3 ms·
Well half the cause (or probably two thirds), as the post shows there’s a second quadratic blowup when filling the not-a-hashmap thing after the very slow parsi
by masklinn 6y ago
Well half the cause (or probably two thirds), as the post shows there’s a second quadratic blowup when filling the not-a-hashmap thing after the very slow parsing has completed.
- crowbahr 6y agoA filling that wasn't even required because it was de-duping already de-duped data instead of just inserting it into a list (or a hashmap, as suggested in the post).
- masklinn 6y agoThe filling was necessary, but the deduplication (in linear time as it works over an array) indeed was not as the input is unique. I can see why the devs would not trust that though. So it’s the inefficiency of the dedup’ which bugs me more than the dedup itself.