Summary:	ASTERISK-00012: app_voicemail2 became a bit silent, lately
Reporter:	siggi (siggi)	Labels:	patch
Date Opened:	2003-07-26 10:07:57	Date Closed:	2004-09-25 02:22:25
Priority:	Major	Regression?	No
Status:	Closed/Complete	Components:	Core/General
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) breaking.diff
Description:	The prompts in both VoiceMail2 and VoiceMailMain2 have become silent. All I get is the initial "blip" followed by the Voice taking breath and being cut off before she has a chance to say "Comedian Mail". ****** ADDITIONAL INFORMATION ****** All other prompts (ie the Playback application) seem to work fine. I can still login to VoiceMailMain2, however, each prompt hangs immediately, until I enter some DTMF, so after entering user and password, the console says:    -- Playing 'vm-youhave' of which I hear just a slight "tick", and then it sits there forever... If I press "2" at this point, the console says    -- Playing 'vm-changeto' again, it only plays a "tick" and hangs till I press the next button. It doesn't matter how the call comes in (I've tried OH323, IAX and IAX2).
Comments:	By: siggi (siggi) 2003-07-27 06:47:39 to make this an "official" report, according to Marks's posting, these pieces need to be added: CVS date: CVS-05/07/03-16:01:05          (the issue does _not_ occur with CVS-05/07/03-16:01:05) uname -a says: asterisk 2.4.18-bf2.4 #1 Son Apr 14 09:53:28 CEST 2002 i686 unknown   (this is Debian's vanilla 2.4.18 kernel, the system is basically Debian stable) hardware: no telephony hardware (and no ztdummy, maybe that's a new requirement?) VoIP environment: - H.323 connection to CCM, via chan_oh323                  - IAX and IAX2 to another (similar) * gateway core file: none, it just hangs as described above config files: can't see which ones are relevant... By: siggi (siggi) 2003-07-27 11:46:13 Hmmm, as those timestamps don't really seem to change: current CVS (2003-07-27, 00:00:00 UTC) does expose this behaviour. This is CVS version 1.32 of app_voicemail2.c. app_voicemail2.c 1.29 (updated on June 27th) did work fine. further testing shows: downgrading app_voicemail2 alone does not help, this must be caused by some other change within asterisk... By: siggi (siggi) 2003-07-30 14:27:35 Ooops, looks like the Playback application is affected as well. (I must have mixed up * versions when trying the first time.) So this is _not specific to app_voicemail2. By: Malcolm Davenport (mdavenport) 2003-08-08 15:50:35 If you checkout the *new* CVS code of asterisk and compile it ..does it still have this problem ? By: siggi (siggi) 2003-08-10 06:17:27 Yes, it does. I've done a "cvs upgrade" and, just to be sure, another completely fresh checkout (cvs -d :pserver:anoncvs@cvs.digium.com/usr/cvsroot co asterisk) The symptoms are still there: All prompts (app_voicemail2 as well as the Playback application) just seem to lock up after sending the first packet out. (ie. I do get a bit of breath of vm-login and "beep" is almost played completely.) Normal calls are working just fine... I did a little bit of testing with older CVS versions, and playback broke with the CVS commit made on 2003-06-29 20:32:25 UTC ie. cvs up -D '2003-06-29 20:32:26 UTC' => playback breaks    cvs up -D '2003-06-29 20:32:25 UTC' => prompts play fine I'm going to upload a "cvs diff -u" between those two versions. My guess is that some of the timing changes in channel.c break this stuff, at least for non-zaptel setups like this one... By: x martinp (martinp) 2003-08-21 14:44:44 Can you get a fresh copy of zaptel/libpri/asterisk and comment out the ZAPTEL_OPTIMIZATIONS and tell us if that still is there ? Are you modifying any other settings in zaptel/libpri/asterisk ? By: siggi (siggi) 2003-08-21 17:26:07 Not quite. I don't have (or need) libpri, and I don't have any zaptel devices, either. However, I did have a stale zaptel installation, which caused ZAPTEL_OPTIMIZATIONS  to be defined. Removing /usr/include/linux/zaptel.h did fix the issue. I guess you can change this bug's severity to "minor" until there's a way to disable zaptel optimizations at runtime, if there's no zaptel device...