Re: [jay_finger@hotmail.com: Two problems in Emacs-21.2.91 on Windows]

[Juanma Barranquero <lektu@terra.es>]

On Wed, 23 Oct 2002 19:41:33 +0300, "Eli Zaretskii" <eliz@is.elta.co.il> wrote:

> msdos.h will do, I think (if it is #include'd by xfaces.c).

Ok.

> But I don't really understand why should the definition in xfaces.c be
> removed.  xfaces.c is mostly platform-independent, so it would make
> sense to actually leave that definition, perhaps with some more
> #ifdef's if Windows and Mac need that, and remoev the other bunch.  Am
> I missing something?

Well, on one hand I don't see the point of having XColor defined in
xfaces.c *and* other .c files (like w32term.c, for example). That's what
include files are for, aren't? But if I put XColor in a w32*.h, I had to
protect it from being redefined in xfaces.c, and that's just weird.

XColor is needed in some platforms (like Mac OS 9, Windows and DOS) and
not in others (like X or Mac OS X), so either you put it in a include
used by every build and guard it through #ifdef's, or put it in
platform-specific includes. That's what I chose because it seems
cleaner/simpler.

                                                           /L/e/k/t/u