Allow mismatched element types for weighted means

[ararslan]

We've recently inadvertently disallowed a mismatch between the array's element type and the element type of the weights when computing a weighted mean. This leads to bizarre behavior, as it dispatches to a nonsensical method where the array is treated as a function.

Fixes #475

[codecov]

Codecov Report

Merging #476 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master     #476   +/-   ##
=======================================
  Coverage   85.11%   85.11%           
=======================================
  Files          21       21           
  Lines        2083     2083           
=======================================
  Hits         1773     1773           
  Misses        310      310

Impacted Files	Coverage Δ
src/weights.jl	78.26% <100%> (ø)	⬆️

---

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 5c9ca27...b570162. Read the comment docs.

[codecov]

Codecov Report

Merging #476 into master will not change coverage.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master     #476   +/-   ##
=======================================
  Coverage   85.11%   85.11%           
=======================================
  Files          21       21           
  Lines        2083     2083           
=======================================
  Hits         1773     1773           
  Misses        310      310

Impacted Files	Coverage Δ
src/weights.jl	78.26% <100%> (ø)	⬆️

---

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 5c9ca27...b570162. Read the comment docs.

[nalimilan]

Good catch. But AFAICT the problem isn't the type mismatch, it's just that #442 added the T<:Number requirement (which previously applied only when dims was provided, which didn't really make sense).

[ararslan]

But AFAICT the problem isn't the type mismatch ...

Oh yeah true. We can just change the commit message on merge to make it more accurate.