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 PR c++/27805


The value of ptrmem_type is not what causes the ICE in this particular
testcase. Upon entry to build_m_component_ref datum has
error_mark_node as it's tree type which causes the failure on the
following line:

objtype = TYPE_MAIN_VARIANT (TREE_TYPE (datum));

Cheers,
Lee.

On 6/17/06, Andrew Pinski <pinskia@physics.uc.edu> wrote:

On Jun 17, 2006, at 8:11 AM, Lee Millward wrote:


> Bootstrapped and regression tested with no new failures on
> i686-pc-linux-gnu. Ok for mainline?

+  if (error_operand_p(datum) || error_operand_p(component))
Whitespace is missing.

Also why can't you use ptrmem_type == error_mark_node after the
setting of ptrmem_type?

-- Pinski



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