text/template: missingkey=zero behaves unexpectedly with map[string]any

[mjwhitta]

Go version

go1.23.6

Output of go env in your module/workspace:

GO111MODULE=''
GOARCH='amd64'
GOBIN=''
GOCACHE='/home/gopher/.cache/go-build'
GOENV='/home/gopher/.config/go/env'
GOEXE=''
GOEXPERIMENT=''
GOFLAGS=''
GOHOSTARCH='amd64'
GOHOSTOS='linux'
GOINSECURE=''
GOMODCACHE='/home/gopher/.go/v1.23.6/pkg/mod'
GONOPROXY=''
GONOSUMDB=''
GOOS='linux'
GOPATH='/home/gopher/.go/v1.23.6'
GOPRIVATE=''
GOPROXY='https://proxy.golang.org,direct'
GOROOT='/home/gopher/.cache/gvm/use'
GOSUMDB='sum.golang.org'
GOTMPDIR=''
GOTOOLCHAIN='auto'
GOTOOLDIR='/home/gopher/.cache/gvm/use/pkg/tool/linux_amd64'
GOVCS=''
GOVERSION='go1.23.6'
GODEBUG=''
GOTELEMETRY='local'
GOTELEMETRYDIR='/home/gopher/.config/go/telemetry'
GCCGO='gccgo'
GOAMD64='v1'
AR='ar'
CC='gcc'
CXX='g++'
CGO_ENABLED='1'
GOMOD='/dev/null'
GOWORK=''
CGO_CFLAGS='-O2 -g'
CGO_CPPFLAGS=''
CGO_CXXFLAGS='-O2 -g'
CGO_FFLAGS='-O2 -g'
CGO_LDFLAGS='-O2 -g'
PKG_CONFIG='pkg-config'
GOGCCFLAGS='-fPIC -m64 -pthread -Wl,--no-gc-sections -fmessage-length=0 -ffile-prefix-map=/tmp/go-build1773965080=/tmp/go-build -gno-record-gcc-switches'

What did you do?

package main

import (
	"os"
	"text/template"
)

var tmpl string = "var = \"{{ .var }}\"\n"

func main() {
	t, _ := template.New("t").Option("missingkey=zero").Parse(tmpl)
	t.Execute(os.Stdout, map[string]bool{})

	t, _ = template.New("t").Option("missingkey=zero").Parse(tmpl)
	t.Execute(os.Stdout, map[string]int{})

	t, _ = template.New("t").Option("missingkey=zero").Parse(tmpl)
	t.Execute(os.Stdout, map[string]string{})

	t, _ = template.New("t").Option("missingkey=zero").Parse(tmpl)
	t.Execute(os.Stdout, map[string][]string{})

	t, _ = template.New("t").Option("missingkey=zero").Parse(tmpl)
	t.Execute(os.Stdout, map[string][]string{"var": nil})

	t, _ = template.New("t").Option("missingkey=zero").Parse(tmpl)
	t.Execute(os.Stdout, map[string]any{})

	t, _ = template.New("t").Option("missingkey=zero").Parse(tmpl)
	t.Execute(os.Stdout, map[string]any{"var": nil})
}

What did you see happen?

var = "false"
var = "0"
var = ""
var = "[]"
var = "[]"
var = "<no value>"
var = "<no value>"

What did you expect to see?

var = "false"
var = "0"
var = ""
var = "[]"
var = "[]"
var = ""
var = ""

[ianlancetaylor]

The text/template package prints a nil value of type any as "<no value>". You will see the same thing for a struct with a field of type any whose value is nil. I don't think that we can change that behavior now; it seems likely that it would break existing use cases.

[gabyhelp]

Related Issues

• text/template: missingkey=zero behaves different from html/template #70681 (closed)
• text/template: missingkey=zero ignored for map[string]interface{} #24963 (closed)
• text/template: key/value assignment is reversed within range loop #60801 (closed)
• text/template: missingkey option lost when cloning a template #60100 (closed)
• text/template: Unexpected assignment #56490 (closed)
• html/template: Unexpected result #58460 (closed)
• the strange first line of template is blank #67194 (closed)
• text/template: pointer to empty slice/map considered truthy #51816 (closed)
• text/template: handling of `nil` keyword #58326 (closed)
• map: a key-value pair may be GC #38976 (closed)

_{(Emoji vote if this was helpful or unhelpful; more detailed feedback welcome in this discussion.)}

[mjwhitta]

I think it would help if the existing documentation at least mentioned this behavior, so that users don't waste time debugging. If I had searched for map[string]interface in the issues, it looks like I would have found the details I needed. But GitHub issues should not be a replacement for documentation. A simple sentence mentioning nil values will still output <no value> is probably good enough. Alternatively, could we not get something like missingkey=zeronil with new functionality? This would not break existing use cases.

[ianlancetaylor]

The behavior isn't really related to missingkey or maps. It's how text/template always handles nil values of type any.

Perhaps the docs should have a section on how values are printed.