This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [patch] fold-const.c: Don't transform X & C into (X >> C') &1 in fold_binary.

From: Paolo Bonzini <bonzini at gnu dot org>
To: gcc-patches at gcc dot gnu dot org
Date: Thu, 21 Apr 2005 17:38:53 +0200
Subject: Re: [patch] fold-const.c: Don't transform X & C into (X >> C') &1 in fold_binary.
References: <20050421.105441.20504321.kazu@cs.umass.edu> <Pine.LNX.4.44.0504210921060.15848-100000@www.eyesopen.com>

My personal preference would be to leave this code in do_jump,
even if it never triggers.  Some day it might.  Indeed, the only
reason we can safely change the tree-level canonical form is that
the reverse-sense code was left in expand_expr.  If this code had
been removed as "dead", your recent changes would have been much
harder.

I think that CVS is what this is for... I am against leaving archaeology in the source base. But I do not want to start yet another philosophic discussion! :-)

Paolo

References:
- Re: [patch] fold-const.c: Don't transform X & C into (X >> C') & 1in fold_binary.
  - From: Kazu Hirata
- Re: [patch] fold-const.c: Don't transform X & C into (X >> C') & 1in fold_binary.
  - From: Roger Sayle

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]