Re: a review of the merge (Re: Emacs.app merged)

[Jason Rumney <jasonr@gnu.org>]

Dan Nicolaescu wrote:

> Index: src/keyboard.h
> ===================================================================
> RCS file: /sources/emacs/emacs/src/keyboard.h,v
> retrieving revision 1.85
> diff -a -u -r1.85 keyboard.h
> --- src/keyboard.h	8 Jun 2008 08:59:47 -0000	1.85
> +++ src/keyboard.h	15 Jul 2008 17:01:00 -0000

> +#if defined (HAVE_NS) || defined (HAVE_NTGUI)
> +
> +typedef void * XtPointer;
> +typedef unsigned char Boolean;
> 
> This looks strange, please get it approved by one of the Windows
> maintainers as it affects that platform.

I think it is OK, those definitions were formerly in w32gui.h when it
was needed on windows only.

> Index: src/w32gui.h
> ===================================================================
> RCS file: /sources/emacs/emacs/src/w32gui.h,v
> retrieving revision 1.34
> diff -a -u -r1.34 w32gui.h
> --- src/w32gui.h	26 Jun 2008 10:48:28 -0000	1.34
> +++ src/w32gui.h	15 Jul 2008 17:03:07 -0000
> @@ -21,5 +21,5 @@
>  #define EMACS_W32GUI_H
>  #include <windows.h>
>  
> +#include "w32bdf.h"
>  
> -/* Emulate widget_value from ../lwlib/lwlib.h, modified for Windows.  */
> 
> This looks very suspicious, why touch the Windows code?

It looks like the merge encountered conflicts here, with the code that
was moved to keyboard.h (above) being in the vicinity of some code that
was removed on 26 June when I tidied up the font backend code. The
conflict was resolved in the wrong way, reintroducing the old code.
As long as that only happened here (and you seem to be checking the
patch thoroughly, so hopefully any further occurences will be caught),
it should be simple to fix this.