On Saturday 09 August 2008 19:54:20, Eli Zaretskii wrote:
> > From: Pedro Alves <pedro@codesourcery.com>
> > Date: Sat, 9 Aug 2008 18:53:44 +0100
> >
> >   GCC complains about attribute packed like so:
> >
> >  cc1.exe: warnings being treated as errors
> >  ../../gdb/go32-nat.c:1292: warning: 'packed' attribute ignored for field
> > of type 'unsigned char' ../../gdb/go32-nat.c:1301: warning: 'packed'
> > attribute ignored for field of type 'unsigned char'
>
> This one is the only one that bother me a little, since these
> structures need to match the memory layouts expected by the system
> calls to which we pass them.  Does the warning mean that GCC will
> never add any padding between the previous field and the `unsigned
> char' field?  If so, the change is safe; if not, we need to find some
> way of enforcing the no-padding layouts.
>

On Saturday 09 August 2008 20:16:09, Mark Kettenis wrote:
> Instead of
>
> struct foo {
>   char c __attribute__((packed));
>   int d __attribute__((packed));
> };
>
> you can (should?) write
>
> struct foo {
>   char c;
>   int d;
> } __attribute__((packed));
>
> I think that'll get rid of the warning.

I considered changing it that way, but resisted because I 
assumed there was some obscure reason this wasn't done this way,
that I was just failing to see.  I've always used the latter form.

Indeed, from GDB's manual:

"Specifying this attribute for struct and union types is equivalent
to specifying the packed attribute on each of the structure or
union members."

> The go32-nat.c changes are fine with me.

Thanks, I've checked in the go32-nat.c changes, as attached.

I'll post the rest of the changes in another thread.

-- 
Pedro Alves