This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] PR18191
- From: Steven Bosscher <stevenb at suse dot de>
- To: Nathan Sidwell <nathan at codesourcery dot com>
- Cc: gcc-patches at gcc dot gnu dot org, pinskia at physics dot uc dot edu
- Date: Sat, 18 Dec 2004 12:18:01 +0100
- Subject: Re: [PATCH] PR18191
- Organization: SUSE Labs
- References: <200412180115.44441.stevenb@suse.de> <41C3FA6E.9020106@codesourcery.com>
On Saturday 18 December 2004 10:37, Nathan Sidwell wrote:
> Steven Bosscher wrote:
> > --- 4252,4275 ----
> > tree lo_index = TREE_OPERAND (purpose, 0);
> > tree hi_index = TREE_OPERAND (purpose, 1);
> >
> > + /* We can silently generate wrong code when the indices of the
> > + range are not representable on the host machine. */
> > if (host_integerp (lo_index, 1) && host_integerp (hi_index, 1))
> > mult = (tree_low_cst (hi_index, 1)
> > - tree_low_cst (lo_index, 1) + 1);
> > + else
> > + abort ();
>
> shouldn't this be a sorry?
Good point. How about the following:
sorry ("constructor with more than " HOST_WIDE_INT_PRINT
"elements not supported on this host.", MAX_HOST_WIDE_INT);
Is there also a sign issue, ie. is tree_low_cst (...) an unsigned
HOST_WIDE_INT?
Gr.
Steven