[Zlib-devel] zlib 1.2.0.1 available for testing

[Dmitriy Anisimkov]

Cosmin Truta wrote:

>So, Mark, you don't have to apply Christian's patches separately.
>They are already included in my update (the newer makefiles, and the
>WIN32_LEAN_AND_MEAN thing).
>
WIN32_LEAN_AND_MEAN is not used at all at zlib-1.2.0.1-cos2, only defined.

>Here is the zconf.h update, but you can find the entire archive update
>(the reversion back to the old *ZEXPORT prototypes, Christian's updates,
>the deflate.c fix), at
>  http://www.cs.toronto.edu/~cosmin/pngtech/z_rle/zlib-1.2.0.1-cos2.zip
>
>I tested the entire new update using Borland C, Microsoft C and MinGW
>under Win32.
>
Cosmin,

Is it possible to support CDECL calling convention for build zlib under 
Win32 ?

I see too much *WINDOWS* and *WIN32* words in the zcong.h. I feel that 
it is too much attension to MS.
And it is hard to understand for me, what should I propose to make zlib 
CDECL in Windows, i mazed between
*WIN32*  and *WINDOWS* like defines ;-).
I would prefer to do not have it at all.
The CDECL is a default calling convention for C, and I think it is 
strange, that we have to use so complex logic to
do simple think, to keep all routines in the default calling convension.

Maybe we should have configuration parameter for define the calling 
convention for Win32 ?