Scanner for UTF-8 byte arrays

sdk/lib/_internal/compiler/implementation/scanner/utf8_bytes_scanner.dart

Index: sdk/lib/_internal/compiler/implementation/scanner/utf8_bytes_scanner.dart
diff --git a/sdk/lib/_internal/compiler/implementation/scanner/utf8_bytes_scanner.dart b/sdk/lib/_internal/compiler/implementation/scanner/utf8_bytes_scanner.dart
new file mode 100644
index 0000000000000000000000000000000000000000..7e7cdf713146df5f0fb869c99f7b68a23b784159
--- /dev/null
+++ b/sdk/lib/_internal/compiler/implementation/scanner/utf8_bytes_scanner.dart
@@ -0,0 +1,194 @@
+// Copyright (c) 2011, the Dart project authors.  Please see the AUTHORS file

[ngeoffray, 2013/10/18 10:19:37]

2011 -> 2013

[lukas, 2013/10/24 16:48:36]

Done.

+// for details. All rights reserved. Use of this source code is governed by a
+// BSD-style license that can be found in the LICENSE file.
+
+part of scanner;
+
+/**
+ * Scanner that reads from a UTF-8 encoded list of bytes and creates tokens
+ * that points to substrings.
+ */
+class Utf8BytesScanner extends ArrayBasedScanner {
+  /** The file content. */
+  List<int> bytes;
+
+  /**
+   * Points to the offset of the byte last returned by [advance].

[ngeoffray, 2013/10/18 10:19:37]

byte last -> last byte?

[lukas, 2013/10/24 16:48:36]

Done.

+   *
+   * After invoking [currentAsUnicode], the [byteOffset] points to the last
+   * byte that is part of the (unicode or ASCII) character. That way, [advance]
+   * can always increase the byte offset by 1.
+   */
+  int byteOffset = -1;
+
+  /**
+   * The getter [scanOffset] is expected to return the index where the current
+   * character *starts*. In case of a non-ascii character, after invoking
+   * [currentAsUnicode], the byte offset points to the *last* byte.
+   *
+   * This field keeps track of the number of bytes for the current unicode
+   * character. For example, if bytes 7,8,9 encode one unicode character, the
+   * [byteOffset] is 9 (after invoking [currentAsUnicode]). The [scanSlack]
+   * will be 2, so that [scanOffset] returns 7.
+   */
+  int scanSlack = 0;
+
+  /**
+   * Holds the [byteOffset] value for which the current [scanSlack] is valid.
+   */
+  int scanSlackOffset = -1;
+
+  /**
+   * Returns the byte offset of the first byte that belongs to the current
+   * character.
+   */
+  int get scanOffset {
+    if (byteOffset == scanSlackOffset) {
+      return byteOffset - scanSlack;
+    } else {
+      return byteOffset;
+    }
+  }
+
+  /**
+   * The difference between the number of bytes and the number of corresponding
+   * string characters, up to the current [byteOffset].
+   */
+  int utf8Slack = 0;
+
+  /**
+   * Creates a new Utf8BytesScanner. The source file is expected to be a
+   * [Utf8BytesSourceFile] that holds a list of UTF-8 bytes. Otherwise the
+   * string text of the source file is decoded.
+   *
+   * The list of UTF-8 bytes [file.slowUtf8Bytes()] is expected to return an
+   * array whose last element is '0' to signal the end of the file. If this
+   * is not the case, the entire array is copied before scanning.
+   */
+  Utf8BytesScanner(SourceFile file, {bool includeComments: false})
+      : bytes = file.slowUtf8Bytes(),
+        super(file, includeComments) {
+    ensureZeroTermination();
+  }
+
+  /**
+   * Creates a new Utf8BytesScanner from a list of UTF-8 bytes.
+   *
+   * The last element of the list is expected to be '0' to signal the end of
+   * the file. If this is not the case, the entire array is copied before
+   * scanning.
+   */
+  Utf8BytesScanner.fromBytes(this.bytes, {bool includeComments: false})
+      : super(null, includeComments) {
+    ensureZeroTermination();
+  }
+
+  void ensureZeroTermination() {
+    if (bytes.isEmpty || bytes[bytes.length - 1] != 0) {
+      // TODO(lry): abort instead of copying the array, or warn?
+      var newBytes =  new Uint8List(bytes.length + 1);
+      for (int i = 0; i < bytes.length; i++) {
+        newBytes[i] = bytes[i];
+      }
+      newBytes[bytes.length] = 0;
+      bytes = newBytes;
+    }
+  }
+
+  int advance() => bytes[++byteOffset];
+
+  int peek() => bytes[byteOffset + 1];
+
+  /**
+   * Returns the unicode code point starting at the byte offset [startOffset]
+   * with the byte [nextByte]. If [advance] is true the current [byteOffset]
+   * is advanced to the last byte of the code point.
+   */
+  int nextCodePoint(int startOffset, int nextByte, bool advance) {
+    // The number of 1s in the first byte indicate the number of bytes, at
+    // least 2.
+    int numBytes = 2;
+    int bit = 0x20;
+    while ((nextByte & bit) != 0) {
+      numBytes++;
+      bit >>= 1;
+    }
+    int end = startOffset + numBytes;
+    if (advance) {
+      byteOffset = end - 1;
+    }
+    // TODO(lry): measurably slow, decode creates first a Utf8Decoder and a
+    // _Utf8Decoder instance. Also the sublist is eagerly allocated.
+    String codePoint = UTF8.decode(bytes.sublist(startOffset, end));
+    if (codePoint.length == 1) {
+      if (advance) {
+        utf8Slack += (numBytes - 1);
+        scanSlack = numBytes - 1;
+        scanSlackOffset = byteOffset;
+      }
+      return codePoint.codeUnitAt(0);
+    } else if (codePoint.length == 2) {
+      if (advance) {
+        utf8Slack += (numBytes - 2);
+        scanSlack = numBytes - 1;
+        scanSlackOffset = byteOffset;
+        stringOffsetSlackOffset = byteOffset;
+      }
+      // In case of a surrogate pair, return a single code point.
+      return codePoint.runes.single;
+    } else {
+      throw "Invalid UTF-8 byte sequence: ${bytes.sublist(startOffset, end)}";

[ngeoffray, 2013/10/18 10:19:37]

Will users face this error? If yes, we should throw differently.

[lukas, 2013/10/24 16:48:36]

That could only happen in a file with wrong UTF-8 encoding, in which case
UTF8.decode would probably fail earlier. I could also just remove the check.

+    }
+  }
+
+  int lastUnicodeOffset = -1;
+  int currentAsUnicode(int next) {
+    if (next < 128) return next;
+    // Check if currentAsUnicode was already invoked.
+    if (byteOffset == lastUnicodeOffset) return next;
+    int res = nextCodePoint(byteOffset, next, true);
+    lastUnicodeOffset = byteOffset;
+    return res;
+  }
+
+  void handleUnicode(int startScanOffset) {
+    int end = byteOffset;
+    // TODO(lry): this measurably slows down the scanner for files with unicode.
+    String s = UTF8.decode(bytes.sublist(startScanOffset, end));
+    utf8Slack += (end - startScanOffset) - s.length;
+  }
+
+  /**
+   * This field remembers the byte offset of the last character decoded with
+   * [nextCodePoint] that used two code units in UTF-16.
+   *
+   * [nextCodePoint] returns a single code point for each unicode character,
+   * even if it needs two code units in UTF-16.
+   *
+   * For example, '\u{1d11e}' uses 4 bytes in UTF-8, and two code units in
+   * UTF-16. The [utf8Slack] is therefore 2. After invoking [nextCodePoint], the
+   * [byteOffset] points to the last (of 4) bytes. The [stringOffset] should
+   * return the offset of the first one, which is one position more left than
+   * the [utf8Slack].
+   */
+  int stringOffsetSlackOffset = -1;
+
+  int get stringOffset {
+    if (stringOffsetSlackOffset == byteOffset) {
+      return byteOffset - utf8Slack - 1;
+    } else {
+      return byteOffset - utf8Slack;
+    }
+  }
+
+  Token firstToken() => tokens.next;
+  Token previousToken() => tail;
+
+

[ngeoffray, 2013/10/18 10:19:37]

Extra line,

[lukas, 2013/10/24 16:48:36]

Done.

+  void appendSubstringToken(PrecedenceInfo info, int start, bool asciiOnly,
+                            [int extraOffset = 0]) {
+    tail.next = new StringToken.fromUtf8Bytes(
+        info, bytes, start, byteOffset + extraOffset, asciiOnly, tokenStart);
+    tail = tail.next;
+  }
+}