On Thu, Sep 27, 2012 at 9:07 AM, Paul Eggert <egg...@cs.ucla.edu> wrote: > It might be better to not have the tablesize > arg, at least in a variant of the function.
This would result to 4 functions instead: 1. the original hash_pjw() 2. hash_pjw_no_tablesize() 3. hash_pjw_s() 4. hash_pjw_s_no_tablesize() We can eliminate (2), but still there are 3 variants of the same function. Maybe it is better to keep only (4) and (1) for backwards compatibility, and anybody who wants to do the % tablesize to do it after calling the hash_pjw_s_no_tablesize() function? regards, Nikos
