Re: separating gdb & inferior output

[Bob Rossi <bob@brasko.net>]

On Wed, Aug 25, 2004 at 07:01:32AM +0300, Eli Zaretskii wrote:
> > Date: Tue, 24 Aug 2004 15:09:12 -0400
> > From: Bob Rossi <bob@brasko.net>
> > Cc: gdb@sources.redhat.com
> > 
> > > Perhaps you could state what are the problems with `tty'?
> > 
> > O, you need to actually get a new tty.
> 
> Sorry, I still don't get it: I know what the `tty' command does, but I
> don't understand what is its problem on non-Unix platforms.  I don't
> see anything dubious in the code that implements the command.
> 
> > Windows doesn't have the concept of a tty.
> 
> Of course, it does.
> 
> > File descriptor redirection is fine.
> 
> And that's exactly what win32-nat.c does, please take a look.
> 
> > Maybe we could have something like
> > 
> >    gdb --i=mi --out_fd=n
> > where n is the descriptor you plan on reading from GDB.
> 
> How is this different from what GDB already does?
> 
> > What does it mean to open a 'tty' on a windows platform, or some other
> > non-unix platform?
> 
> Again, see win32-nat.c (search for "inferior_io_terminal").
> 
> I'm afraid I'm missing something.

OK, lately I have been making assertions when I probably should be
asking questions. Shame on me.

I understand on Unix, the client can open a tty, and direct the output of the inferior to it.
On windows, how would the client do that?

Does native windows actually support opening a 'tty'? Something I didn't
know.

Finally, using cygwin windows, do I use the tty option as if I was on
UNIX?

Thanks,
Bob Rossi