Forrest webapp caching mechanism (was: Running forrest as a webapp)

[Ferdinand Soethe <sa...@soethe.net>]

>>One other thing which confuses me, as I understood it running Forrest as a
>>webapp allows you to update the site in real time. However I have an FAQ
>>page and changes are only displayed if I re-start the forrest site running
>>as a webapp, is this a bug?

I also noticed that some changes to pages (and pipelines) seem to
require a restart, but I couldn't discern a pattern behind it.

Is it correct to assume that Forrest's webapp uses Cocoons caching
scheme which (in theory) is supposed to cache each pipeline up to the
first dynamic component and will assemble the response to all
further requests starting from the point where a component of the
pipeline was last modified.

So what about changes to the sitemap? It would seem that the above
approach could even cache requests through successive changes in
(other) parts of a sitemap, since verification of a cached pipeline
is based on checksums calculated for each step.

And these (in theory) should remain the same if our particular part of
the sitemap was not changed.

Pls feel free to correct me if this is total bullsh..

--
Ferdinand Soethe

Re: Forrest webapp caching mechanism

[Ross Gardler <rg...@apache.org>]

Ferdinand Soethe wrote:
> 
>>>One other thing which confuses me, as I understood it running Forrest as a
>>>webapp allows you to update the site in real time. However I have an FAQ
>>>page and changes are only displayed if I re-start the forrest site running
>>>as a webapp, is this a bug?
> 
> 
> I also noticed that some changes to pages (and pipelines) seem to
> require a restart, but I couldn't discern a pattern behind it.
> 
> Is it correct to assume that Forrest's webapp uses Cocoons caching
> scheme which (in theory) is supposed to cache each pipeline up to the
> first dynamic component and will assemble the response to all
> further requests starting from the point where a component of the
> pipeline was last modified.
> 
> So what about changes to the sitemap? It would seem that the above
> approach could even cache requests through successive changes in
> (other) parts of a sitemap, since verification of a cached pipeline
> is based on checksums calculated for each step.

It's all to do with the "check-reload" parameter in the sitemap element 
or the sitemap parameter:

http://cocoon.apache.org/2.1/faq/faq-sitemap.html#faq-N10067

Ross