When evaluating a conditional, the expression is first evaluated and some further verifications and processing are done if the returned type is not NULL. However, the returned type can also be 'bad_ctype' and if it is the case, the additional verifications will just give meaningless additional warnings. Fix this by using the new helper valid_type() instead of just testing for a null ctype. Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@xxxxxxxxx> --- evaluate.c | 2 +- validation/bad-type-twice1.c | 1 - 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/evaluate.c b/evaluate.c index 1fc6111e4..3c4a25325 100644 --- a/evaluate.c +++ b/evaluate.c @@ -879,7 +879,7 @@ static struct symbol *evaluate_conditional(struct expression *expr, int iterator warning(expr->pos, "assignment expression in conditional"); ctype = evaluate_expression(expr); - if (!ctype) + if (!valid_type(ctype)) return NULL; if (is_safe_type(ctype)) warning(expr->pos, "testing a 'safe expression'"); diff --git a/validation/bad-type-twice1.c b/validation/bad-type-twice1.c index 2f4e2838f..95cfd9e01 100644 --- a/validation/bad-type-twice1.c +++ b/validation/bad-type-twice1.c @@ -7,7 +7,6 @@ static unsigned long foo(unsigned long val, void *ref) /* * check-name: bad-type-twice1 - * check-known-to-fail * * check-error-start bad-type-twice1.c:3:17: error: incompatible types for operation (>=) -- 2.16.0 -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
