[RFA/Obvious] Fix go32v2 compilation.

[RFA/Obvious] Fix go32v2 compilation.

[Pierre Muller]

  Go32v2 native target doesn't compile anymore...
since a patch by Pedro dating back 2010-08-20 :(

  Which sadly means that not many persons currently still try to use
current CVS tree for go32v2...

  I almost checked this in as obvious, but
I didn't really get if this function was necessary or not.

  Pedro, is setting this function to NULL enough to
maintain old functionality for DJGPP
or should a dos specific function be implemented?


Pierre Muller
Pascal language support maintainer for GDB

ChangeLog entry:

2010-10-12  Pierre Muller  <muller@ics.u-strasbg.fr>

	* ser-go32.c (struct dos_ops) Add missing fdopen field.

Index: ser-go32.c
===================================================================
RCS file: /cvs/src/src/gdb/ser-go32.c,v
retrieving revision 1.24
diff -u -p -r1.24 ser-go32.c
--- ser-go32.c  20 Aug 2010 18:49:20 -0000      1.24
+++ ser-go32.c  12 Oct 2010 14:54:53 -0000
@@ -843,6 +843,7 @@ static struct serial_ops dos_ops =
   0,
   dos_open,
   dos_close,
+  NULL,                                /* fdopen, not implemented */
   dos_readchar,
   dos_write,
   dos_noop,                    /* flush output */

Re: [RFA/Obvious] Fix go32v2 compilation.

[Pedro Alves]

On Tuesday 12 October 2010 16:01:10, Pierre Muller wrote:
>   Pedro, is setting this function to NULL enough to
> maintain old functionality for DJGPP
> or should a dos specific function be implemented?

Yes, it's enough.  Aparently, ser-go32.c has the only serial_ops instance
that isn't malloc+memset'ed, and I missed that.

-- 
Pedro Alves

Re: [RFA/Obvious] Fix go32v2 compilation.

[Eli Zaretskii]

> From: "Pierre Muller" <pierre.muller@ics-cnrs.unistra.fr>
> Cc: "'Eli Zaretskii'" <eliz@gnu.org>, "'Pedro Alves'" <pedro@codesourcery.com>
> Date: Tue, 12 Oct 2010 17:01:10 +0200
> 
>   Go32v2 native target doesn't compile anymore...
> since a patch by Pedro dating back 2010-08-20 :(
> 
>   Which sadly means that not many persons currently still try to use
> current CVS tree for go32v2...

That's true.  I don't have enough time to compile it except
occasionally.  Too busy developing bidirectional editing support for
Emacs.

Thanks for fixing this.