Re: Windows 64 port

[Eli Zaretskii]

> From: Fabrice Popineau <address@hidden>
> Date: Sun, 19 Feb 2012 23:05:52 +0100
> Cc: AJMR <address@hidden>, address@hidden
> 
> Here is my Windows 64 bits patch. It is quite crude and only attempts to
> make it possible to compile a 64bits emacs with the MS Sdk.

I have a few questions about the patch (below).

> the heap allocation scheme should probably be reworked

Can you tell which parts of the current scheme need to be reworked,
and why do you think so?

> === modified file 'lib/strftime.c'
> --- lib/strftime.c    2011-03-31 04:24:03 +0000
> +++ lib/strftime.c    2011-12-05 13:09:35 +0000
> @@ -36,9 +36,13 @@
>  #include <ctype.h>
>  #include <time.h>
>  
> +#ifdef _MSC_VER
> +#define tzname _tzname
> +#else
>  #if HAVE_TZNAME && !HAVE_DECL_TZNAME
>  extern char *tzname[];
>  #endif
> +#endif

This seems to be unrelated to 64-bit hosts.  Why is it needed, when
s/ms-w32.h has the same #define (which your patch #ifdef's away)?
What is the issue here?

> === modified file 'src/editfns.c'
> --- src/editfns.c     2012-01-19 07:21:25 +0000
> +++ src/editfns.c     2012-02-05 20:06:34 +0000
> @@ -82,6 +82,8 @@
>  
>  #ifdef WINDOWSNT
>  extern Lisp_Object w32_get_internal_run_time (void);
> +
> +extern struct tm *localtime (const time_t *t);
>  #endif

Why is this needed?  It seems also unrelated to 64-bit Windows.

> -internal_lisp_condition_case (volatile Lisp_Object var, Lisp_Object bodyform,
> +internal_lisp_condition_case (/* volatile */ Lisp_Object var, Lisp_Object 
> bodyform,

Does MSVC have problems with the `volatile' qualifier?

> --- src/frame.c       2012-01-19 07:21:25 +0000
> +++ src/frame.c       2012-02-05 20:06:34 +0000
> @@ -812,6 +812,7 @@
>  #ifdef HAVE_WINDOW_SYSTEM
>    if (track && FRAME_WINDOW_P (XFRAME (frame)))
>      {
> +      extern Lisp_Object x_get_focus_frame(struct frame *);

This should go into w32term.h (it is already in xterm.h and nsterm.h).

> --- src/m/amdx86-64.h 2012-01-19 07:21:25 +0000
> +++ src/m/amdx86-64.h 2012-02-05 20:06:34 +0000
> @@ -17,7 +17,8 @@
>  You should have received a copy of the GNU General Public License
>  along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.  */
>  
> -#define BITS_PER_LONG           64
> +#define BITS_PER_LONG           32

Here Windows and Posix systems differ, so it's wrong to do this
unconditionally.  Does the 64-bit Windows define _LLP64 or some such
somewhere on its headers?  If so, perhaps we could condition the above
on that macro, vs _LP64 on Posix hosts.

>  /* Define the type to use.  */
> -#define EMACS_INT               long
> -#define pI                   "l"
> -#define EMACS_UINT              unsigned long
> +#define EMACS_INT               __int64
> +#define EMACS_UINT              unsigned __int64
> +#define pI                   "ll"
> +#define VIRT_ADDR_VARIES
> +#define DATA_START   get_data_start ()

Likewise here.

Thanks.