Merge pull request #3376 from yurydelendik/bad-pdfs

Misc fixes for corrupted PDFs

src/chunked_stream.js

@@ -141,6 +141,12 @@ var ChunkedStream = (function ChunkedStreamClosure() {
       return bytes.subarray(pos, end);
     },
 
+    peekBytes: function ChunkedStream_peekBytes(length) {
+      var bytes = this.getBytes(length);
+      this.pos -= bytes.length;
+      return bytes;
+    },
+
     getByteRange: function ChunkedStream_getBytes(begin, end) {
       this.ensureRange(begin, end);
       return this.bytes.subarray(begin, end);

src/evaluator.js

@@ -124,7 +124,7 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
     // Images
     BI: { fnName: 'beginInlineImage', numArgs: 0, variableArgs: false },
     ID: { fnName: 'beginImageData', numArgs: 0, variableArgs: false },
-    EI: { fnName: 'endInlineImage', numArgs: 0, variableArgs: false },
+    EI: { fnName: 'endInlineImage', numArgs: 1, variableArgs: false },
 
     // XObjects
     Do: { fnName: 'paintXObject', numArgs: 1, variableArgs: false },

src/obj.js

@@ -266,6 +266,18 @@ var Catalog = (function CatalogClosure() {
       return shadow(this, 'toplevelPagesDict', pagesObj);
     },
     get documentOutline() {
+      var obj = null;
+      try {
+        obj = this.readDocumentOutline();
+      } catch (ex) {
+        if (ex instanceof MissingDataException) {
+          throw ex;
+        }
+        warn('Unable to read document outline');
+      }
+      return shadow(this, 'documentOutline', obj);
+    },
+    readDocumentOutline: function Catalog_readDocumentOutline() {
       var xref = this.xref;
       var obj = this.catDict.get('Outlines');
       var root = { items: [] };
@@ -316,8 +328,7 @@ var Catalog = (function CatalogClosure() {
           }
         }
       }
-      obj = root.items.length > 0 ? root.items : null;
+      return root.items.length > 0 ? root.items : null;
-      return shadow(this, 'documentOutline', obj);
     },
     get numPages() {
       var obj = this.toplevelPagesDict.get('Count');
@@ -598,6 +609,12 @@ var XRef = (function XRefClosure() {
         delete tableState.entryCount;
       }
 
+      // Per issue 3248: hp scanners generate bad XRef
+      if (first === 1 && this.entries[1] && this.entries[1].free) {
+        // shifting the entries
+        this.entries.shift();
+      }
+
       // Sanity check: as per spec, first object must be free
       if (this.entries[0] && !this.entries[0].free)
         error('Invalid XRef table: unexpected first object');

src/parser.js

@@ -217,9 +217,47 @@ var Parser = (function ParserClosure() {
       stream.pos = pos + length;
       this.shift(); // '>>'
       this.shift(); // 'stream'
-      if (!isCmd(this.buf1, 'endstream'))
+      if (!isCmd(this.buf1, 'endstream')) {
-        error('Missing endstream');
+        // bad stream length, scanning for endstream
-      this.shift();
+        stream.pos = pos;
+        var SCAN_BLOCK_SIZE = 2048;
+        var ENDSTREAM_SIGNATURE_LENGTH = 9;
+        var ENDSTREAM_SIGNATURE = [0x65, 0x6E, 0x64, 0x73, 0x74, 0x72, 0x65,
+                                   0x61, 0x6D];
+        var skipped = 0, found = false;
+        while (stream.pos < stream.end) {
+          var scanBytes = stream.peekBytes(SCAN_BLOCK_SIZE);
+          var scanLength = scanBytes.length - ENDSTREAM_SIGNATURE_LENGTH;
+          var found = false, i, ii, j;
+          for (i = 0, j = 0; i < scanLength; i++) {
+            var b = scanBytes[i];
+            if (b !== ENDSTREAM_SIGNATURE[j]) {
+              i -= j;
+              j = 0;
+            } else {
+              j++;
+              if (j >= ENDSTREAM_SIGNATURE_LENGTH) {
+                found = true;
+                break;
+              }
+            }
+          }
+          if (found) {
+            skipped += i - ENDSTREAM_SIGNATURE_LENGTH;
+            stream.pos += i - ENDSTREAM_SIGNATURE_LENGTH;
+            break;
+          }
+          skipped += scanLength;
+          stream.pos += scanLength;
+        }
+        if (!found) {
+          error('Missing endstream');
+        }
+        length = skipped;
+        this.shift();
+        this.shift();
+      }
+      this.shift(); // 'endstream'
 
       stream = stream.makeSubStream(pos, length, dict);
       if (cipherTransform)

src/stream.js

@@ -504,8 +504,11 @@ var FlateStream = (function FlateStreamClosure() {
       if (typeof (b = bytes[bytesPos++]) == 'undefined')
         error('Bad block header in flate stream');
       check |= (b << 8);
-      if (check != (~blockLen & 0xffff))
+      if (check != (~blockLen & 0xffff) &&
+          (blockLen !== 0 || check !== 0)) {
+        // Ignoring error for bad "empty" block (see issue 1277)
         error('Bad uncompressed block length in flate stream');
+      }
 
       this.codeBuf = 0;
       this.codeSize = 0;