Avoid exception from Continuous scale for normed property with non-float data

[mwaskom]

Fixes #3106 in a narrow way by avoiding a TypeError when sorting out the normalization:

This result isn't ideal, we'd like to propagate the boolean-ness of the data through to the visualization, but this is consistent with other numpy / matplotlib behavior where boolean data are upcast to 0/1 floats in numeric contexts. So boolean data should still get a non-continuous scale by default but contra #3135 I think there should be a separate Boolean scale concept rather than default to Nominal.

[codecov]

Codecov Report

Merging #3190 (19c607b) into master (87e0972) will decrease coverage by 0.00%.
The diff coverage is 100.00%.

❗ Current head 19c607b differs from pull request most recent head 0df8277. Consider uploading reports for the commit 0df8277 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3190      +/-   ##
==========================================
- Coverage   98.42%   98.42%   -0.01%     
==========================================
  Files          76       76              
  Lines       24115    24114       -1     
==========================================
- Hits        23735    23734       -1     
  Misses        380      380              

Impacted Files	Coverage Δ
seaborn/_core/scales.py	90.09% <100.00%> (ø)
tests/_core/test_scales.py	100.00% <100.00%> (ø)
tests/_core/test_plot.py	98.68% <0.00%> (-0.01%)	⬇️
seaborn/_core/plot.py	99.33% <0.00%> (ø)