Fix *_DAYS_FROM_YEAR_0 calculation

[pitdicker]

The MIN_DAYS_FROM_YEAR_0 and MAX_DAYS_FROM_YEAR_0 constants are only used in tests.

As described in #1048 there is something subtly wrong in the derivation of them. But I was not smart enough to put my finger on it 😄.

The formula was only correct if MIN_YEAR was a leap year. I changed it to a correct formula that tries to be less smart, and tested it to be correct for all possible years.

test_date_bounds helps to verify NaiveDate::MIN and NaiveDate::MAX are correct. I changed it a little to also report the last ordinal day in NaiveDate::MAX.

Split out from #1048.

[codecov]

Codecov Report

Merging #1312 (913c502) into 0.4.x (a47e0e3) will decrease coverage by 0.01%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##            0.4.x    #1312      +/-   ##
==========================================
- Coverage   91.24%   91.23%   -0.01%     
==========================================
  Files          38       38              
  Lines       17062    17063       +1     
==========================================
  Hits        15568    15568              
- Misses       1494     1495       +1     

Files Changed	Coverage Δ
src/naive/date.rs	96.20% <0.00%> (-0.05%)	⬇️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more