Fix ROS finding for CID Font detection

veraPDF · Mar 20, 2024 · 7f7e009 · 7f7e009
1 parent 484d14b
commit 7f7e009
Show file tree

Hide file tree

Showing 4 changed files with 43 additions and 27 deletions.
diff --git a/src/main/java/org/verapdf/pd/font/cff/CFFCIDFontProgram.java b/src/main/java/org/verapdf/pd/font/cff/CFFCIDFontProgram.java
@@ -57,13 +57,8 @@ public class CFFCIDFontProgram extends CFFFontBaseParser implements FontProgram
     CFFCIDFontProgram(SeekableInputStream stream, CFFIndex definedNames, CFFIndex globalSubrs,
                       long topDictBeginOffset, long topDictEndOffset, CMap externalCMap,
                       boolean isSubset) {
-        super(stream);
-        this.definedNames = definedNames;
-        this.globalSubrs = globalSubrs;
-        this.topDictBeginOffset = topDictBeginOffset;
-        this.topDictEndOffset = topDictEndOffset;
+        super(stream, definedNames, globalSubrs, topDictBeginOffset, topDictEndOffset, isSubset);
         this.externalCMap = externalCMap;
-        this.isSubset = isSubset;
     }
 
     /**

diff --git a/src/main/java/org/verapdf/pd/font/cff/CFFFontBaseParser.java b/src/main/java/org/verapdf/pd/font/cff/CFFFontBaseParser.java
@@ -36,7 +36,7 @@
  *
  * @author Sergey Shemyakov
  */
-abstract class CFFFontBaseParser extends CFFFileBaseParser {
+class CFFFontBaseParser extends CFFFileBaseParser {
 
     private static final Logger LOGGER = Logger.getLogger(CFFFontBaseParser.class.getCanonicalName());
 
@@ -72,13 +72,39 @@ abstract class CFFFontBaseParser extends CFFFileBaseParser {
     //Subrs
     protected long subrsOffset = -1;
 
+    private boolean isContainsROS = false;
+
     public CFFFontBaseParser(SeekableInputStream source) {
         super(source);
         stack = new ArrayList<>(48);
         this.charStringType = 2;
         this.charSetOffset = 0; // default
     }
 
+    public CFFFontBaseParser(SeekableInputStream stream, CFFIndex definedNames, CFFIndex globalSubrs,
+                        long topDictBeginOffset, long topDictEndOffset, boolean isSubset) {
+        this(stream);
+        this.definedNames = definedNames;
+        this.globalSubrs = globalSubrs;
+        this.topDictBeginOffset = topDictBeginOffset;
+        this.topDictEndOffset = topDictEndOffset;
+        this.isSubset = isSubset;
+    }
+
+    protected boolean isContainsROS() {
+        try {
+            this.source.seek(topDictBeginOffset);
+            while (this.source.getOffset() < topDictEndOffset) {
+                readTopDictUnit();
+            }     
+            if (isContainsROS) {
+                return true;
+            }
+        } catch (IOException ignored) {
+        }
+        return false;
+    }
+
     protected void readTopDictUnit() throws IOException {
         try {
             int next = this.source.peek();
@@ -125,6 +151,9 @@ protected void readTopDictUnit() throws IOException {
                                             this.stack.get(this.stack.size() - 1).getInteger();
                                     this.stack.clear();
                                     break;
+                                case 30:
+                                    this.isContainsROS = true;
+                                    break;
                                 default:
                                     readTopDictTwoByteOps(next);
                             }

diff --git a/src/main/java/org/verapdf/pd/font/cff/CFFFontProgram.java b/src/main/java/org/verapdf/pd/font/cff/CFFFontProgram.java
@@ -76,7 +76,7 @@ public void parseFont() throws IOException {
             }
             this.definedNames = this.readIndex();
             CFFIndex globalSubrs = this.readIndex();
-            if (isCIDFont(top.get(0))) {
+            if (isCIDFont(top, globalSubrs, topOffset)) {
                 font = new CFFCIDFontProgram(this.source, this.definedNames, globalSubrs,
                         topOffset + top.getOffset(0) - 1 + top.getOffsetShift(),
                         topOffset + top.getOffset(1) - 1 + top.getOffsetShift(),
@@ -91,9 +91,10 @@ public void parseFont() throws IOException {
         }
     }
 
-    private boolean isCIDFont(byte[] topDict) {
+    private boolean isCIDFont(CFFIndex top, CFFIndex globalSubrs, long topOffset) {
         try {
             byte rosOffset;
+            byte[] topDict = top.get(0);
             int supplementFirstByte = topDict[4] & 0xFF;    // checking if first operator is really ROS
             if (supplementFirstByte < 247 && supplementFirstByte > 31) {
                 rosOffset = 5;
@@ -104,27 +105,23 @@ private boolean isCIDFont(byte[] topDict) {
             } else if (supplementFirstByte == 29) {
                 rosOffset = 9;
             } else {
-                return isContainsROS(topDict);
+                return isContainsROS(top, globalSubrs, topOffset);
             }
             if (topDict[rosOffset] == 12 && topDict[rosOffset + 1] == 30) {
                 isCIDFont = true;
                 return true;
             }
-            return isContainsROS(topDict);
+            return isContainsROS(top, globalSubrs, topOffset);
         } catch (ArrayIndexOutOfBoundsException ex) {
-            return isContainsROS(topDict);
+            return isContainsROS(top, globalSubrs, topOffset);
         }
     }
 
-    private boolean isContainsROS(byte[] topDict) {
-        for (int rosOffset = 0; rosOffset < topDict.length - 2; rosOffset++) {
-            if (topDict[rosOffset] == 12 && topDict[rosOffset + 1] == 30) {
-                LOGGER.log(Level.WARNING, "The Top DICT does not begin with ROS operator");
-                isCIDFont = true;
-                return true;
-            }
-        }
-        return false;
+    private boolean isContainsROS(CFFIndex top, CFFIndex globalSubrs, long topOffset) {
+        return new CFFFontBaseParser(this.source, this.definedNames, globalSubrs,
+                topOffset + top.getOffset(0) - 1 + top.getOffsetShift(),
+                topOffset + top.getOffset(1) - 1 + top.getOffsetShift(),
+                this.isSubset).isContainsROS();
     }
 
     /**

diff --git a/src/main/java/org/verapdf/pd/font/cff/CFFType1FontProgram.java b/src/main/java/org/verapdf/pd/font/cff/CFFType1FontProgram.java
@@ -58,15 +58,10 @@ public class CFFType1FontProgram extends CFFFontBaseParser implements FontProgra
     CFFType1FontProgram(SeekableInputStream stream, CFFIndex definedNames, CFFIndex globalSubrs,
                         long topDictBeginOffset, long topDictEndOffset,
                         CMap externalCMap, boolean isSubset) {
-        super(stream);
+        super(stream, definedNames, globalSubrs, topDictBeginOffset, topDictEndOffset, isSubset);
         encodingOffset = 0;
         encoding = new int[256];
-        this.definedNames = definedNames;
-        this.globalSubrs = globalSubrs;
-        this.topDictBeginOffset = topDictBeginOffset;
-        this.topDictEndOffset = topDictEndOffset;
         this.externalCMap = externalCMap;
-        this.isSubset = isSubset;
         fontMatrix = new float[6];
         System.arraycopy(DEFAULT_FONT_MATRIX, 0, this.fontMatrix, 0,
                 DEFAULT_FONT_MATRIX.length);