Adds newline counts within long strings for location metadata

src/lazy/text/buffer.rs

@@ -281,33 +281,39 @@ impl<'top> TextBuffer<'top> {
     /// Matches one or more whitespace characters.
     pub fn match_whitespace1(&mut self) -> IonMatchResult<'top> {
         let result = take_while(1.., WHITESPACE_BYTES).parse_next(self)?;
-        self.update_location_metadata(result);
+        self.update_location_metadata(result.data);
        Ok(*self)
     }
 
     /// Updates the location metadata based on the matched whitespace bytes in the consumed buffer
-    fn update_location_metadata(&mut self, result: TextBuffer<'top>) {
-        // If the bytes contain '\r\n' in this order then this must be coming from windows line ending pattern and hence should be counted as 1.
-        let crlf_count = result.data.windows(2).filter(|window| window == b"\r\n").count();
-
-        // Subtract the crlf_count from total count of all newline characters to get the correct number of newline match count.
-        let newline_match_count = result.data.iter().filter(|b| NEWLINE_BYTES.contains(b)).count() - crlf_count;
-
-        // Gets index for the last occurrence of the newline byte and subtracts from the result length to get non newline bytes length
-        let last_index_newline_byte = result.data.iter().rposition(|b| NEWLINE_BYTES.contains(b)).unwrap_or(0);
-        let non_newline_match_length = result.data.len() - last_index_newline_byte - 1;
-        self.row += newline_match_count;
+    fn update_location_metadata(&mut self, data: &'top [u8]) {
+        if !data.is_empty() {
+            // If the bytes contain '\r\n' in this order then this must be coming from windows line ending pattern and hence should be counted as 1.
+            let crlf_count = data.windows(2).filter(|window| window == b"\r\n").count();
+
+            // Subtract the crlf_count from total count of all newline characters to get the correct number of newline match count.
+            let newline_match_count = data.iter().filter(|b| NEWLINE_BYTES.contains(b)).count() - crlf_count;
+
+            // Gets index for the last occurrence of the newline byte and subtracts from the result length to get non newline bytes length
+            let last_index_newline_byte = data.iter().rposition(|b| NEWLINE_BYTES.contains(b)).unwrap_or(0);
+            let non_newline_match_length = data.len() - last_index_newline_byte - 1;
+            self.row += newline_match_count;
+
+            // Stores this newline offset as previous newline offset for calculating column position since this has already been matched/parsed
+            if self.offset < non_newline_match_length {
+                // this means that the input is not yet consumed hence get the input length + the current offset
+                self.prev_newline_offset = self.offset + data.len() - non_newline_match_length;
+            } else {
+                self.prev_newline_offset = self.offset - non_newline_match_length;
 
-        // Stores this newline offset as previous newline offset for calculating column position since this has already been matched/parsed
-        self.prev_newline_offset = self.offset - non_newline_match_length;
+            }
+        }
     }
 
     /// Matches zero or more whitespace characters.
     pub fn match_whitespace0(&mut self) -> IonMatchResult<'top> {
         let result = take_while(0.., WHITESPACE_BYTES).parse_next(self)?;
-        if !result.data.is_empty() {
-            self.update_location_metadata(result);
-        }
+        self.update_location_metadata(result.data);
         Ok(*self)
     }
 
@@ -1645,6 +1651,8 @@ impl<'top> TextBuffer<'top> {
         let delimiter_head = delimiter.as_bytes()[0];
         // Whether we've encountered any escapes while looking for the delimiter
         let mut contained_escapes = false;
+        // This input may contain newline characters hence update the location metadata.
+        self.update_location_metadata(self.bytes());
         // The input left to search
         let mut remaining = *self;
         loop {
@@ -2905,6 +2913,11 @@ mod tests {
         MatchTest::new_1_0(input).expect_match_location(match_length(TextBuffer::match_whitespace0), expected_location);
     }
 
+    #[test]
+    fn expect_newline_long_text() {
+        MatchTest::new_1_0("'''long \n\r\n\t hello'''").expect_match_location(match_length(TextBuffer::match_string), (3, 11));
+    }
+
     #[test]
     fn expect_foo() {
         MatchTest::new_1_0("\"hello\"").expect_match(match_length(TextBuffer::match_string));