I read this over. I agree that current language is not symmetric for unparsing with parsing, and in this case it should be symmetric. If the parser is going to stop looking for more instances when maxOccurs is reached, then the unparser should stop output of those instances when maxOccurs is reached.

There is a fair amount of complexity to assigning the proper schema component for use during unparsing, given an infoset, but counting numbers of occurrences is certainly not the most complex such thing, and parsing has to do this counting. So unparsing should as well.

Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | www.tresys.com
Please note: Contributions to the DFDL Workgroup's email discussions are subject to the OGF Intellectual Property Policy


On Mon, Jan 19, 2015 at 1:01 PM, Steve Hanson <smh@uk.ibm.com> wrote:
Please have a position on the below proposal from IBM for this week's WG call.
Regards
 
Steve Hanson
Architect, IBM DFDL
Co-Chair, OGF DFDL Working Group
IBM SWG, Hursley, UK
smh@uk.ibm.com
tel:+44-1962-815848



From:        Steve Hanson/UK/IBM
To:        DFDL-WG <dfdl-wg@ogf.org>
Date:        12/01/2015 16:56
Subject:        Unparser maxOccurs issue


Possible change to spec needed., where it describes what happens when maxOccurs is exceeded during unparsing for occursCountKind 'fixed' and 'implicit' (and by implication scalar elements).

It currently says it is a processing error. I think it is better to say that the unparser moves on when maxOccurs is reached. This makes the behaviour analogous to parsing, when it does not try to parse beyond maxOccurs and moves on.  The current unparser wording is based on the assumption that any next occurrence of the element in the infoset must be an error, but this is not true - the next occurrence could be an occurrence of a same named element later in the schema.

An obvious example is:

<xs:element name="data" minOccurs="2"maxOccurs="2" dfdl:occursCountKind="fixed" ... />
<xs:element name="stuff" minOccurs="0" dfdl:occursCountKind="implicit" ... />
<xs:element name="data" maxOccurs="2" maxOccurs="2" dfdl:occursCountKind="fixed" ... />

with an infoset where 'stuff' is missing:

message_data
  data - xx1
  data - xx2
  data - yy1
  data - yy2

A more interesting example is this, taken from MIL-STD-2045 schema (my bold comments added):

<xsd:sequence dfdl:separator="">
<!-- Element Value1 -->
        <xsd:choice>
          <xsd:sequence dfdl:separator="">
            <xsd:group ref="FPI_true"/>
            <xsd:element dfdl:length="4" dfdl:lengthKind="explicit" name="Value1" type="xsd:string"/>
          </xsd:sequence>
          <xsd:sequence dfdl:separator="">
            <xsd:group ref="FPI_false"/>
          </xsd:sequence>
        </xsd:choice>
<!-- Element Value2 -->
        <xsd:choice>
          <xsd:sequence dfdl:separator="">
            <xsd:group ref="FPI_true"/>
            <xsd:element dfdl:length="4" dfdl:lengthKind="explicit" name="Value2" type="xsd:string"/>
          </xsd:sequence>
          <xsd:sequence dfdl:separator="">
            <xsd:group ref="FPI_false"/>
          </xsd:sequence>
        </xsd:choice>
<!-- Element Value3 -->
        <xsd:choice>
          <xsd:sequence dfdl:separator="">
            <xsd:group ref="FPI_true"/>
            <xsd:element dfdl:length="4" dfdl:lengthKind="explicit" name="Value3" type="xsd:string"/>
          </xsd:sequence>
          <xsd:sequence dfdl:separator="">
            <xsd:group ref="FPI_false"/>
          </xsd:sequence>
        </xsd:choice>
</xsd:sequence>

...where the FPI_true and FPI_false elements are defined in their own global groups.

  <xsd:group name="FPI_true">
    <xsd:sequence dfdl:separator="">
      <xsd:element default="true" dfdl:length="1" dfdl:lengthKind="explicit" dfdl:textBooleanFalseRep="0" dfdl:textBooleanTrueRep="1" name="FPI_true" type="xsd:boolean">
        <xsd:annotation>
          <xsd:appinfo source="http://www.ogf.org/dfdl/">
            <dfdl:discriminator>{. eq fn:true()}</dfdl:discriminator>
          </xsd:appinfo>
        </xsd:annotation>
      </xsd:element>
    </xsd:sequence>
  </xsd:group>

  <xsd:group name="FPI_false">
    <xsd:sequence dfdl:separator="">
      <xsd:element default="false" dfdl:length="1" dfdl:lengthKind="explicit" dfdl:textBooleanFalseRep="0" dfdl:textBooleanTrueRep="1" name="FPI_false" type="xsd:boolean">
      </xsd:element>
    </xsd:sequence>
  </xsd:group>

If the infoset looked like the following an error would be given, whereas it is valid because the second FPI_false is for Value2:

message_prefixedOccurs
  FPI_false - false
  FPI_false - false
  FPI_true - true
  Value3 - 9999

Regards
 
Steve Hanson
Architect, IBM DFDL
Co-Chair, OGF DFDL Working Group
IBM SWG, Hursley, UK
smh@uk.ibm.com
tel:+44-1962-815848
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

--
  dfdl-wg mailing list
  dfdl-wg@ogf.org
  https://www.ogf.org/mailman/listinfo/dfdl-wg