Support ball in at origin with autodiff

[dlfivefifty]

norms with ForwardDiff's Dual are flaky:

julia> norm([Dual(0.0, (1.0,0.0)), Dual(0.0,(0.0,1.0))])
Dual{Nothing}(NaN,NaN,NaN)

This works around this issue.

[dlfivefifty]

Note approx_in will also have the same issue but it's less clear how to resolve it, if tolerance is meant literally....

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.86%. Comparing base (f095869) to head (2bc720d).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #184   +/-   ##
=======================================
  Coverage   93.86%   93.86%           
=======================================
  Files          27       27           
  Lines        1843     1843           
=======================================
  Hits         1730     1730           
  Misses        113      113           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[daanhb]

If this works it is worth tagging a new release, there have been some other minor things.

[dlfivefifty]

Yes it fixes my issue.

There's a small chance that there is a use-case where a type has norm overloaded which this won't pick up. But we can burn that bridge when we get to it.

The patch version is already incremented so I'll merge and tag.