Removing site name from URLs - $URL_PREFIX

Hi, we have a requirement so that site/folder names are removed from generated URL's (as well as removing the .page extension for pages).

Like many other implementations, a number of different sites are provided per branch and these sites have their own domain names. So currently we'll have for example:

www.examplesite1.com/example1/index.page

www.examplesite2.com/example2/index.page

We want this to be:

www.examplesite1.com/index

www.examplesite2.com/index

We sit Apache in front of LSDS as a reverse proxy and I'm completely comfourtable configuring incoming URL's for each host to proxy/rewrite to the correct location.

Links generated from content is what I'm concerned about. Another colleague (who's left the company) previously wrote a custom page token for links and that gives us the ability to write rules which do regex matches/replaces on certain types of links using bean configuration. For example in preview we can render different URLs.

All of these rules currently use the $PAGE_LINK token as part of the regex replace, however I figure I can instead use $URL_PREFIX instead for my new rules which appears to generate the desired URL i.e. no site/folder name and no .page extension.

It also means I can choose to keep using $PAGE_LINK in preview so that links will work correctly in preview.

I don't see any issues with this, can anyone else?

Find more posts tagged with

Comments

nipper

Agreed, Use Apache rewrites/redirects. I like the features that HP includes, but stuff like this should be done at the web server layer and not at the app server layer.

The change to URL_PREFIX should make your SEO guys happy. I am dealing with fixing this after the fact.

mstradling

I'd go with a custom token and use that for all paths, you just need run the token value through a conditional block to work out what logic to run. You can also remove the need to test the link type in your XSL:

Starts "http://" or "https://" -> Fully qualified absolute URL, just pass through.
Starts "/sites/" and ends ".page" -> Page link, remove "/sites/<sitename>" prefix and ".page" suffix.
Doesn't start "/" -> Badly entered URL, prefix with "/".
Ends in "/" -> Folder path, append with "index" (or whatever your default page name is).
Anything else you need.

If you have multiple sites that do/could be cross-linked then you could easilly extend this to check if it's an on-site link (so just strip "/sites/<site-name>" and ".page") or cross-site link (so replace "/sites/<site-name>" with <other-domain-name>" and strip ".page"). Editors can then just browse the link, not have to remember (or more likely forget) to add the hostname.
If you don't want to apply the logic in TeamSite preview, just wrap everything in a context.isPreview() consitional block.
If you want to run your "/images" folder from a media domain, check for "/images/" and prefix with "http://media.mydomain.com".
Anything you can think of.