On Jul 10, 2014, at 8:07 PM, Yann Ylavic <[email protected]> wrote:
> > I propose to garantee the ordering presumably at low coding and > performance (if any) cost, so that one can choose between apr_table or > apr_skiplist to do similar things (iterate over all the values of a > key is one such thing, in a deterministic order is an other, both not > feasible with the current skiplist implementation). > ++1... For me, performance is key. Or, at least, some sort of option to allow for a skiplist to be created that guarantees order or not, depending on whether performance or ordering is more important. After all, the intent of skiplist *is* in performance, and not ordering, so sacrificing performance for ordering seems kinda wonky, unless one can turn that off :)
