[Accessibility] Synthesizer choice api

[Jeremy Whiting]

Hey all,

Today I took another look at qtexttospeech_unix.cpp and realized we are
only exposing the default speech dispatcher output module's voices to the
api. I think either one of two things needs to happen. Which sounds more
usable?

1. When the backend starts up it not only gets the output modules and
create a QVoice for each one. This would require QVoice to have additional
data about which output module it is using, but wouldn't require any api to
switch output modules, just switch voices and it would switch the output
module for you.

2. Add api to QTextToSpeech to get and set the synthesizer. Whenever the
synthesizer is set the list of QVoice objects (availableVoices) will likely
change, but QVoice itself wouldn't need to contain which synthesizer the
voice is for.

I'm a bit torn myself, I think 1 is a better fit for the unix backend, but
it's probably not as usable/meaningful on other platforms. Though maybe all
platforms QVoice will need to have custom data associated with it, so we
could add a QVariant or something to QVoice to hold it.

thoughts?

thanks,
Jeremy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.qt-project.org/pipermail/accessibility/attachments/20150210/06844820/attachment.html>