Re: Dumb identifier name question

[Matthias Clasen <maclas gmx de>]

On Mon, 2003-01-27 at 23:57, Rachel Hestilow wrote:
> As a related issue, what would you think of moving all function argument
> names as declared in .h to their own namespace? I ask this because I'm
> getting tired of seeing warning spew when I turn on -Wshadow (mostly
> because of the libc "index" function).
> 
> Pros:
>  1) Could be done with a perl script
>  2) Won't incur any code changes, as it would it only be changing the
> header prototypes, not the actual implementations.
>  3) Gtk-doc, et al, should still work, since gtk-doc statements are in
> sources, not headers.
> 
> Cons:
>  1) Would cause unnecessary header churn, which could be adverse for CVS
> history, etc.
>  2) Might break programs which read in the Gtk+ headers and do clever
> things with them, such as bindings that don't use .defs. The only such
> program I am aware of is the Gtk# binding generator, and I am a
> contributor to that.
>  3) As you said, would break consistency with the old style.
> 
> -- Rachel

Pro 3) is wrong, since unfortunately, gtk-doc falls under the Cons 2)
category...

Matthias