[hlink] Skip tests if hlink couldn't be created

[Dan Kegel]

On Jan 14, 2008 1:50 PM, Robert Shearman <rob at codeweavers.com> wrote:
> Dan Kegel wrote:
> > This makes the test slightly more robust to operator error.
>
> If we're striving to get zero test failures then converting a crash to
> some failures isn't really much of an improvement.

The crash was in test code, and only happened for
me because I had done something stupid
(probably "wineserver -k" before wineprefixcreate's
server had flushed its registry cache).

Since that's the only situation I can think of this
crash would happen in, it's kind of a silly patch
all around, but it does have the virtue of trying
to give the confused developer a clue
(as in "you did something to screw up wineprefixcreate").
- Dan