Bugfix in overflow, CartesianIndices for BigInt ranges

[jishnub]

There was a bug introduced by #226 , this PR fixes that.

On master:

julia> OffsetVector(big(1):big(2), 0)
ERROR: MethodError: no method matching overflow_check(::Base.OneTo{BigInt}, ::Int64)

After this PR:

julia> OffsetVector(big(1):big(2), 0)
1:2 with indices 1:2

Also fixes a bug in CartesianIndices caused by a missing method for IdOffsetRange. This is broken on master anyway because of the previous bug, so we need to fix that to see this. After adding the fix:

julia> a = OffsetArray(big(1):big(2), 0);

julia> CartesianIndices(a)
ERROR: MethodError: no method matching AbstractUnitRange{Int64}(::OffsetArrays.IdOffsetRange{BigInt,Base.OneTo{BigInt}})

After this PR:

julia> CartesianIndices(a)
2-element CartesianIndices{1,Tuple{OffsetArrays.IdOffsetRange{Int64,Base.OneTo{Int64}}}} with indices 1:2:
 CartesianIndex(1,)
 CartesianIndex(2,)

[codecov]

Codecov Report

Merging #227 (f9479f4) into master (b33bc00) will decrease coverage by 4.70%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #227      +/-   ##
==========================================
- Coverage   99.04%   94.33%   -4.71%     
==========================================
  Files           5        5              
  Lines         314      318       +4     
==========================================
- Hits          311      300      -11     
- Misses          3       18      +15     

Impacted Files	Coverage Δ
src/OffsetArrays.jl	98.12% <100.00%> (-0.47%)	⬇️
src/axes.jl	100.00% <100.00%> (ø)
src/precompile.jl	0.00% <0.00%> (-100.00%)	⬇️

---

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 b33bc00...f9479f4. Read the comment docs.

[jishnub]

Test failure on nightly is probably a breakage. JuliaLang/julia#40035

Can be fixed by defining the conversion of an IdOffsetRange to an OrdinalRange

[jishnub]

4eebd90 fixes overflow checks to use the axes and not the offset (which is usually an Int). This means that we may skip overflow checks for BigInt axes.

Before:

julia> OffsetArray(1:big(2)^65, 4000)
ERROR: OverflowError: offset should be <= -27670116110564327425 corresponding to the axis Base.OneTo(36893488147419103232), received an offset 4000

Now:

julia> OffsetArray(1:big(2)^65, 4000)
1:36893488147419103232 with indices 4001:36893488147419107232

[jishnub]

Whoops I committed the offsetranges.jl file by mistake

[johnnychen94]

If I understand it correctly, except for the compat patch for JuliaLang/julia#40038, this PR is ready? So far it looks good to me.

[jishnub]

Yes this is ready

[johnnychen94]

Perhaps add a comment to JuliaLang/julia#40038 so that people know this can be removed in future Julia versions.