clean up reader interface

1 files changed, 13 insertions, 28 deletions

diff --git a/src/reader.rs b/src/reader.rs
index d2cae92..7246129 100644
--- a/src/reader.rs
+++ b/src/reader.rs
@@ -4,9 +4,8 @@
 pub trait Reader {
     /// Return a new character from the input stream.
     ///
-    /// Newlines have to be normalized as described in [Preprocessing the input
-    /// stream](https://html.spec.whatwg.org/#preprocessing-the-input-stream), however error
-    /// emission is done within the tokenizer.
+    /// The input stream does **not** have to be preprocessed in any way, it can contain standalone
+    /// surrogates and have inconsistent newlines.
     fn read_char(&mut self) -> Option<char>;
 
     /// Attempt to read an entire string at once, either case-insensitively or not.
@@ -46,49 +45,35 @@ impl<'a, R: 'a + Reader> Readable<'a> for R {
 /// from strings.
 pub struct StringReader<'a> {
     input: &'a str,
+    cursor: std::str::Chars<'a>,
     pos: usize,
 }
 
 impl<'a> StringReader<'a> {
     fn new(input: &'a str) -> Self {
-        StringReader { input, pos: 0 }
-    }
-
-    fn peek_char(&self) -> Option<char> {
-        self.input.get(self.pos..)?.chars().next()
+        let cursor = input.chars();
+        StringReader {
+            input,
+            cursor,
+            pos: 0,
+        }
     }
 }
 
 impl<'a> Reader for StringReader<'a> {
     fn read_char(&mut self) -> Option<char> {
-        let mut r1 = match self.peek_char() {
-            Some(x) => x,
-            None => {
-                self.pos += 1;
-                return None;
-            }
-        };
-
-        self.pos += r1.len_utf8();
-
-        if r1 == '\r' {
-            r1 = '\n';
-            let r2 = self.peek_char();
-            if r2 == Some('\n') {
-                self.pos += r2.map(char::len_utf8).unwrap_or(0);
-            }
-        }
-
-        Some(r1)
+        let c = self.cursor.next()?;
+        self.pos += c.len_utf8();
+        Some(c)
     }
 
     fn try_read_string(&mut self, s1: &str, case_sensitive: bool) -> bool {
         // we do not need to call validate_char here because `s` hopefully does not contain invalid
         // characters
-
         if let Some(s2) = self.input.get(self.pos..self.pos + s1.len()) {
             if s1 == s2 || (!case_sensitive && s1.eq_ignore_ascii_case(s2)) {
                 self.pos += s1.len();
+                self.cursor = self.input[self.pos..].chars();
                 return true;
             }
         }