Inlining or de-xmlifying xml using XSLT

S

Simon Brooke

I'm trying to write an XSL template to generate Google's data input
format for blogger, which is documented here: http://goo.gl/tydYA

As you can see, it has the particularly delightful property that the
content of the 'content' element actually /is/ markup, but has been
inlined or entified so that it appears as raw text. Yes, I know this is
bizarre and ugly, but I don't control it - Google do. And as there's no
documentation on how to import normally-well-formed XML, this is what I
need to generate.

I've been trying to write a template which does this awful mangling,
and what I've come up with seems to work:

<xsl:template name="mangle-xml">
<xsl:param name="content"/>
<xsl:choose>
<xsl:when test="$content/*">
&lt;<xsl:value-of select="local-name()"/>
<xsl:for-each select="$content/@*">
<xsl:value-of select="concat(' ', local-name(),
'=&quot;', ., '&quot;')"/>
</xsl:for-each>&gt;
<xsl:for-each select="$content/node()">
<xsl:call-template name="mangle-xml">
<xsl:with-param name="content" select="."/>
</xsl:call-template>
</xsl:for-each>
&lt;/<xsl:value-of select="local-name()"/>&gt;
</xsl:when>
<xsl:eek:therwise>?text?
<xsl:value-of select="$content"/>
</xsl:eek:therwise>
</xsl:choose>
</xsl:template>

This mangles the XML correctly(!), but it's pretty ugly, dark and
mysterious. Is there a cleaner way of doing this?

--
http://www.journeyman.cc/~simon/ :: PGP public key on home page

;; USER ERROR: replace user and press any key to continue


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAk00WPoACgkQPj28Ek2lI8X4rwCgsA+l1yGGMV2pXho124NuDzBP
0yMAn23M0QVt2FDVn2EMGcSY179rwyqW
=0j07
-----END PGP SIGNATURE-----
 
M

Mayeul

I'm trying to write an XSL template to generate Google's data input
format for blogger, which is documented here: http://goo.gl/tydYA

As you can see, it has the particularly delightful property that the
content of the 'content' element actually /is/ markup, but has been
inlined or entified so that it appears as raw text. Yes, I know this is
bizarre and ugly, but I don't control it - Google do. And as there's no
documentation on how to import normally-well-formed XML, this is what I
need to generate.

I've been trying to write a template which does this awful mangling,
and what I've come up with seems to work:

<xsl:template name="mangle-xml">
<xsl:param name="content"/>
<xsl:choose>
<xsl:when test="$content/*">
&lt;<xsl:value-of select="local-name()"/>
<xsl:for-each select="$content/@*">
<xsl:value-of select="concat(' ', local-name(),
'=&quot;', ., '&quot;')"/>
</xsl:for-each>&gt;
<xsl:for-each select="$content/node()">
<xsl:call-template name="mangle-xml">
<xsl:with-param name="content" select="."/>
</xsl:call-template>
</xsl:for-each>
&lt;/<xsl:value-of select="local-name()"/>&gt;
</xsl:when>
<xsl:eek:therwise>?text?
<xsl:value-of select="$content"/>
</xsl:eek:therwise>
</xsl:choose>
</xsl:template>

This mangles the XML correctly(!), but it's pretty ugly, dark and
mysterious. Is there a cleaner way of doing this?
Click to expand...

Ugh. Personally I'd renounce making it in one integrated XSL
transformation. I'd make one for the <content>, another one for the
whole document, and feed the result from the first as a string parameter
to the second.

In idea, this is probably what was meant.
 
M

Martin Honnen

Simon said:
I'm trying to write an XSL template to generate Google's data input
format for blogger, which is documented here: http://goo.gl/tydYA

As you can see, it has the particularly delightful property that the
content of the 'content' element actually /is/ markup, but has been
inlined or entified so that it appears as raw text. Yes, I know this is
bizarre and ugly, but I don't control it - Google do. And as there's no
documentation on how to import normally-well-formed XML, this is what I
need to generate.

I've been trying to write a template which does this awful mangling,
and what I've come up with seems to work:

<xsl:template name="mangle-xml">
<xsl:param name="content"/>
<xsl:choose>
<xsl:when test="$content/*">
&lt;<xsl:value-of select="local-name()"/>
<xsl:for-each select="$content/@*">
<xsl:value-of select="concat(' ', local-name(),
'=&quot;', ., '&quot;')"/>
</xsl:for-each>&gt;
<xsl:for-each select="$content/node()">
<xsl:call-template name="mangle-xml">
<xsl:with-param name="content" select="."/>
</xsl:call-template>
</xsl:for-each>
&lt;/<xsl:value-of select="local-name()"/>&gt;
</xsl:when>
<xsl:eek:therwise>?text?
<xsl:value-of select="$content"/>
</xsl:eek:therwise>
</xsl:choose>
</xsl:template>

This mangles the XML correctly(!), but it's pretty ugly, dark and
mysterious. Is there a cleaner way of doing this?
Click to expand...

Some XSLT processors supply extension functions to serialize nodes as
XML, for instance Saxon has
http://www.saxonica.com/documentation/extensions/functions/serialize.xml. I
would use that if available. If it needs to be done in XSLT itself then
I would probably do it with templates in a particular mode, not with a
single named template. And sophisticated approaches like
http://lenzconsulting.com/xml-to-string/xml-to-string.xsl to deal with
more complex problems like namespaces exist. But I am not sure you need
that, the markup is probably escaped as it is supposed to be some
text/html tag soup and not clean XML.
 
S

Simon Brooke

But I am not sure you need
that, the markup is probably escaped as it is supposed to be some
text/html tag soup and not clean XML.
Click to expand...

Yes, probably the only reason they adopted this horrible kluge in the
first place was to deal with tag soup.

[fx: a look of ineffable disgust and derision plays across his features]

--
http://www.journeyman.cc/~simon/ :: PGP public key on home page

;; USER ERROR: replace user and press any key to continue


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAk00lbQACgkQPj28Ek2lI8V2pwCePqu7t/RpOf2R928SCWvOUXdf
yDYAmQExmFeJIRxccn9B3YbRty0rqrNs
=OboJ
-----END PGP SIGNATURE-----
 
J

Joe Kesselman

Yes, probably the only reason they adopted this horrible kluge in the
first place was to deal with tag soup.
Click to expand...

.... It would accomplish that, I suppose. I've also seen this sort of
thing done simply because people didn't understand that a single tree --
especially if namespaced -- is actually _easier_ to handle than
reparsing the content.

Second the suggestion that this be generated via a Mode. Much of it
could be handled by rewriting the Identity Transform to output text
rather than XML. The really hideous thing is handling the case of nexted
<[![CDATA]]> sections; detecting and handling those requires string
manipulation on the contained text, which is not XSLT's strongest point.


--
Joe Kesselman,
http://www.love-song-productions.com/people/keshlam/index.html

{} ASCII Ribbon Campaign | "may'ron DaroQbe'chugh vaj bIrIQbej" --
/\ Stamp out HTML mail! | "Put down the squeezebox & nobody gets hurt."
 
P

Peter Flynn

But I am not sure you need
that, the markup is probably escaped as it is supposed to be some
text/html tag soup and not clean XML.
Click to expand...

Yes, probably the only reason they adopted this horrible kluge in the
first place was to deal with tag soup.

[fx: a look of ineffable disgust and derision plays across his features]
Click to expand...

Cheer up, the recent announcements about HTML5 will only make things
worse :)

///Peter
 
J

Joe Kesselman

Cheer up, the recent announcements about HTML5 will only make things
worse :)
Click to expand...

Yeah, it seems the term "HTML5" has been hijacked from its original
intent, which was to bring HTML back to being well-formed, make it
XML-based rather than SGML-based, and make it extendable via namespaces.
Sigh. That's the web for ya -- never mind what would be useful and
robust, just let me make it pretty.

--
Joe Kesselman,
http://www.love-song-productions.com/people/keshlam/index.html

{} ASCII Ribbon Campaign | "may'ron DaroQbe'chugh vaj bIrIQbej" --
/\ Stamp out HTML mail! | "Put down the squeezebox & nobody gets hurt."
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Similar Threads

Problem in xml transformation using xslt 0
remove specail character from xml using xslt 0
XSL node as string 1
newbie xsl / xslt question 2
XSLT Question 1
xslt and namespaces 2
XSLT question 4
XSLT filter nodes containing attributes with known values 0

Members online

Total: 332 (members: 3, guests: 329)
Robots: 205

Forum statistics

Threads
473,968
Messages
2,570,150
Members
46,696
Latest member
BarbraOLog

Latest Threads

Top