1 | use regex_automata::{dense, DFA}; |
2 | |
3 | // A regression test for checking that minimization correctly translates |
4 | // whether a state is a match state or not. Previously, it was possible for |
5 | // minimization to mark a non-matching state as matching. |
6 | #[test] |
7 | fn minimize_sets_correct_match_states() { |
8 | let pattern = |
9 | // This is a subset of the grapheme matching regex. I couldn't seem |
10 | // to get a repro any smaller than this unfortunately. |
11 | r"(?x) |
12 | (?: |
13 | \p{gcb=Prepend}* |
14 | (?: |
15 | (?: |
16 | (?: |
17 | \p{gcb=L}* |
18 | (?:\p{gcb=V}+|\p{gcb=LV}\p{gcb=V}*|\p{gcb=LVT}) |
19 | \p{gcb=T}* |
20 | ) |
21 | | |
22 | \p{gcb=L}+ |
23 | | |
24 | \p{gcb=T}+ |
25 | ) |
26 | | |
27 | \p{Extended_Pictographic} |
28 | (?:\p{gcb=Extend}*\p{gcb=ZWJ}\p{Extended_Pictographic})* |
29 | | |
30 | [^\p{gcb=Control}\p{gcb=CR}\p{gcb=LF}] |
31 | ) |
32 | [\p{gcb=Extend}\p{gcb=ZWJ}\p{gcb=SpacingMark}]* |
33 | ) |
34 | " ; |
35 | |
36 | let dfa = dense::Builder::new() |
37 | .minimize(true) |
38 | .anchored(true) |
39 | .build(pattern) |
40 | .unwrap(); |
41 | assert_eq!(None, dfa.find(b" \xE2" )); |
42 | } |
43 | |