gpi-hash.c segfaults_(re)

Wed, 23 Jul 1997 14:11:56 +1100

On 23 Jul 97 at 3:08, Peter Gerwinski wrote:
> I'd normally agree, but one of my latest changes was to zero out the
> hash table at the beginning of each GPI file ...

Yes, but what about any trees that are destroyed in the process of 
parsing the interface section (parameters and anything else that has 
temporary scopage)?  I think it might be these that are causing the 
problem.

Bill
Bill
--
Leave others their otherness.

Bill Currie (billc@blackmagic.tait.co.nz)