proposal: cmd/vet: vet should warn when time.Time type (or types embed it) is used as map keys.

[go101]

What version of Go are you using (go version)?

go version go1.9.2 linux/amd64

Does this issue reproduce with the latest release?

yes

What did you do?

package main

import "time"

type T struct {
	time.Time
}

var a map[T]string

func main() {
}

$ go vet a.go
$ [nothing reported]

What did you expect to see?

$ go vet a.go
$ time.Time is not a good map key type.

Like time.Time, there may be some other types, which are not capable to be used as map key types.

What did you see instead?

nothing reported

[bradfitz]

I don't think this meets the bar for a vet check.

It's perfectly valid to use a time.Time as a map key, if you know what you're doing. I think it would be too noisy for valid code.

But somebody could do some analysis on GitHub to see what percentage of time.Time-keyed maps are valid.

[richardartoul]

I wrote a linter that detects this type of issue if you're interested: https://github.com/m3db/build-tools/blob/master/linters/badtime/README.md

I'm trying to get it integrated with gometalinter right now.

[rsc]

Now that vet is part of go test, the bar is even higher for new checks. Because it is valid sometimes to have a time.Time as a map key, rejecting it in vet is not OK.

You might talk to @dominikh about adding this to megacheck or one of the other linters.