cmd/go: clarify best practice for tool dependencies

[myitcv]

Please answer these questions before submitting your issue. Thanks!

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

go version go1.10.3 linux/amd64
vgo commit 22e23900224f03be49670113d5781e4d89090f45

Does this issue reproduce with the latest release?

Yes; and latest vgo commit (per above)

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

GOARCH="amd64"
GOBIN="/tmp/tmp.VQw1O3x8Wy/hello/bin"
GOCACHE="/home/myitcv/.cache/go-build"
GOEXE=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOOS="linux"
GOPATH="/tmp/tmp.VQw1O3x8Wy"
GORACE=""
GOROOT="/home/myitcv/gos"
GOTMPDIR=""
GOTOOLDIR="/home/myitcv/gos/pkg/tool/linux_amd64"
GCCGO="gccgo"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
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-build414355570=/tmp/go-build -gno-record-gcc-switches"

What did you do?

Off the back of #25624 (comment), I'd like to confirm that the following represents the "best practice" advice when adding and installing tool dependencies:

cd `mktemp -d`
export GOPATH=$PWD

mkdir hello
cd hello
vgo mod -init -module example.com/hello

# this could be anywhere but for convenience...
export GOBIN=$PWD/bin

# add a dependency on golang.org/x/tools/cmd/stringer
cat <<EOD > tools.go
// +build tools

package tools

import (
        _ "golang.org/x/tools/cmd/stringer"
)
EOD

vgo install golang.org/x/tools/cmd/stringer

The go.mod and .Target for stringer look fine:

$ cat go.mod
module example.com/hello

require golang.org/x/tools v0.0.0-20180615195736-465e6f399236
$ vgo list -f "{{.Target}}" golang.org/x/tools/cmd/stringer
/tmp/tmp.VQw1O3x8Wy/hello/bin/stringer

The issue however is that running vgo mod -sync then removes our module requirement on golang.org/x/tools - I suspect this is a bug:

$ vgo mod -json
{
        "Module": {
                "Path": "example.com/hello",
                "Version": ""
        },
        "Require": [
                {
                        "Path": "golang.org/x/tools",
                        "Version": "v0.0.0-20180615195736-465e6f399236"
                }
        ],
        "Exclude": null,
        "Replace": null
}
$ vgo mod -sync
warning: "ALL" matched no packages
$ vgo mod -json
{
        "Module": {
                "Path": "example.com/hello",
                "Version": ""
        },
        "Require": null,
        "Exclude": null,
        "Replace": null
}

If we assume this is a bug and ignore it for now, I also wonder whether we can improve this workflow for adding tool dependencies somehow. The following steps feel a bit "boilerplate" and unnecessary:

• define a fake tools.go file
• add a dependency on the tool (which will never compile because it's a main package, so I'm not sure we can ever safely verify tools.go is "good"?)
• set GOBIN to an appropriate location
• vgo install tool
• ensure PATH includes GOBIN

I wonder whether we could in fact obviate all of this by having something like:

vgo install -tool golang.org/x/tools/cmd/stringer
vgo run golang.org/x/tools/cmd/stringer

Thoughts?

vgo run tool is possible as a result of #22726, but because of #25416 it effectively requires a link step each time.

What did you expect to see?

With respect to what I think is a bug with vgo mod -sync

go.mod unchanged by the vgo mod -sync

What did you see instead?

The golang.org/x/tools requirement removed.

/cc @rsc @bcmills

[bcmills]

Let's take a step back: what do you mean by “tool dependency”?

The stringer example suggests that you mean “a tool required for use with go generate”, but note that generate is intended to be run by the author of the module, not its users: a tool needed to run generate shouldn't be a hard requirement for the module or its tests.

[myitcv]

Let's take a step back: what do you mean by “tool dependency”?

Roughly:

• I'm writing a module; or, more generally, me and my team are writing a module
• we want to use a tool (e.g. stringer) whilst writing that code
• we want to reproducibly re-generate code, or put another way, we want the whole team to consistently generate the same code at any given point in time
• hence I want to ensure that everyone is using the same version of any tool we require

but note that generate is intended to be run by the author of our module, not its users: a tool needed to run generate shouldn't be a hard requirement for the module or its tests.

Correct, this would only be a dependency for the authors of the module (as you point out, the users of our module probably don't care whether or not we've used stringer or other tools).

As I understand it, these tool dependencies do not introduce any additional requirements for users of our module: they (the tool dependencies) can only ever be referenced by running vgo install or vgo run in the context of our module. So as a user of our module, I would, by definition be in a different context (specifically the module that references our module) when running vgo install or vgo run.

I'm using the term "context" here to mean "the go.mod that vgo uses." I'm sure there must be a better, more precise term

[jimmyfrasche]

I haven't played around with vgo much yet or looked under the hood.

I know a lot of package managers handle this by sorting dependencies into buckets: regular dependencies, test dependencies, dev dependencies. That seems like overkill and there can also be dependencies by build tag, like, say, importing a module for internal use but only on windows, and dependencies may only be for a certain package in a module that you may not be using.

Could vgo record all the dependencies for a project—for all build tags, for all packages, for tests, for development—but lazily fetch them as needed? (This may require more invasive integration but it would be an optimization as the semantics are the same for lazy-fetch and fetch-everything-upfront)

[myitcv]

@jimmyfrasche

Could vgo record all the dependencies for a project—for all build tags, for all packages, for tests, for development—but lazily fetch them as needed?

My understanding is that this is exactly what vgo does today, for any dependency, test/tool or otherwise.

And, to your earlier point, I agree it means the delineation between the buckets is unnecessary (for anyone wanting to know why a dependency exists, vgo can tell us that answer).

Just to clarify one point above. My proposal for vgo install -tool golang.org/x/tools/cmd/stringer would, however, add a section to go.mod (or equivalent) where the tool packages are listed. This is simply to avoid the need for the tools.go file referenced in my example.

[bcmills]

A couple of observations:

1. Some tools required by Go builds are not themselves built using the go tool. (For example, Google's protocol compiler is written in C++; only the Go code generator plugin is written in Go.) So even if we add support for tooling to the go tool, it will be at best a partial solution.

2. If you are assuming the use of the go tool anyway, then you can write a Go program that invokes it with appropriate arguments.

3. Because a command-line tool is a top-level package, its versions are not context-sensitive: if you're not mixing in other packages to customize the tool, you don't need to resolve version constraints from those packages.

To me, those suggest a fairly straightforward solution: go run pkg@version. In combination with //go:generate, that would allow you to express the dependency (at the exact version used) directly in the source file:

//go:generate go run golang.org/x/tools/cmd/[email protected] -type=Pill

In contrast, if you are building a custom version of the tool, then you should be able to customize it using blank-imports as you describe.

[jimmyfrasche]

@bcmills that sounds good but it would be awkward to write that each time and easy for things to go out of sync and you have some code using 1.11.0 and some using 1.12.3 but probably want .

go generate let's you define aliases with //go:generate -command alias the-command (I've not used it personally and unsure of the scope: file? package? (I assume file.))

Perhaps a more vgo-aware variant would be helpful, like

//go:generate -require golang.org/x/tools/cmd/[email protected]

which is module scoped. It falls somewhere between an ephemeral go install -tool and a fictitious blank import.

There are a some details to iron out in such an approach, of course.

[myitcv]

@bcmills

Some tools required by Go builds are not themselves built using the go tool.

Completely. As you say, this can either be solved entirely orthogonally, or with Go wrappers (wrappers that could even take care of installing the right version).

Because a command-line tool is a top-level package, its versions are not context-sensitive: if you're not mixing in other packages to customize the tool, you don't need to resolve version constraints from those packages.

I'm not entirely sure I follow your point here. The point I was trying to make is that where (i.e. in what directory) you run vgo install golang.org/x/tools/cmd/stringer is relevant because that will determine which go.mod you reference and therefore which version of the golang.org/x/tools module you use.

To me, those suggest a fairly straightforward solution: go run pkg@version.

As @jimmyfrasche pointed out this has the major downside of littering the required version across all the places you require go:generate directives, which seems to be somewhat counter to having go.mod be the single point of reference for versions.

Just one point to note about the "proposal":

vgo install -tool golang.org/x/tools/cmd/stringer
vgo run golang.org/x/tools/cmd/stringer

Despite the explanation in #25416, I think vgo run golang.org/x/tools/cmd/stringer could actually cache the linked binary, because the cache can very precisely be trimmed given we know exactly what version is being referenced at a given point in time.

[jimmyfrasche]

Running a tool at a specific version depending on the module would have to go through vgo.

Running stringer lets $SHELL pick which stringer to run, so it would need to be something like go run-dev-tool stringer so that vgo could dispatch the correct version of the command.

Something like that could be stuffed into go run but that doesn't seem to fit with the spirit of that command. Arguably go tool would be a better fit but that could cause namespacing issues so it seems like a separate command built to purpose would be best.

go generate could handle this transparently but not all tools are necessarily for code generation. You may also want a specific version of, say, a linter.

An IDE/editor could use tool dependencies to figure out if this is a tool dependency or a regular program and run it through vgo if need be. From the command line, though, it would be up to the user to know whether to run stringer vs go run-dev-tool stringer so there would still be potential skew with the version documented. I don't think there's much to do about that, though.

A tool module could have multiple commands. Look up would be slow if these weren't included somewhere in the metadata.

A tool module could have multiple dependencies, some shared with the module under development. Unlike with other dependencies, you probably don't want those to be solved together: using foobar shouldn't affect your version baz or the version of baz used by other tools you depend on. They should all be siloed.

This is all starting to sound mostly unrelated to vgo but perhaps it's something that can be built on top of vgo.

Going back on what I said about buckets, let's say there was as separate tool godev and a separate file dev.mod.

godev install foo updates dev.mod recording the version and the exposed commands. Instead of installing the command globally it installs the binaries outside of $PATH but somewhere it can easily grab the correct version. It updates dev.mod with the vgo style import and version path but also records the names of the commands. For the most part it's a wrapper around go get that does some extra work.

godev run foo invokes the appropriate foo from dev.mod.

Is there anything that would need to change in vgo to allow something like this?

A downside of not having it integrated in vgo itself would be that go:generate directives would need to explicitly prefix godev run.

[rsc]

What is the decision to be made here?
I think the tools.go file is in fact the best practice for tool dependencies, certainly for Go 1.11.
I like it because it does not introduce new mechanisms.
It simply reuses existing ones.

[myitcv]

I think the tools.go file is in fact the best practice for tool dependencies, certainly for Go 1.11.

Indeed, we don't need to "solve" this now.

What is the decision to be made here?

For Go 1.11

• is the vgo mod -sync behaviour observed in the description a bug or not? It feels like it is.

For later:

• is there a better way to ensure the fake tools.go is valid? Because it will never compile. We could do something with vgo list and a bit of shell, but it feels like we're side stepping the issue
• is there a way to reduce the "boilerplate" setup described in the issue, to make the experience for module maintainers nicer?

[myitcv]

Adding another comment to this thread to give the equivalent, and current, go commands

This is a go-based rewrite of #25922 (comment)

cd $(mktemp -d)
mkdir hello
cd hello
go mod init example.com/hello

# Either rely on GOBIN=GOPATH/bin or set it explicitly
export GOBIN=$PWD/bin

# add a dependency on golang.org/x/tools/cmd/stringer
# the build constraint ensures this file is ignored
cat <<EOD > tools.go
// +build tools

package tools

import (
        _ "golang.org/x/tools/cmd/stringer"
)
EOD

go install golang.org/x/tools/cmd/stringer

cat go.mod

results in:

module example.com/hello

require golang.org/x/tools v0.0.0-20180813205110-a434f64ace81 // indirect

[joeshaw]

Similarly, converting an existing project from dep to the modules system and running go mod vendor results in the stringer tool and its immediate dependencies being removed from the vendor directory.

Sorry, this is actually working. My tools.go file was missing the package directive at the top of the file and because it's never built the compiler never got a chance to complain. This, ironically, is one of @myitcv's concerns in his initial comment.

[rsc]

Best practice remains creating the tools.go file.
It's true that go mod init does not auto-create a tools.go from dep's config,
but I think doing so is getting a bit beyond scope.

[AlekSi]

Probably it's worth to highlight why tools.go should contain some build tag like tools (but not ignore): to avoid side-effects of init() functions in tool dependencies.

[thepudds]

A related comment: as far as I am aware, creating a tools.go as described above in #25922 (comment) seems to work with go mod vendor for Go-based tool dependencies.

In addition, there is also this separate comment in #26244 (comment) :

vgo vendor (now vgo mod -vendor or go mod -vendor) does not copy entire repositories. By design, it copies only the directories needed to satisfy imports in builds of the top-level module. If your module is not importing github.com/antlr/grammars-v4 then that directory will not be copied. And if there are directories with only non-Go files, there is no way to copy those directories into vendor. That's not what vendor is for. Vendor is only about preserving the code needed to build the main module

...which on the one hand seems to say go mod vendor is not intended to cover all use cases around non-Go pieces of a repository...

...but that comment also seems to suggest that if there is a piece of Go code that will be vendored, the operation seems to function at the granularity of directories (e.g., it "copies only the directories needed to satisfy imports in builds of the top-level module").

In practice, this currently seems to mean that non-Go tool dependencies (such as a shell script) are copied into the vendor directory by go mod vendor if they are in the same directory as a vendored piece of Go code. So a tools.go-based approach could in theory be extended to pulling in non-Go based tool dependencies (which would happen naturally if the non-Go tools are intermixed at the directory level with Go code, or perhaps even by adding something like a vendorme.go to a particular directory if needed).

However, as far as I can see, the go mod vendor documentation currently does not say it operates at the directory level, so perhaps that is behavior that should not be relied upon, or perhaps this would otherwise be considered an undesirable approach.

EDIT: note that a fair amount of care is required if you are pulling in additional directories beyond what go mod vendor does naturally, and this is less about vendoring behavior and more about following the go tool's build model. See for example comment from Russ in #26366 (comment), as well as https://golang.org/cl/125297, and #26366 (comment))

[caarlos0]

I've tried the solution proposed by @myitcv, it more or less works, seems like stringer can't solve the dependencies in the package that has the //go:generate:

~/Code/goreleaser/goreleaser vgo*
λ go list -f "{{.Target}}" golang.org/x/tools/cmd/stringer
/Users/carlos/Code/goreleaser/goreleaser/bin/stringer

~/Code/goreleaser/goreleaser vgo*
λ go generate ./...
stringer: checking package: artifact.go:11:2: could not import github.com/apex/log (type-checking package "github.com/apex/log" failed (/Users/carlos/go/pkg/mod/github.com/apex/[email protected]/stack.go:3:8: could not import github.com/pkg/errors (cannot find package "github.com/pkg/errors" in any of:
	/usr/local/Cellar/go/1.11/libexec/src/github.com/pkg/errors (from $GOROOT)
	/Users/carlos/go/src/github.com/pkg/errors (from $GOPATH))))
internal/artifact/artifact.go:15: running "stringer": exit status 1

Not sure if it is a bug in stringer or somewhere else, happy to provide more info if needed though.

---

Not sure if useful, but:

• PR migrating to vgo: migrating to vgo goreleaser/goreleaser#779
• Specific type that with the go generate:
  https://github.com/goreleaser/goreleaser/blob/master/internal/artifact/artifact.go#L15

[mewmew]

re: @rsc in #25922 (comment)

Best practice remains creating the tools.go file.
It's true that go mod init does not auto-create a tools.go from dep's config,
but I think doing so is getting a bit beyond scope.

The best practice described currently triggers a go vet warning (see e.g. llir/llvm#196)

From llir/llvm/ir/enum/tools.go:

//go:build tools

package enum

import (
	_ "golang.org/x/tools/cmd/stringer"
)

Running go vet -tags tools github.com/llir/llvm/ir/enum results in the following warning:

ir/enum/tools.go:6:2: import "golang.org/x/tools/cmd/stringer" is a program, not an importable package

Is using a tools.go file with a go:build tools build tag still the recommended best practice for handling tools dependencies? Or is there another approach that is now recommended?

Cheers,
Robin

[jayconrod]

@mewmew In this case, don't run go vet with -tags tools. The tools.go file is not meant to be part of a buildable package, so vet won't provide useful feedback on it. Without -tags tools, everything except go mod tidy will ignore tools.go.

[andig]

I've moved to installing tools.go dependencies using make with the power of go list:

install:
	go install $$(go list -f '{{join .Imports " "}}' tools.go)

I've not looked into this, but I feel //go:build tools should become obsolete doing so.

[SamWhited]

This will actually install to the users home directory somewhere, polluting the machine outside the repo and is generally a bad practice regardless of the build system being used (and especially with Make where "install" is a phony target and the outputs can't be kept track of). The benefit of using a tools.go and then running the deps with go run is that you always wind up using the correct version and don't have to do the install every single time (once the temporary files are downloaded and properly versioned/hashed Go is happy to skip fetching them). This should work fine with a properly written Makefile because tools.go will be a dependency of whatever your final build target is, eg. if you do something like this: GOFILES!=find . -name '*.go' tool_generated_file: tools.go go run sometool mybinary: tool_generated_file $(GOFILES) go build -o $@ Then when the tools.go file changes the tool will be re-run and your files regenerated. —Sam

…

On Thu, Oct 14, 2021, at 08:40, andig wrote: I've moved to installing `tools.go` dependencies using `make` with the power of `go list`: install: go install $$(go list -f '{{join .Imports " "}}' tools.go) I've not looked into this, but I feel `//go:build tools` should become obsolete doing so.

-- Sam Whited

[egonk]

@rsc

Best practice remains creating the tools.go file. It's true that go mod init does not auto-create a tools.go from dep's config, but I think doing so is getting a bit beyond scope.

Another data point: I'm urgently porting an older large project from git submodules to go modules and tools.go approach breaks go list -deps, which I'm using as a precise audit step together with a custom GOPROXY with a more relaxed general allow list:

// +build tools

package x

import (
	_ "golang.org/x/tools/cmd/stringer"
)

C:\Users\egon\Desktop\x>go list -deps -tags tools -f "{{with .Module}}{{.}}{{end}}"
tools.go:6:2: import "golang.org/x/tools/cmd/stringer" is a program, not an importable package
golang.org/x/sys v0.0.0-20211019181941-9d821ace8654
golang.org/x/tools v0.1.8
...

I can get rid of the error by using -e, but I prefer hard errors. An alternative is to check the final binaries with go version -m or maybe complicate the GOPROXY infrastructure with more precise rules. Not a smooth experience...

@bcmills

[antichris]

I think I'll just summarize my findings down here; it didn't all seem obvious and took some googling and tinkering, as some sources are a bit ambiguous on some of the points.

All native: tools.go

The best practice as endorsed by the Go team (#25922 (comment)).

//go:build tools
// +build tools

package yourpackage

import (
	_ "golang.org/x/tools/cmd/stringer"
	// ...
)

Change yourpackage to the name of the package that you put this in, e.g., use qux if you put it at the root of your module foo/bar/qux. Only use tools when you actually put it in a tools package (directory).

If you're on Go 1.17+, you don't need the // +build tools line.

go.mod

To add the dependencies and the latest versions of these tools to your go.mod and freshen the module cache, run

go mod tidy

Configure gopls

If you use gopls, configure it to include the tools build tag, e.g., for VSCode, add the following to your settings.json:

"gopls": {
  "build.buildFlags": ["-tags=tools"],
}

Install globally

You can install the go.mod versions of all the tools to your $GOBIN directory (probably not the best idea, expand for details)

go install $(go list -f '{{join .Imports " "}}' tools.go)

Note that this can break things for you well outside the module that you run this in, so you're likely going to be better off installing specific versions of tools manually, at your discretion.

If you still feel like you want to have the tools built (they really do start faster that way), either of the following sections (on Makefile and bingo) might be right up your alley.

Makefile

Makefiles are very project-specific and everyone's setup is different, but I've found it useful to add this to my Makefile.

toolsGo := tools.go
toolsDir := bin/tools
toolPkgs := $(shell go list -f '{{join .Imports " "}}' ${toolsGo})
toolCmds := $(foreach tool,$(notdir ${toolPkgs}),${toolsDir}/${tool})
$(foreach cmd,${toolCmds},$(eval $(notdir ${cmd})Cmd := ${cmd}))

go.mod: ${toolsGo}
	go mod tidy
	touch go.mod

${toolCmds}: go.mod
	go build -o $@ $(filter %/$(@F),${toolPkgs})

tools: ${toolCmds}
.PHONY: tools

If you decided to put your tools.go in a tools package, you would have to change the toolsGo variable to tools/tools.go. It may also make sense to flip the toolsDir variable to tools/bin. And maybe put this snippet in a tools/tools.mk, that you could include to reduce the clutter in your main Makefile.

You'd probably want to add the toolsDir directory to your .gitignore.

The magic $(foreach ... $(eval ... line defines <toolname>Cmd variables (e.g. stringerCmd, mockgenCmd, etc.) to be used in other recipes in your Makefile. For (a very rudimentary) example:

wire: ${wireCmd}
	${wireCmd} ./...
.PHONY: wire

Running make wire would now

1. run go mod tidy to ensure go.mod is up to date with tools.go (if the latter is modified more recently than the former)
2. build wire at the version that's defined in your go.mod (if it's not already built), and, finally,
3. recursively generate all your provider and injector wiring.

You can also execute

make tools

and, if go.mod is older than tools.go, it will run go mod tidy, after which all the tools that you have defined in tools.go will get built under the toolsDir (as bin/tools/stringer, for example).

3rd party: bingo

You can use bingo to install version-suffixed executables of your module's tools in the global $GOBIN directory. This is an entirely different approach that avoids conflicts among tools and tool dependencies in larger projects, but it doesn't integrate with your module's go.mod (which might also be a good thing).

The author introduced this tool in #25922 (comment) and further expanded on it in #25922 (comment)

[silverwind]

I think golang needs to:

1. Provide a way to define tool dependencies in go.mod only.
2. Provide a way to run tool binaries, e.g. go execute <tool> <args>.

Edit: It appears go run package@version already provides point 2, so really only point 1 is missing for complete management of tool dependencies.

[bluebrown]

@antichris, I still dont understand how I get the binary from that tools file and can use it in go generate?

I have this in the main package.

//go:build tools

package main

import (
	_ "github.com/golang-migrate/migrate/v4"
	_ "github.com/golang-migrate/migrate/v4/database/postgres"
	_ "github.com/golang-migrate/migrate/v4/source/github"
)

I ran go mod tidy but how can I use the binary now?

package main

//go:generate migrate -source=db
func main() {

}

$ go generate
main.go:3: running "migrate": exec: "migrate": executable file not found in $PATH

[seankhliao]

Unlike many projects, the Go project does not use GitHub Issues for general discussion or asking questions. GitHub Issues are used for tracking bugs and proposals only.

For questions please refer to https://github.com/golang/go/wiki/Questions

[antichris]

@bluebrown In general, the common practice would be running the tool using full package path in a go:generate directive, like so:

//go:generate go run golang.org/x/tools/cmd/stringer -type Foo

This would work if you have imported the respective package in your tools.go, e.g.:

import _ "golang.org/x/tools/cmd/stringer"

In your specific case, you'd have to import github.com/golang-migrate/migrate/v4/cmd/migrate:

import _ "github.com/golang-migrate/migrate/v4/cmd/migrate"

And, according to their documentation, add the appropriate build tags when running the COMMAND that you need:

//go:generate go run -tags postgres github.com/golang-migrate/migrate/v4/cmd/migrate -source=db COMMAND