[MPlayer-dev-eng] latest mingw network patch

[Joey Parrish]

On Thu, Jun 05, 2003 at 03:56:36PM -0400, D Richard Felker III wrote:
> > Once the mingw port is complete, then cygwin should be deprecated.
> > Then cygwin-specific support itself can be removed and -mno-cygwin
> > (IIRC) can be added to Makefile for cygwin systems to make it mingw
> > compatible.  Or is it -mmingw?  I forget exactly, but there is a gcc
> > option or two on cygwin to emulate mingw.  Anyhow, I think this should
> > be our eventual goal for windows systems.
> 
> IMO this is a bad idea. No reason to remove cygwin support if it
> works. Some people already use cygwin regularly under windows, and
> want to built a program that will operate properly in their cygwin
> environment. So as long as it's no extra work to support cygwin
> (shouldn't be), don't arbitrarily remove support, and PLEASE don't do
> like the SDL fools and make it force -mno-cygwin on cygwin (which kept
> people from compiling posix-dependent progs against SDL, e.g.
> mplayer).

This should also answer (hopefully) Diego's mail.

What I'm saying is that since Cygwin is an emulation environment.  If we
have a working cygwin port, then the cygwin DLL is calling certain
native windows functions.  If we port to mingw, then we are calling all
these things ourselves directly.  This leads to greater efficiency IMHO,
and also more portable code in general by abstracting the
operating-system specific things.

Since the cygwin environment can compile code that is independant of the
cygwin DLL by using the mingw runtime, there's no reason to use anything
else.  Native windows binaries run fine in cygwin, too.  :)

As for being like SDL, the problem with SDL lies in the fact that it is
a library that tries to force -mno-cygwin on it's apps.  That is not at
all the case for mplayer, so I believe that this will not be a problem.

Anyone have other thoughts?  Did I miss anything?

--Joey