Understanding action pipeline and form content

I'm trying to get clear on how a pipeline called on an page-flow
"action" element relates to (1) the instance form content for that page
and (2) the instance for content for the result page. So, given:

Page 1 form instance of (assume that the content of "selection" is the
result of user action and an xforms "setvalue" element):

<instance>
  <selection>XMLEditor</selection>
</instance>

An action that is called:

<action when="/instance/action = 'next'"
action="/ToolEvalApp/searchEvaluation2/search.xpl">
  <result page="page2" transform="oxf:xslt">
    <search1 xsl:version="2.0" xmlns="">
        <xsl:copy-of select="doc('input:action')/" />
   </search1>
 </result>
</action>

(Yes, I know I haven't specified what "search.xpl" has in it, but I
can't get anything in it to work so imagine something simple, like some
xslt transformation that takes "XMLEditor" and rewrites it as, say, "xml
editor".)

My question is: what is informing what here? Is the pipeline
"search.xpl" informed by the result of the xslt transformation in the
<result> element? Or is it the other way round, so that the result of
the pipeline informs the xslt transform in the <result> element?

When page2 comes up, what will be available to it: the result of the
pipeline or the result of the xslt transform in the <result> element?

Thanks,

Don




--
You receive this message as a subscriber of the [hidden email] mailing list.
To unsubscribe: mailto:[hidden email]
For general help: mailto:[hidden email]?subject=help
ObjectWeb mailing lists service home page: http://www.objectweb.org/wws

Don,

 > I'm trying to get clear on how a pipeline called on an page-flow
 > "action" element relates to (1) the instance form content for that page
 > and (2) the instance for content for the result page. So, given:
 >
 > Page 1 form instance of (assume that the content of "selection" is the
 > result of user action and an xforms "setvalue" element):
 >
 > <instance>
 >   <selection>XMLEditor</selection>
 > </instance>
 >
 > An action that is called:
 >
 > <action when="/instance/action = 'next'"
 > action="/ToolEvalApp/searchEvaluation2/search.xpl">
 >   <result page="page2" transform="oxf:xslt">
 >     <search1 xsl:version="2.0" xmlns="">
 > <xsl:copy-of select="doc('input:action')/" />
 >    </search1>
 >  </result>
 > </action>
 >
 > (Yes, I know I haven't specified what "search.xpl" has in it, but I
 > can't get anything in it to work so imagine something simple, like some
 > xslt transformation that takes "XMLEditor" and rewrites it as, say, "xml
 > editor".)
 >
 > My question is: what is informing what here? Is the pipeline
 > "search.xpl" informed by the result of the xslt transformation in the
 > <result> element? Or is it the other way round, so that the result of
 > the pipeline informs the xslt transform in the <result> element?

What happens, in order:

o An XML submission occurs on the page. This can be an internal
   (within the PFC) or external (e.g. HTTP POST) submission, see:

   http://www.orbeon.com/ops/doc/reference-page-flow#xml-submission

   Either way, this results in an XML document being submitted to a
   particular path handled by the PFC.

o Note that this XML submission, for historical reasons (XForms was
   tied to the PFC in 2.8 and earlier), is usually referred to as
   "instance" in PFC pipeline inputs.

o <action> sections test their XPath expressions on the submitted XML
   document. The first matching action, if any, is selected.

o The selected <action> section's "action" pipeline, if any, is
   run. Its "instance" input receives the submitted XML document.

o If the action pipeline has a "data" output, that "data" output is
   fed to the <result> sections.

o <result> sections test their XPath expressions on the document
   produced by the action pipeline on its "data" output. The first
   matching result, if any, is selected.

o If the selected <result> section contains a transformation, that
   transformation is used to build an internal XML submission that will
   be sent to the resulting page, here "page2".

So the bottom line: the result of the action pipeline may be used by
the XSLT transform in the <result> element by using
doc('input:action'), *not* the other way around.

 > When page2 comes up, what will be available to it: the result of the
 > pipeline or the result of the xslt transform in the <result> element?

The result of the XSLT transform.

Things are reasonably easy once you understand that the intent of the
transform on the <result> element is just to allow doing the
equivalent of an HTTP POST from within a page flow. This is what we
call an "internal" submission.

I hope this is now clearer ;-)

-Erik




--
You receive this message as a subscriber of the [hidden email] mailing list.
To unsubscribe: mailto:[hidden email]
For general help: mailto:[hidden email]?subject=help
ObjectWeb mailing lists service home page: http://www.objectweb.org/wws

Erik,

Yes, thanks. That helps. I'm now able to run a pipeline as I should be
able to.

Now, if you could clarify something else. It seems that sometimes in a
pipeline the input to the pipeline must be named "instance" and
sometimes not; likewise with outputs: sometimes it must be named "data"
and sometimes not.

Believe it or not, I have actually read and re-read the documentation
many times, and I just can't get clear on when/why names for pipeline
inputs and outputs are already designated, and when they may be user
defined. Any light you can shed on this will be much appreciated.

Thanks,

Don
 

-----Original Message-----
From: Erik Bruchez [mailto:[hidden email]] On Behalf Of Erik Bruchez
Sent: Monday, April 10, 2006 8:26 PM
To: [hidden email]
Subject: Re: [ops-users] Understanding action pipeline and form content

Don,

 > I'm trying to get clear on how a pipeline called on an page-flow  >
"action" element relates to (1) the instance form content for that page
> and (2) the instance for content for the result page. So, given:
 >
 > Page 1 form instance of (assume that the content of "selection" is
the  > result of user action and an xforms "setvalue" element):
 >
 > <instance>
 >   <selection>XMLEditor</selection>
 > </instance>
 >
 > An action that is called:
 >
 > <action when="/instance/action = 'next'"
 > action="/ToolEvalApp/searchEvaluation2/search.xpl">
 >   <result page="page2" transform="oxf:xslt">
 >     <search1 xsl:version="2.0" xmlns="">
 > <xsl:copy-of select="doc('input:action')/" />
 >    </search1>
 >  </result>
 > </action>
 >
 > (Yes, I know I haven't specified what "search.xpl" has in it, but I
> can't get anything in it to work so imagine something simple, like
some  > xslt transformation that takes "XMLEditor" and rewrites it as,
say, "xml  > editor".)  >  > My question is: what is informing what
here? Is the pipeline  > "search.xpl" informed by the result of the xslt
transformation in the  > <result> element? Or is it the other way round,
so that the result of  > the pipeline informs the xslt transform in the
<result> element?

What happens, in order:

o An XML submission occurs on the page. This can be an internal
   (within the PFC) or external (e.g. HTTP POST) submission, see:

   http://www.orbeon.com/ops/doc/reference-page-flow#xml-submission

   Either way, this results in an XML document being submitted to a
   particular path handled by the PFC.

o Note that this XML submission, for historical reasons (XForms was
   tied to the PFC in 2.8 and earlier), is usually referred to as
   "instance" in PFC pipeline inputs.

o <action> sections test their XPath expressions on the submitted XML
   document. The first matching action, if any, is selected.

o The selected <action> section's "action" pipeline, if any, is
   run. Its "instance" input receives the submitted XML document.

o If the action pipeline has a "data" output, that "data" output is
   fed to the <result> sections.

o <result> sections test their XPath expressions on the document
   produced by the action pipeline on its "data" output. The first
   matching result, if any, is selected.

o If the selected <result> section contains a transformation, that
   transformation is used to build an internal XML submission that will
   be sent to the resulting page, here "page2".

So the bottom line: the result of the action pipeline may be used by the
XSLT transform in the <result> element by using doc('input:action'),
*not* the other way around.

 > When page2 comes up, what will be available to it: the result of the
> pipeline or the result of the xslt transform in the <result> element?

The result of the XSLT transform.

Things are reasonably easy once you understand that the intent of the
transform on the <result> element is just to allow doing the equivalent
of an HTTP POST from within a page flow. This is what we call an
"internal" submission.

I hope this is now clearer ;-)

-Erik





--
You receive this message as a subscriber of the [hidden email] mailing list.
To unsubscribe: mailto:[hidden email]
For general help: mailto:[hidden email]?subject=help
ObjectWeb mailing lists service home page: http://www.objectweb.org/wws

Smith, Donald wrote:

 > Now, if you could clarify something else. It seems that sometimes in
 > a pipeline the input to the pipeline must be named "instance" and
 > sometimes not; likewise with outputs: sometimes it must be named
 > "data" and sometimes not.
 >
 > Believe it or not, I have actually read and re-read the
 > documentation many times, and I just can't get clear on when/why
 > names for pipeline inputs and outputs are already designated, and
 > when they may be user defined. Any light you can shed on this will
 > be much appreciated.

That's an easy one:

o Pipelines _called by the PFC_ follow a naming convention for inputs
   and outputs. This is where you use "instance" and "data". "instance"
   is used for the submitted XML document, and "data" is used to pass
   data between action, model, view, and epilogue pipelines.

o Your own pipelines just have to follow your own convention.

This is no different from functions or methods/interfaces in other
programming languages: if your API requires me to provide a callback
function, I better make sure that the function I provide takes and
returns the paramters you are expecting. This is exactly how it works
with the PFC: it expects a "view" pipeline, for example, to have an
"instance" input, a "data" input, and a "data"output.

Now if you write a pipeline, say save.xpl, and you call that pipeline
from another pipeline, you are free to define the names of the inputs
and outputs of that pipeline, as long as you use the same names when
you call it.

Anyway here the important part is to understand that the naming
convention is strictly PFC-related. XPL the pipeline language does not
itself know about any such convention.

-Erik




--
You receive this message as a subscriber of the [hidden email] mailing list.
To unsubscribe: mailto:[hidden email]
For general help: mailto:[hidden email]?subject=help
ObjectWeb mailing lists service home page: http://www.objectweb.org/wws