Re: Make p:http-request required?

On 5/8/07, Jeni Tennison <jeni@jenitennison.com> wrote:
>
>
> Alex Milowski wrote:
> > There are other things to set than HTTP headers as well as entity bodies
> > that
> > you might want to post that aren't XML (e.g.
> > application/x-www-form-urlencoded)
> > that can be embedded inside the c:body element.
>
> Can't the "other things" be set with options? I'm quite happy with
> providing the entity wrapped in a <c:body> element, either always or
> only when the content type isn't application/xml.
>
> It feels like you have an underlying objection to making p:http-request
> more consistent with other steps, or perhaps a particular use case, but
> you're not saying what it is.


My basic objection to using options is that the input to
formulating a request is a complex and structured object.  In recent
times, we use XML to structure such inputs.

The typical way I do this in my smallx pipeline is that there is a preceding
XSLT step that formulates the request by transforming some set of
parameters and input to the pipeline into a specific request against
a REST service.  I want to be able to dynamically produce the
request from something like XSLT.  Having to then disassemble the
complex structure into a simple set of name/value pairs is a lot
of extra work with plenty of room for error.

Perhaps it's just that you know the current p:http-request spec isn't
> complete and want to be given a chance to complete it before it's
> commented on.


There is certainly more work to do to completely specify the semantics
of the XML.


-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics