Add unit of year

[briochemc]

I think that expressing quantities in units of years is so ubiquitous in science that it would be nice to have it in the default Unitful.

FWIW I used 365.25 days = 31557600 seconds for the conversion like in UnitfulAstro.jl.

There may have been previous discussions about this but I could not find them. (Maybe they disappeared when Unitful moved to PainterQubits?)

[codecov-io]

Codecov Report

Merging #288 into master will increase coverage by 3.91%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #288      +/-   ##
==========================================
+ Coverage   78.13%   82.05%   +3.91%     
==========================================
  Files          15       15              
  Lines        1075     1198     +123     
==========================================
+ Hits          840      983     +143     
+ Misses        235      215      -20

Impacted Files	Coverage Δ
src/pkgdefaults.jl	33.92% <ø> (+13.52%)	⬆️
src/conversion.jl	84.21% <0%> (-0.41%)	⬇️
src/user.jl	94.44% <0%> (-0.13%)	⬇️
src/Unitful.jl	100% <0%> (ø)	⬆️
src/dimensions.jl	94.59% <0%> (+0.15%)	⬆️
src/units.jl	86.56% <0%> (+0.3%)	⬆️
src/fastmath.jl	82.53% <0%> (+2.53%)	⬆️
src/logarithm.jl	66.81% <0%> (+2.72%)	⬆️
src/quantities.jl	94.22% <0%> (+4.86%)	⬆️
src/range.jl	83.33% <0%> (+5.28%)	⬆️
... and 5 more

---

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 51a8650...1edf03b. Read the comment docs.

[briochemc]

I realize now that one might want to change the example exteding Unitful because it extends it with "yr".

[cstjean]

I also defined year to be 365.25 days, but wouldn't it make sense to keep it application-specific, since different domains might have different requirements?

Furthermore, if #125 is still planned, we would have 1year != Year(1).

Those are minor points though, I wouldn't mind.

[briochemc]

wouldn't it make sense to keep it application-specific, since different domains might have different requirements?

My opinion is that it is still better for (most?) Unitful users to have a default u"yr" available. If a specific application really needs something different then I believe it should precisely care about making its own precise definition for it, no?

[ajkeller34]

I think I agree with these changes. For calculations with physical quantities it seems like the most generically useful definition is the one in this PR. Other definitions could shadow if needed. Will close and reopen to retrigger CI.