Simple identifier lexer

Cargo.toml

@@ -4,6 +4,8 @@ version = "0.1.0"
 authors = ["Aleksey Kladov <aleksey.kladov@gmail.com>"]
 
 [dependencies]
+unicode-xid = "0.1.0"
+
 serde = "1.0.26"
 serde_derive = "1.0.26"
 file = "1.1.1"

src/lexer.rs

@@ -1,10 +0,0 @@
-use {Token, TextUnit};
-use syntax_kinds::*;
-
-pub fn next_token(text: &str) -> Token {
-    let c = text.chars().next().unwrap();
-    Token {
-        kind: IDENT,
-        len: TextUnit::len_of_char(c),
-    }
-}

src/lexer/mod.rs

@@ -0,0 +1,64 @@
+use unicode_xid::UnicodeXID;
+
+use {Token, SyntaxKind};
+use syntax_kinds::*;
+
+mod ptr;
+use self::ptr::Ptr;
+
+pub fn next_token(text: &str) -> Token {
+    assert!(!text.is_empty());
+    let mut ptr = Ptr::new(text);
+    let c = ptr.bump().unwrap();
+    let kind = next_token_inner(c, &mut ptr);
+    let len = ptr.into_len();
+    Token { kind, len }
+}
+
+fn next_token_inner(c: char, ptr: &mut Ptr) -> SyntaxKind {
+    // Note: r as in r" or r#" is part of a raw string literal,
+    // b as in b' is part of a byte literal.
+    // They are not identifiers, and are handled further down.
+    let ident_start = ident_start(c) && !string_literal_start(c, ptr.next(), ptr.nnext());
+    if ident_start {
+        loop {
+            match ptr.next() {
+                Some(c) if ident_continue(c) => {
+                    ptr.bump();
+                },
+                _ => break,
+            }
+        }
+        IDENT
+    } else {
+        WHITESPACE
+    }
+}
+
+fn ident_start(c: char) -> bool {
+    (c >= 'a' && c <= 'z')
+        || (c >= 'A' && c <= 'Z')
+        || c == '_'
+        || (c > '\x7f' && UnicodeXID::is_xid_start(c))
+}
+
+fn ident_continue(c: char) -> bool {
+    (c >= 'a' && c <= 'z')
+        || (c >= 'A' && c <= 'Z')
+        || (c >= '0' && c <= '9')
+        || c == '_'
+        || (c > '\x7f' && UnicodeXID::is_xid_continue(c))
+}
+
+
+fn string_literal_start(c: char, c1: Option<char>, c2: Option<char>) -> bool {
+    match (c, c1, c2) {
+        ('r', Some('"'), _) |
+        ('r', Some('#'), _) |
+        ('b', Some('"'), _) |
+        ('b', Some('\''), _) |
+        ('b', Some('r'), Some('"')) |
+        ('b', Some('r'), Some('#')) => true,
+        _ => false
+    }
+}

src/lexer/ptr.rs

@@ -0,0 +1,38 @@
+use {TextUnit};
+
+use std::str::Chars;
+
+pub(crate) struct Ptr<'s> {
+    text: &'s str,
+    len: TextUnit,
+}
+
+impl<'s> Ptr<'s> {
+    pub fn new(text: &'s str) -> Ptr<'s> {
+        Ptr { text, len: TextUnit::new(0) }
+    }
+
+    pub fn into_len(self) -> TextUnit {
+        self.len
+    }
+
+    pub fn next(&self) -> Option<char> {
+        self.chars().next()
+    }
+
+    pub fn nnext(&self) -> Option<char> {
+        let mut chars = self.chars();
+        chars.next()?;
+        chars.next()
+    }
+
+    pub fn bump(&mut self) -> Option<char> {
+        let ch = self.chars().next()?;
+        self.len += TextUnit::len_of_char(ch);
+        Some(ch)
+    }
+
+    fn chars(&self) -> Chars {
+        self.text[self.len.0 as usize ..].chars()
+    }
+}

src/lib.rs

@@ -1,3 +1,5 @@
+extern crate unicode_xid;
+
 mod text;
 mod tree;
 mod lexer;

src/text.rs

@@ -1,7 +1,10 @@
 use std::fmt;
+use std::ops;
 
 #[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash)]
-pub struct TextUnit(u32);
+pub struct TextUnit(
+    pub(crate) u32
+);
 
 impl TextUnit {
     pub fn len_of_char(c: char) -> TextUnit {
@@ -30,3 +33,29 @@ impl From<TextUnit> for u32 {
         tu.0
     }
 }
+
+impl ops::Add<TextUnit> for TextUnit {
+    type Output = TextUnit;
+    fn add(self, rhs: TextUnit) -> TextUnit {
+        TextUnit(self.0 + rhs.0)
+    }
+}
+
+impl ops::AddAssign<TextUnit> for TextUnit {
+    fn add_assign(&mut self, rhs: TextUnit) {
+        self.0 += rhs.0
+    }
+}
+
+impl ops::Sub<TextUnit> for TextUnit {
+    type Output = TextUnit;
+    fn sub(self, rhs: TextUnit) -> TextUnit {
+        TextUnit(self.0 - rhs.0)
+    }
+}
+
+impl ops::SubAssign<TextUnit> for TextUnit {
+    fn sub_assign(&mut self, rhs: TextUnit) {
+        self.0 -= rhs.0
+    }
+}

tests/data/lexer/0001_hello.txt

@@ -1,3 +1,3 @@
 IDENT 5
 WHITESPACE 1
-IDENT 5
+IDENT 5

tests/lexer.rs

@@ -41,13 +41,15 @@ fn lexer_test_case(path: &Path) {
         dump_tokens(&tokens)
     };
     let expected = file::get_text(&path.with_extension("txt")).unwrap();
-
+    let expected = expected.as_str();
-    assert_diff!(
+    let actual = actual.as_str();
-        expected.as_str(),
+    if expected == actual {
-        actual.as_str(),
+        return
-        "\n",
+    }
-        0
+    if expected.trim() == actual.trim() {
-    )
+        panic!("Whitespace difference!")
+    }
+    assert_diff!(expected, actual, "\n", 0)
 }
 
 fn tokenize(text: &str) -> Vec<Token> {