[DFDL-WG] Tracker 299 and clarifications needed for dfdl:initiatedContent="yes" behaviors

[Steve Hanson]

Hi Mike

Yes I think your interpretation of the spec is correct.  If you removed 
dfdl:initiatedContent and used a discriminator on foo instead then you 
have an analogous situation.

The awkward scenario is 4, when the points of uncertainty vary depending 
on the occurrences.

Possibly writing Scenario 4 as follows makes things clearer:

<xs:choice dfdl:initiatedContent="yes">
   ....
   <xs:element name="foo" minOccurs="5" maxOccurs="5" 
dfdl:initiator="foo:" dfdl:occursCountKind="fixed".../>
   <xs:element name="foo" minOccurs="0" dfdl:initiator="foo:" 
dfdl:occursCountKind="implicit".../>
   ...
</xs:choice>

When I get a chance I will create a schema for your scenarios, and see 
what IBM DFDL does and what it outputs in its trace. 

Regards
 
Steve Hanson
IBM Hybrid Integration, Hursley, UK
Architect, IBM DFDL
Co-Chair, OGF DFDL Working Group
smh at uk.ibm.com
tel:+44-1962-815848
mob:+44-7717-378890
Note: I work Tuesday to Friday 



From:   Mike Beckerle <mbeckerle.dfdl at gmail.com>
To:     DFDL-WG <dfdl-wg at ogf.org>
Date:   14/06/2019 22:57
Subject:        [DFDL-WG] Tracker 299 and clarifications needed for 
dfdl:initiatedContent="yes" behaviors
Sent by:        "dfdl-wg" <dfdl-wg-bounces at ogf.org>



Clarifications around initiatedContent="yes", choices,  and array elements

DFDL spec says in Section 15 that the root of a choice branch must not be 
optional, meaning minOccurs cannot be 0.
This was questioned and resulted in tracker 299. The last comment there 
suggests a need to revisit.

Here's the link: https://redmine.ogf.org/issues/299

The additional clarification I seek is whether I have properly understood 
how dfdl:initiatedContent works in scenarios involving a combination of 
choices with 
array elements.

Consider this schema fragment:

<!-- Scenario 1 -->
<xs:choice dfdl:initiatedContent="yes">
   ....
   <xs:element name="foo" minOccurs="5" dfdl:initiator="foo:" 
dfdl:occursCountKind="parsed".../>
   ...
</xs:choice>
When parsing this choice, if you encounter "foo:" in the data stream, that 
discriminates two things: the choice itself, and the first occurrence of 
the array foo. Both that choice branch (aka the array), and the first 
occurrence of the array are "known to exist" per section 9.3.1.1. 
Subsequent occurrences of additional foo elements with their initiators 
"foo:" discriminate only the array occurrences. The minOccurs="5" is 
ignored (except for validation)

<!-- Scenario 2 -->
<xs:choice dfdl:initiatedContent="yes">
   ....
   <xs:element name="foo" minOccurs="5" dfdl:initiator="foo:" 
dfdl:occursCountKind="fixed".../>
   ...
</xs:choice>
When parsing this choice, if you encounter "foo:" in the data stream, that 
discriminates the choice, but only for the first occurrence of element 
foo. There is no point of uncertainty for the array. Subsequent 
occurrences of element foo do not discriminate anything. There must be 
exactly 5 occurrences of element foo with initiators. 

<!-- Scenario 3 -->
<xs:choice dfdl:initiatedContent="yes">
   ....
   <xs:element name="foo" minOccurs="5" dfdl:initiator="foo:" 
dfdl:occursCountKind="expression".../>
   ...
</xs:choice>
When parsing this choice, if you encounter "foo:" that discriminates the 
choice, but there are no points of uncertainty for the array. MinOccurs is 
ignored (except for validation).
Subsequent occurrences of element "foo" do not discriminate anything.

<!-- Scenario 4 -->
<xs:choice dfdl:initiatedContent="yes">
   ....
   <xs:element name="foo" minOccurs="5" dfdl:initiator="foo:" 
dfdl:occursCountKind="implicit".../>
   ...
</xs:choice>
When parsing this choice, if you encounter "foo:" that discriminates the 
choice, but there is no point of uncertainty for the array until 5 
occurrences have been parsed.
Once the occurs index is 6 then if you encounter "foo:" that discriminates 
the occurrence of the array element foo, but the choice is already 
decided. 

I raise these scenarios because they show that there is quite a bit of 
complexity associated with dfdl:initiatedContent if viewed as a mechanism 
for setting discriminators. 
Sometimes more than one discriminator must be set, other times only one, 
and sometimes it depends on the array index. 

So the question is: Is my understanding of the above scenarios correct?


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
--
  dfdl-wg mailing list
  dfdl-wg at ogf.org
  
https://urldefense.proofpoint.com/v2/url?u=https-3A__www.ogf.org_mailman_listinfo_dfdl-2Dwg&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=AJa9ThEymJXYnOqu84mJuw&m=c8OLhOSmTRRIX8nVxttaTdItRxbGNaUo-2-F4C6IKkE&s=ppz2ALYc76pJ5da777Me47TDy4GCTDyHHl3QNWaYneA&e= 


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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ogf.org/pipermail/dfdl-wg/attachments/20190620/ca9d68c0/attachment.html>