Alok,
On 10/01/10 16:35, Alok Aggarwal wrote:
Hi Dermot,
On Fri, 1 Oct 2010, Dermot McCluskey wrote:
No objection, but for clarity can you spell out exactly what changes
you are proposing to the schema. It appears to me you are suggesting
something like this:
Current Schema (execution.dtd):
<!ELEMENT kwargs (arg+)>
<!ELEMENT arg (#PCDATA)>
<!ATTLIST arg name CDATA #REQUIRED>
Proposed Schema:
<!ELEMENT kwargs ((arg | arglist)+)>
<!ELEMENT arg (#PCDATA)>
<!ATTLIST arg name CDATA #REQUIRED>
<!ELEMENT arglist (argitem+)>
<!ATTLIST arglist name CDATA #REQUIRED>
<!ELEMENT argitem (#PCDATA)>
Here are the changes that I've made to the execution
schema:
<!ELEMENT kwargs (arg*, arglist*)>
This implies that neither <arg> nor <arglist> may be present.
Is that what you intend?
My version requires that, if <kwargs> is present, then at least
one of <arg> or <arglist> must be specified.
<!ELEMENT arg (#PCDATA)>
<!ATTLIST arg name CDATA #REQUIRED>
<!ELEMENT arglist (item*)>
This allows for there to be no <item> sub-elements.
I would have thought there should be at least one <item>
present if <arglist> is used?
- Dermot
<!ATTLIST arglist name CDATA #REQUIRED>
<!ELEMENT item (#PCDATA)>
- this implies that <arg> and <arglist> are not mutually exclusive
and that you
can have either or both, multiple times (but must have at least one).
Is that correct? eg:
<kwargs>
<arg name="foo">bar</arg>
<arglist name="foo_list">
<argitem>bar1</argitem>
<argitem>bar2</argitem>
</arglist>
<arglist name="foo_list2">
<argitem>bar3</argitem>
<argitem>bar4</argitem>
</arglist>
<arg name="foo2">bar2</arg>
</kwargs>
That is correct, that is what I'm intending with
the above changes.
- seeing as how we already have "args", "arg" and "kwargs", and have
generally agreed on "arglist", I suggest "argitem" for the other new
tag.
(there is already a sub-element called "<name>" under
<software_data>, so
we should not have another similarly-named element)
That's probably not a bad idea. What I have right now
is named 'item' which doesn't actually clash with anything
else in the rest of the schema.
- the other possible sub-element of checkpoint, "args" (with an "s")
is unchanged
by all this, correct?
Correct.
Alok
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
