First commit

3rdparty/pdf.js/external/cmapscompress/README.md

@@ -0,0 +1,171 @@
+# Quick notes about binary CMap format (bcmap)
+
+The format is designed to package some information from the CMap files located at external/cmap. Please notice for size optimization reasons, the original information blocks can be changed (split or joined) and items in the blocks can be swaped.
+
+The data stored in binary format in network byte order (big-endian).
+
+# Data primitives
+
+The following primitives used during encoding of the file:
+  - byte (B) – a byte, bits are numbered from 0 (less significant) to 7 (most significant)
+  - bytes block (B[n])  – a sequence of n bytes
+  - unsigned number (UN) – the number is encoded as sequence of bytes, bit 7 is flag to continue decoding the byte, bits 6-0 store number information, e.g. bytes 0x818407 will represent 16903 (0x4207). Limited to the 32 bit.
+  - signed number (SN) – the number is encoded as sequence of bytes, as UN, however shall be transformed before encoding: if n < 0, the n shall be encoded as (-2*n-1) using UN encoding, other n shall be encoded as (2*n) using UN encoding. So the lowest bit of the number indicates the sign of the initial number
+  - unsigned fixed number (UB[n]) – similar to the UN, but it represents an unsigned number that is stored in B[n]
+  - signed fixed number (SB[n]) – similar to the SN, but it represents a signed number that is stored in B[n]
+  - string (S) – the string is encoded as sequence of bytes. First comes length is characters encoded as UN, when UTF16 characters encoded as UN.
+
+# File structure
+
+The first byte is a header:
+  - bits 2-1 – indicate a CMapType. Valid values are 1 and 2
+  - bit 0 – indicate WMode. Valid values are 0 and 1.
+
+Then records follow. The records starts from the record header encoded as B, where bits 7-5 indicate record type (see description of other bits below):
+  - 0 – codespacerange
+  - 1 – notdefrange
+  - 2 – cidchar
+  - 3 – cidrange
+  - 4 – bfchar
+  - 5 – bfrange
+  - 6 – reserved
+  - 7 – metadata
+
+## Metadata record
+
+The metadata record header bit 4-0 contain id of the metadata:
+  - 0 – comment, body of the record is encoded comment string (S) 
+  - 1 – UseCMap, body of the record is usecmap id string (S)
+
+## Data records
+
+The records that have types 0 – 5, have the following fields in the header:
+  - bit 4 – indicate the char or start/end entries are stored in a sequence in this block
+  - bits 3-0 – contain length of the data size minus 1 in this block (dataSize)
+
+The amount of entries encoded as UN follows the header. The items records follow (see below).
+
+
+### codespacerange (0)
+
+Represents the following CMap block:
+
+  n begincodespacerange
+  <start> <end>
+  endcodespacerange
+
+First record format is:
+
+  - start as B[dataSize]
+  - endDelta as UB[dataSize], end is calculated as (start + endDelta)
+
+Next record format is:
+
+  - startDelta as UB[dataSize], start = end + startDelta
+  - endDelta as UB[dataSize], end = start + endDelta
+
+
+### notdefrange (1)
+
+Represents the following CMap block:
+
+  n beginnotdefrange
+  <start> <end> code
+  endnotdefrange
+
+First record format is:
+
+  - start as B[dataSize]
+  - endDelta as UB[dataSize], end is calculated as (start + endDelta)
+  - code as UN
+
+Next record format is:
+
+  - startDelta as UB[dataSize], start = end + startDelta
+  - endDelta as UB[dataSize], end = start + endDelta
+  - code as UN
+
+
+### cidchar (2)
+
+Represents the following CMap block:
+
+  n begincidchar
+  <char> code
+  endcidchar
+
+First record format is:
+
+  - char as B[dataSize]
+  - code as UN
+
+Next record format is:
+
+  - if sequence = 0, charDelta as UB[dataSize], char = char + charDelta + 1
+  - if sequence = 1, char = char + 1
+  - codeDelta as SN, code = code + codeDelta
+
+
+### cidrange (3)
+
+Represents the following CMap block:
+
+  n begincidrange
+  <start> <end> code
+  endcidrange
+
+First record format is:
+
+  - start as B[dataSize]
+  - endDelta as UN[dataSize], end is calculated as (start + endDelta)
+  - code as UN
+
+Next record format is:
+
+  - if sequence = 0, startDelta as UB[dataSize], start = end + startDelta + 1
+  - if sequence = 1, start = end + 1
+  - endDelta as UN[dataSize], end = start + endDelta
+  - code as UN
+
+
+### bfchar (4)
+
+Represents the following CMap block:
+
+  n beginbfchar
+  <char> <code>
+  endbfchar
+
+First record format is:
+
+  - char as B[ucs2Size], where ucs2Size = 2 (here and below)
+  - code as B[dataSize]
+
+Next record format is:
+
+  - if sequence = 0, charDelta as UN[ucs2Size], char = charDelta + charDelta + 1
+  - if sequence = 1, char = char + 1
+  - codeDelta as SB[dataSize], code = code + codeDelta
+
+
+### bfrange (5)
+
+Represents the following CMap block:
+
+  n beginbfrange
+  <start> <end> <code>
+  endbfrange
+
+First record format is:
+
+  - start as B[ucs2Size]
+  - endDelta as UB[ucs2Size], end is calculated as (start + endDelta)
+  - code as B[dataSize]
+
+Next record format is:
+
+  - if sequence = 0, startDelta as UB[ucs2Size], start = end + startDelta + 1
+  - if sequence = 1, start = end + 1
+  - endDelta as UB[ucs2Size], end = start + endDelta
+  - code as B[dataSize]
+

3rdparty/pdf.js/external/cmapscompress/compress.js

@@ -0,0 +1,437 @@
+/* Copyright 2014 Mozilla Foundation
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+var fs = require('fs');
+var path = require('path');
+var parseAdobeCMap = require('./parse.js').parseAdobeCMap;
+var optimizeCMap = require('./optimize.js').optimizeCMap;
+
+function compressCmap(srcPath, destPath, verify) {
+  var content = fs.readFileSync(srcPath).toString();
+  var inputData = parseAdobeCMap(content);
+  optimizeCMap(inputData);
+
+  var out = writeByte((inputData.type << 1) | inputData.wmode);
+  if (inputData.comment) {
+    out += writeByte(0xE0) + writeString(inputData.comment);
+  }
+  if (inputData.usecmap) {
+    out += writeByte(0xE1) + writeString(inputData.usecmap);
+  }
+  var i = 0;
+  while (i < inputData.body.length) {
+    var item = inputData.body[i++], subitems = item.items;
+    var first = item.items[0];
+    var sequence = item.sequence === true;
+    var flags = (item.type << 5) | (sequence ? 0x10 : 0);
+    var nextStart, nextCode;
+    switch (item.type) {
+      case 0:
+        out += writeByte(flags | getHexSize(first.start)) + writeNumber(subitems.length);
+        out += first.start + writeNumber(subHex(first.end, first.start));
+        nextStart = incHex(first.end);
+        for (var j = 1; j < subitems.length; j++) {
+          out += writeNumber(subHex(subitems[j].start, nextStart)) +
+            writeNumber(subHex(subitems[j].end, subitems[j].start));
+          nextStart = incHex(subitems[j].end);
+        }
+        break;
+      case 1:
+        out += writeByte(flags | getHexSize(first.start)) + writeNumber(subitems.length);
+        out += first.start + writeNumber(subHex(first.end, first.start)) + writeNumber(first.code);
+        nextStart = incHex(first.end);
+        for (var j = 1; j < subitems.length; j++) {
+          out += writeNumber(subHex(subitems[j].start, nextStart)) +
+            writeNumber(subHex(subitems[j].end, subitems[j].start)) +
+            writeNumber(subitems[j].code);
+          nextStart = incHex(subitems[j].end);
+        }
+        break;
+      case 2:
+        out += writeByte(flags | getHexSize(first.char)) + writeNumber(subitems.length);
+        out += first.char + writeNumber(first.code);
+        nextStart = incHex(first.char);
+        nextCode = first.code + 1;
+        for (var j = 1; j < subitems.length; j++) {
+          out += (sequence ? '' : writeNumber(subHex(subitems[j].char, nextStart))) +
+            writeSigned(subitems[j].code - nextCode);
+          nextStart = incHex(subitems[j].char);
+          nextCode = item.items[j].code + 1;
+        }
+        break;
+      case 3:
+        out += writeByte(flags | getHexSize(first.start)) + writeNumber(subitems.length);
+        out += first.start + writeNumber(subHex(first.end, first.start)) + writeNumber(first.code);
+        nextStart = incHex(first.end);
+        for (var j = 1; j < subitems.length; j++) {
+          out += (sequence ? '' : writeNumber(subHex(subitems[j].start, nextStart))) +
+            writeNumber(subHex(subitems[j].end, subitems[j].start)) +
+            writeNumber(subitems[j].code);
+          nextStart = incHex(subitems[j].end);
+        }
+        break;
+      case 4:
+        out += writeByte(flags | getHexSize(first.code)) + writeNumber(subitems.length);
+        out += first.char + first.code;
+        nextStart = incHex(first.char);
+        nextCode = incHex(first.code);
+        for (var j = 1; j < subitems.length; j++) {
+          out += (sequence ? '' : writeNumber(subHex(subitems[j].char, nextStart))) +
+            writeSigned(subHex(subitems[j].code, nextCode));
+          nextStart = incHex(subitems[j].char);
+          nextCode = incHex(subitems[j].code);
+        }
+        break;
+      case 5:
+        out += writeByte(flags | getHexSize(first.code)) + writeNumber(subitems.length);
+        out += first.start + writeNumber(subHex(first.end, first.start)) + first.code;
+        nextStart = incHex(first.end);
+        for (var j = 1; j < subitems.length; j++) {
+          out += (sequence ? '' : writeNumber(subHex(subitems[j].start, nextStart))) +
+            writeNumber(subHex(subitems[j].end, subitems[j].start)) +
+            subitems[j].code;
+          nextStart = incHex(subitems[j].end);
+        }
+        break;
+
+    }
+  }
+
+  fs.writeFileSync(destPath, new Buffer(out, 'hex'));
+
+  if (verify) {
+    var result2 = parseCMap(out);
+    var isGood = JSON.stringify(inputData) == JSON.stringify(result2);
+    if (!isGood) {
+      throw new Error('Extracted data does not match the expected result');
+    }
+  }
+
+  return {
+    orig: fs.statSync(srcPath).size,
+    packed: out.length >> 1
+  };
+}
+
+function parseCMap(binaryData) {
+  var reader = {
+    buffer: binaryData,
+    pos: 0,
+    end: binaryData.length,
+    readByte: function () {
+      if (this.pos >= this.end) {
+        return -1;
+      }
+      var d1 = fromHexDigit(this.buffer[this.pos]);
+      var d2 = fromHexDigit(this.buffer[this.pos + 1]);
+      this.pos += 2;
+      return (d1 << 4) | d2;
+    },
+    readNumber: function () {
+      var n = 0;
+      var last;
+      do {
+        var b = this.readByte();
+        last = !(b & 0x80);
+        n = (n << 7) | (b & 0x7F);
+      } while (!last);
+      return n;
+    },
+    readSigned: function () {
+      var n = this.readNumber();
+      return (n & 1) ? -(n >>> 1) - 1 : n >>> 1;
+    },
+    readHex: function (size) {
+      var lengthInChars = (size + 1) << 1;
+      var s = this.buffer.substr(this.pos, lengthInChars);
+      this.pos += lengthInChars;
+      return s;
+    },
+    readHexNumber: function (size) {
+      var lengthInChars = (size + 1) << 1;
+      var stack = [];
+      do {
+        var b = this.readByte();
+        last = !(b & 0x80);
+        stack.push(b & 0x7F);
+      } while (!last);
+      var s = '', buffer = 0, bufferSize = 0;
+      while (s.length < lengthInChars) {
+        while (bufferSize < 4 && stack.length > 0) {
+          buffer = (stack.pop() << bufferSize) | buffer;
+          bufferSize += 7;
+        }
+        s = toHexDigit(buffer & 15) + s;
+        buffer >>= 4;
+        bufferSize -= 4;
+      }
+      return s;
+    },
+    readHexSigned: function (size) {
+      var num = this.readHexNumber(size);
+      var sign = fromHexDigit(num[num.length - 1]) & 1 ? 15 : 0;
+      var c = 0;
+      var result = '';
+      for (var i = 0; i < num.length; i++) {
+        c = (c << 4) | fromHexDigit(num[i]);
+        result += toHexDigit(sign ? (c >> 1) ^ sign : (c >> 1));
+        c &= 1;
+      }
+      return result;
+    },
+    readString: function () {
+      var len = this.readNumber();
+      var s = '';
+      for (var i = 0; i < len; i++) {
+        s += String.fromCharCode(this.readNumber());
+      }
+      return s;
+    }
+  };
+
+  var header = reader.readByte();
+  var result = {
+    type: header >> 1,
+    wmode: header & 1,
+    comment: null,
+    usecmap: null,
+    body: []
+  };
+
+  var b;
+  while ((b = reader.readByte()) >= 0) {
+    var type = b >> 5;
+    if (type === 7) {
+      switch (b & 0x1F) {
+        case 0:
+          result.comment = reader.readString();
+          break;
+        case 1:
+          result.usecmap = reader.readString();
+          break;
+      }
+      continue;
+    }
+    var sequence = !!(b & 0x10);
+    var dataSize = b & 15;
+    var subitems = [];
+    var item = {
+      type: type,
+      items: subitems
+    };
+    if (sequence) {
+      item.sequence = true;
+    }
+    var ucs2DataSize = 1;
+    var subitemsCount = reader.readNumber();
+    var start, end, code, char;
+    switch (type) {
+      case 0:
+        start = reader.readHex(dataSize);
+        end = addHex(reader.readHexNumber(dataSize), start);
+        subitems.push({start: start, end: end});
+        for (var i = 1; i < subitemsCount; i++) {
+          start = addHex(reader.readHexNumber(dataSize), incHex(end));
+          end = addHex(reader.readHexNumber(dataSize), start);
+          subitems.push({start: start, end: end});
+        }
+        break;
+      case 1:
+        start = reader.readHex(dataSize);
+        end = addHex(reader.readHexNumber(dataSize), start);
+        code = reader.readNumber();
+        subitems.push({start: start, end: end, code: code});
+        for (var i = 1; i < subitemsCount; i++) {
+          start = addHex(reader.readHexNumber(dataSize), incHex(end));
+          end = addHex(reader.readHexNumber(dataSize), start);
+          code = reader.readNumber();
+          subitems.push({start: start, end: end, code: code});
+        }
+        break;
+      case 2:
+        char = reader.readHex(dataSize);
+        code = reader.readNumber();
+        subitems.push({char: char, code: code});
+        for (var i = 1; i < subitemsCount; i++) {
+          char = sequence ? incHex(char) : addHex(reader.readHexNumber(dataSize), incHex(char));
+          code = reader.readSigned() + (code + 1);
+          subitems.push({char: char, code: code});
+        }
+        break;
+      case 3:
+        start = reader.readHex(dataSize);
+        end = addHex(reader.readHexNumber(dataSize), start);
+        code = reader.readNumber();
+        subitems.push({start: start, end: end, code: code});
+        for (var i = 1; i < subitemsCount; i++) {
+          start = sequence ? incHex(end) : addHex(reader.readHexNumber(dataSize), incHex(end));
+          end = addHex(reader.readHexNumber(dataSize), start);
+          code = reader.readNumber();
+          subitems.push({start: start, end: end, code: code});
+        }
+        break;
+      case 4:
+        char = reader.readHex(ucs2DataSize);
+        code = reader.readHex(dataSize);
+        subitems.push({char: char, code: code});
+        for (var i = 1; i < subitemsCount; i++) {
+          char = sequence ? incHex(char) : addHex(reader.readHexNumber(ucs2DataSize), incHex(char));
+          code = addHex(reader.readHexSigned(dataSize), incHex(code));
+          subitems.push({char: char, code: code});
+        }
+        break;
+      case 5:
+        start = reader.readHex(ucs2DataSize);
+        end = addHex(reader.readHexNumber(ucs2DataSize), start);
+        code = reader.readHex(dataSize);
+        subitems.push({start: start, end: end, code: code});
+        for (var i = 1; i < subitemsCount; i++) {
+          start = sequence ? incHex(end) : addHex(reader.readHexNumber(ucs2DataSize), incHex(end));
+          end = addHex(reader.readHexNumber(ucs2DataSize), start);
+          code = reader.readHex(dataSize);
+          subitems.push({start: start, end: end, code: code});
+        }
+        break;
+      default:
+        throw new Error('Unknown type: ' + type)
+    }
+    result.body.push(item);
+  }
+
+  return result;
+}
+
+function toHexDigit(n) {
+  return n.toString(16);
+}
+function fromHexDigit(s) {
+  return parseInt(s, 16);
+}
+function getHexSize(s) {
+  return (s.length >> 1) - 1;
+}
+function writeByte(b) {
+  return toHexDigit((b >> 4) & 15) + toHexDigit(b & 15);
+}
+function writeNumber(n) {
+  if (typeof n === 'string') {
+    var s = '', buffer = 0, bufferSize = 0;
+    var i = n.length;
+    while (i > 0) {
+      --i;
+      buffer = (fromHexDigit(n[i]) << bufferSize) | buffer;
+      bufferSize += 4;
+      if (bufferSize >= 7) {
+        s = writeByte((buffer & 0x7f) | (s.length > 0 ? 0x80 : 0)) + s;
+        buffer >>>= 7;
+        bufferSize -= 7;
+      }
+    }
+    if (buffer > 0) {
+      s = writeByte((buffer & 0x7f) | (s.length > 0 ? 0x80 : 0)) + s;
+    }
+    while (s.indexOf('80') === 0) {
+      s = s.substr(2);
+    }
+    return s;
+  } else {
+    var s = writeByte(n & 0x7F);
+    n >>>= 7;
+    while (n > 0) {
+      s = writeByte((n & 0x7F) | 0x80) + s;
+      n >>>= 7;
+    }
+    return s;
+  }
+}
+function writeSigned(n) {
+  if (typeof n === 'string') {
+    var t = '';
+    var c = fromHexDigit(n[0]);
+    var neg = c >= 8;
+    c = neg ? (c ^ 15) : c;
+    for (var i = 1; i < n.length; i++) {
+      var d = fromHexDigit(n[i]);
+      c = (c << 4) | (neg ? (d ^ 15) : d);
+      t += toHexDigit(c >> 3);
+      c = c & 7;
+    }
+    t += toHexDigit((c << 1) | (neg ? 1 : 0));
+    return writeNumber(t);
+  }
+  return n < 0 ? writeNumber(-2 * n - 1) : writeNumber(2 * n);
+}
+function writeString(s) {
+  var t = writeNumber(s.length);
+  for (var i = 0; i < s.length; i++) {
+    t += writeNumber(s.charCodeAt(i));
+  }
+  return t;
+}
+function addHex(a, b) {
+  var c = 0, s = '';
+  for (var i = a.length - 1; i >= 0; i--) {
+    c += fromHexDigit(a[i]) + fromHexDigit(b[i]);
+    if (c >= 16) {
+      s = toHexDigit(c - 16) + s;
+      c = 1;
+    } else {
+      s = toHexDigit(c) + s;
+      c = 0;
+    }
+  }
+  return s;
+}
+function subHex(a, b) {
+  var c = 0, s = '';
+  for (var i = a.length - 1; i >= 0; i--) {
+    c += fromHexDigit(a[i]) - fromHexDigit(b[i]);
+    if (c < 0) {
+      s = toHexDigit(c + 16) + s;
+      c = -1;
+    } else {
+      s = toHexDigit(c) + s;
+      c = 0;
+    }
+  }
+  return s;
+}
+function incHex(a) {
+  var c = 1, s = '';
+  for (var i = a.length - 1; i >= 0; i--) {
+    c += fromHexDigit(a[i]);
+    if (c >= 16) {
+      s = toHexDigit(c - 16) + s;
+      c = 1;
+    } else {
+      s = toHexDigit(c) + s;
+      c = 0;
+    }
+  }
+  return s;
+}
+
+exports.compressCmaps = function (src, dest, verify) {
+  var files = fs.readdirSync(src).filter(function (fn) {
+    return fn.indexOf('.') < 0; // skipping files with the extension
+  });
+  files.forEach(function (fn) {
+    var srcPath = path.join(src, fn);
+    var destPath = path.join(dest, fn + '.bcmap');
+    var stats = compressCmap(srcPath, destPath, verify);
+    console.log('Compressing ' + fn + ': ' + stats.orig + ' vs ' + stats.packed +
+      ' ' + (stats.packed / stats.orig * 100).toFixed(1) + '%');
+  });
+};

3rdparty/pdf.js/external/cmapscompress/optimize.js

@@ -0,0 +1,211 @@
+/* Copyright 2014 Mozilla Foundation
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+exports.optimizeCMap = function (data) {
+  var i = 1;
+  while (i < data.body.length) {
+    if (data.body[i - 1].type === data.body[i].type) {
+      data.body[i - 1].items = data.body[i - 1].items.concat(data.body[i].items);
+      data.body.splice(i, 1);
+    } else {
+      i++;
+    }
+  }
+  // split into groups with different lengths
+  var i = 0;
+  while (i < data.body.length) {
+    var item = data.body[i];
+    var keys = Object.keys(item.items[0]).filter(function (i) {
+      return typeof item.items[0][i] === 'string';
+    });
+    var j = 1;
+    while (j < item.items.length) {
+      var different = false;
+      for (var q = 0; q < keys.length && !different; q++) {
+        different = item.items[j - 1][keys[q]].length !== item.items[j][keys[q]].length;
+      }
+      if (different) {
+        break;
+      }
+      j++;
+    }
+    if (j < item.items.length) {
+      data.body.splice(i + 1, 0, {
+        type: item.type,
+        items: item.items.splice(j, item.items.length - j)
+      });
+    }
+    i++;
+  }
+  // find sequences of single char ranges
+  var i = 0;
+  while (i < data.body.length) {
+    var item = data.body[i];
+    if (item.type === 3 || item.type === 5) {
+      var j = 0;
+      while (j < item.items.length) {
+        var q = j;
+        while (j < item.items.length && item.items[j].start === item.items[j].end) {
+          j++;
+        }
+        if ((j - q) >= 9) {
+          if (j < item.items.length) {
+            data.body.splice(i + 1, 0, {
+              type: item.type,
+              items: item.items.splice(j, item.items.length - j)
+            });
+          }
+          if (q > 0) {
+            data.body.splice(i + 1, 0, {
+              type: item.type - 1,
+              items: item.items.splice(q, j - q).map(function (i) {
+                return {char: i.start, code: i.code };
+              })
+            });
+            i++;
+          } else {
+            item.type -= 1;
+            item.items = item.items.map(function (i) {
+              return {char: i.start, code: i.code };
+            });
+          }
+          continue;
+        }
+        j++;
+      }
+    }
+    i++;
+  }
+
+  //  find sequences of increasing code/ranges order
+  var i = 0;
+  while (i < data.body.length) {
+    var item = data.body[i];
+    if (item.type >= 2 && item.type <= 5) {
+      var j = 1;
+      var startProp = item.type === 2 || item.type === 4 ? 'char' : 'start';
+      var endProp = item.type === 2 || item.type === 4 ? 'char' : 'end';
+      while (j < item.items.length) {
+        var q = j - 1;
+        while (j < item.items.length && incHex(item.items[j - 1][endProp]) === item.items[j][startProp]) {
+          j++;
+        }
+        if ((j - q) >= 9) {
+          if (j < item.items.length) {
+            data.body.splice(i + 1, 0, {
+              type: item.type,
+              items: item.items.splice(j, item.items.length - j)
+            });
+          }
+          if (q > 0) {
+            data.body.splice(i + 1, 0, {
+              type: item.type,
+              items: item.items.splice(q, j - q),
+              sequence: true
+            });
+            i++;
+          } else {
+            item.sequence = true;
+          }
+          continue;
+        }
+        j++;
+      }
+    }
+    i++;
+  }
+
+  // split non-sequences two groups where codes are close
+  var i = 0;
+  while (i < data.body.length) {
+    var item = data.body[i];
+    if (!item.sequence && (item.type === 2 || item.type === 3)) {
+      var subitems = item.items;
+      var codes = subitems.map(function (i) {
+        return i.code;
+      });
+      codes.sort(function (a, b) {
+        return a - b;
+      });
+      var maxDistance = 100, minItems = 10, itemsPerBucket = 50;
+      if (subitems.length > minItems && codes[codes.length - 1] - codes[0] > maxDistance) {
+        var gapsCount = Math.max(2, (subitems.length / itemsPerBucket) | 0);
+        var gaps = [];
+        for (var q = 0; q < gapsCount; q++) {
+          gaps.push({length: 0});
+        }
+        for (var j = 1; j < codes.length; j++) {
+          var gapLength = codes[j] - codes[j - 1];
+          var q = 0;
+          while (q < gaps.length && gaps[q].length > gapLength) {
+            q++;
+          }
+          if (q >= gaps.length) {
+            continue;
+          }
+          var q0 = q;
+          while (q < gaps.length) {
+            if (gaps[q].length < gaps[q0].length) {
+              q0 = q;
+            }
+            q++;
+          }
+          gaps[q0] = {length: gapLength, boundary: codes[j]};
+        }
+        var groups = gaps.filter(function (g) {
+          return g.length >= maxDistance;
+        }).map(function (g) {
+            return g.boundary;
+          });
+        groups.sort(function (a, b) {
+          return a - b;
+        });
+        if (groups.length > 1) {
+          var buckets = [item.items = []];
+          for (var j = 0; j < groups.length; j++) {
+            var newItem = {type: item.type, items: []}
+            buckets.push(newItem.items);
+            i++;
+            data.body.splice(i, 0, newItem);
+          }
+          for (var j = 0; j < subitems.length; j++) {
+            var code = subitems[j].code;
+            var q = 0;
+            while (q < groups.length && groups[q] <= code) {
+              q++;
+            }
+            buckets[q].push(subitems[j]);
+          }
+        }
+      }
+    }
+    i++;
+  }
+};
+
+function incHex(a) {
+  var c = 1, s = '';
+  for (var i = a.length - 1; i >= 0; i--) {
+    c += parseInt(a[i], 16);
+    if (c >= 16) {
+      s = '0' + s;
+      c = 1;
+    } else {
+      s = c.toString(16) + s;
+      c = 0;
+    }
+  }
+  return s;
+}

3rdparty/pdf.js/external/cmapscompress/parse.js

@@ -0,0 +1,101 @@
+/* Copyright 2014 Mozilla Foundation
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+exports.parseAdobeCMap = function (content) {
+  var m = /(\bbegincmap\b[\s\S]*?)\bendcmap\b/.exec(content);
+  if (!m) {
+    throw new Error('cmap was not found');
+  }
+
+  var body = m[1].replace(/\r\n?/g, '\n');
+  var result = {
+    type: 1,
+    wmode: 0,
+    comment: 'Copyright 1990-2009 Adobe Systems Incorporated.\nAll rights reserved.\nSee ./LICENSE',
+    usecmap: null,
+    body: []
+  };
+  m = /\/CMapType\s+(\d+)+\s+def\b/.exec(body);
+  result.type = +m[1];
+  m = /\/WMode\s+(\d+)+\s+def\b/.exec(body);
+  result.wmode = +m[1];
+  m = /\/([\w\-]+)\s+usecmap\b/.exec(body);
+  if (m) {
+    result.usecmap = m[1];
+  }
+  var re = /(\d+)\s+(begincodespacerange|beginnotdefrange|begincidchar|begincidrange|beginbfchar|beginbfrange)\n([\s\S]*?)\n(endcodespacerange|endnotdefrange|endcidchar|endcidrange|endbfchar|endbfrange)/g;
+  while (m = re.exec(body)) {
+    var lines = m[3].toLowerCase().split('\n');
+    var m2;
+    switch (m[2]) {
+      case 'begincodespacerange':
+        result.body.push({
+          type: 0,
+          items: lines.map(function (line) {
+            var m = /<(\w+)>\s+<(\w+)>/.exec(line);
+            return {start: m[1], end: m[2]};
+          })
+        });
+        break;
+      case 'beginnotdefrange':
+        result.body.push({
+          type: 1,
+          items: lines.map(function (line) {
+            var m = /<(\w+)>\s+<(\w+)>\s+(\d+)/.exec(line);
+            return {start: m[1], end: m[2], code: +m[3]};
+          })
+        });
+        break;
+      case 'begincidchar':
+        result.body.push({
+          type: 2,
+          items: lines.map(function (line) {
+            var m = /<(\w+)>\s+(\d+)/.exec(line);
+            return {char: m[1], code: +m[2]};
+          })
+        });
+        break;
+      case 'begincidrange':
+        result.body.push({
+          type: 3,
+          items: lines.map(function (line) {
+            var m = /<(\w+)>\s+<(\w+)>\s+(\d+)/.exec(line);
+            return {start: m[1], end: m[2], code: +m[3]};
+          })
+        });
+        break;
+      case 'beginbfchar':
+        result.body.push({
+          type: 4,
+          items: lines.map(function (line) {
+            var m = /<(\w+)>\s+<(\w+)>/.exec(line);
+            return {char: m[1], code: m[2]};
+          })
+        });
+        break;
+      case 'beginbfrange':
+        result.body.push({
+          type: 5,
+          items: lines.map(function (line) {
+            var m = /<(\w+)>\s+<(\w+)>\s+<(\w+)>/.exec(line);
+            return {start: m[1], end: m[2], code: m[3]};
+          })
+        });
+        break;
+    }
+  }
+
+  return result;
+};