Re: [Bayonne-devel] Re: how to test the soundcard module

[David Sugar]

I have 1.1.0 cvs and ccscript 1.0.1 working better in this regard.  I 
can now run the debug mode server with the testit.bat file again, and it 
executes scripts passed correctly.  I backported those changes in the 
loadDriver/loadBinder sequences into what will be 1.0.1.

golfong woo wrote:
> Yes, there's some bugs in function ScriptCompiler::compile(const char
> *scrname) of ccscript3/compiler.cpp at line 428:
>
>         token = strrchr(buffer, '\\');
>
> Here  just cut the parts of pathname with "\\", while the parts of "/"
> is still  remain. So change it to following is ok.
>         token = strrchr(buffer, '\\');
>         token = strrchr(buffer, '/');
>
> But it seem has nothing to do with the problem of "start failed" .    
> C:\Program Files\Common Files\GNU Telephony\Runtime>server
> starting 0.3.1 on x86 w32; timeslots=32
> loading english translator
> soundcard/0: session starting
> soundcard/0: state=initial, event=100, seq=0
> soundcard/0: state=idle, event=100, seq=0
> binding ivrscript...
> compiler scanning C:\Program Files\GNU Telephony\Bayonne IVRScript1
> soundcard msgport starting
> compiled test; 4 steps
> compiled test; 4 steps
> compiler scanning C:\Program Files\GNU Telephony\Bayonne Scripts
> 2 applications compiled
> driver(s) started; 1 timeslot(s) used
> tcpmon service started
> soundcard/0: state=idle, event=105, seq=1
> soundcard/0: start failed
>
> 在 05-10-19，David Sugar<address@hidden> 写道：
>
> > We do a transform of the path because the server does a chdir.  But I
> > think I see
> > the real problem:
> >
> > compiled Bayonne IVRScript1/test; 3 steps
> >
> > That should have been "compiled test; x steps"
> >
> > This means somewhere at least part of the pathname is getting into the
> > script
> > name.  Hence, Bayonne internally has compiled this under "Bayonne
> > IVRScript1/test"
> > rather than "test", and this is probably why start fails; it is probably
> > using and
> > looking for "test".  This may even be a bug in ccscript, or a confusion
> > somewhere
> > as a result of perhaps \\ and / both being used as path separators as it
> > processes the pathname internally, and it only pealed off the path for
> > the \\ part.
> > At least that is the sort of problem I suspect and I would look for.
> >
> > golfong woo wrote:
> >
> > > When execute it as: "server.exe -vvvv test.scr"
> > > it's output:
> > > C:\Program Files\Common Files\GNU Telephony\Runtime>server -vvvv test.scr
> > > bayonne: test.scr: cannot access
> > > starting 0.3.1 on x86 w32; timeslots=32
> > > loading english translator
> > > soundcard/0: session starting
> > > soundcard/0: state=initial, event=100, seq=0
> > > soundcard/0: state=idle, event=100, seq=0
> > > binding ivrscript...
> > > compiler scanning C:\Program Files\GNU Telephony\Bayonne IVRScript1
> > > soundcard msgport starting
> > > test.scr(1): : unknown keyword
> > > compiled Bayonne IVRScript1/test; 3 steps
> > > test.scr(1): : unknown keyword
> > > compiled Bayonne IVRScript1/test; 3 steps
> > > compiler scanning C:\Program Files\GNU Telephony\Bayonne Scripts
> > > 2 applications compiled
> > > driver(s) started; 1 timeslot(s) used
> > > tcpmon service started
> > > soundcard/0: state=idle, event=105, seq=1
> > > soundcard/0: start failed
> > >
> > > It indicate "bayonne: test.scr: cannot access". When debug into
> > > function parseConfig(char **argv) in config.cpp at line 782, here is
> > > the code:
> > > if(!prefix)
> > >      {
> > >      getcwd(path, sizeof(path));
> > >      addString(path, sizeof(path), "/");
> > >      addString(path, sizeof(path), *argv);
> > >      prefix = path;
> > >    }
> > > prefix has the value as "C:\Program Files\Common Files\GNU
> > > Telephony\Runtime/test.scr",
> > > while the test.scr is in the directory of "C:\Program Files\GNU
> > > Telephony\Bayonne IVRScript1".
> > >
> > > 在 05-10-19，David Sugar<address@hidden> 写道：
> > >
> > >
> > > > Let's see on posix:
> > > >
> > > > ...
> > > > soundcard/0: state=idle, event=100, seq=0
> > > > soundcard msgport starting
> > > > binding ivrscript...
> > > > compiled ring; 4 steps
> > > > compiler scanning /usr/local/share/bayonne/ivrscript1
> > > > 1 applications compiled
> > > > driver(s) started; 1 timeslot(s) used
> > > > soundcard/0: state=idle, event=105, seq=1
> > > > soundcard/0: state=pickup, event=100, seq=1
> > > > soundcard/0: state=pickup, event=400, seq=2
> > > > soundcard/0: state=run, event=100, seq=2
> > > > HELLO
> > > > soundcard/0: state=hangup, event=100, seq=2
> > > > soundcard/0: state=hangup, event=400, seq=3
> > > > soundcard/0: state=idle, event=100, seq=3
> > > >
> > > > You had a "start failed"...
> > > >
> > > > This could be because you did not directly execute the script
> > > > (bayonne2.exe -vvvv test.scr)
> > > > and did not have an "assign" in your script.  It could also be a bug in
> > > > the way the
> > > > win32 parses and transcribes filenames for filenames passed to the
> > > > server.  I
> > > > would need to know which start method you used, whether you passed the
> > > > script name or not.
> > > >
> > > > The "start failed" is ultimately from atttachStart(), and can be found
> > > > in engine/states.cpp
> > > > at 1390.  The "attachStart" is in session.cpp, and uses several methods
> > > > to locate a
> > > > script to run if none is passed, including scripts assigned through the
> > > > assign command.
> > > > You might try "assign.timeslot 1" in your script.
> > > >
> > > > I also found the unknown keyword/3 steps compiled strange.  I did get 4
> > > > when I cut and pasted your example.
> > > >
> > > > Actually I did not intend to "officially" distribute the 1.0 release for
> > > > windows,
> > > > simply because there was some issues (such as running as NT service) left
> > > > incomplete and inadquately tested, so I did not feel that met the
> > > > criteria for
> > > > "stable".  There should be nothing broken over the previous 0.x
> > > > releases, though.
> > > > With a little work by someone that knows w32 well, I do think it should
> > > > be possible to
> > > > make the current Bayonne release acceptably stable for w32 as well.
> > > >
> > > > Yes, you needed the cape-1.3.20 because the version of ccscript3 was
> > > > upleveled
> > > > to 1.0.0, and that had a few changes that were not backward compatible.
> > > >
> > > > address@hidden wrote:
> > > >
> > > >
> > > > > hi, all
> > > > >     according to the comment, i have installed cape-1.3.20.exe, then compiled 
> > > > > bayonne2-1.0.0 with soundcard successfully. when the "space" key of keyboard is 
> > > > > down,
> > > > > bayonne cannot process this event correctly.
> > > > >
> > > > > starting 0.3.1 on x86 w32; timeslots=32
> > > > > loading english translator
> > > > > soundcard/0: session starting
> > > > > soundcard/0: state=initial, event=100, seq=0
> > > > > soundcard/0: state=idle, event=100, seq=0
> > > > > binding ivrscript...
> > > > > compiler scanning C:\Program Files\GNU Telephony\Bayonne IVRScript1
> > > > > test.scr(1): : unknown keyword
> > > > > compiled Bayonne IVRScript1/test; 3 steps
> > > > > compiler scanning C:\Program Files\GNU Telephony\Bayonne Scripts
> > > > > 1 applications compiled
> > > > > driver(s) started; 1 timeslot(s) used
> > > > > soundcard msgport starting
> > > > > tcpmon service started
> > > > > soundcard/0: state=idle, event=105, seq=1
> > > > > soundcard/0: start failed
> > > > > soundcard/0: state=idle, event=407, seq=2
> > > > > soundcard/0: state=ringing, event=100, seq=2
> > > > > soundcard/0: state=ringing, event=408, seq=3
> > > > >
> > > > > following is the test.scr
> > > > >     echo "HELLO"
> > > > >     exit
> > > > > ^1
> > > > >     echo "1"
> > > > >     exit
> > > >
> > > >
> > > > _______________________________________________
> > > > Bayonne-devel mailing list
> > > > address@hidden
> > > > http://lists.gnu.org/mailman/listinfo/bayonne-devel
> >
> >
> >
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Bayonne-devel mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/bayonne-devel

dyfet.vcf
Description: Vcard