Add timestamp_nanos_opt, deprecate timestamp_nanos

[pitdicker]

NaiveDateTime::timestamp_nanos and DateTime::timestamp_nanos are two of the few methods that can panic on somewhat reasonable values and have no alternative that returns Option.

I added timestamp_nanos_opt and went ahead and deprecated the the old methods, like all other fallible methods.

The new method lead to a nice little simplification in duration_round and duration_trunc.

[codecov]

Codecov Report

Merging #1275 (7d20d22) into 0.4.x (77d0bec) will decrease coverage by 0.03%.
Report is 6 commits behind head on 0.4.x.
The diff coverage is 88.23%.

@@            Coverage Diff             @@
##            0.4.x    #1275      +/-   ##
==========================================
- Coverage   91.40%   91.37%   -0.03%     
==========================================
  Files          38       38              
  Lines       16932    16957      +25     
==========================================
+ Hits        15477    15495      +18     
- Misses       1455     1462       +7     

Files Changed	Coverage Δ
src/naive/datetime/mod.rs	97.34% <42.85%> (-0.37%)	⬇️
src/datetime/mod.rs	86.08% <100.00%> (-0.15%)	⬇️
src/datetime/serde.rs	87.56% <100.00%> (+0.06%)	⬆️
src/naive/datetime/serde.rs	85.65% <100.00%> (+0.08%)	⬆️
src/naive/datetime/tests.rs	98.63% <100.00%> (-0.03%)	⬇️
src/round.rs	97.72% <100.00%> (+0.14%)	⬆️

... and 3 files with indirect coverage changes

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

[djc]

For all of these, why not yield errors here?

(Let's not have a . in expect() messages.)

[pitdicker]

Good point, they should just return an error.