operand scan speedups for PR 18587

[Andrew MacLeod]

On Fri, 2004-11-26 at 01:48, Jeffrey A Law wrote:
> On Thu, 2004-11-25 at 14:28 -0500, Andrew MacLeod wrote:

> > 
> > PS. The var-ann patch is responsible for the lion share of the speedup
> > to cplusplus-grammer.ii, and the call clobber patch responsible for most
> > of the rest of the gains.
> FWIW, I was pondering using hash tables to build up the operand lists.  
> Did you happen to look at that solution, or in your opinion is that
> overkill?
> 

no, I didn't look at hash tables.  Pinski was talking about hash tables
too, but never clarified what he meant by it.

Perhaps I'm dense these days, but what exactly do you want to do with
hash tables? 

Andrew