Add a benchmark for compiling a file with 10000 items

Prior to this change, all of the benchmarks were micro benchmarks.
However, I work with a project that contains some large fluent files,
and we have observed some slowness when compiling them.

This change adds a benchmark for compiling a 10000 item file, which can
be used to verify performance optimisations in other commits.

Author: leamingrad

MANIFEST.in

@@ -17,4 +17,5 @@ recursive-include tests *.py
 recursive-include tools *.md
 recursive-include tools *.py
 recursive-include tools *.txt
+recursive-include tools *.ftl
 prune docs/_build