Le Tue, Jan 13, 2004 at 10:34:14PM +0100, le valeureux mongueur Stéphane Payrard a dit:
> Le Tue, Jan 13, 2004 at 03:06:18PM -0600, le valeureux mongueur Robert Eaglestone a
> dit:
> > OK, I'm looking at the Parrot String documentation, and I've
> > got questions. It's not like the docs are a total mess, they
> > just need some fleshing out. Yeah, that's it. So here I go.
> >
> > Here's the page I'm looking at:
> >
> > http://www.parrotcode.org/docs/strings.pod.html
> >
> > And here are my questions. Or, rather, notes which have
> > questions in them.
> >
> >
> > * OK, I'm game, is 'String' a new thing that's been added to
> > C in the last ten years? I can't find it defined anywhere;
> > my brain must have gone to mush.
>
>
> <STRING>, C<String> are different names for a C<struct
> parrot_string_t>.
I forgot to mention.
You will find the definitions in F<include/parrot/string.h>.
typedef struct parrot_string_t String;
...
#define STRING struct parrot_string_t
etags or ctags (depending if you are an emacs or vi user) is your
friend. I use the following to avoid gathering definitions outside
the core of parrot.
etags `find -name '*.pmc' -o -name '*.c' -o -name '*.h' | grep -v languages/ | grep
-v icu/`
Strings are garbage collected. All garbage
> collected entities are accessible thru a pointer to a C<struct
> pobj_t> that is an union discriminated by the member C<flags>.
>
> Relevant code in F<include/parrot/pobj.h>:
>
> typedef union UnionVal {
> ...
> struct parrot_string_t * string_val;
> }
>
> typedef struct pobj_t {
> UnionVal u;
> Parrot_UInt flags;
> }
>
> typedef enum PObj_enum {
> ...
> PObj_is_string_FLAG = 1 << 8,
> }
>
>
> But this is mostly irrelevant to the string user that will use
> the API you document. If you care about internals, see more info
> about garbage collection in F<docs/memory_internals.pod>.
>
>
> --
> stef
