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] Fix PR26898, 2nd try

From: Richard Guenther <rguenther at suse dot de>
To: "Joseph S. Myers" <joseph at codesourcery dot com>
Cc: Richard Guenther <richard dot guenther at gmail dot com>, gcc-patches at gcc dot gnu dot org
Date: Fri, 31 Mar 2006 14:31:43 +0200 (CEST)
Subject: Re: [PATCH] Fix PR26898, 2nd try
References: <Pine.LNX.4.64.0603310959540.6703@t148.fhfr.qr> <Pine.LNX.4.64.0603311204090.27869@digraph.polyomino.org.uk> <84fc9c000603310414u25e79097pb23a079029b7dfb8@mail.gmail.com> <Pine.LNX.4.64.0603311226220.27869@digraph.polyomino.org.uk>

On Fri, 31 Mar 2006, Joseph S. Myers wrote:

> On Fri, 31 Mar 2006, Richard Guenther wrote:
> 
> > On 3/31/06, Joseph S. Myers <joseph@codesourcery.com> wrote:
> > > I take it that at this point in folding it is guaranteed that the
> > > constants are non-negative, i.e. you don't have to deal with "X + -1" or
> > > "X - -1"?
> > 
> > Hmm I thought they were canonicalized - but now I verified they are not.
> > *sigh*  Are there any problems with treating X - -1 as X + 1?
> 
> No, though you can't fold X + INT_MIN or X - INT_MIN that way.  (But 
> the optimization is still valid as applied to X + INT_MIN CMP Y - INT_MAX 
> converted to X - 1 CMP Y.)

Based on the reasoning that abs(-1) <= abs(INT_MIN), correct?

Richard.

References:
- [PATCH] Fix PR26898, 2nd try
  - From: Richard Guenther
- Re: [PATCH] Fix PR26898, 2nd try
  - From: Joseph S. Myers
- Re: [PATCH] Fix PR26898, 2nd try
  - From: Richard Guenther
- Re: [PATCH] Fix PR26898, 2nd try
  - From: Joseph S. Myers

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