actually getting linked to the recognizer's settings in that case.
Also add a summary line to the line for choosing the recognizer, so you
can see which one is currently selected, in the case that more than one
is installed.
Fix the TTS settings strings to match what we do for voice input.
only showing the option to choose if there is more than one to choose from. Use
the new settingsActivity meta-data to target the appropriate settings activity
for the chosen recognizer.
This replaces the existing "Text-to-speech" section, placing that
inside, along with a link to the newly exposed voice search settings
(which will soon be expanded) and a link to the LatinIME settings
(pending a link to a new section in LatinIME for just voice settings,
including the comma setting and a new input language setting).
The VoiceSearch and LatinIME links are smart enough to remove themselves
if those don't exist on the device.
