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: [ping] Re: c_common_type_for_mode: pass precision, not mode


On Fri, 2005-01-07 at 22:34 -0500, DJ Delorie wrote:
> http://gcc.gnu.org/ml/gcc-patches/2004-12/msg01186.html
> 
> > X-Sieve: CMU Sieve 2.2
> > Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
> > Sender: gcc-patches-owner@gcc.gnu.org
> > Date: Wed, 15 Dec 2004 22:46:58 -0500
> > From: DJ Delorie <dj@redhat.com>
> > X-RedHat-Spam-Score: 0 
> > 
> > 
> > I found this with an internal port, so if someone suggests a port with
> > a pointer mode that doesn't match any "normal" integer type modes I
> > can test it officially.
The last one I was aware of was the mn102, which we deprecated a while
back.  It had 24bit pointers.

> > 
> > I'm also unsure whether the mode's bitsize or precision is the
> > appropriate value to pass.  Consider, for example, using a fractional
> > int mode for Pmode.
Based on the comments, I would think precision.

> > 
> > 2004-12-15  DJ Delorie  <dj@redhat.com>
> > 
> > 	* c-common.c (c_common_type_for_mode): Pass the mode's precision
> > 	to make_[un]signed_type, not the mode itself.
Presumably this bootstrapped and comparison tested?  If so, then it's
fine (the code in mainline right now is clearly incorrect).

Do you have a testcase?  Even if it only triggers on the internal port
it would be good to get the test added to the testsuite.

jeff



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