Johannes Schaefer wrote:
Ross, Just to get my understanding right:
<snip/>
A better approach is to add your own DTD that supports the elements
you want to use, see
http://forrest.apache.org/docs_0_90/your-project.html#adding_new_content_type
Adding your own DTD with transformation to XDoc means you still
validate the XML, which will prevent one set of funnies (i.e. badly
formed XML), but it still leaves you open to breaking some output
formats if your XSL keeps non-XDoc elements in the output.
The transformation is from myDTD to Forrest's internal format.
The <span> tag still isn't present in xdoc.
So this works because Forrest isn't so strict to validate the xml further
down its processing pipeline. One may use the same feature (bug?)
writing an input plugin thus bypassing Forrests validation of its
internal format.
With other words:
Forrest validates the input files when read "from disk"
but afterwards is "sloppy" about validation and passes
unknown tags on to the output.
In the end <span> gets passed on to HTML and the browser knows how to
handle it. Luckily. With a the tag named <funny-span> your content would
not show up: browsers tend to ignore unknown tags (1).
Breaking other output formats means that either further processing steps
fail because of unknown tags or these tags get ignored.
Am I right?
You are 100% correct :-)
Ross
This is because Forrest's xdoc is not very "rich" semantically.
A switch to XHTML as internal format is envisioned for Forrest.
This will offer more flexibility.
Cheers
Johannes
(1) You may see this effect when inspecting any HTML page of your Forrest
site with "body-" prefixed: there are <tocitems> in the source that
the browser ignore.
Ross
