Merge pull request #11680 from Snuffleupagus/hasBlendModes-fetch-errors

Prevent lookup errors in `PartialEvaluator.hasBlendModes` from breaking all parsing/rendering of a page (issue 11678)

src/core/evaluator.js

@@ -61,6 +61,7 @@ import {
   WinAnsiEncoding,
   ZapfDingbatsEncoding,
 } from "./encodings.js";
+import { getLookupTableFactory, MissingDataException } from "./core_utils.js";
 import {
   getNormalizedUnicodes,
   getUnicodeForGlyph,
@@ -77,7 +78,6 @@ import { bidi } from "./bidi.js";
 import { ColorSpace } from "./colorspace.js";
 import { DecodeStream } from "./stream.js";
 import { getGlyphsUnicode } from "./glyphlist.js";
-import { getLookupTableFactory } from "./core_utils.js";
 import { getMetrics } from "./metrics.js";
 import { isPDFFunction } from "./function.js";
 import { JpegStream } from "./jpeg_stream.js";
@@ -266,7 +266,27 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
               if (processed[graphicState.toString()]) {
                 continue; // The ExtGState has already been processed.
               }
-              graphicState = xref.fetch(graphicState);
+              try {
+                graphicState = xref.fetch(graphicState);
+              } catch (ex) {
+                if (ex instanceof MissingDataException) {
+                  throw ex;
+                }
+                if (this.options.ignoreErrors) {
+                  if (graphicState instanceof Ref) {
+                    // Avoid parsing a corrupt ExtGState more than once.
+                    processed[graphicState.toString()] = true;
+                  }
+                  // Error(s) in the ExtGState -- sending unsupported feature
+                  // notification and allow parsing/rendering to continue.
+                  this.handler.send("UnsupportedFeature", {
+                    featureId: UNSUPPORTED_FEATURES.unknown,
+                  });
+                  warn(`hasBlendModes - ignoring ExtGState: "${ex}".`);
+                  continue;
+                }
+                throw ex;
+              }
             }
             if (!(graphicState instanceof Dict)) {
               continue;
@@ -308,7 +328,27 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
               // time for badly generated PDF files (fixes issue6961.pdf).
               continue;
             }
-            xObject = xref.fetch(xObject);
+            try {
+              xObject = xref.fetch(xObject);
+            } catch (ex) {
+              if (ex instanceof MissingDataException) {
+                throw ex;
+              }
+              if (this.options.ignoreErrors) {
+                if (xObject instanceof Ref) {
+                  // Avoid parsing a corrupt XObject more than once.
+                  processed[xObject.toString()] = true;
+                }
+                // Error(s) in the XObject -- sending unsupported feature
+                // notification and allow parsing/rendering to continue.
+                this.handler.send("UnsupportedFeature", {
+                  featureId: UNSUPPORTED_FEATURES.unknown,
+                });
+                warn(`hasBlendModes - ignoring XObject: "${ex}".`);
+                continue;
+              }
+              throw ex;
+            }
           }
           if (!isStream(xObject)) {
             continue;

src/core/parser.js

@@ -842,6 +842,7 @@ class Lexer {
     // other commands or literals as a prefix. The knowCommands is optional.
     this.knownCommands = knownCommands;
 
+    this._hexStringNumWarn = 0;
     this.beginInlineImagePos = -1;
   }
 
@@ -1099,12 +1100,32 @@ class Lexer {
     return Name.get(strBuf.join(""));
   }
 
+  /**
+   * @private
+   */
+  _hexStringWarn(ch) {
+    const MAX_HEX_STRING_NUM_WARN = 5;
+
+    if (this._hexStringNumWarn++ === MAX_HEX_STRING_NUM_WARN) {
+      warn("getHexString - ignoring additional invalid characters.");
+      return;
+    }
+    if (this._hexStringNumWarn > MAX_HEX_STRING_NUM_WARN) {
+      // Limit the number of warning messages printed for a `this.getHexString`
+      // invocation, since corrupt PDF documents may otherwise spam the console
+      // enough to affect general performance negatively.
+      return;
+    }
+    warn(`getHexString - ignoring invalid character: ${ch}`);
+  }
+
   getHexString() {
     const strBuf = this.strBuf;
     strBuf.length = 0;
     let ch = this.currentChar;
     let isFirstHex = true;
     let firstDigit, secondDigit;
+    this._hexStringNumWarn = 0;
 
     while (true) {
       if (ch < 0) {
@@ -1120,14 +1141,14 @@ class Lexer {
         if (isFirstHex) {
           firstDigit = toHexDigit(ch);
           if (firstDigit === -1) {
-            warn(`Ignoring invalid character "${ch}" in hex string`);
+            this._hexStringWarn(ch);
             ch = this.nextChar();
             continue;
           }
         } else {
           secondDigit = toHexDigit(ch);
           if (secondDigit === -1) {
-            warn(`Ignoring invalid character "${ch}" in hex string`);
+            this._hexStringWarn(ch);
             ch = this.nextChar();
             continue;
           }

test/pdfs/issue11678.pdf.link

@@ -0,0 +1 @@
+https://github.com/mozilla/pdf.js/files/4304559/default.pdf

test/test_manifest.json

@@ -3064,6 +3064,15 @@
        "link": false,
        "type": "eq"
     },
+    {  "id": "issue11678",
+       "file": "pdfs/issue11678.pdf",
+       "md5": "e2efadeb91932f4c21e4fc682cce7de9",
+       "rounds": 1,
+       "link": true,
+       "firstPage": 2,
+       "lastPage": 2,
+       "type": "eq"
+    },
     {  "id": "issue4890",
        "file": "pdfs/issue4890.pdf",
        "md5": "1666feb4cd26318c2bdbea6a175dce87",