de.intarsys.pdf.parser.PDFParser Maven / Gradle / Ivy
Show all versions of jpod Show documentation
/*
* Copyright (c) 2007, intarsys consulting GmbH
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* - Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* - Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* - Neither the name of intarsys nor the names of its contributors may be used
* to endorse or promote products derived from this software without specific
* prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
package de.intarsys.pdf.parser;
import de.intarsys.pdf.cos.COSArray;
import de.intarsys.pdf.cos.COSDictionary;
import de.intarsys.pdf.cos.COSDocumentElement;
import de.intarsys.pdf.cos.COSFalse;
import de.intarsys.pdf.cos.COSFixed;
import de.intarsys.pdf.cos.COSIndirectObject;
import de.intarsys.pdf.cos.COSInteger;
import de.intarsys.pdf.cos.COSName;
import de.intarsys.pdf.cos.COSNull;
import de.intarsys.pdf.cos.COSNumber;
import de.intarsys.pdf.cos.COSObject;
import de.intarsys.pdf.cos.COSObjectKey;
import de.intarsys.pdf.cos.COSStream;
import de.intarsys.pdf.cos.COSString;
import de.intarsys.pdf.cos.COSTrue;
import de.intarsys.pdf.crypt.COSSecurityException;
import de.intarsys.pdf.crypt.ISystemSecurityHandler;
import de.intarsys.pdf.st.STDocType;
import de.intarsys.tools.hex.HexTools;
import de.intarsys.tools.randomaccess.IRandomAccess;
import de.intarsys.tools.randomaccess.RandomAccessByteArray;
import de.intarsys.tools.stream.FastByteArrayOutputStream;
import de.intarsys.tools.string.StringTools;
import java.io.IOException;
import java.util.Arrays;
import java.util.List;
/**
* An abstract superclass for our two flavours of PDF Parsers.
*/
public abstract class PDFParser {
public static char CHAR_CR = '\r';
public static char CHAR_LF = '\n';
public static char CHAR_HT = '\t';
public static char CHAR_BS = '\b';
public static char CHAR_FF = '\f';
public static final byte[] TOKEN_PDFHEADER = "%PDF".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_FDFHEADER = "%FDF".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_EOF = "%%EOF".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_obj = "obj".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_endobj = "endobj".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_false = "false".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_true = "true".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_null = "null".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_startxref = "startxref".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_trailer = "trailer".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_xref = "xref".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_stream = "stream".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_s_tream = "tream".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_endstream = "endstream".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_ndstream = "ndstream".getBytes(); //$NON-NLS-1$
public static final byte[] TOKEN_R = "R".getBytes(); //$NON-NLS-1$
public static final String C_WARN_UNEVENHEX = "616a"; //$NON-NLS-1$
public static final String C_WARN_ILLEGALHEX = "616b"; //$NON-NLS-1$
public static final String C_WARN_STRING_TOO_LONG = "ImplLimitString"; //$NON-NLS-1$
public static final String C_WARN_NAME_TOO_LONG = "ImplLimitName"; //$NON-NLS-1$
public static final String C_WARN_ARRAYSIZE = "ImplLimitArray"; //$NON-NLS-1$
public static final String C_WARN_SINGLESPACE = "614a"; //$NON-NLS-1$
public static final String C_WARN_SINGLEEOL = "614b"; //$NON-NLS-1$
public static final String C_WARN_STREAMEOL = "617a"; //$NON-NLS-1$
public static final String C_WARN_ENDSTREAMEOL = "617b"; //$NON-NLS-1$
public static final String C_WARN_ENDSTREAMCORRUPT = "617c"; //$NON-NLS-1$
public static final String C_WARN_STREAMEXTERNAL = "617d"; //$NON-NLS-1$
public static final String C_WARN_STREAMLENGTH = "617e"; //$NON-NLS-1$
public static final String C_WARN_SINGLESPACE_OBJ = "618a"; //$NON-NLS-1$
public static final String C_WARN_SINGLEEOL_OBJ = "618b"; //$NON-NLS-1$
public static final String C_WARN_ENDOBJ_MISSING = "618c"; //$NON-NLS-1$
public static final String C_WARN_LARGE_INT = "6112a"; //$NON-NLS-1$
protected static final String C_TOKEN_ADDWSB = "additional whitespace before"; //$NON-NLS-1$
protected static final String C_TOKEN_WSB = "whitespace before"; //$NON-NLS-1$
protected static final String C_TOKEN_ADDWSA = "additional whitespace after"; //$NON-NLS-1$
protected static final String C_TOKEN_ADDWSA2 = "second add whitespace after"; //$NON-NLS-1$
protected static final String C_TOKEN_COMMENT = "comment"; //$NON-NLS-1$
protected static final String C_TOKEN_NOWSA = "no whitespace after"; //$NON-NLS-1$
protected static final byte[] characterClass = new byte[256];
protected static final byte CHARCLASS_ANY = 0;
protected static final byte CHARCLASS_DELIMITER = 1;
protected static final byte CHARCLASS_WHITESPACE = 2;
protected static final byte CHARCLASS_TOKEN = 3;
protected static final byte CHARCLASS_DIGIT = 4;
protected static final byte CHARCLASS_NUMBERSPECIAL = 5;
public static final byte[] TOKEN_def = "def".getBytes(); //$NON-NLS-1$
static {
for (int i = 0; i < 256; i++) {
characterClass[i] = CHARCLASS_ANY;
}
// delimiters
characterClass['('] = CHARCLASS_DELIMITER;
characterClass[')'] = CHARCLASS_DELIMITER;
characterClass['<'] = CHARCLASS_DELIMITER;
characterClass['>'] = CHARCLASS_DELIMITER;
characterClass['['] = CHARCLASS_DELIMITER;
characterClass[']'] = CHARCLASS_DELIMITER;
characterClass['{'] = CHARCLASS_DELIMITER;
characterClass['}'] = CHARCLASS_DELIMITER;
characterClass['/'] = CHARCLASS_DELIMITER;
characterClass['%'] = CHARCLASS_DELIMITER;
// whitespace
characterClass[' '] = CHARCLASS_WHITESPACE;
characterClass['\t'] = CHARCLASS_WHITESPACE;
characterClass['\r'] = CHARCLASS_WHITESPACE;
characterClass['\n'] = CHARCLASS_WHITESPACE;
characterClass[12] = CHARCLASS_WHITESPACE;
characterClass[0] = CHARCLASS_WHITESPACE;
// digits
characterClass['0'] = CHARCLASS_DIGIT;
characterClass['1'] = CHARCLASS_DIGIT;
characterClass['2'] = CHARCLASS_DIGIT;
characterClass['3'] = CHARCLASS_DIGIT;
characterClass['4'] = CHARCLASS_DIGIT;
characterClass['5'] = CHARCLASS_DIGIT;
characterClass['6'] = CHARCLASS_DIGIT;
characterClass['7'] = CHARCLASS_DIGIT;
characterClass['8'] = CHARCLASS_DIGIT;
characterClass['9'] = CHARCLASS_DIGIT;
// number special
characterClass['.'] = CHARCLASS_NUMBERSPECIAL;
characterClass['-'] = CHARCLASS_NUMBERSPECIAL;
characterClass['+'] = CHARCLASS_NUMBERSPECIAL;
// alpha
for (int i = 'a'; i <= 'z'; i++) {
characterClass[i] = CHARCLASS_TOKEN;
}
for (int i = 'A'; i <= 'Z'; i++) {
characterClass[i] = CHARCLASS_TOKEN;
}
// contentstream allowed token characters
characterClass['\''] = CHARCLASS_TOKEN;
characterClass['"'] = CHARCLASS_TOKEN;
}
/**
* evaluate to true if i is a PDF Delimiter char.
*
*
* See pdf spec delimiter characters.
*
*
* @param i i a byte representation
* @return true if i is a PDF delimiter char
*/
public static boolean isDelimiter(int i) {
return characterClass[i] == CHARCLASS_DELIMITER;
}
/**
* evaluate to true if i is a valid digit.
*
* @param i i a byte representation
* @return true if i is a valid digit
*/
public static boolean isDigit(int i) {
return characterClass[i] == CHARCLASS_DIGIT;
}
/**
* evaluate to true if i is a valid line terminator.
*
* @param i i a byte representation
* @return true if i is a valid line terminator
*/
public static boolean isEOL(int i) {
return (i == CHAR_CR) || (i == CHAR_LF) || (i == 12);
}
/**
* evaluate to true if i is a valid first char for a number token.
*
* @param i i a byte representation
* @return true if i is a valid first char for a number token
*/
public static boolean isNumberStart(int i) {
int cc = characterClass[i];
return (cc == CHARCLASS_DIGIT) || (cc == CHARCLASS_NUMBERSPECIAL);
}
/**
* evaluate to true if i is a valid octal digit.
*
* @param i i a byte representation
* @return true if i is a valid octal digit
*/
public static boolean isOctalDigit(int i) {
return ((i >= '0') && (i <= '7'));
}
/**
* evaluate to true if i is a valid string token start.
*
* @param i i a byte representation
* @return true if i is a valid string token start
*/
public static boolean isTokenStart(int i) {
return characterClass[i] == CHARCLASS_TOKEN;
}
/**
* evaluate to true if i is a valid whitespace.
*
*
* See pdf spec "white space characters"
*
*
* @param i i a byte representation
* @return true if i is a valid whitespace
*/
public static boolean isWhitespace(int i) {
return characterClass[i] == CHARCLASS_WHITESPACE;
}
/**
* parse the given byte array to a valid COSObject.
*
* @param data a byte array containing COS encoded objects
* @return a COSObject
* @throws IOException
* @throws COSLoadException
*/
public static COSObject toCOSObject(byte[] data) throws IOException, COSLoadException {
COSDocumentParser docParser = new COSDocumentParser(null);
return (COSObject) docParser.parseElement(new RandomAccessByteArray(data));
}
/**
* A list for object lookahead (needed with PDF references)
*/
private COSObject[] lookahead = {null, null, null};
/**
* The number of elements currently in the lookahead buffer.
*/
private int lookaheadCount = 0;
private ISystemSecurityHandler securityHandler;
/**
* A flag indicating we should flush the lookahead
*/
private boolean flushLookahead = false;
private FastByteArrayOutputStream localStream = new FastByteArrayOutputStream();
/**
* an exception handler for handling messages eg PDFA compliance checks *
*/
private IPDFParserExceptionHandler exceptionHandler;
private COSObjectKey objectKey;
protected boolean check = false;
protected abstract COSIndirectObject createObjectReference(IRandomAccess input)
throws IOException, COSLoadException;
public IPDFParserExceptionHandler getExceptionHandler() {
return exceptionHandler;
}
protected COSObjectKey getObjectKey() {
return objectKey;
}
protected ISystemSecurityHandler getSecurityHandler() {
return securityHandler;
}
/**
* Handle an error if an exceptionHandler is set.
*
* @param error
* @throws COSLoadException
*/
public void handleError(COSLoadError error) throws COSLoadException {
if (exceptionHandler != null) {
exceptionHandler.error(error);
} else {
throw error;
}
}
/**
* Handle a warning if an exceptionHandler is set.
*
* @param warning
* @throws COSLoadException
*/
public void handleWarning(COSLoadWarning warning) throws COSLoadException {
if (exceptionHandler != null) {
exceptionHandler.warning(warning);
} else {
// it is just a warning...
}
}
/**
* in order to read references we need a two object lookahead for the
* integer numbers this method pops the first object from the fifo
* structure.
*
* @return The topmost {@link COSObject}in the object lookahead buffer.
*/
protected COSObject lookaheadPop() {
COSObject result = lookahead[0];
lookahead[0] = lookahead[1];
lookahead[1] = lookahead[2];
lookahead[2] = null;
lookaheadCount--;
if (lookaheadCount <= 0) {
// everything flushed now
lookaheadCount = 0;
this.flushLookahead = false;
}
return result;
}
/**
* in order to read references we need a two object lookahead for the
* integer numbers this method pushes an object in the fifo structure.
*
* @param obj The {@link COSObject}to push in the buffer.
*/
protected void lookaheadPush(COSObject obj) {
lookahead[lookaheadCount++] = obj;
}
/**
* comment see PDF Reference v1.4, chapter 3.1.2 comments Comment ::= "%"
* anyChar EOL read until end of line.
*
* @throws IOException
*/
protected void parseComment(IRandomAccess input) throws IOException {
int next;
while (true) {
next = input.read();
if (next == -1) {
break;
}
if (isEOL(next)) {
break;
}
}
}
/**
* parse the basic elements from the current stream position.
*
*
* see PDF Reference v1.4, chapter 3.2 Objects
*
*
*
* COSObject ::= COSToken | COSBoolean | COSString | COSNumber | COSName |
* COSNull | COSArray | COSDictionary | COSStream
*
*
* @return the object parsed
* @throws IOException
* @throws COSLoadException
*/
public Object parseElement(IRandomAccess input) throws IOException, COSLoadException {
int next;
do {
next = input.read();
if (next == -1) {
// thats a legal end
return null;
}
// we have found a non-whitespace character
if (isNumberStart(next)) {
return parseOnObjectNumber(input, next);
}
if (next == '(') {
return parseOnObjectString(input);
}
if (isTokenStart(next)) {
byte[] token = readTokenElement(input, next);
if (token.length == 1) {
if (token[0] == TOKEN_R[0]) {
return TOKEN_R;
}
} else if (token.length == 4) {
if ((token[0] == TOKEN_true[0]) && (token[1] == TOKEN_true[1]) && (token[2] == TOKEN_true[2]) && (
token[3]
== TOKEN_true[3])) {
return COSTrue.create();
}
if ((token[0] == TOKEN_null[0]) && (token[1] == TOKEN_null[1]) && (token[2] == TOKEN_null[2]) && (
token[3]
== TOKEN_null[3])) {
return COSNull.create();
}
} else if (token.length == 5) {
if ((token[0] == TOKEN_false[0])
&& (token[1] == TOKEN_false[1])
&& (token[2] == TOKEN_false[2])
&& (token[3] == TOKEN_false[3])
&& (token[4] == TOKEN_false[4])) {
return COSFalse.create();
}
}
return token;
}
if (next == '/') {
return parseOnObjectName(input);
}
// performance shortcut for simple space
if ((next == ' ') || isWhitespace(next)) {
continue;
}
if (next == '%') {
parseComment(input);
continue;
}
// before we start parsing a container we must flush lookahead
if (lookaheadCount > 0) {
input.seekBy(-1);
return null;
}
if (next == '<') {
return parseOnObjectStreamOrDictionaryOrHexString(input);
}
if (next == '[') {
return parseOnObjectArray(input);
}
// unread, i do not understand...
// return null if char unexpected, if this is an error depends on
// context
input.seekBy(-1);
return null;
} while (true);
}
/**
* pdf header see PDF Reference v1.4, chapter 3.4.1 Header COSHEader ::=
* "%PDF-" version.
*
* @throws IOException
* @throws COSLoadException
*/
public STDocType parseHeader(IRandomAccess input) throws IOException, COSLoadException {
int next;
boolean errHeader = false;
while (true) {
next = input.read();
if (next == -1) {
break;
}
// read up to %
if ((next != '%')) {
errHeader = true;
continue;
}
break;
}
STDocType docType = new STDocType();
if (next == -1) {
COSLoadError e = new COSLoadError("file format error. document must start with %PDF or %FDF");
handleError(e);
} else {
byte[] token = new byte[4];
token[0] = (byte) next;
input.read(token, 1, 3);
if (Arrays.equals(token, TOKEN_PDFHEADER)) {
docType.setTypeName("PDF");
} else if (Arrays.equals(token, TOKEN_FDFHEADER)) {
docType.setTypeName("FDF");
} else {
input.seekBy(-token.length);
COSLoadError e =
new COSLoadError("file format error. document must start with %PDF or %FDF at character index "
+ input.getOffset());
handleError(e);
}
if (errHeader) {
COSLoadWarning w = new COSLoadWarning(
"file format error. document must start with %PDF or %FDF at character index "
+ input.getOffset());
handleWarning(w);
}
input.read();
byte[] version = readToken(input);
if (version == null) {
COSLoadError e = new COSLoadError("file format error. no pdf/fdf version info found at character index "
+ input.getOffset());
handleError(e);
} else {
docType.setVersion(StringTools.toString(version));
}
}
return docType;
}
/**
* Parse a valid COS object for use in document context from the current
* stream position.
*
*
* see PDF Reference v1.4, chapter 3.2 Objects
*
*
*
* this implementation is a little more complicated, as we hava a two object
* lookahead to detect references.
*
* {@code
* COSObject ::= COSReference |
* COSBoolean |
* COSString |
* COSNumber |
* COSName |
* COSNull |
* COSArray |
* COSDictionary |
* COSStream
*
* }
*
*
*
* @return the object parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSDocumentElement parseObject(IRandomAccess input) throws IOException, COSLoadException {
if (flushLookahead) {
return lookaheadPop();
}
// parse another element
Object parsedElement = parseElement(input);
if (parsedElement == null) {
flushLookahead = true;
return lookaheadPop();
}
// try to detect reference "R"
COSObject resultObject;
if (parsedElement instanceof byte[]) {
if (TOKEN_R == parsedElement) {
// reference detected, clean up lookahed and return
return createObjectReference(input);
}
// we have found a token that has to be re-read in another context
// take care of consumed whitespace!
input.seekBy(-1);
int next = input.read();
// performance shortcut for simple space
if ((next == ' ') || isWhitespace(next)) {
input.seekBy(-1);
}
input.seekBy(-((byte[]) parsedElement).length);
this.flushLookahead = true;
return lookaheadPop();
}
resultObject = (COSObject) parsedElement;
// build up lookahead stack
if (resultObject instanceof COSNumber) {
lookaheadPush(resultObject);
// return one object if lookahead larger than 2
if (lookaheadCount > 2) {
return lookaheadPop();
}
// enter parse recursive
return parseObject(input);
}
// shortcut to avoid building entry in lookahead
if (lookaheadCount > 0) {
lookaheadPush(resultObject);
this.flushLookahead = true;
return lookaheadPop();
}
return resultObject;
}
protected COSObject parseObjectDictionary(IRandomAccess input) throws IOException, COSLoadException {
int next;
next = input.read();
if (next != '<') {
input.seekBy(-1);
COSLoadError e = new COSLoadError("'<' expected at character index " + input.getOffset());
handleError(e);
}
next = input.read();
if (next != '<') {
input.seekBy(-1);
COSLoadError e = new COSLoadError("'<' expected at character index " + input.getOffset());
handleError(e);
}
return parseOnObjectDictionary(input);
}
/**
* parse a COS array from the current stream position. see PDF Reference
* v1.4, chapter 3.2.5 Array objects COSArray ::= "[" (COSObject) "]"
*
* @return the array parsed
* @throws IOException
* @throws IOException
*/
protected COSObject parseOnObjectArray(IRandomAccess input) throws COSLoadException, IOException {
try {
int next;
COSArray result = COSArray.create();
if (securityHandler != null) {
securityHandler.pushContextObject(result);
}
while (true) {
COSDocumentElement element = parseObject(input);
if (element == null) {
next = input.read();
if (next == -1) {
unexpectedEndOfInput(input);
}
if (next != ']') {
byte[] badElement = readTokenElement(input, next);
if (check) {
COSLoadWarning pwarn =
new COSLoadWarning("bad array element (" + new String(badElement) + ")");
pwarn.setHint(result);
handleWarning(pwarn);
}
continue;
}
break;
}
result.basicAddSilent(element);
}
if (check && (result.size() > 8191)) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_ARRAYSIZE);
pwarn.setHint(result);
handleWarning(pwarn);
}
return result;
} finally {
if (securityHandler != null) {
securityHandler.popContextObject();
}
}
}
/**
* parse a COS dictionary from the current stream position. see PDF
* Reference v1.4, chapter 3.2.6 Dictionary objects
*
* {@code
* COSDictionary ::= "<<" (COSName COSObject)* ">>"
* }
*
* @return the dictionary parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectDictionary(IRandomAccess input) throws IOException, COSLoadException {
try {
int next;
COSDictionary dict = COSDictionary.create();
if (securityHandler != null) {
securityHandler.pushContextObject(dict);
}
try {
while (true) {
COSDocumentElement keyObject = parseObject(input);
if (keyObject == null) {
// when parsing dictionaries in CMap we may encounter
// the
// keyword "def" - don't know if this is legal, but
// happens...
input.mark();
Object tempElement = parseElement(input);
if (tempElement != null) {
// try to detect reference "def"
if (tempElement instanceof byte[]) {
if (Arrays.equals(TOKEN_def, (byte[]) tempElement)) {
// this is no-op
continue;
}
}
}
input.reset();
break;
}
COSName dictKey = (COSName) keyObject;
COSDocumentElement value = parseObject(input);
if (value == null) {
COSLoadError e =
new COSLoadError("missing value for key '" + keyObject + "' at character index " + input
.getOffset());
handleError(e);
} else {
dict.basicPutSilent(dictKey, value);
}
}
} catch (ClassCastException ignored) {
COSLoadError e = new COSLoadError("name expected at character index " + input.getOffset());
handleError(e);
}
next = input.read();
if (next != '>') {
COSLoadError e = new COSLoadError("unexpected character ("
+ (char) next
+ ") at character index "
+ input.getOffset());
handleError(e);
}
next = input.read();
if (next != '>') {
COSLoadError e = new COSLoadError("unexpected character ("
+ (char) next
+ ") at character index "
+ input.getOffset());
handleError(e);
}
return dict;
} finally {
if (securityHandler != null) {
securityHandler.popContextObject();
}
}
}
/**
* parse a COS string encoded in hex from the current stream position. see
* PDF Reference v1.4, chapter 3.2.3 String objects
*
* {@code
* COSString ::= COSString | COSHexString
* COSHexString ::= "<" (hexChar)* ">"
* }
*
* @return the string parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectHexString(IRandomAccess input, int next) throws IOException, COSLoadException {
localStream.reset();
boolean secondDigit = false;
int digitValue = 0;
int charValue = 0;
while (true) {
digitValue = HexTools.hexDigitToInt((char) next);
if (digitValue == -1) {
if (next == -1) {
break;
}
if (next == '>') {
break;
}
if (!isWhitespace(next)) {
IOException ioe = new IOException("<" + next + "> '" + (char) next + "' not a valid hex char");
// todo 3 @mit Warning is useless. Due to the IOException, such documents cannot be loaded anyway. Remove it?
// a warning for PDF/A related checks will be triggered
// exception is handled right on track
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_ILLEGALHEX);
pwarn.setHint(Long.valueOf(input.getOffset()));
handleWarning(pwarn);
throw ioe;
}
} else {
if (secondDigit) {
charValue = (charValue << 4) + digitValue;
localStream.write(charValue);
secondDigit = false;
} else {
secondDigit = true;
charValue = digitValue;
}
}
next = input.read();
}
if (secondDigit) {
// this is a warning for uneven numbers on hex codes
if (check) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_UNEVENHEX);
pwarn.setHint(Long.valueOf(input.getOffset()));
handleWarning(pwarn);
}
// assume trailing "0"
charValue = charValue << 4;
localStream.write(charValue);
}
COSString result;
if ((securityHandler == null) || (objectKey == null)) {
result = COSString.createHex(localStream.toByteArray());
} else {
byte[] bytes = localStream.toByteArray();
try {
byte[] decrypted = securityHandler.decryptString(objectKey, bytes);
result = COSString.createHex(decrypted);
} catch (COSSecurityException e) {
result = COSString.createHex(bytes);
COSLoadWarning warning = new COSLoadWarning("error decrypting string " + objectKey, e);
handleWarning(warning);
}
}
if (check && (result.stringValue().length() > 32767)) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_STRING_TOO_LONG);
pwarn.setHint(result);
handleWarning(pwarn);
}
return result;
}
/**
* parse a COS name from the current stream position. see PDF Reference
* v1.4, chapter 3.2.4 Name Objects COSName ::= "/" nameChars
*
* @return the name parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectName(IRandomAccess input) throws IOException, COSLoadException {
int next;
localStream.reset();
do {
next = input.read();
if (next == -1) {
break;
}
// performance shortcut for simple space
if ((next == ' ') || isWhitespace(next)) {
break;
}
if (isDelimiter(next)) {
input.seekBy(-1);
break;
}
if (next == '#') {
next = input.read();
int digit1 = HexTools.hexDigitToInt((char) next);
if (digit1 == -1) {
COSLoadError e = new COSLoadError("<"
+ next
+ "> not a valid hex char at character index "
+ input.getOffset());
handleError(e);
}
next = input.read();
int digit2 = HexTools.hexDigitToInt((char) next);
if (digit2 == -1) {
COSLoadError e = new COSLoadError("<"
+ next
+ "> not a valid hex char at character index "
+ input.getOffset());
handleError(e);
}
localStream.write((digit1 << 4) + digit2);
} else {
localStream.write(next);
}
} while (true);
byte[] bytes = localStream.toByteArray();
COSName result = COSName.create(bytes);
if (check && (result.stringValue().length() > 127)) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_NAME_TOO_LONG);
pwarn.setHint(result);
handleWarning(pwarn);
}
return result;
}
/**
* parse a COS number from the current stream position. see PDF Reference
* v1.4, chapter 3.2.2 Numeric objects COSNumber ::= COSFixed | COSInteger
* COSFixed ::= (+ | -)? (digit) "." (digit) COSInteger ::= (+ | -)? (digit)
*
* @return the number parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectNumber(IRandomAccess input, int next) throws IOException, COSLoadException {
boolean isFixed = false;
localStream.reset();
isFixed = next == '.';
localStream.write((byte) next);
do {
next = input.read();
if (next == -1) {
break;
} else if (isDigit(next)) {
localStream.write((byte) next);
} else if (next == '.') {
isFixed = true;
localStream.write((byte) '.');
} else if ((next == ' ') || isWhitespace(next)) {
break;
} else {
input.seekBy(-1);
break;
}
} while (true);
if (isFixed) {
return COSFixed.create(localStream.getBytes(), 0, localStream.size());
}
byte[] streamBytes = localStream.getBytes();
int streamSize = localStream.size();
if (exceptionHandler != null) {
COSInteger result = COSInteger.createStrict(streamBytes, 0, streamSize);
if (result != null) {
return result;
}
COSLoadWarning warning = new COSLoadWarning(C_WARN_LARGE_INT);
handleWarning(warning);
}
return COSInteger.create(streamBytes, 0, streamSize);
}
/**
* parse a COS stream from the current stream position. see PDF Reference
* v1.4, chapter 3.2.7 Stream objects COSStream ::= COSDictionary "stream"
* bytes "endstream"
*
* @param dict The object that should be filled with the dictionary entries.
* @return The stream parsed.
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectStream(IRandomAccess input, COSDictionary dict)
throws IOException, COSLoadException {
COSStream stream = COSStream.create(dict);
byte[] token = new byte[5];
// read "tream", "s" already consumed
input.read(token);
if (!Arrays.equals(token, TOKEN_s_tream)) {
input.seekBy(-token.length - 1);
COSLoadError e =
new COSLoadError("file format error. 'stream' expected at character index " + input.getOffset());
handleError(e);
}
// allow for at max two separator chars after "stream"
int next;
next = input.read();
if (next == -1) {
unexpectedEndOfInput(input);
}
if (next == CHAR_CR) {
next = input.read();
}
if (next != CHAR_LF) {
// ?? its legal to have NO separator
// ?? there are testdocuments that provide only a single CR
if (check) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_STREAMEOL);
pwarn.setHint(Long.valueOf(input.getOffset()));
handleWarning(pwarn);
}
input.seekBy(-1);
}
long offset = input.getOffset();
int length = -1;
COSNumber cosLength = dict.get(COSStream.DK_Length).asInteger();
if (cosLength == null) {
// warning for pdfa
if (check) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_STREAMLENGTH);
pwarn.setHint(Long.valueOf(input.getOffset()));
handleWarning(pwarn);
}
} else {
length = cosLength.intValue();
}
// may be moved by reading indirect /Length !
input.seek(offset);
byte[] bytes = null;
if (length < 0) {
bytes = readStream(input);
} else {
bytes = new byte[length];
int count = input.read(bytes);
if (count < length) {
if (check) {
// get additional warning for pdfa
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_STREAMLENGTH);
pwarn.setHint(Long.valueOf(input.getOffset()));
handleWarning(pwarn);
}
unexpectedEndOfInput(input);
}
}
if (check) {
// pdfa compliance check
int test = readEOL(input);
if (test != 1) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_ENDSTREAMEOL);
pwarn.setHint(Long.valueOf(input.getOffset()));
handleWarning(pwarn);
}
} else {
// be lazy with pdf spec and accept any whitespace before
// 'endstream'
readSpaces(input);
}
// read "endstream"
token = new byte[9];
input.read(token);
if (!Arrays.equals(token, TOKEN_endstream)) {
input.seekBy(-token.length - 1);
// a warning for PDF/A related checks will be triggered
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_ENDSTREAMCORRUPT);
pwarn.setHint(Long.valueOf(input.getOffset()));
handleWarning(pwarn);
if (length > 0) {
// retry from the beginning with undeterminate length
input.seek(offset);
bytes = readStream(input);
// read "endstream"
token = new byte[9];
input.read(token);
if (!Arrays.equals(token, TOKEN_endstream)) {
COSLoadError e = new COSLoadError("file format error. 'endstream' expected at character index "
+ input.getOffset());
handleError(e);
}
} else {
COSLoadError e = new COSLoadError("file format error. 'endstream' expected at character index "
+ input.getOffset());
handleError(e);
}
}
if ((securityHandler == null) || (objectKey == null)) {
stream.basicSetEncodedBytes(bytes);
} else {
try {
byte[] decrypted = securityHandler.decryptStream(objectKey, dict, bytes);
stream.basicSetEncodedBytes(decrypted);
} catch (COSSecurityException e) {
stream.basicSetEncodedBytes(bytes);
COSLoadWarning warning = new COSLoadWarning("error decrypting stream " + objectKey, e);
handleWarning(warning);
}
}
return stream;
}
/**
* parse a COS stream or dictionary from the current stream position.
* COSStreamOrDict ::= COSStream | COSDict
*
* @return the object parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectStreamOrDictionary(IRandomAccess input) throws IOException, COSLoadException {
COSObject dict = parseOnObjectDictionary(input);
int next;
boolean lastWasEOL = false;
while (true) {
next = input.read();
if (next == -1) {
return dict;
}
// performance shortcut for simple space
if ((next == ' ') || isWhitespace(next)) {
lastWasEOL = next == '\n' || next == '\r';
continue;
}
break;
}
if (next == 's') {
return parseOnObjectStream(input, (COSDictionary) dict);
}
if (next == 'e' && check && !lastWasEOL) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_SINGLEEOL_OBJ);
pwarn.setHint(new Long(input.getOffset()));
handleWarning(pwarn);
}
input.seekBy(-1);
return dict;
}
/**
* parse a COS stream or dictionary or hex string from the current stream
* position. COSStreamOrDictOrHex ::= COSStream | COSDict | COSHexString
*
* @return the object parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectStreamOrDictionaryOrHexString(IRandomAccess input)
throws IOException, COSLoadException {
int next;
next = input.read();
if (next == '<') {
return parseOnObjectStreamOrDictionary(input);
}
return parseOnObjectHexString(input, next);
}
/**
* parse a COS string from the current stream position. see PDF Reference
* v1.4, chapter 3.2.3. String objects COSString ::= "(" stringData ")"
*
* @return the string parsed
* @throws IOException
* @throws COSLoadException
*/
protected COSObject parseOnObjectString(IRandomAccess input) throws IOException, COSLoadException {
int next;
int paraCount = 0;
localStream.reset();
while (true) {
next = input.read();
if (next == '\\') {
int c = readEscape(input);
if (c != -1) {
localStream.write(c);
}
} else if (next == ')') {
if (paraCount > 0) {
paraCount--;
localStream.write(')');
} else {
break;
}
} else if (next == CHAR_CR) {
// eol is always \n in a string
next = input.read();
if (next != -1 && next != CHAR_LF) {
input.seekBy(-1);
}
localStream.write(CHAR_LF);
} else if (next == '(') {
paraCount++;
localStream.write('(');
} else if (next == -1) {
unexpectedEndOfInput(input);
} else {
localStream.write(next);
}
}
COSString result;
if ((securityHandler == null) || (objectKey == null)) {
result = COSString.create(localStream.toByteArray());
} else {
byte[] bytes = localStream.toByteArray();
try {
byte[] decrypted = securityHandler.decryptString(objectKey, bytes);
result = COSString.create(decrypted);
} catch (COSSecurityException e) {
result = COSString.create(bytes);
COSLoadWarning warning = new COSLoadWarning("error decrypting string " + objectKey, e);
handleWarning(warning);
}
}
if (check && (result.stringValue().length() > 32767)) {
COSLoadWarning pwarn = new COSLoadWarning(C_WARN_STRING_TOO_LONG);
pwarn.setHint(result);
handleWarning(pwarn);
}
return result;
}
/**
* determine number of EOL sequences
*
* @param input
* @return {@code number of EOL}
* @throws IOException
*/
protected int readEOL(IRandomAccess input) throws IOException {
int next = input.read();
if (next == -1) {
return 0;
}
if (next == CHAR_CR) {
next = input.read();
if (next == -1) {
return 1;
} else if (next == CHAR_LF) {
next = input.read();
if (next == -1) {
return 1;
} else if (isWhitespace(next)) {
readSpaces(input);
return 2;
} else {
input.seekBy(-1);
return 1;
}
} else if (isWhitespace(next)) {
readSpaces(input);
return 2;
} else {
input.seekBy(-1);
return 1;
}
}
if (next == CHAR_LF) {
next = input.read();
if (next == -1) {
return 1;
} else if (isWhitespace(next)) {
readSpaces(input);
return 2;
} else {
input.seekBy(-1);
return 1;
}
}
if (isWhitespace(next)) {
readSpaces(input);
return 2;
}
input.seekBy(-1);
return 0;
}
/**
* read an esacped char from the stream.
*
* @return the character corresponding to the escape code
* @throws IOException
*/
protected int readEscape(IRandomAccess input) throws IOException {
int next = 0;
next = input.read();
if (next == -1) {
return -1;
}
if (isOctalDigit(next)) {
input.seekBy(-1);
return readOctalChar(input);
}
if (next == CHAR_LF) {
return -1;
}
if (next == CHAR_CR) {
next = input.read();
if (next != -1 && next != CHAR_LF) {
input.seekBy(-1);
}
return -1;
}
if (next == 'n') {
return CHAR_LF;
}
if (next == 'r') {
return CHAR_CR;
}
if (next == 't') {
return CHAR_HT;
}
if (next == 'b') {
return CHAR_BS;
}
if (next == 'f') {
return CHAR_FF;
}
return next;
}
/**
* reads the next integer on input. consumes one trailing space if
* consumeSpaceAfter is set to true. Consumes leading spaces and comments.
*
* @param input
* @param consumeSpaceAfter
* @return The integer read.
* @throws IOException
*/
public int readInteger(IRandomAccess input, boolean consumeSpaceAfter) throws IOException {
int result = 0;
int next;
while (true) {
next = input.read();
if (next == -1) {
return result;
} else if ((next == ' ') || isWhitespace(next)) {
continue;
} else if (next == '%') {
parseComment(input);
} else {
break;
}
}
// avoid returning 0 for degenerate case
boolean digitFound = false;
while (true) {
if (isDigit(next)) {
digitFound = true;
result = ((result * 10) + next) - '0';
} else {
if (!digitFound) {
throw new IOException("digit expected at " + input.getOffset());
}
input.seekBy(-1);
break;
}
next = input.read();
if (next == -1) {
break;
} else if ((next == ' ') || isWhitespace(next)) {
if (!consumeSpaceAfter) {
input.seekBy(-1);
}
break;
}
}
return result;
}
/**
* read an octal character from the stream.
*
* @return the integer value of the character read or -1
* @throws IOException
*/
protected int readOctalChar(IRandomAccess input) throws IOException {
int result = -1;
int c = 0;
c = input.read();
if (isOctalDigit(c)) {
result = c - '0';
c = input.read();
if (isOctalDigit(c)) {
result = ((result << 3) + c) - '0';
c = input.read();
if (isOctalDigit(c)) {
result = ((result << 3) + c) - '0';
} else {
if (c == -1) {
return result;
}
input.seekBy(-1);
}
} else {
if (c == -1) {
return result;
}
input.seekBy(-1);
}
} else {
if (c == -1) {
return result;
}
input.seekBy(-1);
}
return result;
}
/**
* read all characters until EOF or non space char appears. the first non
* space char is pushed back so the next char read is the first non space
* char.
*
* @throws IOException
*/
public void readSpaces(IRandomAccess input) throws IOException {
int next = 0;
while (true) {
next = input.read();
if (next == -1) {
break;
}
// performance shortcut for simple space
if ((next == ' ') || isWhitespace(next)) {
continue;
}
input.seekBy(-1);
break;
}
}
/**
* Read all characters up to "endstream" and assume them belonging to the
* stream.
*
* ATTENTION this is a heuristic approach as the tag "endstream" may be part
* of the stream data!
*
* @return All characters up to "endstream"
* @throws IOException
*/
protected byte[] readStream(IRandomAccess input) throws IOException {
byte[] token = new byte[8];
localStream.reset();
int next;
while (true) {
next = input.read();
if (next == 'e') {
input.read(token);
if (Arrays.equals(token, TOKEN_ndstream)) {
input.seekBy(-TOKEN_endstream.length);
return localStream.toByteArray();
}
input.seekBy(-token.length);
} else if (next == -1) {
break;
}
localStream.write(next);
}
if (localStream.size() == 0) {
return null;
}
return localStream.toByteArray();
}
/**
* read a single token.
*
* @return the array of characters belonging to the token
* @throws IOException
*/
public byte[] readToken(IRandomAccess input) throws IOException {
//
int next;
while (true) {
next = input.read();
if (next == -1) {
return null;
} else if ((next == ' ') || isWhitespace(next)) {
continue;
} else if (next == '%') {
parseComment(input);
} else {
break;
}
}
return readTokenElement(input, next);
}
/**
* derive of readToken, populates the messages list with non-fatal error
* messages
*
* @param input
* @param messages
* @return token bytes
* @throws IOException
*/
public byte[] readToken(IRandomAccess input, List messages, boolean strict) throws IOException {
int next;
int countWS = 0;
boolean crEol = false;
while (true) {
next = input.read();
if (next == -1) {
return null;
} else if ((next == ' ') || isWhitespace(next)) {
if (!strict && !crEol && (next == ' ')) {
// ignore ignorable space
} else {
countWS++;
}
if (countWS > 1 && !(crEol && next == '\n')) {
messages.add(C_TOKEN_ADDWSB);
} else if (strict && (next == ' ')) {
messages.add(C_TOKEN_WSB);
} else { // may be CR+EOL
crEol = next == '\r';
}
continue;
} else if (next == '%') {
messages.add(C_TOKEN_COMMENT);
parseComment(input);
} else {
break;
}
}
return readTokenElement(input, next, messages);
}
protected byte[] readTokenElement(IRandomAccess input, int next) throws IOException {
localStream.reset();
localStream.write(next);
do {
next = input.read();
if (next == -1 || (next == ' ') || isWhitespace(next)) {
// performance shortcut
break;
}
if (isDelimiter(next)) {
input.seekBy(-1);
break;
}
localStream.write(next);
} while (true);
return localStream.toByteArray();
}
/**
* derive of readToken, populates the messages list with non-fatal error
* messages
*
* @param input
* @param next
* @param messages
* @return token bytes
* @throws IOException
*/
protected byte[] readTokenElement(IRandomAccess input, int next, List messages) throws IOException {
localStream.reset();
localStream.write(next);
do {
next = input.read();
if (next == -1) {
break;
}
if ((next == ' ') || isWhitespace(next)) { // performance
// shortcut
if (next == ' ') {
messages.add(C_TOKEN_ADDWSA);
}
next = input.read();
if ((next == ' ')) { // performance
// shortcut
messages.add(C_TOKEN_ADDWSA2);
}
if (next != -1) {
input.seekBy(-1);
}
break;
}
if (isDelimiter(next)) {
messages.add(C_TOKEN_NOWSA);
input.seekBy(-1);
break;
}
localStream.write(next);
} while (true);
return localStream.toByteArray();
}
public void setExceptionHandler(IPDFParserExceptionHandler exceptionHandler) {
this.exceptionHandler = exceptionHandler;
check = exceptionHandler != null;
}
protected void setObjectKey(COSObjectKey objectKey) {
this.objectKey = objectKey;
}
protected void setSecurityHandler(ISystemSecurityHandler securityHandler) {
this.securityHandler = securityHandler;
}
protected void unexpectedEndOfInput(IRandomAccess input) throws IOException, COSLoadException {
COSLoadError e =
new COSLoadError("file format error. unexpected end of input at character index " + input.getOffset());
handleError(e);
}
}