document-uri() is not working when using collection()

Question

I'm using saxonche version 12.3.0 installed using pip in Python 3.11.5.

When using collection() in XSLT, document-uri() doesn't seem to be working.

I've created a small test:

Python

import saxonche

with saxonche.PySaxonProcessor(license=False) as proc:
    xsltproc = proc.new_xslt30_processor()
    executable = xsltproc.compile_stylesheet(stylesheet_file="test.xsl")
    content = executable.call_template_returning_string()

    print(f"\ncontent:\n{content}")

XSLT 3.0 (test.xsl)

<xsl:stylesheet version="3.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" expand-text="yes">
    <xsl:output indent="yes" omit-xml-declaration="yes"/>
    <xsl:strip-space elements="*"/>
        
    <xsl:template name="xsl:initial-template">
        <xsl:for-each select="collection('input?select=*.(xml|XML)&amp;content-type=application/xml')">
            <xsl:message>Processing "{document-uri()}"...</xsl:message>
            <xsl:copy-of select="."/>
        </xsl:for-each>
    </xsl:template>
    
</xsl:stylesheet>

Output

Processing ""...
Processing ""...
Processing ""...

content:
<doc>First test XML instance.</doc>
<doc>Second test XML instance.</doc>
<doc>Third test XML instance.</doc>

Expected Output

Processing "file:/C:/test_saxonche_collection/input/test_01.xml"...
Processing "file:/C:/test_saxonche_collection/input/test_02.xml"...
Processing "file:/C:/test_saxonche_collection/input/test_03.xml"...

content:
<doc>First test XML instance.</doc>
<doc>Second test XML instance.</doc>
<doc>Third test XML instance.</doc>

Is this a bug in Saxon?

Answer 1

This is not a bug in Saxon.

The issue is that by default the collection is not stable¹ and documents read using collection() are only added to the document pool if the collection is stable.

From the docs:

Documents read using the collection() function are added to the document pool (and therefore have a document-uri() property) if and only if the collection is stable.

I was able to resolve this two different ways.

The first was by adding the stable=yes query param to my collection() call:

collection('input?select=*.(xml|XML)&content-type=application/xml&stable=yes')

The second was by using base-uri() instead of document-uri():

<xsl:message>Processing "{base-uri()}"...</xsl:message>

I'm going to stick with using base-uri() since there are performance concerns making a collection stable. (Per the docs linked above.)

Additional Info

• https://saxonica.plan.io/issues/6227
• https://saxonica.plan.io/issues/5640
• https://saxonica.plan.io/issues/4837

¹: For more information on what a "stable" collection is, take a look at the the Saxon documentation. I've summarized a couple of key points here:

From Collection catalogs:

If stable="false" is specified, however, the URI is dereferenced directly, and the document is not added to the document pool, which means that a subsequent retrieval of the same document will not return the same node [emphasis added].

From Directories as collections:

Making a collection stable has the effect that the entire result of the collection() function is retained in a cache for the duration of the query or transformation, and any further calls on collection() with the same absolute URI return this saved collection retrieved from this cache.