Steve Fink wrote: > On Mon, Oct 07, 2002 at 11:25:31PM +0200, Leopold Toetsch wrote:
>>If conserving memory is an issue, list.c is much more efficient for very >>small and very big arrays. We will see. >> > > Sounds good to me. I originally wrote, and am currently using, intlist > for regular expressions. (The stuff checked into languages/regex > generates code that uses it, although I haven't committed the hooks to > allow languages/perl6 to use languages/regex instead of its native > code.) The access pattern for regexes is very simple: it only uses it > as a stack. .... where depth depends on the backtracking behaviour of the rule and isn't know in advance. So the small items would be a win for probably 90% of the usage. > ... It *might* be useful to be able to pop off multiple > entries at once (or equivalently, set the size of the array/list), list_delete(interp, list, idx, n_to_delete) would do this, if (length-1-idx) == n_to_delete. But just adjusting the array length is probably simpler. list_delete isn't ready yet, but the infrastructure is almost finished. > ... and > it might be useful to be able to append another array onto the end. These are just list_assigns in a loop. > But that's it. So I know of no reason for intlist.c to be kept > separate; I'm perfectly happy having it replaced by list.c. Fine, $code_duplication-- - soon. leo
