routes: Actually match metadata when finding routes, fix problems with paths. When we look for a route that matches a given page, we now look at the source metadata that comes with that page, and compare it to the metadata we need to build URIs. Also, when matching URIs, we handle the case where a 'path'-component in our pattern may be completely empty, and thus we may be missing some trailing slashes in the URI.

[pytest]
norecursedirs = .* _darcs CVS .svn .git .hg *.egg-info dist build node_modules venv tmp* {args}