On Thursday 19 June 2008 07:28:55 [EMAIL PROTECTED] wrote:
> Log:
> [rakudo] Implement basic use of does operator for mixing in a role to an
> existing object at runtime. This was most neatly done with a dynop, so this
> check-in also adds a dynops file for Rakudo. make makefile will be required
> to build after this check-in.
> Added: trunk/languages/perl6/src/ops/perl6.ops
> ===========================================================================
>=== --- (empty file)
> +++ trunk/languages/perl6/src/ops/perl6.ops Thu Jun 19 07:28:54 2008
> @@ -0,0 +1,77 @@
> +/*
> + * $Id:$
> + * Copyright (C) 2008, The Perl Foundation.
> + */
> +
> +#include "parrot/dynext.h"
> +#include "../../../../src/pmc/pmc_object.h"
> +VERSION = PARROT_VERSION;
> +
> +/* */
> +inline op rebless_subclass(in PMC, in PMC) :base_core {
> + PMC *current_class;
> + PMC *parent_list;
> + PMC *value;
> + int i, num_parents, in_parents, new_attribs;
> +
> + /* First verify that the object's class is a superclass of the one
> we're + * to re-bless it into. While we're at it, count the number of
> attributes + * the current class has that the parent class does not. */
> + current_class = VTABLE_get_class(interp, $1);
> + parent_list = VTABLE_inspect_str(interp, $2,
> string_from_literal(interp, "all_parents"));
> + num_parents = VTABLE_elements(interp, parent_list);
> + in_parents = 0;
> + new_attribs = 0;
The initializations and declarations can go on the same lines.
I'd like to replace all of the string_from_literal with const_string. While
the former calculates the string length at compile time (and saves a very
short strlen call at runtime), const_string uses a cache of constant string
headers, trading a lookup cost to retrieve or store a singletone string
header for not allocating a new non-constant string header on each
invocation.
We'll already have "Mutable" and "attributes" in the cache, so this is
effectively a free optimization.
-- c
