mirror of
https://github.com/SerenityOS/serenity
synced 2026-05-13 02:16:39 +02:00
b46bb32584
Instead of recomputing boundaries on every iteration, this now computes boundaries once and then searches that vector. This is still O(n^2) for now, but already much faster (at the expense of using O(n) memory): `BenchmarkSegmenter.for_each_boundary` stays at around 3ms, but `forward` and `backward` go from ~13s to ~60ms on my system.