Re: _WIN32_WINNT redefined?

[Eli Zaretskii <eliz@gnu.org>]

> Date: Wed, 02 Nov 2022 15:51:05 +0000
> From: i.nixman@autistici.org
> 
> On 2022-11-02 15:18, Eli Zaretskii wrote:
> 
> > Then the stuff in common-defs.h should be augmented to define
> > _WIN32_WINNT to the value 0x0600 or greater, if it isn't already high
> > enough, but only if the patch for gthreads is being used.  How exactly
> > to write the cpp conditional for that, I don't know, but hopefully you
> > will be able to figure that out.
> 
> no.
> just for eyes:
> 74 #if defined (__MINGW32__) || defined (__CYGWIN__)
> 75 # ifdef _WIN32_WINNT
> 76 #  if _WIN32_WINNT < 0x0501
> 77 #   undef _WIN32_WINNT
> 78 #   define _WIN32_WINNT 0x0501
> 79 #  endif
> 80 # else
> 81 #  define _WIN32_WINNT 0x0501
> 82 # endif
> 83 #endif  /* __MINGW32__ || __CYGWIN__ */
> 
> the condition on line 75 is always false, because none of the inclusions 
> above include neither `windows.h` nor `winver.h`.

That's on your system, with your MinGW w32api headers.  But that's not
the only game in town.

> it can be solved by inclusion `windows.h` after line 74.
> but I did it and faced with another error:

That's not what I suggested.  I suggested that you augment the above
cpp conditionals such that when the gthread patch is in use,
_WIN32_WINNT is defined to 0x0600, and otherwise to 0x0501, as before.

The question is: what could be a preprocessor conditional to determine
whether the gthread patch is used, so that it could be used to augment
the above.  This is something I cannot answer, but I hope you can.