feat: allow the use of the {,n} syntax in patterns (#41)

The implementation of this feature for regexps used with the `match` operator is still pending.

Cargo.toml

@@ -46,7 +46,7 @@ lazy_static = "1.4.0"
 line-span = "0.1.3"
 linkme = "0.3"
 log = "0.4"
-memchr = "2.5.0"
+memchr = "2.6.3"
 memx = "0.1.28"
 num = "0.4.0"
 pest = "2.5.5"
@@ -55,8 +55,8 @@ pretty_assertions = "1.3.0"
 protobuf = "3.2.0"
 protobuf-codegen = "3.2.0"
 protobuf-parse = "3.2.0"
-regex = "1.9.1"
-regex-syntax = "0.7.4"
+regex = { git = "https://github.com/plusvic/regex.git", rev="423493d" }
+regex-syntax = { git = "https://github.com/plusvic/regex.git", rev="423493d" }
 rustc-hash = "1.1.0"
 smallvec = "1.10.0"
 serde = "1.0.156"

yara-x/src/compiler/rules.rs

@@ -207,6 +207,7 @@ impl Rules {
     /// If no regular expression with such [`RegexpId`] exists.
     #[inline]
     pub(crate) fn get_regexp(&self, regexp_id: RegexpId) -> Regex {
+        // TODO: when compiling regexps allow the use of `{,n}` syntax
         let re = Regexp::new(self.regexp_pool.get(regexp_id).unwrap());
         RegexBuilder::new(re.naked())
             .case_insensitive(re.case_insensitive())

yara-x/src/re/parser.rs

@@ -61,12 +61,14 @@ impl Parser {
 
     /// Parses the regexp and returns its HIR.
     pub fn parse(&self, regexp: &ast::Regexp) -> Result<Hir, Error> {
-        let ast = re::ast::parse::Parser::new().parse(regexp.src).map_err(
-            |err| Error::SyntaxError {
+        let mut parser =
+            re::ast::parse::ParserBuilder::new().empty_min_range(true).build();
+
+        let ast =
+            parser.parse(regexp.src).map_err(|err| Error::SyntaxError {
                 msg: err.kind().to_string(),
                 span: *err.span(),
-            },
-        )?;
+            })?;
 
         let greedy = Validator::new().validate(&ast);
 

yara-x/src/tests/mod.rs

@@ -916,6 +916,7 @@ fn regexp_patterns_2() {
     pattern_match!(r#"/ab{2,3}c/"#, b"abbbc", b"abbbc");
     pattern_match!(r#"/ab{2,3}?c/"#, b"abbbc", b"abbbc");
     pattern_match!(r#"/ab{0,1}?c/"#, b"abc", b"abc");
+    pattern_match!(r#"/ab{,1}?c/"#, b"abc", b"abc");
     pattern_match!(r#"/a{0,1}bc/"#, b"bbc", b"bc");
     pattern_match!(r#"/ab{0,}c/"#, b"ac", b"ac");
     pattern_match!(r#"/ab{0,}c/"#, b"abc", b"abc");
@@ -935,7 +936,9 @@ fn regexp_patterns_2() {
     pattern_false!(r#"/ab{1,}b/"#, b"ab");
     pattern_match!(r#"/ab{1,1}c/"#, b"abc", b"abc");
     pattern_match!(r#"/ab{0,3}c/"#, b"abbbc", b"abbbc");
+    pattern_match!(r#"/ab{,3}c/"#, b"abbbc", b"abbbc");
     pattern_false!(r#"/ab{0,2}c/"#, b"abbbc");
+    pattern_false!(r#"/ab{,2}c/"#, b"abbbc");
     pattern_false!(r#"/ab{4,5}c/"#, b"abbbc");
     pattern_false!(r#"/ab{3}c/"#, b"abbbbc");
     pattern_false!(r#"/ab{4}c/"#, b"abbbbbc");