deltacloud-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tong Li <liton...@us.ibm.com>
Subject Re: Model layer for CIMI
Date Thu, 10 Nov 2011 13:36:19 GMT

I do not think DMTF cimi spec strongly suggested the sequence of the
elements. At least it did not say in the spec. although most of the samples
do show a patten of the elements. When I used XmlSimple (), it actually
keeps the right order with the xml it reads in. The difficulty is when you
have a hash (produced by XmlSimple from an XML doc), you most likely will
add or remove things off of the hash, that is where it sequence really get
messed up, I do not consider it has anything to do with the XmlSimple, it
is just when hash get serialized to xml or json, probably the serializer
does not know how to deal with the sequence since the hash certainly does
not have the sequence information of its members.

just my 2cents.

Tong Li
Emerging Technologies & Standards
Building 501/B205
litong01@us.ibm.com



From:	lutter@redhat.com
To:	deltacloud-dev@incubator.apache.org
Date:	11/09/2011 04:19 PM
Subject:	Model layer for CIMI




This is a model layer for CIMI that does both serialization and
deserialization of JSON and XML based on a simple metadata description of
the CIMI model expressed in a custom DSL. In particular, we won't need HAML
templates to format objects, and I am reasonably sure that the metadata
makes sure we serialize in a predictable manner to XML.

Patch 1/2 contains the basics, including the DSL, 2/2 shows how this can be
used for a specific class.

There's a few convenience additions that should be made to the DSL, in
particular making it possible to reuse nested structures that are used in
several places, something like

  array :operations, :class => Operation

but we can add that later.

One small fly in the ointment is that the use of XmlSimple causes child
elements in XML to be rendered in an unpredictable order; the CIMI standard
doesn't say anything about reordering elements, though with a strict
interpretation of XML, the rendered XML does not conform to the XML schema
implicitly defined in the CIMI standard.

David


Mime
  • Unnamed multipart/related (inline, None, 0 bytes)
View raw message