domus

commit 62301a1218328c7c1933f88d641ec94d5ec90dbd
parent d591c694dd3cb93112029c43cf2ee574a1417351
Author: William Casarin <jb55@jb55.com>
Date:   Fri, 30 Jun 2023 16:30:48 -0700

parser: add fast utf8 parsing

Diffstat:
M
src/parser.rs
 | 
136
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-----------

1 file changed, 118 insertions(+), 18 deletions(-)
diff --git a/src/parser.rs b/src/parser.rs
@@ -1,27 +1,109 @@
 use log::info;
 
+#[derive(Debug, PartialEq)]
 struct Parser<'a> {
-    data: &'a str,
+    data: &'a [u8],
     pos: usize,
 }
 
+#[derive(Debug, PartialEq)]
+enum ParseError {
+    NotFound,
+    BadUtf8Encoding,
+    EOF,
+}
+
+type Result<T> = std::result::Result<T, ParseError>;
+
 impl<'a> Parser<'a> {
-    fn new(data: &'a str) -> Parser {
+    fn new(data: &'a [u8]) -> Parser {
         Parser { data: data, pos: 0 }
     }
 
-    fn parse_until(&mut self, needle: char) -> bool {
-        let mut count = 0;
-        for c in self.data[self.pos..].chars() {
-            if c == needle {
-                self.pos += count - 1;
-                return true;
-            } else {
-                count += 1;
+    fn pull_byte(&mut self) -> Result<u8> {
+        if self.pos + 1 > self.data.len() {
+            return Err(ParseError::EOF);
+        }
+
+        let c = self.data[self.pos];
+        self.pos += 1;
+        return Ok(c);
+    }
+
+    pub fn peek_char(&mut self) -> Result<char> {
+        let peek = true;
+        self.pull_or_peek_char(peek)
+    }
+
+    pub fn pull_char(&mut self) -> Result<char> {
+        let peek = false;
+        self.pull_or_peek_char(peek)
+    }
+
+    fn pull_or_peek_char(&mut self, peek: bool) -> Result<char> {
+        let mut codepoint: u32 = 0;
+
+        let start = self.pos;
+        let b0 = self.pull_byte()? as u32;
+
+        if b0 & 0x80 != 0 {
+            if (b0 & 0xE0) == 0xC0 {
+                // Two-byte sequence
+                let b1 = self.pull_byte()? as u32;
+                codepoint = ((b0 & 0x1F) << 6) | (b1 & 0x3F);
+            } else if (b0 & 0xF0) == 0xE0 {
+                // Three-byte sequence
+                let b1 = self.pull_byte()? as u32;
+                let b2 = self.pull_byte()? as u32;
+                codepoint = ((b0 & 0x0F) << 12) | ((b1 & 0x3F) << 6) | (b2 & 0x3F);
+            } else if (b0 & 0xF8) == 0xF0 {
+                // Four-byte sequence
+                let b1 = self.pull_byte()? as u32;
+                let b2 = self.pull_byte()? as u32;
+                let b3 = self.pull_byte()? as u32;
+                codepoint =
+                    ((b0 & 0x07) << 18) | ((b1 & 0x3F) << 12) | ((b2 & 0x3F) << 6) | (b3 & 0x3F);
             }
+        } else {
+            // Single-byte ASCII character
+            codepoint = b0;
+        }
+
+        if peek {
+            self.pos = start;
         }
 
-        return false;
+        match std::char::from_u32(codepoint) {
+            Some(c) => Ok(c),
+            None => Err(ParseError::BadUtf8Encoding),
+        }
+    }
+
+    fn current(&mut self) -> Result<char> {
+        let last_pos = self.pos;
+        let c = self.pull_char();
+        if c.is_ok() {
+            self.pos = last_pos;
+        }
+        return c;
+    }
+
+    fn parse_until_char(&mut self, needle: char) -> Result<()> {
+        self.parse_until(|c| c == needle)?;
+        Ok(())
+    }
+
+    fn parse_until<F: Fn(char) -> bool>(&mut self, matches: F) -> Result<()> {
+        let len = self.data.len();
+        while self.pos < len {
+            let prev_pos = self.pos;
+            if matches(self.pull_char()?) {
+                self.pos = prev_pos;
+                return Ok(());
+            }
+        }
+
+        Err(ParseError::NotFound)
     }
 }
 
@@ -30,12 +112,30 @@ mod test {
     use super::*;
 
     #[test]
-    fn test_parser() {
-        let s = "hey there #hashtag";
-        let mut parser = Parser::new(s);
-        parser.parse_until('#');
-        assert_eq!(parser.pos, 9);
-        parser.parse_until('t');
-        assert_eq!(parser.pos, 14);
+    fn test_parser() -> Result<()> {
+        //             v alien  v
+        // 00000000: 20f0 9f91 bd23 6861 7368 7461 670a       _....#hashtag.
+        let s = " #hashtag ";
+        let mut parser = Parser::new(s.as_bytes());
+        let mut res = parser.parse_until_char('#');
+        assert_eq!(res, Ok(()));
+        assert_eq!(parser.pos, 1);
+        res = parser.parse_until_char('t');
+        assert_eq!(res, Ok(()));
+        assert_eq!(parser.pos, 6);
+        Ok(())
+    }
+
+    #[test]
+    fn test_utf8_parsing() -> Result<()> {
+        let s = "hey there #👽.";
+        let mut parser = Parser::new(s.as_bytes());
+        let _ = parser.parse_until_char('👽');
+        assert_eq!(parser.current(), Ok('👽'));
+        assert_eq!(parser.pos, 11);
+        let res = parser.parse_until(|c| c.is_ascii_whitespace() || c.is_ascii_punctuation());
+        assert_eq!(res, Ok(()));
+        assert_eq!(parser.current(), Ok('.'));
+        Ok(())
     }
 }