cocoon-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vadim Gritsenko" <>
Subject RE: Relationship between pipelines and match nodes
Date Sat, 09 Mar 2002 16:10:53 GMT
> From: Peter Robins []
> I'm getting rather confused on what a pipeline actually consists of,

Pipeline object consist from:
 - 1 generator
 - 0..N transformers
 - 1 serializer
 - 1 reader
 - 1 redirect

And, pipeline object is not the same as <map:pipeline/> element in the
sitemap file.

> and can't find any docu on this.
> In the sample sitemap, some match nodes are in
> their own pipeline, other pipelines contain many different matches.

It does not mean anything really.

> One pipeline contains
> its own error-handling definition, which implies that you
> can have different pipelines with differing error-handling routines
(altho in
> this case it seems to be identical, so quite what purpose . . .).

Yes. Take a look at sub/sitemap.xmap, it specifically have a sample of
customized 404 (IIRC).

> If so, what happens
> for those pipelines that don't have an error-handling definition?

They will get default, built-in error handler.

> I eventually found a definition of the 'cocoon:/' syntax buried
> in the changes log,

Were do you want it to be buried? ;)

> which says that it accesses another pipeline,

It accesses another *pipeline*, not *<map:pipeline/>* - these are
different things. <map:pipeline/> defines how pipeline is assembled.

> yet in the documentation sitemap,
> the example doesn't go to another pipeline but to
> another match in the same pipeline.

See previous comment. Don't mix sitemap file syntax and the sitemap
object being created as a result of processing sitemap file. Also, take
a look at changes entry saying about "Additions to sitemap syntax:".
Matchers can be nested in other matchers, selectors, actions, etc. So
your sentence is not correct.

> In any case, afaics, there is no
> identifier for pipelines.

Exactly. Sitemap is build dynamically by sitemap engine based on current
execution environment, and the resulting pipeline object does not bear
any identifier.

> If these matches were all in their own separate pipelines, would the
> behaviour be different, and if so how?

<map:pipeline/> has two things which could be customized:
 - @internal-only=true: makes whole <map:pipeline> inaccessible from
outside world,
 - <map:handle-errors/>: changes error handling routine.

If you move some of the code from one <map:pipeline/> two another,
behavior of the system will change only if these <map:pipeline/>s have
different customizations.


Please check that your question has not already been answered in the
FAQ before posting. <>

To unsubscribe, e-mail: <>
For additional commands, e-mail: <>

View raw message