[mythtv] Ticket #4270: MythTV on Windows (with MinGW)

[Daniel Kristjansson]

On Tue, 2007-12-11 at 11:10 -0500, Andrei Tanas wrote:
> > As I general rule I try to keep these types of defines limited to
> > only where they are needed so that ccache can do a better job on
> > recompiles after you change a compile time option. But in this
> > case it isn't really relevant since you won't enable/disable MINGW
> > between compiles on the same machine. But I'm thinking forward to
> > a MinGW less Windows port at some time in the future.
> 
> Well, compat.h is needed almost everywhere for the MinGW build, and we check
> for USING_MINGW there. It is still restricted to mingw {} sections in the
> project files, I think it's cleaner to define it in one place.

You've convinced me, go ahead and make a patch.

> MinGW-less - as in Visual C++ build? I'd love to do that, I would very much
> prefer to be able to use Visual Studio debugger, however I considered it
> briefly, it just seems to be too much work. A lot of code uses posix API,
> which is not exactly there.

Yeah, someone tried a Visual C++ port a few months ago and gave up.
But if we get it running well under MinGW, a Visual C++ port might
be easier; especially after MythTV switches to Qt4.

I could tell reviewing this that you've done a lot of work to get this
going, but I'm sure you noticed there have been a few earlier attempts,
and you benefited from the earlier OS X port as well.

-- Daniel