[widget Audit] toga.DetailedList

[freakboy3742]

Audit of DetailedList.

Builds on #2017 because of dependencies on Table and Sources (especially on GTK).

• Adds the ability to customise the accessors for data.
• Adds the ability to customise the missing value for data.
• Changes the default interpretation order when populating with a list of tuples, or a list of lists. This is to ensure that the first value in each item is the "title", which means a list of strings will produce a usable DetailedList. This shouldn't impact too many people as the only documented way of feeding a DetailedList was a list of dicts, which is unaffected by this change.
• On iOS, the names "Delete" and "Remove" are interpreted as "destructive" actions, and rendered with a red background in accordance with iOS style guides. There's potential to do a similar thing on GTK, showing icons rather than text for certain "magic" words, but I've left that as future work.
• I've added support for a --coverage flag when running the testbed; I was finding it cumbersome to have to run the entire test suite just to get coverage of a single widget.

Fixes:

• Fixes (GTK) Changing DetailedList's Source makes rows invisible #2026
• Fixes The beeliza and detailedlist examples do not show a DetailedList on Android #1997
• Fixes Slow when triggering the "on_refresh" of the DetailedList toga-android==0.3.0.dev39 #1630
• Fixes Error AttributeError: 'NoneType' object has no attribute '_impl' #1473
• Fixes DetailedList: add a selection attribute #962
• Fixes DetailedList, documentation, special keys, multiline text #772
• Fixes Examples do not support new ListSource APIs #2071

Audit checklist

• Core tests ported to pytest
• Core tests at 100% coverage
• Core docstrings complete, and in Sphinx format
• Documentation complete and consistently formatted
• cocoa backend at 100% coverage
• winforms backend at 100% coverage
• gtk backend at 100% coverage
• iOS backend at 100% coverage
• android backend at 100% coverage
• Widget support table updated
• Widget Issue backlog reviewed

[freakboy3742]

A couple of minor suggestions, but otherwise looking good (other than the single failing test)

[mhsmith]

This made the test unreliable in --slow mode, because the handler and the test function are running in separate asyncio tasks and there's no guarantee of which one will finish awaiting first.