I found another workaround which works
fine: - the first transfo creates PCDATA into
the <content> tag - the second transfo is a XSl copy transfo
which simply requires CDATA into the <content> tag. This works fine from Python, Still, any comments are welcome with
respect to the behavior of xsltproc and the difference with the same XSL ran
from Python. Thanks JL From:
xslt-bounces gnome org [mailto:xslt-bounces gnome org] On Behalf Of Meunier, Jean-Luc Hi, My XSLT stylesheet produces strange results when it
comes to CDATA. And in Python it’s getting worse. First, instead of encapsulating all the data of my
<content> element into one CDATA, I see many contiguous CDATA: (My stylesheet says at the beginning:
<xsl:output method="xml"
cdata-section-elements="content" indent="yes"/> ) <content><![CDATA[Contract Change
Order]]><![CDATA[ Application and Certificate for
Payment]]><![CDATA[ Progress Schedule and Report]]><![CDATA[ Inventory of Stored Materials]]> </content> It seems that because multiple templates generate the
data of the <content> element, there are many CDATA. BTW, Saxon creates only one CDATA As a workaround, I pass the following additional
XSLT: <xsl:output method="xml"
cdata-section-elements="content" indent="yes"/> <xsl:template
match="@*|node()"> <xsl:copy> <xsl:apply-templates
select="@*|node()"/> </xsl:copy> </xsl:template> Then contiguous CDATA are merged. Ok. **BUT**, my really big problem is that this
workaround doesn’t work in Python. The workaround XSL keeps things
unchanged. Any idea/help would be really appreciated, Thanks JL |