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] tree-vrp.c: Work around a bug in the front end thatproduces a comparison of mismatched types.


On Fri, 2005-04-15 at 12:43 -0400, Kazu Hirata wrote:
> Hi Jeff,
> 
> > > > 	PR tree-optimization/21021
> > > > 	* tree-vrp.c (compare_values): Work around a bug in the front
> > > > 	end that produces a comparison of mismatched types.
> > > > 
> > > OK.
> > Err, umm, I DO NOT like this.
> 
> I do not either.
> 
> > We're working around a bug in some bogus front-end.  Let's
> > fix that rather than hacking up the tree optimizers to deal with
> > clearly bogus and broken code.
> > 
> > I'd really rather see this patch come out and the real bug fixed.
> 
> Yes, that's what I am planning to do.  Indeed, I just found the source
> of this type mismatch.  fold_binary is stripping what I think are
> useful conversions.
> 
> I added more detailed analysis in PR 21024.  If you could give us a
> thought on this, that would be greatly appreciated.
> 
> Maybe the right fix is as simple as teaching STRIP_SIGN_NOPS not to
> strip an integer-pointer type conversion.
It's my understanding that this problem has been resolved by the
recent change to expand_builtin_strcat and find_binary, right?

jeff



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