After integrating voice over through accessibilityLabel
s, and testing the interaction alone, it was time to try turning on the voice. Fortunately, it worked perfectly well for english text... But wasn't so lucky with Arabic.
Apparently, voice over utters "unpronounceable" when it reaches Easten Arabic-Indic numerals:
١ ، ٢ ، ٣
It is really inefficient to keep listening to each accessibility label to make sure, so I thought there would be some sort of query we can do to the TTS engine, and write tests around that.
All I know after research is that the underlying TTS engine is AVSpeechSynthesis
, but that doesn't seem to have anything of that sort.