[Bug 29162] New: [xslt30] Parameter-document: static or dynamic?


            Bug ID: 29162
           Summary: [xslt30] Parameter-document: static or dynamic?
           Product: XPath / XQuery / XSLT
           Version: Last Call drafts
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: XSLT 3.0
          Assignee: mike@saxonica.com
          Reporter: mike@saxonica.com
        QA Contact: public-qt-comments@w3.org
  Target Milestone: ---

xsl:output and xsl:result-document have an attribute parameter-document which
is the URI of a file containing serialization parameters.

In the case of xsl:result-document, the attribute is an AVT, so there is a
clear implication that we don't know the URI until run-time and therefore we
don't read the parameter document until run-time.

For xsl:output the attribute is known statically, but we don't give any
indication of whether we expect the parameter document to be read at compile
time or at run-time. If we leave this question open, we create a potential
interoperability problem, especially in the context of distributable packages
(this also affects the location of the parameter file).

For some reason I don't fully understand, we always resisted making the
serialisation attributes on xsl:output dynamic (AVTs). Is the introduction of
parameter documents supposed to circumvent that restriction?

In keeping with the current status of serialization properties, I'm inclined to
say that we add the rules:

* For xsl:output, processors should read the supplied parameter file during
static analysis of the stylesheet; subsequent changes to the content of the
file should have no effect.

* For xsl:result-document, processors should read the supplied parameter
document during the evaluation of the xsl:result-document instruction. If the
same parameter file (that is, the same absolute URI) is used in the course of
repeated evaluation of one or more xsl:result-document instructions, it is
implementation-dependent whether any changes to the content of the file after
the first use have any effect.

(Alternative resolution: scrap the feature. We currently have no tests, it
delivers very little benefit, and it's complexity we can do without.)

You are receiving this mail because:
You are the QA Contact for the bug.

Received on Wednesday, 30 September 2015 08:04:11 UTC