PATCH: C++ front end & register vars

[Richard Henderson]

On Sun, Jan 16, 2000 at 12:06:06PM -0000, Andrew Haley wrote:
> IMO the stupid allocator is more trouble than it's worth, but we
> had that argument last year.

My last patch to kill stupid sustained a 6% slowdown at -O0
for whatever test I'd selected.  Do folks still think this
is too much?  It'd be a pity to obfuscate the C++ front end 
because of it...


r~