instantiate4.C testcase succeeds on subsequent make check runs

[Mike Stump]

> To: oliva@dcc.unicamp.br (Alexandre Oliva), egcs-bugs@cygnus.com
> From: Jason Merrill <jason@cygnus.com>
> Date: 01 Dec 1998 20:56:57 -0800

> It looks like it does; the problem is that the test harness treats the
> informational messages from collect as errors.

Sounds like a prune_system_crud type fix may be the best way to fix
this type of problem, though we all hate that.  The other way, is to
only give informational messages when -v is given, and to not give -v
in the testcase.  gcc and g++ do quite a bit of work, but usually they
never give informational messages, maybe collect2 should do likewise.