On Sat, 15 Sep 2007, Allison Randal wrote:
> I just launched PDD 17 out of the drafts directory, which marks it as ready
> for implementation. Jonathan Worthington and chromatic have volunteered to
> work with me on implementing the refinements to the PMC design.
I had two questions:
1. Is the order of elements in struct PMC deliberate?
struct PMC {
Parrot_UInt flags;
UnionVal cache;
VTABLE *vtable;
DPOINTER *data;
struct PMC_EXT *pmc_ext;
};
In particular, on 32-bit SPARC, sizeof(flag) will be 4, but cache will
have to be aligned on an 8-byte boundary. Thus there will be 4 "wasted"
bytes of padding. Swapping the order (putting cache first and flags
second) will reduce the size of a pmc by 4 bytes. Not crucial, but
nice.
2. Did you deliberately drop _int_val2 from UnionVal?
Specifically, the former
struct _i {
INTVAL _int_val;
INTVAL _int_val2;
} _i;
has been replaced by a single _int_val entry. However, since the
UnionVal already contains 2 pointers, keeping the second _int_val2 only
costs space on systems where sizeof(INTVAL) > sizeof(void *). Are such
platforms a significant concern?
Or were there other good reasons for dropping the entry? I do note that
it's used heavily in the various array pmcs (including
src/pmc/resizeablebooleanarry.pmc, which I know you've worked on
relatively recently.)
--
Andy Dougherty [EMAIL PROTECTED]
