invalid XML character

[Marco Montel]

I have two applications that should comunicate through an xml file. This
xml will contain a CDATA section with a digital signature.

The problem is that the digital signature is composed of special
character that are nor correctly recognized by the xml parser.

When you try to open the follow file with an xml editor, like jedit, you
will see that the CDATA block is marked with follow error:


"An invalid XML character (Unicode: 0x6) was found in the CDATA section."

This is the xml file:

<?xml version="1.0" encoding="utf-8"?>
<Document>
<DataBlock>
<![CDATA[0‚V_ *†H†÷

 ‚VP0‚VL

1 0+0‚FO *†H†÷


‚F@‚F<Content-Type: multipart/mixed;
boundary="----=_NextPart_717_3066_3832508151693287"
MIME-Version: 1.0
X-Mailer: EldoS MIMEBlackbox Library, version: 2004.04.16
Date: Sat, 4 Dec 2004 19:33:50 +0100
Message-ID: <200412041933500630@69768441>

This is a multi-part message in MIME format.

------=_NextPart_717_3066_3832508151693287
Content-Type: text/html;
charset="utf-8"
Content-Transfer-Encoding: base64
]]>
</DataBlock>
</Document>


How can I handle this special content in an xml file ? Should I use an
other encoding ?


bye marco

 

[Martin Honnen]

Marco Montel wrote:

The problem is that the digital signature is composed of special
character that are nor correctly recognized by the xml parser.

When you try to open the follow file with an xml editor, like jedit, you
will see that the CDATA block is marked with follow error:


"An invalid XML character (Unicode: 0x6) was found in the CDATA section."

In XML 1.0 0x6 is indeed not an allowed character so you need to encode
it, using Base64 encoding for instance.

 

[Marco Montel]

Marco Montel wrote:




In XML 1.0 0x6 is indeed not an allowed character so you need to encode
it, using Base64 encoding for instance.

This is a great idea ! I didn't know that the xml 1.0 specification does
not allow some control caracter. I see that xml 1.1 allow that, but my
parser support only the XML 1.0 specification.

thanks a lot

bye marco

 

[David Carlisle]

This is a great idea ! I didn't know that the xml 1.0 specification does
not allow some control caracter. I see that xml 1.1 allow that, but my
parser support only the XML 1.0 specification.

thanks a lot

bye marco

Even XML 1.1 wouldn't allow the character in a CDATA section.

in XML 1.1 you can encode the character as a character reference, but
you can't have character references in a CDATA section (as & has no
special meaning in a CDATA section)

David

 

[lkrubner]

I'm a total beginner, and I hope you won't mind a stupid question. I
don't understand your point about the ampersand not having a special
meaning in the CDATA section. Can you elaborate on that? Would it have
a special meaning elsewhere? Why would it need a special meaning in the
CDATA section? I thought CDATA sections were there when you wanted the
parser to ignore everything in the secton and just pass the text
through.

 

[Roland Eriksson]

...I don't understand your point about the ampersand not
having a special meaning in the CDATA section ... Would it have
a special meaning elsewhere?

Yes; in a PCDATA section (e.g. inside "normal" marked up content) the
ampersand does have a special meaning, it signals to the parser that
"here is a possible start of an entity, or a numeric character,
reference".

Why would it need a special meaning in the CDATA section?

It does not "need" to have a special meaning and, per definition of
CDATA, it _can_not_ have a special meaning in CDATA.

I thought CDATA sections were there when you wanted the parser to
ignore everything in the section and just pass the text through.

Yep, that's about it.

 

[David Carlisle]

Roland's pretty much answered this already but just to be explict.

the point I was making was that you can't have character 1 in XML 1 as
it's a banned control character, but in XML 1.1 (if your application
supports it, which most don't yet) you can not have the literal control
character but you can have " & # 1 ; " (without the spaces) to refer to
that character which will then appear as the character with unicode
number 1 if you look at the document using and API. DOM, SAX, XPath etc.

However the error message in the original posting refered to a CDATA
section but you can not refer to character 1 in a CDATA section even in
XML 1.1 as in:

"< [ C D A T A & # 1 ; ] ] > "

The & does not start a character reference as it is no longer special,
it just starts itself so the above (again without the spaces) is the
equvalent to

"& a m p ; # 1 ; "


David


(Spaces added as some (bad) newsreaders/mailreaders are confused by
character references and try to interpret them even in a plain text
message)