🚀 go-pugleaf

I found a bug in GCC 2.2.2, PL2.  When I try to compile the following
program

        #include <float.h>

        main()
        {
            float x;

            x = -DBL_MAX;
        }
        
the compiler aborts with an `address error' in cc1.  Obviosly DBL_MAX cannot
be stored in a float, but I think it is not ok for the compiler to crash
in such a case.

I run GCC on an Atari ST.  The version I use is said to be compiled with
the MiNT libs (I don't know whether this makes a difference).

Jochen.

---
You can mail me at `kring@physik.uni-kl.de'.
Please use `c/o Jochen' in the Subject.