proposal: sort: add InvertSlice

[codic12]

Golang should have an InvertSlice() function in some module (preferably sort), which essentially reverses all slice elements. This would allow some neat features like string reversal.

This is easy to implement and will not break code. Please consider the proposal.

[networkimprov]

I agree that it should be in the stdlib, both for strings and slices, e.g. sort.InvertSlice(slice)

Note the comparison in the loop can be simpler than i < len(r)/2:

   for a1, a2 := 0, len(S)-1; a1 < a2; a1, a2 = a1+1, a2-1 {
      S[a1], S[a2] = S[a2], S[a1]
   }

[ALTree]

This was rejected in the past: #14777. A few comments:

Even C and C++ have this

This is usually not considered a valid argument for inclusion of a new feature in the standard library. There are many things that other languages have, and Go has not.

This is easy to implement and will not break code.

This is not enough: the addition also needs to be useful enough to a great number of Go programmer to justify its inclusion in the standard library. As noted in #14777, how often would the typical Go programmer use Reverse (except when implementing it as a programming exercise)?

I'm closing here as a dup of #14777.

[networkimprov]

Pls reopen. #14777 was closed after 2.5h and a single comment by a Go team member, without hearing the author's use case. Proposals like this are now reviewed by committee.

Even if strings.Reverse() isn't accepted, the widely applicable sort.InvertSlice() should be considered, and that would allow string(sort.InvertSlice([]rune(s))).

[ianlancetaylor]

I'm not aware of a standard C function that reverses a string. Do you have a pointer? C++ has reverse iterators, but again I'm not aware of a function or method that reverses a string, unless you mean std::reverse which is a generic operation.

In Go, strings are a sequence of bytes but many of the functions in the strings package treat them as a sequence of characters. A reverse operation is inherently ambiguous: does it reverse bytes or characters?

[ianlancetaylor]

It's fine to make this a proposal, but it should be specific about exactly what strings.Reverse should do, and why that is better than the other choice.

[networkimprov]

@ianlancetaylor could you reopen the issue, as it's under discussion?

[codic12]

@ianlancetaylor for C++, std::reverse (from iostream.h or string.h, not sure). For C, I'm talking about strrev from string.h.
I am talking about reversing runes, aka characters.

[codic12]

and why that is better than the other choice.

which other choice?

This is not enough: the addition also needs to be useful enough to a great number of Go programmer to justify its inclusion in the standard library. As noted in #14777, how often would the typical Go programmer use Reverse (except when implementing it as a programming exercise)?

This is a useful core language feature, cryptography programs could benefit from it.

[networkimprov]

@gsbhasin123, a crypto algorithm would use []byte, not string. So I think you have a stronger case by proposing for package sort:

func InvertSlice(s interface{}) interface{} // returns its argument ('Reverse' is taken)

That enables a string inversion via string(sort.InvertSlice([]rune(str)).([]rune))

[codic12]

@networkimprov "package sort" meaning?
sorry, I'm new to Golang

[networkimprov]

https://golang.org/pkg/sort/

[codic12]

Ah.

[codic12]

I see, I thought you were talking about sorting packages lol.
An InvertSlice function would make this easy, yes