src/runtime_strings.js

/**
 * @license
 * Copyright 2019 The Emscripten Authors
 * SPDX-License-Identifier: MIT
 */

// runtime_strings.js: Strings related runtime functions that are part of both MINIMAL_RUNTIME and regular runtime.

// Given a pointer 'ptr' to a null-terminated UTF8-encoded string in the given array that contains uint8 values, returns
// a copy of that string as a Javascript String object.

#if USE_PTHREADS && TEXTDECODER
// UTF8Decoder.decode may not work with a view of a SharedArrayBuffer, see
// https://github.com/whatwg/encoding/issues/172
// To avoid that, we wrap around it and add a copy into a normal ArrayBuffer,
// which can still be much faster than creating a string character by
// character.
function TextDecoderWrapper(encoding) {
  var textDecoder = new TextDecoder(encoding);
  this.decode = function(data) {
#if ASSERTIONS
    assert(data instanceof Uint8Array);
#endif
    // While we compile with pthreads, this method can be called on side buffers
    // as well, such as the stdout buffer in the filesystem code. Only copy when
    // we have to.
    if (data.buffer instanceof SharedArrayBuffer) {
      data = new Uint8Array(data);
    }
    return textDecoder.decode.call(textDecoder, data);
  };
}
#endif

#if TEXTDECODER == 2
var UTF8Decoder = new TextDecoder{{{ USE_PTHREADS ? 'Wrapper' : ''}}}('utf8');
#else // TEXTDECODER == 2
#if TEXTDECODER
var UTF8Decoder = typeof TextDecoder !== 'undefined' ? new TextDecoder{{{ USE_PTHREADS ? 'Wrapper' : ''}}}('utf8') : undefined;
#endif // TEXTDECODER
#endif // TEXTDECODER == 2

/**
 * @param {number} idx
 * @param {number=} maxBytesToRead
 * @return {string}
 */
function UTF8ArrayToString(heap, idx, maxBytesToRead) {
#if CAN_ADDRESS_2GB
  idx >>>= 0;
#endif
  var endIdx = idx + maxBytesToRead;
#if TEXTDECODER
  var endPtr = idx;
  // TextDecoder needs to know the byte length in advance, it doesn't stop on null terminator by itself.
  // Also, use the length info to avoid running tiny strings through TextDecoder, since .subarray() allocates garbage.
  // (As a tiny code save trick, compare endPtr against endIdx using a negation, so that undefined means Infinity)
  while (heap[endPtr] && !(endPtr >= endIdx)) ++endPtr;
#endif // TEXTDECODER

#if TEXTDECODER == 2
  return UTF8Decoder.decode(
    heap.subarray ? heap.subarray(idx, endPtr) : new Uint8Array(heap.slice(idx, endPtr))
  );
#else // TEXTDECODER == 2
#if TEXTDECODER
  if (endPtr - idx > 16 && heap.subarray && UTF8Decoder) {
    return UTF8Decoder.decode(heap.subarray(idx, endPtr));
  } else {
#endif // TEXTDECODER
    var str = '';
#if TEXTDECODER
    // If building with TextDecoder, we have already computed the string length above, so test loop end condition against that
    while (idx < endPtr) {
#else
    while (!(idx >= endIdx)) {
#endif
      // For UTF8 byte structure, see:
      // http://en.wikipedia.org/wiki/UTF-8#Description
      // https://www.ietf.org/rfc/rfc2279.txt
      // https://tools.ietf.org/html/rfc3629
      var u0 = heap[idx++];
#if !TEXTDECODER
      // If not building with TextDecoder enabled, we don't know the string length, so scan for \0 byte.
      // If building with TextDecoder, we know exactly at what byte index the string ends, so checking for nulls here would be redundant.
      if (!u0) return str;
#endif
      if (!(u0 & 0x80)) { str += String.fromCharCode(u0); continue; }
      var u1 = heap[idx++] & 63;
      if ((u0 & 0xE0) == 0xC0) { str += String.fromCharCode(((u0 & 31) << 6) | u1); continue; }
      var u2 = heap[idx++] & 63;
      if ((u0 & 0xF0) == 0xE0) {
        u0 = ((u0 & 15) << 12) | (u1 << 6) | u2;
      } else {
#if ASSERTIONS
        if ((u0 & 0xF8) != 0xF0) warnOnce('Invalid UTF-8 leading byte 0x' + u0.toString(16) + ' encountered when deserializing a UTF-8 string in wasm memory to a JS string!');
#endif
        u0 = ((u0 & 7) << 18) | (u1 << 12) | (u2 << 6) | (heap[idx++] & 63);
      }

      if (u0 < 0x10000) {
        str += String.fromCharCode(u0);
      } else {
        var ch = u0 - 0x10000;
        str += String.fromCharCode(0xD800 | (ch >> 10), 0xDC00 | (ch & 0x3FF));
      }
    }
#if TEXTDECODER
  }
#endif // TEXTDECODER
  return str;
#endif // TEXTDECODER == 2
}

// Given a pointer 'ptr' to a null-terminated UTF8-encoded string in the emscripten HEAP, returns a
// copy of that string as a Javascript String object.
// maxBytesToRead: an optional length that specifies the maximum number of bytes to read. You can omit
//                 this parameter to scan the string until the first \0 byte. If maxBytesToRead is
//                 passed, and the string at [ptr, ptr+maxBytesToReadr[ contains a null byte in the
//                 middle, then the string will cut short at that byte index (i.e. maxBytesToRead will
//                 not produce a string of exact length [ptr, ptr+maxBytesToRead[)
//                 N.B. mixing frequent uses of UTF8ToString() with and without maxBytesToRead may
//                 throw JS JIT optimizations off, so it is worth to consider consistently using one
//                 style or the other.
/**
 * @param {number} ptr
 * @param {number=} maxBytesToRead
 * @return {string}
 */
function UTF8ToString(ptr, maxBytesToRead) {
#if CAN_ADDRESS_2GB
  ptr >>>= 0;
#endif
#if TEXTDECODER == 2
  if (!ptr) return '';
  var maxPtr = ptr + maxBytesToRead;
  for (var end = ptr; !(end >= maxPtr) && HEAPU8[end];) ++end;
  return UTF8Decoder.decode(HEAPU8.subarray(ptr, end));
#else
  return ptr ? UTF8ArrayToString(HEAPU8, ptr, maxBytesToRead) : '';
#endif
}

// Copies the given Javascript String object 'str' to the given byte array at address 'outIdx',
// encoded in UTF8 form and null-terminated. The copy will require at most str.length*4+1 bytes of space in the HEAP.
// Use the function lengthBytesUTF8 to compute the exact number of bytes (excluding null terminator) that this function will write.
// Parameters:
//   str: the Javascript string to copy.
//   heap: the array to copy to. Each index in this array is assumed to be one 8-byte element.
//   outIdx: The starting offset in the array to begin the copying.
//   maxBytesToWrite: The maximum number of bytes this function can write to the array.
//                    This count should include the null terminator,
//                    i.e. if maxBytesToWrite=1, only the null terminator will be written and nothing else.
//                    maxBytesToWrite=0 does not write any bytes to the output, not even the null terminator.
// Returns the number of bytes written, EXCLUDING the null terminator.

function stringToUTF8Array(str, heap, outIdx, maxBytesToWrite) {
#if CAN_ADDRESS_2GB
  outIdx >>>= 0;
#endif
  if (!(maxBytesToWrite > 0)) // Parameter maxBytesToWrite is not optional. Negative values, 0, null, undefined and false each don't write out any bytes.
    return 0;

  var startIdx = outIdx;
  var endIdx = outIdx + maxBytesToWrite - 1; // -1 for string null terminator.
  for (var i = 0; i < str.length; ++i) {
    // Gotcha: charCodeAt returns a 16-bit word that is a UTF-16 encoded code unit, not a Unicode code point of the character! So decode UTF16->UTF32->UTF8.
    // See http://unicode.org/faq/utf_bom.html#utf16-3
    // For UTF8 byte structure, see http://en.wikipedia.org/wiki/UTF-8#Description and https://www.ietf.org/rfc/rfc2279.txt and https://tools.ietf.org/html/rfc3629
    var u = str.charCodeAt(i); // possibly a lead surrogate
    if (u >= 0xD800 && u <= 0xDFFF) {
      var u1 = str.charCodeAt(++i);
      u = 0x10000 + ((u & 0x3FF) << 10) | (u1 & 0x3FF);
    }
    if (u <= 0x7F) {
      if (outIdx >= endIdx) break;
      heap[outIdx++] = u;
    } else if (u <= 0x7FF) {
      if (outIdx + 1 >= endIdx) break;
      heap[outIdx++] = 0xC0 | (u >> 6);
      heap[outIdx++] = 0x80 | (u & 63);
    } else if (u <= 0xFFFF) {
      if (outIdx + 2 >= endIdx) break;
      heap[outIdx++] = 0xE0 | (u >> 12);
      heap[outIdx++] = 0x80 | ((u >> 6) & 63);
      heap[outIdx++] = 0x80 | (u & 63);
    } else {
      if (outIdx + 3 >= endIdx) break;
#if ASSERTIONS
      if (u >= 0x200000) warnOnce('Invalid Unicode code point 0x' + u.toString(16) + ' encountered when serializing a JS string to a UTF-8 string in wasm memory! (Valid unicode code points should be in range 0-0x1FFFFF).');
#endif
      heap[outIdx++] = 0xF0 | (u >> 18);
      heap[outIdx++] = 0x80 | ((u >> 12) & 63);
      heap[outIdx++] = 0x80 | ((u >> 6) & 63);
      heap[outIdx++] = 0x80 | (u & 63);
    }
  }
  // Null-terminate the pointer to the buffer.
  heap[outIdx] = 0;
  return outIdx - startIdx;
}

// Copies the given Javascript String object 'str' to the emscripten HEAP at address 'outPtr',
// null-terminated and encoded in UTF8 form. The copy will require at most str.length*4+1 bytes of space in the HEAP.
// Use the function lengthBytesUTF8 to compute the exact number of bytes (excluding null terminator) that this function will write.
// Returns the number of bytes written, EXCLUDING the null terminator.

function stringToUTF8(str, outPtr, maxBytesToWrite) {
#if ASSERTIONS
  assert(typeof maxBytesToWrite == 'number', 'stringToUTF8(str, outPtr, maxBytesToWrite) is missing the third parameter that specifies the length of the output buffer!');
#endif
  return stringToUTF8Array(str, {{{ heapAndOffset('HEAPU8', 'outPtr') }}}, maxBytesToWrite);
}

// Returns the number of bytes the given Javascript string takes if encoded as a UTF8 byte array, EXCLUDING the null terminator byte.
function lengthBytesUTF8(str) {
  var len = 0;
  for (var i = 0; i < str.length; ++i) {
    // Gotcha: charCodeAt returns a 16-bit word that is a UTF-16 encoded code unit, not a Unicode code point of the character! So decode UTF16->UTF32->UTF8.
    // See http://unicode.org/faq/utf_bom.html#utf16-3
    var u = str.charCodeAt(i); // possibly a lead surrogate
    if (u >= 0xD800 && u <= 0xDFFF) u = 0x10000 + ((u & 0x3FF) << 10) | (str.charCodeAt(++i) & 0x3FF);
    if (u <= 0x7F) ++len;
    else if (u <= 0x7FF) len += 2;
    else if (u <= 0xFFFF) len += 3;
    else len += 4;
  }
  return len;
}