kernel32/tests: remove win9x hacks (try 2)

[Saulius Krasuckas]

* On Thu, 24 Feb 2011, Alexandre Julliard wrote:
> 
> The win9x support makes the tests less strict, by allowing additional 
> behaviors, and that only when running on Windows.

Is that a problem?

> Running them on Wine is pointless since these code paths are never 
> executed.

I may be missing the point, but Wine uses broken() function already.  
Similar function, eg. dead_9x(), could be easily created to check for such 
cases.

> > And does this mean that Windows 3.1 (and DOS?) tests aren't welcome 
> > now?
> 
> As part of the winetest infrastructure, they are not welcome, and never 
> have been.

I have at least one Win16 app which behaves differently on Win16, on XP 
and on Wine [*].
What's the right way to get it running OK on Wine then (without 
appropriate Win16API tests) ?

S.

[*] http://bugs.winehq.org/show_bug.cgi?id=5278#c27