Document URL handling/matching

[myniva]

Weblounge comes with a complex URL handling/matching mechanism which also takes the resource's flavor, version or language into account (there are even diferent ways how you can provide this additional parameters). Beside the complexity of the URL itself there are also "intelligent" url matchers, which map extended urls to basic ones.

Currently there's no developer documentation which explains the whole url handling/machting magic, which makes it hard to implement new features or find/fix bugs in the current code related to url handling.

Maybe we should also think about simplify the the url handling - at least in my eyes it's a bit too inscrubtable and a lot of it is not realy in use on public sites.

[myniva]

Here's a real-world example which illustrates the complexity. All these URLs point to a sole page:

http://www.swissunihockey.ch/news/2014/03/06
http://www.swissunihockey.ch/news/2014/03/06/de
http://www.swissunihockey.ch/news/2014/03/06/de/html
http://www.swissunihockey.ch/news/2014/03/06/html
http://www.swissunihockey.ch/news/2014/03/06/html/de
http://www.swissunihockey.ch/news/2014/03/06/index.html
http://www.swissunihockey.ch/news/2014/03/06/index_de.html
http://www.swissunihockey.ch/news/2014/03/06/live.html
http://www.swissunihockey.ch/news/2014/03/06/live_de.html

[myniva]

I propose to change the behavior like this:

• Only language may be added as path-parameter
• Request flavor is selected by HTTP request accept header field
• Version (which is of interest only to a logged in editor) can solely be requested by query-parameter
• Remove legacy support of index.html & version.html

So, the only URLs of the above example would then be:

http://www.swissunihockey.ch/news/2014/03/06
http://www.swissunihockey.ch/news/2014/03/06/de