This commit allows pre-compiled templates to be executed without
keeping the original template around. It is still required when using
higher-order functions.

Fixes #262

Fixes #274

clean up and standardize UMD wrapper

Add compiled templates overview and examples to readme

Fixes #267

Don't evaluate lambdas that return falsy values as an empty string.

Performance fix for iphones and ipads

Issue: If I create a Mustache lambda to return numbers and I expect 0 to be of
those numbers, I'll instead only receive an empty string for 0 because it
evaluates to falsy.

Fix: Check the result of the lambda to be undefined or null. Only then should
it return the empty string.

Sections - undefined keys

Fixes #258

Fixes #257

Fixes #247

Text tokens now accurately reflect their content's exact position
in the original template.

Also, changed order of arguments to internal render function.

Fixes #207 #210

This is a strict port of the existing process - there are many opportunities for optimization based on this.

Also, allow partials to be passed to the function that is returned
for pre-compiled templates. This makes the API more consistent all
around, but comes with the caveat that renderers are much more
aggressive at caching. Now, any time a template or partial is
compiled with the same name as a previously compiled template the
cached version is used.

Fixes #240