[olug] OT: Seeking Information on Borland C++ CLI for M$

William E. Kempf wekempf at cox.net
Wed Nov 20 23:52:00 UTC 2002


Don Kauffman said:
> I've tried Cygwin but found that it doesn't compile the M$ headers. (M$
> is not necessarily ANSI compliant.) I know, I know,  why am  I wanting
> to do this when I have Linux as an option? Because the online  course
> I'm taking uses them

If you already have Cygwin, you should be able to do this still with out
mingw.  The Cygwin compiler has a mingw mode.  The only minor trick is
that it will have to use some special header files for Win32, i.e. the
w32api package.  I don't know off the top of my head if this is included
with Cygwin, but if not you can d/l just that from the mingw site instead
of the entire package.

> I'm not a big fan of M$ but I will use it. I will try the mingw compiler
>  first. But I would still like to get Borland to working.

Which Borland compiler?  The free one?  I use that one for my Boost
development (http://www.boost.com) and have never encountered the problem
you describe.

In any event, if you're coding Windows specific stuff (as evidenced by
your using the MS headers), the MS compiler is your best bet.  It
optimizes better than most of the competitors, is sure to be 100%
compatible with all the headers, is link compatible with most Windows C++
libraries, etc.  The only issue with it is standards conformance, but in
this area it does at least as good, if not better than, the free Borland
compiler any way (as in they both suck in specific areas, but the Borland
one sucks in areas you're more likely to encounter since most of MS's
issues are in template areas, while Borland's are in plenty of other areas
as well).  And the next release of the compiler should be one of the more
standards conformant compilers available.  Don't avoid it just because
it's a MS product, especially if your goal is to write Windows programs.

William E. Kempf






More information about the OLUG mailing list