Is the TTS Cache logic correct?

Development
,
1

In TextToSpeechEntity

         # Is speech already in memory
         if cache_key in self.mem_cache:
             filename = self.mem_cache[cache_key]["filename"]
         # Is file store in file cache
         elif use_cache and cache_key in self.file_cache:
             filename = self.file_cache[cache_key]
             self.hass.async_create_task(self._async_file_to_mem(cache_key))
         # Load speech from engine into memory
         else:
             filename = await self._async_get_tts_audio(
                 engine_instance, cache_key, message, use_cache, language, options
             )

Seems like the cache=F only disables the long term cache. The short term memory cache is still used. I would imagine one of the primary use case is to disable the cahce is the development of TTS generator.