| 02:52.49 | *** join/#asterisk-dev infobot (~infobot@rikers.org) |
| 02:52.49 | *** topic/#asterisk-dev is Asterisk Development Discussion -=- http://www.asterisk.org/developers -=- Tier 2 and 3.14159265 support is in #asterisk -=- Tier 4 support is in #2,000 |
| 04:29.48 | *** join/#asterisk-dev moy_ (~moy@70.48.10.32) |
| 04:41.36 | *** join/#asterisk-dev irroot (~gregory@dsl-185-76-175.dynamic.wa.co.za) |
| 07:22.07 | *** join/#asterisk-dev schmidts (~schmidts@Lo20-rt01.lm33.sil.at) |
| 07:22.08 | schmidts | good morning |
| 07:40.38 | *** join/#asterisk-dev tzafrir (~tzafrir@local.xorcom.com) |
| 07:40.38 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 07:46.53 | *** join/#asterisk-dev hehol (~Adium@2a01:198:71d:0:21f:d0ff:fea1:568e) |
| 07:49.09 | *** join/#asterisk-dev sezuan (bouncer@irc.scheff32.de) |
| 08:03.48 | *** join/#asterisk-dev irroot (~gregory@196-210-238-87.dynamic.isadsl.co.za) |
| 08:29.00 | *** join/#asterisk-dev tamiel (~tamiel@213.30.183.226) |
| 09:35.47 | *** join/#asterisk-dev hehol (~hehol@2001:1438:1009:200:6912:49e2:2d71:3f22) |
| 09:39.17 | *** join/#asterisk-dev happylife (~happylife@212.92.145.7) |
| 10:07.44 | *** join/#asterisk-dev Malak123 (~chatzilla@77.42.241.114) |
| 10:08.30 | *** join/#asterisk-dev irroot (~gregory@196-210-238-87.dynamic.isadsl.co.za) |
| 10:08.55 | Malak123 | im facing the following issue on my local server https://issues.asterisk.org/view.php?id=10709, could someone tell me if it's resolved? |
| 10:25.45 | wdoekes2 | Malak123: there is no support for 1.4 |
| 10:25.57 | wdoekes2 | if you face the same issue in 1.8, create a new ticket |
| 10:42.00 | schmidts | irroot it really looks my problems are solved ;) 1 week uptime and 100.000 calls processed without any problems \o/ |
| 10:42.20 | irroot | yo that is super cool |
| 10:42.52 | schmidts | i only had some hanging calls but i also had them with 1.2 so i guess this is more a client problem than from asterisk itself |
| 11:01.35 | *** join/#asterisk-dev cemeh4 (~cemeh4@193.19.228.130) |
| 12:07.28 | schmidts | irroot what do i need to get fax_receive running? |
| 12:12.08 | schmidts | and how far are you with your T38<-> t30 gateway? is it allready in 10? |
| 12:16.17 | schmidts | ok i allready found ReceiveFax ;) why is this not called fax_receive? :D |
| 12:38.24 | irroot | schmidts lol dude T38 Gw is in 10 |
| 12:38.54 | irroot | there is a backport branch in svn team/irroot |
| 12:39.19 | schmidts | irroot thanks thats what i want to hear :D |
| 12:39.47 | schmidts | and what i need to use this stuff? |
| 12:40.50 | irroot | its quite simple a well clocked line ie ISDN to be the T.30 side and then SET(FAXOPT(gateway)=....) to attach the gateway |
| 12:41.00 | irroot | thats about it |
| 12:45.38 | schmidts | and with gateway it means like Dahdi/g1 for example? |
| 13:22.36 | irroot | <PROTECTED> |
| 13:22.38 | irroot | <PROTECTED> |
| 13:22.40 | irroot | schmidtas |
| 13:22.47 | irroot | s/schmidtas/schmidts/ |
| 13:34.08 | schmidts | :D |
| 13:41.29 | irroot | schmidts works with OOH323/SIP ie t38modem + gnugk + hylafax = Astrifax Server |
| 13:58.30 | *** join/#asterisk-dev mjordan (~mjordan@nat/digium/x-yuyuynnmamheuccz) |
| 13:58.30 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 14:16.43 | *** join/#asterisk-dev irroot (~gregory@196-210-238-87.dynamic.isadsl.co.za) |
| 14:21.44 | *** join/#asterisk-dev irroot (~gregory@196-210-238-87.dynamic.isadsl.co.za) |
| 14:24.05 | *** join/#asterisk-dev eliel_ (~quassel@200.32.121.66) |
| 15:08.32 | *** join/#asterisk-dev sruffell (~Adium@asterisk/the-kernel-guy/sruffell) |
| 15:08.32 | *** mode/#asterisk-dev [+o sruffell] by ChanServ |
| 15:12.43 | *** join/#asterisk-dev dwmw2 (~dwmw2@twosheds.infradead.org) |
| 15:18.49 | *** join/#asterisk-dev The_Boy_Wonder (~vossel@asterisk/batman-developer/dvossel) |
| 15:18.49 | *** mode/#asterisk-dev [+o The_Boy_Wonder] by ChanServ |
| 15:33.01 | *** join/#asterisk-dev putnopvut (~putnopvut@asterisk/master-of-queues/mmichelson) |
| 15:33.01 | *** mode/#asterisk-dev [+o putnopvut] by ChanServ |
| 15:34.05 | tzanger | hm, is there a trick in getting gmenuselect to NOT build? I'm specifically saying --without-gtk2 but it still finds GTK libraries and thus says "ZOMG I CAN BUILD A GUI FOR YOU..." followed by an earth shatterring kaboom |
| 15:55.52 | *** part/#asterisk-dev Malak123 (~chatzilla@77.42.241.114) |
| 16:28.08 | *** join/#asterisk-dev vpopov (~happylife@dyn-56-47.dzer.fttbee.kis.ru) |
| 16:28.11 | jrose_atDigium | I don't even know why we have gmenuselect. It's ugly, uninformative, and generally useless. |
| 16:28.21 | Qwell | yes. |
| 16:28.29 | jrose_atDigium | Russell was suggesting that I cull it. |
| 16:28.33 | Qwell | same reason we used to have pbx_gtkconsole. |
| 16:29.22 | jsmith | Shoot it in the head! |
| 16:30.35 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 16:30.35 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 16:30.55 | Qwell | yeah +1 for that |
| 16:31.11 | file | STONITH? |
| 16:31.27 | Qwell | SGMITH |
| 16:31.38 | Qwell | (GM=gmenuselect) |
| 16:31.50 | jrose_atDigium | I wonder if that would count as a tool... so that I can make getting rid of it something on my tool sprint. |
| 16:32.03 | jrose_atDigium | I doubt it'll take more than a little while of course. |
| 16:32.08 | Qwell | it shouldn't take that long |
| 16:32.35 | Qwell | there are a few projects that use menuselect outside of Asterisk though. dahdi-tools for example. |
| 16:33.21 | jrose_atDigium | Yeah, ripple effects always make me worry a little. |
| 16:37.33 | jrose_atDigium | Kobaz: Just about finished with this little deal. http://i.eho.st/pptlvsbp.png |
| 16:37.52 | jrose_atDigium | Just need to figure out how I would save it to the makeopts. |
| 16:38.03 | jrose_atDigium | Which is moderately trivial really... |
| 16:38.29 | jrose_atDigium | A little different in cmenuselect... there it's just controlled with + and - |
| 16:38.52 | jrose_atDigium | But yeah, numerical entries in menuselect should be possible if and when we need them. |
| 16:39.21 | tzanger | woot, I think I got the menuselect working |
| 16:39.33 | tzanger | (cd menuselect ; PKGCONFIG=No ./configure) |
| 16:39.39 | tzanger | that seems to be the only way to kill it |
| 16:39.49 | tzanger | I can't seem to say "JUST IGNORE GTK FFS" |
| 16:40.31 | tzanger | I've never, ever used gtkconsole or any GUI ocnsole. I should have :-) |
| 16:40.47 | Qwell | tzafrir: install the newt devel libs. It has higher priority than gtk |
| 16:40.57 | tzanger | orly |
| 16:41.04 | tzanger | interesting |
| 16:41.08 | jrose_atDigium | It's also prettier than the curses interface. |
| 16:41.22 | tzanger | I've always been happy with curses |
| 16:41.27 | tzanger | but I'm a crotchety old fart |
| 16:41.33 | tzanger | just ask kpfleming or file |
| 16:41.35 | jrose_atDigium | Though has a conspicuous lack of invaders, presumably from space. |
| 16:42.09 | jrose_atDigium | I wonder if it would be at all possible to get our interns to port that to nmenuselect... |
| 16:42.14 | file | it's true |
| 16:42.21 | tzanger | ... invaders? |
| 16:42.28 | tzanger | am I missing something? |
| 16:42.37 | Qwell | tzanger: You are missing invaders, clearly. |
| 16:42.43 | tzanger | indeed |
| 16:42.50 | jrose_atDigium | I presume you've never pressed 'i' from cmenuselect then. |
| 16:42.57 | tzanger | indeed I haven't |
| 16:43.11 | Kobaz | i've beeen seeing these deadlocks for a few months now |
| 16:43.13 | Kobaz | http://pastebin.com/JmheFKNa |
| 16:43.41 | tzanger | how are you getting them? |
| 16:43.51 | Kobaz | no idea yet |
| 16:43.53 | Kobaz | lots of calls |
| 16:43.58 | Kobaz | lots of transfers |
| 16:44.17 | *** join/#asterisk-dev outtolunc (~outtolunc@adsl-66-218-53-172.dslextreme.com) |
| 16:44.21 | jrose_atDigium | Confounding confounders... |
| 16:45.02 | *** join/#asterisk-dev Katty (~Katty@mail.copi-rite.com) |
| 16:45.15 | Katty | Corydon76-home: that's a mighty nice lookin bun you've got there! |
| 16:45.26 | Katty | Corydon76-home: a fine rug |
| 16:48.04 | Corydon76-home | Katty: he's moved since. Probably under or near the Christmas tree |
| 16:48.24 | Qwell | You have a christmas tree already? It's not even thanksgiving. :( |
| 16:48.50 | Katty | my christmas tree is going up this weekend |
| 16:48.57 | Katty | target has a black friday sale. |
| 16:49.01 | Katty | a very pretty one for 30 bucks |
| 16:49.04 | Qwell | as long as it's after thursday, it's all good |
| 16:49.15 | Corydon76-home | Teryl decorates the tree, and since he works in retail, he won't have time after Thanksgiving |
| 16:49.37 | Kobaz | i'm thinking of a longterm project to redo the asterisk threading model to be much more stateful |
| 16:49.49 | Kobaz | and avoid a boatload of deadlock issues |
| 16:49.56 | Katty | hmm yeah. retails going to be a pain this weekend |
| 16:50.00 | Katty | and next week |
| 16:50.06 | Katty | well, even this week it's been rather busy at the mall. |
| 16:50.43 | Corydon76-home | He goes in at 6 am on Friday. His immediate boss has to be there at 1 am. The mall is fining any store that isn't open by 4 am at $1,000/hour. |
| 16:50.58 | Qwell | wtf |
| 16:50.58 | Katty | yeouch. |
| 16:51.05 | Katty | most are opening at midnight this year tho |
| 16:51.20 | Corydon76-home | The department stores will be open at midnight |
| 16:51.27 | jrose_atDigium | Malls are allowed to fine stores for ot being open at 4AM? |
| 16:51.28 | jrose_atDigium | Man |
| 16:51.31 | jrose_atDigium | That's crazy. |
| 16:51.47 | Corydon76-home | On Black Friday, yes. Their lease, their rules. |
| 16:51.52 | jrose_atDigium | True |
| 16:52.16 | Qwell | I'm sure Mastercuts is thrilled to be open at 4. |
| 16:52.33 | Corydon76-home | So naturally, Coach won't open until 6 am. |
| 16:52.33 | Qwell | Because obviously people will need haircuts that early. |
| 16:52.45 | jrose_atDigium | Technically not a store, so maybe they are exempt? |
| 16:52.52 | Qwell | They do sell things. |
| 16:52.55 | jrose_atDigium | True. |
| 16:53.15 | Corydon76-home | Because Coach can make that $2k back in 1 sale |
| 16:55.42 | Kobaz | each channel should have a very specific and well defined path through the code |
| 16:56.02 | Kobaz | my thought is each channel should have a bridging thread and a control thread |
| 16:56.55 | Kobaz | the bridging would go along its merry way passing audio, and if you wanted to do things like set channel variables or process control frames, etc, you would put them in a command queue for the channel, that the control thread would process |
| 16:57.06 | Kobaz | so no a locking b, and b locking a, and deadlock issues |
| 16:58.28 | *** join/#asterisk-dev anthm (~anthm@freeswitch/developer/anthm) |
| 16:58.57 | Qwell | Kobaz: we've been talking about it a bit lately. It's an insanely huge job. |
| 16:59.05 | Corydon76-home | Kobaz: how would you do transfers? |
| 16:59.12 | Kobaz | Corydon76-home: haven't thought of that yet |
| 17:00.00 | Kobaz | but off the top of my head, the transferinitiator (as i call it), would pass a command to the command queue to say, hey start transfering |
| 17:00.08 | Corydon76-home | Kobaz: or worse, colliding transfers (e.g. two people attempting to pick up the same call at the same time). |
| 17:00.22 | Kobaz | first one wins |
| 17:00.48 | Kobaz | they would both put in the request to pick up, the command queue processes one, the other one is ignored since it's already processing one |
| 17:00.55 | Corydon76-home | Kobaz: actually, the second would win. |
| 17:01.08 | Kobaz | well you could set timeouts |
| 17:01.10 | Corydon76-home | if you had no deadlock potential |
| 17:01.56 | Corydon76-home | timeouts will solve all deadlock potentials... |
| 17:02.16 | Corydon76-home | and you can do them without any changes to the model |
| 17:02.25 | Kobaz | mm, true |
| 17:02.42 | Corydon76-home | pthread_timedwrlock() or whatever it's called |
| 17:46.37 | tzafrir | pabelanger, are there any changes in the gigthub starpy not included in the SourceForge SVN? |
| 17:49.24 | tzafrir | I guess it is |
| 17:49.35 | tzafrir | err, I guess there are |
| 17:53.20 | pabelanger | tzafrir: yes, we moved the project to github and now maintain it there |
| 17:54.39 | Kobaz | the thing with adding timeouts for locks, is yeah it would avoid deadlocks, but not by fixing the original problem causing the deadlock locking order |
| 17:54.55 | Qwell | pabelanger: we maintain it now? |
| 18:13.41 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 18:13.41 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 18:14.24 | *** join/#asterisk-dev malcolmd (~malcolmd@pdpc/sponsor/digium/malcolmd) |
| 18:14.24 | *** mode/#asterisk-dev [+o malcolmd] by ChanServ |
| 18:55.14 | Kobaz | so looking at the lock output |
| 18:55.31 | Kobaz | there's a deadlock between ast_do_masq and autoservice |
| 18:56.05 | Kobaz | i would figure the do_masq should always get right of way, so the autoservice should timeout or deadlockavoid magic |
| 18:58.25 | mjordan | Kobaz: what are you autoservicing, and why? |
| 18:59.17 | Kobaz | Dial(...,,U() |
| 18:59.26 | Kobaz | nothing fancy |
| 18:59.35 | Kobaz | gosub execute on pickup |
| 19:00.43 | Kobaz | i always get these |
| 19:00.44 | Kobaz | [2011-11-21 11:36:45] ERROR[23379] app_dial.c: <SIP/368-00000011> Could not stop autoservice on calling channel |
| 19:01.04 | mjordan | what version? |
| 19:01.16 | Kobaz | but it doesn't seem to be a huge problem since i've had systems running for months and months getting those errors constantly without an issue |
| 19:01.19 | Kobaz | 1.8.5 |
| 19:01.48 | Kobaz | but certain systems are hitting this deadlock |
| 19:02.06 | mjordan | Can you file a bug report? |
| 19:04.29 | Kobaz | yeap sure |
| 19:04.44 | mjordan | I'd also be curious to see what the gosub is doing, so the dialplan might be needed as well |
| 19:04.46 | Kobaz | i usually wait till i have a fix to file |
| 19:05.09 | mjordan | fair enough :-) But I would guess that being in autoservice and trying to masquerade the channel is a bad thing. |
| 19:05.13 | Kobaz | but this has been going on long enough that obviously i need someone to look at it |
| 19:05.16 | Kobaz | yeah |
| 19:05.29 | mjordan | we do put the channel in autoservice in dial when we do a gosub |
| 19:05.34 | Kobaz | it's probably a timing issue |
| 19:05.53 | Kobaz | like, i have enough code that runs on pickup that you could transfer while the autoservice is still running and handling the pickup gosub |
| 19:06.40 | Kobaz | it's a mix of pickup macros and Bridge() stealing the call |
| 19:07.14 | Kobaz | it's definitely reproducable, every single autoservice that's started for me, will complain it cant be stopped |
| 19:07.53 | Kobaz | === Thread ID: 0xb5bffb70 (do_monitor started at [25356] chan_sip.c restart_monitor()) |
| 19:07.57 | Kobaz | === ---> Lock #0 (chan_sip.c): MUTEX 24871 handle_request_do &netlock 0xb69dcdc0 (1) |
| 19:08.00 | Kobaz | that's the bad spot |
| 19:08.03 | mjordan | see what happens if you stop the autoservice in Dial right before you execute pbx_exec on the gosub |
| 19:08.19 | Kobaz | because restart monitor is waiting for a lock |
| 19:08.25 | Kobaz | that autoservice locked |
| 19:08.39 | *** join/#asterisk-dev moy (~moy@209.226.201.250) |
| 19:08.46 | Kobaz | autoservice has three locks at that point |
| 19:08.53 | Kobaz | channels, original, and clonechan |
| 19:09.12 | mjordan | autoservice is going to be processing frames on the channel, so I would expect it would be doing all sorts of stuff |
| 19:09.18 | Kobaz | yeah |
| 19:09.28 | Kobaz | it needs to release them all at some point |
| 19:09.32 | Kobaz | to let masq carry on |
| 19:09.43 | Kobaz | and then check if the channel is byebye and then bail if needed |
| 19:09.53 | mjordan | which it should if you stop it |
| 19:09.59 | mjordan | but we don't stop it until after the gosub is completed |
| 19:10.03 | Kobaz | yeah |
| 19:10.22 | Kobaz | it's most likly very much still running when a Bridge() comes along to steal it |
| 19:10.49 | mjordan | right. So what might be interesting to see is if Bridge can detect that the channel is in autoservice and request a stop before it goes and does its thing |
| 19:11.13 | Kobaz | i'll try and make a nice small dialplan to reproduce it |
| 19:21.12 | *** join/#asterisk-dev tamiel (~tamiel@ip-183.net-81-220-86.toulouse.rev.numericable.fr) |
| 19:24.21 | *** join/#asterisk-dev Deeewayne (~dwayne@c-76-29-230-45.hsd1.al.comcast.net) |
| 19:24.21 | *** mode/#asterisk-dev [+o Deeewayne] by ChanServ |
| 19:25.47 | Kobaz | ooh |
| 19:25.48 | Kobaz | yeah |
| 19:25.53 | Kobaz | here's what i'm doing |
| 19:25.58 | Kobaz | i'm doing a Bridge() in a pickup macro |
| 19:27.42 | *** join/#asterisk-dev hehol (~Adium@2a01:198:71d:0:21f:d0ff:fea1:568e) |
| 19:30.40 | mjordan | Kobaz: I would look at bridge_exec in features.c and see if you can detect if a channel is currently in autoservice. You'd want to do that before the call to do_bridge_masquerade. This may work around what you're looking at - it may have some other interesting side effects though too. |
| 19:32.14 | Kobaz | can't reproduce the autoservice stop problem with my mini dialplan |
| 19:32.15 | Kobaz | k |
| 19:46.06 | *** join/#asterisk-dev basti1101 (~basti1101@dslb-178-008-095-207.pools.arcor-ip.net) |
| 19:48.46 | *** join/#asterisk-dev basti1101 (~basti1101@dslb-178-008-095-207.pools.arcor-ip.net) |
| 19:52.13 | Kobaz | can more than one autoservice be running for a channel |
| 20:02.03 | Kobaz | i guess not |
| 20:04.04 | mjordan | nope, but autoservice is a single thread that can service many channels |
| 20:04.32 | mjordan | but there's only ever one 'autoservice' thread running in asterisk |
| 20:06.35 | Kobaz | ah okay |
| 20:06.47 | Kobaz | so it's the autoservicing for a particular channel |
| 20:07.33 | Kobaz | okay so |
| 20:07.42 | Kobaz | i added a flag to the channel to note if autoservice is running on it |
| 20:07.55 | Kobaz | and then in ast_bridge i stop it if it's running |
| 20:15.36 | *** join/#asterisk-dev mpe (~mpe@31.25.22.225) |
| 20:15.59 | *** join/#asterisk-dev serafie (~erin@nat/digium/x-rlzjktazykthjkym) |
| 20:17.13 | Kobaz | hmm, well that didn't end well, got a nice seg fault |
| 21:27.35 | *** join/#asterisk-dev philipp64|laptop (~chatzilla@mail.redfish-solutions.com) |
| 21:29.25 | pabelanger | Anybody else having problems commenting on reviewboard? |
| 21:29.30 | pabelanger | I assume it is my browser |
| 21:38.30 | Kobaz | hey hmm |
| 21:38.40 | Kobaz | i want to add something new to ast_channel |
| 21:39.26 | Kobaz | i'm adding just an int field, and i rebuild, and run, and all of a suddon i'm getting really bad memory accesses, reads of weird data |
| 21:39.40 | Kobaz | i've never had problems adding new stuff to ast_channel before |
| 21:40.13 | Kobaz | ==31599== Invalid read of size 4 |
| 21:40.13 | Kobaz | ==31599== at 0x80AA575: __ast_queue_frame (channel.c:1774) |
| 21:40.14 | Kobaz | etc |
| 21:42.48 | file | make clean |
| 21:45.38 | Kobaz | yeah just did |
| 21:45.42 | Kobaz | seems to work |
| 22:15.05 | Kobaz | Bridge() uses ast_channel_bridge() in channel.c |
| 22:32.08 | *** join/#asterisk-dev bklang (~bklang@184-192-86-89.pools.spcsdns.net) |
| 22:47.06 | Kobaz | mjordan: so autoservice isn't running when Bridge() runs |
| 22:54.48 | mjordan | Kobaz: Dial uses autoservice before jumping out to a GoSub or a macro |
| 22:57.49 | Kobaz | yeah |
| 22:57.53 | Kobaz | it turns it on and the back off |
| 22:58.01 | Kobaz | before the pickup macro runs |
| 23:01.53 | *** part/#asterisk-dev sruffell (~Adium@asterisk/the-kernel-guy/sruffell) |
| 23:02.01 | Kobaz | it's probably due to my convoluted setup |
| 23:02.17 | *** join/#asterisk-dev citywok (~citywok@67-134-194-33.dia.static.qwest.net) |
| 23:25.01 | Kobaz | okay so |
| 23:25.02 | Kobaz | what does this mean |
| 23:25.18 | Kobaz | http://pastebin.com/hgwqPiaC |
| 23:25.31 | Kobaz | ast_do_masquerade was called from autoservice? |
| 23:44.36 | mjordan | Kobaz: I'll take a peek tomorrow, but that is odd |
| 23:44.42 | mjordan | or at least, not what I would have expected |
| 23:45.27 | *** part/#asterisk-dev mjordan (~mjordan@nat/digium/x-yuyuynnmamheuccz) |
