fix a[..] .= something
#27
Conversation
…ntained already in test/Project.toml)
| # avoid copying if indexing with .. alone, see | ||
| # https://github.com/JuliaDiffEq/OrdinaryDiffEq.jl/issues/214 | ||
| @inline Base.getindex(A::AbstractArray, ::Ellipsis) = A | ||
| @inline ArrayInterface.getindex(A, ::Ellipsis) = A |
There was a problem hiding this comment.
is the timing okay without this?
There was a problem hiding this comment.
The issue mentioned there is still open. We discussed streamlining the interpolation codes using .., but didn't implement it (as far as I remember). The only difference will be when calling A[..]. If it's on the left-hand side, it can subtly break (as fixed in this PR). If it's on the right-hand side, it's an indexing operation anyways and should be consistent in creating a copy of the data. So I think this should be fine.
There was a problem hiding this comment.
Do you have some specific tests in mind?
There was a problem hiding this comment.
What does it do without these dispatches? Does it make a copy?
There was a problem hiding this comment.
Yes, without these definitions A[..] behaves as A[:].
I noticed that ab assignment of the form
a[..] .= somethingdid not work. Hence, I deleted the offendinggetindexoverload in EllipsisNotation. The comment there referred to a still open issue that does not seem to be active right now.If these changes are okay, it would be nice to have a new release.