[
https://issues.apache.org/jira/browse/PDFBOX-1073?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13069473#comment-13069473
]
Mehdi Houshmand commented on PDFBOX-1073:
-----------------------------------------
I had a little help from a colleague, and we decoded the TIFF image in
question, there seem to be two problems:
Firstly, the TIFF begins with an EOL, which breaks the TIFF decoder. However,
ignoring that, it seems that rather than having each code being white then
black sequentially (-[White|Black]-) this particular TIFF has white followed by
white. In fact, there is a long stream of White codes. In this case the TIFF
decoder throws an Error.
I'll try and upload the problem image, it seems to be re-creatable since Gimp
produces a similarly White-White encoded image.
> Error when handling a TIFF image
> --------------------------------
>
> Key: PDFBOX-1073
> URL: https://issues.apache.org/jira/browse/PDFBOX-1073
> Project: PDFBox
> Issue Type: Bug
> Components: PDModel
> Affects Versions: 1.6.0
> Environment: Ubuntu 11/04
> Reporter: Mehdi Houshmand
> Priority: Trivial
> Attachments: tiffErrorFix.patch
>
>
> There is an error in a TIFF image within a PDF document and when
> PDCcitt.getRGBImage() is called, there is a disparity between current PDFBox
> and v0.8.0. The issue is that there now TIFF decoding features whereas the
> older PDFBox used to hand that off to java.imageio.
> The bug here isn't in the Ccitt decoder, but rather in how
> PDCcitt.getRGBImage() handles an error. Rather than bombing out, I have
> reimplemented the old feature when an error is thrown.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
