I'm not sure I understand.. The page at http://dlang.org/windows.html says

  Instead of the:
  #include <windows.h>
  of C, in D there is:
  import std.c.windows.windows;

However, the samples at https://github.com/AndrejMitrovic/DWinProgramming/tree/master/Samples use

  import win32.windef;
  import win32.winuser;

My question is which one should I use in my programs?

-- 

  D

On 1/10/2012 10:24 PM, DNewbie wrote:
> I'm not sure I understand.. The page at http://dlang.org/windows.html says
>
>    Instead of the:
>    #include<windows.h>
>    of C, in D there is:
>    import std.c.windows.windows;
>
> However, the samples at https://github.com/AndrejMitrovic/DWinProgramming/tree/master/Samples
> use
>
>    import win32.windef;
>    import win32.winuser;
>
> My question is which one should I use in my programs?
>

Those samples use a binding to the Win32 API[1] that does not ship with DMD. So in your case, std.c.windows.windows is probably what you want for now.


[1]http://dsource.org/projects/bindings/wiki/WindowsApi

On Tue, Jan 10, 2012, at 10:37 PM, Mike Parker wrote:
> 
> Those samples use a binding to the Win32 API[1] that does not ship with DMD. So in your case, std.c.windows.windows is probably what you want for now.
> 
> 
> [1]http://dsource.org/projects/bindings/wiki/WindowsApi
> 

Ok.. I still don't understand why such a 'binding' is needed. Is std.c.windows.windows not enough for everyone?

-- 

  D

std.c.windows.windows is missing *a lot* of definitions. It also doesn't provide aliases to ASCII/UTF16 functions like WindowsAPI does via the Unicode flag, so you have to explicitly use e.g. MessageBoxA/MessageBoxW instead of MessageBox. WindowsAPI is nicely modularized, and is based on existing MinGW headers.

On 1/10/2012 11:44 PM, Andrej Mitrovic wrote:
> std.c.windows.windows is missing *a lot* of definitions. It also
> doesn't provide aliases to ASCII/UTF16 functions like WindowsAPI does
> via the Unicode flag, so you have to explicitly use e.g.
> MessageBoxA/MessageBoxW instead of MessageBox. WindowsAPI is nicely
> modularized, and is based on existing MinGW headers.

Yes, but it's not necessarily the best place to start for someone who hasn't figured out the toolchain yet.

On 1/10/2012 10:57 PM, DNewbie wrote:
> On Tue, Jan 10, 2012, at 10:37 PM, Mike Parker wrote:
>>
>> Those samples use a binding to the Win32 API[1] that does not ship with
>> DMD. So in your case, std.c.windows.windows is probably what you want
>> for now.
>>
>>
>> [1]http://dsource.org/projects/bindings/wiki/WindowsApi
>>
>
> Ok.. I still don't understand why such a 'binding' is needed.
> Is std.c.windows.windows not enough for everyone?
>

Unfortunately, no. It is not a complete binding of the Win32 API. It has a lot of stuff, but if you need more you need to look elsewhere. Andrej hit on some of the deficiencies in his post.

On Wed, Jan 11, 2012, at 01:16 PM, Mike Parker wrote:
> On 1/10/2012 10:57 PM, DNewbie wrote:
> > On Tue, Jan 10, 2012, at 10:37 PM, Mike Parker wrote:
> >>
> >> Those samples use a binding to the Win32 API[1] that does not ship with DMD. So in your case, std.c.windows.windows is probably what you want for now.
> >>
> >>
> >> [1]http://dsource.org/projects/bindings/wiki/WindowsApi
> >>
> >
> > Ok.. I still don't understand why such a 'binding' is needed. Is std.c.windows.windows not enough for everyone?
> >
> 
> Unfortunately, no. It is not a complete binding of the Win32 API. It has a lot of stuff, but if you need more you need to look elsewhere. Andrej hit on some of the deficiencies in his post.
> 

Understood. I think I'll go with the binding. Thnk you.