> Luke Palmer <[EMAIL PROTECTED]> wrote: > > +op pop_pad(out PMC) { > > + $1 = new_pmc_header(interpreter); > > + stack_pop(interpreter, &interpreter->ctx.pad_stack, > > + &$1, STACK_ENTRY_PMC); > > goto NEXT(); > > Do we really need a new PMC header here? The PMC already must have one.
Duh. I even looked at the source, and it was clear that one wasn't needed... but I wrote it anyway. It was late. > > +rotate_entries(Interp *interpreter, Stack_Chunk_t *stack, INTVAL num_entries) > > { > > Stack_Entry_t temp; > > - Intval i; > > - Intval depth = num_entries - 1; > > + INTVAL i; > > + INTVAL depth = num_entries - 1; > > And these are typical cases, where a plain int or size_t seems more > appropriate, as Brian did show. But of course the whole source tree is > full of such (ab)usage of INTVALs. Ahh, I was just doing a lexical transformation. I didn't actually look at what they were used for. The source tree is now free of Intval (and Floatval), at least. Luke