Hi James, Joseph and all,
Don't worry, we're here to help each other.
If you allow me, I could help you make the speechHistory add-on compatible nvda-2019.3, there is not much to do.
You just have to add the import of the sys module and modify line 35 as follows:
text = u''.join([x for x in sequence if isinstance(x, basestring)])
text = u''.join([x for x in sequence if isinstance(x, basestring)]) if sys.version_info.major == 2 else u''.join([x for x in sequence if isinstance(x, str)])
The basestring type doesn't exist with Python 3.
It will also be necessary to define minimumNVDAVersion in the manifest file and set lastTestedNVDAVersion to 2019.3.0.
Optionally, it will also be preferable to define tab as indentation, Joseph and Noelia have always advised me to do it.
For the documentation of the add-on, it is absent, it would be necessary to take that appearing on the community website.
The local folder is also missing, it should be added, in order to add the documentation of the used scripts to translate.
I think I have forgotten nothing, do not hesitate to tell me if something is missing.
Download link of the version tested with nvda-2019.3beta1, I have not yet added the local folder, nor the documentation: