proposal: net/url: add (Values).AddAll(key string, values []string)

[tomocy]

The current (Values).Add, you know, enables us to add one value to one key, but is less convenient sometimes, especially when you want to add multiple values to one key (etc: coping one Values to another Values).

func copyValues(dest, src url.Value) {
    for k, vs := range src {
        for _, v := range vs {
            dest.Add(k, v)
        }
    }
}

If (Values).AddAll is available, you can write like the blow.

func copyValues(dest, src url.Value) {
    for k, vs := range src {
        dest.AddAll(k, vs)
    }
}

It may make code clearer and improve it. (The code seems to be O(N).)

Is it good idea to (Values).AddAll to add multiple values to one key at the same time?

[OneOfOne]

vals[key] = append(vals[key], newkeys...)?

[Freeaqingme]

Could you perhaps elaborate how much this would improve a real-world application?

[rsc]

This is Add:

func (v Values) Add(key, value string) {
	v[key] = append(v[key], value)
}

It's only a convenience function; it would be fine to write the append.

It is also fine to write @OneOfOne's version when you have multiple values.
It would make sense to have a convenience function if this happened a lot but does it (@Freeaqingme's question)?

The complexity analysis in the top comment is not accurate. Assuming AddAll uses @OneOfOne's append, the loops are identical in runtime - both are O(N), because append is amortized O(1).

[tomocy]

I just found two cases of this in golang/go

• go/src/net/http/request.go

  Line 1166 in e7c51ca

  func copyValues(dst, src url.Values) {

• go/src/net/http/request_test.go

  Line 199 in 62bfa69

  for k, v := range initialFormItems {

[rsc]

Thanks @tomocy. I filed #33744 for the internal helper in net/http.
Internal helpers, of course, don't warrant external API.

Given the lack of evidence that this happens a lot, this seems like a likely decline.
Will leave open for a week for final comment period.

[rsc]

Marked this last week as likely decline w/ call for last comments (#33498 (comment)).
Declining now.