[DFDL-WG] Discriminator example, alternative presentation
Simon Parker
simon.parker at polarlake.com
Fri Feb 15 07:33:49 CST 2008
Although I don't yet grasp all the nuances of 'assert' and
'discriminator', your point is valid.
This potential capability is too useful to lose, and I feel we should
make the specification change you suggest.
Thanks for the thoughtful feedback,
Simon
________________________________
From: Steve Hanson [mailto:smh at uk.ibm.com]
Sent: 15 February 2008 11:47
To: Simon Parker
Cc: dfdl-wg at ogf.org; dfdl-wg-bounces at ogf.org
Subject: Re: [DFDL-WG] Discriminator example, alternative
presentation
Hi Simon
I see where you are coming from, but what you propose does not
agree with the spec (section 5.2.8). This states that use of 'fixed' is
the equivalent to dfdl:assert, not dfdl:discriminator. The difference
between the two is that encountering a dfdl:discriminator that evaluates
to 'true' in a choice branch switches off speculative parsing and a
subsequent processing error in that branch of the choice will not cause
backtracking. Using dfdl:assert does not provide that same semantic as a
subsequent processing error in that branch of the choice will cause
backtracking . I think all we need to do is change 5.2.8 to say that
use of 'fixed' is equivalent to dfdl:discriminator instead.
This semantic difference is the reason why we have both
dfdl:assert and dfdl:discriminator. As I pointed out on the call, it's
been lost from sections 7.3 and 7.4.
Here's 5.2.8 fyi.
The 'fixed' attribute is used:
to constrain the logical value of a required element while
parsing.
to provide the logical value of a required element when
unparsing when the DFDL information set does not have a value for the
element.
Note that a 'fixed' attribute can cause parsing to backtrack and
try other alternatives (see footnote). When data is encountered and it
does not match what is specified on the 'fixed' attribute, then it is a
parse error. That is, the fixed value is used for parsing, not only for
validation checking.
Footnote:
The semantics of fixed='$$$' is as if it were translated into an
assertion and a dfdl:outputValueCalc property:
<element name="x" type="string" dfdl:outputValueCalc='$$$' >
<annotation><appinfo>
<dfdl:assert>$. = '$$$'</dfdl:assert>
</appinfo></annotation>
</element>
Assertion failure causes a processing error. The outputValueCalc
fills in the value when writing out data.
Regards, Steve
Steve Hanson
WebSphere Message Brokers
Hursley, UK
Internet: smh at uk.ibm.com
Phone (+44)/(0) 1962-815848
"Simon Parker" <simon.parker at polarlake.com>
Sent by: dfdl-wg-bounces at ogf.org
13/02/2008 17:43
To
<dfdl-wg at ogf.org>
cc
Subject
[DFDL-WG] Discriminator example, alternative presentation
Good evening.
As discussed at the conference, here's a more succinct way to
express this simple requirement.
Perhaps discriminator's motivating example should be more
complex.
Comments and corrections welcome,
Simon
Simon Parker
Software Consultant
PolarLake
80 Harcourt Street
Dublin 2, Ireland
Phone +353 1 449-1075
Fax +353 1 449-1011
Web http://www.polarlake.com/ <http://www.polarlake.com/>
The information transmitted in this email is intended for the
addressee only and may contain confidential and/or privileged material.
Any review, retransmission, dissemination, reliance upon or other use of
this information by persons or entities other than the addressee is
prohibited. If you think, for any reason, that this message may have
been addressed to you in error, we would ask you to notify the sender
immediately by return email and delete the material. PolarLake Limited |
Registered in Dublin, Ireland | Number 357324 | Registered office as
above.
--
dfdl-wg mailing list
dfdl-wg at ogf.org
http://www.ogf.org/mailman/listinfo/dfdl-wg
________________________________
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/20080215/1b559308/attachment.html
More information about the dfdl-wg
mailing list