drop rendered simplates

[chadwhitacre]

Rendered simplates are really negotiated simplates with only one content page allowed. On our February 2014 call, @pjz convinced me to do away with them, so that we only have one type of simplate (JSON simplates are going away in #293).

[chadwhitacre]

We decided not to do this on the March call. Due to the deep-seated conflation of file extension and content type that exists throughout computing, we need to treat filename.html specially.

[chadwhitacre]

IRC

[chadwhitacre]

Reopening after spending time with the dispatcher (#389) and thinking through refactoring the whole of aspen for better modularity (so we can use parts of aspen in other frameworks; also so we can make aspen nice and documented like other libraries—a parallel strategy to calving off libraries).

I think we should honor filename extensions in URL paths, but not in filesystem paths. That is, we should keep supporting indirect content negotiation, but we should drop the implicit specline media type based on filesystem file extension.

[pjz]

I'm okay with that. It doesn't change the dispatcher much if at all, actually, since foo.html.spt is a 'more specific' than foo.spt for a request for /foo.html. The main change would be that mimetype must be specified on the specline now.

[chadwhitacre]

The main change would be that mimetype must be specified on the specline now.

Yeah, that's the main change from a dev/user pov. For us this means we get to clean up a lot of code. :-)

[chadwhitacre]

I'm diving in on this.

[chadwhitacre]

After attempting this on #390, we decided on the October 2014 call to keep rendered resource behavior after all. We are going to simplify the class hierarchy, however. That falls under #341.