Add a the beggining of a Type1 font reader

diff --git a/PDFFont.js b/PDFFont.js
new file mode 100644 (file)
index 0000000..8328c8a
--- /dev/null
+++ b/PDFFont.js
@@ -0,0 +1,226 @@
+
+
+var Type1Parser = function(aLexer) {
+  var lexer = aLexer;
+
+  
+  /* 
+   * The operand stack holds arbitrary PostScript objects that are the operands
+   * and results of PostScript operators being executed. The interpreter pushes
+   * objects on the operand stack when it encounters them as literal data in a
+   * program being executed. When an operator requires one or more operands, it
+   * obtains them by popping them off the top of the operand stack. When an
+   * operator returns one or more results, it does so by pushing them on the
+   * operand stack.
+   */
+  var operandStack = [];
+
+
+  /*
+   * The dictionary stack holds only dictionary objects. The current set of
+   * dictionaries on the dictionary stack defines the environment for all
+   * implicit name searches, such as those that occur when the interpreter
+   * encounters an executable name. The role of the dictionary stack is
+   * introduced in Section 3.3, “Data Types and Objects,” and is further
+   * explained in Section 3.5, “Execution.” of the PostScript Language
+   * Reference.
+   */
+  var systemDict = new Dict(),
+      globalDict = new Dict(),
+      userDict   = new Dict();
+      
+
+  var dictionaryStack = {
+    __innerStack__: [systemDict, globalDict],
+
+    push: function(aDictionary) {
+      this.__innerStack__.push(aDictionary);
+    },
+
+    pop: function() {
+      if (this.__innerStack__.length == 2)
+        return null;
+
+      return this.__innerStack__.pop();
+    },
+
+    peek: function() {
+      return this.__innerStack__[this.__innerStack__.length - 1];
+    }
+  }
+  var currentDict = dictionaryStack.peek();
+
+
+  /*
+   * The execution stack holds executable objects (mainly procedures and files)
+   * that are in intermediate stages of execution. At any point in the
+   * execution of a PostScript program, this stack represents the program’s
+   * call stack. Whenever the interpreter suspends execution of an object to
+   * execute some other object, it pushes the new object on the execution
+   * stack. When the interpreter finishes executing an object, it pops that
+   * object off the execution stack and resumes executing the suspended object
+   * beneath it.
+   */
+  var isExecutionStack = false;
+  var executionStack = [];
+
+  /* Stub to inhibit the logs */
+
+  this.getObj = function() {
+    var obj = lexer.getObj();
+    if (isExecutionStack && !IsCmd(obj, "}") && !IsCmd(obj, "]")) {
+      executionStack.push(obj);
+      this.getObj();
+    } else if (IsBool(obj) || IsInt(obj) || IsNum(obj) || IsString(obj)) {
+      log("Value: " + obj);
+      operandStack.push(obj);
+      this.getObj();
+    } else if (IsCmd(obj, "dup") || IsCmd(obj, "readonly") ||
+               IsCmd(obj, "currentdict") || IsCmd(obj, "currentfile")) {
+      // Do nothing for the moment
+      this.getObj();
+    } else if (IsName(obj)) {
+      log("Name: " + obj.name);
+      operandStack.push(obj.name);
+      this.getObj();
+    } else if (IsCmd(obj, "dict")) {
+      log("Dict: " + obj);
+
+      // XXX handling of dictionary is wrong here
+      var size = operandStack.pop();
+      var name = operandStack.pop();
+      if (!name) {
+        log ("Creating the global dict");
+        currentDict = dictionaryStack.peek();
+      } else {
+        var dict = new Dict();
+        log("Assign name: " + name + " for the dictionary");
+        currentDict.set(name, dict);
+        dictionaryStack.push(dict);
+      }
+
+      this.getObj();
+    } else if (IsCmd(obj, "begin")) {
+      log("begin a dictionary");
+      currentDict = dictionaryStack.peek();
+      this.getObj();
+    } else if (IsCmd(obj, "end")) {
+      log("Ending a dictionary");
+      dictionaryStack.pop();
+      currentDict = dictionaryStack.peek();
+      this.getObj();
+    } else if (IsCmd(obj, "def")) {
+      if (executionStack.length) {
+        var value = [];
+        while (executionStack.length)
+          value.push(executionStack.shift());
+      } else {
+        var value = operandStack.pop();
+      }
+
+      var key = operandStack.pop();
+      // XXX this happen because of a bad way to handle dictionary
+      if (key) {
+        log("def: " + key + " = " + value);
+        currentDict.set(key, value);
+      }
+      this.getObj();
+    } else if (IsCmd(obj, "{")) {
+      log("Start Proc: " + obj);
+      executionStack = [];
+      isExecutionStack = true;
+      this.getObj();
+    } else if (IsCmd(obj, "}")) {
+      log("End Proc: " + obj);
+      isExecutionStack = false;
+      this.getObj();
+    } else if (IsCmd(obj, "[")) {
+      isExecutionStack = true;
+      executionStack = [];
+      this.getObj();
+      log("Start array: " + obj);
+    } else if (IsCmd(obj, "]")) {
+      log("End array: " + obj);
+      isExecutionStack = false;
+      this.getObj();
+    } else if (IsCmd(obj, "eexec")) {
+      return; // end of the ASCII header
+    } else {
+      log("Getting an unknow token, adding it to the stack just in case");
+      log(obj);
+      operandStack.push(obj);
+      this.getObj();
+    }
+    return currentDict;
+  }
+};
+
+var hack = false;
+
+var Type1Font = function(aFontName, aFontFile) {
+  // All type1 font program should begin with the comment %!
+  var validHeader = aFontFile.getByte() == 0x25 && aFontFile.getByte() == 0x21;
+  if (!validHeader)
+    error("Invalid file header");
+
+  var programType = "PS-AdobeFont";
+  for (var i = 0; i< programType.length; i++)
+    aFontFile.getChar();
+
+  // Ignore the '-' separator
+  aFontFile.getChar();
+
+  var version = parseFloat(aFontFile.getChar() + aFontFile.getChar() + aFontFile.getChar());
+    
+  if (!hack) {
+    log(aFontName);
+    log("Version is: " + version);
+
+    var ASCIIStream = aFontFile.makeSubStream(0, aFontFile.dict.get("Length1"), aFontFile.dict);
+    this.parser = new Type1Parser(new Lexer(ASCIIStream));
+
+    var fontDictionary = this.parser.getObj();
+    log(fontDictionary + "\t" + 
+        "fontInfo: " + fontDictionary.get("FontInfo") + "\t" +
+        "charStrings: " + fontDictionary.get("charStrings"));
+
+    var binaryStream = aFontFile.makeSubStream(aFontFile.dict.get("Length1"), aFontFile.dict.get("Length2"), aFontFile.dict);
+    function decrypt(aBinaryStream, aKey) {
+      var c1 = 52845, c2 = 22719;
+      var R = aKey;
+
+      var streamSize = aBinaryStream.length;
+      var decryptedString = [];
+
+      var value = null;
+      for (var i = 0; i < streamSize; i++) {
+        value = aBinaryStream.getByte();
+        decryptedString[i] = String.fromCharCode(value ^ (R >> 8));
+        R = ((value + R) * c1 + c2) & ((1 << 16) - 1);
+      }
+      return decryptedString.slice(4);
+    }
+
+    var eexecString = decrypt(binaryStream, 55665).join("");
+    log(eexecString);
+
+    TODO("decrypt charStrings data with the key 4330");
+    hack = true;
+  }
+
+
+  this.info = {};
+  this.name = aFontName;
+  this.encoding = [];
+  this.paintType = 0;
+  this.fontType = 0;
+  this.fontMatrix = [];
+  this.fontBBox = [];
+  this.uniqueID = 0;
+  this.metrics = {};
+  this.strokeWidth = 0.0;
+  this.private = {};
+  this.charStrings = {}
+  this.FID = 0;
+};
+

diff --git a/pdf.js b/pdf.js
index 0628cfe32b662bfed3b7e795d6b381b36c7f09cf..8caac0362282fbfe724bd2a5e064fa2c5de1233f 100644 (file)
--- a/pdf.js
+++ b/pdf.js
@@ -1032,6 +1032,12 @@ var Dict = (function() {
         },
         set: function(key, value) {
             this.map[key] = value;
+        },
+        toString: function() {
+          var keys = [];
+          for (var key in this.map)
+            keys.push(key);
+          return "Dict with " + keys.length + " keys: " + keys;
         }
     };
 
@@ -1360,10 +1366,11 @@ var Lexer = (function() {
                     stream.skip();
                     return new Cmd(">>");
                 }
+            case "{":
+            case "}":
+              return new Cmd(ch);
                // fall through
             case ')':
-            case '{':
-            case '}':
                 error("Illegal character");
                 return Error;
             }
@@ -2258,11 +2265,34 @@ var CanvasGraphics = (function() {
         },
         setFont: function(fontRef, size) {
             var font = this.res.get("Font").get(fontRef.name);
+            font = this.xref.fetchIfRef(font);
             if (!font)
                 return;
+
+            var fontName = "Nimbus Roman No9 L";
+            var subtype = font.get("Subtype").name;
+            switch (subtype) {
+              case "Type1":
+                var fontDescriptor = font.get("FontDescriptor");
+                if (fontDescriptor.num) {
+                  var fontDescriptor = this.xref.fetchIfRef(fontDescriptor);
+                  var fontFile = this.xref.fetchIfRef(fontDescriptor.get("FontFile"));
+                  font = new Type1Font(fontDescriptor.get("FontName").name, fontFile);
+                }
+                break;
+
+              case "Type3":
+                TODO("support Type3 font");
+                break;
+
+              default:
+                error("Unsupported font type: " + subtype);
+                break;
+            }
+
             this.current.fontSize = size;
             TODO("using hard-coded font for testing");
-            this.ctx.font = this.current.fontSize +'px "Nimbus Roman No9 L"';
+            this.ctx.font = this.current.fontSize +'px "' + fontName + '"';
         },
         moveText: function (x, y) {
             this.current.x = this.current.lineX += x;

diff --git a/test.html b/test.html
index f14533add94ee86eb0f0389c1c4f77e24ffbca18..ac568473a71b9770d1e972c4b83bd2d50743fc4a 100644 (file)
--- a/test.html
+++ b/test.html
@@ -5,6 +5,7 @@
 
         <script type="text/javascript" src="pdf.js"></script>
         <script type="text/javascript" src="test.js"></script>
+        <script type="text/javascript" src="PDFFont.js"></script>
   </head>
 
   <body onload="load();">