When I fixed PR 102622, I accidently left behind a TYPE_PRECISION
check which I had there for checking before hand. This check
is not needed as the code will handle it correctly anyways.
Committed as obvious after a bootstrap/test on x86_64-linux-gnu.
PR tree-optimization/10352
gcc/ChangeLog:
* match.pd: Remove check of TYPE_PRECISION for
the a?0:pow2 case.
gcc/testsuite/ChangeLog:
* gcc.c-torture/execute/pr10352-1.c: New test.
(if (integer_onep (@2))
(convert (bit_xor (convert:boolean_type_node @0) { booltrue; } )))
/* a ? powerof2cst : 0 -> (!a) << (log2(powerof2cst)) */
- (if (INTEGRAL_TYPE_P (type) && integer_pow2p (@2)
- && TYPE_PRECISION (type) != 1)
+ (if (INTEGRAL_TYPE_P (type) && integer_pow2p (@2))
(with {
tree shift = build_int_cst (integer_type_node, tree_log2 (@2));
}
--- /dev/null
+/* this is another case where phiopt
+ would create -signed1bit which is undefined. */
+struct {
+ int a:1;
+} b;
+int *c = (int *)&b, d;
+int main() {
+ d = c && (b.a = (d < 0) ^ 3);
+ if (d != 1)
+ __builtin_abort();
+ return 0;
+}