Re: Windows 64 port

[AJMR <ajmr@ilovetortilladepatatas.com>]

Hi everyone,

First of all, my gratitude for all the replies to the thread.

Secondly, it is not that I got scared (although, to be honest I feel
a bit overwhelmed for the depth of some details in this first contact).
However, most of time I can make nowadays is on weekends.

Thirdly, I have been reading any developers guides / instructions that
I could find and diving into the code. However, I am not sure whether
this is the correct approach. Any comments related to this topic will
be extra welcomed.

Finally, mainly to Eli and Fabrice, would it be OK if I contacted you
(via direct e-mail) in order to plan a more task-oriented roadmap? 

======== This is probably quite off topic ======
Dear Dr. Richard Stallman,

First of all, many, many, many thanks for all your contributions
already made (and hopefully more to come). I tip my hat to you.

I also appreciate your concern about porting myself completely to
GNU/Linux and even becoming a 100% free software user. However, not
being free is a sacrifice I have chosen to make. Temporary. The target
is to be once again really free.

Meanwhile, I would love to add something to the GNU and free
software world. Of course, I would love even more to contribute with an
extremely useful application which would improve people's life.
Nevertheless, I thought about going step by step and join into an
already existing project. I even thought that a port would be well
defined and narrowly scoped so I could gain momentum.

Of course, I may be wrong.

With all that in mind, I thought about helping with the free software I
have used the most. I guessed that it would not be a piece of cake, but
nothing is ever easy.
================================================


Thanks,
Alfredo.

On Mon, 20 Feb 2012 22:47:23 +0200
Eli Zaretskii <eliz@gnu.org> wrote:

> > From: Fabrice Popineau <fabrice.popineau@supelec.fr>
> > Date: Sun, 19 Feb 2012 23:05:52 +0100
> > Cc: AJMR <ajmr@ilovetortilladepatatas.com>, emacs-devel@gnu.org
> > 
> > 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.
>