To create the PdfReader:
if (strPassword != null)
{
reader = new PdfReader(strFile, strPassword.getBytes());
}
else
{
reader = new PdfReader(strFile);
}
Then, I get a copy of the RandomAccessFileOrArray:
file = reader.getSafeFile();
file.reOpen();
Extracting the stream is kinda complex. I use reader.getPageN() to get
the PdfDictionary for each page, and pass it to the following recursive
function which walks through the pdf until it gets to the single image
stream for the page, and returns it.
private PRStream getImageStream(PdfObject obj, int page)
{
PdfName im = new PdfName("Im" + page);
PdfName xf = new PdfName("Xf" + page);
PdfDictionary dict;
if (obj.isStream())
{
dict = (PdfDictionary)((PRStream)obj).get(PdfName.RESOURCES);
}
else
{
dict = (PdfDictionary)((PdfDictionary)obj).get(PdfName.RESOURCES);
}
dict = (PdfDictionary)dict.get(PdfName.XOBJECT);
if (dict.contains(im))
{
return (PRStream)PdfReader.getPdfObject(dict.get(im));
}
return getImageStream(PdfReader.getPdfObject(dict.get(xf)), page);
}
Next, I use PdfReader.getStreamBytesRaw() to get the raw bytes for the
image stream.
Then, I use TIFFFaxDecoder to decode the byte stream (our images are
CCITT fax encoded).
Finally, I use JAI functionality to wrap a BufferedImage around the byte
stream and use TIFFImageEncoder to encode the image to a FileOutputStream.
Works like a charm except for the "too many files open" exception every
few thousand or so pdfs.
John Pruitt
Paulo Soares wrote:
How are you creating the PdfReader? How are you extracting the stream?
-----Original Message-----
From: John Pruitt [mailto:[EMAIL PROTECTED]
Sent: Tuesday, December 20, 2005 4:38 PM
To: Paulo Soares
Cc: [email protected]
Subject: Re: [iText-questions] Possible File Resource Leak
Actually, I am converting pdfs to tiff. We have a special
case where our
pdfs all consist of full-page images. I use iText to get to the image
streams in the pdf, and use JAI to encode the images to tiff.
I am closing both the RandomAccessFileOrArray, and the
PdfReader I use
(although I think PdfReader.close() also closes the
RandomAccessFileOrArray). Is there anything else I need to close?
John Pruitt
Paulo Soares wrote:
I suppose you are converting tiff to pdf and not the other
way around.
Check that your code is something like this:
RandomAccessFileOrArray ra = null;
try {
ra = new RandomAccessFileOrArray(imgb);
Image img = TiffImage.getTiffImage(ra, 1);
return img;
} finally {
if (ra != null)
ra.close();
}
It will make sure that the file is closed if some exception
is throwned.
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On
Behalf Of John Pruitt
Sent: Tuesday, December 20, 2005 3:15 PM
To: [email protected]
Subject: [iText-questions] Possible File Resource Leak
I have a program that basically batch converts pdfs to
multi-page tiffs
using iText. I am running into a problem. In my testing, I had the
program batch convert 12,000 pdf files to multi-page tiff.
The program
worked like a charm for 8 hours, then it hit an exception:
"To many open
files." The program logged the exception and continued for 8
more hours
before hitting the same exception again. After that, the
exception came
more frequently, every 2 hours or so.
I have checked my code, and am closing file streams as I
finish with
them. I was wondering how likely it is that iText may not
be properly
releasing resources, and where I should look in the code to see.
Thank you,
John Pruitt
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep
through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web.
DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
iText-questions mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/itext-questions
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
iText-questions mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/itext-questions
