XML Brainteaser

[centizen]

Preface to the is problem: I am coding in a framework that blends jsp
and xml in a heathen manner. Meaning I have to nearly sacrifice a
virgin to get anything to work.

I want to load the character '<' into a java string. Only problem is
that that string is inside a bit of XML. And to compound the problem
the string is sent to a browser as part of an html page. The line
something looks like: (with a '_' in where the '<' needs be)

out.print("_title>Hello, World_/title>");

We were aiming to have the title of the page be "Hello, World". The
problem is that if the blank is filled with '<' the servlet fails to
compile because there is an XML error and nothing is output where this
should be in the html source. If the blank is filled with "&ltf;" then
java interprets this as 5 characters. This outputs to the screen the
phrase:

<title>Hello, World</title>

and the source of the html page looks exactly like the Java string:

&ltf;title>Hello, World&ltf;/title>

Any idea how to get a '<' into a java string inside XML?

(The solution that I am using involved virgin sacrifice, we used some
custom tags from the framework. Somthing along the lines of <c:title
value="Hello, World"> )

[Joe Kesselman]

> Any idea how to get a '<' into a java string inside XML?

The right answer would simply be &lt; or < or &#x3c;, all of which
are ways that XML escapes the < character. If you're accessing your XML
through any of the standard tools/APIs, that does the Right Thing.

If you really insist on needlessly sacrificing virgins -- if you won't
accept anything but a dumb cut-and-paste of the Java directly into the
XML at a textual level -- well, that's a Bad Idea, but that's what CDATA
Sections are for. The XML fragment
<foo><![CDATA[out.print("<title>Hello, World</title>");]]></foo>
is semantically identical to
<foo>out.print("&lt;title>Hello, World&lt;/title>");</foo>

Of course it's up to your system to manage the CDATA section delimiters.
And there are limitations; if your Java code ever contains the ]]>
sequence the workarounds are arguably uglier than individually escaping
the < characters would have been.

custom tags from the framework. Somthing along the lines of <c:title
value="Hello, World"> )

That isn't necessarily inferior, depending on what you're doing.

For a completely different approach, you might want to take a look at
IBM's patent describing the BML Scripting Pages concept, which embeds
scripts written in the Bean Markup Language into documents written in
other XML-based languages. There are some significant advantages to
having both the surrounding markup and the script be written in XML;
among other things, it means one pass through a parser can deal with both.
--
() ASCII Ribbon Campaign | Joe Kesselman
/\ Stamp out HTML e-mail! | System architexture and kinetic poetry

[Simon Brooke]

in message <wa********************@comcast.com>, Joe Kesselman
('k*************@comcast.net') wrote:

Any idea how to get a '<' into a java string inside XML?

The right answer would simply be &lt; or < or &#x3c;, all of which
are ways that XML escapes the < character. If you're accessing your XML
through any of the standard tools/APIs, that does the Right Thing.

If you really insist on needlessly sacrificing virgins -- if you won't
accept anything but a dumb cut-and-paste of the Java directly into the
XML at a textual level -- well, that's a Bad Idea, but that's what
CDATA Sections are for. The XML fragment
<foo><![CDATA[out.print("<title>Hello, World</title>");]]></foo>
is semantically identical to
<foo>out.print("&lt;title>Hello, World&lt;/title>");</foo>

Strictly,

<foo>out.print("&lt;title&gt;Hello, World&lt;/title&gt;");</foo>

The unescaped right angle bracket is just as verboten as the unescaped
left angle bracket.

--
si***@jasmine.org.uk (Simon Brooke) http://www.jasmine.org.uk/~simon/

;; All in all you're just another nick in the ball
-- Think Droid

[Joe Kesselman]

Simon Brooke wrote:

The unescaped right angle bracket is just as verboten as the unescaped
left angle bracket.

Actually, no. XML can easily distinguish whether > is intended to be
part of a tag or part of text, by simply looking at whether it is
currently in the middle of reading a tag. It is *common* to escape text
instances as &gt; so human readers can determine this more easily, but
never necessary... and there is a "for compatibility" requirement that
text instances of ]]> be escaped so there's no risk that an SGML
processor would misread them as ending CDATA Sections.

See http://www.w3.org/TR/2004/REC-xml11-20040204/#syntax

Note also that inside attribute values < does not need to be escaped,
because the fact that it's a quoted string means there's no risk of it
being misinterpreted. But there's no harm in escaping it, and once again
you may want to do so if you expect humans to be reading your XML document.

--
() ASCII Ribbon Campaign | Joe Kesselman
/\ Stamp out HTML e-mail! | System architexture and kinetic poetry

[Richard Tobin]

In article <Fa********************@comcast.com>,
Joe Kesselman <ke************@comcast.net> wrote:

Note also that inside attribute values < does not need to be escaped,
because the fact that it's a quoted string means there's no risk of it
being misinterpreted.

Actually they *must* be escaped, even though there is no possibility
of misinterpretation:

http://www.w3.org/TR/2004/REC-xml11-...#CleanAttrVals

One practical advantage of this is that missing close-quotes on
attribute values are often detected nearer to the location of the
error.

-- Richard

[Joseph Kesselman]

Richard Tobin wrote:

Actually they *must* be escaped, even though there is no possibility
of misinterpretation:

http://www.w3.org/TR/2004/REC-xml11-...#CleanAttrVals

.... True. And the same appears in XML 1.0. OK, that's tie score again.
<grin/>

--
Joe Kesselman / Beware the fury of a patient man. -- John Dryden

[centizen]

Thanks for your help! The application was frozen today for demoing. So
I try these out next week and let you know how they worked.