Ferdinand Soethe wrote:
David Crossley wrote:
I gather that you are saying that the XML Processing Instructions
(PI) get stripped somewhere by one of the transformations.
No, they pass through ok if I embedd them into an html-page. Problem
is that the final page needs to have the .php-extension for the Server
to process it correctly. And when I change the extension of the file
from html to php our pipelines don't want to skin it anymore. Too many
extension-based matchers along the route I suppose.
If you have control over the server you can control which pages ares
processed by the PHP engine. However, changing it to process all html
pages would be a performance problem since I guess many HTML files will
not contain PHP elements.
If I understand correctly you want to have Forrest generate *.php files
which will then be processed by the PHP engine. That is Forrest is *not*
processing the PHP itself. It is easy to create a pipeline for this.
Rather than sharing the code offlist can I suggest that the relevant
pipeline snippets are shared online so that others can benefit (better
still submit a plugin to the project).
In the absence of this, all you (should) need to do is copy the two HTML
matchers in sitemap.xmap (pattern="*.html" and pattern="**/*.html") with
identical matchers put looking for the php extension, i.e.
pattern="*.php" and pattern="**/*.php"
Then, in each place where {0} appears change it to {1}.html (this is
because {0} includes the extension of the match). This assumes that our
internal processing allows the PHP tags to pass through corectly. You
appear to confirm this above.
Note, this will result in any document that would normally be processed
by Forrest to create an HTML page being able to include PHP elements as
well (assuming that you can encode them in the source format of course).
FOr example, if you turn of validation you could include the elements in
an XDoc.
WARNING
This is untested, but would be where I would start experimenting and I'm
pretty sure is close to correct (maybe even correct ;-).
Ross
