Mail Archives: cygwin-developers/2002/09/05/21:10:01
> On Thu, Sep 05, 2002 at 08:25:44AM -0400, Earnie Boyd wrote:
> >Well, yes, but for w32api that would not be logical. At least this
> >thread has pointed some doco weaknesses that I need to address. Yes,
> >if you want to use the newer API functions then you must be aware to
> >set WINVER to the version you wish to support. MS must do this for
> >each generation of new compilers to support newer OS's. If give the
> >developer the freedom to not think about backward compatibility, and as
> >evidenced there is little. We are smarter than that, so we choose to
> >let the developer set the version h[im|er]self, I.E.: the developer is
> >forced to think about the new functionality w.r.t. backward
> >compatibility of older OS versions.
>
> Is this decision documented somewhere? It doesn't make sense to me
> that w32api would be like Microsoft in every other way but this one.
> And, somehow, I thought that as one of the main users of w32api, we
> might get a vote in this.
>
> Why throw a barrier in the way of someone who thinks that they've
> found an alternative to MSVC? You're guaranteeing that someone will
> have to find the mingw mailing list and ask questions because things
> are working differently than they expect.
>
I gotta agree here. All this is going to do is result in endless "Why can't
mingw find APIOnlyInNTSeries()? My program compiles on VC7 just fine!"'s. That
sounds to me to be exactly not what the whole idea of mingw is about; isn't the
ultimate mingw ideal to be a (vastly superior) "drop-in" for MSVC++? Personally
I sure hope so!
--
Gary R. Van Sickle
Brewer. Patriot.
- Raw text -