[
https://issues.apache.org/jira/browse/PDFBOX-3645?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15826394#comment-15826394
]
Tilman Hausherr commented on PDFBOX-3645:
-----------------------------------------
This was introduced with PDFBOX-318 /
https://sourceforge.net/p/pdfbox/bugs/471/ . I tried reading this file
("exception_version1.pdf") and got a correct version 1.3.
I suspect that the intention wasn't to allow split headers, but to prevent the
loop from taking too long. So my intention would be to leave this unchanged.
Do you have a PDF with a split header, or was this rather an observation you
made while looking through the code?
> Probably not handling header read on bad PDFs
> ---------------------------------------------
>
> Key: PDFBOX-3645
> URL: https://issues.apache.org/jira/browse/PDFBOX-3645
> Project: PDFBox
> Issue Type: Bug
> Components: Parsing
> Affects Versions: 2.0.4
> Reporter: G. Ralph Kuntz
> Priority: Minor
>
> In COSParser.parseHeader, the code on line 1893 tries to handle PDFs where
> the version number is split on a separate line from the "%PDF-". It assigns
> the newly read line to `header`, detects a line that starts with a digit,
> then breaks.
> On line 1902, the code tries again to match the header "%PDF-xxx", but will
> not succeed in the case where the version was on a separate line, because
> `header` will contain only the version number and not the leading `%PDF-`.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
