[geos-devel] Win32 porting

[Frank Warmerdam]

On 11/7/05, strk at refractions.net <strk at refractions.net> wrote:
> > Is the VisualStudio directory being maintained?  If not, I would
> > like permission to remove it.  It confused me, so I imagine
> > it will confuse others.
>
> Since nobody answered I think you can safely remove it.

Strk,

OK, I think I have removed this subtree from CVS successfully.

> > I will note that win32 builds are complicated by the need to manually
> > add version information in version.h and geos_c.h.  Will these
> > files be distributed with version info embedded in release snapshots?
>
> Currently not, they are created by ./configure.

OK, well for the time being I think win32 builds will remain partially
manual since these have to be processed by hand.

> What I'd really like to do is provide a narrower API for C++
> as well. I hate being unable to touch anything w/out breaking
> the *huge* API. We might move Doxygen documentation, DLL
> exports and all to geos.h and completely avoiding to install
> other internal headers. How many users will be influenced by
> such a change (I can only think of FME) ?

Is there any particular reason not to just have most client applications
use the C API instead of adding another restricted C++ API?  The
position I took with GDAL was that the C API is ABI stable over time
but the C++ API is not guaranteed to be so.  Applications wanting to
stay at arms length are encouraged to use the C API.

Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent