On Wed, 23 Apr 2014, Marek Polacek wrote:
> diff --git gcc/testsuite/c-c++-common/attributes-1.c
> gcc/testsuite/c-c++-common/attributes-1.c
> index af4dd12..8458e47 100644
> --- gcc/testsuite/c-c++-common/attributes-1.c
> +++ gcc/testsuite/c-c++-common/attributes-1.c
> @@ -9,7 +9,7 @@ typedef char vec __attribute__((vector_size(bar))); /* {
> dg-warning "ignored" }
> void f1(char*) __attribute__((nonnull(bar))); /* { dg-error "invalid
> operand" } */
> void f2(char*) __attribute__((nonnull(1,bar))); /* { dg-error "invalid
> operand" } */
>
> -void g() __attribute__((aligned(bar))); /* { dg-error "invalid value|not an
> integer" } */
> +void g() __attribute__((aligned(bar)));
I don't think it's appropriate to remove any test assertion that this
invalid code gets diagnosed.
If the only diagnostic is now one swallowed by the dg-prune-output in this
test, either that dg-prune-output needs to be removed (and corresponding
more detailed error expectations added), or a separate test needs adding
for this erroneous use of this attribute (that separate test not using
dg-prune-output).
--
Joseph S. Myers
[email protected]
