The reason this works is that saxon:get-pseudo-attribute() expands entity and
character references found in the PI content, so the resulting string contains
angle brackets in place of "& lt ; " and "& gt ;".

Michael Kay
Saxonica

> On 15 Jan 2021, at 15:36, Eliot Kimber ekimber@xxxxxxxxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Yes, I'm using the Saxon extension function to get the value of the @content
pseudo-attribute.
>
> What I didn't realize was that parse-xml and parse-xml-fragment() will
handle the string as it comes, so this works:
>
> <xsl:variable name="rawMarkup"
select="saxon:get-pseudo-attribute('content')/>
> <xsl:variable name="deletedContent" as="element()">
>  <deletedContent xmlns="">
>   <xsl:sequence select="parse-xml-fragment($rawMarkup)/>
>  </deletedContent>
> </xsl:variable>
>
> So it's all good.
>
> Cheers,
>
> E.
>
> --
> Eliot Kimber
> http://contrext.com
>
>
> o;?On 1/15/21, 9:22 AM, "Martin Honnen martin.honnen@xxxxxx"
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
>    Am 15.01.2021 um 15:58 schrieb Eliot Kimber ekimber@xxxxxxxxxxxx:
>> I researched this question in this list but I didn't find a recent response
so I'm asking it again.
>>
>> My specific requirement is to take the value of an Oxygen deleted content
processing instruction and parse it back into elements in the context of an
XSLT 3 transform.
>>
>> The PI is:
>>
>> <?oxy_delete author="ekimber" timestamp="20210115T085216-0600"
content="&lt;body&gt;&lt;p&gt;&lt;ul
id=&quot;ul_c5k_mhh_34b&quot;&gt;&lt;li&gt;li one&lt;/li&gt;&lt;li&gt;li
two&lt;/li&gt;&lt;/ul&gt;&lt;/p&gt;&lt;/body&gt;"?>
>>
>> What I'm failing to work out is how to convert the value of the @content
pseudo-attribute back into a proper XML string that I can pass to
parse-xml().
>
>    I think Saxon PE and EE have an extension function to access the pseudo
>    attributes of a processing instruction value.
>
>    So inside of oXygen with Saxon PE or EE this should be easz.