Deprecate SingletonHasTraits & related classes

[midhun-pm]

Fix #881
PR updates creates a Deprecated Classes section in the has_traits docs.

[mdickinson]

I think we could replace "classes and instances" with just "subclasses" here.

[mdickinson]

This LGTM as far as it goes. Do you think we can find a way for these classes to issue a DeprecationWarning at instantiation time? If there are users of these classes out there, it's unlikely that they'll notice the doc change - a warning would be more useful, if we can find a way to do it.

[mdickinson]

Can we look into ways to have these issue a DeprecationWarning on use?

[midhun-pm]

Made the doc changes.
Also issuing a DeprecationWarning on instantiation.
Added tests for deprecation.

Had to force push after messing up the history.

[codecov-io]

Codecov Report

Merging #887 into master will increase coverage by 0.3%.
The diff coverage is 100%.

@@            Coverage Diff            @@
##           master     #887     +/-   ##
=========================================
+ Coverage   72.58%   72.89%   +0.3%     
=========================================
  Files          51       51             
  Lines        6471     6474      +3     
  Branches     1302     1302             
=========================================
+ Hits         4697     4719     +22     
+ Misses       1378     1355     -23     
- Partials      396      400      +4

Impacted Files	Coverage Δ
traits/trait_types.py	71.07% <100%> (-0.09%)	⬇️
traits/has_traits.py	70.72% <100%> (+0.33%)	⬆️
traits/editor_factories.py	81.25% <0%> (-3.13%)	⬇️
traits/traits.py	76.92% <0%> (+4.04%)	⬆️
traits/etsconfig/etsconfig.py	63.8% <0%> (+6.13%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2815c85...b0c1b98. Read the comment docs.

[mdickinson]

LGTM. The deprecation warning gets issued at instantiation time rather than subclassing time, which is fine. (We probably could find a way to issue it at subclassing time, but then we'd get DeprecationWarnings being issued in imports, which is always awkward to deal with.)