[
https://issues.apache.org/jira/browse/PDFBOX-6242?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18108205#comment-18108205
]
Tilman Hausherr commented on PDFBOX-6242:
-----------------------------------------
ok with replacement character because we'll get less support requests about
PDFs that can't be changed.
> XFDF export allows characters illegal in XML 1.0
> ------------------------------------------------
>
> Key: PDFBOX-6242
> URL: https://issues.apache.org/jira/browse/PDFBOX-6242
> Project: PDFBox
> Issue Type: Bug
> Components: PDModel
> Affects Versions: 2.0.37, 3.0.8 PDFBox, 4.0.0
> Reporter: Maruan Sahyoun
> Assignee: Maruan Sahyoun
> Priority: Minor
>
> XFDF export is intended to produce XML 1.0-safe output but has two gaps:
> * C0 control characters (e.g. 0x00–0x08, 0x0B, 0x0C, 0x0E–0x1F) are not
> filtered. They're written verbatim since they aren't one of the five escaped
> special characters. These code points are outside the legal Char production
> in the XML 1.0 spec, and per the XML errata, a numeric character reference to
> an illegal code point (e.g. ) is equally non-well-formed — so there's no fix
> by escaping alone; these characters must be rejected, stripped, or replaced.
> * The method iterates by UTF-16 char, not by Unicode code point.
> Supplementary-plane characters are represented as surrogate pairs in Java;
> each surrogate half is > 0x7e and gets emitted as its own &#N; reference,
> where N falls in 0xD800–0xDFFF. Surrogate values are explicitly excluded from
> the legal Char range, so this also produces non-well-formed XML.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
