This is the mail archive of the
mailing list for the GCC project.
Re: Uninitialized memory access in final.c
- To: Richard Henderson <rth at cygnus dot com>
- Subject: Re: Uninitialized memory access in final.c
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Fri, 22 Sep 2000 18:56:03 -0600
- cc: Bernd Schmidt <bernds at redhat dot co dot uk>, gcc-patches at gcc dot gnu dot org
- Reply-To: law at cygnus dot com
In message <20000919091025.B26169@cygnus.com>you write:
> On Tue, Sep 19, 2000 at 03:45:03PM +0100, Bernd Schmidt wrote:
> > Is it ever valid for an insn's length to be increased during
> > shorten_branches? It seems to me that this would cause the
> > algorithm to be no longer guaranteed to terminate.
> I thought we started with smallest lengths and always expanded.
> Otherwise you wouldn't find the minimal solution.
Nyet. We start pessimistic and improve, at least last time I looked
at that code. Someone (I don't remember who) suggested we change it
to start optimistic and lengthen as needed to find the best solution.