3/18/2026 at 11:12:35 AM
> input size normal hardened speedup w/ hardened> 1,000 0.7ms 28us 25x
> 5,000 18ms 146us 123x
> 10,000 73ms 303us 241x
> 50,000 1.8s 1.6ms 1,125x
Why is there a normal mode if hardened mode is faster for all input sizes?
by nextaccountic
3/18/2026 at 12:53:55 PM
Sorry, finished the post just now with more comparisons on other inputsThe reason is just that the normal mode is faster in average non pathological cases
by ieviev
3/18/2026 at 1:38:15 PM
Could you have a heuristics based on the input size and the pattern to decide what to use?by tracnar
3/18/2026 at 1:39:33 PM
Yes, this is entirely possible. you can even explore the automaton eagerly and detect if it's possible to loop from an accepting state to a nonaccepting one.Exciting stuff for future work
by ieviev
3/18/2026 at 9:59:15 PM
Ripgrep does something like thhis. It has a meta regex engine that switches engine when it finds what looks like pathological cases (or rather, the regex-automata crate does, which is used by the regex crate, which powers ripgrep).https://docs.rs/regex-automata/latest/regex_automata/meta/st...
Ripgrep in turn exposes some knobs to tweak the heuristics
https://github.com/BurntSushi/ripgrep/blob/master/FAQ.md#how...
by nextaccountic