similar(T, axes) does not accept colons

[jishnub]

Looks like this was missed out in #126 . This improves the error message (it's still a MethodError).

before:

julia> similar(Matrix{Float64}, (:, :))
ERROR: MethodError: no method matching Array{Float64,2}(::UndefInitializer, ::Tuple{Colon,Colon})

after:

julia> similar(Matrix{Float64}, (:, :))
ERROR: MethodError: no method matching similar(::Type{Array{Float64,2}}, ::Tuple{Colon,Colon})

[codecov]

Codecov Report

Merging #165 into master will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #165   +/-   ##
=======================================
  Coverage   99.21%   99.21%           
=======================================
  Files           4        4           
  Lines         256      256           
=======================================
  Hits          254      254           
  Misses          2        2           

Impacted Files	Coverage Δ
src/OffsetArrays.jl	99.43% <100.00%> (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 c841a99...2b4c73d. Read the comment docs.

[johnnychen94]

Looks nice. Perhaps you might also want to test the error message content, since this is what this PR does. For me it doesn't matter much as long as it errors...

[jishnub]

Ideally yes, but that might be going too far, don't want tests to break if Base changes the error message. This is type-piracy anyway

[jishnub]

There appears to be a way to test this by checking the function field of the raised MethodError and making sure that it's similar, but I wonder if it might be too internal? That's exactly what this PR does

[johnnychen94]

Probably just redirect the error io: https://github.com/JuliaGraphics/ColorTypes.jl/blob/bd31741d162361ebd44ed05ae532266998d9ce9f/test/error_hints.jl#L1-L15

(Your new avatar looks nice!)