On Thu, Apr 16, 2015 at 3:54 PM, Jason Merrill wrote:
> OK.
Thanks, committed as revision 222176.
>
> Jason
OK.
Jason
On Thu, 16 Apr 2015, Jason Merrill wrote:
On 04/15/2015 09:00 PM, Patrick Palka wrote:
- if (!cleared || num_nonzero_elements > 0)
How about adding || TREE_SIDE_EFFECTS (TREE_OPERAND (*expr_p), 1) to this
test rather than removing it entirely?
That works too. Does the following patc
On 04/15/2015 09:00 PM, Patrick Palka wrote:
- if (!cleared || num_nonzero_elements > 0)
How about adding || TREE_SIDE_EFFECTS (TREE_OPERAND (*expr_p), 1) to
this test rather than removing it entirely?
Jason
gimplify_init_constructor() only attempts to gimplify a CONSTRUCTOR's
elts under two conditions: if the object has not been zero-initialized,
or if categorize_ctor_elements() detected more than one "nonzero" elt in
the CONSTRUCTOR. With the PR's testcase, both of these conditions are
false: the gi
