Closed lkdvos closed 6 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Comparison is base (
1a19cd2
) 82.00% compared to head (aa342d9
) 82.69%.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Thanks. In principle floating point numbers are not allowed as labels:
julia> TensorOperations.isindex(3.0)
false
but clearly this is not checked by the add_indices
function 😄 . Can I merge or do you want to change the test in order not to confuse people?
I've changed the tests to improve clarity, although now I am wondering if these functions should assert that the labels are valid? I think @assert
s are disabled when optimisation is set to high, so this should not impact performance?
[edit] currently @assert
only states in its documentation that it may be optimized away, but I don't think this actually happens. With that in mind, probably the best option is to merge this as-is, and not do actual checking in these methods. I don't think there is enough use to warrant a rewrite in terms of unsafe_
variants.
Some simple tests to improve test coverage