chore: upgrade regex-syntax and regex-automata crates.

This solves an issue that prevented using unicode characters in regular expressions.
See: rust-lang/regex@04f5d7b

Cargo.toml

@@ -77,8 +77,8 @@ protobuf-codegen = { git = "https://github.com/plusvic/rust-protobuf.git", rev =
 protobuf-json-mapping = { git = "https://github.com/plusvic/rust-protobuf.git", rev = "b484d8a7" }
 protobuf-parse = { git = "https://github.com/plusvic/rust-protobuf.git", rev = "b484d8a7" }
 protobuf-support = { git = "https://github.com/plusvic/rust-protobuf.git", rev = "b484d8a7" }
-regex-syntax = { git = "https://github.com/plusvic/regex.git", rev = "423493d" }
-regex-automata = { git = "https://github.com/plusvic/regex.git", rev = "423493d" }
+regex-syntax = { git = "https://github.com/plusvic/regex.git", rev = "50a708b" }
+regex-automata = { git = "https://github.com/plusvic/regex.git", rev = "50a708b" }
 roxmltree = "0.19.0"
 rustc-hash = "1.1.0"
 smallvec = "1.10.0"

lib/src/re/parser.rs

@@ -52,7 +52,7 @@ impl Parser {
     /// If true, allows regular expressions that mixes greedy and non-greedy
     /// quantifiers (e.g: `/ab.*cd.*?ef/`). When mixed greediness is not allowed
     /// [`Parser::parse`] returns an error if the regular expression contains
-    /// both greedy and non-greedy quantifiers. By default mixed greediness is
+    /// both greedy and non-greedy quantifiers. By default, mixed greediness is
     /// allowed.
     pub fn allow_mixed_greediness(mut self, yes: bool) -> Self {
         self.allow_mixed_greediness = yes;
@@ -89,13 +89,12 @@ impl Parser {
             regexp.case_insensitive
         };
 
-        let mut translator =
-            regex_syntax::hir::translate::TranslatorBuilder::new()
-                .case_insensitive(case_insensitive)
-                .dot_matches_new_line(regexp.dot_matches_new_line)
-                .unicode(false)
-                .utf8(false)
-                .build();
+        let mut translator = re::hir::translate::TranslatorBuilder::new()
+            .case_insensitive(case_insensitive)
+            .dot_matches_new_line(regexp.dot_matches_new_line)
+            .unicode(false)
+            .utf8(false)
+            .build();
 
         let hir = translator.translate(regexp.src, &ast).map_err(|err| {
             Error::SyntaxError {

lib/src/tests/mod.rs

@@ -271,6 +271,8 @@ fn string_operations() {
     condition_true!(r#""foo\nbar" matches /foo.*bar/s"#);
     condition_false!(r#""foo\nbar" matches /foo.*bar/"#);
     condition_true!(r#""foobar" matches /fo{,2}bar/"#);
+    condition_true!(r#""タイトル" matches /タイトル/"#);
+    condition_true!(r#""🙈🙉🙊" matches /.../"#);
 }
 
 #[test]
@@ -1309,17 +1311,10 @@ fn regexp_patterns_4() {
 
     pattern_match!(r"/foo\x01bar/", b"foo\x01bar", b"foo\x01bar");
 
-    /*
-    TODO: YARA accepts unicode characters in regexps but regexp_syntax either
-    accepts unicode characters or escape sequences like \x01, but not both
-    at the same time. Presumably this is because with escape sequence you can't
-    create non-valid unicode codepoints, so when you enable unicode it disables
-    escape sequences.
     pattern_true!(
         r#"/🙈🙉🙊/i"#,
         b"\xF0\x9F\x99\x88\xF0\x9F\x99\x89\xF0\x9F\x99\x8A"
     );
-    */
 }
 
 #[test]