Stripping empty tags from Plone content with Diazo - tinymce

Question

I have a Plone site, themed with plone.app.theming. The problem I have is that the design is quite strict and doesn't assume any empty <p> elements or any other nonsense TinyMCE outputs. Such elements break the intended design. So I want to strip the empty elements from the content. I have tried inline xslt (that, according to http://diazo.org/advanced.html#inline-xsl-directives should be supported) like:
<xsl:output omit-xml-declaration="yes" indent="yes"/>
<xsl:strip-space elements="*"/>
<xsl:template match="node()|#*">
<xsl:copy>
<xsl:apply-templates select="node()|#*"/>
</xsl:copy>
</xsl:template>
<xsl:template match="*[not(*) and not(text()[normalize-space()])]"/>
But it didn't do the trick. In fact it made something weird. The empty p tags that I wanted to get rid of stayed intact but some other elements like
<img src="../++theme++jarn.com/whatever.png" />
turned into
with the image being striped out. Replacing match="*[… in the second template to match="p[… didn't strip out the images, but those nasty <p> were still in the output.
Any hints on how to get rid of the empty elements using Diazo rules?
UPDATE January 31, 2012
Here is the content from which I need the empty p tags to be stripped off:
<div id="parent-fieldname-text">
<p></p>
<p> </p>
<p> </p>
<p><section id="what-we-do">
<p class="visualClear summary">Not empty Paragraph</p>
<ul class="thumbsList">
<li> <img src="../++theme++jarn.com/whatever.png" /></li>
<li> <img src="../++theme++jarn.com/whatever.png" /></li>
</ul>
</section></p>
</div>
The Diazo transformation rules:
<?xml version="1.0" encoding="UTF-8"?>
<rules
xmlns="http://namespaces.plone.org/diazo"
xmlns:css="http://namespaces.plone.org/diazo/css"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output omit-xml-declaration="yes" indent="yes"/>
<xsl:strip-space elements="*"/>
<xsl:template match="node()|#*">
<xsl:copy>
<xsl:apply-templates select="node()|#*"/>
</xsl:copy>
</xsl:template>
<xsl:template match="p[not(*) and not(normalize-space())]"/>
<!-- The default theme, used for standard Plone web pages -->
<theme href="index.html" css:if-content="#visual-portal-wrapper" />
<replace css:theme-children="div.contentWrapper" css:content-children="#content" />
</rules>
The output I get after applying the transformations to the Plone site is absolutely identical to the input while I would expect to get those 3 empty <p> tags after opening <div> to go away.
If I change the second template to match all elements like match="*… then the images get stripped out, but the empty <p> tags are still there.

Answer 1

Just have this:
<xsl:template match="p[not(*) and not(normalize-space())]"/>
A complete transformation:
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output omit-xml-declaration="yes" indent="yes"/>
<xsl:strip-space elements="*"/>
<xsl:template match="node()|#*">
<xsl:copy>
<xsl:apply-templates select="node()|#*"/>
</xsl:copy>
</xsl:template>
<xsl:template match="p[not(*) and not(normalize-space())]"/>
</xsl:stylesheet>
when this transformation is applied on this XML document:
<div>
<p/>
<p> </p>
<p><img src="..."/></p>
<img src="..."/>
</div>
the wanted, correct result is produced:
<div>
<p>
<img src="..."/>
</p>
<img src="..."/>
</div>

Answer 2

Works for me. I've added an example of using Dimitre's xpath in a drop content rule at https://github.com/plone/diazo/commit/94ddff7117d25d3a8a89457eeb272b5500ec21c5 but it also works as the equivalent xsl:template. The example is pared down to the basics but it works using the complete example content in the question too.