PDFBox image extraction fails with an ArrayOutOfBoundsException in
PDPixelMap.getRGBImage()
-------------------------------------------------------------------------------------------
Key: PDFBOX-574
URL: https://issues.apache.org/jira/browse/PDFBOX-574
Project: PDFBox
Issue Type: Bug
Components: Utilities
Affects Versions: 0.8.0-incubator
Environment: Java
Reporter: Ian Kaplan
The project that I'm working on has been using PDFBox for both text extraction
and image extraction from PDF documents. We wrote a class, PDFImageStripper,
which extends PDFStreamEngine:
{code:Java}
public class PDFImageStripper extends PDFStreamEngine
{code}
{code:Java}
public List<ExtractedImage> getImages(PDDocument document, String
documentFilename, File targetDirectory) throws IOException {
resetEngine();
this.document = document;
this.documentFilename = documentFilename;
this.targetDirectory = targetDirectory;
currentImageNumber = 1;
images.clear();
writeImages();
return images;
}
{code}
{code:Java}
private void writeImages() throws IOException {
List<PDPage> pages = (List<PDPage>)
document.getDocumentCatalog().getAllPages();
for (PDPage page : pages) {
if (page != null) {
processStream(page, page.findResources(),
page.getContents().getStream());
}
}
}
{code}
The call chain is shown below:
{noformat}
None.decode(byte[], byte[]) line: 57
PDPixelMap.getRGBImage() line: 182
PDPixelMap.write2OutputStream(OutputStream) line: 209
PDPixelMap(PDXObjectImage).write2file(File) line: 142
PDFImageStripper.saveImage(PDXObjectImage, String, File) line: 208
PDFImageStripper.processOperator(PDFOperator, List) line: 155
PDFImageStripper(PDFStreamEngine).processSubStream(PDPage, PDResources,
COSStream) line: 229
PDFImageStripper(PDFStreamEngine).processStream(PDPage, PDResources, COSStream)
line: 188
PDFImageStripper.writeImages() line: 113
{noformat}
There is an ArrayOutOfBoundsException in the decode method. The decode method
is nothing more than a wrapper for a call to System.arraycopy():
{code:Java}
public void decode(byte[] src, byte[] dest)
{
System.arraycopy(src,0,dest,0,src.length);
}
{code}
The problem is, the source array is larger than the destination array. This is
show (from the Eclipse debugger) below:
{noformat}
src byte[455112] (id=171)
dest byte[435456] (id=175)
{noformat}
The code that seems to be causing the problem is shown below. The branch that
this bug shows up on is the LZW_DECODE branch. Note that in the other code
branch, the code makes sure that there is no size problem.
{code:Java}
if( predictor < 10 ||
filters == null || !(filters.contains(
COSName.LZW_DECODE.getName()) ||
filters.contains( COSName.FLATE_DECODE.getName()) ) )
{
PredictorAlgorithm filter =
PredictorAlgorithm.getFilter(predictor);
filter.setWidth(width);
filter.setHeight(height);
filter.setBpp((bpc * 3) / 8);
filter.decode(array, bufferData);
}
else
{
System.arraycopy( array, 0,bufferData, 0,
(array.length<bufferData.length?array.length:
bufferData.length) );
}
{code}
One fix may be to simply change the code as follows (again, recall that the
"decode" method is nothing but a wrapper for System.arraycopy()):
{code:Java}
if( predictor < 10 ||
filters == null || !(filters.contains(
COSName.LZW_DECODE.getName()) ||
filters.contains( COSName.FLATE_DECODE.getName()) ) )
{
PredictorAlgorithm filter =
PredictorAlgorithm.getFilter(predictor);
filter.setWidth(width);
filter.setHeight(height);
filter.setBpp((bpc * 3) / 8);
}
System.arraycopy( array, 0,bufferData, 0,
(array.length<bufferData.length?array.length:
bufferData.length) );
{code}
If Jira allows me to attach a file that causes this problem I will do so.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
