | *** join/#asterisk-dev tzafrir (~tzafrir@212.179.75.202) |
| 05:29.01 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 06:03.12 | *** join/#asterisk-dev schmidts (~schmidts@Lo20-rt01.lm33.sil.at) |
| 06:03.13 | schmidts | good morning |
| 06:14.45 | kaii | hey schmidts |
| 07:05.34 | *** join/#asterisk-dev sgimeno (~sgimeno@163.117.206.10) |
| 07:20.03 | schmidts | hey folks is there a know bug with 1.8 with a lot of sip peers (> 4000) and sip peer objects get inserted multiple times? when i do a sip reload every time a new sip object for every peer is allocated and the refcounter rises by 2 for all objects |
| 07:21.04 | schmidts | it looks like this happens only with a lot of sip peers with 10 i wasnt able to reproduce this |
| 07:25.16 | schmidts | please take a look here: http://pastebin.com/6kv8AMjc i guess this describes better what i mean |
| 07:36.04 | wdoekes2 | schmidts: qualify/register going on for any/all of them? |
| 07:38.38 | schmidts | i havent tried if everything works but normally these peers are behind a proxy and only the proxy communicates with asterisk, sip directly is blocked |
| 07:38.54 | schmidts | but for my gateway servers, qualify works fine |
| 07:39.23 | schmidts | here is a pastebin of my sip.conf maybe i have something very wrong in it, but i dont think so: http://pastebin.com/6kv8AMjc |
| 07:41.03 | schmidts | i am trying now if trunk works or if this problem exists there too |
| 07:43.36 | schmidts | same for trunk :( |
| 07:45.42 | schmidts | and it looks like these objects really get allocated: 161923720 memory before and after sip reload: 161951348 |
| 07:47.15 | wdoekes2 | bad pb paste? |
| 07:47.26 | schmidts | which one? |
| 07:53.08 | schmidts | ouch |
| 07:53.51 | schmidts | it looks like the part of build_peer which should find old peers and remove and readd it to the list doesnt work |
| 08:01.28 | *** join/#asterisk-dev irroot (~irroot@196-215-124-168.dynamic.isadsl.co.za) |
| 08:01.59 | snuff-work | schmidts: did u try and find a min number of peers that this happens on? |
| 08:04.33 | schmidts | snuff-work not now, i am allready in the code where and why this happens ;) but i will try to find a value |
| 08:05.43 | irroot | morning folks |
| 08:06.03 | schmidts | morning ;) |
| 08:07.24 | irroot | it should be criminal to stay < 20m from beach and have to come into the office ... |
| 08:08.40 | schmidts | it looks like its happens here: http://www.asterisk.org/doxygen/asterisk1.8/chan__sip_8c-source.html#l26589 peer->ha looks like its empty |
| 08:18.01 | schmidts | snuff-work bad news :( it doesnt depends on the amount of peers, it looks like when i include a file into sip.conf this happens |
| 08:20.56 | wdoekes2 | mm.. it looks like there are several early returns that can cause oldha to not be freed (might be unrelated) |
| 08:22.10 | schmidts | not in this case, i have added some debug output and the build_peer function runs down everytime to where it should be freed but oldha is NULL at the moment, but i might be wrong at all of this oldha thing cause this is only the access list for a peer |
| 08:23.30 | schmidts | ao2_find increase the ref counter right? maybe its just a refcounter problem |
| 08:23.46 | schmidts | cause i didnt see a refcounter decrease for this peer |
| 08:23.46 | wdoekes2 | yes.. I think it's a different issue -- if it is one. but I do wonder why oldha and olddirectmedia is nulled but freed first at end of function |
| 08:23.49 | snuff-work | yes it does |
| 08:24.11 | schmidts | wdoekes2 maybe this is only cause i dont use access lists on peers so i guess its ok that its null |
| 08:24.17 | snuff-work | gotta go |
| 08:25.17 | wdoekes2 | "just" a refcounter problem.. sounds plausible that it is one |
| 08:25.44 | schmidts | what i can see is that the refcounter is the same for every object and thats increased with every sip reload |
| 08:26.41 | schmidts | but i dont understand where this peer object is allocated |
| 08:27.48 | schmidts | ah the peer which is found in the function is unlinked from the peers container, maybe thats why i can only see them in the object list but not in the sip peers list |
| 08:31.17 | schmidts | man, i have tried to add a refcounter decrease at the end of the function to make sure it will go away and the refcounter is now only increased by 1 but the objects still stay there |
| 08:40.13 | schmidts | it really looks like an unref problem but i didnt get it why a new object is added |
| 08:42.12 | wdoekes2 | I suggest you reduce the config to something minimal first |
| 08:43.27 | schmidts | 4 peers ;) |
| 08:48.44 | *** join/#asterisk-dev oej (~olle@static-213-115-251-100.sme.bredbandsbolaget.se) |
| 08:48.44 | *** mode/#asterisk-dev [+o oej] by ChanServ |
| 08:48.58 | oej | Morning folks. |
| 08:49.38 | schmidts | morning olle |
| 08:52.20 | schmidts | ok i have found it :D |
| 08:52.47 | schmidts | build_peer use ao2_find with flag OBJ_UNLINK to the peers container but its never unlinked from the peers_by_ip container so the refcounter stays wrong |
| 08:53.02 | schmidts | by adding a ao2_unlink from peers_by_ip container of this peer its fine |
| 08:56.50 | irroot | schmidts +1 |
| 08:57.03 | oej | Great |
| 08:57.40 | irroot | job well done |
| 08:57.54 | irroot | ~beer shmidts |
| 08:57.54 | infobot | ACTION deftly decants a fine Piraat for shmidts |
| 09:22.31 | *** join/#asterisk-dev pnlarsson (~niklas@fw1.gml.g.icnet.infracom.se) |
| 09:28.23 | *** join/#asterisk-dev sgimeno (~sgimeno@163.117.206.10) |
| 09:47.15 | *** join/#asterisk-dev frawd (~francois@23.Red-81-38-28.dynamicIP.rima-tde.net) |
| 09:49.27 | irroot | schmidts hi there ill look at the code latter but it looks like youve nailed it |
| 09:54.02 | schmidts | i guess so ;) |
| 09:59.32 | schmidts | i love this machine :D 1980 active channels |
| 09:59.32 | schmidts | 1980 active calls |
| 09:59.32 | schmidts | Load average: 0.20 0.11 0.09 |
| 09:59.50 | irroot | nice |
| 09:59.52 | wdoekes2 | nice |
| 10:00.18 | schmidts | dual 6 core with hyper threading ;) |
| 10:01.05 | irroot | the eu green cops are coming looking for you :P |
| 10:02.00 | schmidts | not really ;) this is one of my cluster maschines and right now there are 6 production servers running on it. |
| 10:02.14 | schmidts | After i have virtualized a lot of services i have turned off 4 servers |
| 10:14.26 | irroot | schmidts looked at the source your patch makes sense and works for me |
| 10:15.11 | *** join/#asterisk-dev tzafrir_laptop (~tzafrir@local.xorcom.com) |
| 10:15.11 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 10:19.37 | *** join/#asterisk-dev seanbright (~sean@asterisk/contributor-and-bug-marshal/seanbright) |
| 10:19.37 | *** mode/#asterisk-dev [+o seanbright] by ChanServ |
| 10:25.50 | wdoekes2 | schmidts: (1) what happens if the peer was not in the peers_by_ip? (2) you may want to alter the comments to: /* we've unlinked the peer from the peers container but not unlinked from the peers_by_ip container yet */ (see leifs "segfault in ..." asterisk-dev post) |
| 10:27.40 | schmidts | wdoekes2 nothing will happens, ao2_unlink is only a callback with FLAG OBJ_UNLINK |
| 10:27.54 | schmidts | if this will not find it, nothing will happens but i will change the comment |
| 10:30.33 | *** join/#asterisk-dev irroot (~irroot@196-215-124-168.dynamic.isadsl.co.za) |
| 10:31.45 | schmidts | wdoekes2 which post of leif do you mean? havent found it in my mail history ;) |
| 10:32.05 | *** join/#asterisk-dev irroot (~irroot@196-215-124-168.dynamic.isadsl.co.za) |
| 10:35.23 | wdoekes2 | hm.. now I cannot find it either :D I'll check later |
| 10:36.16 | schmidts | ok :D but thanks for your suggestion, it really makes more sense this way |
| 11:14.07 | *** join/#asterisk-dev irroot (~irroot@196-215-124-168.dynamic.isadsl.co.za) |
| 11:15.32 | schmidts | hah this time i have done it wrong in the first step ;) |
| 11:15.34 | oej | That was a very important fix. |
| 11:16.38 | schmidts | especially for me :D reloading > 4000 peers every 15 minutes will burn memory faster than anything ;) |
| 11:17.20 | schmidts | even on a machine with 12GB of ram :D |
| 11:19.35 | oej | Agreed |
| 11:28.42 | *** join/#asterisk-dev hehol (~hehol@2001:1438:1009:200:cb:75ea:9af6:3e16) |
| 11:31.44 | *** join/#asterisk-dev snuff-home (~snuffy@203-219-29-182.tpgi.com.au) |
| 11:31.44 | *** mode/#asterisk-dev [+o snuff-home] by ChanServ |
| 11:42.58 | *** join/#asterisk-dev kpfleming (~kpfleming@asterisk/digium-director-of-software-tech/kpfleming) |
| 11:42.58 | *** mode/#asterisk-dev [+o kpfleming] by ChanServ |
| 11:56.59 | *** join/#asterisk-dev cemeh4 (~cemeh4@193.19.228.130) |
| 12:00.39 | *** join/#asterisk-dev cemeh4 (~anonymous@193.19.228.130) |
| 12:00.58 | cemeh4 | hi all |
| 12:02.29 | cemeh4 | i'm creating custom application for ast |
| 12:02.56 | cemeh4 | 2 questions |
| 12:04.32 | cemeh4 | 1) one channel in more than one bridge - is it normal? |
| 12:05.01 | russellb | no. |
| 12:05.03 | cemeh4 | 2) if not: how can i take data from bridged channel and send it to another channel? |
| 12:05.56 | russellb | look it up, ast_channel_get_by_name() or similar functions in channel.h |
| 12:06.17 | *** join/#asterisk-dev anonymouz666 (~anonymouz@187-28-37-118.poolip.RJO.embratel.net.br) |
| 12:06.26 | cemeh4 | i have channel pointer already |
| 12:07.34 | cemeh4 | i need a way to read frames from it and send them to another channel but first channel need to stay in bridge |
| 12:07.56 | russellb | well i dunno, i need coffee i think |
| 12:08.08 | schmidts | russellb good morning :D |
| 12:08.16 | schmidts | cemeh4 take a look at the framehook api |
| 12:09.34 | cemeh4 | can you say a few words about that api? |
| 12:10.29 | schmidts | cemeh4 here you go: http://www.asterisk.org/doxygen/asterisk1.8/AstFrameHookAPI.html |
| 12:10.36 | cemeh4 | ok thx |
| 12:10.44 | schmidts | your welcome |
| 12:21.33 | *** part/#asterisk-dev cemeh4 (~anonymous@193.19.228.130) |
| 12:26.35 | schmidts | wdoekes2 about your review 1395 i think my patch has solved a little bit of the memory leak you have found |
| 13:01.55 | *** join/#asterisk-dev mjordan (~mjordan@nat/digium/x-eknenhabzhuvpeiq) |
| 13:01.55 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 13:05.12 | *** join/#asterisk-dev serafie (~erin@nat/digium/x-mhevgqczoxscuxdo) |
| 13:10.14 | oej | russellb: Thanks for the quick review. I've updated now. |
| 13:10.36 | russellb | np |
| 13:17.48 | *** join/#asterisk-dev anny__ (~chatzilla@77.42.241.114) |
| 13:18.57 | anny__ | hey all, I have a problem with asterisk not being able to handle a high limit of inbound calls, when these limits are reached clients calls will begin to drop |
| 13:22.38 | kpfleming | anny__: all software has limitations |
| 13:23.50 | anny__ | kpfleming: :) yeah i know, but are you aware of the limitations asterisk can handle, is it hardware dependent ? |
| 13:24.11 | kpfleming | of course it is |
| 13:24.26 | kpfleming | asterisk has no inherent limitation on the number of calls it can handle |
| 13:25.06 | anonymouz666 | in the praticse I wouldn't recommend more than 200 calls in one machine, even with a great hardware. |
| 13:25.22 | anonymouz666 | very strange things happens |
| 13:25.52 | kpfleming | that depends a lot on what you are doing with the calls, and what version of asterisk you are running |
| 13:25.57 | oej | We are running constantly with more than 400 calls |
| 13:26.06 | kpfleming | there are a large number of factors that influence how many calls can be handled |
| 13:26.23 | kpfleming | for example, if you are recording 50% of your calls, then the number you can handle will drop dramatically |
| 13:26.53 | anonymouz666 | oej: what rtp mode? |
| 13:27.01 | oej | p2p bridge |
| 13:27.04 | oej | Mostly |
| 13:27.09 | anonymouz666 | we can't do that :-) |
| 13:27.27 | oej | . . . using Asterisk 1.4 |
| 13:27.54 | anonymouz666 | we can't run p2p bridge in inboud callcenter setup |
| 13:28.18 | kpfleming | why? |
| 13:28.19 | anonymouz666 | oej: asterisk 1.8 can handle the same 200 calls here with lower load. |
| 13:28.40 | anonymouz666 | but I almost lost my job because other issues ;-) |
| 13:29.21 | oej | That's bad. YOu should be able to handle much more. There are reasons why I am not using 1.8 though... |
| 13:29.29 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 13:29.29 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 13:32.15 | anonymouz666 | kpfleming: how can we do native bridge (p2p is native bridge for rtp channels, right) with different techs? |
| 13:33.14 | anonymouz666 | oej: very hard to setup big callcenters without distribute device state. |
| 13:33.19 | anonymouz666 | is a must |
| 13:33.51 | kpfleming | if you are using two channel technologies that both use RTP, then p2p bridging should work. |
| 13:34.00 | kpfleming | if one of your channel types is not using RTP, then obviously it can't be done. |
| 13:34.08 | anonymouz666 | that's why. |
| 13:34.31 | oej | anonymouz: There's a reason why I am working on Pinana |
| 13:34.42 | anonymouz666 | there are only 240 dahdi channels getting calls :-) on each machine |
| 13:35.13 | anonymouz666 | oej: res_ais stuff works FINE |
| 13:35.17 | anonymouz666 | (also) |
| 13:35.21 | oej | The wiki is unclear on merging from 1.8 to trunk. Which script should I use. |
| 13:35.34 | oej | anonymous: I prefer using SIP - other apps can use it. |
| 13:36.11 | *** join/#asterisk-dev merbzt (~benjamin@193.12.106.2) |
| 13:37.03 | anonymouz666 | can't imagine you doing that without a SIP PROXY involved. |
| 13:37.15 | oej | Asterisk2asterisk works fine |
| 13:37.21 | oej | But I'm also using Kamailio |
| 13:37.53 | anonymouz666 | that was my first thought |
| 13:37.55 | oej | Anyone that can enlighten me in which repotools script to use for merging from 1.8 upwards? My old svnmerge command doesn't work. |
| 13:38.11 | russellb | 1.8 -> 10 -> trunk |
| 13:38.20 | russellb | so, merge810 <rev>, probably |
| 13:39.02 | oej | Ok, feels dangerous. Trying it |
| 13:39.38 | russellb | you'll have to update your repotools install if you haven't updated since the 10 branch was created |
| 13:39.53 | oej | It's always updated :-) But thanks for the reminder |
| 13:40.04 | russellb | k |
| 13:40.34 | anonymouz666 | oej: you have the pinana-subscrive and pinana-publish? |
| 13:40.58 | oej | anonymous: Still working on it. Dangerous code. Not ready :-) |
| 13:41.34 | oej | pinana-publish-1.4 is where the cool stuff is. PUBLISH/SUBSCRIBE works, but I have some issues with the new thread I'm trying to implement. |
| 13:41.56 | oej | Will resume work with it this week |
| 13:45.00 | anonymouz666 | using SIP to do that I think the delay between the machines could be higher |
| 13:45.32 | oej | I don't see commit message file |
| 13:48.02 | oej | Found it. |
| 13:49.20 | leifmadsen | ya I've used merge810 and merge10trunk recently -- they work great ;) |
| 13:49.32 | schmidts | i have used it today :D |
| 13:50.26 | opticron | ditto :) |
| 13:51.10 | oej | Ok, no more crashes in voicemail notifications. Seems like a GOOD thing (tm) |
| 13:58.40 | schmidts | oej +1 |
| 14:01.14 | *** join/#asterisk-dev putnopvut (~putnopvut@asterisk/master-of-queues/mmichelson) |
| 14:01.14 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 14:13.58 | *** join/#asterisk-dev The_Boy_Wonder (~vossel@asterisk/batman-developer/dvossel) |
| 14:13.58 | *** mode/#asterisk-dev [+o The_Boy_Wonder] by ChanServ |
| 14:23.52 | *** join/#asterisk-dev otherwiseguy (~terry@69.76.217.40) |
| 14:25.59 | *** join/#asterisk-dev irroot (~irroot@196-215-124-168.dynamic.isadsl.co.za) |
| 14:27.37 | irroot | putnopvut thx for the info on the locking in app_queue turns out with ao2 there should be not need to lock queues container |
| 14:28.34 | *** join/#asterisk-dev sruffell (~Adium@asterisk/the-kernel-guy/sruffell) |
| 14:28.34 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 14:37.32 | *** join/#asterisk-dev jeffg (~jeffg@pdpc/supporter/professional/jeffg) |
| 14:43.28 | oej | should sayunixtime add a lot of prompts outside of my own format? |
| 14:43.44 | oej | Today, it does in many languages, so it's kind of out of control. |
| 14:43.55 | oej | I thought the idea was to leave the dialplan in control FULLY |
| 14:44.05 | oej | I need a sayunixtimenofrills() |
| 14:54.28 | Kobaz | it doesn't just use lang/digits/x ? |
| 14:55.30 | Corydon76-home | oej: %q and %Q are designed to be "easy" format options. |
| 14:56.26 | Corydon76-home | Kobaz: the unixtime is the input, date and time are the output |
| 14:56.33 | oej | sayunixtime adds words in the middle of formats |
| 14:56.50 | Corydon76-home | oej: should only for %q and %Q |
| 14:56.54 | Kobaz | ah, i don't think i've ever used sayunixtime |
| 14:57.11 | Kobaz | oh, yeah |
| 14:57.12 | oej | In Danish, it adds stuff. Need to review this again. |
| 14:57.17 | oej | Bye for now. |
| 14:57.20 | Kobaz | i have used it, i remember it adding filler words |
| 14:57.35 | oej | Yes it does, which means we have no "clean" format |
| 14:57.39 | Corydon76-home | Kobaz: only if you don't specify the format |
| 14:57.51 | Kobaz | i remember putting in a format, and having it add words |
| 14:57.59 | Corydon76-home | Really? |
| 14:58.01 | Kobaz | yeah |
| 14:58.11 | Corydon76-home | It's the core that is doing all the work... |
| 14:59.06 | Corydon76-home | ast_say_date_with_format() |
| 14:59.08 | anonymouz666 | Corydon76-home: do you mind if I ask you a user question? I can't see you in #asterisk anymore |
| 14:59.39 | anonymouz666 | it's about realtime queue members |
| 15:02.00 | Corydon76-home | anonymouz666: No |
| 15:02.30 | Corydon76-home | anonymouz666: ask someone in #asterisk. I am not there because my present job isn't paying me to answer Asterisk questions |
| 15:03.11 | anonymouz666 | alright, thanks to clarify. |
| 15:17.30 | *** join/#asterisk-dev kpfleming (~kpfleming@asterisk/digium-director-of-software-tech/kpfleming) |
| 15:17.30 | *** mode/#asterisk-dev [+o kpfleming] by ChanServ |
| 15:22.19 | acidfoo | someone have any idea about this: http://forums.digium.com/viewtopic.php?f=1&t=79463 |
| 15:22.30 | opticron | is there anyone here well versed in the new extensions pattern matching engine? I'd appreciate a review on https://reviewboard.asterisk.org/r/1367/ |
| 15:22.32 | acidfoo | im not the one who wrote the post, but it seems that we have the same problem here |
| 15:25.36 | opticron | I'd use genconf from a slightly older version of dahdi if possible |
| 15:25.50 | opticron | or maybe from dahdi trunk |
| 15:27.42 | acidfoo | we are using 2.5.1 |
| 15:28.15 | *** join/#asterisk-dev anthm (~anthm@freeswitch/developer/anthm) |
| 15:35.57 | pabelanger | mjordan: psst... #asterisk-testing |
| 15:38.33 | merbzt | how does a channel handle dtmf input from an endpoint ? |
| 15:38.44 | merbzt | or dtmf events |
| 15:38.59 | kpfleming | however it wants... the answer is dependent on the technology it implements |
| 15:41.54 | merbzt | I'm looking through some channels but I cant figure out how to form a sequence of numbers from the dtmf events |
| 15:42.23 | kpfleming | channel drivers don't do that, at least none of the ones currently implemented do |
| 15:42.52 | anonymouz666 | there's an old bug report (https://issues.asterisk.org/view.php?id=10017), that Qwell's says "Fixed in svn branches 1.4 and trunk in revisions 71915 and 71916" - how can I see today the diff in these revisions? I am getting a core exactly the same way as is written in this issue. |
| 15:43.17 | kpfleming | svn diff -c71915 http://svn.digium.com/svn/asterisk |
| 15:43.47 | merbzt | ok, so the channels only relay a dtmf event at a time ? |
| 15:43.57 | kpfleming | yes |
| 15:44.08 | kpfleming | or they turn them into audio, if that is required |
| 15:44.15 | kpfleming | channel drivers don't interpret DTMF sequences at all |
| 15:44.53 | merbzt | so I just make an ast_frame of it then ? |
| 15:45.48 | kpfleming | you haven't described what you are trying to do. your first question was about dtmf events, which i assumed was dtmf frames you received from the asterisk core. |
| 15:46.14 | kpfleming | if you are talking about how to handle dtmf input you've received from an external endpoint, yes, you create the appropriate dtmf frames and feed them to the core. |
| 15:47.44 | merbzt | the endpoint driver has an ioctl that when polled will return events, dtmf is among them |
| 15:48.13 | kpfleming | ok, then yes, this is similar to the way DAHDI can work if the hardware reports DTMF events (but most DAHDI hardware doesn't) |
| 15:49.16 | merbzt | kpfleming: does the core have logic to decide when to initiate a call ? |
| 15:49.33 | kpfleming | applications that initiate calls do |
| 15:49.40 | kpfleming | app_dial, app_queue, etc. |
| 15:52.13 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 15:52.13 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 15:53.16 | merbzt | tnx, I'll go back to reading the code |
| 15:53.38 | *** join/#asterisk-dev moy (~moy@173.239.155.74) |
| 16:16.06 | anny__ | if a sip account connected to pri is calling an IAX client through asterisk, does this need any transcoding? |
| 16:18.35 | Kobaz | depends on the codecs |
| 16:19.01 | kpfleming | how is a sip account connected to a pri? that doesn't make any sense. |
| 16:19.23 | kpfleming | but as Kobaz says, the channel technology is irrelevant, all that matters is the codec in use on each channel. |
| 16:19.30 | Kobaz | he means sip peer whose context has a Dial(DAHDI/xyz) |
| 16:19.48 | anny__ | yes, exactly |
| 16:19.49 | kpfleming | but then it's not calling an IAX client |
| 16:19.55 | kpfleming | it's calling a DAHDI channel |
| 16:19.57 | anny__ | no |
| 16:20.17 | anny__ | it's a DAHDI Channel being rerouted to an IAX client |
| 16:20.21 | Kobaz | sip peer -> dial(dahdi) -> other end of dahdi Dial(IAX..) |
| 16:20.36 | kpfleming | pictures make things better :-) |
| 16:21.48 | *** join/#asterisk-dev moy (~moy@CPE002719f00364-CM0026f3a1701d.cpe.net.cable.rogers.com) |
| 16:35.56 | *** join/#asterisk-dev mnicholson (~mnicholso@nat/digium/x-kmdqqqzsjwekbzlw) |
| 16:35.56 | *** mode/#asterisk-dev [+o mnicholson] by ChanServ |
| 18:10.41 | seanbright | so who do i beg for hardware discounts around here? |
| 18:10.59 | Corydon76-home | That would be malcolmd |
| 18:11.31 | seanbright | malcolmd: i hereby request a hardware discount |
| 18:11.32 | seanbright | heh |
| 18:11.47 | kpfleming | malcolmd: i authorize a 0.05% discount for seanbright |
| 18:11.53 | seanbright | oh snap |
| 18:12.05 | file | I authorize a 25% markup for seanbright |
| 18:12.39 | pabelanger | divides by zero |
| 18:12.48 | malcolmd | that should save him a bundle. if he buys a hundred million bucks worth of stuff, he might be able to fence it and make a profit |
| 18:13.18 | seanbright | they're on to me |
| 18:13.45 | malcolmd | we'll have to wait for the check to clear first, naturally |
| 18:13.46 | Qwell | malcolmd: We should charge him more in that case. He can just make it up on volume. |
| 18:14.26 | seanbright | was hoping to save a bit on a new transcoder purchase |
| 18:14.52 | seanbright | the money i would save would serve as... say... 'sponsorship' for my continued asterisk development efforts |
| 18:14.59 | seanbright | stops trolling |
| 18:16.06 | Qwell | mjordan: Can you make that more global? |
| 18:16.20 | Qwell | That's still being done per test, when it doesn't need to (I don't think?) |
| 18:21.21 | mjordan | Qwell: its going to be really hard to get rid of that one folder |
| 18:21.39 | mjordan | the Dependency class has no knowledge of the name or relative path of the test about to be performed |
| 18:21.52 | mjordan | What's more, the Asterisk class creates all of its directories during object instantiation |
| 18:22.10 | Qwell | oh well |
| 18:22.20 | mjordan | We could change either point, but then we'd effect everything else using it. I think that's the right thing to do, but the quicker fix that at least got us down to one folder was two lines of code, so I got it in. |
| 18:22.35 | mjordan | The other options I'd rather put up for review :-) |
| 18:52.54 | *** join/#asterisk-dev irroot (~irroot@41.53.130.200) |
| 19:02.39 | *** join/#asterisk-dev philipp64|laptop (~chatzilla@50-39-153-143.bvtn.or.frontiernet.net) |
| 19:18.24 | jmls | shouldn't make kick in the knackers if you haven't yet done a ./configure ? |
| 19:18.43 | jmls | make as in "make update, make clean etc" |
| 19:19.27 | jmls | there's a missing "you" as well -> "kick you in the knackers" |
| 19:20.05 | opticron | yes and it does as far as I've seen |
| 19:21.16 | jmls | 's privates are still intact. something has gone wrong ;) |
| 19:21.51 | jmls | fresh svn checkout of -10, make clean and make update both crapped out with all sorts of errors because ./configure had not been run |
| 19:24.54 | *** join/#asterisk-dev snuff-work (~snuff-wor@210.9.82.197) |
| 19:24.54 | *** join/#asterisk-dev Kobaz (~kobaz@its.kobaz.net) |
| 19:24.55 | *** join/#asterisk-dev file (~file@2001:4978:2d7:0:20c:29ff:feff:49c0) |
| 19:25.02 | *** mode/#asterisk-dev [+o snuff-work] by ChanServ |
| 19:25.19 | *** join/#asterisk-dev Corydon76-home (eight@c-69-137-80-31.hsd1.tn.comcast.net) |
| 19:25.19 | *** mode/#asterisk-dev [+o Corydon76-home] by ChanServ |
| 19:34.37 | *** join/#asterisk-dev elguero (~miguel323@2001:470:1f06:12c4::2) |
| 19:39.59 | *** join/#asterisk-dev Dovid (~Dovid@213.8.121.90) |
| 19:42.25 | Dovid | anyone know how hard it would be to create patch for asterisk that will re-invite a call with a specific codec. For instance say a call is using G729 and I want to re-invite if I want the call to use G711U |
| 19:43.42 | opticron | you mean change the codec mid-call? or do you just want to always use a specific code? |
| 19:43.45 | opticron | *codec |
| 19:43.57 | Dovid | mid call. pre-answer i know how to do it |
| 19:44.10 | Dovid | set(SIP_CODEC=G711u) |
| 19:44.16 | Dovid | it's once the call is answered |
| 19:45.51 | opticron | I don't have enough experience with SIP to answer that |
| 19:46.46 | opticron | I want to say that asterisk doesn't currently support mid-call codec renegotiation, but I'm not absolutely sure |
| 19:47.40 | jsmith | It does on a re-invite... |
| 19:49.48 | Dovid | jsmith: afaik it will accept a re-invite if given one but is there a way for Asterisk to intiate the re-invite with a specific codec "on demand" |
| 19:50.45 | russellb | nope. |
| 19:51.10 | leifmadsen | ASTERISK-18240 |
| 19:51.13 | DocAwesome | [ASTERISK-18240] [Status: Triage] UniqueID is posted twice to CDR when and CDR(userfield) is extended/broken in the process. - https://issues.asterisk.org/jira/browse/ASTERISK-18240 |
| 19:51.19 | leifmadsen | I can't determine if this is expected/normal or a bug... |
| 19:51.23 | Dovid | russelb: How hard and how involved do you think it would be to create such a patch |
| 19:51.55 | russellb | hard. |
| 19:52.19 | leifmadsen | you could stick a SIP proxy in front to handle that right? |
| 19:55.08 | leifmadsen | pabelanger: hey! |
| 19:57.12 | pabelanger | leifmadsen: word |
| 19:57.20 | leifmadsen | pabelanger: hey can I get you to do me a quick favour? |
| 19:57.21 | anonymouz666 | yes, SIP PROXY can do that easily. |
| 19:57.26 | pabelanger | leifmadsen: sup |
| 19:57.30 | kpfleming | but something still has to trigger the renegotiation |
| 19:57.31 | leifmadsen | pabelanger: can you turn down the suck? |
| 19:57.40 | pabelanger | leifmadsen: yes |
| 19:57.43 | leifmadsen | pabelanger: thanks! |
| 19:57.51 | pabelanger | turns up the suck |
| 19:57.55 | kpfleming | asterisk will respond to a reinvite and change codecs, but it will not ask the bridged channel(s) to do the same |
| 19:58.00 | anonymouz666 | Dovid: you can look at rtpmap and rewrite it if you want |
| 19:58.39 | kpfleming | actually... it really shouldn't be all that hard. we already have the ability to accept a request to switch to T.38 and then forward that across the bridge to await a response. |
| 19:58.55 | kpfleming | doing a similar thing for voice/video codecs would not be terrible. |
| 20:03.37 | Dovid | kpfleming: Where is the code for t.38 and how is it triggerd ? i don't know c but I can take a peak |
| 20:05.03 | kpfleming | there is no 'the code for T.38'... there is T.38 related code in chan_sip and other places (all channel drivers that understand T.38, res_fax, etc.) |
| 20:05.28 | kpfleming | but it's not something that could just be duplicated and modified for this purpose |
| 20:05.37 | kpfleming | it's just a hint about how it could be handled |
| 20:06.26 | Dovid | kpfleming: Ok. That is what I was asking. I assume that one can look at the code for T.38 in chan_sip that triggers the re-invite and have something else call it from the dial plan. |
| 20:06.53 | Dovid | kpgelming: If you had to "guess" and a real rough guess, how hard would it be to create such a patch for 1.8.x ? |
| 20:08.04 | kpfleming | nothing in chan_sip triggers the reinvite... the triggering is done by the external endpoint, or by an application in asterisk noticing FAX answering tones. |
| 20:08.23 | kpfleming | it can't be done from the dialplan, because the dialplan is not 'running' while a call is up and bridged |
| 20:09.18 | kpfleming | if the goal of the patch was to allow codec renegotiation invites to make their way from one SIP endpoint through Asterisk to another SIP endpoint and back... i would think that could be done by an accomplished asterisk/sip developer in probably a few weeks |
| 20:10.42 | Dovid | kpfleming: Ok. My reason for this is because on a "normal system" we want to use G729 because it's apss thorugh and saves on bandwdith. The issue is say with a conference call. We don't want to bring the machine to it's knee's. we want to re-invite to ulaw before going to meetme |
| 20:11.58 | irroot | Dovid if you transfering to the conf via dialplan there is some hope for you\ |
| 20:12.02 | Dovid | so actually it would be before the call is brifgen |
| 20:12.13 | Dovid | irroot: How's that? |
| 20:12.15 | kpfleming | ok, yeah, that would be slightly easier, but of less general use :-) |
| 20:12.42 | irroot | Dovid you could have "reinvite dialplan trigger" |
| 20:12.54 | Dovid | kpfelimg: Not if we record the call or use meetme. It will save a lot on the CPU. would it then be easier to write something or is it as complex ? |
| 20:13.03 | Dovid | iroot: Thats what I want |
| 20:13.21 | jmls | how does ./configure determine if pgsql is installed ? |
| 20:13.24 | irroot | some func i would guess that will be set with the codec list for invite |
| 20:13.30 | jmls | I have /usr/pgsql-9.1/bin/psql |
| 20:13.49 | irroot | jmls i would guess the include files |
| 20:14.17 | jmls | Package postgresql91-devel-9.1.0-2PGDG.rhel5.i386 already installed and latest version |
| 20:14.21 | kpfleming | the configure script checks for headers and libraries, and ensures it can compile and link against them |
| 20:16.25 | jmls | right. so I think that I have the -devel package installed |
| 20:16.36 | anonymouz666 | I love the issues that you can't reproduce in any way :/ |
| 20:16.58 | jmls | anonymouz666: that's what makes this so ... much ... fun ... |
| 20:17.04 | Corydon76-home | jmls: Did that install an executable named "pg_config"? |
| 20:17.26 | kpfleming | jmls: did you look in config.log to see what it is complaining about? |
| 20:17.26 | jmls | yup. /usr/pgsql-9.1/bin/pg_config |
| 20:18.06 | kpfleming | what distro are you running that has postgres 9.1 already? |
| 20:18.08 | Corydon76-home | jmls: Okay, so: ./configure --with-postgresql=/usr/pgsql-9.1 |
| 20:18.20 | jmls | oooooohhhh |
| 20:18.32 | Corydon76-home | Actually, --with-postgres=... |
| 20:18.58 | Corydon76-home | Should probably make that an alias |
| 20:19.43 | jmls | looking good .... |
| 20:20.01 | jmls | oh! oh! Bingo! |
| 20:20.12 | jmls | points awarded to Corydon76-home |
| 20:20.24 | jmls | thanks |
| 20:20.28 | kpfleming | for that to work at runtime you are probably going to have to have /usr/pgsql-9.1/lib in your ld configuration, or asterisk won't be able to find the libraries |
| 20:21.34 | Corydon76-home | kpfleming: I'm thinking we already do that |
| 20:21.58 | kpfleming | i don't follow |
| 20:22.00 | jmls | so the file "postgresql-9.1-libs.conf " in ld.so.conf.d would do, right ? |
| 20:22.03 | Corydon76-home | The pg_config utility specifies the necessary -L and -l flags to ensure that it links |
| 20:22.19 | kpfleming | yes, but that doesn't put anything into the final binary to tell it where to find the libraries |
| 20:22.25 | kpfleming | that only tells the linker where to find them |
| 20:22.29 | kpfleming | jmls: yes, probably |
| 20:22.50 | jmls | cool. |
| 20:23.01 | jmls | is glad that the file was already there ;) |
| 20:23.09 | file | is always there |
| 20:23.17 | jmls | "the" file ? |
| 20:23.30 | jmls | the "one and only" file ? |
| 20:23.36 | file | the one file to rule them all |
| 20:23.42 | jmls | lol |
| 20:24.01 | jmls | frodo tosses file into the fires of Mt Doom |
| 20:25.46 | file | yay fire |
| 20:28.13 | Dovid | haha |
| 20:38.31 | irroot | thx jmls ill enjoy file company down here in hades |
| 20:50.23 | *** join/#asterisk-dev chazzam (~chazz@173-24-236-90.client.mchsi.com) |
| 21:08.23 | *** join/#asterisk-dev otherwiseguy_ (~terry@69.76.217.40) |
| 21:10.24 | *** join/#asterisk-dev otherwiseguy__ (~terry@69.76.217.40) |
| 21:11.27 | *** join/#asterisk-dev otherwiseguy_ (~terry@69.76.217.40) |
| 21:11.55 | *** join/#asterisk-dev otherwiseguy__ (~terry@69.76.217.40) |
| 21:12.37 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 21:12.37 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 21:28.44 | *** join/#asterisk-dev wdoekes2 (~walter@wjd.osso.nl) |
| 21:30.23 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 21:30.23 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 21:35.33 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 21:35.33 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 21:45.07 | malcolmd | leifmadsen: why you gotta fill up my inbox at 4:45pm? ;) |
| 21:45.23 | leifmadsen | malcolmd: because it's 5:45pm here? :) |
| 21:47.16 | chazzam | leifmadsen does not believe in freedom |
| 21:47.24 | leifmadsen | true dat |
| 21:53.26 | *** part/#asterisk-dev mjordan (~mjordan@nat/digium/x-eknenhabzhuvpeiq) |
| 22:11.12 | *** join/#asterisk-dev lucifurr (~lucifurr@65.172.54.254) |
| 22:18.02 | kpfleming | pabelanger: you can use a hardlink instead of a symlink, as long as the directories where you need the log files to appear are on the same filesystem |
| 22:26.21 | wdoekes2 | leifmadsen: I thought you posted a "please do not use commit message like \"segfault when ...\", like you're introducing bugs"-post recently, but I cannot seem to find it |
| 22:26.42 | leifmadsen | wdoekes2: oh I think I just said that on IRC :) |
| 22:27.37 | *** join/#asterisk-dev seraphie (~erin@75.76.38.159) |
| 22:28.24 | wdoekes2 | that would explain why I didn't find it =) |
| 22:28.28 | leifmadsen | found it |
| 22:28.29 | leifmadsen | <leifmadsen> A gentle reminder to everyone: when writing your commit messages, please word them in such a way that it reads that you're fixing an issue, and not introducing an issue. For example, don't word your commits like this: Segfault when publishing device states via XMPP and not connected. |
| 22:29.00 | wdoekes2 | thanks |
| 22:29.15 | leifmadsen | wdoekes2: also for reference: https://wiki.asterisk.org/wiki/display/AST/Commit+Messages |
| 22:29.47 | leifmadsen | First line should be an <= 80 char one line message that says what you fixed/changed. More verbose message below that. |
| 22:35.12 | *** part/#asterisk-dev lucifurr (~lucifurr@65.172.54.254) |