On 11/18/2013 01:54 PM, David Malcolm wrote:
On Thu, 2013-11-14 at 00:37 -0700, Jeff Law wrote:
On 10/31/13 10:26, David Malcolm wrote:
* gimple.h (is_a_helper <const gimple_statement_with_ops>::test): New.
(is_a_helper <gimple_statement_with_ops>::test): New.
(is_a_helper <const gimple_statement_with_memory_ops>::test): New.
(is_a_helper <gimple_statement_with_memory_ops>::test): New.
OK with the usual conditions. Check with Andrew as to the location of
these helpers since he's in the middle of ripping apart gimple.h.
The helpers become used in accessors such as gimple_use_ops,
gimple_vuse, gimple_vdef etc as of the automated patch (patch 3 in the
series [1]), so presumably the helpers need to be in the same file as
those accessors. The accessors are currently in gimple.h, and hence I
also put the helpers within gimple.h (earlier within the header).
Andrew, do you plan to move the accessor functions out of gimple.h? If
so, where should the is_a_helper fns land?
Not currently. They are a component of the gimple statement, and
gimple.[ch] is now just the statement and related accessors. Since your
changes are just for gimple-stmt, in theory everything should go there.
Andrew