Text To Speech enactor should be supplied with a mock implementation to facilitate integration testing

We need to implement a mock version of the fluid.textToSpeech component for integration testing purposes. This will enable us to create a linear log of all the items that were queued for speech, for which we could then issue a jqUnit.assertDeepEq assertion for verification. Issues such as FLOE-363 ( https://github.com/fluid-project/first-discovery/pull/77 ) could then straightforwardly be extended with integration tests which target this API.

See channel discussion: https://botbot.me/freenode/fluid-work/2015-06-02/?msg=40811279&page=2

This will also allow us to run some TTS tests on browsers where the API is currently not supported, e.g. Firefox.