- From: Norman Walsh <ndw@nwalsh.com>
- Date: Sun, 28 Jun 2009 17:06:11 -0400
- To: XProc Dev <xproc-dev@w3.org>
- Message-ID: <m2ab3sxe4c.fsf@nwalsh.com>
"Costello, Roger L." <costello@mitre.org> writes:
> This pipeline outputs the content of the parameter input port:
>
> <p:declare-step xmlns:p="http://www.w3.org/ns/xproc"
> xmlns:c="http://www.w3.org/ns/xproc-step"
> name="myPipeline">
>
> <p:input port="parameters" kind="parameter">
> <p:inline>
> <c:param-set>
> <c:param name="Title" value="First and Last Freedom" />
> <c:param name="Author" value="J. Krishnamurti" />
> <c:param name="Date" value="1954" />
> <c:param name="ISBN" value="0-06-064831-7" />
> <c:param name="Publisher" value="Harper & Row" />
> </c:param-set>
> </p:inline>
> </p:input>
>
> <p:output port="result" sequence="true" />
>
> <p:identity>
> <p:input port="source">
> <p:pipe step="myPipeline" port="parameters" />
> </p:input>
> </p:identity>
>
> </p:declare-step>
>
> Surprisingly, I get as output two <c:param-set> elements:
That's a bug. But it raises an interesting question. What should happen
if you run the above pipeline like this:
$ calabash pipeline.xpl name=value
Should the parameters port contain a parameter named "name" or not?
My first thought is that it should not. But I could be persuaded
otherwise, since we already treat parameters somewhat specially.
Be seeing you,
norm
--
Norman Walsh <ndw@nwalsh.com> | In theory, there is no difference
http://nwalsh.com/ | between theory and practice, but not in
| practice.
Received on Sunday, 28 June 2009 21:06:55 UTC