Add subsec_micros and subsec_millis methods to TimeDelta

[ggoetz]

This PR revives @botahamec 's PR #1351, which has been abandoned, and adds subsec_micros and subsec_millis to the public API for chrono::TimeDelta, for consistency with the API of std::time::Duration.

Summary of changes:

• update the docstring for TimeDelta::subsec_nanos() and add a test for the method.
• add subsec_millis and subsec_micros, along with unit tests.

Fixes: #1348

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.07%. Comparing base (2b7a28e) to head (0d4b934).
Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1668      +/-   ##
==========================================
+ Coverage   91.05%   91.07%   +0.01%     
==========================================
  Files          37       37              
  Lines       17402    17431      +29     
==========================================
+ Hits        15846    15875      +29     
  Misses       1556     1556              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[djc]

Honestly I'm not convinced by the argument that these consts should be made public just so two of them can make a single appearance in a docstring. IMO it would be fine to leave the constants alone and use literals to explain the methods.

[ggoetz]

Honestly I'm not convinced by the argument that these consts should be made public just so two of them can make a single appearance in a docstring. IMO it would be fine to leave the constants alone and use literals to explain the methods.

Makes sense to me -- and there was precedent with the previous docstring for subsec_nanos to only use a literal without exposing the corresponding const. I've reverted the commit that exposed the constants.

[djc]

Sorry, by literal I meant to replace NANOS_PER_SEC with 1_000_000_000 here (and same for the new methods). Other than that this seems fine!

(Maybe also update the PR description to describe the current approach.)

[ggoetz]

Updated the 3 comments and the PR description.