2 files changed, 102 insertions, 0 deletions

diff --git a/src/lexer/mod.rs b/src/lexer/mod.rs
new file mode 100644
index 000000000..136afb7b8
--- /dev/null
+++ b/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
+    }
+}
diff --git a/src/lexer/ptr.rs b/src/lexer/ptr.rs
new file mode 100644
index 000000000..4638dac21
--- /dev/null
+++ b/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()
+    }
+}