I have a bunch of patches that I'd like to submit to enable XEmacs
to be
compiled by a native MinGW gcc (rather than Cygwin's cross-compiler gcc
-mno-cygwin). A MinGW installation gcc (
www.mingw.org) requires different
include paths than gcc -mno-cygwin. There are also some #defines in the
wrong places and some typedefs that need to be altered.
Comparing the preprocessor directives defined by MinGW's Windows native
gcc, and the ones defined by Cygwin's gcc -mno-cygwin, the only difference
is that __MINGW__ is merely defined -D__MINGW__ by cygwin, whereas in
mingw gcc, it's defined -D__MINGW__=1.0.
Would it be a bad thing to add in a new #define to XEmacs that would
discriminate between these two compilers, and then modify the places in
the code (and configure script) that care about the difference to get the
differences right? It would enable me to get rid of a whole load of
patches I apply to each new XEmacs release to make it compile correctly
under MinGW.
Sounds fine. The hard part is coming up with a name everyone agrees on.
andy