diff --git a/binding.gyp b/binding.gyp
index 74256d45..5d7349c6 100644
--- a/binding.gyp
+++ b/binding.gyp
@@ -9,7 +9,7 @@
       "sources": [
         "src/parser.c",
         "bindings/node/binding.cc",
-        "src/scanner.cc"
+        "src/scanner.c"
       ],
       "cflags_c": [
         "-std=c99",
diff --git a/bindings/rust/build.rs b/bindings/rust/build.rs
index 44501668..8851fed1 100644
--- a/bindings/rust/build.rs
+++ b/bindings/rust/build.rs
@@ -1,28 +1,19 @@
-use std::path::Path;
-extern crate cc;
-
 fn main() {
-    let src_dir = Path::new("src");
+    let src_dir = std::path::Path::new("src");
 
     let mut c_config = cc::Build::new();
-    c_config.include(&src_dir);
+    c_config.include(src_dir);
     c_config
         .flag_if_supported("-Wno-unused-parameter")
         .flag_if_supported("-Wno-unused-but-set-variable")
         .flag_if_supported("-Wno-trigraphs");
     let parser_path = src_dir.join("parser.c");
     c_config.file(&parser_path);
-    println!("cargo:rerun-if-changed={}", parser_path.to_str().unwrap());
-    c_config.compile("parser");
 
-    let mut cpp_config = cc::Build::new();
-    cpp_config.cpp(true);
-    cpp_config.include(&src_dir);
-    cpp_config
-        .flag_if_supported("-Wno-unused-parameter")
-        .flag_if_supported("-Wno-unused-but-set-variable");
-    let scanner_path = src_dir.join("scanner.cc");
-    cpp_config.file(&scanner_path);
+    let scanner_path = src_dir.join("scanner.c");
+    c_config.file(&scanner_path);
     println!("cargo:rerun-if-changed={}", scanner_path.to_str().unwrap());
-    cpp_config.compile("scanner");
+
+    c_config.compile("parser");
+    println!("cargo:rerun-if-changed={}", parser_path.to_str().unwrap());
 }
diff --git a/src/scanner.c b/src/scanner.c
new file mode 100644
index 00000000..40c5e2cf
--- /dev/null
+++ b/src/scanner.c
@@ -0,0 +1,485 @@
+#include <assert.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <string.h>
+#include <tree_sitter/parser.h>
+
+#define MAX(a, b) ((a) > (b) ? (a) : (b))
+
+#define VEC_RESIZE(vec, _cap)                                                  \
+    void *tmp = realloc((vec).data, (_cap) * sizeof((vec).data[0]));           \
+    assert(tmp != NULL);                                                       \
+    (vec).data = tmp;                                                          \
+    (vec).cap = (_cap);
+
+#define VEC_GROW(vec, _cap)                                                    \
+    if ((vec).cap < (_cap)) {                                                  \
+        VEC_RESIZE((vec), (_cap));                                             \
+    }
+
+#define VEC_PUSH(vec, el)                                                      \
+    if ((vec).cap == (vec).len) {                                              \
+        VEC_RESIZE((vec), MAX(16, (vec).len * 2));                             \
+    }                                                                          \
+    (vec).data[(vec).len++] = (el);
+
+#define VEC_POP(vec) (vec).len--;
+
+#define VEC_NEW                                                                \
+    { .len = 0, .cap = 0, .data = NULL }
+
+#define VEC_BACK(vec) ((vec).data[(vec).len - 1])
+
+#define VEC_FREE(vec)                                                          \
+    {                                                                          \
+        if ((vec).data != NULL)                                                \
+            free((vec).data);                                                  \
+    }
+
+#define VEC_CLEAR(vec) (vec).len = 0;
+
+enum TokenType {
+    NEWLINE,
+    INDENT,
+    DEDENT,
+    STRING_START,
+    STRING_CONTENT,
+    STRING_END,
+    COMMENT,
+    CLOSE_PAREN,
+    CLOSE_BRACKET,
+    CLOSE_BRACE,
+};
+
+typedef enum {
+    SingleQuote = 1 << 0,
+    DoubleQuote = 1 << 1,
+    BackQuote = 1 << 2,
+    Raw = 1 << 3,
+    Format = 1 << 4,
+    Triple = 1 << 5,
+    Bytes = 1 << 6,
+} Flags;
+
+typedef struct {
+    char flags;
+} Delimiter;
+
+static inline Delimiter new_delimiter() { return (Delimiter){0}; }
+
+static inline bool is_format(Delimiter *delimiter) {
+    return delimiter->flags & Format;
+}
+
+static inline bool is_raw(Delimiter *delimiter) {
+    return delimiter->flags & Raw;
+}
+
+static inline bool is_triple(Delimiter *delimiter) {
+    return delimiter->flags & Triple;
+}
+
+static inline bool is_bytes(Delimiter *delimiter) {
+    return delimiter->flags & Bytes;
+}
+
+static inline int32_t end_character(Delimiter *delimiter) {
+    if (delimiter->flags & SingleQuote) {
+        return '\'';
+    }
+    if (delimiter->flags & DoubleQuote) {
+        return '"';
+    }
+    if (delimiter->flags & BackQuote) {
+        return '`';
+    }
+    return 0;
+}
+
+static inline void set_format(Delimiter *delimiter) {
+    delimiter->flags |= Format;
+}
+
+static inline void set_raw(Delimiter *delimiter) { delimiter->flags |= Raw; }
+
+static inline void set_triple(Delimiter *delimiter) {
+    delimiter->flags |= Triple;
+}
+
+static inline void set_bytes(Delimiter *delimiter) {
+    delimiter->flags |= Bytes;
+}
+
+static inline void set_end_character(Delimiter *delimiter, int32_t character) {
+    switch (character) {
+        case '\'':
+            delimiter->flags |= SingleQuote;
+            break;
+        case '"':
+            delimiter->flags |= DoubleQuote;
+            break;
+        case '`':
+            delimiter->flags |= BackQuote;
+            break;
+        default:
+            assert(false);
+    }
+}
+
+static inline const char *delimiter_string(Delimiter *delimiter) {
+    if (delimiter->flags & SingleQuote) {
+        return "\'";
+    }
+    if (delimiter->flags & DoubleQuote) {
+        return "\"";
+    }
+    if (delimiter->flags & BackQuote) {
+        return "`";
+    }
+    return "";
+}
+
+typedef struct {
+    uint32_t len;
+    uint32_t cap;
+    uint16_t *data;
+} indent_vec;
+
+indent_vec indent_vec_new() {
+    indent_vec vec = VEC_NEW;
+    vec.data = calloc(1, sizeof(uint16_t));
+    vec.cap = 1;
+    return vec;
+}
+
+typedef struct {
+    uint32_t len;
+    uint32_t cap;
+    Delimiter *data;
+} delimiter_vec;
+
+delimiter_vec delimiter_vec_new() {
+    delimiter_vec vec = VEC_NEW;
+    vec.data = calloc(1, sizeof(Delimiter));
+    vec.cap = 1;
+    return vec;
+}
+
+typedef struct {
+    indent_vec indents;
+    delimiter_vec delimiters;
+} Scanner;
+
+static inline void advance(TSLexer *lexer) { lexer->advance(lexer, false); }
+
+static inline void skip(TSLexer *lexer) { lexer->advance(lexer, true); }
+
+bool tree_sitter_python_external_scanner_scan(void *payload, TSLexer *lexer,
+                                              const bool *valid_symbols) {
+    Scanner *scanner = (Scanner *)payload;
+
+    bool error_recovery_mode =
+        valid_symbols[STRING_CONTENT] && valid_symbols[INDENT];
+    bool within_brackets = valid_symbols[CLOSE_BRACE] ||
+                           valid_symbols[CLOSE_PAREN] ||
+                           valid_symbols[CLOSE_BRACKET];
+
+    if (valid_symbols[STRING_CONTENT] && scanner->delimiters.len > 0 &&
+        !error_recovery_mode) {
+        Delimiter delimiter = VEC_BACK(scanner->delimiters);
+        int32_t end_char = end_character(&delimiter);
+        bool has_content = false;
+        while (lexer->lookahead) {
+            if ((lexer->lookahead == '{' || lexer->lookahead == '}') &&
+                is_format(&delimiter)) {
+                lexer->mark_end(lexer);
+                lexer->result_symbol = STRING_CONTENT;
+                return has_content;
+            }
+            if (lexer->lookahead == '\\') {
+                if (is_raw(&delimiter)) {
+                    // Step over the backslash.
+                    lexer->advance(lexer, false);
+                    // Step over any escaped quotes.
+                    if (lexer->lookahead == end_character(&delimiter) ||
+                        lexer->lookahead == '\\') {
+                        lexer->advance(lexer, false);
+                    }
+                    continue;
+                }
+                if (is_bytes(&delimiter)) {
+                    lexer->mark_end(lexer);
+                    lexer->advance(lexer, false);
+                    if (lexer->lookahead == 'N' || lexer->lookahead == 'u' ||
+                        lexer->lookahead == 'U') {
+                        // In bytes string, \N{...}, \uXXXX and \UXXXXXXXX are
+                        // not escape sequences
+                        // https://docs.python.org/3/reference/lexical_analysis.html#string-and-bytes-literals
+                        lexer->advance(lexer, false);
+                    } else {
+                        lexer->result_symbol = STRING_CONTENT;
+                        return has_content;
+                    }
+                } else {
+                    lexer->mark_end(lexer);
+                    lexer->result_symbol = STRING_CONTENT;
+                    return has_content;
+                }
+            } else if (lexer->lookahead == end_char) {
+                if (is_triple(&delimiter)) {
+                    lexer->mark_end(lexer);
+                    lexer->advance(lexer, false);
+                    if (lexer->lookahead == end_char) {
+                        lexer->advance(lexer, false);
+                        if (lexer->lookahead == end_char) {
+                            if (has_content) {
+                                lexer->result_symbol = STRING_CONTENT;
+                            } else {
+                                lexer->advance(lexer, false);
+                                lexer->mark_end(lexer);
+                                VEC_POP(scanner->delimiters);
+                                lexer->result_symbol = STRING_END;
+                            }
+                            return true;
+                        }
+                        lexer->mark_end(lexer);
+                        lexer->result_symbol = STRING_CONTENT;
+                        return true;
+                    }
+                    lexer->mark_end(lexer);
+                    lexer->result_symbol = STRING_CONTENT;
+                    return true;
+                }
+                if (has_content) {
+                    lexer->result_symbol = STRING_CONTENT;
+                } else {
+                    lexer->advance(lexer, false);
+                    VEC_POP(scanner->delimiters);
+                    lexer->result_symbol = STRING_END;
+                }
+                lexer->mark_end(lexer);
+                return true;
+
+            } else if (lexer->lookahead == '\n' && has_content &&
+                       !is_triple(&delimiter)) {
+                return false;
+            }
+            advance(lexer);
+            has_content = true;
+        }
+    }
+
+    lexer->mark_end(lexer);
+
+    bool found_end_of_line = false;
+    uint32_t indent_length = 0;
+    int32_t first_comment_indent_length = -1;
+    for (;;) {
+        if (lexer->lookahead == '\n') {
+            found_end_of_line = true;
+            indent_length = 0;
+            skip(lexer);
+        } else if (lexer->lookahead == ' ') {
+            indent_length++;
+            skip(lexer);
+        } else if (lexer->lookahead == '\r' || lexer->lookahead == '\f') {
+            indent_length = 0;
+            skip(lexer);
+        } else if (lexer->lookahead == '\t') {
+            indent_length += 8;
+            skip(lexer);
+        } else if (lexer->lookahead == '#') {
+            if (first_comment_indent_length == -1) {
+                first_comment_indent_length = (int32_t)indent_length;
+            }
+            while (lexer->lookahead && lexer->lookahead != '\n') {
+                skip(lexer);
+            }
+            skip(lexer);
+            indent_length = 0;
+        } else if (lexer->lookahead == '\\') {
+            skip(lexer);
+            if (lexer->lookahead == '\r') {
+                skip(lexer);
+            }
+            if (lexer->lookahead == '\n' || lexer->eof(lexer)) {
+                skip(lexer);
+            } else {
+                return false;
+            }
+        } else if (lexer->eof(lexer)) {
+            indent_length = 0;
+            found_end_of_line = true;
+            break;
+        } else {
+            break;
+        }
+    }
+
+    if (found_end_of_line) {
+        if (scanner->indents.len > 0) {
+            uint16_t current_indent_length = VEC_BACK(scanner->indents);
+
+            if (valid_symbols[INDENT] &&
+                indent_length > current_indent_length) {
+                VEC_PUSH(scanner->indents, indent_length);
+                lexer->result_symbol = INDENT;
+                return true;
+            }
+
+            if ((valid_symbols[DEDENT] ||
+                 (!valid_symbols[NEWLINE] && !valid_symbols[STRING_START] &&
+                  !within_brackets)) &&
+                indent_length < current_indent_length &&
+
+                // Wait to create a dedent token until we've consumed any
+                // comments
+                // whose indentation matches the current block.
+                first_comment_indent_length < (int32_t)current_indent_length) {
+                VEC_POP(scanner->indents);
+                lexer->result_symbol = DEDENT;
+                return true;
+            }
+        }
+
+        if (valid_symbols[NEWLINE] && !error_recovery_mode) {
+            lexer->result_symbol = NEWLINE;
+            return true;
+        }
+    }
+
+    if (first_comment_indent_length == -1 && valid_symbols[STRING_START]) {
+        Delimiter delimiter = new_delimiter();
+
+        bool has_flags = false;
+        while (lexer->lookahead) {
+            if (lexer->lookahead == 'f' || lexer->lookahead == 'F') {
+                set_format(&delimiter);
+            } else if (lexer->lookahead == 'r' || lexer->lookahead == 'R') {
+                set_raw(&delimiter);
+            } else if (lexer->lookahead == 'b' || lexer->lookahead == 'B') {
+                set_bytes(&delimiter);
+            } else if (lexer->lookahead != 'u' && lexer->lookahead != 'U') {
+                break;
+            }
+            has_flags = true;
+            advance(lexer);
+        }
+
+        if (lexer->lookahead == '`') {
+            set_end_character(&delimiter, '`');
+            advance(lexer);
+            lexer->mark_end(lexer);
+        } else if (lexer->lookahead == '\'') {
+            set_end_character(&delimiter, '\'');
+            advance(lexer);
+            lexer->mark_end(lexer);
+            if (lexer->lookahead == '\'') {
+                advance(lexer);
+                if (lexer->lookahead == '\'') {
+                    advance(lexer);
+                    lexer->mark_end(lexer);
+                    set_triple(&delimiter);
+                }
+            }
+        } else if (lexer->lookahead == '"') {
+            set_end_character(&delimiter, '"');
+            advance(lexer);
+            lexer->mark_end(lexer);
+            if (lexer->lookahead == '"') {
+                advance(lexer);
+                if (lexer->lookahead == '"') {
+                    advance(lexer);
+                    lexer->mark_end(lexer);
+                    set_triple(&delimiter);
+                }
+            }
+        }
+
+        if (end_character(&delimiter)) {
+            VEC_PUSH(scanner->delimiters, delimiter);
+            lexer->result_symbol = STRING_START;
+
+            return true;
+        }
+        if (has_flags) {
+            return false;
+        }
+    }
+
+    return false;
+}
+
+unsigned tree_sitter_python_external_scanner_serialize(void *payload,
+                                                       char *buffer) {
+    Scanner *scanner = (Scanner *)payload;
+
+    size_t size = 0;
+
+    size_t delimiter_count = scanner->delimiters.len;
+    if (delimiter_count > UINT8_MAX) {
+        delimiter_count = UINT8_MAX;
+    }
+    buffer[size++] = (char)delimiter_count;
+
+    if (delimiter_count > 0) {
+        memcpy(&buffer[size], scanner->delimiters.data, delimiter_count);
+    }
+    size += delimiter_count;
+
+    int iter = 1;
+    for (; iter < scanner->indents.len &&
+           size < TREE_SITTER_SERIALIZATION_BUFFER_SIZE;
+         ++iter) {
+        // yeah, it narrows the value but it's fine?
+        buffer[size++] = (char)scanner->indents.data[iter];
+    }
+
+    return size;
+}
+
+void tree_sitter_python_external_scanner_deserialize(void *payload,
+                                                     const char *buffer,
+                                                     unsigned length) {
+    Scanner *scanner = (Scanner *)payload;
+
+    VEC_CLEAR(scanner->delimiters);
+    VEC_CLEAR(scanner->indents);
+    VEC_PUSH(scanner->indents, 0);
+
+    if (length > 0) {
+        size_t size = 0;
+
+        size_t delimiter_count = (uint8_t)buffer[size++];
+        if (delimiter_count > 0) {
+            VEC_GROW(scanner->delimiters, delimiter_count);
+            scanner->delimiters.len = delimiter_count;
+            memcpy(scanner->delimiters.data, &buffer[size], delimiter_count);
+            size += delimiter_count;
+        }
+
+        for (; size < length; size++) {
+            VEC_PUSH(scanner->indents, (unsigned char)buffer[size]);
+        }
+    }
+}
+
+void *tree_sitter_python_external_scanner_create() {
+#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)
+    static_assert(sizeof(Delimiter) == sizeof(char), "");
+#else
+    assert(sizeof(Delimiter) == sizeof(char));
+#endif
+    Scanner *scanner = calloc(1, sizeof(Scanner));
+    scanner->indents = indent_vec_new();
+    scanner->delimiters = delimiter_vec_new();
+    tree_sitter_python_external_scanner_deserialize(scanner, NULL, 0);
+    return scanner;
+}
+
+void tree_sitter_python_external_scanner_destroy(void *payload) {
+    Scanner *scanner = (Scanner *)payload;
+    VEC_FREE(scanner->indents);
+    VEC_FREE(scanner->delimiters);
+    free(scanner);
+}
diff --git a/src/scanner.cc b/src/scanner.cc
deleted file mode 100644
index e9ba26b3..00000000
--- a/src/scanner.cc
+++ /dev/null
@@ -1,414 +0,0 @@
-#include <tree_sitter/parser.h>
-#include <vector>
-#include <cwctype>
-#include <cstring>
-#include <cassert>
-#include <stdio.h>
-namespace {
-
-using std::vector;
-using std::iswspace;
-using std::memcpy;
-
-enum TokenType {
-  NEWLINE,
-  INDENT,
-  DEDENT,
-  STRING_START,
-  STRING_CONTENT,
-  STRING_END,
-  COMMENT,
-  CLOSE_PAREN,
-  CLOSE_BRACKET,
-  CLOSE_BRACE,
-};
-
-struct Delimiter {
-  enum {
-    SingleQuote = 1 << 0,
-    DoubleQuote = 1 << 1,
-    BackQuote = 1 << 2,
-    Raw = 1 << 3,
-    Format = 1 << 4,
-    Triple = 1 << 5,
-    Bytes = 1 << 6,
-  };
-
-  Delimiter() : flags(0) {}
-
-  bool is_format() const {
-    return flags & Format;
-  }
-
-  bool is_raw() const {
-    return flags & Raw;
-  }
-
-  bool is_triple() const {
-    return flags & Triple;
-  }
-
-  bool is_bytes() const {
-    return flags & Bytes;
-  }
-
-  int32_t end_character() const {
-    if (flags & SingleQuote) return '\'';
-    if (flags & DoubleQuote) return '"';
-    if (flags & BackQuote) return '`';
-    return 0;
-  }
-
-  void set_format() {
-    flags |= Format;
-  }
-
-  void set_raw() {
-    flags |= Raw;
-  }
-
-  void set_triple() {
-    flags |= Triple;
-  }
-
-  void set_bytes() {
-    flags |= Bytes;
-  }
-
-  void set_end_character(int32_t character) {
-    switch (character) {
-      case '\'':
-        flags |= SingleQuote;
-        break;
-      case '"':
-        flags |= DoubleQuote;
-        break;
-      case '`':
-        flags |= BackQuote;
-        break;
-      default:
-        assert(false);
-    }
-  }
-
-  char flags;
-};
-
-struct Scanner {
-  Scanner() {
-    assert(sizeof(Delimiter) == sizeof(char));
-    deserialize(NULL, 0);
-  }
-
-  unsigned serialize(char *buffer) {
-    size_t i = 0;
-
-    size_t delimiter_count = delimiter_stack.size();
-    if (delimiter_count > UINT8_MAX) delimiter_count = UINT8_MAX;
-    buffer[i++] = delimiter_count;
-
-    if (delimiter_count > 0) {
-      memcpy(&buffer[i], delimiter_stack.data(), delimiter_count);
-    }
-    i += delimiter_count;
-
-    vector<uint16_t>::iterator
-      iter = indent_length_stack.begin() + 1,
-      end = indent_length_stack.end();
-
-    for (; iter != end && i < TREE_SITTER_SERIALIZATION_BUFFER_SIZE; ++iter) {
-      buffer[i++] = *iter;
-    }
-
-    return i;
-  }
-
-  void deserialize(const char *buffer, unsigned length) {
-    delimiter_stack.clear();
-    indent_length_stack.clear();
-    indent_length_stack.push_back(0);
-
-    if (length > 0) {
-      size_t i = 0;
-
-      size_t delimiter_count = (uint8_t)buffer[i++];
-      delimiter_stack.resize(delimiter_count);
-      if (delimiter_count > 0) {
-        memcpy(delimiter_stack.data(), &buffer[i], delimiter_count);
-      }
-      i += delimiter_count;
-
-      for (; i < length; i++) {
-        indent_length_stack.push_back(buffer[i]);
-      }
-    }
-  }
-
-  void advance(TSLexer *lexer) {
-    lexer->advance(lexer, false);
-  }
-
-  void skip(TSLexer *lexer) {
-    lexer->advance(lexer, true);
-  }
-
-  bool scan(TSLexer *lexer, const bool *valid_symbols) {
-    bool error_recovery_mode = valid_symbols[STRING_CONTENT] && valid_symbols[INDENT];
-    bool within_brackets = valid_symbols[CLOSE_BRACE] || valid_symbols[CLOSE_PAREN] || valid_symbols[CLOSE_BRACKET];
-    
-    if (valid_symbols[STRING_CONTENT] && !delimiter_stack.empty() && !error_recovery_mode) {
-      Delimiter delimiter = delimiter_stack.back();
-      int32_t end_character = delimiter.end_character();
-      bool has_content = false;
-      while (lexer->lookahead) {
-        if ((lexer->lookahead == '{' || lexer->lookahead == '}') && delimiter.is_format()) {
-          lexer->mark_end(lexer);
-          lexer->result_symbol = STRING_CONTENT;
-          return has_content;
-        } else if (lexer->lookahead == '\\') {
-          if (delimiter.is_raw()) {
-            // Step over the backslash.
-            lexer->advance(lexer, false);
-            // Step over any escaped quotes.
-            if (lexer->lookahead == delimiter.end_character() || lexer->lookahead == '\\') {
-              lexer->advance(lexer, false);
-            }
-            continue;
-          } else if (delimiter.is_bytes()) {
-              lexer->mark_end(lexer);
-              lexer->advance(lexer, false);
-              if (lexer->lookahead == 'N' || lexer->lookahead == 'u' || lexer->lookahead == 'U') {
-                // In bytes string, \N{...}, \uXXXX and \UXXXXXXXX are not escape sequences
-                // https://docs.python.org/3/reference/lexical_analysis.html#string-and-bytes-literals
-                lexer->advance(lexer, false);
-              } else {
-                  lexer->result_symbol = STRING_CONTENT;
-                  return has_content;
-              }
-          } else {
-            lexer->mark_end(lexer);
-            lexer->result_symbol = STRING_CONTENT;
-            return has_content;
-          }
-        } else if (lexer->lookahead == end_character) {
-          if (delimiter.is_triple()) {
-            lexer->mark_end(lexer);
-            lexer->advance(lexer, false);
-            if (lexer->lookahead == end_character) {
-              lexer->advance(lexer, false);
-              if (lexer->lookahead == end_character) {
-                if (has_content) {
-                  lexer->result_symbol = STRING_CONTENT;
-                } else {
-                  lexer->advance(lexer, false);
-                  lexer->mark_end(lexer);
-                  delimiter_stack.pop_back();
-                  lexer->result_symbol = STRING_END;
-                }
-                return true;
-              } else {
-                lexer->mark_end(lexer);
-                lexer->result_symbol = STRING_CONTENT;
-                return true;
-              }
-            } else {
-              lexer->mark_end(lexer);
-              lexer->result_symbol = STRING_CONTENT;
-              return true;
-            }
-          } else {
-            if (has_content) {
-              lexer->result_symbol = STRING_CONTENT;
-            } else {
-              lexer->advance(lexer, false);
-              delimiter_stack.pop_back();
-              lexer->result_symbol = STRING_END;
-            }
-            lexer->mark_end(lexer);
-            return true;
-          }
-        } else if (lexer->lookahead == '\n' && has_content && !delimiter.is_triple()) {
-          return false;
-        }
-        advance(lexer);
-        has_content = true;
-      }
-    }
-
-    lexer->mark_end(lexer);
-
-    bool found_end_of_line = false;
-    uint32_t indent_length = 0;
-    int32_t first_comment_indent_length = -1;
-    for (;;) {
-      if (lexer->lookahead == '\n') {
-        found_end_of_line = true;
-        indent_length = 0;
-        skip(lexer);
-      } else if (lexer->lookahead == ' ') {
-        indent_length++;
-        skip(lexer);
-      } else if (lexer->lookahead == '\r') {
-        indent_length = 0;
-        skip(lexer);
-      } else if (lexer->lookahead == '\t') {
-        indent_length += 8;
-        skip(lexer);
-      } else if (lexer->lookahead == '#') {
-        if (first_comment_indent_length == -1) {
-          first_comment_indent_length = (int32_t)indent_length;
-        }
-        while (lexer->lookahead && lexer->lookahead != '\n') {
-          skip(lexer);
-        }
-        skip(lexer);
-        indent_length = 0;
-      } else if (lexer->lookahead == '\\') {
-        skip(lexer);
-        if (lexer->lookahead == '\r') {
-          skip(lexer);
-        }
-        if (lexer->lookahead == '\n') {
-          skip(lexer);
-        } else {
-          return false;
-        }
-      } else if (lexer->lookahead == '\f') {
-        indent_length = 0;
-        skip(lexer);
-      } else if (lexer->lookahead == 0) {
-        indent_length = 0;
-        found_end_of_line = true;
-        break;
-      } else {
-        break;
-      }
-    }
-
-    if (found_end_of_line) {
-      if (!indent_length_stack.empty()) {
-        uint16_t current_indent_length = indent_length_stack.back();
-
-        if (
-          valid_symbols[INDENT] &&
-          indent_length > current_indent_length
-        ) {
-          indent_length_stack.push_back(indent_length);
-          lexer->result_symbol = INDENT;
-          return true;
-        }
-
-        if (
-          (valid_symbols[DEDENT] || (!valid_symbols[NEWLINE] && !within_brackets)) &&
-          indent_length < current_indent_length &&
-
-          // Wait to create a dedent token until we've consumed any comments
-          // whose indentation matches the current block.
-          first_comment_indent_length < (int32_t)current_indent_length
-        ) {
-          indent_length_stack.pop_back();
-          lexer->result_symbol = DEDENT;
-          return true;
-        }
-      }
-
-      if (valid_symbols[NEWLINE] && !error_recovery_mode) {
-        lexer->result_symbol = NEWLINE;
-        return true;
-      }
-    }
-
-    if (first_comment_indent_length == -1 && valid_symbols[STRING_START]) {
-      Delimiter delimiter;
-
-      bool has_flags = false;
-      while (lexer->lookahead) {
-        if (lexer->lookahead == 'f' || lexer->lookahead == 'F') {
-          delimiter.set_format();
-        } else if (lexer->lookahead == 'r' || lexer->lookahead == 'R') {
-          delimiter.set_raw();
-        } else if (lexer->lookahead == 'b' || lexer->lookahead == 'B') {
-          delimiter.set_bytes();
-        } else if (lexer->lookahead != 'u' && lexer->lookahead != 'U') {
-          break;
-        }
-        has_flags = true;
-        advance(lexer);
-      }
-
-      if (lexer->lookahead == '`') {
-        delimiter.set_end_character('`');
-        advance(lexer);
-        lexer->mark_end(lexer);
-      } else if (lexer->lookahead == '\'') {
-        delimiter.set_end_character('\'');
-        advance(lexer);
-        lexer->mark_end(lexer);
-        if (lexer->lookahead == '\'') {
-          advance(lexer);
-          if (lexer->lookahead == '\'') {
-            advance(lexer);
-            lexer->mark_end(lexer);
-            delimiter.set_triple();
-          }
-        }
-      } else if (lexer->lookahead == '"') {
-        delimiter.set_end_character('"');
-        advance(lexer);
-        lexer->mark_end(lexer);
-        if (lexer->lookahead == '"') {
-          advance(lexer);
-          if (lexer->lookahead == '"') {
-            advance(lexer);
-            lexer->mark_end(lexer);
-            delimiter.set_triple();
-          }
-        }
-      }
-
-      if (delimiter.end_character()) {
-        delimiter_stack.push_back(delimiter);
-        lexer->result_symbol = STRING_START;
-        return true;
-      } else if (has_flags) {
-        return false;
-      }
-    }
-
-    return false;
-  }
-
-  vector<uint16_t> indent_length_stack;
-  vector<Delimiter> delimiter_stack;
-};
-
-}
-
-extern "C" {
-
-void *tree_sitter_python_external_scanner_create() {
-  return new Scanner();
-}
-
-bool tree_sitter_python_external_scanner_scan(void *payload, TSLexer *lexer,
-                                            const bool *valid_symbols) {
-  Scanner *scanner = static_cast<Scanner *>(payload);
-  return scanner->scan(lexer, valid_symbols);
-}
-
-unsigned tree_sitter_python_external_scanner_serialize(void *payload, char *buffer) {
-  Scanner *scanner = static_cast<Scanner *>(payload);
-  return scanner->serialize(buffer);
-}
-
-void tree_sitter_python_external_scanner_deserialize(void *payload, const char *buffer, unsigned length) {
-  Scanner *scanner = static_cast<Scanner *>(payload);
-  scanner->deserialize(buffer, length);
-}
-
-void tree_sitter_python_external_scanner_destroy(void *payload) {
-  Scanner *scanner = static_cast<Scanner *>(payload);
-  delete scanner;
-}
-
-}