You have found a bug but it is not related to *your* XML declaration,
which is, of course, correct.
All the following documents work fine:
---
<?xml version="1.0"?>
<article2>
<title>Cannot be opened by XXE</title>
<para>
For whatever reason, the xml declaration in this file
causes problems for XXE (or XXE's parser).
</para>
</article2>
---
---
<?xml version="1.0"?>
<!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd";>
<article>
<title>Cannot be opened by XXE</title>
<para>
For whatever reason, the xml declaration in this file
causes problems for XXE (or XXE's parser).
</para>
</article>
---
---
<?xml version="1.0" encoding="US-ASCII" ?>
<article>
<title>Cannot be opened by XXE</title>
<para>
For whatever reason, the xml declaration in this file
causes problems for XXE (or XXE's parser).
</para>
</article>
---
---
<article>
<title>Cannot be opened by XXE</title>
<para>
For whatever reason, the xml declaration in this file
causes problems for XXE (or XXE's parser).
</para>
</article>
---
Here's what happens:
* The DocBook configuration is used to recognize your article as being
DocBook.
* Because your article has no <!DOCTYPE>, XXE creates a temporary master
document with a DocBook <!DOCTYPE>, including your article as an
external entity. Then XXE loads this temporary master document.
* The problem is that, now, your article is treated by the parser as an
external entity.
* ``<?xml version="1.0"?>'' is an incorrect *TextDecl* for an external
entity (an encoding must be specified in this case).
---
[77] TextDecl ::= '<?xml' VersionInfo? EncodingDecl S? '?>'
---
This bug will be fixed in next version. The workaround is to specify an
encoding in your XML declaration.
John L. Clark wrote:
> The attached file, a very simple DocBook file, is well-formed as far as
> I can tell, but XXE refuses to open it, giving the error:
>
> parse error in "<file/>", line 1, column 0: invalid XML declaration
>
> The XML declaration is:
>
> <?xml version="1.0"?>
>
> which looks like it matches the production for XMLdecl given in the XML
> 1.0 Spec[0]. Also, this file is parsed by `xmllint` with no errors.
> This seems to me either a bug in the XXE suite or a problem with the
> file. If it is the latter, what am I doing wrong? If the former,
> consider this a bug report.
>
>
> [0] http://www.w3.org/TR/REC-xml/#NT-XMLDecl
>
>
> ------------------------------------------------------------------------
>
> <?xml version="1.0"?>
> <article>
> <title>Cannot be opened by XXE</title>
>
> <para>
> For whatever reason, the xml declaration in this file
> causes problems for XXE (or XXE's parser).
> </para>
> </article>
>
>
> ------------------------------------------------------------------------
