iaxclient-devel — Development Discussion List

[Re] Re: [Re] Re: [Iaxclient-devel] Setting output level on windows

[Cyril V. <cyr...@me...>]

De : mailto:st...@st...
Emission : 15/11/2004 21:10:15
> Cyril VELTER wrote:
> 
> >	I have a patch for alaw support (I just reworted it from scratch, inspired 
by 
> >the ulaw codec and some sun free code). Is it possible to include this in 
CVS ? 
> >and where should I post the patch ?
> >  
> >
> 
> I got your patch off-list, and it's now included in CVS.  I just had to 
> un-MSDOS-ify the files and merge new diffs.
> 
> It compiles clean, but I haven't actually tried it, but I'm sure people 
> on the list will let us know how it works out..


	Very nice,

	I just updated my local copy with cvs recompiled everthing and tested it with 
simpleclient/testcall (modified to force alaw usage) and it works.

> 
> If someone wants to, you can probably make decoding much faster by 
> compiling a table at initialization time, and then just doing a table 
> lookup at run time, like we do for ulaw.  There's really no downside to 
> this, because there are only 256 2-byte entries in the table, so it only 
> takes up 512 bytes.  OTOH, for a client, this all doesn't make much 
> difference, because we're unlikely to tax anything with encoding or 
> decoding.
> 
> Encoding is a bit different, because even with the method we use in 
> ulaw, there might still be some quantization.  (and a 64K table does 
> take more memory, doesn't always fit in cache well, etc).

	You are right. I'll test to see if there is a real difference.


	cyril velter

Re: [Re] Re: [Iaxclient-devel] Setting output level on windows

[Steve K. <st...@st...>]

Cyril VELTER wrote:

>	I have a patch for alaw support (I just reworted it from scratch, inspired by 
>the ulaw codec and some sun free code). Is it possible to include this in CVS ? 
>and where should I post the patch ?
>  
>

I got your patch off-list, and it's now included in CVS.  I just had to 
un-MSDOS-ify the files and merge new diffs.

It compiles clean, but I haven't actually tried it, but I'm sure people 
on the list will let us know how it works out..

If someone wants to, you can probably make decoding much faster by 
compiling a table at initialization time, and then just doing a table 
lookup at run time, like we do for ulaw.  There's really no downside to 
this, because there are only 256 2-byte entries in the table, so it only 
takes up 512 bytes.  OTOH, for a client, this all doesn't make much 
difference, because we're unlikely to tax anything with encoding or 
decoding.

Encoding is a bit different, because even with the method we use in 
ulaw, there might still be some quantization.  (and a 64K table does 
take more memory, doesn't always fit in cache well, etc).



-SteveK

Re: [Re] Re: [Iaxclient-devel] Setting output level on windows

[Steve K. <st...@st...>]

Cyril VELTER wrote:

>	I agree it depends on the application. iaxclient is a library and should not 
>IMHO modify the process priorityclass because it has no knowledge of what the 
>application do. Modifying the thread priority is OK because the library know 
>what the the thread is doing and that it needs a high priority to service 
>audio.
>  
>
Fair enough.  I'll change this..  It's easy enough for people to set the 
priority of the rest of their application if they need to.

>
>	I have a patch for alaw support (I just reworted it from scratch, inspired by 
>the ulaw codec and some sun free code). Is it possible to include this in CVS ? 
>and where should I post the patch ?
>  
>

Thanks! You can send it to me, or to the list..

-SteveK

[Re] Re: [Iaxclient-devel] Setting output level on windows

[Cyril V. <cyr...@me...>]

De : mailto:je...@ru...
Emission : 15/11/2004 18:24:27

> 
> On Nov 15, 2004, at 11:46 AM, Steve Kann wrote:
> 
> > Cyril VELTER wrote:
> >
> >> 	While evaluating iaxclient, I noticed a small bug in px_win_wmme.c 
> >> line 735 (_Px_InitOutputVolumeControls).
> >> mixerOpen( 		( LPHMIXER )( &mixer->hInputMixer ), 		hWaveOut, 0, 0,
> >> 		MIXER_OBJECTF_HWAVEOUT 	) ;
> >>
> >>
> >> should be
> >> mixerOpen( 		( LPHMIXER )( &mixer->hOutputMixer ), 		hWaveOut, 0, 0,
> >> 		MIXER_OBJECTF_HWAVEOUT 	) ;
> >>
> >>
> >> 	the cirrent code does not set the output volume of the right 
> >> interface (in the case the current device is not the first one)
> >>
> > I didn't write that code, but AFAIK it seemed to be working correctly 
> > for us.  I'll ask the original author, who knows this code better than 
> > I do.
> 
> hrm.
> 
> yeah, seems it _should_ be "&mixer->hOutputMixer", but it doesn't seem
> to have an effect one way or another.
> 
> i'll test more with the corrected call, and then commit it - for 
> consistency's sake -
> if all goes well.

	I have a configuration where it does make a difference :


	WinXP pro, SoundMAX integrated audio + Plantronics DA50 USB adapter. With this 
configuration, where I use the USB audio, setting output volume will set the 
output volume of the soundmax device.


	cyril velter.

[Re] Re: [Iaxclient-devel] Setting output level on windows

[Cyril V. <cyr...@me...>]

> Cyril VELTER wrote:
> 
> >Regarding thread priority, it is better to set the thread priority directly. 
On 
> >windows, scheduling is done thread by thread and SetPriorityClass influence 
> >every thread in a process which is not good. It would be better to replace
> >
> >	SetPriorityClass
> >
> >by 
> >
> >	SetThreadPriority(GetCurrentThread(),THREAD_PRIORITY_TIME_CRITICAL)
> >
> >this will only change the priority of the thread servicing audio and not any 

> >other threads (user interface ...)
> >  
> >
> It depends on what you're actually trying to do, or if your UI threads 
> can end up blocking your audio/network threads or now.   I think that in 
> practice, it shouldn't make too much difference, because the UI thread 
> for a softphone will generally be simple enough that it won't use much 
> CPU anyway.  However, if your UI thread can block processing in some way 
> (i.e. it might hold a mutex that the callbacks need to get), then having 
> it also run with elevated priority will prevent the audio/network thread 
> from blocking on that mutex.  (I suppose, that a smart scheduler _might_ 
> be smart enough to see that a high-priority thread is blocking on a 
> mutex held by a lower-priority thread, and it could increase the 
> lower-priority thread's priority until it drops the mutex or something).
> 
> In my application, I actually increase the priority of the whole 
> application when I launch it from another application, so this priority 
> stuff is redundant, and I just added it for the community.  I'm more 
> than open to making it thread-specific if others agree..

	I agree it depends on the application. iaxclient is a library and should not 
IMHO modify the process priorityclass because it has no knowledge of what the 
application do. Modifying the thread priority is OK because the library know 
what the the thread is doing and that it needs a high priority to service 
audio.

> 
> >Is there any plan ta support Alaw (I'm in france, and it's the native pstn 
> >codec here) ? I've hacked some support base on asterisk code, but there 
might 
> >be licencing issues.
> >  
> >
> 
> I don't plan on writing it, but it's pretty simple, so someone could if 
> they had a few hours and some motivation.  If it's based on asterisk 
> code, that's not OK licensing wise, unless you get a disclaimer from 
> digium to relicense it LGPL.  


	I have a patch for alaw support (I just reworted it from scratch, inspired by 
the ulaw codec and some sun free code). Is it possible to include this in CVS ? 
and where should I post the patch ?

	cyril velter

Re: [Iaxclient-devel] Setting output level on windows

[jeff o. <je...@ru...>]

On Nov 15, 2004, at 11:46 AM, Steve Kann wrote:

> Cyril VELTER wrote:
>
>> 	While evaluating iaxclient, I noticed a small bug in px_win_wmme.c 
>> line 735 (_Px_InitOutputVolumeControls).
>> mixerOpen( 		( LPHMIXER )( &mixer->hInputMixer ), 		hWaveOut, 0, 0,
>> 		MIXER_OBJECTF_HWAVEOUT 	) ;
>>
>>
>> should be
>> mixerOpen( 		( LPHMIXER )( &mixer->hOutputMixer ), 		hWaveOut, 0, 0,
>> 		MIXER_OBJECTF_HWAVEOUT 	) ;
>>
>>
>> 	the cirrent code does not set the output volume of the right 
>> interface (in the case the current device is not the first one)
>>
> I didn't write that code, but AFAIK it seemed to be working correctly 
> for us.  I'll ask the original author, who knows this code better than 
> I do.

hrm.

yeah, seems it _should_ be "&mixer->hOutputMixer", but it doesn't seem
to have an effect one way or another.

i'll test more with the corrected call, and then commit it - for 
consistency's sake -
if all goes well.

j-

Re: [Iaxclient-devel] Setting output level on windows

[Steve K. <st...@st...>]

Cyril VELTER wrote:

>	While evaluating iaxclient, I noticed a small bug in px_win_wmme.c line 735 
>(_Px_InitOutputVolumeControls). 
>
>mixerOpen( 
>		( LPHMIXER )( &mixer->hInputMixer ), 
>		hWaveOut, 0, 0,
>		MIXER_OBJECTF_HWAVEOUT 
>	) ;
>
>
>should be 
>
>mixerOpen( 
>		( LPHMIXER )( &mixer->hOutputMixer ), 
>		hWaveOut, 0, 0,
>		MIXER_OBJECTF_HWAVEOUT 
>	) ;
>
>
>	the cirrent code does not set the output volume of the right interface (in the 
>case the current device is not the first one)
>  
>
I didn't write that code, but AFAIK it seemed to be working correctly 
for us.  I'll ask the original author, who knows this code better than I do.

>Regarding thread priority, it is better to set the thread priority directly. On 
>windows, scheduling is done thread by thread and SetPriorityClass influence 
>every thread in a process which is not good. It would be better to replace
>
>	SetPriorityClass
>
>by 
>
>	SetThreadPriority(GetCurrentThread(),THREAD_PRIORITY_TIME_CRITICAL)
>
>this will only change the priority of the thread servicing audio and not any 
>other threads (user interface ...)
>  
>
It depends on what you're actually trying to do, or if your UI threads 
can end up blocking your audio/network threads or now.   I think that in 
practice, it shouldn't make too much difference, because the UI thread 
for a softphone will generally be simple enough that it won't use much 
CPU anyway.  However, if your UI thread can block processing in some way 
(i.e. it might hold a mutex that the callbacks need to get), then having 
it also run with elevated priority will prevent the audio/network thread 
from blocking on that mutex.  (I suppose, that a smart scheduler _might_ 
be smart enough to see that a high-priority thread is blocking on a 
mutex held by a lower-priority thread, and it could increase the 
lower-priority thread's priority until it drops the mutex or something).

In my application, I actually increase the priority of the whole 
application when I launch it from another application, so this priority 
stuff is redundant, and I just added it for the community.  I'm more 
than open to making it thread-specific if others agree..

>Is there any plan ta support Alaw (I'm in france, and it's the native pstn 
>codec here) ? I've hacked some support base on asterisk code, but there might 
>be licencing issues.
>  
>

I don't plan on writing it, but it's pretty simple, so someone could if 
they had a few hours and some motivation.  If it's based on asterisk 
code, that's not OK licensing wise, unless you get a disclaimer from 
digium to relicense it LGPL.  

-SteveK

[Iaxclient-devel] Setting output level on windows

[Cyril V. <cyr...@me...>]

	While evaluating iaxclient, I noticed a small bug in px_win_wmme.c line 735 
(_Px_InitOutputVolumeControls). 

mixerOpen( 
		( LPHMIXER )( &mixer->hInputMixer ), 
		hWaveOut, 0, 0,
		MIXER_OBJECTF_HWAVEOUT 
	) ;


should be 

mixerOpen( 
		( LPHMIXER )( &mixer->hOutputMixer ), 
		hWaveOut, 0, 0,
		MIXER_OBJECTF_HWAVEOUT 
	) ;


	the cirrent code does not set the output volume of the right interface (in the 
case the current device is not the first one)


Regarding thread priority, it is better to set the thread priority directly. On 
windows, scheduling is done thread by thread and SetPriorityClass influence 
every thread in a process which is not good. It would be better to replace

	SetPriorityClass

by 

	SetThreadPriority(GetCurrentThread(),THREAD_PRIORITY_TIME_CRITICAL)

this will only change the priority of the thread servicing audio and not any 
other threads (user interface ...)


Is there any plan ta support Alaw (I'm in france, and it's the native pstn 
codec here) ? I've hacked some support base on asterisk code, but there might 
be licencing issues.


	cyril velter

Re: [Iaxclient-devel] app_conference compile error

[xavier d. <xa...@sy...>]

Thanks for your help.
 > Maybe try removing -std=c99?

No difference. I'm stupid, I forgot to mention than I don't have only 
one error but a lots. I only shown the first ones.

Any idea ?

Ast the stddef:212-213
#if !(defined (__GNUG__) && defined (size_t))
typedef __SIZE_TYPE__ size_t;

/usr/include/time.h:76: error: syntax error before "typedef"
__BEGIN_NAMESPACE_STD
/* Returned by `time'.  */

/usr/include/time.h:92: error: syntax error before "typedef"
# include <bits/types.h>

/

make
gcc -pipe -Wall -Wmissing-prototypes -Wmissing-declarations -g  
-I/usr/src/asterisk/include/asterisk/  -D_REENTRANT -D_GNU_SOURCE -O3 
-ffast-math -funroll-all-loops -fprefetch-loop-arrays 
-fsingle-precision-constant  -DCRYPTO -DAPP_CONFERENCE_DEBUG -Ilibspeex 
-DSILDET=2   -c -o app_conference.o app_conference.c
cc1: warning: -fprefetch-loop-arrays not supported for this target (try 
-march switches)
In file included from /usr/include/bits/types.h:31,
                 from /usr/include/sys/types.h:31,
                 from /usr/include/asterisk/frame.h:23,
                 from /usr/include/asterisk/channel.h:17,
                 from /usr/include/asterisk/pbx.h:17,
                 from app_conference.h:23,
                 from app_conference.c:19:
/usr/lib/gcc-lib/i486-linux/3.3.4/include/stddef.h:213: error: syntax 
error before "typedef"
In file included from /usr/include/sys/types.h:133,
                 from /usr/include/asterisk/frame.h:23,
                 from /usr/include/asterisk/channel.h:17,
                 from /usr/include/asterisk/pbx.h:17,
                 from app_conference.h:23,
                 from app_conference.c:19:
/usr/include/time.h:76: error: syntax error before "typedef"
/usr/include/time.h:92: error: syntax error before "typedef"
In file included from /usr/include/asterisk/frame.h:23,
                 from /usr/include/asterisk/channel.h:17,
                 from /usr/include/asterisk/pbx.h:17,
                 from app_conference.h:23,
                 from app_conference.c:19:
/usr/include/sys/types.h:158:20: missing binary operator before token "("
[...]

Re: [Iaxclient-devel] app_conference compile error

[Steve K. <st...@st...>]

On Nov 14, 2004, at 3:30 PM, xavier dutoit wrote:
> # make
> gcc -pipe -std=c99 -Wall -Wmissing-prototypes -Wmissing-declarations 
> -g  -I/usr/src/asterisk/include/asterisk/  -D_REENTRANT -D_GNU_SOURCE 
> -O3 -ffast-math -funroll-all-loops -fprefetch-loop-arrays 
> -fsingle-precision-constant  -DCRYPTO -DAPP_CONFERENCE_DEBUG 
> -Ilibspeex -DSILDET=2   -c -o app_conference.o app_conference.c
> cc1: warning: -fprefetch-loop-arrays not supported for this target 
> (try -march switches)
> In file included from /usr/include/bits/types.h:31,
>                 from /usr/include/sys/types.h:31,
>                 from /usr/include/asterisk/frame.h:23,
>                 from /usr/include/asterisk/channel.h:17,
>                 from /usr/include/asterisk/pbx.h:17,
>                 from app_conference.h:23,
>                 from app_conference.c:19:
> /usr/lib/gcc-lib/i486-linux/3.3.4/include/stddef.h:213: error: syntax 
> error before "typedef"
> In file included from /usr/include/sys/types.h:133,
>                 from /usr/include/asterisk/frame.h:23,
>                 from /usr/include/asterisk/channel.h:17,
>                 from /usr/include/asterisk/pbx.h:17,
>                 from app_conference.h:23,
>                 from app_conference.c:19:
> /usr/include/time.h:76: error: syntax error before "typedef"

I have a debian unstable box, and a sarge box (but it might be a few 
weeks behind), and I don't have those files..

What's on the lines in question: 
/usr/lib/gcc-lib/i486-linux/3.3.4/include/stddef.h:213
and usr/include/time.h:76


Maybe try removing -std=c99?



>
>
> Any idea ?
>
> Second question: On the makefile, what does mean :
> # 0 = OFF 1 = astdsp 2 = speex
> SILDET := 2
>

It specifies which kind of VAD (silence detection) you might want to 
use.  It is generally only effective to use this if you want to get 
indications of active speakers from the manager interface.

-SteveK

[Iaxclient-devel] app_conference compile error

[xavier d. <xa...@sy...>]

Hello,

I have some problems compiling app_conference (both asterisk and it 
fresh updated from the CVS) on my debian sarge.
I've modified the Makefile to change the ASTERISK_INCLUDE_DIR and to 
comment out "CFLAGS += -mcpu=7450 -faltivec -mabi=altivec -mdynamic-no-pic".

When I make, both from the app_conference folder or from asterisk (I've 
modified apps/Makefile to include the app_conference makefile and added 
app_conference.so), I got the same result:


# make
gcc -pipe -std=c99 -Wall -Wmissing-prototypes -Wmissing-declarations -g  
-I/usr/src/asterisk/include/asterisk/  -D_REENTRANT -D_GNU_SOURCE -O3 
-ffast-math -funroll-all-loops -fprefetch-loop-arrays 
-fsingle-precision-constant  -DCRYPTO -DAPP_CONFERENCE_DEBUG -Ilibspeex 
-DSILDET=2   -c -o app_conference.o app_conference.c
cc1: warning: -fprefetch-loop-arrays not supported for this target (try 
-march switches)
In file included from /usr/include/bits/types.h:31,
                 from /usr/include/sys/types.h:31,
                 from /usr/include/asterisk/frame.h:23,
                 from /usr/include/asterisk/channel.h:17,
                 from /usr/include/asterisk/pbx.h:17,
                 from app_conference.h:23,
                 from app_conference.c:19:
/usr/lib/gcc-lib/i486-linux/3.3.4/include/stddef.h:213: error: syntax 
error before "typedef"
In file included from /usr/include/sys/types.h:133,
                 from /usr/include/asterisk/frame.h:23,
                 from /usr/include/asterisk/channel.h:17,
                 from /usr/include/asterisk/pbx.h:17,
                 from app_conference.h:23,
                 from app_conference.c:19:
/usr/include/time.h:76: error: syntax error before "typedef"


Any idea ?

Second question: On the makefile, what does mean :
# 0 = OFF 1 = astdsp 2 = speex
SILDET := 2
?

FYI, I'm not very happy with the meetme app, where I got a very crappy 
sound with two iaxcomm and a zap FXO channel (however, it seems to work 
better with a SIP phone...).

Xavier

Re: [Iaxclient-devel] Implementing IAX_COMMAND_REGREL?

[Steven S. <ss...@so...>]

Steve Kann wrote:

>
>
> Steven Sokol wrote:
>
>> Ok, I've been playing in the library again trying to get REGREL to 
>> work and I've found that I can send it all I want but that the curren 
>> chan_iax2 does not seem to do anything with it.  Of course I am 
>> sending it with only a bare minimum of additional information -- I 
>> toss in most of the same IEs that are sent along with a  REGREQ.
>>
>> I see the request hit Asterisk in the iax debug output, but it never 
>> seems to hit the proper point in the chan_iax2 code to unregister the 
>> host.  I've added verbose statemements to chan_iax2 at line 5791 (the 
>> case statement handling IAX commands) but it never gets called.
>>
>> Does anybody have any idea what is going on with this?  Here's my 
>> iaxc_unregister and the equivalent code in libiax2.  Keep in mind 
>> that I have hacked the iaxc_register into iaxc_register2 and am 
>> keeping a list of valid registrations...
>
>
>
>
> I'm not sure why it's not working, but I guess my question is -- is it 
> important to notify the server that you want to unregister, or would 
> it be sufficient to just let the registration expire?  If you have 
> registrations last for say, 2 minutes, and then refresh every 30 
> seconds, if you tell iaxclient to stop re-registering, the old 
> registration should only last between 90 and 120 seconds..  (change 
> numbers as appropriate).
> Is there a real need for _immediate_ unregistration?
>
> -SteveK
>
I have one situation where it is, at least sort-of critical.  I have a 
client who uses IAX Phone for home office/road warrior stuff, and a SIP 
hard-phone for the desk at work.  He wants to have calls go to the IAX 
Phone if it is on-line, and to the SIP device if not.  The code in his 
dial plan checks for an available IAX channel with ChanIsAvail and, if 
the device is registered, calls the IAX phone.  If not, it checks the 
SIP phone and trys to call it.  Unfortunately, there is a 90 - 120 
second gap between shutting down the IAX phone and the system knowing 
that it is gone.

While I know there are a number of ways of overcoming this, but it seems 
most elegant to be able to "sign off" at the end of a session.

- SteveS

>
>
>>
>> [in iaxclient_lib.c]:
>>
>> int iaxc_unregister2(int regid)
>> {
>>    struct iaxc_registration *prereg = NULL;
>>    struct iaxc_registration *oldreg = NULL;
>>    struct iaxc_registration *curreg = NULL;
>>          // Bail if the registration value is zero
>>    if (!registrations)
>>        return -1;
>>      // Set pointers to the base registrations value
>>    curreg = registrations;
>>    prereg = registrations;
>>      // Look for the registration by regid
>>    while (1) {
>>              if (curreg->regid == regid) {
>>            oldreg = curreg;
>>            break;
>>        }
>>              // cache the current registration for future use
>>        prereg = curreg;
>>              // get the next registration structure
>>        if (curreg->next)
>>            curreg = curreg->next;
>>        else
>>            return -2;  // not found
>>    }
>>      // Hmmm... not found.
>>    if(!oldreg) {
>>        iaxc_usermsg(IAXC_ERROR, "Can't find registration to cancel.");
>>        return -3;
>>    }
>>      // Create a new session for this (should be every time)
>>    if(!oldreg->session) {
>>        oldreg->session = iax_session_new();      }
>>      // send out the unregister request
>>    iax_unregister(oldreg->session, oldreg->host, oldreg->user, 
>> oldreg->pass);
>>      // remove it from the list;
>>    if (oldreg->next) {
>>        if (prereg)
>>            prereg->next == oldreg->next;
>>    }
>>      // Destroy it
>>    free(oldreg);
>>      // return zero for success
>>    return 0;
>> }
>>
>> [in iax.c (in libiax2)]:
>>
>> int iax_unregister(struct iax_session *session, char *server, char 
>> *peer, char *secret)
>> {
>>      /* Send an un-registration request */
>>    char tmp[256];
>>    char *p;
>>    int res;
>>    int portno = IAX_DEFAULT_PORTNO;
>>    struct iax_ie_data ied;
>>    struct hostent *hp;
>>      tmp[255] = '\0';
>>    strncpy(tmp, server, sizeof(tmp) - 1);
>>    p = strchr(tmp, ':');
>>    if (p)
>>        portno = atoi(p);
>>      memset(&ied, 0, sizeof(ied));
>>      if (secret)
>>        strncpy(session->secret, secret, sizeof(session->secret) - 1);
>>    else
>>        strcpy(session->secret, "");
>>      /* Connect first */
>>    hp = gethostbyname(tmp);
>>    if (!hp) {
>>        snprintf(iax_errstr, sizeof(iax_errstr), "Invalid hostname: 
>> %s", tmp);
>>        return -1;
>>    }
>>      memcpy(&session->peeraddr.sin_addr, hp->h_addr, 
>> sizeof(session->peeraddr.sin_addr));
>>    session->peeraddr.sin_port = htons(portno);
>>    session->peeraddr.sin_family = AF_INET;
>>    strncpy(session->username, peer, sizeof(session->username) - 1);
>>    iax_ie_append_str(&ied, IAX_IE_USERNAME, peer);
>>    res = send_command(session, AST_FRAME_IAX, IAX_COMMAND_REGREL, 0, 
>> ied.buf, ied.pos, -1);
>>      DEBU(G "Sent Reg Release Message\n");
>>      return res;
>> }
>
>
>

Re: [Iaxclient-devel] Implementing IAX_COMMAND_REGREL?

[Steve K. <st...@st...>]

Steven Sokol wrote:

> Ok, I've been playing in the library again trying to get REGREL to 
> work and I've found that I can send it all I want but that the curren 
> chan_iax2 does not seem to do anything with it.  Of course I am 
> sending it with only a bare minimum of additional information -- I 
> toss in most of the same IEs that are sent along with a  REGREQ.
>
> I see the request hit Asterisk in the iax debug output, but it never 
> seems to hit the proper point in the chan_iax2 code to unregister the 
> host.  I've added verbose statemements to chan_iax2 at line 5791 (the 
> case statement handling IAX commands) but it never gets called.
>
> Does anybody have any idea what is going on with this?  Here's my 
> iaxc_unregister and the equivalent code in libiax2.  Keep in mind that 
> I have hacked the iaxc_register into iaxc_register2 and am keeping a 
> list of valid registrations...



I'm not sure why it's not working, but I guess my question is -- is it 
important to notify the server that you want to unregister, or would it 
be sufficient to just let the registration expire?  If you have 
registrations last for say, 2 minutes, and then refresh every 30 
seconds, if you tell iaxclient to stop re-registering, the old 
registration should only last between 90 and 120 seconds..  (change 
numbers as appropriate). 

Is there a real need for _immediate_ unregistration?

-SteveK



>
> [in iaxclient_lib.c]:
>
> int iaxc_unregister2(int regid)
> {
>    struct iaxc_registration *prereg = NULL;
>    struct iaxc_registration *oldreg = NULL;
>    struct iaxc_registration *curreg = NULL;
>          // Bail if the registration value is zero
>    if (!registrations)
>        return -1;
>      // Set pointers to the base registrations value
>    curreg = registrations;
>    prereg = registrations;
>      // Look for the registration by regid
>    while (1) {
>              if (curreg->regid == regid) {
>            oldreg = curreg;
>            break;
>        }
>              // cache the current registration for future use
>        prereg = curreg;
>              // get the next registration structure
>        if (curreg->next)
>            curreg = curreg->next;
>        else
>            return -2;  // not found
>    }
>      // Hmmm... not found.
>    if(!oldreg) {
>        iaxc_usermsg(IAXC_ERROR, "Can't find registration to cancel.");
>        return -3;
>    }
>      // Create a new session for this (should be every time)
>    if(!oldreg->session) {
>        oldreg->session = iax_session_new();      }
>      // send out the unregister request
>    iax_unregister(oldreg->session, oldreg->host, oldreg->user, 
> oldreg->pass);
>      // remove it from the list;
>    if (oldreg->next) {
>        if (prereg)
>            prereg->next == oldreg->next;
>    }
>      // Destroy it
>    free(oldreg);
>      // return zero for success
>    return 0;
> }
>
> [in iax.c (in libiax2)]:
>
> int iax_unregister(struct iax_session *session, char *server, char 
> *peer, char *secret)
> {
>      /* Send an un-registration request */
>    char tmp[256];
>    char *p;
>    int res;
>    int portno = IAX_DEFAULT_PORTNO;
>    struct iax_ie_data ied;
>    struct hostent *hp;
>      tmp[255] = '\0';
>    strncpy(tmp, server, sizeof(tmp) - 1);
>    p = strchr(tmp, ':');
>    if (p)
>        portno = atoi(p);
>      memset(&ied, 0, sizeof(ied));
>      if (secret)
>        strncpy(session->secret, secret, sizeof(session->secret) - 1);
>    else
>        strcpy(session->secret, "");
>      /* Connect first */
>    hp = gethostbyname(tmp);
>    if (!hp) {
>        snprintf(iax_errstr, sizeof(iax_errstr), "Invalid hostname: 
> %s", tmp);
>        return -1;
>    }
>      memcpy(&session->peeraddr.sin_addr, hp->h_addr, 
> sizeof(session->peeraddr.sin_addr));
>    session->peeraddr.sin_port = htons(portno);
>    session->peeraddr.sin_family = AF_INET;
>    strncpy(session->username, peer, sizeof(session->username) - 1);
>    iax_ie_append_str(&ied, IAX_IE_USERNAME, peer);
>    res = send_command(session, AST_FRAME_IAX, IAX_COMMAND_REGREL, 0, 
> ied.buf, ied.pos, -1);
>      DEBU(G "Sent Reg Release Message\n");
>      return res;
> }

[Iaxclient-devel] Implementing IAX_COMMAND_REGREL?

[Steven S. <ss...@so...>]

Ok, I've been playing in the library again trying to get REGREL to work 
and I've found that I can send it all I want but that the curren 
chan_iax2 does not seem to do anything with it.  Of course I am sending 
it with only a bare minimum of additional information -- I toss in most 
of the same IEs that are sent along with a  REGREQ.

I see the request hit Asterisk in the iax debug output, but it never 
seems to hit the proper point in the chan_iax2 code to unregister the 
host.  I've added verbose statemements to chan_iax2 at line 5791 (the 
case statement handling IAX commands) but it never gets called.

Does anybody have any idea what is going on with this?  Here's my 
iaxc_unregister and the equivalent code in libiax2.  Keep in mind that I 
have hacked the iaxc_register into iaxc_register2 and am keeping a list 
of valid registrations...

[in iaxclient_lib.c]:

int iaxc_unregister2(int regid)
{
    struct iaxc_registration *prereg = NULL;
    struct iaxc_registration *oldreg = NULL;
    struct iaxc_registration *curreg = NULL;
       
    // Bail if the registration value is zero
    if (!registrations)
        return -1;
   
    // Set pointers to the base registrations value
    curreg = registrations;
    prereg = registrations;
   
    // Look for the registration by regid
    while (1) {
       
        if (curreg->regid == regid) {
            oldreg = curreg;
            break;
        }
       
        // cache the current registration for future use
        prereg = curreg;
       
        // get the next registration structure
        if (curreg->next)
            curreg = curreg->next;
        else
            return -2;  // not found
    }
   
    // Hmmm... not found.
    if(!oldreg) {
        iaxc_usermsg(IAXC_ERROR, "Can't find registration to cancel.");
        return -3;
    }
   
    // Create a new session for this (should be every time)
    if(!oldreg->session) {
        oldreg->session = iax_session_new();   
    }
   
    // send out the unregister request
    iax_unregister(oldreg->session, oldreg->host, oldreg->user, 
oldreg->pass);
   
    // remove it from the list;
    if (oldreg->next) {
        if (prereg)
            prereg->next == oldreg->next;
    }
   
    // Destroy it
    free(oldreg);
   
    // return zero for success
    return 0;
}

[in iax.c (in libiax2)]:

int iax_unregister(struct iax_session *session, char *server, char 
*peer, char *secret)
{
   
    /* Send an un-registration request */
    char tmp[256];
    char *p;
    int res;
    int portno = IAX_DEFAULT_PORTNO;
    struct iax_ie_data ied;
    struct hostent *hp;
   
    tmp[255] = '\0';
    strncpy(tmp, server, sizeof(tmp) - 1);
    p = strchr(tmp, ':');
    if (p)
        portno = atoi(p);
   
    memset(&ied, 0, sizeof(ied));
   
    if (secret)
        strncpy(session->secret, secret, sizeof(session->secret) - 1);
    else
        strcpy(session->secret, "");
   
    /* Connect first */
    hp = gethostbyname(tmp);
    if (!hp) {
        snprintf(iax_errstr, sizeof(iax_errstr), "Invalid hostname: %s", 
tmp);
        return -1;
    }
   
    memcpy(&session->peeraddr.sin_addr, hp->h_addr, 
sizeof(session->peeraddr.sin_addr));
    session->peeraddr.sin_port = htons(portno);
    session->peeraddr.sin_family = AF_INET;
    strncpy(session->username, peer, sizeof(session->username) - 1);
    iax_ie_append_str(&ied, IAX_IE_USERNAME, peer);
    res = send_command(session, AST_FRAME_IAX, IAX_COMMAND_REGREL, 0, 
ied.buf, ied.pos, -1);
   
    DEBU(G "Sent Reg Release Message\n");
   
    return res;
}

Re: [Iaxclient-devel] Crash w/ IAXComm talking to another IAXComm

[Nathan S. <iax...@np...>]

Michael Van Donselaar wrote:

> 
>
>  
>
>>-----Original Message-----
>>From: iax...@li... 
>>[mailto:iax...@li...] On 
>>Behalf Of Nathan Sharp
>>Sent: Tuesday, November 09, 2004 11:06 AM
>>To: mi...@va...
>>Cc: iax...@li...
>>Subject: Re: [Iaxclient-devel] Crash w/ IAXComm talking to 
>>another IAXComm
>>
>>Michael,
>>  Thanks so much for the quick reply!  To answer your 
>>questions, I am calling through an Asterisk 1.0.1 server 
>>(with Speex disabled).  I am not using the echo cancellation 
>>feature.  
>>
>>  Could this be the same problem?
>>http://sourceforge.net/mailarchive/forum.php?thread_id=5836576
>>&forum_id=29380
>>
>>    
>>
>
>I'm really not totally sure, but I think there's still a problem with codec
>negotiation.
>
>Could you try calling peer to peer and see what you get?
>  
>
Sure enough, I did just find something that works based on that comment:
disallow=all
allow=gsm

in the iax.conf setting for this client.  I can now dial to FWD fine (I 
can't try the RH8 box until I get home).  As an aside, I tried 
allow=ulaw first (so my asterisk box wouldn't have to convert when 
connecting to FWD), and it crashed as well.  Thanks for the hint!  That 
will get me running for now (although I await a better version soon!)

>I am also going to post new binaries tonite or tomorrow compiled on RH9 and
>WinXP.  They will have CVS updates for iaxclient, and be set to prefer iLBC.
>
>  
>
>>  Only the existence of NAT or not doesn't seem to matter for 
>>me.  I have hooked up my asterisk server to a FWD account 
>>using iax2.fwdnet.net and I've found that the client crashes 
>>also if I try to connect to FWD through my asterisk server.   
>>
>>Feature Requests?  My top priorities, in order:
>>1) Secure against malicious incoming packets
>>    
>>
>
>I don't know what to do here
>  
>
Code review and time will find these.  It is a small target for malware 
writers right now too.

>  
>
>>2) Low Latency (10-2004 version seems much better than the 
>>earlier version, when it doesn't crash). 
>>    
>>
>
>See Steve Kann's comments about this.  I'll recompile with a lower latency,
>and see how it works.
>  
>
Yea, I found some stuff on the environment variable.  The Linux version 
seems much better than the Windows version.  I mention it because it 
makes the most difference to the "naturalness" of phone conversations.  
It is one of the main reasons I don't use a cell phone.  I appreciate 
that the software has gotten better and that you guys are still working 
on it :-)

>  
>
>>3) Reliable.  The earlier version has never crashed yet on 
>>me, it just has much higher latency.
>>    
>>
>
>That's a feature everyone can use!
>
>  
>
>>4) Easy to install/use (so I can hand it out to family and 
>>friends).  You are pretty close on this front already.  There 
>>are a couple of minor GUI issues which could be cleaned up.
>>    
>>
>
>You shoulda seen the earlier versions!  I think I've got the install down to
>the easiest I can without going to the point of using an installer:  a
>single file that you can run from anywhere.
>  
>
Good Work!

>Quick hint that I should put in the QUICKSTART:  I trim down the registry
>and export the HKEY/CU/Software/iaxcomm node and ship my family the binary
>and the reg key. "Double click the .reg file then double click the phone."
>  
>
Great idea!

>Since this list is for all iaxclient stuff, email me off list with your GUI
>ideas.
>  
>
Will do.

>  
>
>>5) Control over the encoding format
>>    
>>
>
>Yes, I need to add controls to override the defaults.
> 
>  
>
>>And as an aside for later: Some type of network level 
>>encryption would be great (if it didn't make latency bad).
>>    
>>
>
>Any ideas on this, guys?
>
>I know just enough about encryption to know that ROT-13 ing the encoded
>payload is probably not the route to take.
>  
>
This is somewhat of a pipe dream, I think.  I dream of software that 
makes encryption so easy to use that I can use it with my mother.  I 
have no expectations of the such anytime soon.  I just saw the tunneling 
idea on the listserve and that will work for the time being.

Re: [Iaxclient-devel] Crash w/ IAXComm talking to another IAXComm

[Steve K. <st...@st...>]

Michael Van Donselaar wrote:

> 
>
>  
>
>>-----Original Message-----
>>From: iax...@li... 
>>[mailto:iax...@li...] On 
>>Behalf Of Nathan Sharp
>>Sent: Tuesday, November 09, 2004 11:06 AM
>>To: mi...@va...
>>Cc: iax...@li...
>>Subject: Re: [Iaxclient-devel] Crash w/ IAXComm talking to 
>>another IAXComm
>>
>>Michael,
>>  Thanks so much for the quick reply!  To answer your 
>>questions, I am calling through an Asterisk 1.0.1 server 
>>(with Speex disabled).  I am not using the echo cancellation 
>>feature.  
>>
>>  Could this be the same problem?
>>http://sourceforge.net/mailarchive/forum.php?thread_id=5836576
>>&forum_id=29380
>>
>>    
>>
>
>I'm really not totally sure, but I think there's still a problem with codec
>negotiation.
>
>Could you try calling peer to peer and see what you get?
>  
>
It could be another problem in native transfer, though, like the one we 
fixed a couple of weeks ago (Are you binaries from past that date?).  I 
think this is the problem referred to in that post.  I fixed that on 10/29.



>I am also going to post new binaries tonite or tomorrow compiled on RH9 and
>WinXP.  They will have CVS updates for iaxclient, and be set to prefer iLBC.
>
>  
>
>>  Only the existence of NAT or not doesn't seem to matter for 
>>me.  I have hooked up my asterisk server to a FWD account 
>>using iax2.fwdnet.net and I've found that the client crashes 
>>also if I try to connect to FWD through my asterisk server.   
>>
>>Feature Requests?  My top priorities, in order:
>>1) Secure against malicious incoming packets
>>    
>>
>
>I don't know what to do here
>  
>
This, and encryption below, seem to be libiax2 and IAX2 protocol 
issues.  There's a bug in digium's CVS discussing IAX2 encryption.

Of course, if you control both endpoints, you could always run IAX2 (and 
iaxclient) through an encrypted tunnel.  I will probably offer an API in 
iaxclient so that you can integrate something like this at the code 
level soon, but even without that, you can tunnel easily because IAX2 is 
so nat-friendly.

I.e. you run like this:  [iaxclient] <localhost interface> [tunnel 
program] <encrypted tunneled link over internet] <tunnel program> 
[localhost interface] <asterisk>.

You'd make calls, register, etc via a localhost URL, I.e. dial 
foo:bar@localhost:port/extension.

This all works fine.  I think zebedee is a program that could do the 
tunneling for you.

[...]

>
>Any ideas on this, guys?
>
>I know just enough about encryption to know that ROT-13 ing the encoded
>payload is probably not the route to take.
>  
>

See above, and bugs.digium.com for info..

-SteveK

Re: [Iaxclient-devel] IAX Client Library on win32

[Dennis W. <den...@ya...>]

I know that some work has been done in the past for
MSVC support but am not completely sure of its outcome
or current status.

I (and I think many others) would appreciate knowing
the current status of things.

Depending on the level of needed work to complete this
task I would be willing to pay a bounty for this work
to be done.

/Dennis

--- Iftikhar Qureshi <ift...@ya...>
wrote:

> 
> Can someone point to current library source code
> that I can compile in VC++ 6.0 on win32.
> 
> The code I got from from cvs doesnt compile
> properly...
> 
> I'm using the code that comes in the iaxclient
> module.
> 
> 			
> ---------------------------------
> Do you Yahoo!?
>  Check out the new Yahoo! Front Page. www.yahoo.com


=====
Dennis White
Telefonik - "Unleashing Powers of Technology"
http://www.telefonik.com

RE: [Iaxclient-devel] Crash w/ IAXComm talking to another IAXComm

[Michael V. D. <mi...@va...>]

 

> -----Original Message-----
> From: iax...@li... 
> [mailto:iax...@li...] On 
> Behalf Of Nathan Sharp
> Sent: Tuesday, November 09, 2004 11:06 AM
> To: mi...@va...
> Cc: iax...@li...
> Subject: Re: [Iaxclient-devel] Crash w/ IAXComm talking to 
> another IAXComm
> 
> Michael,
>   Thanks so much for the quick reply!  To answer your 
> questions, I am calling through an Asterisk 1.0.1 server 
> (with Speex disabled).  I am not using the echo cancellation 
> feature.  
> 
>   Could this be the same problem?
> http://sourceforge.net/mailarchive/forum.php?thread_id=5836576
> &forum_id=29380
> 

I'm really not totally sure, but I think there's still a problem with codec
negotiation.

Could you try calling peer to peer and see what you get?

I am also going to post new binaries tonite or tomorrow compiled on RH9 and
WinXP.  They will have CVS updates for iaxclient, and be set to prefer iLBC.

>   Only the existence of NAT or not doesn't seem to matter for 
> me.  I have hooked up my asterisk server to a FWD account 
> using iax2.fwdnet.net and I've found that the client crashes 
> also if I try to connect to FWD through my asterisk server.   
> 
> Feature Requests?  My top priorities, in order:
> 1) Secure against malicious incoming packets

I don't know what to do here

> 2) Low Latency (10-2004 version seems much better than the 
> earlier version, when it doesn't crash). 

See Steve Kann's comments about this.  I'll recompile with a lower latency,
and see how it works.

> 3) Reliable.  The earlier version has never crashed yet on 
> me, it just has much higher latency.

That's a feature everyone can use!

> 4) Easy to install/use (so I can hand it out to family and 
> friends).  You are pretty close on this front already.  There 
> are a couple of minor GUI issues which could be cleaned up.

You shoulda seen the earlier versions!  I think I've got the install down to
the easiest I can without going to the point of using an installer:  a
single file that you can run from anywhere.

Quick hint that I should put in the QUICKSTART:  I trim down the registry
and export the HKEY/CU/Software/iaxcomm node and ship my family the binary
and the reg key. "Double click the .reg file then double click the phone."

Since this list is for all iaxclient stuff, email me off list with your GUI
ideas.

> 5) Control over the encoding format

Yes, I need to add controls to override the defaults.
 
> And as an aside for later: Some type of network level 
> encryption would be great (if it didn't make latency bad).

Any ideas on this, guys?

I know just enough about encryption to know that ROT-13 ing the encoded
payload is probably not the route to take.
 
>   Thanks again!
>    Nathan
> 
> Michael Van Donselaar wrote:
> 
> 
> 	
> 	Are you calling peer-to -peer (XP to RH8), or are both 
> phones registered with an
> 	asterisk server?
> 	
> 	Do you have echo cancel turned on?  I think there's 
> still a problem in there.  I
> 	probably ought to disable that.
> 	
> 	  
> 	
> 	Feature requests are welcome too!
> 	  
> 
> 
> 
> ---
> Incoming mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.791 / Virus Database: 535 - Release Date: 11/8/2004
> 
> 
> 

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.791 / Virus Database: 535 - Release Date: 11/8/2004
 

RE: [Iaxclient-devel] Feature request

[Michael V. D. <mi...@va...>]

Dan,

please email the xrc to me off-list along with the best translations for
remote, state, and any others.

I need to look into the best way to do this, and can't promise how soon I
can get to it, but will work on it. 

> -----Original Message-----
> From: iax...@li... 
> [mailto:iax...@li...] On 
> Behalf Of administrator tootai
> Sent: Tuesday, November 09, 2004 10:59 AM
> To: Iaxclient-list
> Subject: [Iaxclient-devel] Feature request
> 
> Hi,
> 
> I finished the french translation for iaxComm (rc directory. 
> xrc file) but I'm now sticked with all the text which is 
> hardcoded. Could you try to think about this? For instance, 
> State and Remote in the main window, the default values, aso. 
> They are a few of them.
> 
> Thanks
> 
> --
> Daniel
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Sybase ASE Linux Express Edition - download now for FREE
> LinuxWorld Reader's Choice Award Winner for best database on Linux.
> http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
> _______________________________________________
> Iaxclient-devel mailing list
> Iax...@li...
> https://lists.sourceforge.net/lists/listinfo/iaxclient-devel
> 
> ---
> Incoming mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.791 / Virus Database: 535 - Release Date: 11/8/2004
>  
> 

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.791 / Virus Database: 535 - Release Date: 11/8/2004
 

[Iaxclient-devel] Feature request

[administrator t. <ad...@to...>]

Hi,

I finished the french translation for iaxComm (rc directory. xrc file) 
but I'm now sticked with all the text which is hardcoded. Could you try 
to think about this? For instance, State and Remote in the main window, 
the default values, aso. They are a few of them.

Thanks

-- 
Daniel

Re: [Iaxclient-devel] Crash w/ IAXComm talking to another IAXComm

[Nathan S. <iax...@np...>]

Michael,
  Thanks so much for the quick reply!  To answer your questions, I am 
calling through an Asterisk 1.0.1 server (with Speex disabled).  I am 
not using the echo cancellation feature. 

  Could this be the same problem?
http://sourceforge.net/mailarchive/forum.php?thread_id=5836576&forum_id=29380

  Only the existence of NAT or not doesn't seem to matter for me.  I 
have hooked up my asterisk server to a FWD account using iax2.fwdnet.net 
and I've found that the client crashes also if I try to connect to FWD 
through my asterisk server.  

Feature Requests?  My top priorities, in order:
1) Secure against malicious incoming packets
2) Low Latency (10-2004 version seems much better than the earlier 
version, when it doesn't crash).
3) Reliable.  The earlier version has never crashed yet on me, it just 
has much higher latency.
4) Easy to install/use (so I can hand it out to family and friends).  
You are pretty close on this front already.  There are a couple of minor 
GUI issues which could be cleaned up.
5) Control over the encoding format

And as an aside for later: Some type of network level encryption would 
be great (if it didn't make latency bad).

  Thanks again!
   Nathan

Michael Van Donselaar wrote:

>
>Are you calling peer-to -peer (XP to RH8), or are both phones registered with an
>asterisk server?
>
>Do you have echo cancel turned on?  I think there's still a problem in there.  I
>probably ought to disable that.
>
>  
>
>
>Feature requests are welcome too!
>  
>

Re: [Iaxclient-devel] Crash w/ IAXComm talking to another IAXComm

[Michael V. D. <mic...@va...>]

On Tue, 09 Nov 2004 07:02:28 -0500, Nathan Sharp =
<iax...@np...>
wrote:

>Hi,
>  I recently found your IAXComm client and for the most part think it is=
=20
>great!  I am using Asterisk 1.0.1 and IAXComm 10-2004 and I can use the=20
>echo test and voicemail features of my server just fine.  However,=20
>whenever I try to call from my Windows XP box to my RedHat 8 box, both=20
>clients crash instantly when I click to answer the call.  Unfortunately=20
>I don't seem to have all the necessary devel libraries to compile a=20
>newer version.  Is there any debugging in particular that might be=20
>useful?  The only thing I can think of that is unusual is that I had to=20
>explicitly disable the speex codec on the server because otherwise it=20
>would accept the call in that format, then bail out because I don't have=
=20
>speex installed there.

Are you calling peer-to -peer (XP to RH8), or are both phones registered =
with an
asterisk server?

Do you have echo cancel turned on?  I think there's still a problem in =
there.  I
probably ought to disable that.

>
>  Kind regards, and thanks for such great software!
>   Nathan

=46eature requests are welcome too!
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by:
>Sybase ASE Linux Express Edition - download now for FREE
>LinuxWorld Reader's Choice Award Winner for best database on Linux.
>http://ads.osdn.com/?ad_id=3D5588&alloc_id=3D12065&op=3Dclick
>_______________________________________________
>Iaxclient-devel mailing list
>Iax...@li...
>https://lists.sourceforge.net/lists/listinfo/iaxclient-devel

[Iaxclient-devel] Crash w/ IAXComm talking to another IAXComm

[Nathan S. <iax...@np...>]

Hi,
  I recently found your IAXComm client and for the most part think it is 
great!  I am using Asterisk 1.0.1 and IAXComm 10-2004 and I can use the 
echo test and voicemail features of my server just fine.  However, 
whenever I try to call from my Windows XP box to my RedHat 8 box, both 
clients crash instantly when I click to answer the call.  Unfortunately 
I don't seem to have all the necessary devel libraries to compile a 
newer version.  Is there any debugging in particular that might be 
useful?  The only thing I can think of that is unusual is that I had to 
explicitly disable the speex codec on the server because otherwise it 
would accept the call in that format, then bail out because I don't have 
speex installed there.

  Kind regards, and thanks for such great software!
   Nathan

Re: [Iaxclient-devel] Next to Audio Caller problem

[daniel h. <de...@to...>]

Steve Kann a =E9crit :

> daniel huhardeaux wrote:
>
>>
>> Ok, I got it: problem is echo cancelation. If it's activated I have=20
>> my audio problems.
>
>
>
> OK, so, if you turn off EC, everything works right?

Yes
--=20
Daniel

Re: [Iaxclient-devel] Next to Audio Caller problem

[Steve K. <st...@st...>]

daniel huhardeaux wrote:

>
> Ok, I got it: problem is echo cancelation. If it's activated I have my 
> audio problems.


OK, so, if you turn off EC, everything works right?

EC was actually _really_ broken recently, as the libspeex API to it 
changed, and our code didn't. So we were passing in arrays of float, but 
it now expects arrays of short.

I actually just changed this about 10 minutes before I read this mail; I 
was going through things to eliminate the warnings in the original 
iaxclient code (as opposed to the tons of warnings thrown by GSM and 
libiax2).

However, the EC is still experimental at best.  It probably won't help 
you to use it, and it will often make things worse.   It's a great area 
for someone with some audio processing experience to contribute (by 
working with the speex project primarily).

-SteveK