Author: lehmi
Date: Mon Apr 8 05:54:20 2024
New Revision: 1916856
URL: http://svn.apache.org/viewvc?rev=1916856&view=rev
Log:
PDFBOX-5032: implement a separate parser for the xref information.
BruteForceParser no longer extends COSParser
Added:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/XrefParser.java
(with props)
Modified:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BaseParser.java
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BruteForceParser.java
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/COSParser.java
Modified:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BaseParser.java
URL:
http://svn.apache.org/viewvc/pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BaseParser.java?rev=1916856&r1=1916855&r2=1916856&view=diff
==============================================================================
---
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BaseParser.java
(original)
+++
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BaseParser.java
Mon Apr 8 05:54:20 2024
@@ -62,7 +62,7 @@ public abstract class BaseParser
private static final long GENERATION_NUMBER_THRESHOLD = 65535;
- static final int MAX_LENGTH_LONG = Long.toString(Long.MAX_VALUE).length();
+ private static final int MAX_LENGTH_LONG =
Long.toString(Long.MAX_VALUE).length();
private static final Charset ALTERNATIVE_CHARSET;
@@ -133,13 +133,25 @@ public abstract class BaseParser
private static final char[] NULL = { 'n', 'u', 'l', 'l' };
/**
+ * ASCII code for Null.
+ */
+ private static final byte ASCII_NULL = 0;
+ /**
+ * ASCII code for horizontal tab.
+ */
+ private static final byte ASCII_TAB = 9;
+ /**
* ASCII code for line feed.
*/
- protected static final byte ASCII_LF = 10;
+ private static final byte ASCII_LF = 10;
+ /**
+ * ASCII code for form feed.
+ */
+ private static final byte ASCII_FF = 12;
/**
* ASCII code for carriage return.
*/
- protected static final byte ASCII_CR = 13;
+ private static final byte ASCII_CR = 13;
private static final byte ASCII_ZERO = 48;
private static final byte ASCII_NINE = 57;
private static final byte ASCII_SPACE = 32;
@@ -251,8 +263,7 @@ public abstract class BaseParser
{
if (document == null)
{
- throw new IOException("object reference " + key + " at offset " +
source
- .getPosition()
+ throw new IOException("object reference " + key + " at offset " +
source.getPosition()
+ " in content stream");
}
return document.getObjectFromPool(key);
@@ -389,32 +400,35 @@ public abstract class BaseParser
return true;
}
+ /**
+ * Skip the upcoming CRLF or LF which are supposed to follow a stream.
+ *
+ * @throws IOException
+ */
protected void skipWhiteSpaces() throws IOException
{
//PDF Ref 3.2.7 A stream must be followed by either
//a CRLF or LF but nothing else.
-
int whitespace = source.read();
-
//see brother_scan_cover.pdf, it adds whitespaces
//after the stream but before the start of the
//data, so just read those first
- while (ASCII_SPACE == whitespace)
+ while (isSpace(whitespace))
{
whitespace = source.read();
}
- if (ASCII_CR == whitespace)
+ if (isCR(whitespace))
{
whitespace = source.read();
- if (ASCII_LF != whitespace)
+ if (!isLF(whitespace))
{
source.rewind(1);
//The spec says this is invalid but it happens in the real
//world so we must support it.
}
}
- else if (ASCII_LF != whitespace)
+ else if (!isLF(whitespace))
{
//we are in an error.
//but again we will do a lenient parsing and just assume that
everything
@@ -460,10 +474,10 @@ public abstract class BaseParser
// 4. COSDictionary ends in the next line: LF + '>'
// 5. Next line contains another COSObject: CR + '/'
// 6. COSDictionary ends in the next line: CR + '>'
- if (((nextThreeBytes[0] == ASCII_CR || nextThreeBytes[0] == ASCII_LF)
+ if (((isCR(nextThreeBytes[0]) || isLF(nextThreeBytes[0]))
&& (nextThreeBytes[1] == '/' || nextThreeBytes[1] == '>')) //
|| //
- (nextThreeBytes[0] == ASCII_CR && nextThreeBytes[1] == ASCII_LF
+ (isCR(nextThreeBytes[0]) && isLF(nextThreeBytes[1])
&& (nextThreeBytes[2] == '/' || nextThreeBytes[2] ==
'>')) //
)
{
@@ -781,14 +795,14 @@ public abstract class BaseParser
* @param ch The character
* @return true if the character terminates a PDF name, otherwise false.
*/
- protected boolean isEndOfName(int ch)
+ protected static boolean isEndOfName(int ch)
{
switch (ch)
{
case ASCII_SPACE:
case ASCII_CR:
case ASCII_LF:
- case 9:
+ case ASCII_TAB:
case '>':
case '<':
case '[':
@@ -796,7 +810,7 @@ public abstract class BaseParser
case ']':
case ')':
case '(':
- case 0:
+ case ASCII_NULL:
case '\f':
case '%':
case -1:
@@ -872,8 +886,8 @@ public abstract class BaseParser
}
/**
- * Tries to decode the buffer cotent to an UTF-8 String.
- * If that fails, tries the alternative Encoding.
+ * Tries to decode the buffer content to an UTF-8 String. If that fails,
tries the alternative Encoding.
+ *
* @param buffer the {@link ByteArrayOutputStream} containing the bytes to
decode
* @return the decoded String
*/
@@ -935,7 +949,7 @@ public abstract class BaseParser
case (char)-1:
return null;
default:
- if( Character.isDigit(c) || c == '-' || c == '+' || c == '.')
+ if (isDigit(c) || c == '-' || c == '+' || c == '.')
{
return parseCOSNumber();
}
@@ -1086,22 +1100,10 @@ public abstract class BaseParser
/**
* This will tell if the next character is a closing brace( close of PDF
array ).
*
- * @return true if the next byte is ']', false otherwise.
- *
- * @throws IOException If an IO error occurs.
- */
- protected boolean isClosing() throws IOException
- {
- return isClosing(source.peek());
- }
-
- /**
- * This will tell if the next character is a closing brace( close of PDF
array ).
- *
* @param c The character to check against end of line
* @return true if the next byte is ']', false otherwise.
*/
- protected boolean isClosing(int c)
+ protected static boolean isClosing(int c)
{
return c == ']';
}
@@ -1144,18 +1146,6 @@ public abstract class BaseParser
}
/**
- * This will tell if the next byte to be read is an end of line byte.
- *
- * @return true if the next byte is 0x0A or 0x0D.
- *
- * @throws IOException If there is an error reading from the stream.
- */
- protected boolean isEOL() throws IOException
- {
- return isEOL(source.peek());
- }
-
- /**
* This will tell if the end of the data is reached.
*
* @return true if the end of the data is reached.
@@ -1172,17 +1162,29 @@ public abstract class BaseParser
* @param c The character to check against end of line
* @return true if the next byte is 0x0A or 0x0D.
*/
- protected boolean isEOL(int c)
+ protected static boolean isEOL(int c)
{
return isLF(c) || isCR(c);
}
- private boolean isLF(int c)
+ /**
+ * This will tell if the next byte to be read is a line feed.
+ *
+ * @param c The character to check against line feed
+ * @return true if the next byte is 0x0A.
+ */
+ private static boolean isLF(int c)
{
return ASCII_LF == c;
}
- private boolean isCR(int c)
+ /**
+ * This will tell if the next byte to be read is a carriage return.
+ *
+ * @param c The character to check against carriage return
+ * @return true if the next byte is 0x0D.
+ */
+ private static boolean isCR(int c)
{
return ASCII_CR == c;
}
@@ -1209,9 +1211,9 @@ public abstract class BaseParser
{
switch (c)
{
- case 0:
- case 9:
- case 12:
+ case ASCII_NULL:
+ case ASCII_TAB:
+ case ASCII_FF:
case ASCII_LF:
case ASCII_CR:
case ASCII_SPACE:
@@ -1239,7 +1241,7 @@ public abstract class BaseParser
* @param c The character to check against space
* @return true if the next byte in the stream is a space character.
*/
- protected boolean isSpace(int c)
+ private static boolean isSpace(int c)
{
return ASCII_SPACE == c;
}
@@ -1327,7 +1329,8 @@ public abstract class BaseParser
int retval = readInt();
if(retval < 0 || retval > GENERATION_NUMBER_THRESHOLD)
{
- throw new IOException("Generation Number '" + retval + "' has more
than 5 digits");
+ throw new IOException(
+ "Generation Number '" + retval + "' has more than 5 digits
or is negative");
}
return retval;
}
@@ -1399,7 +1402,7 @@ public abstract class BaseParser
{
int lastByte;
StringBuilder buffer = new StringBuilder();
- while ((lastByte = source.read()) >= '0' && lastByte <= '9')
+ while (isDigit(lastByte = source.read()))
{
buffer.append( (char)lastByte );
if (buffer.length() > MAX_LENGTH_LONG)
Modified:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BruteForceParser.java
URL:
http://svn.apache.org/viewvc/pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BruteForceParser.java?rev=1916856&r1=1916855&r2=1916856&view=diff
==============================================================================
---
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BruteForceParser.java
(original)
+++
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/BruteForceParser.java
Mon Apr 8 05:54:20 2024
@@ -42,7 +42,7 @@ import org.apache.pdfbox.pdmodel.encrypt
/**
* Brute force parser to be used as last resort if a malformed pdf can't be
read.
*/
-public class BruteForceParser extends COSParser
+public class BruteForceParser
{
private static final char[] XREF_TABLE = { 'x', 'r', 'e', 'f' };
private static final char[] XREF_STREAM = { '/', 'X', 'R', 'e', 'f' };
@@ -77,18 +77,23 @@ public class BruteForceParser extends CO
private boolean bfSearchTriggered = false;
+ private final COSParser parser;
+ private final COSDocument document;
+ private final RandomAccessRead source;
+
/**
* Constructor. Triggers a brute force search for all objects of the
document.
*
- * @param source input representing the pdf.
- * @param document the corresponding COS document
+ * @param cosDocument the corresponding COS document
+ * @param cosParser the COSParser to be used for reading the pdf
*
* @throws IOException if the source data could not be read
*/
- public BruteForceParser(RandomAccessRead source, COSDocument document)
throws IOException
+ public BruteForceParser(COSDocument cosDocument, COSParser cosParser)
throws IOException
{
- super(source);
- this.document = document;
+ document = cosDocument;
+ parser = cosParser;
+ source = parser.source;
}
/**
@@ -140,25 +145,25 @@ public class BruteForceParser extends CO
source.seek(currentOffset);
int nextChar = source.read();
currentOffset++;
- if (isWhitespace(nextChar) && isString(OBJ_MARKER))
+ if (COSParser.isWhitespace(nextChar) &&
parser.isString(OBJ_MARKER))
{
long tempOffset = currentOffset - 2;
source.seek(tempOffset);
int genID = source.peek();
// is the next char a digit?
- if (isDigit(genID))
+ if (COSParser.isDigit(genID))
{
genID -= 48;
tempOffset--;
source.seek(tempOffset);
- if (isWhitespace())
+ if (parser.isWhitespace())
{
- while (tempOffset > MINIMUM_SEARCH_OFFSET &&
isWhitespace())
+ while (tempOffset > MINIMUM_SEARCH_OFFSET &&
parser.isWhitespace())
{
source.seek(--tempOffset);
}
boolean objectIDFound = false;
- while (tempOffset > MINIMUM_SEARCH_OFFSET && isDigit())
+ while (tempOffset > MINIMUM_SEARCH_OFFSET &&
parser.isDigit())
{
source.seek(--tempOffset);
objectIDFound = true;
@@ -166,7 +171,7 @@ public class BruteForceParser extends CO
if (objectIDFound)
{
source.read();
- long objectId = readObjectNumber();
+ long objectId = parser.readObjectNumber();
if (lastObjOffset > 0)
{
// add the former object ID only if there was
a subsequent object ID
@@ -185,21 +190,21 @@ public class BruteForceParser extends CO
// check for "endo" as abbreviation for "endobj", as the pdf may
be cut off
// in the middle of the keyword, see PDFBOX-3936.
// We could possibly implement a more intelligent algorithm if
necessary
- else if (nextChar == 'e' && isString(endobjString))
+ else if (nextChar == 'e' && parser.isString(endobjString))
{
currentOffset += endobjString.length;
source.seek(currentOffset);
- if (source.isEOF())
+ if (parser.isEOF())
{
endOfObjFound = true;
}
- else if (isString(endobjRemainingString))
+ else if (parser.isString(endobjRemainingString))
{
currentOffset += endobjRemainingString.length;
endOfObjFound = true;
}
}
- } while (currentOffset < lastEOFMarker && !source.isEOF());
+ } while (currentOffset < lastEOFMarker && !parser.isEOF());
if ((lastEOFMarker < Long.MAX_VALUE || endOfObjFound) && lastObjOffset
> 0)
{
// if the pdf wasn't cut off in the middle or if the last object
ends with a "endobj" marker
@@ -291,15 +296,12 @@ public class BruteForceParser extends CO
/**
* Brute force search for all objects streams of a pdf.
*
- * @param trailerResolver the trailer resolver of the document
- * @param securityHandler security handler to be used to decrypt encrypted
documents
+ * @param xrefTable the cross reference table of the document
+ *
* @throws IOException if something went wrong
*/
- protected void bfSearchForObjStreams(XrefTrailerResolver trailerResolver,
- SecurityHandler<? extends ProtectionPolicy> securityHandler)
throws IOException
+ protected void bfSearchForObjStreams(Map<COSObjectKey, Long> xrefTable)
throws IOException
{
- // update security handler
- this.securityHandler = securityHandler;
// save origin offset
long originOffset = source.getPosition();
@@ -318,24 +320,24 @@ public class BruteForceParser extends CO
.map(Map.Entry::getKey) //
.collect(Collectors.toList());
// add all found compressed objects to the brute force search result
+ SecurityHandler<? extends ProtectionPolicy> securityHandler =
parser.getSecurityHandler();
for (Long offset : objStreamOffsets)
{
source.seek(offset);
- long stmObjNumber = readObjectNumber();
- int stmGenNumber = readGenerationNumber();
- readExpectedString(OBJ_MARKER, true);
+ long stmObjNumber = parser.readObjectNumber();
+ int stmGenNumber = parser.readGenerationNumber();
+ parser.readExpectedString(OBJ_MARKER, true);
COSStream stream = null;
try
{
- COSDictionary dict = parseCOSDictionary(false);
- stream = parseCOSStream(dict);
+ COSDictionary dict = parser.parseCOSDictionary(false);
+ stream = parser.parseCOSStream(dict);
if (securityHandler != null)
{
securityHandler.decryptStream(stream, stmObjNumber,
stmGenNumber);
}
PDFObjectStreamParser objStreamParser = new
PDFObjectStreamParser(stream, document);
Map<Long, Integer> objectNumbers =
objStreamParser.readObjectNumbers();
- Map<COSObjectKey, Long> xrefOffset =
trailerResolver.getXrefTable();
for (Long objNumber : objectNumbers.keySet())
{
COSObjectKey objKey = new COSObjectKey(objNumber, 0);
@@ -349,7 +351,7 @@ public class BruteForceParser extends CO
if (existingOffset == null || offset > existingOffset)
{
bfCOSObjectOffsets.put(objKey, -stmObjNumber);
- xrefOffset.put(objKey, -stmObjNumber);
+ xrefTable.put(objKey, -stmObjNumber);
}
}
}
@@ -389,8 +391,8 @@ public class BruteForceParser extends CO
{
boolean rootFound = false;
boolean infoFound = false;
- skipSpaces();
- COSDictionary trailerDict = parseCOSDictionary(true);
+ parser.skipSpaces();
+ COSDictionary trailerDict = parser.parseCOSDictionary(true);
COSObject rootObj = trailerDict.getCOSObject(COSName.ROOT);
if (rootObj != null)
{
@@ -532,11 +534,11 @@ public class BruteForceParser extends CO
// check if the following data is some valid pdf content
// which most likely indicates that the pdf is linearized,
// updated or just cut off somewhere in the middle
- skipSpaces();
- if (!isString(XREF_TABLE))
+ parser.skipSpaces();
+ if (!parser.isString(XREF_TABLE))
{
- readObjectNumber();
- readGenerationNumber();
+ parser.readObjectNumber();
+ parser.readGenerationNumber();
}
}
catch (IOException exception)
@@ -583,21 +585,20 @@ public class BruteForceParser extends CO
source.seek(currentOffset);
for (int j = 0; j < 10; j++)
{
- if (isString(string))
+ if (parser.isString(string))
{
long tempOffset = currentOffset - 1;
source.seek(tempOffset);
- int genID = source.peek();
// is the next char a digit?
- if (isDigit(genID))
+ if (parser.isDigit())
{
tempOffset--;
source.seek(tempOffset);
- if (isSpace())
+ if (parser.isSpace())
{
int length = 0;
source.seek(--tempOffset);
- while (tempOffset > MINIMUM_SEARCH_OFFSET
&& isDigit())
+ while (tempOffset > MINIMUM_SEARCH_OFFSET
&& parser.isDigit())
{
source.seek(--tempOffset);
length++;
@@ -606,8 +607,8 @@ public class BruteForceParser extends CO
{
source.read();
newOffset = source.getPosition();
- long objNumber = readObjectNumber();
- int genNumber = readGenerationNumber();
+ long objNumber =
parser.readObjectNumber();
+ int genNumber =
parser.readGenerationNumber();
COSObjectKey streamObjectKey = new
COSObjectKey(objNumber,
genNumber);
bfSearchObjStreamsOffsets.put(newOffset, streamObjectKey);
@@ -648,7 +649,7 @@ public class BruteForceParser extends CO
{
source.seek(newOffset - 1);
// ensure that we don't read "startxref" instead of "xref"
- if (isWhitespace())
+ if (parser.isWhitespace())
{
bfSearchXRefTablesOffsets.add(newOffset);
}
@@ -685,21 +686,20 @@ public class BruteForceParser extends CO
source.seek(currentOffset);
for (int j = 0; j < 10; j++)
{
- if (isString(string))
+ if (parser.isString(string))
{
long tempOffset = currentOffset - 1;
source.seek(tempOffset);
- int genID = source.peek();
// is the next char a digit?
- if (isDigit(genID))
+ if (parser.isDigit())
{
tempOffset--;
source.seek(tempOffset);
- if (isSpace())
+ if (parser.isSpace())
{
int length = 0;
source.seek(--tempOffset);
- while (tempOffset > MINIMUM_SEARCH_OFFSET
&& isDigit())
+ while (tempOffset > MINIMUM_SEARCH_OFFSET
&& parser.isDigit())
{
source.seek(--tempOffset);
length++;
@@ -759,7 +759,7 @@ public class BruteForceParser extends CO
/**
* Tell if the dictionary is a PDF or FDF catalog.
*
- * @param dictionary
+ * @param dictionary the dictionary to be tested
* @return true if the given dictionary is a root dictionary
*/
private boolean isCatalog(COSDictionary dictionary)
@@ -810,20 +810,17 @@ public class BruteForceParser extends CO
/**
* Rebuild the trailer dictionary if startxref can't be found.
+ *
+ * @param xrefTable the cross reference table of the pdf
*
- * @param trailerResolver the trailer resolver of the document
- * @param securityHandler security handler to be used to decrypt encrypted
documents
* @return the rebuild trailer dictionary
*
* @throws IOException if something went wrong
*/
- protected COSDictionary rebuildTrailer(XrefTrailerResolver trailerResolver,
- SecurityHandler<? extends ProtectionPolicy> securityHandler)
throws IOException
+ protected COSDictionary rebuildTrailer(Map<COSObjectKey, Long> xrefTable)
throws IOException
{
- // update security handler
- this.securityHandler = securityHandler;
- // reset trailer resolver
- trailerResolver.reset();
+ // use a new trailer resolver
+ XrefTrailerResolver trailerResolver = new XrefTrailerResolver();
// use the found objects to rebuild the trailer resolver
trailerResolver.nextXrefObj(0, XRefType.TABLE);
getBFCOSObjectOffsets().forEach(trailerResolver::setXRef);
@@ -840,22 +837,23 @@ public class BruteForceParser extends CO
COSDictionary trailer = trailerResolver.getTrailer();
document.setTrailer(trailer);
+ xrefTable.putAll(trailerResolver.getXrefTable());
boolean searchForObjStreamsDone = false;
if (!bfSearchForTrailer(trailer) && !searchForTrailerItems(trailer))
{
// root entry wasn't found, maybe it is part of an object stream
// brute force search for all object streams.
- bfSearchForObjStreams(trailerResolver, securityHandler);
+ bfSearchForObjStreams(xrefTable);
searchForObjStreamsDone = true;
// search again for the root entry
searchForTrailerItems(trailer);
}
// prepare decryption if necessary
- prepareDecryption();
+ parser.prepareDecryption();
if (!searchForObjStreamsDone)
{
// brute force search for all object streams.
- bfSearchForObjStreams(trailerResolver, securityHandler);
+ bfSearchForObjStreams(xrefTable);
}
return trailer;
}
Modified:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/COSParser.java
URL:
http://svn.apache.org/viewvc/pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/COSParser.java?rev=1916856&r1=1916855&r2=1916856&view=diff
==============================================================================
---
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/COSParser.java
(original)
+++
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/COSParser.java
Mon Apr 8 05:54:20 2024
@@ -25,8 +25,6 @@ import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
-import java.util.Optional;
-import java.util.Map.Entry;
import java.util.Set;
import org.apache.logging.log4j.Logger;
@@ -46,7 +44,6 @@ import org.apache.pdfbox.io.IOUtils;
import org.apache.pdfbox.io.RandomAccessRead;
import org.apache.pdfbox.io.RandomAccessReadView;
import org.apache.pdfbox.io.RandomAccessStreamCache.StreamCacheCreateFunction;
-import org.apache.pdfbox.pdfparser.XrefTrailerResolver.XRefType;
import org.apache.pdfbox.pdmodel.encryption.AccessPermission;
import org.apache.pdfbox.pdmodel.encryption.DecryptionMaterial;
import org.apache.pdfbox.pdmodel.encryption.InvalidPasswordException;
@@ -55,7 +52,6 @@ import org.apache.pdfbox.pdmodel.encrypt
import org.apache.pdfbox.pdmodel.encryption.PublicKeyDecryptionMaterial;
import org.apache.pdfbox.pdmodel.encryption.SecurityHandler;
import org.apache.pdfbox.pdmodel.encryption.StandardDecryptionMaterial;
-import org.apache.pdfbox.util.StringUtil;
/**
* COS-Parser which first reads startxref and xref tables in order to know
valid objects and parse only these objects.
@@ -71,17 +67,14 @@ public class COSParser extends BaseParse
private static final String PDF_DEFAULT_VERSION = "1.4";
private static final String FDF_DEFAULT_VERSION = "1.0";
- private static final char[] XREF_TABLE = { 'x', 'r', 'e', 'f' };
private static final char[] STARTXREF = {
's','t','a','r','t','x','r','e','f' };
private static final byte[] ENDSTREAM = { E, N, D, S, T, R, E, A, M };
private static final byte[] ENDOBJ = { E, N, D, O, B, J };
- private static final long MINIMUM_SEARCH_OFFSET = 6;
+ protected static final long MINIMUM_SEARCH_OFFSET = 6;
- private static final int X = 'x';
-
private static final int STRMBUFLEN = 2048;
private final byte[] strmBuf = new byte[ STRMBUFLEN ];
@@ -105,16 +98,16 @@ public class COSParser extends BaseParse
/**
* EOF-marker.
*/
- protected static final char[] EOF_MARKER = { '%', '%', 'E', 'O', 'F' };
+ private static final char[] EOF_MARKER = { '%', '%', 'E', 'O', 'F' };
/**
* obj-marker.
*/
- protected static final char[] OBJ_MARKER = { 'o', 'b', 'j' };
+ private static final char[] OBJ_MARKER = { 'o', 'b', 'j' };
/**
* file length.
*/
- protected long fileLen;
+ private final long fileLen;
/**
* is parser using auto healing capacity ?
@@ -127,6 +120,7 @@ public class COSParser extends BaseParse
private BruteForceParser bruteForceParser = null;
private PDEncryption encryption = null;
+ private final Map<COSObjectKey, Long> xrefTable = new HashMap<>();
/**
* Intermediate cache. Contains all objects of already read compressed
object streams. Objects are removed after
@@ -137,7 +131,7 @@ public class COSParser extends BaseParse
/**
* The security handler.
*/
- protected SecurityHandler<? extends ProtectionPolicy> securityHandler =
null;
+ private SecurityHandler<? extends ProtectionPolicy> securityHandler = null;
/**
* how many trailing bytes to read for EOF marker.
@@ -146,12 +140,6 @@ public class COSParser extends BaseParse
private static final Logger LOG = LogManager.getLogger(COSParser.class);
- /**
- * Collects all Xref/trailer objects and resolves them into single
- * object using startxref reference.
- */
- protected XrefTrailerResolver xrefTrailerResolver = new
XrefTrailerResolver();
-
/**
* Default constructor.
*
@@ -261,7 +249,9 @@ public class COSParser extends BaseParse
long startXRefOffset = getStartxrefOffset();
if (startXRefOffset > -1)
{
- trailer = parseXref(startXRefOffset);
+ XrefParser xrefParser = new XrefParser(document, this);
+ trailer = xrefParser.parseXref(startXRefOffset);
+ xrefTable.putAll(xrefParser.getXrefTable());
}
else
{
@@ -286,15 +276,10 @@ public class COSParser extends BaseParse
}
if (rebuildTrailer)
{
- trailer =
getBruteForceParser().rebuildTrailer(xrefTrailerResolver, null);
+ // reset cross reference table
+ xrefTable.clear();
+ trailer = getBruteForceParser().rebuildTrailer(xrefTable);
trailerWasRebuild = true;
- // transfer encryption information from BruteForceParser
- encryption = getBruteForceParser().getEncryption();
- if (encryption != null)
- {
- securityHandler = encryption.getSecurityHandler();
- accessPermission =
securityHandler.getCurrentAccessPermission();
- }
}
else
{
@@ -303,194 +288,13 @@ public class COSParser extends BaseParse
// don't use the getter as it creates an instance of
BruteForceParser
if (bruteForceParser != null &&
bruteForceParser.bfSearchTriggered())
{
-
getBruteForceParser().bfSearchForObjStreams(xrefTrailerResolver,
securityHandler);
+ getBruteForceParser().bfSearchForObjStreams(xrefTable);
}
}
- if (resetTrailerResolver())
- {
- xrefTrailerResolver.reset();
- xrefTrailerResolver = null;
- }
return trailer;
}
/**
- * Indicates whether the xref trailer resolver should be reset or not.
Should be overwritten if the xref trailer
- * resolver is needed after the initial parsing.
- *
- * @return true if the xref trailer resolver should be reset
- */
- protected boolean resetTrailerResolver()
- {
- return true;
- }
-
- /**
- * Parses cross reference tables.
- *
- * @param startXRefOffset start offset of the first table
- * @return the trailer dictionary
- * @throws IOException if something went wrong
- */
- private COSDictionary parseXref(long startXRefOffset) throws IOException
- {
- source.seek(startXRefOffset);
- long startXrefOffset = Math.max(0, parseStartXref());
- // check the startxref offset
- long fixedOffset = checkXRefOffset(startXrefOffset);
- if (fixedOffset > -1)
- {
- startXrefOffset = fixedOffset;
- }
- document.setStartXref(startXrefOffset);
- long prev = startXrefOffset;
- // ---- parse whole chain of xref tables/object streams using PREV
reference
- Set<Long> prevSet = new HashSet<>();
- COSDictionary trailer = null;
- while (prev > 0)
- {
- // save expected position for loop detection
- prevSet.add(prev);
- // seek to xref table
- source.seek(prev);
- // skip white spaces
- skipSpaces();
- // save current position as well due to skipped spaces
- prevSet.add(source.getPosition());
- // -- parse xref
- if (source.peek() == X)
- {
- // xref table and trailer
- // use existing parser to parse xref table
- if (!parseXrefTable(prev) || !parseTrailer())
- {
- throw new IOException("Expected trailer object at offset "
- + source.getPosition());
- }
- trailer = xrefTrailerResolver.getCurrentTrailer();
- // check for a XRef stream, it may contain some object ids of
compressed objects
- if(trailer.containsKey(COSName.XREF_STM))
- {
- int streamOffset = trailer.getInt(COSName.XREF_STM);
- // check the xref stream reference
- fixedOffset = checkXRefOffset(streamOffset);
- if (fixedOffset > -1 && fixedOffset != streamOffset)
- {
- LOG.warn("/XRefStm offset {} is incorrect, corrected
to {}", streamOffset,
- fixedOffset);
- streamOffset = (int)fixedOffset;
- trailer.setInt(COSName.XREF_STM, streamOffset);
- }
- if (streamOffset > 0)
- {
- source.seek(streamOffset);
- skipSpaces();
- try
- {
- parseXrefObjStream(prev, false);
- document.setHasHybridXRef();
- }
- catch (IOException ex)
- {
- if (isLenient)
- {
- LOG.error("Failed to parse /XRefStm at offset
{}", streamOffset,
- ex);
- }
- else
- {
- throw ex;
- }
- }
- }
- else
- {
- if(isLenient)
- {
- LOG.error("Skipped XRef stream due to a corrupt
offset:{}",
- streamOffset);
- }
- else
- {
- throw new IOException("Skipped XRef stream due to
a corrupt offset:"+streamOffset);
- }
- }
- }
- prev = trailer.getLong(COSName.PREV);
- }
- else
- {
- // parse xref stream
- prev = parseXrefObjStream(prev, true);
- trailer = xrefTrailerResolver.getCurrentTrailer();
- }
- if (prev > 0)
- {
- // check the xref table reference
- fixedOffset = checkXRefOffset(prev);
- if (fixedOffset > -1 && fixedOffset != prev)
- {
- prev = fixedOffset;
- trailer.setLong(COSName.PREV, prev);
- }
- }
- if (prevSet.contains(prev))
- {
- throw new IOException("/Prev loop at offset " + prev);
- }
- }
- // ---- build valid xrefs out of the xref chain
- xrefTrailerResolver.setStartxref(startXrefOffset);
- trailer = xrefTrailerResolver.getTrailer();
- document.setTrailer(trailer);
- document.setIsXRefStream(XRefType.STREAM ==
xrefTrailerResolver.getXrefType());
- // check the offsets of all referenced objects
- if (isLenient)
- {
- checkXrefOffsets();
- }
- // copy xref table
- document.addXRefTable(xrefTrailerResolver.getXrefTable());
-
- // remember the highest XRef object number to avoid it being reused in
incremental saving
- Optional<Long> maxValue = document.getXrefTable().keySet().stream() //
- .map(COSObjectKey::getNumber) //
- .reduce(Long::max);
- document.setHighestXRefObjectNumber(maxValue.isPresent() ?
maxValue.get() : 0);
-
- return trailer;
- }
-
- /**
- * Parses an xref object stream starting with indirect object id.
- *
- * @return value of PREV item in dictionary or <code>-1</code> if no such
item exists
- */
- private long parseXrefObjStream(long objByteOffset, boolean isStandalone)
throws IOException
- {
- // ---- parse indirect object head
- readObjectNumber();
- readGenerationNumber();
- readExpectedString(OBJ_MARKER, true);
-
- COSDictionary dict = parseCOSDictionary(false);
- try (COSStream xrefStream = parseCOSStream(dict))
- {
- // the cross reference stream of a hybrid xref table will be added
to the existing one
- // and we must not override the offset and the trailer
- if ( isStandalone )
- {
- xrefTrailerResolver.nextXrefObj( objByteOffset,
XRefType.STREAM );
- xrefTrailerResolver.setTrailer(xrefStream);
- }
- PDFXrefStreamParser parser = new PDFXrefStreamParser(xrefStream,
document);
- parser.parse(xrefTrailerResolver);
- }
-
- return dict.getLong(COSName.PREV);
- }
-
- /**
* Looks for and parses startxref. We first look for last '%%EOF' marker
(within last
* {@link #DEFAULT_TRAIL_BYTECOUNT} bytes (or range set via {@link
#setEOFLookupRange(int)}) and go back to find
* <code>startxref</code>.
@@ -564,7 +368,7 @@ public class COSParser extends BaseParse
*
* @return start offset of pattern within buffer or <code>-1</code> if
pattern could not be found
*/
- protected int lastIndexOf(final char[] pattern, final byte[] buf, final
int endOff)
+ private int lastIndexOf(final char[] pattern, final byte[] buf, final int
endOff)
{
final int lastPatternChOff = pattern.length - 1;
@@ -658,7 +462,7 @@ public class COSParser extends BaseParse
*
* @throws IOException If an IO error occurs.
*/
- protected synchronized COSBase parseObjectDynamically(COSObjectKey objKey,
+ private synchronized COSBase parseObjectDynamically(COSObjectKey objKey,
boolean requireExistingNotCompressedObj) throws IOException
{
COSObject pdfObject = document.getObjectFromPool(objKey);
@@ -727,7 +531,7 @@ public class COSParser extends BaseParse
// an indirect object starts with the object number/generation number
final long readObjNr = readObjectNumber();
final int readObjGen = readGenerationNumber();
- readExpectedString(OBJ_MARKER, true);
+ readObjectMarker();
// consistency check
if (readObjNr != objKey.getNumber() || readObjGen !=
objKey.getGeneration())
@@ -876,7 +680,7 @@ public class COSParser extends BaseParse
}
if (COSNull.NULL == length)
{
- LOG.warn("Length object ({} {}) not found",
lengthObj.getKey());
+ LOG.warn("Length object ({}) not found", lengthObj.getKey());
return null;
}
if (length instanceof COSNumber)
@@ -961,7 +765,7 @@ public class COSParser extends BaseParse
{
throw new IOException(
"Error reading stream, expected='endstream' actual='"
- + endStream + "' at offset " + source.getPosition());
+ + endStream + "' at offset " +
source.getPosition());
}
return document.createCOSStream(dic, streamStartPosition,
streamLength);
}
@@ -1102,287 +906,11 @@ public class COSParser extends BaseParse
return streamLengthIsValid;
}
- /**
- * Check if the cross reference table/stream can be found at the current
offset.
- *
- * @param startXRefOffset
- * @return the revised offset
- * @throws IOException
- */
- private long checkXRefOffset(long startXRefOffset) throws IOException
- {
- // repair mode isn't available in non-lenient mode
- if (!isLenient)
- {
- return startXRefOffset;
- }
- source.seek(startXRefOffset);
- skipSpaces();
- if (isString(XREF_TABLE))
- {
- return startXRefOffset;
- }
- if (startXRefOffset > 0)
- {
- if (checkXRefStreamOffset(startXRefOffset))
- {
- return startXRefOffset;
- }
- else
- {
- return calculateXRefFixedOffset(startXRefOffset);
- }
- }
- // can't find a valid offset
- return -1;
- }
-
- /**
- * Check if the cross reference stream can be found at the current offset.
- *
- * @param startXRefOffset the expected start offset of the XRef stream
- * @return the revised offset
- * @throws IOException if something went wrong
- */
- private boolean checkXRefStreamOffset(long startXRefOffset) throws
IOException
- {
- // repair mode isn't available in non-lenient mode
- if (!isLenient || startXRefOffset == 0)
- {
- return true;
- }
- // seek to offset-1
- source.seek(startXRefOffset-1);
- int nextValue = source.read();
- // the first character has to be a whitespace, and then a digit
- if (isWhitespace(nextValue))
- {
- skipSpaces();
- if (isDigit())
- {
- try
- {
- // it's a XRef stream
- readObjectNumber();
- readGenerationNumber();
- readExpectedString(OBJ_MARKER, true);
- // check the dictionary to avoid false positives
- COSDictionary dict = parseCOSDictionary(false);
- source.seek(startXRefOffset);
- if ("XRef".equals(dict.getNameAsString(COSName.TYPE)))
- {
- return true;
- }
- }
- catch (IOException exception)
- {
- // there wasn't an object of a xref stream
- LOG.debug("No Xref stream at given location {}",
startXRefOffset, exception);
- source.seek(startXRefOffset);
- }
- }
- }
- return false;
- }
-
- /**
- * Try to find a fixed offset for the given xref table/stream.
- *
- * @param objectOffset the given offset where to look at
- * @return the fixed offset
- *
- * @throws IOException if something went wrong
- */
- private long calculateXRefFixedOffset(long objectOffset) throws IOException
- {
- if (objectOffset < 0)
- {
- LOG.error("Invalid object offset {} when searching for a xref
table/stream",
- objectOffset);
- return 0;
- }
- // search for the offset of the given xref table/stream among those
found by a brute force search.
- long newOffset = getBruteForceParser().bfSearchForXRef(objectOffset);
- if (newOffset > -1)
- {
- LOG.debug("Fixed reference for xref table/stream {} -> {}",
objectOffset, newOffset);
- return newOffset;
- }
- LOG.error("Can't find the object xref table/stream at offset {}",
objectOffset);
- return 0;
- }
-
- private boolean validateXrefOffsets(Map<COSObjectKey, Long> xrefOffset)
throws IOException
- {
- if (xrefOffset == null)
- {
- return true;
- }
- Map<COSObjectKey, COSObjectKey> correctedKeys = new HashMap<>();
- HashSet<COSObjectKey> validKeys = new HashSet<>();
- for (Entry<COSObjectKey, Long> objectEntry : xrefOffset.entrySet())
- {
- COSObjectKey objectKey = objectEntry.getKey();
- Long objectOffset = objectEntry.getValue();
- // a negative offset number represents an object number itself
- // see type 2 entry in xref stream
- if (objectOffset != null && objectOffset >= 0)
- {
- COSObjectKey foundObjectKey = findObjectKey(objectKey,
objectOffset, xrefOffset);
- if (foundObjectKey == null)
- {
- LOG.debug(
- "Stop checking xref offsets as at least one ({})
couldn't be dereferenced",
- objectKey);
- return false;
- }
- else if (foundObjectKey != objectKey)
- {
- // Generation was fixed - need to update map later, after
iteration
- correctedKeys.put(objectKey, foundObjectKey);
- }
- else
- {
- validKeys.add(objectKey);
- }
- }
- }
- Map<COSObjectKey, Long> correctedPointers = new HashMap<>();
- for (Entry<COSObjectKey, COSObjectKey> correctedKeyEntry :
correctedKeys.entrySet())
- {
- if (!validKeys.contains(correctedKeyEntry.getValue()))
- {
- // Only replace entries, if the original entry does not point
to a valid object
- correctedPointers.put(correctedKeyEntry.getValue(),
- xrefOffset.get(correctedKeyEntry.getKey()));
- }
- }
- // remove old invalid, as some might not be replaced
- correctedKeys.forEach((key, value) -> xrefOffset.remove(key));
- xrefOffset.putAll(correctedPointers);
- return true;
- }
-
- /**
- * Check the XRef table by dereferencing all objects and fixing the offset
if necessary.
- *
- * @throws IOException if something went wrong.
- */
- private void checkXrefOffsets() throws IOException
- {
- Map<COSObjectKey, Long> xrefOffset =
xrefTrailerResolver.getXrefTable();
- if (!validateXrefOffsets(xrefOffset))
- {
- Map<COSObjectKey, Long> bfCOSObjectKeyOffsets =
getBruteForceParser()
- .getBFCOSObjectOffsets();
- if (!bfCOSObjectKeyOffsets.isEmpty())
- {
- LOG.debug("Replaced read xref table with the results of a
brute force search");
- xrefOffset.clear();
- xrefOffset.putAll(bfCOSObjectKeyOffsets);
- }
- }
- }
-
- /**
- * Check if the given object can be found at the given offset. Returns the
provided object key if everything is ok.
- * If the generation number differs it will be fixed and a new object key
is returned.
- *
- * @param objectKey the key of object we are looking for
- * @param offset the offset where to look
- * @param xrefOffset a map with with all known xref entries
- * @return returns the found/fixed object key
- *
- * @throws IOException if something went wrong
- */
- private COSObjectKey findObjectKey(COSObjectKey objectKey, long offset,
- Map<COSObjectKey, Long> xrefOffset) throws IOException
- {
- // there can't be any object at the very beginning of a pdf
- if (offset < MINIMUM_SEARCH_OFFSET)
- {
- return null;
- }
- try
- {
- source.seek(offset);
- skipWhiteSpaces();
- if (source.getPosition() == offset)
- {
- // ensure that at least one whitespace is skipped in front of
the object number
- source.seek(offset - 1);
- if (source.getPosition() < offset)
- {
- if (!isDigit())
- {
- // anything else but a digit may be some garbage of
the previous object -> just ignore it
- source.read();
- }
- else
- {
- long current = source.getPosition();
- source.seek(--current);
- while (isDigit())
- source.seek(--current);
- long newObjNr = readObjectNumber();
- int newGenNr = readGenerationNumber();
- COSObjectKey newObjKey = new COSObjectKey(newObjNr,
newGenNr);
- Long existingOffset = xrefOffset.get(newObjKey);
- // the found object number belongs to another
uncompressed object at the same or nearby offset
- // something has to be wrong
- if (existingOffset != null && existingOffset > 0
- && Math.abs(offset - existingOffset) < 10)
- {
- LOG.debug("Found the object {} instead of {} at
offset {} - ignoring",
- newObjKey, objectKey, offset);
- return null;
- }
- // something seems to be wrong but it's hard to
determine what exactly -> simply continue
- source.seek(offset);
- }
- }
- }
- // try to read the given object/generation number
- long foundObjectNumber = readObjectNumber();
- if (objectKey.getNumber() != foundObjectNumber)
- {
- LOG.warn("found wrong object number. expected [{}] found [{}]",
- objectKey.getNumber(), foundObjectNumber);
- if (!isLenient)
- {
- return null;
- }
- else
- {
- objectKey = new COSObjectKey(foundObjectNumber,
objectKey.getGeneration());
- }
- }
-
- int genNumber = readGenerationNumber();
- // finally try to read the object marker
- readExpectedString(OBJ_MARKER, true);
- if (genNumber == objectKey.getGeneration())
- {
- return objectKey;
- }
- else if (isLenient && genNumber > objectKey.getGeneration())
- {
- return new COSObjectKey(objectKey.getNumber(), genNumber);
- }
- }
- catch (IOException exception)
- {
- // Swallow the exception, obviously there isn't any valid object
number
- LOG.debug("No valid object at given location {} - ignoring",
offset, exception);
- }
- return null;
- }
-
- private BruteForceParser getBruteForceParser() throws IOException
+ protected BruteForceParser getBruteForceParser() throws IOException
{
if (bruteForceParser == null)
{
- bruteForceParser = new BruteForceParser(source, document);
+ bruteForceParser = new BruteForceParser(document, this);
}
return bruteForceParser;
}
@@ -1457,25 +985,6 @@ public class COSParser extends BaseParse
}
/**
- * This will parse the startxref section from the stream. The startxref
value is ignored.
- *
- * @return the startxref value or -1 on parsing error
- * @throws IOException If an IO error occurs.
- */
- private long parseStartXref() throws IOException
- {
- long startXref = -1;
- if (isString(STARTXREF))
- {
- readString();
- skipSpaces();
- // This integer is the byte offset of the first object referenced
by the xref or xref stream
- startXref = readLong();
- }
- return startXref;
- }
-
- /**
* Checks if the given string can be found at the current offset.
*
* @param string the bytes of the string to look for
@@ -1521,67 +1030,9 @@ public class COSParser extends BaseParse
return bytesMatching;
}
- /**
- * This will parse the trailer from the stream and add it to the state.
- *
- * @return false on parsing error
- * @throws IOException If an IO error occurs.
- */
- private boolean parseTrailer() throws IOException
+ protected void readObjectMarker() throws IOException
{
- // parse the last trailer.
- long trailerOffset = source.getPosition();
- // PDFBOX-1739 skip extra xref entries in RegisSTAR documents
- if (isLenient)
- {
- int nextCharacter = source.peek();
- while (nextCharacter != 't' && isDigit(nextCharacter))
- {
- if (source.getPosition() == trailerOffset)
- {
- // warn only the first time
- LOG.warn("Expected trailer object at offset {}, keep
trying", trailerOffset);
- }
- readLine();
- nextCharacter = source.peek();
- }
- }
- if(source.peek() != 't')
- {
- return false;
- }
- //read "trailer"
- long currentOffset = source.getPosition();
- String nextLine = readLine();
- if( !nextLine.trim().equals( "trailer" ) )
- {
- // in some cases the EOL is missing and the trailer immediately
- // continues with "<<" or with a blank character
- // even if this does not comply with PDF reference we want to
support as many PDFs as possible
- // Acrobat reader can also deal with this.
- if (nextLine.startsWith("trailer"))
- {
- // we can't just unread a portion of the read data as we don't
know if the EOL consist of 1 or 2 bytes
- int len = "trailer".length();
- // jump back right after "trailer"
- source.seek(currentOffset + len);
- }
- else
- {
- return false;
- }
- }
-
- // in some cases the EOL is missing and the trailer continues with "
<<"
- // even if this does not comply with PDF reference we want to support
as many PDFs as possible
- // Acrobat reader can also deal with this.
- skipSpaces();
-
- COSDictionary parsedTrailer = parseCOSDictionary(true);
- xrefTrailerResolver.setTrailer( parsedTrailer );
-
- skipSpaces();
- return true;
+ readExpectedString(OBJ_MARKER, true);
}
/**
@@ -1690,128 +1141,6 @@ public class COSParser extends BaseParse
}
/**
- * This will parse the xref table from the stream and add it to the state
- * The XrefTable contents are ignored.
- * @param startByteOffset the offset to start at
- * @return false on parsing error
- * @throws IOException If an IO error occurs.
- */
- protected boolean parseXrefTable(long startByteOffset) throws IOException
- {
- if(source.peek() != 'x')
- {
- return false;
- }
- String xref = readString();
- if( !xref.trim().equals( "xref" ) )
- {
- return false;
- }
-
- // check for trailer after xref
- String str = readString();
- byte[] b = str.getBytes(StandardCharsets.ISO_8859_1);
- source.rewind(b.length);
-
- // signal start of new XRef
- xrefTrailerResolver.nextXrefObj( startByteOffset, XRefType.TABLE );
-
- if (str.startsWith("trailer"))
- {
- LOG.warn("skipping empty xref table");
- return false;
- }
-
- // Xref tables can have multiple sections. Each starts with a starting
object id and a count.
- while(true)
- {
- String currentLine = readLine();
- String[] splitString = StringUtil.splitOnSpace(currentLine);
- if (splitString.length != 2)
- {
- LOG.warn("Unexpected XRefTable Entry: {}", currentLine);
- return false;
- }
- // first obj id
- long currObjID;
- try
- {
- currObjID = Long.parseLong(splitString[0]);
- }
- catch (NumberFormatException exception)
- {
- LOG.warn("XRefTable: invalid ID for the first object: {}",
currentLine);
- return false;
- }
-
- // the number of objects in the xref table
- int count = 0;
- try
- {
- count = Integer.parseInt(splitString[1]);
- }
- catch (NumberFormatException exception)
- {
- LOG.warn("XRefTable: invalid number of objects: {}",
currentLine);
- return false;
- }
-
- skipSpaces();
- for(int i = 0; i < count; i++)
- {
- if (source.isEOF() || isEndOfName(source.peek()))
- {
- break;
- }
- if(source.peek() == 't')
- {
- break;
- }
- //Ignore table contents
- currentLine = readLine();
- splitString = StringUtil.splitOnSpace(currentLine);
- if (splitString.length < 3)
- {
- LOG.warn("invalid xref line: {}", currentLine);
- break;
- }
- /* This supports the corrupt table as reported in
- * PDFBOX-474 (XXXX XXX XX n) */
- if(splitString[splitString.length-1].equals("n"))
- {
- try
- {
- long currOffset = Long.parseLong(splitString[0]);
- // skip 0 offsets
- if (currOffset > 0)
- {
- int currGenID = Integer.parseInt(splitString[1]);
- COSObjectKey objKey = new COSObjectKey(currObjID,
currGenID);
- xrefTrailerResolver.setXRef(objKey, currOffset);
- }
- }
- catch (IllegalArgumentException e)
- {
- throw new IOException(e);
- }
- }
- else if(!splitString[2].equals("f"))
- {
- throw new IOException("Corrupt XRefTable Entry - ObjID:" +
currObjID);
- }
- currObjID++;
- skipSpaces();
- }
- skipSpaces();
- if (!isDigit())
- {
- break;
- }
- }
- return true;
- }
-
- /**
* This will get the encryption dictionary. The document must be parsed
before this is called.
*
* @return The encryption dictionary of the document that was parsed.
@@ -1901,4 +1230,15 @@ public class COSParser extends BaseParse
}
}
+ /**
+ * This will get the security handler. The document must be parsed before
this is called.
+ *
+ * @return The security handler of the document that was parsed.
+ *
+ * @throws IOException If there is an error getting the document.
+ */
+ protected SecurityHandler<? extends ProtectionPolicy> getSecurityHandler()
+ {
+ return securityHandler;
+ }
}
Added:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/XrefParser.java
URL:
http://svn.apache.org/viewvc/pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/XrefParser.java?rev=1916856&view=auto
==============================================================================
---
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/XrefParser.java
(added)
+++
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/XrefParser.java
Mon Apr 8 05:54:20 2024
@@ -0,0 +1,694 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.pdfbox.pdfparser;
+
+import java.io.IOException;
+import java.nio.charset.StandardCharsets;
+import java.util.HashMap;
+import java.util.HashSet;
+import java.util.Map;
+import java.util.Optional;
+import java.util.Map.Entry;
+import java.util.Set;
+
+import org.apache.logging.log4j.LogManager;
+import org.apache.logging.log4j.Logger;
+import org.apache.pdfbox.cos.COSDictionary;
+import org.apache.pdfbox.cos.COSDocument;
+import org.apache.pdfbox.cos.COSName;
+import org.apache.pdfbox.cos.COSObjectKey;
+import org.apache.pdfbox.cos.COSStream;
+import org.apache.pdfbox.io.RandomAccessRead;
+import org.apache.pdfbox.pdfparser.XrefTrailerResolver.XRefType;
+
+/**
+ * Parser to be used to read the cross reference table of a pdf. It is either
a simple table or a stream.
+ */
+public class XrefParser
+{
+ private static final Logger LOG = LogManager.getLogger(XrefParser.class);
+
+ private static final int X = 'x';
+ private static final char[] XREF_TABLE = { 'x', 'r', 'e', 'f' };
+ private static final char[] STARTXREF = { 's', 't', 'a', 'r', 't', 'x',
'r', 'e', 'f' };
+
+ /**
+ * Collects all Xref/trailer objects and resolves them into single
+ * object using startxref reference.
+ */
+ private XrefTrailerResolver xrefTrailerResolver = new
XrefTrailerResolver();
+
+ private final COSParser parser;
+ private final COSDocument document;
+ private final RandomAccessRead source;
+
+ /**
+ * Default constructor.
+ *
+ * @param cosDocument the corresponding COS document of the pdf.
+ * @param cosParser the parser to be used to read the pdf.
+ *
+ */
+ public XrefParser(COSDocument cosDocument, COSParser cosParser)
+ {
+ document = cosDocument;
+ parser = cosParser;
+ source = parser.source;
+ }
+
+ /**
+ * Returns the resulting cross reference table.
+ *
+ * @return
+ */
+ public Map<COSObjectKey, Long> getXrefTable()
+ {
+ return xrefTrailerResolver.getXrefTable();
+ }
+
+ /**
+ * Parses cross reference tables.
+ *
+ * @param startXRefOffset start offset of the first table
+ * @return the trailer dictionary
+ * @throws IOException if something went wrong
+ */
+ public COSDictionary parseXref(long startXRefOffset) throws IOException
+ {
+ source.seek(startXRefOffset);
+ long startXrefOffset = Math.max(0, parseStartXref());
+ // check the startxref offset
+ long fixedOffset = checkXRefOffset(startXrefOffset);
+ if (fixedOffset > -1)
+ {
+ startXrefOffset = fixedOffset;
+ }
+ document.setStartXref(startXrefOffset);
+ long prev = startXrefOffset;
+ // ---- parse whole chain of xref tables/object streams using PREV
reference
+ Set<Long> prevSet = new HashSet<>();
+ COSDictionary trailer = null;
+ while (prev > 0)
+ {
+ // save expected position for loop detection
+ prevSet.add(prev);
+ // seek to xref table
+ source.seek(prev);
+ // skip white spaces
+ parser.skipSpaces();
+ // save current position as well due to skipped spaces
+ prevSet.add(source.getPosition());
+ // -- parse xref
+ if (source.peek() == X)
+ {
+ // xref table and trailer
+ // use existing parser to parse xref table
+ if (!parseXrefTable(prev) || !parseTrailer())
+ {
+ throw new IOException("Expected trailer object at offset "
+ + source.getPosition());
+ }
+ trailer = xrefTrailerResolver.getCurrentTrailer();
+ // check for a XRef stream, it may contain some object ids of
compressed objects
+ if(trailer.containsKey(COSName.XREF_STM))
+ {
+ int streamOffset = trailer.getInt(COSName.XREF_STM);
+ // check the xref stream reference
+ fixedOffset = checkXRefOffset(streamOffset);
+ if (fixedOffset > -1 && fixedOffset != streamOffset)
+ {
+ LOG.warn("/XRefStm offset {} is incorrect, corrected
to {}", streamOffset,
+ fixedOffset);
+ streamOffset = (int)fixedOffset;
+ trailer.setInt(COSName.XREF_STM, streamOffset);
+ }
+ if (streamOffset > 0)
+ {
+ source.seek(streamOffset);
+ parser.skipSpaces();
+ try
+ {
+ parseXrefObjStream(prev, false);
+ document.setHasHybridXRef();
+ }
+ catch (IOException ex)
+ {
+ LOG.error("Failed to parse /XRefStm at offset {}",
streamOffset, ex);
+ }
+ }
+ else
+ {
+ LOG.error("Skipped XRef stream due to a corrupt
offset: {}", streamOffset);
+ }
+ }
+ prev = trailer.getLong(COSName.PREV);
+ }
+ else
+ {
+ // parse xref stream
+ prev = parseXrefObjStream(prev, true);
+ trailer = xrefTrailerResolver.getCurrentTrailer();
+ }
+ if (prev > 0)
+ {
+ // check the xref table reference
+ fixedOffset = checkXRefOffset(prev);
+ if (fixedOffset > -1 && fixedOffset != prev)
+ {
+ prev = fixedOffset;
+ trailer.setLong(COSName.PREV, prev);
+ }
+ }
+ if (prevSet.contains(prev))
+ {
+ throw new IOException("/Prev loop at offset " + prev);
+ }
+ }
+ // ---- build valid xrefs out of the xref chain
+ xrefTrailerResolver.setStartxref(startXrefOffset);
+ trailer = xrefTrailerResolver.getTrailer();
+ document.setTrailer(trailer);
+ document.setIsXRefStream(XRefType.STREAM ==
xrefTrailerResolver.getXrefType());
+ // check the offsets of all referenced objects
+ checkXrefOffsets();
+ // copy xref table
+ document.addXRefTable(xrefTrailerResolver.getXrefTable());
+
+ // remember the highest XRef object number to avoid it being reused in
incremental saving
+ Optional<Long> maxValue = document.getXrefTable().keySet().stream() //
+ .map(COSObjectKey::getNumber) //
+ .reduce(Long::max);
+ document.setHighestXRefObjectNumber(maxValue.isPresent() ?
maxValue.get() : 0);
+
+ return trailer;
+ }
+
+ /**
+ * This will parse the trailer from the stream and add it to the state.
+ *
+ * @return false on parsing error
+ * @throws IOException If an IO error occurs.
+ */
+ private boolean parseTrailer() throws IOException
+ {
+ // parse the last trailer.
+ long trailerOffset = source.getPosition();
+ // PDFBOX-1739 skip extra xref entries in RegisSTAR documents
+ int nextCharacter = source.peek();
+ while (nextCharacter != 't' && COSParser.isDigit(nextCharacter))
+ {
+ if (source.getPosition() == trailerOffset)
+ {
+ // warn only the first time
+ LOG.warn("Expected trailer object at offset {}, keep trying",
trailerOffset);
+ }
+ parser.readLine();
+ nextCharacter = source.peek();
+ }
+ if (source.peek() != 't')
+ {
+ return false;
+ }
+ // read "trailer"
+ long currentOffset = source.getPosition();
+ String nextLine = parser.readLine();
+ if (!nextLine.trim().equals("trailer"))
+ {
+ // in some cases the EOL is missing and the trailer immediately
+ // continues with "<<" or with a blank character
+ // even if this does not comply with PDF reference we want to
support as many PDFs as possible
+ // Acrobat reader can also deal with this.
+ if (nextLine.startsWith("trailer"))
+ {
+ // we can't just unread a portion of the read data as we don't
know if the EOL consist of 1 or 2 bytes
+ int len = "trailer".length();
+ // jump back right after "trailer"
+ source.seek(currentOffset + len);
+ }
+ else
+ {
+ return false;
+ }
+ }
+
+ // in some cases the EOL is missing and the trailer continues with "
<<"
+ // even if this does not comply with PDF reference we want to support
as many PDFs as possible
+ // Acrobat reader can also deal with this.
+ parser.skipSpaces();
+
+ COSDictionary parsedTrailer = parser.parseCOSDictionary(true);
+ xrefTrailerResolver.setTrailer(parsedTrailer);
+
+ parser.skipSpaces();
+ return true;
+ }
+
+ /**
+ * Parses an xref object stream starting with indirect object id.
+ *
+ * @return value of PREV item in dictionary or <code>-1</code> if no such
item exists
+ */
+ private long parseXrefObjStream(long objByteOffset, boolean isStandalone)
throws IOException
+ {
+ // ---- parse indirect object head
+ parser.readObjectNumber();
+ parser.readGenerationNumber();
+ parser.readObjectMarker();
+
+ COSDictionary dict = parser.parseCOSDictionary(false);
+ try (COSStream xrefStream = parser.parseCOSStream(dict))
+ {
+ // the cross reference stream of a hybrid xref table will be added
to the existing one
+ // and we must not override the offset and the trailer
+ if ( isStandalone )
+ {
+ xrefTrailerResolver.nextXrefObj( objByteOffset,
XRefType.STREAM );
+ xrefTrailerResolver.setTrailer(xrefStream);
+ }
+ PDFXrefStreamParser parser = new PDFXrefStreamParser(xrefStream,
document);
+ parser.parse(xrefTrailerResolver);
+ }
+
+ return dict.getLong(COSName.PREV);
+ }
+
+ /**
+ * Check if the cross reference table/stream can be found at the current
offset.
+ *
+ * @param startXRefOffset
+ * @return the revised offset
+ * @throws IOException
+ */
+ private long checkXRefOffset(long startXRefOffset) throws IOException
+ {
+ source.seek(startXRefOffset);
+ parser.skipSpaces();
+ if (parser.isString(XREF_TABLE))
+ {
+ return startXRefOffset;
+ }
+ if (startXRefOffset > 0)
+ {
+ if (checkXRefStreamOffset(startXRefOffset))
+ {
+ return startXRefOffset;
+ }
+ else
+ {
+ return calculateXRefFixedOffset(startXRefOffset);
+ }
+ }
+ // can't find a valid offset
+ return -1;
+ }
+
+ /**
+ * Try to find a fixed offset for the given xref table/stream.
+ *
+ * @param objectOffset the given offset where to look at
+ * @return the fixed offset
+ *
+ * @throws IOException if something went wrong
+ */
+ private long calculateXRefFixedOffset(long objectOffset) throws IOException
+ {
+ if (objectOffset < 0)
+ {
+ LOG.error("Invalid object offset {} when searching for a xref
table/stream",
+ objectOffset);
+ return 0;
+ }
+ // search for the offset of the given xref table/stream among those
found by a brute force search.
+ long newOffset =
parser.getBruteForceParser().bfSearchForXRef(objectOffset);
+ if (newOffset > -1)
+ {
+ LOG.debug("Fixed reference for xref table/stream {} -> {}",
objectOffset, newOffset);
+ return newOffset;
+ }
+ LOG.error("Can't find the object xref table/stream at offset {}",
objectOffset);
+ return 0;
+ }
+
+ /**
+ * Check if the cross reference stream can be found at the current offset.
+ *
+ * @param startXRefOffset the expected start offset of the XRef stream
+ * @return the revised offset
+ * @throws IOException if something went wrong
+ */
+ private boolean checkXRefStreamOffset(long startXRefOffset) throws
IOException
+ {
+ if (startXRefOffset == 0)
+ {
+ return true;
+ }
+ // seek to offset-1
+ source.seek(startXRefOffset - 1);
+ int nextValue = source.read();
+ // the first character has to be a whitespace, and then a digit
+ if (COSParser.isWhitespace(nextValue))
+ {
+ parser.skipSpaces();
+ if (parser.isDigit())
+ {
+ try
+ {
+ // it's a XRef stream
+ parser.readObjectNumber();
+ parser.readGenerationNumber();
+ parser.readObjectMarker();
+ // check the dictionary to avoid false positives
+ COSDictionary dict = parser.parseCOSDictionary(false);
+ source.seek(startXRefOffset);
+ if ("XRef".equals(dict.getNameAsString(COSName.TYPE)))
+ {
+ return true;
+ }
+ }
+ catch (IOException exception)
+ {
+ // there wasn't an object of a xref stream
+ LOG.debug("No Xref stream at given location {}",
startXRefOffset, exception);
+ source.seek(startXRefOffset);
+ }
+ }
+ }
+ return false;
+ }
+
+ private boolean validateXrefOffsets(Map<COSObjectKey, Long> xrefOffset)
throws IOException
+ {
+ if (xrefOffset == null)
+ {
+ return true;
+ }
+ Map<COSObjectKey, COSObjectKey> correctedKeys = new HashMap<>();
+ HashSet<COSObjectKey> validKeys = new HashSet<>();
+ for (Entry<COSObjectKey, Long> objectEntry : xrefOffset.entrySet())
+ {
+ COSObjectKey objectKey = objectEntry.getKey();
+ Long objectOffset = objectEntry.getValue();
+ // a negative offset number represents an object number itself
+ // see type 2 entry in xref stream
+ if (objectOffset != null && objectOffset >= 0)
+ {
+ COSObjectKey foundObjectKey = findObjectKey(objectKey,
objectOffset, xrefOffset);
+ if (foundObjectKey == null)
+ {
+ LOG.debug(
+ "Stop checking xref offsets as at least one ({})
couldn't be dereferenced",
+ objectKey);
+ return false;
+ }
+ else if (foundObjectKey != objectKey)
+ {
+ // Generation was fixed - need to update map later, after
iteration
+ correctedKeys.put(objectKey, foundObjectKey);
+ }
+ else
+ {
+ validKeys.add(objectKey);
+ }
+ }
+ }
+ Map<COSObjectKey, Long> correctedPointers = new HashMap<>();
+ for (Entry<COSObjectKey, COSObjectKey> correctedKeyEntry :
correctedKeys.entrySet())
+ {
+ if (!validKeys.contains(correctedKeyEntry.getValue()))
+ {
+ // Only replace entries, if the original entry does not point
to a valid object
+ correctedPointers.put(correctedKeyEntry.getValue(),
+ xrefOffset.get(correctedKeyEntry.getKey()));
+ }
+ }
+ // remove old invalid, as some might not be replaced
+ correctedKeys.forEach((key, value) -> xrefOffset.remove(key));
+ xrefOffset.putAll(correctedPointers);
+ return true;
+ }
+
+ /**
+ * Check the XRef table by dereferencing all objects and fixing the offset
if necessary.
+ *
+ * @throws IOException if something went wrong.
+ */
+ private void checkXrefOffsets() throws IOException
+ {
+ Map<COSObjectKey, Long> xrefOffset =
xrefTrailerResolver.getXrefTable();
+ if (!validateXrefOffsets(xrefOffset))
+ {
+ Map<COSObjectKey, Long> bfCOSObjectKeyOffsets =
parser.getBruteForceParser()
+ .getBFCOSObjectOffsets();
+ if (!bfCOSObjectKeyOffsets.isEmpty())
+ {
+ LOG.debug("Replaced read xref table with the results of a
brute force search");
+ xrefOffset.clear();
+ xrefOffset.putAll(bfCOSObjectKeyOffsets);
+ }
+ }
+ }
+
+ /**
+ * Check if the given object can be found at the given offset. Returns the
provided object key if everything is ok.
+ * If the generation number differs it will be fixed and a new object key
is returned.
+ *
+ * @param objectKey the key of object we are looking for
+ * @param offset the offset where to look
+ * @param xrefOffset a map with with all known xref entries
+ * @return returns the found/fixed object key
+ *
+ * @throws IOException if something went wrong
+ */
+ private COSObjectKey findObjectKey(COSObjectKey objectKey, long offset,
+ Map<COSObjectKey, Long> xrefOffset) throws IOException
+ {
+ // there can't be any object at the very beginning of a pdf
+ if (offset < COSParser.MINIMUM_SEARCH_OFFSET)
+ {
+ return null;
+ }
+ try
+ {
+ source.seek(offset);
+ parser.skipWhiteSpaces();
+ if (source.getPosition() == offset)
+ {
+ // ensure that at least one whitespace is skipped in front of
the object number
+ source.seek(offset - 1);
+ if (source.getPosition() < offset)
+ {
+ if (!parser.isDigit())
+ {
+ // anything else but a digit may be some garbage of
the previous object -> just ignore it
+ source.read();
+ }
+ else
+ {
+ long current = source.getPosition();
+ source.seek(--current);
+ while (parser.isDigit())
+ source.seek(--current);
+ long newObjNr = parser.readObjectNumber();
+ int newGenNr = parser.readGenerationNumber();
+ COSObjectKey newObjKey = new COSObjectKey(newObjNr,
newGenNr);
+ Long existingOffset = xrefOffset.get(newObjKey);
+ // the found object number belongs to another
uncompressed object at the same or nearby offset
+ // something has to be wrong
+ if (existingOffset != null && existingOffset > 0
+ && Math.abs(offset - existingOffset) < 10)
+ {
+ LOG.debug("Found the object {} instead of {} at
offset {} - ignoring",
+ newObjKey, objectKey, offset);
+ return null;
+ }
+ // something seems to be wrong but it's hard to
determine what exactly -> simply continue
+ source.seek(offset);
+ }
+ }
+ }
+ // try to read the given object/generation number
+ long foundObjectNumber = parser.readObjectNumber();
+ if (objectKey.getNumber() != foundObjectNumber)
+ {
+ LOG.warn("found wrong object number. expected [{}] found [{}]",
+ objectKey.getNumber(), foundObjectNumber);
+ objectKey = new COSObjectKey(foundObjectNumber,
objectKey.getGeneration());
+ }
+
+ int genNumber = parser.readGenerationNumber();
+ // finally try to read the object marker
+ parser.readObjectMarker();
+ if (genNumber == objectKey.getGeneration())
+ {
+ return objectKey;
+ }
+ else if (genNumber > objectKey.getGeneration())
+ {
+ return new COSObjectKey(objectKey.getNumber(), genNumber);
+ }
+ }
+ catch (IOException exception)
+ {
+ // Swallow the exception, obviously there isn't any valid object
number
+ LOG.debug("No valid object at given location {} - ignoring",
offset, exception);
+ }
+ return null;
+ }
+
+ /**
+ * This will parse the startxref section from the stream. The startxref
value is ignored.
+ *
+ * @return the startxref value or -1 on parsing error
+ * @throws IOException If an IO error occurs.
+ */
+ private long parseStartXref() throws IOException
+ {
+ long startXref = -1;
+ if (parser.isString(STARTXREF))
+ {
+ parser.readString();
+ parser.skipSpaces();
+ // This integer is the byte offset of the first object referenced
by the xref or xref stream
+ startXref = parser.readLong();
+ }
+ return startXref;
+ }
+
+ /**
+ * This will parse the xref table from the stream and add it to the state
+ * The XrefTable contents are ignored.
+ * @param startByteOffset the offset to start at
+ * @return false on parsing error
+ * @throws IOException If an IO error occurs.
+ */
+ private boolean parseXrefTable(long startByteOffset) throws IOException
+ {
+ if (source.peek() != 'x')
+ {
+ return false;
+ }
+ String xref = parser.readString();
+ if( !xref.trim().equals( "xref" ) )
+ {
+ return false;
+ }
+
+ // check for trailer after xref
+ String str = parser.readString();
+ byte[] b = str.getBytes(StandardCharsets.ISO_8859_1);
+ source.seek(source.getPosition() - b.length);
+
+ // signal start of new XRef
+ xrefTrailerResolver.nextXrefObj( startByteOffset, XRefType.TABLE );
+
+ if (str.startsWith("trailer"))
+ {
+ LOG.warn("skipping empty xref table");
+ return false;
+ }
+
+ // Xref tables can have multiple sections. Each starts with a starting
object id and a count.
+ while(true)
+ {
+ String currentLine = parser.readLine();
+ String[] splitString = currentLine.split("\\s");
+ if (splitString.length != 2)
+ {
+ LOG.warn("Unexpected XRefTable Entry: {}", currentLine);
+ return false;
+ }
+ // first obj id
+ long currObjID;
+ try
+ {
+ currObjID = Long.parseLong(splitString[0]);
+ }
+ catch (NumberFormatException exception)
+ {
+ LOG.warn("XRefTable: invalid ID for the first object: {}",
currentLine);
+ return false;
+ }
+
+ // the number of objects in the xref table
+ int count = 0;
+ try
+ {
+ count = Integer.parseInt(splitString[1]);
+ }
+ catch (NumberFormatException exception)
+ {
+ LOG.warn("XRefTable: invalid number of objects: {}",
currentLine);
+ return false;
+ }
+
+ parser.skipSpaces();
+ for(int i = 0; i < count; i++)
+ {
+ if (parser.isEOF() )
+ {
+ break;
+ }
+ int nextChar = source.peek();
+ if (nextChar == 't' || COSParser.isEndOfName(nextChar))
+ {
+ break;
+ }
+ //Ignore table contents
+ currentLine = parser.readLine();
+ splitString = currentLine.split("\\s");
+ if (splitString.length < 3)
+ {
+ LOG.warn("invalid xref line: {}", currentLine);
+ break;
+ }
+ /* This supports the corrupt table as reported in
+ * PDFBOX-474 (XXXX XXX XX n) */
+ if(splitString[splitString.length-1].equals("n"))
+ {
+ try
+ {
+ long currOffset = Long.parseLong(splitString[0]);
+ // skip 0 offsets
+ if (currOffset > 0)
+ {
+ int currGenID = Integer.parseInt(splitString[1]);
+ COSObjectKey objKey = new COSObjectKey(currObjID,
currGenID);
+ xrefTrailerResolver.setXRef(objKey, currOffset);
+ }
+ }
+ catch (IllegalArgumentException e)
+ {
+ throw new IOException(e);
+ }
+ }
+ else if(!splitString[2].equals("f"))
+ {
+ throw new IOException("Corrupt XRefTable Entry - ObjID:" +
currObjID);
+ }
+ currObjID++;
+ parser.skipSpaces();
+ }
+ parser.skipSpaces();
+ if (!parser.isDigit())
+ {
+ break;
+ }
+ }
+ return true;
+ }
+
+}
Propchange:
pdfbox/trunk/pdfbox/src/main/java/org/apache/pdfbox/pdfparser/XrefParser.java
------------------------------------------------------------------------------
svn:eol-style = native
