</snip>
>
> I'm sorry, but I can't spot a difference between these two options:
>
> <map:serialize type="xhtml" pragma="{cache}"
> cache-control="{cache}"/>
>
> and:
>
> <map:act type="header">
> <map:parameter name="pragma" value="{pragma}"/>
> <map:parameter name="cache-control" value="{cache}"/>
> </map:act>
> <map:serialize type="xhtml"/>
>
> In both cases, sitemap parameter expansion will happen while
> sitemap is executed
> and pipeline is built. In first case you'd presumable set
> response headers in
> Serializer.setOutputStream() method, which is happening a bit
> later comparing
> with Action.act() of the second case, but still it should be
> the same effect,
> isn't it?
In this example, there is really no difference in output indeed, but if you
have one catch all matcher with one single serializer which is always used for
every request (I always use <map:match pattern="**" and then an aggregate below
it that aggregates content with map:parts that delegate to subsitemaps ). So,
all request eventually use the same serializer. I cannot know in this matcher,
that there will be a form with contiuation, so I cannot set the header in this
pipeline (then stateless flat html pages would get a pragma no-cache header for
example).
And, thx Bertrand for finding back the discussion [1], in combination with the
problem that setting http headers on internal pipelines do not make it to the
main pipepipeline, therefor, my suggestion to store it like the status-code in
a variable. Ofcourse, when [1] is easier to fix, that would probably be the
better way to go.
[1] https://issues.apache.org/jira/browse/COCOON-1619
Ard
>
> Vadim
>
