Hi Samantha,
So far as I'm aware we haven't tested this deeply on iOS yet but have done a lot of testing on OS X with VoiceOver. The main problems that I'm unsure about are around navigation, actually. Specifically:
- We've been building screen reader support strictly on top of the keyboard navigation work from v12 rather than relying on a more web-traditional tab-navigation approach. What this practically means for mobile use cases is that their typical "swipe-to-navigate" approach is unlikely to work correctly unless the accessibility service overrides the default WCAG behaviors (which it very well might, I'm not actually sure). More specifically, it's my understanding that swipe-to-navigate works essentially the same as the 'tab' key for desktop users by moving to the next tab index in the accessibility node tree. Blockly doesn't use tabbable indexes for the majority of its components since navigation is manually handled through keyboard navigation shortcuts and keybindings.
- The less commonly used tap & hold to discover (which I know exists on Android, I'm not actually certain if it also does on iOS) might work but a lot of the ARIA labels that provide screen reader feedback are dynamically calculated when that element is first focused (via keyboard), so I expect this feature may also not quite work correctly. I'm also unsure if it relies on tabbable elements as well.
If you're willing and able, it would actually be really helpful if someone could test either
https://raspberrypifoundation.github.io/blockly-keyboard-experimentation/screenreader/ on a mobile device or your own application integrated manually with the 'add-screen-reader-support-experimental' branches of both Blockly core and the keyboard navigation plugin (we definitely recommend
not shipping a production-facing product with these branches). Any problems that are found would be excellent to collect and added to a new Blockly issue on GitHub since this will help us understand the scope of the mobile work.
One caveat I'll give is that we may not be able to fully support mobile use cases in v13. The assumption is that it will require extra work and is a less common use case for Blockly as a whole. If this is a key requirement for you then that feedback can be helpful to influence priorities, especially after v13 (we could, for instance, possibly release a minor version of v13 that brings full mobile support depending on exactly how much work is needed and depending on other Blockly priorities).
Hope the helps,
Ben