On Wednesday 12 March 2008 12:36:08 [EMAIL PROTECTED] wrote:
> Modified:
> trunk/src/io/io.c
>
> Log:
> Not sure of this fix, but one more test passing (test 16 from t/src/io.t)
>
>
> Modified: trunk/src/io/io.c
> ===========================================================================
>=== --- trunk/src/io/io.c (original)
> +++ trunk/src/io/io.c Wed Mar 12 12:36:07 2008
> @@ -1284,6 +1284,7 @@
>
> */
>
> +PARROT_API
> PIOOFF_T
> PIO_make_offset(INTVAL offset)
> {
It's not the right fix, but it's not your fault.
Unless anyone objects, I'm going to rip out all of the silly t/src/ tests
which poke into the guts of Parrot from C through an API that we should never
expose and do not want to support. As a general policy, if the solution to a
problem appears to be putting PARROT_API on something not explictly mentioned
in the embedding/extending PDDs, then first we need to figure out why it's
not in those PDDs and only then consider adding the decorator.
Half of the tests in t/src/ are disasters anyway. If our existing tests of
PMCs and such don't actually test that, for example, hashes work properly,
something's very wrong.
-- c
