[Gdal-dev] Win32: Supported build environments

[Mateusz Loskot]

Frank Warmerdam wrote:
> Folks,
> 
> As I'm sure many have noticed in the past, I've been cool to actively
> supporting MingW builds of GDAL.  Some people might remember that I was
> the same way about Cygwin builds for three or four years, till I finally
> accepted it should be a supported build environment.

Frank,

So, we're going to officially support both: MinGW and Cygwin.

> However, there is clearly a significant interest in MingW builds.  So I'd
> like to open a discussion about what win32 build environments we are
> willing to expect maintainer time on supporting.
> Environments not in this list can
> still be supported by contributors, but we aren't going to commit
> significant maintainer resources to them,
> and I won't feel personally bad if they rot. :-)
> 
> Currently supported:
>  o Visual Studio 6 (via nmakefile)
>  o Visual Studio 2003 (aka VC7.1) (via nmakefile)

I think we also (implicitly) support Visual Studio .NET 2002 with
the Visual C++ 7.0 compiler.

>  o Visual Studio 2005 (aka VC8) (via nmakefile)
>  o Cygwin (via unix build mechanism)
> 
> 
> Some potential candidates that come to mind:
>  o Visual Studio Express Edition (the free download)
>  o Visual Studio via project files (beyond the simple ones
>    we have now that just run the nmakefiles)
>  o MINGW (using the unix build scripts and makefiles)
>  o Borland C

Do we have any version of Borland specified?
BTW, the new version is called Turbo C/C++:

http://www.turboexplorer.com/

>  o WATCOM C (or whatever the free version is called)

Open Watcom C/C++ and it's Open Source project available here:

http://www.openwatcom.org

> My currently thinking is that we should add MINGW, and VS Express
> Edition support,

The Express Edition can be supported similarly to VC++ 7.x/8

> and possibly drop Visual Studio 6 support.  After a bit
> of discussion, my intent would be to make a motion declaring particular
> compiler/build environments supported for win32, and thereafter we would
> make a concerted effort to test them before releases, and fix issues that
> come up.

I'm all for that.

Cheers
-- 
Mateusz Loskot
http://mateusz.loskot.net