[
https://issues.apache.org/jira/browse/PDFBOX-2690?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14330194#comment-14330194
]
Tilman Hausherr commented on PDFBOX-2690:
-----------------------------------------
Workaround until saving in object streams is implemented: use qpdf with the
option "--object-streams=generate".
> Filesize becomes extremely large after saving
> ---------------------------------------------
>
> Key: PDFBOX-2690
> URL: https://issues.apache.org/jira/browse/PDFBOX-2690
> Project: PDFBox
> Issue Type: Bug
> Components: Writing
> Affects Versions: 1.8.8, 1.8.9, 2.0.0
> Environment: PDFBox 1.8.8, Java8u25, Windows 8.1
> Reporter: Brian Liu
> Attachments: input2-after-save.pdf, input2.pdf
>
>
> I am using PDFBox 1.8.8 to manipulate existing PDF files. After saving a
> document, the output file becomes several times larger than the original.
> This is undesirable.
> *How to reproduce my problem:*
> In the following code, PDFBox simply loads an existing PDF and then save it.
> Nothing else is done. Yet the file size still becomes several times larger.
> {code}
> import java.io.*;
> import org.apache.pdfbox.pdmodel.*;
> import org.apache.pdfbox.exceptions.*;
> class Test
> {
> public static void main(String[] args) throws IOException,
> COSVisitorException {
> PDDocument document = PDDocument.load("input2.pdf");
> document.save("input2-after-save.pdf");
> document.close();
> }
> }
> {code}
> Attached are two sample PDF files. input2.pdf is an original, unprocessed
> PDF. input2-after-save.pdf is processed by the code above. After
> processing, file size increases from 416kB to 1.25MB.
> *Possible reason:*
> Tilman Hausherr suggests that there is an enormous amount of "structure"
> information / object stream that is compressed in the input file, but not in
> the output file.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
