crypto: implement Hash.String

[Ma124]

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

$ go version
go version go1.12.1 linux/amd64

Does this issue reproduce with the latest release?

yes

What operating system and processor architecture are you using (go env)?

go env Output

$ go env
GOARCH="amd64"
GOBIN=""
GOCACHE="/home/ma_124/.cache/go-build"
GOEXE=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOOS="linux"
GOPATH="/home/ma_124/.go"
GOPROXY=""
GORACE=""
GOROOT="/usr/local/go"
GOTMPDIR=""
GOTOOLDIR="/usr/local/go/pkg/tool/linux_amd64"
GCCGO="gccgo"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
GOMOD=""
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build821474858=/tmp/go-build -gno-record-gcc-switches"

What did you do?

I wanted to iterate all builtin hash functions by doing something like the following:

var algs map[string]crypto.Hash

var h crypto.Hash
for ; h < crypto.maxHash; h++ {
    if h.Available() {
        algs[h.String()] = h
    }
}

What did you expect to see?

An exported crypto.maxHash and crypto.Hash implementing the fmt.Stringer interface.

[andybons]

@FiloSottile

[andybons]

@Ma124 you mention what you want to do, but don’t explain why you or others would need this functionality. Could you elaborate a bit more on that in your proposal? What problem does this solve?

[Ma124]

The Stringer would be useful for all kinds of logging purposes and I personally see no drawback.

The maxHash would be useful every time you need to show something non-go the options there are for hashing. So for example every time a user can choose a hashing algorithm in a CLI or GUI or maybe you list available hashing options to a program written in another language. And your duplicating code every time when you use the alternative: importing the package and putting it into a map. So if for example one of the choices becomes insecure you need to remove the import and the map entry and if you forget one of them you've got either an unused import which cannot be detected by the compiler or you've got a panic.

So I think there are no reasons against Stringer and very few drawbacks of a renamed maxHash.

[FiloSottile]

I'm not against implementing crypto.Hash.String for logging and the like.

Since hash functions need to be imported anyway to be available, I don't see the advantage of maxHash, and it's leaking an internal detail. Also, I don't want to encourage "pick your own hash" designs. The set of available hashes in a design should be ideally of size 1, and anyway clearly specified, not "whatever Go supports".

[rsc]

Added to proposal minutes, seems headed for likely accept.

[rsc]

Based on the discussion above, this seems like a likely accept.

[rsc]

No change in consensus, so accepted.