: > maybe a regex version could be started later
:
: Don't hurry... I think that's moving in the wrong
direction for good performance
: http://www.nabble.com/NO_NORMS-and-TOKENIZE
D--tf3064721.html#a9047635
i'd completley forgotten about that thread (was it just Feb?
it seems like
a year ago) ... a regex based approach would certainly be
slower then
the prefix/suffix mechanism we have now ... but i'm guessing
for a lot of
use cases it would be fine -- it's also the kind of thing
whee we could
document that they expressions are evaluated in order --
putting the onus
on teh schema creator to order them based on the likelyhood
of matching.
-Hoss
|