Why is it needed to remove this test?

The test itself is incorrect because the 2 should be a 3 (not my best moment, this was). Additionally, this is patched in matplotlib 2.0.1, not matplotlib 2.0.0 as indicated here. So to keep the test around, we would need to modify when it skips.

More substantially, this bug was actually a bug in matplotlib, which I fixed and added a regression test for in their test suite. pandas doesn't really need to duplicate it.

Ah, the reason that it is 2 instead of 3 is the reason it now just passes, while the test should actually fail because it needs >= 2.0.1 instead of >= 2.0.0 ?

I think testing this in pandas is fine, as we should keep tight_layout working with pandas plots (we may change implementation details in the hist method that could cause problems with tight_layout)

Yes, exactly.

Hmm. OK, can do.

Codecov Report

Merging #15671 into master will decrease coverage by 0.02%.
The diff coverage is 66.66%.

@@            Coverage Diff             @@
##           master   #15671      +/-   ##
==========================================
- Coverage   91.03%   91.01%   -0.03%     
==========================================
  Files         143      143              
  Lines       49380    49386       +6     
==========================================
- Hits        44953    44947       -6     
- Misses       4427     4439      +12

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 32df1e6...99cd255. Read the comment docs.

@jreback @jorisvandenbossche Does this look good? Tests pass.

This would close #9351.

lgtm. @jorisvandenbossche

Thanks!