Add {n,m} repetition range support (#278)

* Add {n,m} regex support * Added test for repeating character class
maciejhirsz · Feb 26, 2023 · 1ecd6a4 · 1ecd6a4
1 parent 51c1f8c
commit 1ecd6a4
Show file tree

Hide file tree

Showing 2 changed files with 34 additions and 4 deletions.
diff --git a/logos-codegen/src/mir.rs b/logos-codegen/src/mir.rs
@@ -1,6 +1,6 @@
 use std::convert::TryFrom;
 
-use regex_syntax::hir::{Hir, HirKind, RepetitionKind};
+use regex_syntax::hir::{Hir, HirKind, RepetitionKind, RepetitionRange};
 use regex_syntax::ParserBuilder;
 
 pub use regex_syntax::hir::{Class, ClassUnicode, Literal};
@@ -118,8 +118,34 @@ impl TryFrom<Hir> for Mir {
                     RepetitionKind::OneOrMore => {
                         Ok(Mir::Concat(vec![mir.clone(), Mir::Loop(Box::new(mir))]))
                     }
-                    RepetitionKind::Range(..) => {
-                        Err("#[regex]: {n,m} repetition range is currently unsupported.".into())
+                    RepetitionKind::Range(range) => {
+                        match range {
+                            RepetitionRange::Exactly(n) => {
+                                let mut out = Vec::with_capacity(n as usize);
+                                for _ in 0..n {
+                                    out.push(mir.clone());
+                                }
+                                Ok(Mir::Concat(out))
+                            }
+                            RepetitionRange::AtLeast(n) => {
+                                let mut out = Vec::with_capacity(n as usize );
+                                for _ in 0..n {
+                                    out.push(mir.clone());
+                                }
+                                out.push(Mir::Loop(Box::new(mir)));
+                                Ok(Mir::Concat(out))
+                            }
+                            RepetitionRange::Bounded(n, m) => {
+                                let mut out = Vec::with_capacity(m as usize);
+                                for _ in 0..n {
+                                    out.push(mir.clone());
+                                }
+                                for _ in n..m {
+                                    out.push(Mir::Maybe(Box::new(mir.clone())));
+                                }
+                                Ok(Mir::Concat(out))
+                            }
+                        }
                     }
                 }
             }

diff --git a/tests/tests/lexer_modes.rs b/tests/tests/lexer_modes.rs
@@ -28,6 +28,9 @@ enum Inner {
     #[regex(r"\\u\{[^}]*\}")]
     EscapedCodepoint,
 
+    #[regex(r"\\[0-7]{1,3}")]
+    EscapedOctal,
+
     #[token(r#"\""#)]
     EscapedQuote,
 
@@ -105,7 +108,7 @@ impl<'source> Iterator for ModeBridge<'source> {
 fn iterating_modes() {
     use Inner::*;
     use Tokens::*;
-    let s = r#""Hello W\u{00f4}rld\n""#;
+    let s = r#""Hello W\u{00f4}\162ld\n""#;
     let moded = ModeBridge {
         mode: Modes::new(s),
     };
@@ -115,6 +118,7 @@ fn iterating_modes() {
         OuterToken(Outer::StartString),
         InnerToken(Text),
         InnerToken(EscapedCodepoint),
+        InnerToken(EscapedOctal),
         InnerToken(Text),
         InnerToken(EscapedNewline),
         InnerToken(EndString),