PDFBOX-5997: avoid creation of temporary objects as proposed by Axel Howind

git-svn-id: https://svn.apache.org/repos/asf/pdfbox/trunk@1926545 13f79535-47bb-0310-9956-ffa450edef68

Author: lehmi

pdfbox/src/main/java/org/apache/pdfbox/cos/COSString.java

@@ -135,36 +135,60 @@ public COSString(String text, boolean forceHex)
      */
     public static COSString parseHex(String hex) throws IOException
     {
-        StringBuilder hexBuffer = new StringBuilder(hex.trim());
-
-        // if odd number then the last hex digit is assumed to be 0
-        if (hexBuffer.length() % 2 != 0)
+        // skip leading and trailing whitespace
+        int end = hex.length();
+        while (end > 0 && Character.isWhitespace(hex.charAt(end - 1)))
+        {
+            end--;
+        }
+        int start = 0;
+        while (start < end && Character.isWhitespace(hex.charAt(start)))
         {
-            hexBuffer.append('0');
+            start++;
         }
 
-        int length = hexBuffer.length();
+        int length = end - start;
         ByteArrayOutputStream bytes = new ByteArrayOutputStream((length + 1) / 2);
+
+        boolean isLengthUneven = length % 2 != 0;
+        if (isLengthUneven)
+        {
+            length--;
+        }
         for (int i = 0; i < length; i += 2)
         {
-            try
+            int value = 16 * Hex.getHexValue(hex.charAt(i)) + Hex.getHexValue(hex.charAt(i + 1));
+            if (value >= 0)
+            {
+                bytes.write(value);
+            }
+            else if (FORCE_PARSING)
             {
-                bytes.write(Integer.parseInt(hexBuffer.substring(i, i + 2), 16));
+                LOG.warn("Encountered a malformed hex string");
+                bytes.write('?'); // todo: what does Acrobat do? Any example PDFs?
             }
-            catch (NumberFormatException e)
+            else
             {
-                if (FORCE_PARSING)
-                {
-                    LOG.warn("Encountered a malformed hex string");
-                    bytes.write('?'); // todo: what does Acrobat do? Any example PDFs?
-                }
-                else
-                {
-                    throw new IOException("Invalid hex string: " + hex, e);
-                }
+                throw new IOException("Invalid hex string: " + hex);
+            }
+        }
+        if (isLengthUneven)
+        {
+            int value = 16 * Hex.getHexValue(hex.charAt(length));
+            if (value >= 0)
+            {
+                bytes.write(value);
+            }
+            else if (FORCE_PARSING)
+            {
+                LOG.warn("Encountered a malformed hex string");
+                bytes.write('?'); // todo: what does Acrobat do? Any example PDFs?
+            }
+            else
+            {
+                throw new IOException("Invalid hex string: " + hex);
             }
         }
-
         return new COSString(bytes.toByteArray());
     }
 

pdfbox/src/main/java/org/apache/pdfbox/util/Hex.java

@@ -228,19 +228,46 @@ public static byte[] decodeHex(String s)
             }
             else
             {
-                String hexByte = s.substring(i, i + 2);
-                try
+                int value = 16 * getHexValue(s.charAt(i)) + getHexValue(s.charAt(i + 1));
+                if (value >= 0)
                 {
-                    baos.write(Integer.parseInt(hexByte, 16)); // Byte.parseByte won't work with "9C"
+                    baos.write(value);
                 }
-                catch (NumberFormatException ex)
+                else
                 {
-                    LOG.error(() -> "Can't parse " + hexByte + ", aborting decode", ex);
-                    break;
+                    String hexByte = s.substring(i, i + 2);
+                    LOG.error("Can't parse " + hexByte + ", aborting decode");
                 }
                 i += 2;
             }
         }
         return baos.toByteArray();
     }
+
+    /**
+     * Converts a given character to its corresponding hexadecimal value. Valid characters are '0'-'9', 'A'-'F', or
+     * 'a'-'f'. Returns -256 for invalid characters.
+     * <p>
+     * The value of -256 is chosen so that to hex digits can be combined before checking for an invalid hex string
+     *
+     * @param c the character to be converted to a hexadecimal value
+     * @return the hexadecimal value of the character, or -256 if the character is invalid
+     */
+    public static int getHexValue(char c)
+    {
+        if (c >= '0' && c <= '9')
+        {
+            return c - '0';
+        }
+        else if (c >= 'A' && c <= 'F')
+        {
+            return c - 'A' + 10;
+        }
+        else if (c >= 'a' && c <= 'f')
+        {
+            return c - 'a' + 10;
+        }
+        return -256;
+    }
+
 }