You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cocoon.apache.org by Torsten Curdt <tc...@dff.st> on 2001/05/07 10:13:10 UTC

some aggregation questions

There is the XIncludeSAXconnector and the sitemap aggregation

1. Which one is usually faster? Talking about a local aggregation
   I assume the SAXconnector?

2. I remember having some trouble putting the map:generate etc inside
   the map:part elmenent ommitting the src attribute. But this should
   work, shouldn't it?! (Maybe I had trouble because of the sub sitemap)

3. Looking at the aggregation example I recognized the parts are
   serialized to html before they are aggregated. Why? This does not
   make much sense to me.
--
Torsten



---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Giacomo Pati <gi...@apache.org>.
Quoting Sylvain Wallez <sy...@anyware-tech.com>:

> 
> 
> Giacomo Pati a écrit :
> > 
> > Quoting Torsten Curdt <tc...@dff.st>:
> > 
> > > > > > There aren't any element inside a map:parm element:
> > > > > >
> > > > > >    <map:aggregate ...>
> > > > > >      <map:part .../>
> > > > > >      <map:part .../>
> > > > > >    </map:aggregate>
> > > > >
> > > > > Yes... but I'd like to put elements there...
> > > > >
> > > > >     <map:aggregate ...>
> > > > >       <map:part ... >
> > > > >          <map:generate .../>
> > > > >          <map:transform .../>
> > > > >       </map:part>
> > > > >       <map:part ... >
> > > > >          <map:generate .../>
> > > > >          <map:transform .../>
> > > > >       </map:part>
> > > > >       <map:transform .../>
> > > > >       <map:serialize/>
> > > > >     </map:aggregate>
> > > >
> > > > you can achieve the same results by putting each of the
> > > generate/transform
> > > > blocks in their own 'named' pipeline - but i agree, doing it this
> way
> > > > would be handy sometimes. it doesn't break any seperation of
> concerns
> > > that
> > > > i can see, just an alternate syntax.
> > 
> > Ok, let's discuss this. The initial proposal was like the first sample
> above
> > which nobody objected or came with another suggestion.
> > 
> > Now, after implementing that Torsten came up with another syntax
> (second
> > sample above).
> > 
> > So how should the <map:aggregate> snippet look like:
> > 
> > a) like the initial proposal
> > b) like Torstens proposal
> > c) a combination of both (ie. map:part must be empty if src attribute
> >    specified otherwise there should be generators/transformers inside
> >    map:part).
> > 
> > > Hm, Giacomo, to you think this is hard to change?
> > 
> > Well I don't think it should be that hard :)
> > 
> > > Putting everything
> > > outside can be really anoying and is makeing the sitemap more
> complex
> > > than it should be.
> > 
> > I don't agree here. I think it make sense to define separate pipelines
> somewhere
> > in the sitemap and finally aggregate them into one single resource
> (ie.
> > debugging aspects).
> > 
> > Look, people will soon come with having actions, matcher and selectors
> inside
> > map:part as well which doesn't make it look better. What would you
> tell them
> > than? I feel it make more sense to have them separated. You see each
> single
> > pipeline as it is and also see the aggregation how it is done as
> compact
> > snippets in the sitemap.
> > 
> 
> +1 for a) 
> - if you aggregate parts, it's likely that these parts will be reused
> somewhere else in the sitemap
> - I agree with Giacomo concerning the readability of the map:aggregate
> when you have complex pipelines.
> 
> But there's an issue here : the parts have to be declared as complete
> pipelines, i.e. include a serializer, which means they're potentially
> visible by the user if he knows their URL. There may be cases when
> access to part information should be allowed only to internal requests.
> How can we handle that ?

This is already solved :)

<map:pipeline internal-only="true">
 ...
</map:pipeline>

Every pipeline inside the internal-only pipeline is only accessible by 
aggregator parts or by the "transparent" xinclude facility.

Giacomo

> 
> > > If you can give some hints - maybe I can change it myself...
> > 
> > sitemap.xsl ContentAgregator.java
> > 
> > Giacomo
> > 
> -- 
> Sylvain Wallez
> Anyware Technologies - http://www.anyware-tech.com
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
> For additional commands, email: cocoon-dev-help@xml.apache.org
> 
> 
> 

---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Sylvain Wallez <sy...@anyware-tech.com>.

Giacomo Pati a écrit :
> 
> Quoting Torsten Curdt <tc...@dff.st>:
> 
> > > > > There aren't any element inside a map:parm element:
> > > > >
> > > > >    <map:aggregate ...>
> > > > >      <map:part .../>
> > > > >      <map:part .../>
> > > > >    </map:aggregate>
> > > >
> > > > Yes... but I'd like to put elements there...
> > > >
> > > >     <map:aggregate ...>
> > > >       <map:part ... >
> > > >          <map:generate .../>
> > > >          <map:transform .../>
> > > >       </map:part>
> > > >       <map:part ... >
> > > >          <map:generate .../>
> > > >          <map:transform .../>
> > > >       </map:part>
> > > >       <map:transform .../>
> > > >       <map:serialize/>
> > > >     </map:aggregate>
> > >
> > > you can achieve the same results by putting each of the
> > generate/transform
> > > blocks in their own 'named' pipeline - but i agree, doing it this way
> > > would be handy sometimes. it doesn't break any seperation of concerns
> > that
> > > i can see, just an alternate syntax.
> 
> Ok, let's discuss this. The initial proposal was like the first sample above
> which nobody objected or came with another suggestion.
> 
> Now, after implementing that Torsten came up with another syntax (second
> sample above).
> 
> So how should the <map:aggregate> snippet look like:
> 
> a) like the initial proposal
> b) like Torstens proposal
> c) a combination of both (ie. map:part must be empty if src attribute
>    specified otherwise there should be generators/transformers inside
>    map:part).
> 
> > Hm, Giacomo, to you think this is hard to change?
> 
> Well I don't think it should be that hard :)
> 
> > Putting everything
> > outside can be really anoying and is makeing the sitemap more complex
> > than it should be.
> 
> I don't agree here. I think it make sense to define separate pipelines somewhere
> in the sitemap and finally aggregate them into one single resource (ie.
> debugging aspects).
> 
> Look, people will soon come with having actions, matcher and selectors inside
> map:part as well which doesn't make it look better. What would you tell them
> than? I feel it make more sense to have them separated. You see each single
> pipeline as it is and also see the aggregation how it is done as compact
> snippets in the sitemap.
> 

+1 for a) 
- if you aggregate parts, it's likely that these parts will be reused
somewhere else in the sitemap
- I agree with Giacomo concerning the readability of the map:aggregate
when you have complex pipelines.

But there's an issue here : the parts have to be declared as complete
pipelines, i.e. include a serializer, which means they're potentially
visible by the user if he knows their URL. There may be cases when
access to part information should be allowed only to internal requests.
How can we handle that ?

> > If you can give some hints - maybe I can change it myself...
> 
> sitemap.xsl ContentAgregator.java
> 
> Giacomo
> 
-- 
Sylvain Wallez
Anyware Technologies - http://www.anyware-tech.com

---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Torsten Curdt <tc...@dff.st>.
> > > > > There aren't any element inside a map:parm element:
> > > > >
> > > > >    <map:aggregate ...>
> > > > >      <map:part .../>
> > > > >      <map:part .../>
> > > > >    </map:aggregate>
> > > >
> > > > Yes... but I'd like to put elements there...
> > > >
> > > >     <map:aggregate ...>
> > > >       <map:part ... >
> > > >          <map:generate .../>
> > > >          <map:transform .../>
> > > >       </map:part>
> > > >       <map:part ... >
> > > >          <map:generate .../>
> > > >          <map:transform .../>
> > > >       </map:part>
> > > >       <map:transform .../>
> > > >       <map:serialize/>
> > > >     </map:aggregate>
> > >
> > > you can achieve the same results by putting each of the
> > generate/transform
> > > blocks in their own 'named' pipeline - but i agree, doing it this way
> > > would be handy sometimes. it doesn't break any seperation of concerns
> > that
> > > i can see, just an alternate syntax.
>
> Ok, let's discuss this. The initial proposal was like the first sample above
> which nobody objected or came with another suggestion.
>
> Now, after implementing that Torsten came up with another syntax (second
> sample above).
>
> So how should the <map:aggregate> snippet look like:
>
> a) like the initial proposal
> b) like Torstens proposal
> c) a combination of both (ie. map:part must be empty if src attribute
>    specified otherwise there should be generators/transformers inside
>    map:part).

Well, my proposal was of course c)! ;) Both should be possible!

I think my proposal is quite straight forward
- from a user's point of view. As long as nesting
makes sense it will be not comprehensible to the
user why not to nest the desired tags.
And I think this approach is right - as long it
doesn't break anything in terms of SOC.

And in this case it doesn't IMHO.

> > Hm, Giacomo, to you think this is hard to change?
> Well I don't think it should be that hard :)
>
> > Putting everything
> > outside can be really anoying and is makeing the sitemap more complex
> > than it should be.
>
> I don't agree here. I think it make sense to define separate pipelines somewhere
> in the sitemap and finally aggregate them into one single resource (ie.
> debugging aspects).

I personally think only having the current syntax works against
the sitemap concept. (Sorry, I haven't spoken up before the
implementation. Don't get me wrong!! I'm very(!!) happy to have
what we have right now ;) ) But now working with this stuff
I get more and more sitemap entries even for a single page.

I guess having these matchers that are only used internally
mess things up.

I encounter this when trying to pass a parameter from
the aggregation matcher to part matcher.


   <map:match pattern="included">
         <map:generate type="serverpages" src="included.xsp">
            <parameter name="path" value="{from_the_last_matcher}"/>
         </map:generate>
         <map:serialize/>
   </map:match>

   <map:match pattern="**/test.html">
    <map:aggregate element="page" ns="">
     <map:part src="included" .../>
     <map:part src=... />
    </map:aggregate>
   </map:match>

And even if this might be possible
by using {../1} (is it?) this is not very
nice IMHO because "included" makes only sense
together with the other matcher.

> Look, people will soon come with having actions, matcher and selectors inside
> map:part as well which doesn't make it look better. What would you tell them
> than? I feel it make more sense to have them separated. You see each single
> pipeline as it is and also see the aggregation how it is done as compact
> snippets in the sitemap.

I think it is nothing bad to have actions and stuff inside... Why not?
Maybe again the flexibility syndrom... but I'm a friend of reusing
syntax as much as possible.

> > If you can give some hints - maybe I can change it myself...
>
> sitemap.xsl ContentAgregator.java

I gonna have a look at that ;)
--
Torsten


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: AW: some aggregation questions

Posted by Torsten Curdt <tc...@dff.st>.
> > > > > > There aren't any element inside a map:parm element:
> > > > > >
> > > > > >    <map:aggregate ...>
> > > > > >      <map:part .../>
> > > > > >      <map:part .../>
> > > > > >    </map:aggregate>
> > > > >
> > > > > Yes... but I'd like to put elements there...
> > > > >
> > > > >     <map:aggregate ...>
> > > > >       <map:part ... >
> > > > >          <map:generate .../>
> > > > >          <map:transform .../>
> > > > >       </map:part>
> > > > >       <map:part ... >
> > > > >          <map:generate .../>
> > > > >          <map:transform .../>
> > > > >       </map:part>
> > > > >       <map:transform .../>
> > > > >       <map:serialize/>
> > > > >     </map:aggregate>
> > > >
> > > > you can achieve the same results by putting each of the
> > > generate/transform
> > > > blocks in their own 'named' pipeline - but i agree, doing it this way
> > > > would be handy sometimes. it doesn't break any seperation of concerns
> > > that
> > > > i can see, just an alternate syntax.
> >
> > Ok, let's discuss this. The initial proposal was like the first
> > sample above
> > which nobody objected or came with another suggestion.
> >
> > Now, after implementing that Torsten came up with another syntax (second
> > sample above).
> >
> > So how should the <map:aggregate> snippet look like:
> >
> > a) like the initial proposal
> > b) like Torstens proposal
> > c) a combination of both (ie. map:part must be empty if src attribute
> >    specified otherwise there should be generators/transformers inside
> >    map:part).
> >
> +10 for a)
>
> I see three reasons for a):
> 1. We are going beta this week and we should avoid big changes right now.
>    I personally have a list with others changes like that above, but we
>    should really discuss them after version 2.0.

Well, that's totally reasonable...
+1

> 2. There was enough time to discuss the syntax of the aggregation.

But why not discussing it again if the current implementation
might be as flexible as it could/should be?

(Doesn't mean it should be high priority...)

> 3. I agree with Giacomo that defining separate pipelines and "only"
>    aggregate them is the easier way for now. Otherwise we would really
>    need all possibilities, like selectors, matchers and actions during
>    the aggregation.

Yep! I think we do ;)
--
Torsten


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


AW: some aggregation questions

Posted by Carsten Ziegeler <cz...@sundn.de>.
> Giacomo Pati wrote:
>
> Quoting Torsten Curdt <tc...@dff.st>:
> 
> > > > > There aren't any element inside a map:parm element:
> > > > >
> > > > >    <map:aggregate ...>
> > > > >      <map:part .../>
> > > > >      <map:part .../>
> > > > >    </map:aggregate>
> > > >
> > > > Yes... but I'd like to put elements there...
> > > >
> > > >     <map:aggregate ...>
> > > >       <map:part ... >
> > > >          <map:generate .../>
> > > >          <map:transform .../>
> > > >       </map:part>
> > > >       <map:part ... >
> > > >          <map:generate .../>
> > > >          <map:transform .../>
> > > >       </map:part>
> > > >       <map:transform .../>
> > > >       <map:serialize/>
> > > >     </map:aggregate>
> > >
> > > you can achieve the same results by putting each of the
> > generate/transform
> > > blocks in their own 'named' pipeline - but i agree, doing it this way
> > > would be handy sometimes. it doesn't break any seperation of concerns
> > that
> > > i can see, just an alternate syntax.
> 
> Ok, let's discuss this. The initial proposal was like the first 
> sample above 
> which nobody objected or came with another suggestion.
> 
> Now, after implementing that Torsten came up with another syntax (second 
> sample above).
> 
> So how should the <map:aggregate> snippet look like:
> 
> a) like the initial proposal
> b) like Torstens proposal
> c) a combination of both (ie. map:part must be empty if src attribute 
>    specified otherwise there should be generators/transformers inside 
>    map:part).
>
+10 for a)

I see three reasons for a):
1. We are going beta this week and we should avoid big changes right now.
   I personally have a list with others changes like that above, but we
   should really discuss them after version 2.0.
2. There was enough time to discuss the syntax of the aggregation.
3. I agree with Giacomo that defining separate pipelines and "only"
   aggregate them is the easier way for now. Otherwise we would really
   need all possibilities, like selectors, matchers and actions during
   the aggregation.

> > Hm, Giacomo, to you think this is hard to change? 
> 
> Well I don't think it should be that hard :)
> 
> > Putting everything
> > outside can be really anoying and is makeing the sitemap more complex
> > than it should be.
> 
> I don't agree here. I think it make sense to define separate 
> pipelines somewhere 
> in the sitemap and finally aggregate them into one single resource (ie. 
> debugging aspects).
> 
> Look, people will soon come with having actions, matcher and 
> selectors inside 
> map:part as well which doesn't make it look better. What would 
> you tell them 
> than? I feel it make more sense to have them separated. You see 
> each single 
> pipeline as it is and also see the aggregation how it is done as compact 
> snippets in the sitemap.
> 
> > If you can give some hints - maybe I can change it myself...
> 
> sitemap.xsl ContentAgregator.java
> 
> Giacomo
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
> For additional commands, email: cocoon-dev-help@xml.apache.org
> 

Carsten 

Open Source Group                        sunShine - b:Integrated
================================================================
Carsten Ziegeler, S&N AG, Klingenderstrasse 5, D-33100 Paderborn
www.sundn.de                          mailto: cziegeler@sundn.de 
================================================================


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Donald Ball <ba...@webslingerZ.com>.
On Tue, 8 May 2001, Giacomo Pati wrote:

> Ok, let's discuss this. The initial proposal was like the first sample above
> which nobody objected or came with another suggestion.
>
> Now, after implementing that Torsten came up with another syntax (second
> sample above).
>
> So how should the <map:aggregate> snippet look like:
>
> a) like the initial proposal
> b) like Torstens proposal
> c) a combination of both (ie. map:part must be empty if src attribute
>    specified otherwise there should be generators/transformers inside
>    map:part).

+1 to c. it's a simple mod that doesn't change the existing grammar, just
augments it.

- donald


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Giacomo Pati <gi...@apache.org>.
Quoting Torsten Curdt <tc...@dff.st>:

> > > > There aren't any element inside a map:parm element:
> > > >
> > > >    <map:aggregate ...>
> > > >      <map:part .../>
> > > >      <map:part .../>
> > > >    </map:aggregate>
> > >
> > > Yes... but I'd like to put elements there...
> > >
> > >     <map:aggregate ...>
> > >       <map:part ... >
> > >          <map:generate .../>
> > >          <map:transform .../>
> > >       </map:part>
> > >       <map:part ... >
> > >          <map:generate .../>
> > >          <map:transform .../>
> > >       </map:part>
> > >       <map:transform .../>
> > >       <map:serialize/>
> > >     </map:aggregate>
> >
> > you can achieve the same results by putting each of the
> generate/transform
> > blocks in their own 'named' pipeline - but i agree, doing it this way
> > would be handy sometimes. it doesn't break any seperation of concerns
> that
> > i can see, just an alternate syntax.

Ok, let's discuss this. The initial proposal was like the first sample above 
which nobody objected or came with another suggestion.

Now, after implementing that Torsten came up with another syntax (second 
sample above).

So how should the <map:aggregate> snippet look like:

a) like the initial proposal
b) like Torstens proposal
c) a combination of both (ie. map:part must be empty if src attribute 
   specified otherwise there should be generators/transformers inside 
   map:part).

> Hm, Giacomo, to you think this is hard to change? 

Well I don't think it should be that hard :)

> Putting everything
> outside can be really anoying and is makeing the sitemap more complex
> than it should be.

I don't agree here. I think it make sense to define separate pipelines somewhere 
in the sitemap and finally aggregate them into one single resource (ie. 
debugging aspects).

Look, people will soon come with having actions, matcher and selectors inside 
map:part as well which doesn't make it look better. What would you tell them 
than? I feel it make more sense to have them separated. You see each single 
pipeline as it is and also see the aggregation how it is done as compact 
snippets in the sitemap.

> If you can give some hints - maybe I can change it myself...

sitemap.xsl ContentAgregator.java

Giacomo

---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Torsten Curdt <tc...@dff.st>.
> > > There aren't any element inside a map:parm element:
> > >
> > >    <map:aggregate ...>
> > >      <map:part .../>
> > >      <map:part .../>
> > >    </map:aggregate>
> >
> > Yes... but I'd like to put elements there...
> >
> >     <map:aggregate ...>
> >       <map:part ... >
> >          <map:generate .../>
> >          <map:transform .../>
> >       </map:part>
> >       <map:part ... >
> >          <map:generate .../>
> >          <map:transform .../>
> >       </map:part>
> >       <map:transform .../>
> >       <map:serialize/>
> >     </map:aggregate>
>
> you can achieve the same results by putting each of the generate/transform
> blocks in their own 'named' pipeline - but i agree, doing it this way
> would be handy sometimes. it doesn't break any seperation of concerns that
> i can see, just an alternate syntax.

Hm, Giacomo, to you think this is hard to change? Putting everything
outside can be really anoying and is makeing the sitemap more complex
than it should be.

If you can give some hints - maybe I can change it myself...
--
Torsten


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Donald Ball <ba...@webslingerZ.com>.
On Mon, 7 May 2001, Torsten Curdt wrote:

> > > 2. I remember having some trouble putting the map:generate etc inside
> > >    the map:part elmenent ommitting the src attribute. But this should
> > >    work, shouldn't it?! (Maybe I had trouble because of the sub sitemap)
> >
> > There aren't any element inside a map:parm element:
> >
> >    <map:aggregate ...>
> >      <map:part .../>
> >      <map:part .../>
> >    </map:aggregate>
>
> Yes... but I'd like to put elements there...
>
>     <map:aggregate ...>
>       <map:part ... >
>          <map:generate .../>
>          <map:transform .../>
>       </map:part>
>       <map:part ... >
>          <map:generate .../>
>          <map:transform .../>
>       </map:part>
>       <map:transform .../>
>       <map:serialize/>
>     </map:aggregate>

you can achieve the same results by putting each of the generate/transform
blocks in their own 'named' pipeline - but i agree, doing it this way
would be handy sometimes. it doesn't break any seperation of concerns that
i can see, just an alternate syntax.

- donald


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Torsten Curdt <tc...@dff.st>.
> > There is the XIncludeSAXconnector and the sitemap aggregation
> >
> > 1. Which one is usually faster? Talking about a local aggregation
> >    I assume the SAXconnector?
>
> I have no idea of which one will be faster.

But you doubt there is a major difference?

> > 2. I remember having some trouble putting the map:generate etc inside
> >    the map:part elmenent ommitting the src attribute. But this should
> >    work, shouldn't it?! (Maybe I had trouble because of the sub sitemap)
>
> There aren't any element inside a map:parm element:
>
>    <map:aggregate ...>
>      <map:part .../>
>      <map:part .../>
>    </map:aggregate>

Yes... but I'd like to put elements there...

    <map:aggregate ...>
      <map:part ... >
         <map:generate .../>
         <map:transform .../>
      </map:part>
      <map:part ... >
         <map:generate .../>
         <map:transform .../>
      </map:part>
      <map:transform .../>
      <map:serialize/>
    </map:aggregate>

> > 3. Looking at the aggregation example I recognized the parts are
> >    serialized to html before they are aggregated. Why? This does not
> >    make much sense to me.
>
> Why do you think they get serializer to html? Because there is a serializer in
> the pipeline? Well, the map:serialize element is used as an indication to the
> sitemap when a pipeline has reached it's final component.

Hm.. I thought a serialiaze without type attribute is the default
serializer - which is in fact the html serializer.
Having this only as a EOP (End-Of-Pipeline) marking is very misleading
IMHO
--
Torsten


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Re: some aggregation questions

Posted by Giacomo Pati <gi...@apache.org>.
Quoting Torsten Curdt <tc...@dff.st>:

> There is the XIncludeSAXconnector and the sitemap aggregation
> 
> 1. Which one is usually faster? Talking about a local aggregation
>    I assume the SAXconnector?

I have no idea of which one will be faster.

> 2. I remember having some trouble putting the map:generate etc inside
>    the map:part elmenent ommitting the src attribute. But this should
>    work, shouldn't it?! (Maybe I had trouble because of the sub sitemap)

There aren't any element inside a map:parm element:

   <map:aggregate ...>
     <map:part .../>
     <map:part .../>
   </map:aggregate>

> 3. Looking at the aggregation example I recognized the parts are
>    serialized to html before they are aggregated. Why? This does not
>    make much sense to me.

Why do you think they get serializer to html? Because there is a serializer in 
the pipeline? Well, the map:serialize element is used as an indication to the 
sitemap when a pipeline has reached it's final component. But the map:part as 
well as the XIncludeSAXConnector strips them away because they only use the real 
XML part. You have to note that a pipeline is made up of a EventPipeline (pure 
SAX) and a StreamPipeline (serializer part). The aggregation components only use 
the EventPipelines.

Giacomo

---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org