mirror of
https://github.com/tailscale/tailscale.git
synced 2026-02-09 17:52:57 +01:00
This file was never truly necessary and has never actually been used in the history of Tailscale's open source releases. A Brief History of AUTHORS files --- The AUTHORS file was a pattern developed at Google, originally for Chromium, then adopted by Go and a bunch of other projects. The problem was that Chromium originally had a copyright line only recognizing Google as the copyright holder. Because Google (and most open source projects) do not require copyright assignemnt for contributions, each contributor maintains their copyright. Some large corporate contributors then tried to add their own name to the copyright line in the LICENSE file or in file headers. This quickly becomes unwieldy, and puts a tremendous burden on anyone building on top of Chromium, since the license requires that they keep all copyright lines intact. The compromise was to create an AUTHORS file that would list all of the copyright holders. The LICENSE file and source file headers would then include that list by reference, listing the copyright holder as "The Chromium Authors". This also become cumbersome to simply keep the file up to date with a high rate of new contributors. Plus it's not always obvious who the copyright holder is. Sometimes it is the individual making the contribution, but many times it may be their employer. There is no way for the proejct maintainer to know. Eventually, Google changed their policy to no longer recommend trying to keep the AUTHORS file up to date proactively, and instead to only add to it when requested: https://opensource.google/docs/releasing/authors. They are also clear that: > Adding contributors to the AUTHORS file is entirely within the > project's discretion and has no implications for copyright ownership. It was primarily added to appease a small number of large contributors that insisted that they be recognized as copyright holders (which was entirely their right to do). But it's not truly necessary, and not even the most accurate way of identifying contributors and/or copyright holders. In practice, we've never added anyone to our AUTHORS file. It only lists Tailscale, so it's not really serving any purpose. It also causes confusion because Tailscalars put the "Tailscale Inc & AUTHORS" header in other open source repos which don't actually have an AUTHORS file, so it's ambiguous what that means. Instead, we just acknowledge that the contributors to Tailscale (whoever they are) are copyright holders for their individual contributions. We also have the benefit of using the DCO (developercertificate.org) which provides some additional certification of their right to make the contribution. The source file changes were purely mechanical with: git ls-files | xargs sed -i -e 's/\(Tailscale Inc &\) AUTHORS/\1 contributors/g' Updates #cleanup Change-Id: Ia101a4a3005adb9118051b3416f5a64a4a45987d Signed-off-by: Will Norris <will@tailscale.com>
414 lines
12 KiB
Go
414 lines
12 KiB
Go
// Copyright (c) Tailscale Inc & contributors
|
|
// SPDX-License-Identifier: BSD-3-Clause
|
|
|
|
// Package dist is a release artifact builder library.
|
|
package dist
|
|
|
|
import (
|
|
"bytes"
|
|
"errors"
|
|
"fmt"
|
|
"io"
|
|
"log"
|
|
"os"
|
|
"os/exec"
|
|
"path/filepath"
|
|
"regexp"
|
|
"runtime"
|
|
"sort"
|
|
"strings"
|
|
"sync"
|
|
"time"
|
|
|
|
"tailscale.com/version/mkversion"
|
|
)
|
|
|
|
// A Target is something that can be build in a Build.
|
|
type Target interface {
|
|
String() string
|
|
Build(build *Build) ([]string, error)
|
|
}
|
|
|
|
// Signer is pluggable signer for a Target.
|
|
type Signer func(io.Reader) ([]byte, error)
|
|
|
|
// SignFile signs the file at filePath with s and writes the signature to
|
|
// sigPath.
|
|
func (s Signer) SignFile(filePath, sigPath string) error {
|
|
f, err := os.Open(filePath)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
defer f.Close()
|
|
sig, err := s(f)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
return os.WriteFile(sigPath, sig, 0644)
|
|
}
|
|
|
|
// A Build is a build context for Targets.
|
|
type Build struct {
|
|
// Repo is a path to the root Go module for the build.
|
|
Repo string
|
|
// Out is where build artifacts are written.
|
|
Out string
|
|
// Verbose is whether to print all command output, rather than just failed
|
|
// commands.
|
|
Verbose bool
|
|
// WebClientSource is a path to the source for the web client.
|
|
// If non-empty, web client assets will be built.
|
|
WebClientSource string
|
|
|
|
// Tmp is a temporary directory that gets deleted when the Builder is closed.
|
|
Tmp string
|
|
// Go is the path to the Go binary to use for building.
|
|
Go string
|
|
// Yarn is the path to the yarn binary to use for building the web client assets.
|
|
Yarn string
|
|
// Version is the version info of the build.
|
|
Version mkversion.VersionInfo
|
|
// Time is the timestamp of the build.
|
|
Time time.Time
|
|
|
|
// once is a cache of function invocations that should run once per process
|
|
// (for example building a helper docker container)
|
|
once once
|
|
|
|
extraMu sync.Mutex
|
|
extra map[any]any
|
|
|
|
goBuilds Memoize[string]
|
|
// When running `dist build all` on a cold Go build cache, the fanout of
|
|
// gooses and goarches results in a very large number of compile processes,
|
|
// which bogs down the build machine.
|
|
//
|
|
// This throttles the number of concurrent `go build` invocations to the
|
|
// number of CPU cores, which empirically keeps the builder responsive
|
|
// without impacting overall build time.
|
|
goBuildLimit chan struct{}
|
|
|
|
onCloseFuncs []func() error // funcs to be called when Builder is closed
|
|
}
|
|
|
|
// NewBuild creates a new Build rooted at repo, and writing artifacts to out.
|
|
func NewBuild(repo, out string) (*Build, error) {
|
|
if err := os.MkdirAll(out, 0750); err != nil {
|
|
return nil, fmt.Errorf("creating out dir: %w", err)
|
|
}
|
|
tmp, err := os.MkdirTemp("", "dist-*")
|
|
if err != nil {
|
|
return nil, fmt.Errorf("creating tempdir: %w", err)
|
|
}
|
|
repo, err = findModRoot(repo)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("finding module root: %w", err)
|
|
}
|
|
goTool, err := findTool(repo, "go")
|
|
if err != nil {
|
|
return nil, fmt.Errorf("finding go binary: %w", err)
|
|
}
|
|
yarnTool, err := findTool(repo, "yarn")
|
|
if err != nil {
|
|
return nil, fmt.Errorf("finding yarn binary: %w", err)
|
|
}
|
|
b := &Build{
|
|
Repo: repo,
|
|
Tmp: tmp,
|
|
Out: out,
|
|
Go: goTool,
|
|
Yarn: yarnTool,
|
|
Version: mkversion.Info(),
|
|
Time: time.Now().UTC(),
|
|
extra: map[any]any{},
|
|
goBuildLimit: make(chan struct{}, runtime.NumCPU()),
|
|
}
|
|
|
|
return b, nil
|
|
}
|
|
|
|
func (b *Build) AddOnCloseFunc(f func() error) {
|
|
b.onCloseFuncs = append(b.onCloseFuncs, f)
|
|
}
|
|
|
|
// Close ends the build, cleans up temporary files,
|
|
// and runs any onCloseFuncs.
|
|
func (b *Build) Close() error {
|
|
var errs []error
|
|
errs = append(errs, os.RemoveAll(b.Tmp))
|
|
for _, f := range b.onCloseFuncs {
|
|
errs = append(errs, f())
|
|
}
|
|
return errors.Join(errs...)
|
|
}
|
|
|
|
// Build builds all targets concurrently.
|
|
func (b *Build) Build(targets []Target) (files []string, err error) {
|
|
if len(targets) == 0 {
|
|
return nil, errors.New("no targets specified")
|
|
}
|
|
log.Printf("Building %d targets: %v", len(targets), targets)
|
|
var (
|
|
wg sync.WaitGroup
|
|
errs = make([]error, len(targets))
|
|
buildFiles = make([][]string, len(targets))
|
|
)
|
|
for i, t := range targets {
|
|
wg.Add(1)
|
|
go func(i int, t Target) {
|
|
var err error
|
|
defer func() {
|
|
if err != nil {
|
|
err = fmt.Errorf("%s: %w", t, err)
|
|
}
|
|
errs[i] = err
|
|
wg.Done()
|
|
}()
|
|
fs, err := t.Build(b)
|
|
buildFiles[i] = fs
|
|
}(i, t)
|
|
}
|
|
wg.Wait()
|
|
|
|
for _, fs := range buildFiles {
|
|
files = append(files, fs...)
|
|
}
|
|
sort.Strings(files)
|
|
|
|
return files, errors.Join(errs...)
|
|
}
|
|
|
|
// Once runs fn if Once hasn't been called with name before.
|
|
func (b *Build) Once(name string, fn func() error) error {
|
|
return b.once.Do(name, fn)
|
|
}
|
|
|
|
// Extra returns a value from the build's extra state, creating it if necessary.
|
|
func (b *Build) Extra(key any, constructor func() any) any {
|
|
b.extraMu.Lock()
|
|
defer b.extraMu.Unlock()
|
|
ret, ok := b.extra[key]
|
|
if !ok {
|
|
ret = constructor()
|
|
b.extra[key] = ret
|
|
}
|
|
return ret
|
|
}
|
|
|
|
// GoPkg returns the path on disk of pkg.
|
|
// The module of pkg must be imported in b.Repo's go.mod.
|
|
func (b *Build) GoPkg(pkg string) (string, error) {
|
|
out, err := b.Command(b.Repo, b.Go, "list", "-f", "{{.Dir}}", pkg).CombinedOutput()
|
|
if err != nil {
|
|
return "", fmt.Errorf("finding package %q: %w", pkg, err)
|
|
}
|
|
return strings.TrimSpace(out), nil
|
|
}
|
|
|
|
// TmpDir creates and returns a new empty temporary directory.
|
|
// The caller does not need to clean up the directory after use, it will get
|
|
// deleted by b.Close().
|
|
func (b *Build) TmpDir() string {
|
|
// Because we're creating all temp dirs in our parent temp dir, the only
|
|
// failures that can happen at this point are sequence breaks (e.g. if b.Tmp
|
|
// is deleted while stuff is still running). So, panic on error to slightly
|
|
// simplify callsites.
|
|
ret, err := os.MkdirTemp(b.Tmp, "")
|
|
if err != nil {
|
|
panic(fmt.Sprintf("creating temp dir: %v", err))
|
|
}
|
|
return ret
|
|
}
|
|
|
|
// BuildWebClientAssets builds the JS and CSS assets used by the web client.
|
|
// If b.WebClientSource is non-empty, assets are built in a "build" sub-directory of that path.
|
|
// Otherwise, no assets are built.
|
|
func (b *Build) BuildWebClientAssets() error {
|
|
// Nothing in the web client assets is platform-specific,
|
|
// so we only need to build it once.
|
|
return b.Once("build-web-client-assets", func() error {
|
|
if b.WebClientSource == "" {
|
|
return nil
|
|
}
|
|
dir := b.WebClientSource
|
|
if err := b.Command(dir, b.Yarn, "install").Run(); err != nil {
|
|
return err
|
|
}
|
|
if err := b.Command(dir, b.Yarn, "build").Run(); err != nil {
|
|
return err
|
|
}
|
|
return nil
|
|
})
|
|
}
|
|
|
|
// BuildGoBinary builds the Go binary at path and returns the path to the
|
|
// binary. Builds are cached by path and env, so each build only happens once
|
|
// per process execution.
|
|
func (b *Build) BuildGoBinary(path string, env map[string]string) (string, error) {
|
|
return b.BuildGoBinaryWithTags(path, env, nil)
|
|
}
|
|
|
|
// BuildGoBinaryWithTags builds the Go binary at path and returns the
|
|
// path to the binary. Builds are cached by path, env and tags, so
|
|
// each build only happens once per process execution.
|
|
//
|
|
// The passed in tags override gocross's automatic selection of build
|
|
// tags, so you will have to figure out and specify all the tags
|
|
// relevant to your build.
|
|
func (b *Build) BuildGoBinaryWithTags(path string, env map[string]string, tags []string) (string, error) {
|
|
err := b.Once("init-go", func() error {
|
|
log.Printf("Initializing Go toolchain")
|
|
// If the build is using a tool/go, it may need to download a toolchain
|
|
// and do other initialization. Running `go version` once takes care of
|
|
// all of that and avoids that initialization happening concurrently
|
|
// later on in builds.
|
|
_, err := b.Command(b.Repo, b.Go, "version").CombinedOutput()
|
|
return err
|
|
})
|
|
if err != nil {
|
|
return "", err
|
|
}
|
|
|
|
buildKey := []any{"go-build", path, env, tags}
|
|
return b.goBuilds.Do(buildKey, func() (string, error) {
|
|
b.goBuildLimit <- struct{}{}
|
|
defer func() { <-b.goBuildLimit }()
|
|
|
|
var envStrs []string
|
|
for k, v := range env {
|
|
envStrs = append(envStrs, k+"="+v)
|
|
}
|
|
sort.Strings(envStrs)
|
|
buildDir := b.TmpDir()
|
|
outPath := buildDir
|
|
if env["GOOS"] == "windowsdll" {
|
|
// DLL builds fail unless we use a fully-qualified path to the output binary.
|
|
outPath = filepath.Join(buildDir, filepath.Base(path)+".dll")
|
|
}
|
|
args := []string{"build", "-v", "-o", outPath}
|
|
if len(tags) > 0 {
|
|
tagsStr := strings.Join(tags, ",")
|
|
log.Printf("Building %s (with env %s, tags %s)", path, strings.Join(envStrs, " "), tagsStr)
|
|
args = append(args, "-tags="+tagsStr)
|
|
} else {
|
|
log.Printf("Building %s (with env %s)", path, strings.Join(envStrs, " "))
|
|
}
|
|
args = append(args, path)
|
|
cmd := b.Command(b.Repo, b.Go, args...)
|
|
for k, v := range env {
|
|
cmd.Cmd.Env = append(cmd.Cmd.Env, k+"="+v)
|
|
}
|
|
if err := cmd.Run(); err != nil {
|
|
return "", err
|
|
}
|
|
out := filepath.Join(buildDir, filepath.Base(path))
|
|
if env["GOOS"] == "windows" || env["GOOS"] == "windowsgui" {
|
|
out += ".exe"
|
|
} else if env["GOOS"] == "windowsdll" {
|
|
out += ".dll"
|
|
}
|
|
return out, nil
|
|
})
|
|
}
|
|
|
|
// Command prepares an exec.Cmd to run [cmd, args...] in dir.
|
|
func (b *Build) Command(dir, cmd string, args ...string) *Command {
|
|
ret := &Command{
|
|
Cmd: exec.Command(cmd, args...),
|
|
}
|
|
if b.Verbose {
|
|
ret.Cmd.Stdout = os.Stdout
|
|
ret.Cmd.Stderr = os.Stderr
|
|
} else {
|
|
ret.Cmd.Stdout = &ret.Output
|
|
ret.Cmd.Stderr = &ret.Output
|
|
}
|
|
// dist always wants to use gocross if any Go is involved.
|
|
ret.Cmd.Env = append(os.Environ(), "TS_USE_GOCROSS=1")
|
|
ret.Cmd.Dir = dir
|
|
return ret
|
|
}
|
|
|
|
// Command runs an exec.Cmd and returns its exit status. If the command fails,
|
|
// its output is printed to os.Stdout, otherwise it's suppressed.
|
|
type Command struct {
|
|
Cmd *exec.Cmd
|
|
Output bytes.Buffer
|
|
}
|
|
|
|
// Run is like c.Cmd.Run, but if the command fails, its output is printed to
|
|
// os.Stdout before returning the error.
|
|
func (c *Command) Run() error {
|
|
err := c.Cmd.Run()
|
|
if err != nil {
|
|
// Command failed, dump its output.
|
|
os.Stdout.Write(c.Output.Bytes())
|
|
}
|
|
return err
|
|
}
|
|
|
|
// CombinedOutput is like c.Cmd.CombinedOutput, but returns the output as a
|
|
// string instead of a byte slice.
|
|
func (c *Command) CombinedOutput() (string, error) {
|
|
c.Cmd.Stdout = nil
|
|
c.Cmd.Stderr = nil
|
|
bs, err := c.Cmd.CombinedOutput()
|
|
return string(bs), err
|
|
}
|
|
|
|
func findModRoot(path string) (string, error) {
|
|
for {
|
|
modpath := filepath.Join(path, "go.mod")
|
|
if _, err := os.Stat(modpath); err == nil {
|
|
return path, nil
|
|
} else if !errors.Is(err, os.ErrNotExist) {
|
|
return "", err
|
|
}
|
|
path = filepath.Dir(path)
|
|
if path == "/" {
|
|
return "", fmt.Errorf("no go.mod found in %q or any parent directory", path)
|
|
}
|
|
}
|
|
}
|
|
|
|
// findTool returns the path to the specified named tool.
|
|
// It first looks in the "tool" directory in the provided path,
|
|
// then in the $PATH environment variable.
|
|
func findTool(path, name string) (string, error) {
|
|
tool := filepath.Join(path, "tool", name)
|
|
if _, err := os.Stat(tool); err == nil {
|
|
return tool, nil
|
|
}
|
|
tool, err := exec.LookPath(name)
|
|
if err != nil {
|
|
return "", err
|
|
}
|
|
return tool, nil
|
|
}
|
|
|
|
// FilterTargets returns the subset of targets that match any of the filters.
|
|
// If filters is empty, returns all targets.
|
|
func FilterTargets(targets []Target, filters []string) ([]Target, error) {
|
|
var filts []*regexp.Regexp
|
|
for _, f := range filters {
|
|
if f == "all" {
|
|
return targets, nil
|
|
}
|
|
filt, err := regexp.Compile(f)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("invalid filter %q: %w", f, err)
|
|
}
|
|
filts = append(filts, filt)
|
|
}
|
|
var ret []Target
|
|
for _, t := range targets {
|
|
for _, filt := range filts {
|
|
if filt.MatchString(t.String()) {
|
|
ret = append(ret, t)
|
|
break
|
|
}
|
|
}
|
|
}
|
|
return ret, nil
|
|
}
|