On Fri, Oct 16, 2015 at 5:12 AM, Jan Hubicka <hubi...@ucw.cz> wrote:
> Hi,
> as Richard noticed in my port of the code to operand_equal_p, the checking of
> CONSTURCTOR in ipa-icf-gimple is incomplete missing the index checks.
> It is also unnecesary since non-empty ctors does not happen as gimple
> operands. This patch thus removes the unnecesary code.
Err - they do happen, for vector constructors. Just empty constructors
are not allowed for vector constructors - vector constructors are required
to have elements in proper order and none left out.
Sorry for misleading you.
> Bootstrapped/regtested x86_64-linux, comitted.
this will definitely ICE ...
Richard.
> Honza
>
> * ipa-icf-gimple.c (func_checker::compare_operand): Compare only
> empty constructors.
> Index: ipa-icf-gimple.c
> ===================================================================
> --- ipa-icf-gimple.c (revision 228851)
> +++ ipa-icf-gimple.c (working copy)
> @@ -415,20 +415,9 @@ func_checker::compare_operand (tree t1,
> switch (TREE_CODE (t1))
> {
> case CONSTRUCTOR:
> - {
> - unsigned length1 = vec_safe_length (CONSTRUCTOR_ELTS (t1));
> - unsigned length2 = vec_safe_length (CONSTRUCTOR_ELTS (t2));
> -
> - if (length1 != length2)
> - return return_false ();
> -
> - for (unsigned i = 0; i < length1; i++)
> - if (!compare_operand (CONSTRUCTOR_ELT (t1, i)->value,
> - CONSTRUCTOR_ELT (t2, i)->value))
> - return return_false();
> -
> - return true;
> - }
> + gcc_assert (!vec_safe_length (CONSTRUCTOR_ELTS (t1))
> + && !vec_safe_length (CONSTRUCTOR_ELTS (t2)));
> + return true;
> case ARRAY_REF:
> case ARRAY_RANGE_REF:
> /* First argument is the array, second is the index. */
