On 05/28/2012 01:32 PM, Gabriel Dos Reis wrote:
On Mon, May 28, 2012 at 6:13 AM, Paolo Carlini<paolo.carl...@oracle.com> wrote:
Hi,
apparently we have serious troubles with<cmath> and -std=c++11 and
-fno-trapping-math because LTGT_EXPR is unhandled in
potential_constant_expression_1. Thus, unless we have sound reasons to *not*
handle it together with all the other *_EXPR, I think the below could safely
go in mainline and 4_7-branch. Bootstrapped and tested x86_64-linux.
Ok?
I suspect its missing from the original implementation must have been
my ignorance of its existence. OK.
Thanks, applied.
Paolo.
