Encode XML

[Warrick FitzGerald]

Hi All,

When sending an XML \ SOAP request, if the value looks as follows:

<TEST> My Test & Son </TEST>

How should the & be encoded. What is this "TYPE" of encoding called, and
where can I find a spec on which values should and should not be encoded
.... I know the W3C will have something but I can't for the life of me
decode it from their docs.

Thanks
Warrick

[Bob Foster]

Re: Encode XML

"Warrick FitzGerald" <nntp.wfitzgera ld@crtman.com> wrote in message
news:pan.2003.0 8.05.03.37.02.2 38421@crtman.co m...[color=blue]
> When sending an XML \ SOAP request, if the value looks as follows:
>
> <TEST> My Test & Son </TEST>
>
> How should the & be encoded.[/color]

&amp;
[color=blue]
> What is this "TYPE" of encoding called[/color]

Predefined entities. You could also use numeric character references like
&#38; but that requires looking up (uni) codes.
[color=blue]
> and
> where can I find a spec on which values should and should not be encoded[/color]

XML 1.0 spec.
[color=blue]
> ... I know the W3C will have something but I can't for the life of me
> decode it from their docs.[/color]

To boil it down, two characters need to be escaped if they appear in
attribute values or text outside of CDATA sections:

& as &amp; or equivalent
< as &lt; or equivalent

Two other characters are sometimes handy to escape:

" as &quot; or equivalent
' as &apos; or equivalent

The other predefined entity is &gt; but AFAIK you never _need_ to use it.

Bob Foster

[Arto V. Viitanen]

Re: Encode XML

>>>>> "Warrick" == Warrick FitzGerald <nntp.wfitzgera ld@crtman.com> writes:

Warrick> Hi All, When sending an XML \ SOAP request, if the value looks as
Warrick> follows:

Warrick> <TEST> My Test & Son </TEST>

Warrick> How should the & be encoded. What is this "TYPE" of encoding
Warrick> called, and where can I find a spec on which values should and
Warrick> should not be encoded ... I know the W3C will have something but I
Warrick> can't for the life of me decode it from their docs.

I am not sure if it is enought, but ZSI (a Python SOAP package) uses
following encodings for text:

& &amp;
< &lt;
"> &gt;
\015 &#xD;

and following for attributes

& &amp;
< &lt;
" &quot;
\011 &#x9;
\012 &#xA;
\015 &#xD;

(from file compat.py)


--
Arto V. Viitanen av@cs.uta.fi
University of Tampere, Department of Computer and Information Sciences
Tampere, Finland http://www.cs.uta.fi/~av/

[arnold m. slotnik]

Re: Encode XML

"Bob Foster" <bobkfoster@com cast.net> wrote in
news:8mHXa.4845 8$It4.28798@rwc rnsc51.ops.asp. att.net:
[color=blue]
> The other predefined entity is &gt; but AFAIK you never _need_
> to use it.[/color]

Not quite "never". If you have the string ]]> in content, but not as
the end of a CDATA section, you need to escape the >.

--
a. m. slotnik

[Richard Tobin]

Re: Encode XML

In article <Xns93CE7956F8A 35slotnikattnet @204.127.36.1>,
arnold m. slotnik <slotnik@att.ne t> wrote:
[color=blue]
>Not quite "never". If you have the string ]]> in content, but not as
>the end of a CDATA section, you need to escape the >.[/color]

Or one of the ]s.

-- Richard
--
Spam filter: to mail me from a .com/.net site, put my surname in the headers.

FreeBSD rules!