Merge pull request #9827 from Snuffleupagus/misc-corrupt-pdf-fixes

Fix various corrupt PDF files (issue 9252, issue 9418)

src/core/obj.js

@@ -855,7 +855,16 @@ var XRef = (function XRefClosure() {
       }
       trailerDict.assignXref(this);
       this.trailer = trailerDict;
-      var encrypt = trailerDict.get('Encrypt');
+
+      let encrypt;
+      try {
+        encrypt = trailerDict.get('Encrypt');
+      } catch (ex) {
+        if (ex instanceof MissingDataException) {
+          throw ex;
+        }
+        warn(`XRef.parse - Invalid "Encrypt" reference: "${ex}".`);
+      }
       if (isDict(encrypt)) {
         var ids = trailerDict.get('ID');
         var fileId = (ids && ids.length) ? ids[0] : '';
@@ -868,8 +877,22 @@ var XRef = (function XRefClosure() {
                                                   this.pdfManager.password);
       }
 
-      // get the root dictionary (catalog) object
-      if (!(this.root = trailerDict.get('Root'))) {
+      // Get the root dictionary (catalog) object, and do some basic validation.
+      let root;
+      try {
+        root = trailerDict.get('Root');
+      } catch (ex) {
+        if (ex instanceof MissingDataException) {
+          throw ex;
+        }
+        warn(`XRef.parse - Invalid "Root" reference: "${ex}".`);
+      }
+      if (isDict(root) && root.has('Pages')) {
+        this.root = root;
+      } else {
+        if (!recoveryMode) {
+          throw new XRefParseException();
+        }
         throw new FormatError('Invalid root reference');
       }
     },
@@ -1208,7 +1231,7 @@ var XRef = (function XRefClosure() {
                 break;
               }
             }
-            startPos += contentLength;
+            startPos = endPos;
           }
           let content = buffer.subarray(position, position + contentLength);
 
@@ -1237,7 +1260,7 @@ var XRef = (function XRefClosure() {
         this.readXRef(/* recoveryMode */ true);
       }
       // finding main trailer
-      var dict;
+      let trailerDict;
       for (i = 0, ii = trailers.length; i < ii; ++i) {
         stream.pos = trailers[i];
         var parser = new Parser(new Lexer(stream), /* allowStreams = */ true,
@@ -1247,18 +1270,33 @@ var XRef = (function XRefClosure() {
           continue;
         }
         // read the trailer dictionary
-        dict = parser.getObj();
+        let dict = parser.getObj();
         if (!isDict(dict)) {
           continue;
         }
+        // Do some basic validation of the trailer/root dictionary candidate.
+        let rootDict;
+        try {
+          rootDict = dict.get('Root');
+        } catch (ex) {
+          if (ex instanceof MissingDataException) {
+            throw ex;
+          }
+          continue;
+        }
+        if (!isDict(rootDict) || !rootDict.has('Pages')) {
+          continue;
+        }
         // taking the first one with 'ID'
         if (dict.has('ID')) {
           return dict;
         }
+        // The current dictionary is a candidate, but continue searching.
+        trailerDict = dict;
       }
-      // no tailer with 'ID', taking last one (if exists)
-      if (dict) {
-        return dict;
+      // No trailer with 'ID', taking last one (if exists).
+      if (trailerDict) {
+        return trailerDict;
       }
       // nothing helps
       throw new InvalidPDFException('Invalid PDF structure');

src/core/parser.js

@@ -18,8 +18,8 @@ import {
   PredictorStream, RunLengthStream
 } from './stream';
 import {
-  assert, FormatError, info, isNum, isString, MissingDataException, StreamType,
-  warn
+  assert, FormatError, info, isNum, isSpace, isString, MissingDataException,
+  StreamType, warn
 } from '../shared/util';
 import {
   Cmd, Dict, EOF, isCmd, isDict, isEOF, isName, Name, Ref
@@ -721,7 +721,7 @@ var Lexer = (function LexerClosure() {
       var ch = this.currentChar;
       var eNotation = false;
       var divideBy = 0; // different from 0 if it's a floating point value
-      var sign = 1;
+      var sign = 0;
 
       if (ch === 0x2D) { // '-'
         sign = -1;
@@ -732,10 +732,7 @@ var Lexer = (function LexerClosure() {
           ch = this.nextChar();
         }
       } else if (ch === 0x2B) { // '+'
-        ch = this.nextChar();
-      }
-      if (ch === 0x2E) { // '.'
-        divideBy = 10;
+        sign = 1;
         ch = this.nextChar();
       }
       if (ch === 0x0A || ch === 0x0D) { // LF, CR
@@ -744,11 +741,22 @@ var Lexer = (function LexerClosure() {
           ch = this.nextChar();
         } while (ch === 0x0A || ch === 0x0D);
       }
+      if (ch === 0x2E) { // '.'
+        divideBy = 10;
+        ch = this.nextChar();
+      }
       if (ch < 0x30 || ch > 0x39) { // '0' - '9'
+        if (divideBy === 10 && sign === 0 &&
+            (isSpace(ch) || ch === /* EOF = */ -1)) {
+          // This is consistent with Adobe Reader (fixes issue9252.pdf).
+          warn('Lexer.getNumber - treating a single decimal point as zero.');
+          return 0;
+        }
         throw new FormatError(
           `Invalid number: ${String.fromCharCode(ch)} (charCode ${ch})`);
       }
 
+      sign = sign || 1;
       var baseValue = ch - 0x30; // '0'
       var powerValue = 0;
       var powerValueSign = 1;

test/pdfs/.gitignore

@@ -65,8 +65,10 @@
 !issue8823.pdf
 !issue9084.pdf
 !issue9105_reduced.pdf
+!issue9252.pdf
 !issue9262_reduced.pdf
 !issue9291.pdf
+!issue9418.pdf
 !issue9458.pdf
 !bad-PageLabels.pdf
 !decodeACSuccessive.pdf

test/pdfs/issue9252.pdf

@@ -0,0 +1,90 @@
+%PDF-1.4
+1 0 obj
+<<
+/Type /Catalog
+/Version /1.4
+/Pages 5 0 R
+/Outlines 3 0 R
+>>
+endobj
+2 0 obj
+<<
+/Title (Test)
+/Author (Test)
+/Creator (sharpPDF)
+/CreationDate (2017124)
+>>
+endobj
+3 0 obj
+<<
+/Type /Outlines
+/Count 0
+>>
+endobj
+4 0 obj
+<<
+/Type /Font
+/Subtype /Type1
+/Name /F1
+/BaseFont /Helvetica
+/Encoding /WinAnsiEncoding
+>>
+endobj
+5 0 obj
+<<
+/Type /Pages
+/Count 1
+/Kids [6 0 R ]
+>>
+endobj
+6 0 obj
+<<
+/Type /Page
+/Parent 5 0 R
+/Resources <</Font <</F1 4 0 R >>
+>>
+/MediaBox [0 0 612 792]
+/CropBox [0 0 612 792]
+/Rotate 0
+/ProcSet [/PDF /Text /ImageC]
+/Contents [7 0 R ]
+>>
+endobj
+7 0 obj
+<<
+/Filter [/ASCIIHexDecode]
+/Length 105
+>>
+stream
+710A42540A2F46312031352054660A2E202E3539202E38342072670A3530203735302054640A28546573742920546A0A45540A51>
+endstream
+endobj
+0 19
+0000000000 65535 f
+0000000010 00000 n
+0000000098 00000 n
+0000000202 00000 n
+0000000254 00000 n
+0000000370 00000 n
+0000000491 00000 n
+0000000615 00000 n
+0000000743 00000 n
+0000000857 00000 n
+0000000976 00000 n
+0000001099 00000 n
+0000001226 00000 n
+0000001345 00000 n
+0000001464 00000 n
+0000001585 00000 n
+0000001710 00000 n
+0000001777 00000 n
+0000002102 00000 n
+trailer
+<<
+/Size 102
+/Root 1 0 R
+/Info 2 0 R
+>>
+startxref
+36673
+%%EOF

test/test_manifest.json

@@ -741,6 +741,20 @@
        "lastPage": 1,
        "type": "eq"
     },
+    {  "id": "issue9252",
+       "file": "pdfs/issue9252.pdf",
+       "md5": "c7d039d808d9344a95d2c9cfa7586ca3",
+       "rounds": 1,
+       "link": false,
+       "type": "eq"
+    },
+    {  "id": "issue9418",
+       "file": "pdfs/issue9418.pdf",
+       "md5": "32ecad8098acb1938539d47944ecb54b",
+       "rounds": 1,
+       "link": false,
+       "type": "eq"
+    },
     {  "id": "issue9262",
        "file": "pdfs/issue9262_reduced.pdf",
        "md5": "5347ce2d7b3866625c22e115fd90e0de",

test/unit/parser_spec.js

@@ -14,6 +14,7 @@
  */
 
 import { Lexer, Linearization } from '../../src/core/parser';
+import { FormatError } from '../../src/shared/util';
 import { Name } from '../../src/core/primitives';
 import { StringStream } from '../../src/core/stream';
 
@@ -58,11 +59,32 @@ describe('parser', function() {
 
     it('should ignore line-breaks between operator and digit in number',
         function() {
-      var input = new StringStream('-\r\n205.88');
-      var lexer = new Lexer(input);
-      var result = lexer.getNumber();
+      let minusInput = new StringStream('-\r\n205.88');
+      let minusLexer = new Lexer(minusInput);
 
-      expect(result).toEqual(-205.88);
+      expect(minusLexer.getNumber()).toEqual(-205.88);
+
+      let plusInput = new StringStream('+\r\n205.88');
+      let plusLexer = new Lexer(plusInput);
+
+      expect(plusLexer.getNumber()).toEqual(205.88);
+    });
+
+    it('should treat a single decimal point as zero', function() {
+      let input = new StringStream('.');
+      let lexer = new Lexer(input);
+
+      expect(lexer.getNumber()).toEqual(0);
+
+      let numbers = ['..', '-.', '+.', '-\r\n.', '+\r\n.'];
+      for (let number of numbers) {
+        let input = new StringStream(number);
+        let lexer = new Lexer(input);
+
+        expect(function() {
+          return lexer.getNumber();
+        }).toThrowError(FormatError, /^Invalid number:\s/);
+      }
     });
 
     it('should handle glued numbers and operators', function() {