bytes, strings: add ContainsFunc

[itchyny]

I propose to add ContainsFunc to package strings and bytes.

// ContainsFunc reports whether any Unicode code points c within s satisfy f(c).
func ContainsFunc(s string, f func(rune) bool) bool {
	return IndexFunc(s, f) >= 0
}

The {bytes,strings}.ContainsFunc functions replace codes which only check whether the string contains a Unicode code point satisfying the condition or not using {bytes,strings}.IndexFunc.

In the Go repository, there are 6 codes (out of 9 using $PKG.IndexFunc(...)) to be replaced by the proposed ContainsFunc functions.

 $ semgrep --quiet --lang=go --pattern='$PKG.IndexFunc(...) >= 0'

Findings:

  src/archive/tar/reader.go
        395┆ if bytes.IndexFunc(tr.blk[:], notASCII) >= 0 {


  src/testing/benchmark.go
        353┆ if strings.IndexFunc(unit, unicode.IsSpace) >= 0 {

 $ semgrep --quiet --lang=go --pattern='$PKG.IndexFunc(...) < 0'

Findings:

  src/mime/grammar.go
         31┆ return strings.IndexFunc(s, isNotTokenChar) < 0


  src/net/http/cookie.go
        465┆ return strings.IndexFunc(raw, isNotToken) < 0


  src/text/template/funcs.go
        725┆ if strings.IndexFunc(s, jsIsSpecial) < 0 {

 $ semgrep --quiet --lang=go --pattern='$PKG.IndexFunc(...) == -1'

Findings:

  src/net/http/request.go
        831┆ return len(method) > 0 && strings.IndexFunc(method, isNotToken) == -1

In the Kubernetes repository, I find 2 codes (out of 3 using $PKG.IndexFunc(...)).

 $ semgrep --quiet --lang=go --pattern='$PKG.IndexFunc(...) != -1'

Findings:

  staging/src/k8s.io/client-go/rest/request.go
       1068┆ if bytes.IndexFunc(body, func(r rune) bool {
       1069┆    return r < 0x0a
       1070┆ }) != -1 {


  staging/src/k8s.io/kubectl/pkg/describe/describe.go
        349┆ if strings.IndexFunc(field, func(r rune) bool {
        350┆    return !unicode.IsLetter(r) && r != '-'
        351┆ }) != -1 {

Using ContainsFunc in these if statements will definitely improve the code readability. Also, I believe API symmetry reduces the frustration of coding (I cannot explain why there are Contains, ContainsAny, ContainsRune but not ContainsFunc to Go newbies). Currently, we need to translate what we actually want to check into comparison of returned index against numbers (which may lead to pointless discussion of which is better >= 0 or != -1).

[earthboundkid]

Cf #53983

[rsc]

This was waiting on slices.ContainsFunc. Now that we've accepted that proposal, it seems fine to also have bytes.ContainsFunc and strings.ContainsFunc.

[rsc]

Based on the discussion above, this proposal seems like a likely accept.
— rsc for the proposal review group

[rsc]

No change in consensus, so accepted. 🎉
This issue now tracks the work of implementing the proposal.
— rsc for the proposal review group

[gopherbot]

Change https://go.dev/cl/460216 mentions this issue: bytes, strings: add ContainsFunc