Juan Hernandez has posted comments on this change.
Change subject: codegen: Only top level elements in root class map
......................................................................
Patch Set 1:
This is my understanding of why we need this dictionary:
1. The XML to Python converter needs it because when parsing an XML document it
needs to know what is the class corresponding to the XML tag of the root
element. By default it assumes that the class is named exactly as the tag. For
example, if we give the parser the following XML document:
<cluster ...>
<cpu ...>
</cpu>
..
</cluster>
It will try to instantiate a "cluster" class, which doesn't exist. So we need
to tell the converter to use the "Cluster" class instead.
This isn't needed for non root elements, as the mapping between tag name and
class name is already included in the generated code. For example, when parsing
the above cluster the Cluster class already knows that the "cpu" tag
corresponds to the "CPU" class:
def buildChildren(self, child_, node, nodeName_, fromsubclass_=False):
if nodeName_ == 'cpu':
obj_ = CPU.factory()
obj_.build(child_)
self.set_cpu(obj_)
...
So for this purpose we only need to include in the dictionary entries for the
XML elements that may be the root of a document. A XML document that contains
any other root element isn't valid, it doesn't match the XML schema.
2. You say that we need this dictionary for the conversion of Python to XML,
and that is true, for the same reasons, but again only for elements that may
appear as roots of documents, as for the others the generated code already
knows how to do it. For example, for the CPU inside a Cluster:
def exportChildren(self, outfile, ...):
...
if self.cpu is not None:
self.cpu.export(outfile, level, namespace_, name_='cpu',
pretty_print=pretty_print)
...
As you can see the generated code already knows that in this context the name
to use is "cpu", not "CPU".
So the only place where we need to instruct it to use different names is for
elements that may appear as root of valid XML documents.
--
To view, visit http://gerrit.ovirt.org/23964
To unsubscribe, visit http://gerrit.ovirt.org/settings
Gerrit-MessageType: comment
Gerrit-Change-Id: Ifcdefe71f3d15ab0da841f35b4b94916d19d950d
Gerrit-PatchSet: 1
Gerrit-Project: ovirt-engine-sdk
Gerrit-Branch: master
Gerrit-Owner: Juan Hernandez <juan.hernan...@redhat.com>
Gerrit-Reviewer: Juan Hernandez <juan.hernan...@redhat.com>
Gerrit-Reviewer: Michael Pasternak <mishka8...@yahoo.com>
Gerrit-Reviewer: Ravi Nori <rn...@redhat.com>
Gerrit-HasComments: No
_______________________________________________
Engine-patches mailing list
Engine-patches@ovirt.org
http://lists.ovirt.org/mailman/listinfo/engine-patches
