Here is a complete XSLT 1.0 solution:
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:ext="http://exslt.org/common" xmlns:w="w"
exclude-result-prefixes="ext w">
<xsl:output omit-xml-declaration="yes"/>
<xsl:strip-space elements="*"/>
<xsl:template match="w:p">
<xsl:variable name="vrtfPass1">
<p>
<xsl:apply-templates/>
</p>
</xsl:variable>
<xsl:apply-templates mode="pass2"
select="ext:node-set($vrtfPass1)/*"/>
</xsl:template>
<xsl:template match="w:r">
<xsl:variable name="vrtfProps">
<xsl:for-each select="w:rPr/*">
<xsl:sort select="local-name()"/>
<xsl:copy-of select="."/>
</xsl:for-each>
</xsl:variable>
<xsl:call-template name="toHtml">
<xsl:with-param name="pProps" select=
"ext:node-set($vrtfProps)/*"/>
<xsl:with-param name="pText" select="w:t/text()"/>
</xsl:call-template>
</xsl:template>
<xsl:template name="toHtml">
<xsl:param name="pProps"/>
<xsl:param name="pText"/>
<xsl:choose>
<xsl:when test="not($pProps)">
<xsl:copy-of select="$pText"/>
</xsl:when>
<xsl:otherwise>
<xsl:element name="{local-name($pProps[1])}">
<xsl:call-template name="toHtml">
<xsl:with-param name="pProps" select=
"$pProps[position()>1]"/>
<xsl:with-param name="pText" select="$pText"/>
</xsl:call-template>
</xsl:element>
</xsl:otherwise>
</xsl:choose>
</xsl:template>
<xsl:template match="/*" mode="pass2">
<xsl:copy>
<xsl:copy-of select="@*"/>
<xsl:call-template name="processInner">
<xsl:with-param name="pNodes" select="node()"/>
</xsl:call-template>
</xsl:copy>
</xsl:template>
<xsl:template name="processInner">
<xsl:param name="pNodes"/>
<xsl:variable name="pNode1" select="$pNodes[1]"/>
<xsl:if test="$pNode1">
<xsl:choose>
<xsl:when test="not($pNode1/self::*)">
<xsl:copy-of select="$pNode1"/>
<xsl:call-template name="processInner">
<xsl:with-param name="pNodes" select=
"$pNodes[position()>1]"/>
</xsl:call-template>
</xsl:when>
<xsl:otherwise>
<xsl:variable name="vbatchLength">
<xsl:call-template name="getBatchLength">
<xsl:with-param name="pNodes"
select="$pNodes[position()>1]"/>
<xsl:with-param name="pName"
select="name($pNode1)"/>
<xsl:with-param name="pCount" select="1"/>
</xsl:call-template>
</xsl:variable>
<xsl:element name="{name($pNode1)}">
<xsl:copy-of select="@*"/>
<xsl:call-template name="processInner">
<xsl:with-param name="pNodes" select=
"$pNodes[not(position()>$vbatchLength)]
/node()"/>
</xsl:call-template>
</xsl:element>
<xsl:call-template name="processInner">
<xsl:with-param name="pNodes" select=
"$pNodes[position()>$vbatchLength]"/>
</xsl:call-template>
</xsl:otherwise>
</xsl:choose>
</xsl:if>
</xsl:template>
<xsl:template name="getBatchLength">
<xsl:param name="pNodes"/>
<xsl:param name="pName"/>
<xsl:param name="pCount"/>
<xsl:choose>
<xsl:when test=
"not($pNodes) or not($pNodes[1]/self::*)
or not(name($pNodes[1])=$pName)">
<xsl:value-of select="$pCount"/>
</xsl:when>
<xsl:otherwise>
<xsl:call-template name="getBatchLength">
<xsl:with-param name="pNodes" select=
"$pNodes[position()>1]"/>
<xsl:with-param name="pName" select="$pName"/>
<xsl:with-param name="pCount" select="$pCount+1"/>
</xsl:call-template>
</xsl:otherwise>
</xsl:choose>
</xsl:template>
</xsl:stylesheet>
when this transformation is applied to the following XML document (based on the provided, but made more complicated to show how more edge-cases are covered):
<w:p xmlns:w="w">
<w:r>
<w:rPr>
<w:b/>
</w:rPr>
<w:t xml:space="preserve">This is a </w:t>
</w:r>
<w:r>
<w:rPr>
<w:b/>
</w:rPr>
<w:t xml:space="preserve">bold </w:t>
</w:r>
<w:r>
<w:rPr>
<w:b/>
<w:i/>
</w:rPr>
<w:t>with a bit of italic</w:t>
</w:r>
<w:r>
<w:rPr>
<w:b/>
<w:i/>
</w:rPr>
<w:t> and some more italic</w:t>
</w:r>
<w:r>
<w:rPr>
<w:i/>
</w:rPr>
<w:t> and just italic, no-bold</w:t>
</w:r>
<w:r>
<w:rPr>
<w:b/>
</w:rPr>
<w:t xml:space="preserve"></w:t>
</w:r>
<w:r>
<w:rPr>
<w:b/>
</w:rPr>
<w:t>paragr</w:t>
</w:r>
<w:r>
<w:rPr>
<w:b/>
</w:rPr>
<w:t>a</w:t>
</w:r>
<w:r>
<w:rPr>
<w:b/>
</w:rPr>
<w:t>ph</w:t>
</w:r>
<w:r>
<w:t xml:space="preserve"> with some non-bold in it too.</w:t>
</w:r>
</w:p>
the wanted, correct result is produced:
<p><b>This is a bold <i>with a bit of italic and some more italic</i></b><i> and just italic, no-bold</i><b>paragraph</b> with some non-bold in it too.</p>
Explanation:
- This is a two-pass transformation. The first pass is relatively simple and converts the source XML document (in our specific case) to the following:
pass1 result (indented for readability):
<p>
<b>This is a </b>
<b>bold </b>
<b>
<i>with a bit of italic</i>
</b>
<b>
<i> and some more italic</i>
</b>
<i> and just italic, no-bold</i>
<b/>
<b>paragr</b>
<b>a</b>
<b>ph</b> with some non-bold in it too.</p>
.2. The second pass (executed in mode "pass2") merges any batch of consecutive and identically named elements into a single element with that name. It recursively calls-itself on the children of the merged elements -- thus batches at any depth are merged.
.3. Do note: We do not (and cannot) use the axes following-sibling:: or preceding-sibling, because only the nodes (to be merged) at the top level are really siblings. Due to this reason we process all nodes just as a node-set.
.4. This solution is completely generic -- it merges any batch of consecutive identically-named elements at any depth -- and no specific names are hardcoded.
