4 Jun
2009
4 Jun
'09
7:26 p.m.
On Wed, Jun 3, 2009 at 12:29 PM, Vincent Massol <vincent(a)massol.net> wrote:
Hi,
I think we need to add a plain text parser & a "plain/1.0" syntax id.
Here's a use case: you want to add a groovy class in a wiki page (and
you don't want it to be rendered as wiki syntax of course).
Seems like this is an uneccessary complication, and a design compromise made
from expendiency...it also serves to further complicate the document model.
I think scripting needs to be removed entirely from the document so that it
remains purely declative and not procedural. Procedural content should be
packaged and treated differently. I therefore suggest a different approach,
based on existing precedent:
What about following the precedent of the
$xwiki.jsx.use("$doc.fullName") construct. Have other language-specific
"extensions" objects attached to the document as objects, and give options
in the "XWiki.GroovyExtension" instance (analogous to
XWiki.JavaScriptExtension class) for:
(1) velocity parsing
(2) caching
Stylistically, I don't like the practice of dropping all manner of random
procedural <script> in the middle of a declarative Wiki or HTML document
(though in some cases it's necessary to add some javascript post-processing
at the end of a document for JS that needs access to the DOM).
IMHO, the model of using JSX and SSX to inject JavaScript and CSS into a
document seems cleaner, and should be extended to using other scripting
languages in Xwiki.
Niels
http://nielsmayer.com