I don't really have much of a stake in this, but it seems to be that it's probably a good idea to question the assumption that a single codebase shared between Linux and Windows is really a good idea. It's definitely appealing to be able to reuse the code that has been written for Linux, but in the long term trying to keep a single C codebase for such different OSes is probably way more trouble than it's worth.
The real problem is that no one seems to use C for native Windows code, let alone all the POSIX library environment stuff. If you want to write cross-platform stuff it probably makes more sense to use a language that is really supported on both platforms -- Java, C#, Python or even C++ would probably be better choices. Of course this would require a lot of work rewriting things in a new language, but if you're really serious about supporting Windows as a first-class platform, I don't think trying to get Linux C code working with a bunch of #ifdef-ery is ever going to get you there. - R. _______________________________________________ general mailing list general@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
