provide sub to override AA

[fieker]

No description provided.

[codecov]

Codecov Report

❌ Patch coverage is 18.18182% with 9 lines in your changes missing coverage. Please review.
✅ Project coverage is 88.21%. Comparing base (2a1d29b) to head (5f80e6c).
⚠️ Report is 31 commits behind head on master.

Files with missing lines	Patch %	Lines
src/matrix.jl	18.18%	9 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2194      +/-   ##
==========================================
+ Coverage   88.17%   88.21%   +0.03%     
==========================================
  Files          99       99              
  Lines       37177    37087      -90     
==========================================
- Hits        32782    32716      -66     
+ Misses       4395     4371      -24     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[thofma]

@lgoettgens we cleaned it up (similar to your suggestions).

[lgoettgens]

Looks good, thanks! Two comments below

[lgoettgens]

Note that AA defines https://github.com/Nemocas/AbstractAlgebra.jl/blob/a2325de25fea847b16780935803a30ee84582443/src/Matrix.jl#L476, i.e. the delegation goes from sub to getindex already. This means that this line is superfluous, and the other two special cases should be added to getindex instead.

Suggested change

	# make x[r, c] sugar for sub(x, r, c)
	getindex(x::_MatTypes, r::AbstractVector{Int}, c::AbstractVector{Int}) = sub(x, r, c)

[lgoettgens]

Suggested change

	function sub(x::$T, r::AbstractUnitRange{Int}, c::AbstractUnitRange{Int})
	return deepcopy(view(x, r, c))
	end
	function sub(M::$T, r::AbstractVector{Int}, c::AbstractVector{Int})
	function getindex(x::$T, r::AbstractUnitRange{Int}, c::AbstractUnitRange{Int})
	return deepcopy(view(x, r, c))
	end
	function getindex(M::$T, r::AbstractVector{Int}, c::AbstractVector{Int})

[lgoettgens]

Nevermind, my suggestions contradict with the implementations in Nemo. I think, the cleanest solution would be to switch the delegation direction in AA and then keep everything here as is (just removing the last two lines as they then coincide with AA)

[thofma]

Hm, yeah, there is also still something off. I agree that eventually we should swap it around in AA (make getindex just call sub). But I probably first want get this PR here in.

[fingolfin]

I'd like to make a "breaking" AA release soon (this week, before Christmas?) for other reasons. If someone can make a PR for the AA change discussed here "soonish", it could perhaps also get into that AA release?