Skip commands that have too few arguments

- Commands that have too few args will be skipped
- Commands that have too many args will generate an info, but still
execute

src/evaluator.js

@@ -30,101 +30,104 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
     this.fontIdCounter = 0;
   }
 
+  // Specifies properties for each command
+  //
+  // If variableArgs === true: [0, `numArgs`] expected
+  // If variableArgs === false: exactly `numArgs` expected
   var OP_MAP = {
-    // Graphics state
+    // Graphic state
-    w: 'setLineWidth',
+    w: { fnName: 'setLineWidth', numArgs: 1, variableArgs: false },
-    J: 'setLineCap',
+    J: { fnName: 'setLineCap', numArgs: 1, variableArgs: false },
-    j: 'setLineJoin',
+    j: { fnName: 'setLineJoin', numArgs: 1, variableArgs: false },
-    M: 'setMiterLimit',
+    M: { fnName: 'setMiterLimit', numArgs: 1, variableArgs: false },
-    d: 'setDash',
+    d: { fnName: 'setDash', numArgs: 2, variableArgs: false },
-    ri: 'setRenderingIntent',
+    ri: { fnName: 'setRenderingIntent', numArgs: 1, variableArgs: false },
-    i: 'setFlatness',
+    i: { fnName: 'setFlatness', numArgs: 1, variableArgs: false },
-    gs: 'setGState',
+    gs: { fnName: 'setGState', numArgs: 1, variableArgs: false },
-    q: 'save',
+    q: { fnName: 'save', numArgs: 0, variableArgs: false },
-    Q: 'restore',
+    Q: { fnName: 'restore', numArgs: 0, variableArgs: false },
-    cm: 'transform',
+    cm: { fnName: 'transform', numArgs: 6, variableArgs: false },
 
     // Path
-    m: 'moveTo',
+    m: { fnName: 'moveTo', numArgs: 2, variableArgs: false },
-    l: 'lineTo',
+    l: { fnName: 'lineTo', numArgs: 2, variableArgs: false },
-    c: 'curveTo',
+    c: { fnName: 'curveTo', numArgs: 6, variableArgs: false },
-    v: 'curveTo2',
+    v: { fnName: 'curveTo2', numArgs: 4, variableArgs: false },
-    y: 'curveTo3',
+    y: { fnName: 'curveTo3', numArgs: 4, variableArgs: false },
-    h: 'closePath',
+    h: { fnName: 'closePath', numArgs: 0, variableArgs: false },
-    re: 'rectangle',
+    re: { fnName: 'rectangle', numArgs: 4, variableArgs: false },
-    S: 'stroke',
+    S: { fnName: 'stroke', numArgs: 0, variableArgs: false },
-    s: 'closeStroke',
+    s: { fnName: 'closeStroke', numArgs: 0, variableArgs: false },
-    f: 'fill',
+    f: { fnName: 'fill', numArgs: 0, variableArgs: false },
-    F: 'fill',
+    F: { fnName: 'fill', numArgs: 0, variableArgs: false },
-    'f*': 'eoFill',
+    'f*': { fnName: 'eoFill', numArgs: 0, variableArgs: false },
-    B: 'fillStroke',
+    B: { fnName: 'fillStroke', numArgs: 0, variableArgs: false },
-    'B*': 'eoFillStroke',
+    'B*': { fnName: 'eoFillStroke', numArgs: 0, variableArgs: false },
-    b: 'closeFillStroke',
+    b: { fnName: 'closeFillStroke', numArgs: 0, variableArgs: false },
-    'b*': 'closeEOFillStroke',
+    'b*': { fnName: 'closeEOFillStroke', numArgs: 0, variableArgs: false },
-    n: 'endPath',
+    n: { fnName: 'endPath', numArgs: 0, variableArgs: false },
 
     // Clipping
-    W: 'clip',
+    W: { fnName: 'clip', numArgs: 0, variableArgs: false },
-    'W*': 'eoClip',
+    'W*': { fnName: 'eoClip', numArgs: 0, variableArgs: false },
 
     // Text
-    BT: 'beginText',
+    BT: { fnName: 'beginText', numArgs: 0, variableArgs: false },
-    ET: 'endText',
+    ET: { fnName: 'endText', numArgs: 0, variableArgs: false },
-    Tc: 'setCharSpacing',
+    Tc: { fnName: 'setCharSpacing', numArgs: 1, variableArgs: false },
-    Tw: 'setWordSpacing',
+    Tw: { fnName: 'setWordSpacing', numArgs: 1, variableArgs: false },
-    Tz: 'setHScale',
+    Tz: { fnName: 'setHScale', numArgs: 1, variableArgs: false },
-    TL: 'setLeading',
+    TL: { fnName: 'setLeading', numArgs: 1, variableArgs: false },
-    Tf: 'setFont',
+    Tf: { fnName: 'setFont', numArgs: 2, variableArgs: false },
-    Tr: 'setTextRenderingMode',
+    Tr: { fnName: 'setTextRenderingMode', numArgs: 1, variableArgs: false },
-    Ts: 'setTextRise',
+    Ts: { fnName: 'setTextRise', numArgs: 1, variableArgs: false },
-    Td: 'moveText',
+    Td: { fnName: 'moveText', numArgs: 2, variableArgs: false },
-    TD: 'setLeadingMoveText',
+    TD: { fnName: 'setLeadingMoveText', numArgs: 2, variableArgs: false },
-    Tm: 'setTextMatrix',
+    Tm: { fnName: 'setTextMatrix', numArgs: 6, variableArgs: false },
-    'T*': 'nextLine',
+    'T*': { fnName: 'nextLine', numArgs: 0, variableArgs: false },
-    Tj: 'showText',
+    Tj: { fnName: 'showText', numArgs: 1, variableArgs: false },
-    TJ: 'showSpacedText',
+    TJ: { fnName: 'showSpacedText', numArgs: 1, variableArgs: false },
-    "'": 'nextLineShowText',
+    '\'': { fnName: 'nextLineShowText', numArgs: 1, variableArgs: false },
-    '"': 'nextLineSetSpacingShowText',
+    '"': { fnName: 'nextLineSetSpacingShowText', numArgs: 3,
+      variableArgs: false },
 
     // Type3 fonts
-    d0: 'setCharWidth',
+    d0: { fnName: 'setCharWidth', numArgs: 2, variableArgs: false },
-    d1: 'setCharWidthAndBounds',
+    d1: { fnName: 'setCharWidthAndBounds', numArgs: 6, variableArgs: false },
 
     // Color
-    CS: 'setStrokeColorSpace',
+    CS: { fnName: 'setStrokeColorSpace', numArgs: 1, variableArgs: false },
-    cs: 'setFillColorSpace',
+    cs: { fnName: 'setFillColorSpace', numArgs: 1, variableArgs: false },
-    SC: 'setStrokeColor',
+    SC: { fnName: 'setStrokeColor', numArgs: 4, variableArgs: true },
-    SCN: 'setStrokeColorN',
+    SCN: { fnName: 'setStrokeColorN', numArgs: 33, variableArgs: true },
-    sc: 'setFillColor',
+    sc: { fnName: 'setFillColor', numArgs: 4, variableArgs: true },
-    scn: 'setFillColorN',
+    scn: { fnName: 'setFillColorN', numArgs: 33, variableArgs: true },
-    G: 'setStrokeGray',
+    G: { fnName: 'setStrokeGray', numArgs: 1, variableArgs: false },
-    g: 'setFillGray',
+    g: { fnName: 'setFillGray', numArgs: 1, variableArgs: false },
-    RG: 'setStrokeRGBColor',
+    RG: { fnName: 'setStrokeRGBColor', numArgs: 3, variableArgs: false },
-    rg: 'setFillRGBColor',
+    rg: { fnName: 'setFillRGBColor', numArgs: 3, variableArgs: false },
-    K: 'setStrokeCMYKColor',
+    K: { fnName: 'setStrokeCMYKColor', numArgs: 4, variableArgs: false },
-    k: 'setFillCMYKColor',
+    k: { fnName: 'setFillCMYKColor', numArgs: 4, variableArgs: false },
 
     // Shading
-    sh: 'shadingFill',
+    sh: { fnName: 'shadingFill', numArgs: 1, variableArgs: false },
 
     // Images
-    BI: 'beginInlineImage',
+    BI: { fnName: 'beginInlineImage', numArgs: 0, variableArgs: false },
-    ID: 'beginImageData',
+    ID: { fnName: 'beginImageData', numArgs: 0, variableArgs: false },
-    EI: 'endInlineImage',
+    EI: { fnName: 'endInlineImage', numArgs: 0, variableArgs: false },
 
     // XObjects
-    Do: 'paintXObject',
+    Do: { fnName: 'paintXObject', numArgs: 1, variableArgs: false },
-
+    MP: { fnName: 'markPoint', numArgs: 1, variableArgs: false },
-    // Marked content
+    DP: { fnName: 'markPointProps', numArgs: 2, variableArgs: false },
-    MP: 'markPoint',
+    BMC: { fnName: 'beginMarkedContent', numArgs: 1, variableArgs: false },
-    DP: 'markPointProps',
+    BDC: { fnName: 'beginMarkedContentProps', numArgs: 2, variableArgs: false },
-    BMC: 'beginMarkedContent',
+    EMC: { fnName: 'endMarkedContent', numArgs: 0, variableArgs: false },
-    BDC: 'beginMarkedContentProps',
-    EMC: 'endMarkedContent',
 
     // Compatibility
-    BX: 'beginCompat',
+    BX: { fnName: 'beginCompat', numArgs: 0, variableArgs: false },
-    EX: 'endCompat',
+    EX: { fnName: 'endCompat', numArgs: 0, variableArgs: false },
 
     // (reserved partial commands for the lexer)
     BM: null,
@@ -314,6 +317,8 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
       resources = resources || new Dict();
       var xobjs = resources.get('XObject') || new Dict();
       var patterns = resources.get('Pattern') || new Dict();
+      // TODO(mduan): pass array of knownCommands rather than OP_MAP
+      // dictionary
       var parser = new Parser(new Lexer(stream, OP_MAP), false, xref);
       var res = resources;
       var args = [], obj;
@@ -321,13 +326,42 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
 
       while (true) {
         obj = parser.getObj();
-        if (isEOF(obj))
+        if (isEOF(obj)) {
           break;
+        }
 
         if (isCmd(obj)) {
           var cmd = obj.cmd;
-          var fn = OP_MAP[cmd];
+
-          assertWellFormed(fn, 'Unknown command "' + cmd + '"');
+          // Check that the command is valid
+          var opSpec = OP_MAP[cmd];
+          if (!opSpec) {
+            warn('Unknown command "' + cmd + '"');
+            continue;
+          }
+
+          var fn = opSpec.fnName;
+
+          // Validate the number of arguments for the command
+          if (opSpec.variableArgs) {
+            if (args.length > opSpec.numArgs) {
+              info('Command ' + fn + ': expected [0,' + opSpec.numArgs +
+                  '] args, but received ' + args.length + ' args');
+            }
+          } else {
+            if (args.length < opSpec.numArgs) {
+              // If we receive too few args, it's not possible to possible
+              // to execute the command, so skip the command
+              info('Command ' + fn + ': because expected ' + opSpec.numArgs +
+                  ' args, but received ' + args.length + ' args; skipping');
+              args = [];
+              continue;
+            } else if (args.length > opSpec.numArgs) {
+              info('Command ' + fn + ': expected ' + opSpec.numArgs +
+                  ' args, but received ' + args.length + ' args');
+            }
+          }
+
           // TODO figure out how to type-check vararg functions
 
           if ((cmd == 'SCN' || cmd == 'scn') && !args[args.length - 1].code) {
@@ -509,8 +543,8 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
           argsArray.push(args);
           args = [];
         } else if (obj != null) {
-          assertWellFormed(args.length <= 33, 'Too many arguments');
           args.push(obj instanceof Dict ? obj.getAll() : obj);
+          assertWellFormed(args.length <= 33, 'Too many arguments');
         }
       }
 

test/pdfs/.gitignore

@@ -1,6 +1,7 @@
 *.pdf
 
 !tracemonkey.pdf
+!issue2391-1.pdf
 !ArabicCIDTrueType.pdf
 !ThuluthFeatures.pdf
 !arial_unicode_ab_cidfont.pdf

test/pdfs/issue2391-1.pdf

@@ -0,0 +1,131 @@
+%PDF-1.4
+%âãÏÓ
+1 0 obj 
+<<
+/Pages 2 0 R
+/Metadata 3 0 R
+/Type /Catalog
+>>
+endobj 
+2 0 obj 
+<<
+/Kids [4 0 R]
+/Count 1
+/Type /Pages
+>>
+endobj 
+4 0 obj 
+<<
+/Rotate 0
+/Parent 2 0 R
+/Resources 
+<<
+/ExtGState 5 0 R
+/Font 6 0 R
+/ProcSet [/PDF /Text]
+>>
+/MediaBox [0 0 595 842]
+/Contents 7 0 R
+/Type /Page
+>>
+endobj 
+7 0 obj 
+<<
+/Length 148
+>>
+stream
+q 0.1 0 0 0.1 0 0 cm
+/R7 gs
+0 g
+q
+10 0 0 10 0 0 cm BT
+undefined 10 Tf
+/R8 10 Tf
+1 0 0 1 29 805 Tm
+(test command with wrong number of args)Tj
+ET
+Q
+Q
+
+endstream 
+endobj 
+8 0 obj 
+<<
+/Type /ExtGState
+/OPM 1
+>>
+endobj 
+9 0 obj 
+<<
+/BaseFont /Courier
+/Subtype /Type1
+/Type /Font
+>>
+endobj 
+5 0 obj 
+<<
+/R7 8 0 R
+>>
+endobj 
+6 0 obj 
+<<
+/R8 9 0 R
+>>
+endobj 
+3 0 obj 
+<<
+/Subtype /XML
+/Length 1421
+/Type /Metadata
+>>
+stream
+<?xpacket begin='' id='W5M0MpCehiHzreSzNTczkc9d'?>
+<?adobe-xap-filters esc="CRLF"?>
+<x:xmpmeta xmlns:x='adobe:ns:meta/' x:xmptk='XMP toolkit 2.9.1-13, framework 1.6'>
+<rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#' xmlns:iX='http://ns.adobe.com/iX/1.0/'>
+<rdf:Description rdf:about='uuid:39048ebf-93ad-11ed-0000-3d9120169d5e' xmlns:pdf='http://ns.adobe.com/pdf/1.3/' pdf:Producer='GPL Ghostscript 9.05'/>
+<rdf:Description rdf:about='uuid:39048ebf-93ad-11ed-0000-3d9120169d5e' xmlns:xmp='http://ns.adobe.com/xap/1.0/'><xmp:ModifyDate>2013-01-10T17:46:10-08:00</xmp:ModifyDate>
+<xmp:CreateDate>2013-01-10T17:46:10-08:00</xmp:CreateDate>
+<xmp:CreatorTool>GNU Enscript 1.6.6</xmp:CreatorTool></rdf:Description>
+<rdf:Description rdf:about='uuid:39048ebf-93ad-11ed-0000-3d9120169d5e' xmlns:xapMM='http://ns.adobe.com/xap/1.0/mm/' xapMM:DocumentID='uuid:39048ebf-93ad-11ed-0000-3d9120169d5e'/>
+<rdf:Description rdf:about='uuid:39048ebf-93ad-11ed-0000-3d9120169d5e' xmlns:dc='http://purl.org/dc/elements/1.1/' dc:format='application/pdf'><dc:title><rdf:Alt><rdf:li xml:lang='x-default'>Enscript Output</rdf:li></rdf:Alt></dc:title><dc:creator><rdf:Seq><rdf:li>Mack Duan</rdf:li></rdf:Seq></dc:creator></rdf:Description>
+</rdf:RDF>
+</x:xmpmeta>
+                                                                        
+                                                                        
+<?xpacket end='w'?>
+endstream 
+endobj 
+10 0 obj 
+<<
+/Creator (GNU Enscript 1.6.6)
+/Title (Enscript Output)
+/Producer (GPL Ghostscript 9.05)
+/Author (Mack Duan)
+/ModDate (D:20130110174610-08'00')
+/CreationDate (D:20130110174610-08'00')
+>>
+endobj xref
+0 11
+0000000000 65535 f 
+0000000015 00000 n 
+0000000082 00000 n 
+0000000694 00000 n 
+0000000141 00000 n 
+0000000628 00000 n 
+0000000661 00000 n 
+0000000309 00000 n 
+0000000511 00000 n 
+0000000558 00000 n 
+0000002200 00000 n 
+trailer
+
+<<
+/Info 10 0 R
+/Root 1 0 R
+/Size 11
+/ID [<5807376a23851d4cc6096009432892a3> <5807376a23851d4cc6096009432892a3>]
+>>
+startxref
+2406
+%%EOF

test/test_manifest.json

@@ -17,6 +17,13 @@
        "rounds": 1,
        "type": "text"
     },
+    {  "id": "issue2391-1",
+      "file": "pdfs/issue2391-1.pdf",
+      "md5": "25ae9cb959612e7b343b55da63af2716",
+      "rounds": 1,
+      "pageLimit": 1,
+      "type": "load"
+    },
     {  "id": "html5-canvas-cheat-sheet-load",
        "file": "pdfs/canvas.pdf",
        "md5": "59510028561daf62e00bf9f6f066b033",

test/unit/evaluator_spec.js

@@ -32,13 +32,12 @@ describe('evaluator', function() {
       var evaluator = new PartialEvaluator(new XrefMock(), new HandlerMock(),
                                            'prefix');
       var stream = new StringStream('qTT');
-      var thrown = false;
+      var result = evaluator.getOperatorList(stream, new ResourcesMock(), []);
-      try {
+
-        evaluator.getOperatorList(stream, new ResourcesMock(), []);
+      expect(!!result.fnArray && !!result.argsArray).toEqual(true);
-      } catch (e) {
+      expect(result.fnArray.length).toEqual(1);
-        thrown = e;
+      expect(result.fnArray[0]).toEqual('save');
-      }
+      expect(result.argsArray[0].length).toEqual(0);
-      expect(thrown).toNotEqual(false);
     });
 
     it('should handle one operations', function() {
@@ -84,14 +83,14 @@ describe('evaluator', function() {
                                            'prefix');
       var resources = new ResourcesMock();
       resources.Res1 = {};
-      var stream = new StringStream('B*BBMC');
+      var stream = new StringStream('B*Bf*');
       var result = evaluator.getOperatorList(stream, resources, []);
 
       expect(!!result.fnArray && !!result.argsArray).toEqual(true);
       expect(result.fnArray.length).toEqual(3);
       expect(result.fnArray[0]).toEqual('eoFillStroke');
       expect(result.fnArray[1]).toEqual('fillStroke');
-      expect(result.fnArray[2]).toEqual('beginMarkedContent');
+      expect(result.fnArray[2]).toEqual('eoFill');
     });
 
     it('should handle glued operations and operands', function() {
@@ -112,19 +111,53 @@ describe('evaluator', function() {
     it('should handle glued operations and literals', function() {
       var evaluator = new PartialEvaluator(new XrefMock(), new HandlerMock(),
                                            'prefix');
-      var stream = new StringStream('trueifalserinulli');
+      var stream = new StringStream('trueifalserinullq');
       var result = evaluator.getOperatorList(stream, new ResourcesMock(), []);
 
       expect(!!result.fnArray && !!result.argsArray).toEqual(true);
       expect(result.fnArray.length).toEqual(3);
       expect(result.fnArray[0]).toEqual('setFlatness');
       expect(result.fnArray[1]).toEqual('setRenderingIntent');
-      expect(result.fnArray[2]).toEqual('setFlatness');
+      expect(result.fnArray[2]).toEqual('save');
       expect(result.argsArray.length).toEqual(3);
       expect(result.argsArray[0].length).toEqual(1);
       expect(result.argsArray[0][0]).toEqual(true);
       expect(result.argsArray[1].length).toEqual(1);
       expect(result.argsArray[1][0]).toEqual(false);
+      expect(result.argsArray[2].length).toEqual(0);
+    });
+  });
+
+  describe('validateNumberOfArgs', function() {
+    it('should execute if correct number of arguments', function() {
+      var evaluator = new PartialEvaluator(new XrefMock(), new HandlerMock(),
+                                           'prefix');
+      var stream = new StringStream('5 1 d0');
+      var result = evaluator.getOperatorList(stream, new ResourcesMock(), []);
+
+      expect(result.argsArray[0][0]).toEqual(5);
+      expect(result.argsArray[0][1]).toEqual(1);
+      expect(result.fnArray[0]).toEqual('setCharWidth');
+    });
+    it('should execute if too many arguments', function() {
+      var evaluator = new PartialEvaluator(new XrefMock(), new HandlerMock(),
+                                           'prefix');
+      var stream = new StringStream('5 1 4 d0');
+      var result = evaluator.getOperatorList(stream, new ResourcesMock(), []);
+
+      expect(result.argsArray[0][0]).toEqual(5);
+      expect(result.argsArray[0][1]).toEqual(1);
+      expect(result.argsArray[0][2]).toEqual(4);
+      expect(result.fnArray[0]).toEqual('setCharWidth');
+    });
+    it('should skip if too few arguments', function() {
+      var evaluator = new PartialEvaluator(new XrefMock(), new HandlerMock(),
+                                           'prefix');
+      var stream = new StringStream('5 d0');
+      var result = evaluator.getOperatorList(stream, new ResourcesMock(), []);
+
+      expect(result.argsArray).toEqual([]);
+      expect(result.fnArray).toEqual([]);
     });
   });
 });