Dan Sugalski wrote:
At 9:35 AM +0100 12/4/02, Leopold Toetsch wrote:
Doese odd twiddling include the Buffer/PMC unification?
Yup. This'd be a good time to do that.
So the question is, how should the final "Parrot Object" look like. I did send a test program on Oct. 25th which got warnocked:
Re: [RFC] Buffer/PMC unification, variable/value vtable split, tied scalars
Via save manipulating ops or just pass over e.g. flags?
Everything in the pmc struct will be readable. I think everything'll be writable, though there are some issues with that.
I would encapsulate the access into the ops: set_pmc_internals pmc, what, data e.g. set I0, .PmcFflags set I1, .MemIsMalloced set_pmc_internals P0, I0, I1 This is expandable and save. Setting arbitrary flags/fields could do really Bad Things. leo