Fix histplot/kdeplot normalization with weights

[mwaskom]

Fixes #2655

With the example from that issue:

f, axs = plt.subplots(2, 2, constrained_layout=True, figsize=(7, 6))
params = dict(x=[1, 2, 3], weights=[1, 2, 2], discrete=True)
sns.histplot(**params, ax=axs.flat[0])
sns.histplot(**params, hue=["a", "b", "b"], ax=axs.flat[1])
sns.histplot(**params, hue=["a", "b", "b"], stat="proportion", ax=axs.flat[2])
sns.histplot(**params, hue=["a", "b", "b"], stat="proportion", common_norm=False, ax=axs.flat[3])

[codecov]

Codecov Report

Merging #2812 (603fa57) into master (0550195) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #2812   +/-   ##
=======================================
  Coverage   97.96%   97.96%           
=======================================
  Files          67       67           
  Lines       21719    21745   +26     
=======================================
+ Hits        21276    21302   +26     
  Misses        443      443           

Impacted Files	Coverage Δ
seaborn/distributions.py	96.27% <100.00%> (+0.04%)	⬆️
seaborn/tests/test_distributions.py	99.79% <100.00%> (+<0.01%)	⬆️

---

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 0550195...603fa57. Read the comment docs.