slice.Remove

[MatthewEdge]

• Adds slice.Remove as a safer variant of the slices.Delete() function that avoids the panic case for invalid indecies

This function does fall within the warning about potential memory leaks if the slices modified are long-lived. See slices.Delete() docs for more info

Implementation Notes

• Remove does use the same in-place approach go/x/exp uses to achieve the in-place update.

Benchmarking Checks

On an Intel Mac, to gut check the no-allocation claim:

goos: darwin
goarch: amd64
pkg: github.com/mailgun/holster/v4/slice
cpu: Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
BenchmarkRemove-12       	242117806	         4.944 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	223956164	         4.946 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	244528929	         4.919 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	244414112	         4.878 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	244394427	         5.028 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	229395693	         4.942 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	243575570	         4.853 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	244754611	         5.171 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	234260016	         5.056 ns/op	       0 B/op	       0 allocs/op
BenchmarkRemove-12       	246132768	         4.855 ns/op	       0 B/op	       0 allocs/op

[vtopc]

It's just slices package already - https://pkg.go.dev/slices#Delete

[MatthewEdge]

With one key difference: Delete panics if s[i:j] is not a valid slice of s <- this function does not panic. That's what I was trying to get across with this comment. That snippet is pulled from that link btw

[vtopc]

I just meant that go/x/exp/slices and build-in slices are different packages.

[MatthewEdge]

Ahhh I see. I misread your comment, apologies!