partysip-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Partysip-dev] compiling partysip


From: Aymeric Moizard
Subject: Re: [Partysip-dev] compiling partysip
Date: Thu, 16 May 2002 10:52:42 +0200 (CEST)

Let me give you more informations:

The behavior that I'm expecting is:
 If no remote socket is found, the SECOND sendto call
 returns ECONNREFUSED.

On linux, I have the following behavior:
 connect + sendto  <- stupid, but works fine
 connect + send    <- better, and works fine
 sendto only       <- never returns ECONNREFUSED.

As you said below, the rfc2543 recommands to use connect + send, but
the latest draft appears to bypass this information.

My last proposal is to keep the old behavior (replacing the sendto by
send, but keeping the connect call).

Just to make sure? when all of you get this error, are you sure you are
sending the message to a remote IP/port where a socket is listening???

Can you try this code: (with send instead of sendto)

  connect(sock,(struct sockaddr *) &addr,sizeof(addr));
  if (0  > send (sock, (const void*) buf, strlen (buf), 0))

Let me know if you still get a network error or a wrong behavior
even if a remote application is listening.

Thanks
Aymeric

> Doesn't RFC 2543 (which is deprecated..?) recommand to use connect() for UDP 
> as well?
> Though I do not find that recommandation anymore in the new bis-09 draft, I 
> believe
> using send() function should be the solution...
> Any comment?
>
> Regards,
> Tim Na
>
> > >
> > >   But I write the simple UDP client which send "HELLO
> > > MSG" to port 5060 and it is working without any
> > > problem.
> > >
> > > Can u guys help in this.
> > >
> > > Jayesh
> > >
> > > --- Aymeric Moizard <address@hidden> wrote:
> > > >
> > > >
> > > > On Wed, 15 May 2002, Ralf Mahlo wrote:
> > > >
> > > > > Hi!
> > > > > I'm working with Solaris 2.6 . First when I tried
> > > > to compile I've got the error: in pplsocket.c ,
> > > > > function ppl_gethostbyname : too much arguments
> > > > for gethostbyname_r (in Solaris there are only 5
> > > > > arguments, don't know about other systems...).
> > > > However, I changed result_buffer with result and
> > > > > deleted the old (5.) '&result'. (maybe that's
> > > > wrong...)
> > > >
> > > > ... yes, this is probably wrong.
> > > > gethostbyname is a reentrant method: this is why I
> > > > assume 'result' should
> > > > be a presiously allocated element and not a pointer
> > > > to a non allocated
> > > > 'struct hostent'
> > > >
> > > > Could you try fixing it this way: (
> > > >
> > > > > PPL_DECLARE(int) ppl_gethostbyname(struct
> > > > sockaddr_in *sin, char *hostname, int port)
> > > > > {
> > > > > //  struct hostent result_buffer;
> > > > >   char tmp[GETHOSTBYNAME_BUFLEN];
> > > > >   struct hostent *result;    <<- replace with
> > > > 'struct hostent result'
> > > > >   int my_error;
> > > > >
> > > > >   gethostbyname_r(hostname,       /* the FQDN */
> > > > >     result,  /* the result buffer */     <<-
> > > > replace with '&result'
> > > > >     tmp,
> > > > >     GETHOSTBYNAME_BUFLEN - 1,
> > > > > //    &result,
> > > > >     &my_error);
> > > > >
> > > > >   if (!result)
> > > > >     return my_error;
> > > > > }
> > > > >
> > > > > After that I can compile the program completely,
> > > > but get the warning : in psp_plugin.c , function
> > > > > psp_plugin_load : warning: ANSI forbids assignment
> > > > between function pointer and `void *' for the
> > > > > lines 69, 77 and 85 ...
> > > >
> > > > This is also true on linux... partysip is not ANSI,
> > > > but nearly... If
> > > > somebody know THE proper way to get rid of this
> > > > warning let me know...
> > > >
> > > > > I can run the program, but while loading the
> > > > authentication plugin, I get a bus error...
> > > >
> > > > as you (the method gethostbyname) are wrinting in a
> > > > non-allocated buffer,
> > > > this seems normal to get something like a bus
> > > > error...
> > > > But of course, it may come from somewhere else...
> > > > Let me know if the
> > > > change I propose make it work.
> > > >
> > > > > Another question to the partysip.conf file: is it
> > > > necessary to change the content of the file with
> > > > > my own data or does it work as well with the
> > > > current content?? (okay, when I want to use it, it
> > > > must
> > > > > be changed, but for now (only for testing, if it
> > > > works) ?)
> > > >
> > > > The included file 'partysip.conf' is by default
> > > > given with authentication
> > > > on. But of course, the database of users does not
> > > > contains your user. So
> > > > any request is expected to get a 401 or 407
> > > > response. If you change that
> > > > line (authentication = off), things should work.
> > > >
> > > > Aymeric
> > > >
> > > > > Regards, Ralf
> > > > >
> > > > >
> > > > > _______________________________________________
> > > > > Partysip-dev mailing list
> > > > > address@hidden
> > > > >
> > > >
> > > http://mail.freesoftware.fsf.org/mailman/listinfo/partysip-dev
> > > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > Partysip-dev mailing list
> > > > address@hidden
> > > >
> > > http://mail.freesoftware.fsf.org/mailman/listinfo/partysip-dev
> > >
> > > __________________________________________________
> > > Do You Yahoo!?
> > > LAUNCH - Your Yahoo! Music Experience
> > > http://launch.yahoo.com
> > >
> > > _______________________________________________
> > > Partysip-dev mailing list
> > > address@hidden
> > > http://mail.freesoftware.fsf.org/mailman/listinfo/partysip-dev
> >
> >
> > _______________________________________________
> > Partysip-dev mailing list
> > address@hidden
> > http://mail.freesoftware.fsf.org/mailman/listinfo/partysip-dev
> >
>
> _______________________________________________
> Partysip-dev mailing list
> address@hidden
> http://mail.freesoftware.fsf.org/mailman/listinfo/partysip-dev
>




reply via email to

[Prev in Thread] Current Thread [Next in Thread]