Nabble - SourceForge - Sound/Audio

Re: mp3split shudders

2009-12-19T09:47:24Z

Hello again,

the source file has a synchronisation error at that exact point @5:47.
You can split the file in two at that exact location by using the 'sync
error mode' of mp3splt:
mp3splt -d output_directory -e source.mp3

You will most likely have problems with some players at that point.
However, I don't know yet why you don't have the same problem with the
original file.

--
Alex

On 18:04 Sat 19 Dec , Alexandru Munteanu wrote:

> On 08:07 Sat 19 Dec , Skip Coombe wrote:
> > mp3splt_2.2.6a_i386.exe was the 1st version I ever used. I cannot find
> >
> > > mp3splt_2.2.6_i386 on http://sourceforge.net/projects/mp3splt/files/
> >
> > > under "View all files"
> >
> > >
> > The test case files are:
> >
> > > 14 - Never Call Retreat_ Military an.mp3 // source mp3
> >
> > > 14 - Never Call Retreat_ Military an_05m_00s__10m_00s.mp3// split track
> >
> > Available at:
> > http://dl.dropbox.com/u/1681743/14%20-%20Never%20Call%20Retreat_%20Military%20an.mp3
> > (52 MB)
> >
> > http://dl.dropbox.com/u/1681743/14%20-%20Never%20Call%20Retreat_%20Military%20an_05m_00s__10m_00s.mp3(5
> > MB)
> >
> > Let me know if you have problems accessing them.
>
> Thank you.
> I have downloaded the files.
> I will test and reply after testing.
>
> --
> Alex
>
> > The problem can be seen (heard) in the difference between (source mp3)
> > @5:47 and (split track) @0:47, at which point the split track skips
> >
> > > back to the beginning of the source track.
> >
> > >
> > Note: Never Call Retreat_ Military an_05m_00s__10m_00s.mp3 is 5:26
> > min despite the name assigned by mp3splt.
> >
> > >
> > Produced with
> >
> > > "C:\Program Files\mp3splt\mp3splt.exe" -t 5.0 -a -d split $file.mp3
> >
> > > using mp3splt_2.2.7a_i386.
> >
> > >
> > Skip
> >
> > >
> > --
> >
> > > Skip Coombe
> >
> > > skipcoombe@...
>
>
> ------------------------------------------------------------------------------
> This SF.Net email is sponsored by the Verizon Developer Community
> Take advantage of Verizon's best-in-class app development support
> A streamlined, 14 day to market process makes app distribution fast and easy
> Join now and get one step closer to millions of Verizon customers
> http://p.sf.net/sfu/verizon-dev2dev
> _______________________________________________
> Mp3splt-support mailing list
> Mp3splt-support@...
> https://lists.sourceforge.net/lists/listinfo/mp3splt-support

--
Alexandru Munteanu

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
Mp3splt-support mailing list
Mp3splt-support@...
https://lists.sourceforge.net/lists/listinfo/mp3splt-support

From forum: mp3splt-support

Re: mp3split shudders

2009-12-19T09:04:04Z

On 08:07 Sat 19 Dec , Skip Coombe wrote:

> mp3splt_2.2.6a_i386.exe was the 1st version I ever used. I cannot find
>
> > mp3splt_2.2.6_i386 on http://sourceforge.net/projects/mp3splt/files/
>
> > under "View all files"
>
> >
> The test case files are:
>
> > 14 - Never Call Retreat_ Military an.mp3 // source mp3
>
> > 14 - Never Call Retreat_ Military an_05m_00s__10m_00s.mp3// split track
>
> Available at:
> http://dl.dropbox.com/u/1681743/14%20-%20Never%20Call%20Retreat_%20Military%20an.mp3
> (52 MB)
>
> http://dl.dropbox.com/u/1681743/14%20-%20Never%20Call%20Retreat_%20Military%20an_05m_00s__10m_00s.mp3(5
> MB)
>
> Let me know if you have problems accessing them.

Thank you.
I have downloaded the files.
I will test and reply after testing.

--
Alex

> The problem can be seen (heard) in the difference between (source mp3)
> @5:47 and (split track) @0:47, at which point the split track skips
>
> > back to the beginning of the source track.
>
> >
> Note: Never Call Retreat_ Military an_05m_00s__10m_00s.mp3 is 5:26
> min despite the name assigned by mp3splt.
>
> >
> Produced with
>
> > "C:\Program Files\mp3splt\mp3splt.exe" -t 5.0 -a -d split $file.mp3
>
> > using mp3splt_2.2.7a_i386.
>
> >
> Skip
>
> >
> --
>
> > Skip Coombe
>
> > skipcoombe@...

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
Mp3splt-support mailing list
Mp3splt-support@...
https://lists.sourceforge.net/lists/listinfo/mp3splt-support

From forum: mp3splt-support

Re: mp3split shudders

2009-12-19T05:07:31Z

mp3splt_2.2.6a_i386.exe was the 1st version I ever used. I cannot find

mp3splt_2.2.6_i386 on http://sourceforge.net/projects/mp3splt/files/

under "View all files"

The test case files are:

14 - Never Call Retreat_ Military an.mp3 // source mp3

14 - Never Call Retreat_ Military an_05m_00s__10m_00s.mp3// split track

Available at:

http://dl.dropbox.com/u/1681743/14%20-%20Never%20Call%20Retreat_%20Military%20an.mp3 (52 MB)

http://dl.dropbox.com/u/1681743/14%20-%20Never%20Call%20Retreat_%20Military%20an_05m_00s__10m_00s.mp3 (5 MB)

Let me know if you have problems accessing them.

The problem can be seen (heard) in the difference between (source mp3) @5:47 and (split track) @0:47, at which point the split track skips

back to the beginning of the source track.

Note: Never Call Retreat_ Military an_05m_00s__10m_00s.mp3 is 5:26

min despite the name assigned by mp3splt.

Produced with

"C:\Program Files\mp3splt\mp3splt.exe" -t 5.0 -a -d split $file.mp3

using mp3splt_2.2.7a_i386.

Skip

Skip Coombe

skipcoombe@...

--
Skip Coombe
skipcoombe@...

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
Mp3splt-support mailing list
Mp3splt-support@...
https://lists.sourceforge.net/lists/listinfo/mp3splt-support

From forum: mp3splt-support

Re: mp3split shudders

2009-12-17T16:05:44Z

Hello,

thank you for the feedback.

Does it work (better) with version 2.2.6 ?
The only change between 2.2.6 and 2.2.6a is a bug fix related to the 'gap'
parameter; this impacts on the use of the '-a' option.

I am sorry but I don't understand the behaviour.
Does it skips some audio parts or some audio parts are overlapped ?
Probably the simplest is to post a small example.
(e.g.: it produces files with the following audio parts - in minutes:
1) from 0.0 to 5.0 2) from 4.8 to 9.9 ... instead of:
1) from 0.0 to 5.0 2) from 5.0 to 10.0 ...)

Regards,
--
Alex

On 20:51 Wed 16 Dec , Skip Coombe wrote:

> Of course I meant "stutters".
>
> On Wed, Dec 16, 2009 at 6:09 PM, Skip Coombe <skipcoombe@...> wrote:
>
> > Great SW.
> >
> > I use it in a very simple way to split ~60 min lecture mp3's into tracks
> > that
> > can be burned to an audio CD (provides a navigation method for listening to
> > the
> > lectures in a car CD player).
> >
> > The command line is
> > "C:\Program Files\mp3splt\mp3splt.exe" -t 5.0 -a -d split $file.mp3
> > (XP SP3).
> >
> > The chunk size is not critical - just provides granular navigation.
> >
> > Occasionally the split shudders: after starting a track (40-120 sec) it
> > will drop in
> > an almost complete former track. The next track proceeds correctly.
> >
> > This first appeared under mp3splt_2.2.6a_i386.exe. mp3splt_2.2.7a_i386.exe
> > did not correct it.
> >
> > Can anyone suggest different command line options?
> >
> > Thanks,
> >
> > Skip Coombe
> > skipcoombe@...

From forum: mp3splt-support

Re: mp3split shudders

2009-12-16T17:51:50Z

Of course I meant "stutters".

On Wed, Dec 16, 2009 at 6:09 PM, Skip Coombe <skipcoombe@...> wrote:

Great SW.

I use it in a very simple way to split ~60 min lecture mp3's into tracks that
can be burned to an audio CD (provides a navigation method for listening to the
lectures in a car CD player).

The command line is
"C:\Program Files\mp3splt\mp3splt.exe" -t 5.0 -a -d split $file.mp3
(XP SP3).

The chunk size is not critical - just provides granular navigation.

Occasionally the split shudders: after starting a track (40-120 sec) it will drop in
an almost complete former track. The next track proceeds correctly.

This first appeared under mp3splt_2.2.6a_i386.exe. mp3splt_2.2.7a_i386.exe

did not correct it.

Can anyone suggest different command line options?

Thanks,

Skip Coombe
skipcoombe@...

--
Skip Coombe
skipcoombe@...

From forum: mp3splt-support

mp3split shudders

2009-12-16T15:09:32Z

Great SW.

I use it in a very simple way to split ~60 min lecture mp3's into tracks that

can be burned to an audio CD (provides a navigation method for listening to the

lectures in a car CD player).

The command line is

"C:\Program Files\mp3splt\mp3splt.exe" -t 5.0 -a -d split $file.mp3

(XP SP3).

The chunk size is not critical - just provides granular navigation.

Occasionally the split shudders: after starting a track (40-120 sec) it will drop in

an almost complete former track. The next track proceeds correctly.

This first appeared under mp3splt_2.2.6a_i386.exe. mp3splt_2.2.7a_i386.exe

did not correct it.

Can anyone suggest different command line options?

Thanks,

Skip Coombe

skipcoombe@...

--
Skip Coombe
skipcoombe@...

From forum: mp3splt-support

Re: Access restrictions in eclipse

2009-12-16T02:01:46Z

Hi Spencer,

I am answering to the list since this may be of interest for other users.

Is threading/concurrency already handled for each session -- of does the call manager need to do that?

Yes. This is already supported.

If CallManager does not have to deal with threading , a listener would be nice. Can you add one?

I added a Session Listener, that can be added once you obtained a session. Maybe, I should also add an error notification method?

When I look at the code, I think that I should refactor on thing or two for the client access. SessionListeners will not be supported if they are used remotely. So the remote interface must be different to the one that I use internally or that can be used from a call manager. Maybe I should do not use JVoiceXml, Session, etc when interacting remotely?

The other side of the coin is when the telephone is hung up, I need to stop and cleanup this jvoicxml session. I will use this method.
jvxmlSession.hangup();

Yes, this would be good.

Dirk

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Missing audio play toolbar in k3b 1.68

2009-12-15T14:50:46Z

Hello,

after upgrading to k3b 1.68 I'm missing the audio play toolbar which
used to be available for audio CD projects in 1.0.5. In case it is gone
I'd like to vote for including this kind of functionality again, as it
really helps with splitting audio tracks etc. without having to use an
external application (like audacity).

Also, the play entry in the context menu of audio tracks is greyed out.
Can someone point me into the right direction what might keep k3b from
making this functionality available? Everything in the 'Plugins' and
'Programs' settings looks fine and extracodecs are installed. (Base OS
is a current release of some popular Debian derived Linux distribution)

Thanks in advance,
Arvid

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
K3b-user mailing list
K3b-user@...
https://lists.sourceforge.net/lists/listinfo/k3b-user

From forum: k3b-user

new list dvb-t avellino montevergine

2009-12-15T00:50:26Z

thank for your consideration
antonio galasso
--

--
suse10.3-linux-antonio

# DVB-T MONTEVERGINE (AVELLINO) 83100
# 12/12/2009 BY ANTONIO GALASSO anto1714@...
(RICORDATI DI METTERE IL LETTORE IN POSIZIONE STOP ORIMA DI INIZIARE LA SCANSIONE DEI CANALI)

# T FREQ BW FEC_HI FEC_LO MOD TRANSMISSION-MODE GUARD-INTERVAL HIERARCHY

#C-MUX

T 177500000 7MHz 2/3 AUTO QAM64 8k 1/32 AUTO

T 482000000 8MHz 2/3 AUTO QAM64 8k 1/4 AUTO

T 498000000 8MHz 2/3 AUTO QAM64 8k 1/32 AUTO

T 514000000 8MHz 2/3 AUTO QAM64 8k 1/4 AUTO

T 522000000 8MHz 2/3 AUTO QAM64 8k 1/32 AUTO

T 530000000 8MHz 2/3 AUTO QAM64 8k 1/32 AUTO

T 546000000 8MHz 2/3 AUTO QAM64 8k 1/4 AUTO

T 554000000 8MHz 2/3 AUTO QAM64 8k 1/32 AUTO

T 594000000 8MHz 2/3 AUTO QAM64 8k 1/4 AUTO

T 642000000 8MHz 2/3 AUTO QAM64 8k 1/8 AUTO

T 682000000 8MHz 2/3 AUTO QAM64 8k 1/4 AUTO

T 690000000 8MHz 2/3 AUTO QAM64 8k 1/4 AUTO

T 698000000 8MHz 2/3 AUTO QAM64 8k 1/4 AUTO
------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
kaffeine-user mailing list
kaffeine-user@...
https://lists.sourceforge.net/lists/listinfo/kaffeine-user

From forum: kaffeine-user

Re: K3B will not start?

2009-12-14T08:46:39Z

Hi,

On 14/12/2009, brow96@... <brow96@...> wrote:
> I use Ubuntu for my OS. I haven't had a need to use K3B in the past few
> weeks. During this time I've had several updates of the OS. Apparently
> one of them broke K3B. Because now it simply refuses to run. No errors
> or logs are generated. I've uninstalled and reinstalled to no effect,

Are you sure there's no logs or output?

What if you run k3b from a terminal?
What if you run k3b with strace or ltrace?
eg. strace k3b

Usually updates that break applications will be in the form of some
library is missing or wrong version, or undefined references or a
crash.

But Ubuntu is weird in that it does strange things with your execution
ernvironment - a number of commands could be shell aliases or
symlinks.

Can you also do: which k3b
And also: file `which k3b`
And: ls -l `which k3b`

That might let you know if the weird Ubuntu /etc/alternatives symlinks
is pointing to the wrong file or if you have a shell alias for k3b
that does nothing.

eg.
qmake - a binary from Qt. One would expect this to be a normal
binary in /usr/bin/
but...

ls -l `which qmake`
lrwxrwxrwx 1 root root 23 2007-07-06 16:24 /usr/bin/qmake ->
/etc/alternatives/qmake

If the symlink was accidentally changed to /bin/false... then you'd
never get any output from it. So worth checking.

Good luck,
Srdjan

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
K3b-user mailing list
K3b-user@...
https://lists.sourceforge.net/lists/listinfo/k3b-user

From forum: k3b-user

K3B will not start?

2009-12-14T08:24:37Z

I've been a happy user of K3B for a long tine now. I've had no problems
burning data, video, and music CD's.

I use Ubuntu for my OS. I haven't had a need to use K3B in the past few
weeks. During this time I've had several updates of the OS. Apparently
one of them broke K3B. Because now it simply refuses to run. No errors
or logs are generated. I've uninstalled and reinstalled to no effect,
i.e. no errors on removal or reinstallation.

I'm open to any suggestions on this one.

Bill

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
K3b-user mailing list
K3b-user@...
https://lists.sourceforge.net/lists/listinfo/k3b-user

From forum: k3b-user

K3B will not start?

2009-12-14T08:24:27Z

From forum: k3b-user

Re: Request: kaffeine --dvb start parameter

2009-12-13T07:50:33Z

2009/12/5 Pascal <pascal@...>:
<snip>
> strangely enough this works if I close kaffeine witch Ctrl-Q or close the
> window using the close-button of the window decoration, but it does not work
> when I close kaffeine using "qdbus /MainApplication quit". In this case the
> last channel is not remembered. I dont know what this dbus function really
> does, as it is automatically generated and not directly implemented in the
> kaffeine source... Maybe it simply kills the process???

No. The problem was that the main window wasn't properly deconstructed
in that case (it has to be treated in a different way than all other
widgets); fixed in svn.

> cheers,
> Pascal

Thanks,

Christoph

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
kaffeine-devel mailing list
kaffeine-devel@...
https://lists.sourceforge.net/lists/listinfo/kaffeine-devel

From forum: kaffeine-devel

[SPAM]

2009-12-13T05:00:14Z

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
MPlayerOSX-devel mailing list
MPlayerOSX-devel@...
https://lists.sourceforge.net/lists/listinfo/mplayerosx-devel

disseisins.jpg (15K) Download Attachment

From forum: mplayerosx-devel

WIKI update

2009-12-11T04:13:48Z

Hi all,

I just started to enhance the documentation that is present in the WIKI
https://sourceforge.net/apps/mediawiki/jvoicexml/index.php

There were several requests to explicitly state the implementation progress
of each tag. Currently I described two, but this may be a start.

Please inform me, if you think that there is something missing in the
documentation or that is described not correctly.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

WIKI update

2009-12-11T04:11:52Z

Hi all,

I just started to enhance the documentation that is present in the WIKI
https://sourceforge.net/apps/mediawiki/jvoicexml/index.php

There were several requests to explicitly state the implementation progress of each tag. Currently I described two, but this may be a start.

Please inform me, if you think that there is something missing in the documentation or that is described not correctly.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

MRCPv2 grammars

2009-12-10T11:48:59Z

Hi Spencer,

I just introduced a generic grammar implementation that is based on a
grammar document. In the MRCPv2 layer I used this to implement load
grammar using a string based grammar document.

hth
Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T11:21:24Z

On Wed, Dec 9, 2009 at 10:49 AM, <jvoicexml-developer@...> wrote:

One more:

If all fails and the BaseCallManager does not fit into this concept, it is still possible to simply implement the org.jvoicexml.CallManager interface that has only rudimentary constraints. Maybe this would be easier?

Yes, it would be easier. This is a good fallback.

Dirk

From: jvoicexml-developer@... [mailto:jvoicexml-developer@...]
Sent: Wednesday, December 09, 2009 7:39 PM
To: jvoicexml-developer@...
Subject: Re: [JVoiceXML-developer] MRCP4j

The problem is probably with my implementation of loadGrammar in the implementation's mrcpv2UserInput. What do I need to return in this method and how do I create it? Can you take a look?

Not really, since I have only the jar.

The mrcp2userinput class is not in the znazibar jar. It is in the mrcpv2 implementation project.

org.jvoicexml.implementation.mrcpv2.Mrcpv2SpokenInput

The only component that is used from zanzibar jar is the sipServer (the singleton dicussed below).

OK. I was a bit confused since there is another Mrcpv2UserInput. You are talking about the SpokenInput. The first one is not called. You are right.

I have to delegate the request to the specific implementation platform if a grammar type is requested that does not belong to the supported grammars, e.g. JSGF.It may use classes that the core does not know.

This is still true, if we move the identifier and transformer back to the core. So it will also need jsapi.jar to create a RuleGrammarImplementation as the JSAPI 1.0 platform does.

I need to refactor and reorganize the code and jars so only what is needed is included -- don't need the full zanzibar.jar.

If I understand correctly, you are going to do some refactoring and then let me know what to do to use the jsgf grammars.

The Terminal has the handle since this is the part that manages the connection, e.g. it accepts the call.

I think this is the core of the issue. A terminal receiving calls is a JTAPI concept. I have a singleton (SipServer) recieving calls.

IMHO, the JTAPI/JSAPI concepts should stay in JTAPI and JSAPI implementation platforms -- and all implementation platforms should conform to a more general and abstract framework. This framework should be as simple and felxible as possible.

An alternative is to fully embrace JTAPI (and JSAPI) and make every implementation platform become a JTAPI (and JSAPI) provider.

I think that a singleton is not well suited to work in a general and abstract framework. I still think that it can be solved by the current basis.

How does the sip server behave, if a call arrives?

The sipServer (really an Agent) listens on a port. I am using JAIN SIP. So there are events for things like invite request received, invite response received etc, info requests -- all of the sip messages. It also has methods so that you can initiate messages. In a way it is both a server and a client.

There are transactons on cleinet an server side (for instance the invite request/response/ack is a transaction. There are session and dialogs that last for the call.. I am treating dialogs as session -- although they are technically different.

A typical interaction
   invite comes in (with sdp message (offer) defining media channel requested) handled by sip listener
   invite response sent back with sdp offer accepted or modified
   invite ack received

   sesion occurs (audio channels)

   bye request comes in (on listener).
   byte response sent back.

The design is rather simple. There are two components. SipServer and CallManager
CallManager implements an interface with these methods.
public void startNewMrcpDialog(SipSession pbxSession, SipSession mrcpSession) throws Exception;
    public void StopDialog(SipSession session) throws SipException;
    public void dtmf(SipSession session, char code);

CallManager maintains a list of active calls. The only methd SipServer provides to Callmanager is
   public stopDialog(SipSession session) throws SipException;

I can take some steps to use terminals if you like.

I can preregister phone numbers with the sipserver and pass in a terminal that corresponds to that number.
Or I can use the number that I have in teh pbxSipsession to lookup a terminal.

But what do you expect this terminal to do? I suspect you want to move some of the logic out of SipServer and place it into terminal objects.

Perhasp you can tell me exactly what interfaces a terminal should provide.

I have the impression that the SipServer is what I call a Terminal. Or is it able to listen on more than a single sip endpoint? If yes, the concepts do not fit. If no, we should consider moving the logic to the Terminal implementation.

By single sip endpoint do you mean single call, or single phone numbers (each which could have multiple calls)? I think in either case the answer is yes.

Still, I can move some logic to a terminal objects and an extented BaseCallManager, but I am not sure where to start.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T10:49:43Z

One more:

If all fails and the BaseCallManager does not fit into this concept, it is still possible to simply implement the org.jvoicexml.CallManager interface that has only rudimentary constraints. Maybe this would be easier?

Dirk

From: jvoicexml-developer@... [mailto:jvoicexml-developer@...]
Sent: Wednesday, December 09, 2009 7:39 PM
To: jvoicexml-developer@...
Subject: Re: [JVoiceXML-developer] MRCP4j

The problem is probably with my implementation of loadGrammar in the implementation's mrcpv2UserInput. What do I need to return in this method and how do I create it? Can you take a look?

Not really, since I have only the jar.

The mrcp2userinput class is not in the znazibar jar. It is in the mrcpv2 implementation project.

org.jvoicexml.implementation.mrcpv2.Mrcpv2SpokenInput

The only component that is used from zanzibar jar is the sipServer (the singleton dicussed below).

OK. I was a bit confused since there is another Mrcpv2UserInput. You are talking about the SpokenInput. The first one is not called. You are right.

I have to delegate the request to the specific implementation platform if a grammar type is requested that does not belong to the supported grammars, e.g. JSGF.It may use classes that the core does not know.

This is still true, if we move the identifier and transformer back to the core. So it will also need jsapi.jar to create a RuleGrammarImplementation as the JSAPI 1.0 platform does.

The Terminal has the handle since this is the part that manages the connection, e.g. it accepts the call.

I think this is the core of the issue. A terminal receiving calls is a JTAPI concept. I have a singleton (SipServer) recieving calls.

IMHO, the JTAPI/JSAPI concepts should stay in JTAPI and JSAPI implementation platforms -- and all implementation platforms should conform to a more general and abstract framework. This framework should be as simple and felxible as possible.

An alternative is to fully embrace JTAPI (and JSAPI) and make every implementation platform become a JTAPI (and JSAPI) provider.

I think that a singleton is not well suited to work in a general and abstract framework. I still think that it can be solved by the current basis.

How does the sip server behave, if a call arrives?

The sipServer (really an Agent) listens on a port. I am using JAIN SIP. So there are events for things like invite request received, invite response received etc, info requests -- all of the sip messages. It also has methods so that you can initiate messages. In a way it is both a server and a client.

There are transactons on cleinet an server side (for instance the invite request/response/ack is a transaction. There are session and dialogs that last for the call.. I am treating dialogs as session -- although they are technically different.

A typical interaction
   invite comes in (with sdp message (offer) defining media channel requested) handled by sip listener
   invite response sent back with sdp offer accepted or modified
   invite ack received

   sesion occurs (audio channels)

   bye request comes in (on listener).
   byte response sent back.

The design is rather simple. There are two components. SipServer and CallManager
CallManager implements an interface with these methods.
public void startNewMrcpDialog(SipSession pbxSession, SipSession mrcpSession) throws Exception;
    public void StopDialog(SipSession session) throws SipException;
    public void dtmf(SipSession session, char code);

CallManager maintains a list of active calls. The only methd SipServer provides to Callmanager is
   public stopDialog(SipSession session) throws SipException;

I can take some steps to use terminals if you like.

I can preregister phone numbers with the sipserver and pass in a terminal that corresponds to that number.
Or I can use the number that I have in teh pbxSipsession to lookup a terminal.

But what do you expect this terminal to do? I suspect you want to move some of the logic out of SipServer and place it into terminal objects.

Perhasp you can tell me exactly what interfaces a terminal should provide.

I have the impression that the SipServer is what I call a Terminal. Or is it able to listen on more than a single sip endpoint? If yes, the concepts do not fit. If no, we should consider moving the logic to the Terminal implementation.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T10:38:36Z

The problem is probably with my implementation of loadGrammar in the implementation's mrcpv2UserInput. What do I need to return in this method and how do I create it? Can you take a look?

Not really, since I have only the jar.

OK. I was a bit confused since there is another Mrcpv2UserInput. You are talking about the SpokenInput. The first one is not called. You are right.

This is still true, if we move the identifier and transformer back to the core. So it will also need jsapi.jar to create a RuleGrammarImplementation as the JSAPI 1.0 platform does.

The Terminal has the handle since this is the part that manages the connection, e.g. it accepts the call.

I think this is the core of the issue. A terminal receiving calls is a JTAPI concept. I have a singleton (SipServer) recieving calls.

IMHO, the JTAPI/JSAPI concepts should stay in JTAPI and JSAPI implementation platforms -- and all implementation platforms should conform to a more general and abstract framework. This framework should be as simple and felxible as possible.

An alternative is to fully embrace JTAPI (and JSAPI) and make every implementation platform become a JTAPI (and JSAPI) provider.

I think that a singleton is not well suited to work in a general and abstract framework. I still think that it can be solved by the current basis.

How does the sip server behave, if a call arrives?

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T10:03:22Z

On Wed, Dec 9, 2009 at 8:57 AM, <jvoicexml-developer@...> wrote:

Hi Spencer,

I think that the transformer did not return a valid grammar. If you are using the IdentTransformer, it asks the UserInput implementation to load the grammar. This is the Mrcpv2UserInput located in the Zanzibar package. Should it be there or part of this implementation. We should try to figure this out. I added a check in the core if this happens to issue a warning that should easier lead to the cause.

How can you tell it is using Mrcpv2USerInput in the Zanzibar package? It was not my intention to do so.

Maybe it was by mistake but I was not able to find the java class in the source folder but in zanzibar-SNAPSHOT.jar. So we should also have it here?

There is an mrcp2userinput in zanzibar.jar that is used when it is running stand alone. But it should not be used. I expect that the mrcp2userinput object that is in this implementation is being used, as specified in the mrcpv2-implementation.xml (note the full name of the factory beans). By the way I should probably reorganize the code so that the full zanzibar.jar is not included here. It only uses a subset of the classes that handle call management.

The problem is probably with my implementation of loadGrammar in the implementation's mrcpv2UserInput. What do I need to return in this method and how do I create it? Can you take a look?

Not really, since I have only the jar.

The grammar config file is located in the etc folder of the JSAPI 1.0 implementation. I think that it is a special behavior of mrcp4j to support with JSGF grammars. Usually we have SRGS grammars here that are supported by the core. We should look for a better solution.

It seems to me that JSGF grammars belong in core with the SRGS grammars, since there are more than one implementations that use them.

Does Sphinx4 support SRGS? I am using JSGF becuase it is supportted by Sphinx4 -- There is no reason for MRCP not to support SRGS too.

Here is the grammar I was using, I dont think there is anything wrong with it. Sphinx 4 does not have a problem using it.

#JSGF V1.0;

/**
* JSGF Grammar for demo examples
*/

grammar example;

public <main> = ( [ <pre> ] ( <weather> {WEATHER} | <sports> {SPORTS} | <stocks> {STOCKS} ) ) | <quit> {QUIT};

<pre> = ( I would like [ to hear ] ) | ( hear ) | ( [ please ] get [ me ] ) | ( look up );

<weather> = [ the ] weather;

<sports> = sports [ news ];

<stocks> = ( [ a ] stock ( quote | quotes ) ) | stocks;

<quit> = exit | goodbye | quit;

I think that there is nothing wrong with the grammar. And you are right that sphinx4 supports only JSGF but not SRGS. My point was that MRCPv2 is intended to support SRGS XML in first place. Am I wrong in this point? You are the expert. I thought that MRCP was tightly coupled with this specific grammar format.

No it is not tightly coupled. It is carried as a MIME entity. MRCPv2 must support SRGS and may support others like JSGF and ABNF form of SRGS.

OK. Thanks for the info.

Maybe you are right that the JSGF grammar handling should be located in the core. I moved it to the JSAPI package since it used part of the jsapi.jar. I changed it and now there is no real reason to keep it there.

Yes I can see that it fits nicely there. But IMHO, the complexity introduced to configure other implementations to use JSGF outweighs that benefit.

Yes, you are right. So we should move it back to core.

The Terminal has the handle since this is the part that manages the connection, e.g. it accepts the call.

I think this is the core of the issue. A terminal receiving calls is a JTAPI concept. I have a singleton (SipServer) recieving calls.

IMHO, the JTAPI/JSAPI concepts should stay in JTAPI and JSAPI implementation platforms -- and all implementation platforms should conform to a more general and abstract framework. This framework should be as simple and felxible as possible.

An alternative is to fully embrace JTAPI (and JSAPI) and make every implementation platform become a JTAPI (and JSAPI) provider.

I think that a singleton is not well suited to work in a general and abstract framework. I still think that it can be solved by the current basis.

How does the sip server behave, if a call arrives?

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T08:57:14Z

Hi Spencer,

I think that the transformer did not return a valid grammar. If you are using the IdentTransformer, it asks the UserInput implementation to load the grammar. This is the Mrcpv2UserInput located in the Zanzibar package. Should it be there or part of this implementation. We should try to figure this out. I added a check in the core if this happens to issue a warning that should easier lead to the cause.

How can you tell it is using Mrcpv2USerInput in the Zanzibar package? It was not my intention to do so.

Maybe it was by mistake but I was not able to find the java class in the source folder but in zanzibar-SNAPSHOT.jar. So we should also have it here?

There is an mrcp2userinput in zanzibar.jar that is used when it is running stand alone. But it should not be used. I expect that the mrcp2userinput object that is in this implementation is being used, as specified in the mrcpv2-implementation.xml (note the full name of the factory beans). By the way I should probably reorganize the code so that the full zanzibar.jar is not included here. It only uses a subset of the classes that handle call management.

The problem is probably with my implementation of loadGrammar in the implementation's mrcpv2UserInput. What do I need to return in this method and how do I create it? Can you take a look?

Not really, since I have only the jar.

The grammar config file is located in the etc folder of the JSAPI 1.0 implementation. I think that it is a special behavior of mrcp4j to support with JSGF grammars. Usually we have SRGS grammars here that are supported by the core. We should look for a better solution.

It seems to me that JSGF grammars belong in core with the SRGS grammars, since there are more than one implementations that use them.

Does Sphinx4 support SRGS? I am using JSGF becuase it is supportted by Sphinx4 -- There is no reason for MRCP not to support SRGS too.

Here is the grammar I was using, I dont think there is anything wrong with it. Sphinx 4 does not have a problem using it.

#JSGF V1.0;

/**
* JSGF Grammar for demo examples
*/

grammar example;

public <main> = ( [ <pre> ] ( <weather> {WEATHER} | <sports> {SPORTS} | <stocks> {STOCKS} ) ) | <quit> {QUIT};

<pre> = ( I would like [ to hear ] ) | ( hear ) | ( [ please ] get [ me ] ) | ( look up );

<weather> = [ the ] weather;

<sports> = sports [ news ];

<stocks> = ( [ a ] stock ( quote | quotes ) ) | stocks;

<quit> = exit | goodbye | quit;

I think that there is nothing wrong with the grammar. And you are right that sphinx4 supports only JSGF but not SRGS. My point was that MRCPv2 is intended to support SRGS XML in first place. Am I wrong in this point? You are the expert. I thought that MRCP was tightly coupled with this specific grammar format.

No it is not tightly coupled. It is carried as a MIME entity. MRCPv2 must support SRGS and may support others like JSGF and ABNF form of SRGS.

OK. Thanks for the info.

Maybe you are right that the JSGF grammar handling should be located in the core. I moved it to the JSAPI package since it used part of the jsapi.jar. I changed it and now there is no real reason to keep it there.

Yes I can see that it fits nicely there. But IMHO, the complexity introduced to configure other implementations to use JSGF outweighs that benefit.

Yes, you are right. So we should move it back to core.

The Terminal has the handle since this is the part that manages the connection, e.g. it accepts the call.

I think this is the core of the issue. A terminal receiving calls is a JTAPI concept. I have a singleton (SipServer) recieving calls.

IMHO, the JTAPI/JSAPI concepts should stay in JTAPI and JSAPI implementation platforms -- and all implementation platforms should conform to a more general and abstract framework. This framework should be as simple and felxible as possible.

An alternative is to fully embrace JTAPI (and JSAPI) and make every implementation platform become a JTAPI (and JSAPI) provider.

I think that a singleton is not well suited to work in a general and abstract framework. I still think that it can be solved by the current basis.

How does the sip server behave, if a call arrives?

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T08:38:16Z

On Wed, Dec 9, 2009 at 12:39 AM, Dr. Dirk Schnelle-Walka <dirk@...> wrote:

Hi Spencer,

I think that the transformer did not return a valid grammar. If you are using the IdentTransformer, it asks the UserInput implementation to load the grammar. This is the Mrcpv2UserInput located in the Zanzibar package. Should it be there or part of this implementation. We should try to figure this out. I added a check in the core if this happens to issue a warning that should easier lead to the cause.

How can you tell it is using Mrcpv2USerInput in the Zanzibar package? It was not my intention to do so.

Maybe it was by mistake but I was not able to find the java class in the source folder but in zanzibar-SNAPSHOT.jar. So we should also have it here?

The grammar config file is located in the etc folder of the JSAPI 1.0 implementation. I think that it is a special behavior of mrcp4j to support with JSGF grammars. Usually we have SRGS grammars here that are supported by the core. We should look for a better solution.

It seems to me that JSGF grammars belong in core with the SRGS grammars, since there are more than one implementations that use them.

Does Sphinx4 support SRGS? I am using JSGF becuase it is supportted by Sphinx4 -- There is no reason for MRCP not to support SRGS too.

Here is the grammar I was using, I dont think there is anything wrong with it. Sphinx 4 does not have a problem using it.

#JSGF V1.0;

/**
* JSGF Grammar for demo examples
*/

grammar example;

public <main> = ( [ <pre> ] ( <weather> {WEATHER} | <sports> {SPORTS} | <stocks> {STOCKS} ) ) | <quit> {QUIT};

<pre> = ( I would like [ to hear ] ) | ( hear ) | ( [ please ] get [ me ] ) | ( look up );

<weather> = [ the ] weather;

<sports> = sports [ news ];

<stocks> = ( [ a ] stock ( quote | quotes ) ) | stocks;

<quit> = exit | goodbye | quit;

I think that there is nothing wrong with the grammar. And you are right that sphinx4 supports only JSGF but not SRGS. My point was that MRCPv2 is intended to support SRGS XML in first place. Am I wrong in this point? You are the expert. I thought that MRCP was tightly coupled with this specific grammar format.

No it is not tightly coupled. It is carried as a MIME entity. MRCPv2 must support SRGS and may support others like JSGF and ABNF form of SRGS.

Maybe you are right that the JSGF grammar handling should be located in the core. I moved it to the JSAPI package since it used part of the jsapi.jar. I changed it and now there is no real reason to keep it there.

Yes I can see that it fits nicely there. But IMHO, the complexity introduced to configure other implementations to use JSGF outweighs that benefit.

2. I am not sure how to get the application url given the callerid. It should be a simple map? See the methods startNewMrcpDialog and startNewCloudDialog in SipCallManager

Is it really dependent on the called id? IMHO it should be the called Id.

yes, you are right.

What you put in these methods is what is already in BaseCallManager.createSession(). This is also the reason why the RemoteClientFactory is not protected. This should be called by the terminal. I already started two methods in the SipTerminal that call this method and marked them with a TODO.

Given I have a called ID, how do I get the handle to the correct terminal?

The Terminal has the handle since this is the part that manages the connection, e.g. it accepts the call.

The intention of the terminal is to be a representative for the live connection. Is there really a difference between SipCallManagerSession and SipTerminal or do they share the same goal?

My understanding of terminals is that they represent a called ID that runs an app and can have multiple live connections -- each running the same app. Is this correct?

A simcallmanagersession represents a single live connections.

Please see above. I designed the terminal to represent exact the live connection.

Thanks for your comments. Hope that we get things clearer.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T00:39:29Z

Hi Spencer,

How can you tell it is using Mrcpv2USerInput in the Zanzibar package? It was not my intention to do so.

Maybe it was by mistake but I was not able to find the java class in the source folder but in zanzibar-SNAPSHOT.jar. So we should also have it here?

The grammar config file is located in the etc folder of the JSAPI 1.0 implementation. I think that it is a special behavior of mrcp4j to support with JSGF grammars. Usually we have SRGS grammars here that are supported by the core. We should look for a better solution.

It seems to me that JSGF grammars belong in core with the SRGS grammars, since there are more than one implementations that use them.

Does Sphinx4 support SRGS? I am using JSGF becuase it is supportted by Sphinx4 -- There is no reason for MRCP not to support SRGS too.

Here is the grammar I was using, I dont think there is anything wrong with it. Sphinx 4 does not have a problem using it.

#JSGF V1.0;

/**
* JSGF Grammar for demo examples
*/

grammar example;

public <main> = ( [ <pre> ] ( <weather> {WEATHER} | <sports> {SPORTS} | <stocks> {STOCKS} ) ) | <quit> {QUIT};

<pre> = ( I would like [ to hear ] ) | ( hear ) | ( [ please ] get [ me ] ) | ( look up );

<weather> = [ the ] weather;

<sports> = sports [ news ];

<stocks> = ( [ a ] stock ( quote | quotes ) ) | stocks;

<quit> = exit | goodbye | quit;

I think that there is nothing wrong with the grammar. And you are right that sphinx4 supports only JSGF but not SRGS. My point was that MRCPv2 is intended to support SRGS XML in first place. Am I wrong in this point? You are the expert. I thought that MRCP was tightly coupled with this specific grammar format.

Maybe you are right that the JSGF grammar handling should be located in the core. I moved it to the JSAPI package since it used part of the jsapi.jar. I changed it and now there is no real reason to keep it there.

2. I am not sure how to get the application url given the callerid. It should be a simple map? See the methods startNewMrcpDialog and startNewCloudDialog in SipCallManager

Is it really dependent on the called id? IMHO it should be the called Id.

yes, you are right.

What you put in these methods is what is already in BaseCallManager.createSession(). This is also the reason why the RemoteClientFactory is not protected. This should be called by the terminal. I already started two methods in the SipTerminal that call this method and marked them with a TODO.

Given I have a called ID, how do I get the handle to the correct terminal?

The Terminal has the handle since this is the part that manages the connection, e.g. it accepts the call.

The intention of the terminal is to be a representative for the live connection. Is there really a difference between SipCallManagerSession and SipTerminal or do they share the same goal?

My understanding of terminals is that they represent a called ID that runs an app and can have multiple live connections -- each running the same app. Is this correct?

A simcallmanagersession represents a single live connections.

Please see above. I designed the terminal to represent exact the live connection.

Thanks for your comments. Hope that we get things clearer.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-09T00:22:43Z

On Tue, Dec 8, 2009 at 11:56 PM, Dr. Dirk Schnelle-Walka <dirk@...> wrote:

Hi Spencer,

I am answering also to the list. Maybe this is also helpful for others to understand the concepts.

Please see below.

I did a big check in.

Great!!!

1. I am not sure how to handle including the jsgf grammar capability. Where is the grammar config xml file specified. I was able to sort of get it working, but ran into this exception

     [java] java.lang.NullPointerException
     [java]     at org.jvoicexml.interpreter.formitem.AbstractGrammarContainer.a
ddGrammar(AbstractGrammarContainer.java:78)
     [java]     at org.jvoicexml.interpreter.formitem.FieldFormItem.addGrammar(F
ieldFormItem.java:52)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.process
Grammars(FormInterpretationAlgorithm.java:706)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.activat
eGrammars(FormInterpretationAlgorithm.java:821)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.collect
(FormInterpretationAlgorithm.java:520)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.mainLoo
p(FormInterpretationAlgorithm.java:369)
     [java]     at org.jvoicexml.interpreter.VoiceXmlInterpreter.process(VoiceXm
lInterpreter.java:242)
     [java]     at org.jvoicexml.interpreter.VoiceXmlInterpreterContext.interpre
t(VoiceXmlInterpreterContext.java:521)
     [java]     at org.jvoicexml.interpreter.VoiceXmlInterpreterContext.process(
VoiceXmlInterpreterContext.java:321)
     [java]     at org.jvoicexml.interpreter.JVoiceXmlSession.run(JVoiceXmlSessi
on.java:266)

I think that the transformer did not return a valid grammar. If you are using the IdentTransformer, it asks the UserInput implementation to load the grammar. This is the Mrcpv2UserInput located in the Zanzibar package. Should it be there or part of this implementation. We should try to figure this out. I added a check in the core if this happens to issue a warning that should easier lead to the cause.

How can you tell it is using Mrcpv2USerInput in the Zanzibar package? It was not my intention to do so.

The grammar config file is located in the etc folder of the JSAPI 1.0 implementation. I think that it is a special behavior of mrcp4j to support with JSGF grammars. Usually we have SRGS grammars here that are supported by the core. We should look for a better solution.

2. I am not sure how to get the application url given the callerid. It should be a simple map? See the methods startNewMrcpDialog and startNewCloudDialog in SipCallManager

Is it really dependent on the called id? IMHO it should be the called Id.

yes, you are right.

What you put in these methods is what is already in BaseCallManager.createSession(). This is also the reason why the RemoteClientFactory is not protected. This should be called by the terminal. I already started two methods in the SipTerminal that call this method and marked them with a TODO.

Given I have a called ID, how do I get the handle to the correct terminal?

The intention of the terminal is to be a representative for the live connection. Is there really a difference between SipCallManagerSession and SipTerminal or do they share the same goal?

3. I made two call variables protected in BaseCallManager. I am guessing you will want to change them back...So it is a seperate checkin so you could revert it and find another solution.

   /** Factory to create the {@link org.jvoicexml.RemoteClient} instances. */
    protected RemoteClientFactory clientFactory;

    /** Established sessions. */
    protected final Map<Terminal, Session> sessions;

Please see above.

Thanks for your effort and commiting your stuff.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Re: MRCP4j

2009-12-08T23:56:23Z

Hi Spencer,

I am answering also to the list. Maybe this is also helpful for others to understand the concepts.

Please see below.

I did a big check in.

Great!!!

1. I am not sure how to handle including the jsgf grammar capability. Where is the grammar config xml file specified. I was able to sort of get it working, but ran into this exception

     [java] java.lang.NullPointerException
     [java]     at org.jvoicexml.interpreter.formitem.AbstractGrammarContainer.a
ddGrammar(AbstractGrammarContainer.java:78)
     [java]     at org.jvoicexml.interpreter.formitem.FieldFormItem.addGrammar(F
ieldFormItem.java:52)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.process
Grammars(FormInterpretationAlgorithm.java:706)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.activat
eGrammars(FormInterpretationAlgorithm.java:821)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.collect
(FormInterpretationAlgorithm.java:520)
     [java]     at org.jvoicexml.interpreter.FormInterpretationAlgorithm.mainLoo
p(FormInterpretationAlgorithm.java:369)
     [java]     at org.jvoicexml.interpreter.VoiceXmlInterpreter.process(VoiceXm
lInterpreter.java:242)
     [java]     at org.jvoicexml.interpreter.VoiceXmlInterpreterContext.interpre
t(VoiceXmlInterpreterContext.java:521)
     [java]     at org.jvoicexml.interpreter.VoiceXmlInterpreterContext.process(
VoiceXmlInterpreterContext.java:321)
     [java]     at org.jvoicexml.interpreter.JVoiceXmlSession.run(JVoiceXmlSessi
on.java:266)

The grammar config file is located in the etc folder of the JSAPI 1.0 implementation. I think that it is a special behavior of mrcp4j to support with JSGF grammars. Usually we have SRGS grammars here that are supported by the core. We should look for a better solution.

2. I am not sure how to get the application url given the callerid. It should be a simple map? See the methods startNewMrcpDialog and startNewCloudDialog in SipCallManager

Is it really dependent on the called id? IMHO it should be the called Id.

What you put in these methods is what is already in BaseCallManager.createSession(). This is also the reason why the RemoteClientFactory is not protected. This should be called by the terminal. I already started two methods in the SipTerminal that call this method and marked them with a TODO.

The intention of the terminal is to be a representative for the live connection. Is there really a difference between SipCallManagerSession and SipTerminal or do they share the same goal?

3. I made two call variables protected in BaseCallManager. I am guessing you will want to change them back...So it is a seperate checkin so you could revert it and find another solution.

   /** Factory to create the {@link org.jvoicexml.RemoteClient} instances. */
    protected RemoteClientFactory clientFactory;

    /** Established sessions. */
    protected final Map<Terminal, Session> sessions;

Please see above.

Thanks for your effort and commiting your stuff.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev

_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Reviews

2009-12-08T13:50:16Z

Hi all,

SF offers the ability to post reviews of the hosted projects. Some hours
ago, a user posted a very negative review.
https://sourceforge.net/projects/jvoicexml/reviews/

This is OK and the right of everybody, but IMHO it does not help much to
make things better.

So if this user really spent hours to make this buggy system work it would
be good to try getting help or telling what went wrong. Otherwise I do not
have a chance to improve it.

So I want to encourage you to tell me what is wrong with the current
implementation, where are leaks in the documentation and so on. This will
enable me to help you and improve the situation for everybody.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

Reorganizing the jars in the JSAPI 1.0 platform

2009-12-08T00:36:08Z

Hi all,

I just changed the organization of the created jars from the JSAPI 1.0
platform. I separated the grammar identification and transformation from the
rest and also renamed the jars.

Please check if this still works for you.

Dirk

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer

DVD Menu from remote control

2009-12-06T13:45:03Z

Over the past couple of days I have been configuring my remote control to operate Kaffeine. I've got just about everything I want to work using the "config = dcop kaffeine KaffeineIface ..." commands, but there are two functions I would like to set-up that "decop" doesn't seep to support.

These two functions are related to DVD Menues. I would like to know how to cinfigure the remote to navigate through lists such as "episode list" and then select "X" episode to play. I cannot figure out the programming for those two functions.

I hope I am being clear on what I am asking as I am quite new to Linux.

If you need me to be more clear or know how to help I would greatly appreciate an e-mail.

Thank You
Jeremy
jeremy.j.weimer@...
------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
kaffeine-user mailing list
kaffeine-user@...
https://lists.sourceforge.net/lists/listinfo/kaffeine-user

From forum: kaffeine-user

[SPAM] had to be procured, ready-made, from some other source. That

2009-12-06T06:57:02Z

not in all respects identical with those of the middle-classes. Of
course the unskilled labourer earning eighteen shillings a week has not
equal opportunities with the man who earns thirty-six; he cannot buy the
newspapers and occasional books to which the other treats himself and
his children, and in general he is less well informed. But the same
grave and circumspect talk goes down with the one as with the other; to
both the same topics are interesting. And for me the probability of a
development for our village labourers similar to that of the town
artisans is heightened, by recollection of what artisans themselves were
like, say a quarter of a century ago. I knew a few of these very well.
As craftsmen they were as able as those of to-day; but their crafts had
not taught them to think. While they worked by rule of thumb, outside
their work they were as full of prejudices, and as unable to grasp
reasons, as any of my village neighbours. The most of them, in fact, had
been born in villages near the town, and retained a good deal of the
rural outlook. Their gardens, and the harvest--yes, and odd scraps of
very ancient folk-lore which they still believed--occupied an important
place in their attention. They had quite the old attit

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
MPlayerOSX-devel mailing list
MPlayerOSX-devel@...
https://lists.sourceforge.net/lists/listinfo/mplayerosx-devel

aline.jpg (13K) Download Attachment

From forum: mplayerosx-devel

[SPAM] Rtaining his wishes in the premises, and then pr

2009-12-05T15:23:50Z

Y were killing me by inches," she protested. "Ouch!" he groaned, his leg
giving him a mighty twinge. "What is it?" she cried in alarm. "Why
should we wait for those men? Come, Mr. Bonner, take me to the
village--please do. I am crazy, absolutely crazy, to see Daddy Crow and
mother. I can walk there--how far is it?--please come." She was running
on eagerly in this strain until she saw the look of pain in his
face--the look he tried so hard to conceal. She was standing straight
and strong and eager before him, and he was very pale under the tan. "I
can't, Miss Gray. I'm sorry, you know. See! Where there's smoke there's
fire--I mean, where there's blood there's a wound. I'm done for, in
other words." "Done for? Oh, you're not--not going to die! Are you hurt?
Why didn't you tell me?" Whereupon she dropped to her knees at his side,
her dark eyes searching his intently, despair in them until the winning
smile str

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
MPlayerOSX-devel mailing list
MPlayerOSX-devel@...
https://lists.sourceforge.net/lists/listinfo/mplayerosx-devel

dolmetsch.jpg (11K) Download Attachment

From forum: mplayerosx-devel

Re: Request: kaffeine --dvb start parameter

2009-12-05T14:45:48Z

Christoph Pfister schrieb:

2009/11/17 Boris Cuber kaffeine@...:

Hi kaffeine people,

One feature i miss from the old versions is the possibilty
to start kaffeine from the cmdline in dvb mode with
the last tuned channel.

This would also keep some compatability to ~kaffeine-0.8.x.
I dunno if this is really a "wanted" feature, but it improves my
tv/dvb watching experience. I'm sure this could also be done
by altering the --tv parameter, but i'm no qt/c++ programmer
and my tries ended up in segfaults when starting kaffeine.

Whatever, it'd be cool to have this feature back ;)



A simple patch (applies to trunk), adds a --dvb parameter:


Applied (I've called the option "--lastchannel").

strangely enough this works if I close kaffeine witch Ctrl-Q or close the window using the close-button of the window decoration, but it does not work when I close kaffeine using "qdbus /MainApplication quit". In this case the last channel is not remembered. I dont know what this dbus function really does, as it is automatically generated and not directly implemented in the kaffeine source... Maybe it simply kills the process???

cheers,
Pascal

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
kaffeine-devel mailing list
kaffeine-devel@...
https://lists.sourceforge.net/lists/listinfo/kaffeine-devel

From forum: kaffeine-devel

Re: dbus support for kaffeine

2009-12-05T14:09:42Z

Hi Christoph,

with my old patch the channel ui didn't get updated when the channel was set
using the SetNumber(int) function, i now fixed it. I also cleaned up a
little bit my code in dvbtab.cpp and fixed a possible segmentation fault
when the dbus function SetDvbChannel(Qstring) was called with an invalid
name. The patch is against rev. 1053875.

Finally I managed to review, overwork and commit your patch ...

Thank you Christoph for taking my work into account and for this detailled explanation. I'm not sure I really made your life easier, as you finally had to change much of what I've done, but at least I could give a little push to the whole dbus thing ;) I will not go into all the points you mentioned, as it's all ok and I have nothing to add...

c) continue hacking (if you need a task, you can try to create a
kdelirc profile for kaffeine, that's an xml file which provides
default associations between remote control buttons and kaffeine dbus
interface [2])

I'll continue hacking for sure and will have a look at the kdelirc profiles, I think I can do something in the next time. I'm happy if I can help, you're doing great work and there are really much things to do, this cannot be taken for granted, all this work only because there is a new Qt version... Thumbs up!

ciao,
Pascal

Christoph Pfister schrieb:

Hi,

2009/11/24 Pascal pascal@...:

Hi Christoph,

with my old patch the channel ui didn't get updated when the channel was set
using the SetNumber(int) function, i now fixed it. I also cleaned up a
little bit my code in dvbtab.cpp and fixed a possible segmentation fault
when the dbus function SetDvbChannel(Qstring) was called with an invalid
name. The patch is against rev. 1053875.


Finally I managed to review, overwork and commit your patch ...

cheers,
pascal

+class MprisPlayerObject : public QObject
+{
+       Q_OBJECT
+       Q_CLASSINFO("D-Bus Interface", "org.freedesktop.MediaPlayer")
+public:
+       explicit MprisPlayerObject(MediaWidget *mediaWidget_);
+       ~MprisPlayerObject();
+
+public slots:
+       void Next();
+       void Prev();
+       void Play();
+       void Pause();
+       void Stop();
+       int  VolumeGet();
+       void VolumeSet(int volume);
+       int  PositionGet();
+       void PositionSet(int position);
+       void Repeat();


You forgot the bool parameter.

+       MprisStatusStruct GetStatus();
+       QVariantMap GetMetadata();


MPRIS metadata and Phonon metadata specifications aren't 100%
compatible ("location" means url in mpris terms, while it means
"location where track was recorded" in Phonon terms) ... I've left out
the metadata part for the moment.

+       int GetCaps();


All functions reordered into the order given by the spec.

+       // this functions are not part of the MPRIS specs
+       void PlayPause();


Not needed, Pause() has the same behaviour than mediaWidget's playPause action.

+       void VolumeIncrease();
+       void VolumeDecrease();


Renamed to "IncreaseVolume" and "DecreaseVolume".

+       void Play(const QString &url);


Not needed, TrackList has an AddTrack() function.

+       void PlayAudioCd();
+       void PlayVideoCd();
+       void PlayDvd();
+       void ChangeAudioChannel(int index);
+       void ChangeSubtitle(int index);
+       void LongSkipBackward();
+       void SkipBackward();
+       void SkipForward();
+       void LongSkipForward();
+       void TimeButtonClicked();


Left out for the moment (this has much to do with my workflow, not
with the actual content of the patch ... I'll think about those
functions once I have some spare time again).

+       void ToggleMuted();

+       void UpdateTimeButton();


Not needed, internal function.

+signals:
+       void CapsChange( int );
+       void StatusChange( MprisStatusStruct );
+       void TrackChange( QVariantMap );


Reordered into the order given by the spec.

+private:
+       MediaWidget *mediaWidget;

+class MprisTrackListObject : public QObject
+{
+       Q_OBJECT
+       Q_CLASSINFO("D-Bus Interface", "org.freedesktop.MediaPlayer")
+public:
+       explicit MprisTrackListObject(MediaWidget *mediaWidget_);
+       ~MprisTrackListObject();
+
+public slots:
+       QVariantMap GetMetadata();
+       int GetCurrentTrack();
+       int GetLength();
+       int AddTrack(QString track, bool play);


Use "const QString &", not "QString".

+       void DelTrack(int position);
+       void SetLoop(bool loop);
+       void SetRandom(bool random);
+
+signals:
+       void TrackListChange(int);
+
+private:
+       MediaWidget *mediaWidget;


All those functions have a relation to PlaylistTab, not to MediaWidget.

+class MprisDvbObject : public QObject


Renamed to "DBusTelevisionObject" (nothing to do with mpris).

+{
+       Q_OBJECT
+       Q_CLASSINFO("D-Bus Interface", "org.freedesktop.MediaPlayer")
+public:
+       explicit MprisDvbObject(MediaWidget *mediaWidget_);
+       ~MprisDvbObject();
+
+public slots:
+       void SetDvbChannel(const QString &name);
+       void SetDvbChannelNumber(int number);


Merged into a single function - playChannel(const QString &nameOrNumber);

+       void SetNumber(int number);


Not needed (what's the sense of this function?)
Added a DigitPressed(int) function, for entering channel numbers via
remote control.

+       void SetDvbLastChannel();


Renamed to PlayLastChannel().

+       void StopDvb();


Not needed, internal function.

+       void ToggleInstantRecord();
+       void ToggleOsd();
+
+private:
+       MediaWidget *mediaWidget;


All those functions have a relation to DvbTab, not to MediaWidget.

+       void AspectRatioAuto();
+       void AspectRatio4_3();
+       void AspectRatio16_9();
+       void AspectRatioWidget();


Left out for the moment (this has much to do with my workflow, not
with the actual content of the patch ... I'll think about those
functions once I have some spare time again).

+       void ToggleFullScreen();


Moved to the player object.

+struct MprisStatusStruct
+{
+       quint16 play;
+       quint16 random;
+       quint16 repeatTrack;
+       quint16 repeatAll;
+};


The spec says 32 bit integers (= int).

+Q_DECLARE_METATYPE(MprisStatusStruct)
+
+/**
+* The bit values for the capabilities flags
+*/
+enum Cap {
+       NO_CAPS                    = 0,
+       CAN_GO_NEXT                = 1 << 0,
+       CAN_GO_PREV                = 1 << 1,
+       CAN_PAUSE                        = 1 << 2,
+       CAN_PLAY                          = 1 << 3,
+       CAN_SEEK                          = 1 << 4,
+       CAN_PROVIDE_METADATA  = 1 << 5,
+       CAN_HAS_TRACKLIST        = 1 << 6,
+       ALL_KNOWN_CAPS          = (1 << 7) - 1
+};
+Q_DECLARE_FLAGS(Caps, Cap)


Directly merged into GetCaps (because it is the only function that
will use caps).

+signals:
+       void updateDvbChannelUi(int number);


Solved that problem in a different way (ChannelModel returns the row,
not the channel itself --> the proxy model can map that information to
the correct view model index).

+void DvbTab::toggleInstantRecord()
+{
+       if (instantRecordAction->isChecked()) {
+               instantRecordAction->setChecked(false);
+               instantRecord(false);
+       }
+       else {
+               instantRecordAction->setChecked(true);
+               instantRecord(true);
+       }


Simply use instantRecordAction->trigger();

+}

+bool MediaWidget::isPaused()
+{
+       if (playing && mediaObject->state() == Phonon::PausedState)
+       {
+               return true;
+       }
+       return false;


Query the playPauseAction.

+}
+int MediaWidget::getVolume()
+{
+       int volume = audioOutput->volume() * 100;
+
+       if ( volume >= 0 && volume <= 100 ) {
+               return volume;
+       }
+
+       return 0;


Query the volumeSlider.

+}
+
 void MediaWidget::changeVolume(int volume)
 {
       audioOutput->setVolume(volume * qreal(0.01));
@@ -915,6 +1014,16 @@
       mediaObject->seek(mediaObject->currentTime() + 1000 *
longSkipDuration);
 }

 class MediaWidget : public QWidget
 {
       Q_OBJECT
+       friend class MprisPlayerObject;
+       friend class MprisTrackListObject;
+       friend class MprisDvbObject;
+       friend class MprisDisplayObject;


Do not use friend classes for this purpose because of two reasons:
- there are internal functions that _really_ shouldn't be called from
outside (e.g. volumeChanged or mutedChanged, because they don't update
ui state)
- private functions usually trust their arguments etc, but in this
context more value checking is needed for calls from outside

So I've made the required functions public and rechecked that they
work correctly in all situations.

+       bool hasMedia();
+       void setNumber(int number);
+       void changeDvbChannel(int number);
+       void changeDvbChannel(const QString &name);
+       void toggleOsd();
+       void dvbLastChannel();
+       void toggleInstantRecord();
+       int getCurrentTrack();
+       void playPause();
+       void pause();
+       void setDvbChannel(int number);
+       void setDvbChannel(const QString &name);
+       void setDvbLastChannel();
+       void fullScreen();
+       void delTrack(int position);
+       void setLoop(bool loop);
+       void setRandom(bool random);


Removed all those functions because they are a) redundant (playPause,
pause --> simplified to togglePause()) or b) don't belong here (the
dbus objects call directly the needed functions from PlaylistTab or
DvbTab).

I suggest the following steps to proceed:
a) study the comments above a bit; the patches I've commited are [1];
feel free to ask questions
b) make your svn repo clean and update it
c) continue hacking (if you need a task, you can try to create a
kdelirc profile for kaffeine, that's an xml file which provides
default associations between remote control buttons and kaffeine dbus
interface [2])

Thanks a lot,

Christoph


[1]
http://websvn.kde.org/?view=revision&revision=1055907
http://websvn.kde.org/?view=revision&revision=1056261
http://websvn.kde.org/?view=revision&revision=1056694
[2]
the resulting file would be installed to
/usr/share/kde4/apps/profiles/kaffeine.profile.xml

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
kaffeine-devel mailing list
kaffeine-devel@...
https://lists.sourceforge.net/lists/listinfo/kaffeine-devel

From forum: kaffeine-devel

Re: Monkey audio

2009-12-04T16:24:51Z

Leon Pollak wrote, on 02/12/09 22:25:
> Hello, all.
>
> Is there somebody successfully using monkey audio plugin?
> I still has a problem with it - nothing works with my FC11 installation.
>
> Thanks.

I think that I found a source .deb and compiled it successfully to run
with the kde 3.5.X k3b.

Arthur.

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
K3b-user mailing list
K3b-user@...
https://lists.sourceforge.net/lists/listinfo/k3b-user

From forum: k3b-user

Re: MRCP4j

2009-12-03T08:00:28Z

Hi Spencer,

please see below.

From: spencer lord [mailto:spencer.lord@...]
Sent: Thursday, December 03, 2009 3:30 AM
To: Dr. Dirk Schnelle-Walka
Subject: Re: MRCP4j

I made some progress with the mrcp implementation for jvoicexml.

That is great news!

The current problem I have is:
     [java] org.jvoicexml.event.error.NoresourceError: error.noresource: Pool of type 'sip' is unknown!
     [java]     at org.jvoicexml.implementation.pool.KeyedResourcePool.borrowObject(KeyedResourcePool.java:113)
     [java]     at org.jvoicexml.implementation.jvxml.JVoiceXmlImplementationPlatform.getExternalResourceFromPool(JVoiceXmlImplementationPlatform.java:645)
     [java]     at org.jvoicexml.implementation.jvxml.JVoiceXmlImplementationPla
tform.getCallControl(JVoiceXmlImplementationPlatform.java:420)
     [java]     at org.jvoicexml.interpreter.tagstrategy.PromptStrategy.execute(PromptStrategy.java:141)
     [java]     at org.jvoicexml.interpreter.event.TagStrategyExecutor.executeTagStrategy(TagStrategyExecutor.java:169)
     [java]     at org.jvoicexml.interpreter.event.TagStrategyExecutor.executeChildNodes(TagStrategyExecutor.java:136)
     [java]     at org.jvoicexml.interpreter.event.TagStrategyExecutor.executeChildNodes(TagStrategyExecutor.java:80)

What should I return from the remote client in the getCallControl method? I wont need the play and recod methods to do anything. So maybe I need the dummy?

Well, it should be the dummy platform since there is no dedicated sip platform.

I also am having some difficulty understanding what to do with terminals -- I dont really think they are necessary -- maybe for jtapi -- but it seems like a lot of complexity that is not needed.

When the call comes in I need a way to lookup the application name, given a callee number or address. So I can pass the URI too the jvoicexml session. It is not clear to me how to do that. There is a list of termninals -- maybe that should be a hashmap with address/number as the key?

Maybe it is hard to get the idea. You are right that this concept is highly motivated by the JTAPI. So it might not fit with other platforms. I have similar problems if I think of a socket based call manager I discover similar problems.

Maybe it is time to refactor some things.

I think that we need something like that call manager that starts up things in advance, e.g. the server sockets or the SIP session manager. If I think of sockets, I can imagine multiple server sockets, each being responsible for a certain application. This can be done e.g. by a terminal. There also might be multiple connections to a terminal, each calling the same application. For JTAPI we have the special case to have multiple terminals for the same application. So a terminal must know the URL. Currently there is a mapping in the call manger from a terminal to a URL. This has to be changed so that a terminal also may know the URL.

What do you think?

Dirk

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing.
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Jvoicexml-developer mailing list
Jvoicexml-developer@...
https://lists.sourceforge.net/lists/listinfo/jvoicexml-developer

From forum: jvoicexml-developer