Deprecate broadcast_(get|set)index in favor of broadcasted (get|set)index

[mbauman]

Here's another simple deprecation I'd like to sneak in.

The performance here is comparable in my few quick tests, and only two packages used either of these functions -- both in situations where they would have gained much more by participating in dot-fusion.

julia> A = rand(5000,5000);
julia> @benchmark broadcast_getindex($A, 1:5000, 1:5000)
BenchmarkTools.Trial:
  memory estimate:  39.14 KiB
  allocs estimate:  2
  --------------
  minimum time:     34.535 μs (0.00% GC)
  median time:      35.510 μs (0.00% GC)
  mean time:        36.835 μs (1.72% GC)
  maximum time:     2.742 ms (97.61% GC)
  --------------
  samples:          10000
  evals/sample:     1

julia> @benchmark getindex.(($A,), 1:5000, 1:5000)
BenchmarkTools.Trial:
  memory estimate:  39.14 KiB
  allocs estimate:  2
  --------------
  minimum time:     31.564 μs (0.00% GC)
  median time:      32.596 μs (0.00% GC)
  mean time:        35.887 μs (2.17% GC)
  maximum time:     3.604 ms (0.00% GC)
  --------------
  samples:          10000
  evals/sample:     1

julia> @benchmark broadcast_setindex!($A, 1:5000, 1:5000, 1:5000)
BenchmarkTools.Trial:
  memory estimate:  0 bytes
  allocs estimate:  0
  --------------
  minimum time:     32.503 μs (0.00% GC)
  median time:      33.403 μs (0.00% GC)
  mean time:        33.432 μs (0.00% GC)
  maximum time:     136.444 μs (0.00% GC)
  --------------
  samples:          10000
  evals/sample:     1

julia> @benchmark setindex!.(($A,), 1:5000, 1:5000, 1:5000)
BenchmarkTools.Trial:
  memory estimate:  39.14 KiB
  allocs estimate:  2
  --------------
  minimum time:     35.955 μs (0.00% GC)
  median time:      37.788 μs (0.00% GC)
  mean time:        39.442 μs (2.17% GC)
  maximum time:     642.989 μs (80.10% GC)
  --------------
  samples:          10000
  evals/sample:     1

[JeffBezanson]

So nice!!

[Sacha0]

CI suggests that doc references to broadcast_getindex and broadcast_setindex! need excising?

[mbauman]

FreeBSD CI failure looks unrelated and is flagged as a low-probability failure.