Re: [pdf-devel] Compiling for windows

[jemarch]

    > A possible alternative would be to manage the situation using a more
    > fine-grained approach: to test capabilities for single functions (such
    > as the version of the `open' function existing in the system) using
    > autoconf.
    > 
    > What do you (all) think about this?

   The well-known way to do it is with preprocessor conditionals and
   platform macros (e.g. __WIN32__). It leads to ugly code in certain
   situations but they're not that common.

Yes. It is a common practice and it works.

   The autoconf approach is not bad. But it's capability based and
   we're gonna probably need much more preprocessor conditional checks
   (since we'll also need conditional checks). Correct me If I'm
   wrong.

Well, I would say that we should always try to write generic code
(able to work without changes in all the supported plattform) if
possible.