Leopold Toetsch <[EMAIL PROTECTED]> writes: > Dan Sugalski <[EMAIL PROTECTED]> wrote: >> And semantic differences--don't forget those. > > A keyed "add" vtable doesn't help to provide more semantics. The set vs > assign thread applies here too. On the contrary: to provide all > semantics you would need "add_set_p_k_p_k_p_k" and "add_assign_p_k_p_k_p_k" > and possibly "add_clone_p_k_p_k_p_k".
If I'm on target, what Dan's trying to avoid is having to create magic proxy objects for autovivification. If this is all we want out of keyed ops, then we can get it more cheaply by only adding keyed variants for the _out_ parameters of each op. That way the standard get_*_keyed ops can either return NULL, return a new PMC without insertion, or insert and return a new PMC for read parameters, while the *_pk_*_* ops can do "the right thing" with respect to autovivification. No temporary keys or null keys need be created. I think this covers the semantics. Beyond that, it's just a matter of performance -- extra opcodes vs. extra ops -- that we can fight out in the benchmark arena. /s
