forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Turner <je...@apache.org>
Subject Re: Difficulties upgrading FOP website--compliance.html page
Date Sat, 27 Sep 2003 14:19:49 GMT
On Sat, Sep 27, 2003 at 05:54:58AM -0700, Glen Mazza wrote:
> In the upgrade from Forrest 0.4 to Forrest 0.5 (*very*
> nice speed improvement, BTW!) FOP's compliance.xml
> page is no longer working.
...
> Forrest 0.5:
> 
> <map:match pattern="body-compliance.xml">

Ah, that should be 'body-compliance.html'.  Sorry, that's completely my
fault, as I supplied the sitemap-0.5.xmap in the first place.  Diff
attached.

...
> Also, other questions that will help me better understand the site map:
> 
> 1.)  Why are we matching on "body-xxxx" or
> "**body-*.html" within the Forrest sitemap?  We don't
> have any files named body-compliance.xml, or anything
> with "body" in it.  I'm not sure what "body"
> signifies.

It's the 'body' part of index.html, just as menu-index.html is the menu,
tab-index.html is the tabs.  The body, menu and tab pipelines are all
aggregated to form the final page:

      <map:match pattern="*.html">
        <map:aggregate element="site">
          <map:part src="cocoon:/tab-{0}"/>
          <map:part src="cocoon:/menu-{0}"/>
          <map:part src="cocoon:/body-{0}"/>
        </map:aggregate>
        ...

The "body-compliance.html" pipeline is defined just before the generic
"**body-*.html" matcher, in order to apply special formatting.

> 2.)  A major change between 0.4 and 0.5 sitemaps [2]
> is that the attribute label="content" was removed from
> the <map:generator> definitions.  (Only the
> "file-nolabel" generator in Forrest 0.4 lacked this
> attribute.)  What did the presence/absence of the
> "label" attribute originally mean for a generator? 

It defined the "content view" of Forrest URLs.  "Views" are a neat idea;
that you can obtain a different representation of the same resource by
providing an additional ?cocoon-view=.... parameter.  So
index.pdf?cocoon-view=content will return XML for index.pdf.  The Cocoon
crawler used the content view to get links (so it didn't have to parse
opaque formats like PDF).  Unfortunately, getting the link view
practically doubled the rendering time, so Forrest 0.5 uses a new CLI
implementation that does without views, and is much faster (partially) as
a result.

> Why could it now be removed in Forrest 0.5?

Yes, although they don't hurt.

> 3.) The Forrest 0.4 sitemap had five generators
> defined:  directory, html, libre, nekodtd,
> textparser--Forrest 0.5 has only one:  file, with
> "html" and "directory" commented out.  Why are these
> no longer needed?  (Note, reactivating the HTML
> generator did help me with the compliance.xml
> problem.)

Forrest 0.4 had one file, sitemap.xmap, doing everything.  In 0.5, this
has been split into a bunch of smaller *.xmap files:

http://xml.apache.org/forrest/sitemap-ref.html

So most generators have been moved into subsitemaps, which sitemap.xmap
delegates to with <map:mount>.


--Jeff


> Thanks for any assistance,
> Glen
> 
> [1]
> http://cvs.apache.org/viewcvs.cgi/xml-fop/src/documentation/resources/stylesheets/
> [2]
> http://cvs.apache.org/viewcvs.cgi/xml-fop/src/documentation/
> 
> 
> __________________________________
> Do you Yahoo!?
> The New Yahoo! Shopping - with improved product search
> http://shopping.yahoo.com

Mime
View raw message