improve state management in speech.py

[seanbudd]

Link to issue number:

Fixes #12393, follow up of #12251

Summary of the issue:

Using shift to resume speech after using shift to pause speech does not resume speech.

This is because when isPaused is imported through speech\__init__.py it is does not reference the global variable used in speech.py

Description of how this pull request fixes the issue:

Removes the use of top level variables in speech.py in favour of a private dataclass to manage state, and an enum for certain modes. Constants (or rather, variables with names in all capitals) remain as is.

Testing strategy:

Using reproduction steps in #12393 to manually test the pausing and resuming of speech.

System tests for this may be useful, but require a change to the system test spy synth driver.

Known issues with pull request:

None

Change log entries:

Developer changes:

• speech.curWordChars has been renamed speech._curWordChars
• the following have been removed from speech and can be accessed through speech.getState(). These are readonly values now.
  • speechMode
  • speechMode_beeps_ms
  • beenCanceled
  • isPaused
• to update speech.speechMode use speech.setSpeechMode
• the following have been moved to speech.SpeechMode
  • speech.speechMode_off becomes speech.SpeechMode.off
  • speech.speechMode_beeps becomes speech.SpeechMode.beeps
  • speech.speechMode_talk becomes speech.SpeechMode.talk

Code Review Checklist:

• Pull Request description is up to date.
• Unit tests.
• System (end to end) tests.
• Manual tests.
• User Documentation.
• Change log entry.
• Context sensitive help for GUI changes.

[AppVeyorBot]

• Unit test failure
• Build (for testing PR)

See test results for failed build of commit 46cab6a7b4

[feerrenrut]

Generally looks good!

[feerrenrut]

I think this should be lower case: _speechState