Source file src/cmd/go/internal/vcs/vcs.go

Documentation: cmd/go/internal/vcs

     1  // Copyright 2012 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  package vcs
     6  
     7  import (
     8  	"encoding/json"
     9  	"errors"
    10  	"fmt"
    11  	exec "internal/execabs"
    12  	"internal/lazyregexp"
    13  	"internal/singleflight"
    14  	"io/fs"
    15  	"log"
    16  	urlpkg "net/url"
    17  	"os"
    18  	"path/filepath"
    19  	"regexp"
    20  	"strings"
    21  	"sync"
    22  
    23  	"cmd/go/internal/base"
    24  	"cmd/go/internal/cfg"
    25  	"cmd/go/internal/search"
    26  	"cmd/go/internal/str"
    27  	"cmd/go/internal/web"
    28  
    29  	"golang.org/x/mod/module"
    30  )
    31  
    32  // A vcsCmd describes how to use a version control system
    33  // like Mercurial, Git, or Subversion.
    34  type Cmd struct {
    35  	Name string
    36  	Cmd  string // name of binary to invoke command
    37  
    38  	CreateCmd   []string // commands to download a fresh copy of a repository
    39  	DownloadCmd []string // commands to download updates into an existing repository
    40  
    41  	TagCmd         []tagCmd // commands to list tags
    42  	TagLookupCmd   []tagCmd // commands to lookup tags before running tagSyncCmd
    43  	TagSyncCmd     []string // commands to sync to specific tag
    44  	TagSyncDefault []string // commands to sync to default tag
    45  
    46  	Scheme  []string
    47  	PingCmd string
    48  
    49  	RemoteRepo  func(v *Cmd, rootDir string) (remoteRepo string, err error)
    50  	ResolveRepo func(v *Cmd, rootDir, remoteRepo string) (realRepo string, err error)
    51  }
    52  
    53  var defaultSecureScheme = map[string]bool{
    54  	"https":   true,
    55  	"git+ssh": true,
    56  	"bzr+ssh": true,
    57  	"svn+ssh": true,
    58  	"ssh":     true,
    59  }
    60  
    61  func (v *Cmd) IsSecure(repo string) bool {
    62  	u, err := urlpkg.Parse(repo)
    63  	if err != nil {
    64  		// If repo is not a URL, it's not secure.
    65  		return false
    66  	}
    67  	return v.isSecureScheme(u.Scheme)
    68  }
    69  
    70  func (v *Cmd) isSecureScheme(scheme string) bool {
    71  	switch v.Cmd {
    72  	case "git":
    73  		// GIT_ALLOW_PROTOCOL is an environment variable defined by Git. It is a
    74  		// colon-separated list of schemes that are allowed to be used with git
    75  		// fetch/clone. Any scheme not mentioned will be considered insecure.
    76  		if allow := os.Getenv("GIT_ALLOW_PROTOCOL"); allow != "" {
    77  			for _, s := range strings.Split(allow, ":") {
    78  				if s == scheme {
    79  					return true
    80  				}
    81  			}
    82  			return false
    83  		}
    84  	}
    85  	return defaultSecureScheme[scheme]
    86  }
    87  
    88  // A tagCmd describes a command to list available tags
    89  // that can be passed to tagSyncCmd.
    90  type tagCmd struct {
    91  	cmd     string // command to list tags
    92  	pattern string // regexp to extract tags from list
    93  }
    94  
    95  // vcsList lists the known version control systems
    96  var vcsList = []*Cmd{
    97  	vcsHg,
    98  	vcsGit,
    99  	vcsSvn,
   100  	vcsBzr,
   101  	vcsFossil,
   102  }
   103  
   104  // vcsMod is a stub for the "mod" scheme. It's returned by
   105  // repoRootForImportPathDynamic, but is otherwise not treated as a VCS command.
   106  var vcsMod = &Cmd{Name: "mod"}
   107  
   108  // vcsByCmd returns the version control system for the given
   109  // command name (hg, git, svn, bzr).
   110  func vcsByCmd(cmd string) *Cmd {
   111  	for _, vcs := range vcsList {
   112  		if vcs.Cmd == cmd {
   113  			return vcs
   114  		}
   115  	}
   116  	return nil
   117  }
   118  
   119  // vcsHg describes how to use Mercurial.
   120  var vcsHg = &Cmd{
   121  	Name: "Mercurial",
   122  	Cmd:  "hg",
   123  
   124  	CreateCmd:   []string{"clone -U -- {repo} {dir}"},
   125  	DownloadCmd: []string{"pull"},
   126  
   127  	// We allow both tag and branch names as 'tags'
   128  	// for selecting a version. This lets people have
   129  	// a go.release.r60 branch and a go1 branch
   130  	// and make changes in both, without constantly
   131  	// editing .hgtags.
   132  	TagCmd: []tagCmd{
   133  		{"tags", `^(\S+)`},
   134  		{"branches", `^(\S+)`},
   135  	},
   136  	TagSyncCmd:     []string{"update -r {tag}"},
   137  	TagSyncDefault: []string{"update default"},
   138  
   139  	Scheme:     []string{"https", "http", "ssh"},
   140  	PingCmd:    "identify -- {scheme}://{repo}",
   141  	RemoteRepo: hgRemoteRepo,
   142  }
   143  
   144  func hgRemoteRepo(vcsHg *Cmd, rootDir string) (remoteRepo string, err error) {
   145  	out, err := vcsHg.runOutput(rootDir, "paths default")
   146  	if err != nil {
   147  		return "", err
   148  	}
   149  	return strings.TrimSpace(string(out)), nil
   150  }
   151  
   152  // vcsGit describes how to use Git.
   153  var vcsGit = &Cmd{
   154  	Name: "Git",
   155  	Cmd:  "git",
   156  
   157  	CreateCmd:   []string{"clone -- {repo} {dir}", "-go-internal-cd {dir} submodule update --init --recursive"},
   158  	DownloadCmd: []string{"pull --ff-only", "submodule update --init --recursive"},
   159  
   160  	TagCmd: []tagCmd{
   161  		// tags/xxx matches a git tag named xxx
   162  		// origin/xxx matches a git branch named xxx on the default remote repository
   163  		{"show-ref", `(?:tags|origin)/(\S+)$`},
   164  	},
   165  	TagLookupCmd: []tagCmd{
   166  		{"show-ref tags/{tag} origin/{tag}", `((?:tags|origin)/\S+)$`},
   167  	},
   168  	TagSyncCmd: []string{"checkout {tag}", "submodule update --init --recursive"},
   169  	// both createCmd and downloadCmd update the working dir.
   170  	// No need to do more here. We used to 'checkout master'
   171  	// but that doesn't work if the default branch is not named master.
   172  	// DO NOT add 'checkout master' here.
   173  	// See golang.org/issue/9032.
   174  	TagSyncDefault: []string{"submodule update --init --recursive"},
   175  
   176  	Scheme: []string{"git", "https", "http", "git+ssh", "ssh"},
   177  
   178  	// Leave out the '--' separator in the ls-remote command: git 2.7.4 does not
   179  	// support such a separator for that command, and this use should be safe
   180  	// without it because the {scheme} value comes from the predefined list above.
   181  	// See golang.org/issue/33836.
   182  	PingCmd: "ls-remote {scheme}://{repo}",
   183  
   184  	RemoteRepo: gitRemoteRepo,
   185  }
   186  
   187  // scpSyntaxRe matches the SCP-like addresses used by Git to access
   188  // repositories by SSH.
   189  var scpSyntaxRe = lazyregexp.New(`^([a-zA-Z0-9_]+)@([a-zA-Z0-9._-]+):(.*)$`)
   190  
   191  func gitRemoteRepo(vcsGit *Cmd, rootDir string) (remoteRepo string, err error) {
   192  	cmd := "config remote.origin.url"
   193  	errParse := errors.New("unable to parse output of git " + cmd)
   194  	errRemoteOriginNotFound := errors.New("remote origin not found")
   195  	outb, err := vcsGit.run1(rootDir, cmd, nil, false)
   196  	if err != nil {
   197  		// if it doesn't output any message, it means the config argument is correct,
   198  		// but the config value itself doesn't exist
   199  		if outb != nil && len(outb) == 0 {
   200  			return "", errRemoteOriginNotFound
   201  		}
   202  		return "", err
   203  	}
   204  	out := strings.TrimSpace(string(outb))
   205  
   206  	var repoURL *urlpkg.URL
   207  	if m := scpSyntaxRe.FindStringSubmatch(out); m != nil {
   208  		// Match SCP-like syntax and convert it to a URL.
   209  		// Eg, "git@github.com:user/repo" becomes
   210  		// "ssh://git@github.com/user/repo".
   211  		repoURL = &urlpkg.URL{
   212  			Scheme: "ssh",
   213  			User:   urlpkg.User(m[1]),
   214  			Host:   m[2],
   215  			Path:   m[3],
   216  		}
   217  	} else {
   218  		repoURL, err = urlpkg.Parse(out)
   219  		if err != nil {
   220  			return "", err
   221  		}
   222  	}
   223  
   224  	// Iterate over insecure schemes too, because this function simply
   225  	// reports the state of the repo. If we can't see insecure schemes then
   226  	// we can't report the actual repo URL.
   227  	for _, s := range vcsGit.Scheme {
   228  		if repoURL.Scheme == s {
   229  			return repoURL.String(), nil
   230  		}
   231  	}
   232  	return "", errParse
   233  }
   234  
   235  // vcsBzr describes how to use Bazaar.
   236  var vcsBzr = &Cmd{
   237  	Name: "Bazaar",
   238  	Cmd:  "bzr",
   239  
   240  	CreateCmd: []string{"branch -- {repo} {dir}"},
   241  
   242  	// Without --overwrite bzr will not pull tags that changed.
   243  	// Replace by --overwrite-tags after http://pad.lv/681792 goes in.
   244  	DownloadCmd: []string{"pull --overwrite"},
   245  
   246  	TagCmd:         []tagCmd{{"tags", `^(\S+)`}},
   247  	TagSyncCmd:     []string{"update -r {tag}"},
   248  	TagSyncDefault: []string{"update -r revno:-1"},
   249  
   250  	Scheme:      []string{"https", "http", "bzr", "bzr+ssh"},
   251  	PingCmd:     "info -- {scheme}://{repo}",
   252  	RemoteRepo:  bzrRemoteRepo,
   253  	ResolveRepo: bzrResolveRepo,
   254  }
   255  
   256  func bzrRemoteRepo(vcsBzr *Cmd, rootDir string) (remoteRepo string, err error) {
   257  	outb, err := vcsBzr.runOutput(rootDir, "config parent_location")
   258  	if err != nil {
   259  		return "", err
   260  	}
   261  	return strings.TrimSpace(string(outb)), nil
   262  }
   263  
   264  func bzrResolveRepo(vcsBzr *Cmd, rootDir, remoteRepo string) (realRepo string, err error) {
   265  	outb, err := vcsBzr.runOutput(rootDir, "info "+remoteRepo)
   266  	if err != nil {
   267  		return "", err
   268  	}
   269  	out := string(outb)
   270  
   271  	// Expect:
   272  	// ...
   273  	//   (branch root|repository branch): <URL>
   274  	// ...
   275  
   276  	found := false
   277  	for _, prefix := range []string{"\n  branch root: ", "\n  repository branch: "} {
   278  		i := strings.Index(out, prefix)
   279  		if i >= 0 {
   280  			out = out[i+len(prefix):]
   281  			found = true
   282  			break
   283  		}
   284  	}
   285  	if !found {
   286  		return "", fmt.Errorf("unable to parse output of bzr info")
   287  	}
   288  
   289  	i := strings.Index(out, "\n")
   290  	if i < 0 {
   291  		return "", fmt.Errorf("unable to parse output of bzr info")
   292  	}
   293  	out = out[:i]
   294  	return strings.TrimSpace(out), nil
   295  }
   296  
   297  // vcsSvn describes how to use Subversion.
   298  var vcsSvn = &Cmd{
   299  	Name: "Subversion",
   300  	Cmd:  "svn",
   301  
   302  	CreateCmd:   []string{"checkout -- {repo} {dir}"},
   303  	DownloadCmd: []string{"update"},
   304  
   305  	// There is no tag command in subversion.
   306  	// The branch information is all in the path names.
   307  
   308  	Scheme:     []string{"https", "http", "svn", "svn+ssh"},
   309  	PingCmd:    "info -- {scheme}://{repo}",
   310  	RemoteRepo: svnRemoteRepo,
   311  }
   312  
   313  func svnRemoteRepo(vcsSvn *Cmd, rootDir string) (remoteRepo string, err error) {
   314  	outb, err := vcsSvn.runOutput(rootDir, "info")
   315  	if err != nil {
   316  		return "", err
   317  	}
   318  	out := string(outb)
   319  
   320  	// Expect:
   321  	//
   322  	//	 ...
   323  	// 	URL: <URL>
   324  	// 	...
   325  	//
   326  	// Note that we're not using the Repository Root line,
   327  	// because svn allows checking out subtrees.
   328  	// The URL will be the URL of the subtree (what we used with 'svn co')
   329  	// while the Repository Root may be a much higher parent.
   330  	i := strings.Index(out, "\nURL: ")
   331  	if i < 0 {
   332  		return "", fmt.Errorf("unable to parse output of svn info")
   333  	}
   334  	out = out[i+len("\nURL: "):]
   335  	i = strings.Index(out, "\n")
   336  	if i < 0 {
   337  		return "", fmt.Errorf("unable to parse output of svn info")
   338  	}
   339  	out = out[:i]
   340  	return strings.TrimSpace(out), nil
   341  }
   342  
   343  // fossilRepoName is the name go get associates with a fossil repository. In the
   344  // real world the file can be named anything.
   345  const fossilRepoName = ".fossil"
   346  
   347  // vcsFossil describes how to use Fossil (fossil-scm.org)
   348  var vcsFossil = &Cmd{
   349  	Name: "Fossil",
   350  	Cmd:  "fossil",
   351  
   352  	CreateCmd:   []string{"-go-internal-mkdir {dir} clone -- {repo} " + filepath.Join("{dir}", fossilRepoName), "-go-internal-cd {dir} open .fossil"},
   353  	DownloadCmd: []string{"up"},
   354  
   355  	TagCmd:         []tagCmd{{"tag ls", `(.*)`}},
   356  	TagSyncCmd:     []string{"up tag:{tag}"},
   357  	TagSyncDefault: []string{"up trunk"},
   358  
   359  	Scheme:     []string{"https", "http"},
   360  	RemoteRepo: fossilRemoteRepo,
   361  }
   362  
   363  func fossilRemoteRepo(vcsFossil *Cmd, rootDir string) (remoteRepo string, err error) {
   364  	out, err := vcsFossil.runOutput(rootDir, "remote-url")
   365  	if err != nil {
   366  		return "", err
   367  	}
   368  	return strings.TrimSpace(string(out)), nil
   369  }
   370  
   371  func (v *Cmd) String() string {
   372  	return v.Name
   373  }
   374  
   375  // run runs the command line cmd in the given directory.
   376  // keyval is a list of key, value pairs. run expands
   377  // instances of {key} in cmd into value, but only after
   378  // splitting cmd into individual arguments.
   379  // If an error occurs, run prints the command line and the
   380  // command's combined stdout+stderr to standard error.
   381  // Otherwise run discards the command's output.
   382  func (v *Cmd) run(dir string, cmd string, keyval ...string) error {
   383  	_, err := v.run1(dir, cmd, keyval, true)
   384  	return err
   385  }
   386  
   387  // runVerboseOnly is like run but only generates error output to standard error in verbose mode.
   388  func (v *Cmd) runVerboseOnly(dir string, cmd string, keyval ...string) error {
   389  	_, err := v.run1(dir, cmd, keyval, false)
   390  	return err
   391  }
   392  
   393  // runOutput is like run but returns the output of the command.
   394  func (v *Cmd) runOutput(dir string, cmd string, keyval ...string) ([]byte, error) {
   395  	return v.run1(dir, cmd, keyval, true)
   396  }
   397  
   398  // run1 is the generalized implementation of run and runOutput.
   399  func (v *Cmd) run1(dir string, cmdline string, keyval []string, verbose bool) ([]byte, error) {
   400  	m := make(map[string]string)
   401  	for i := 0; i < len(keyval); i += 2 {
   402  		m[keyval[i]] = keyval[i+1]
   403  	}
   404  	args := strings.Fields(cmdline)
   405  	for i, arg := range args {
   406  		args[i] = expand(m, arg)
   407  	}
   408  
   409  	if len(args) >= 2 && args[0] == "-go-internal-mkdir" {
   410  		var err error
   411  		if filepath.IsAbs(args[1]) {
   412  			err = os.Mkdir(args[1], fs.ModePerm)
   413  		} else {
   414  			err = os.Mkdir(filepath.Join(dir, args[1]), fs.ModePerm)
   415  		}
   416  		if err != nil {
   417  			return nil, err
   418  		}
   419  		args = args[2:]
   420  	}
   421  
   422  	if len(args) >= 2 && args[0] == "-go-internal-cd" {
   423  		if filepath.IsAbs(args[1]) {
   424  			dir = args[1]
   425  		} else {
   426  			dir = filepath.Join(dir, args[1])
   427  		}
   428  		args = args[2:]
   429  	}
   430  
   431  	_, err := exec.LookPath(v.Cmd)
   432  	if err != nil {
   433  		fmt.Fprintf(os.Stderr,
   434  			"go: missing %s command. See https://golang.org/s/gogetcmd\n",
   435  			v.Name)
   436  		return nil, err
   437  	}
   438  
   439  	cmd := exec.Command(v.Cmd, args...)
   440  	cmd.Dir = dir
   441  	cmd.Env = base.AppendPWD(os.Environ(), cmd.Dir)
   442  	if cfg.BuildX {
   443  		fmt.Fprintf(os.Stderr, "cd %s\n", dir)
   444  		fmt.Fprintf(os.Stderr, "%s %s\n", v.Cmd, strings.Join(args, " "))
   445  	}
   446  	out, err := cmd.Output()
   447  	if err != nil {
   448  		if verbose || cfg.BuildV {
   449  			fmt.Fprintf(os.Stderr, "# cd %s; %s %s\n", dir, v.Cmd, strings.Join(args, " "))
   450  			if ee, ok := err.(*exec.ExitError); ok && len(ee.Stderr) > 0 {
   451  				os.Stderr.Write(ee.Stderr)
   452  			} else {
   453  				fmt.Fprintf(os.Stderr, err.Error())
   454  			}
   455  		}
   456  	}
   457  	return out, err
   458  }
   459  
   460  // Ping pings to determine scheme to use.
   461  func (v *Cmd) Ping(scheme, repo string) error {
   462  	return v.runVerboseOnly(".", v.PingCmd, "scheme", scheme, "repo", repo)
   463  }
   464  
   465  // Create creates a new copy of repo in dir.
   466  // The parent of dir must exist; dir must not.
   467  func (v *Cmd) Create(dir, repo string) error {
   468  	for _, cmd := range v.CreateCmd {
   469  		if err := v.run(".", cmd, "dir", dir, "repo", repo); err != nil {
   470  			return err
   471  		}
   472  	}
   473  	return nil
   474  }
   475  
   476  // Download downloads any new changes for the repo in dir.
   477  func (v *Cmd) Download(dir string) error {
   478  	for _, cmd := range v.DownloadCmd {
   479  		if err := v.run(dir, cmd); err != nil {
   480  			return err
   481  		}
   482  	}
   483  	return nil
   484  }
   485  
   486  // Tags returns the list of available tags for the repo in dir.
   487  func (v *Cmd) Tags(dir string) ([]string, error) {
   488  	var tags []string
   489  	for _, tc := range v.TagCmd {
   490  		out, err := v.runOutput(dir, tc.cmd)
   491  		if err != nil {
   492  			return nil, err
   493  		}
   494  		re := regexp.MustCompile(`(?m-s)` + tc.pattern)
   495  		for _, m := range re.FindAllStringSubmatch(string(out), -1) {
   496  			tags = append(tags, m[1])
   497  		}
   498  	}
   499  	return tags, nil
   500  }
   501  
   502  // tagSync syncs the repo in dir to the named tag,
   503  // which either is a tag returned by tags or is v.tagDefault.
   504  func (v *Cmd) TagSync(dir, tag string) error {
   505  	if v.TagSyncCmd == nil {
   506  		return nil
   507  	}
   508  	if tag != "" {
   509  		for _, tc := range v.TagLookupCmd {
   510  			out, err := v.runOutput(dir, tc.cmd, "tag", tag)
   511  			if err != nil {
   512  				return err
   513  			}
   514  			re := regexp.MustCompile(`(?m-s)` + tc.pattern)
   515  			m := re.FindStringSubmatch(string(out))
   516  			if len(m) > 1 {
   517  				tag = m[1]
   518  				break
   519  			}
   520  		}
   521  	}
   522  
   523  	if tag == "" && v.TagSyncDefault != nil {
   524  		for _, cmd := range v.TagSyncDefault {
   525  			if err := v.run(dir, cmd); err != nil {
   526  				return err
   527  			}
   528  		}
   529  		return nil
   530  	}
   531  
   532  	for _, cmd := range v.TagSyncCmd {
   533  		if err := v.run(dir, cmd, "tag", tag); err != nil {
   534  			return err
   535  		}
   536  	}
   537  	return nil
   538  }
   539  
   540  // A vcsPath describes how to convert an import path into a
   541  // version control system and repository name.
   542  type vcsPath struct {
   543  	pathPrefix     string                              // prefix this description applies to
   544  	regexp         *lazyregexp.Regexp                  // compiled pattern for import path
   545  	repo           string                              // repository to use (expand with match of re)
   546  	vcs            string                              // version control system to use (expand with match of re)
   547  	check          func(match map[string]string) error // additional checks
   548  	schemelessRepo bool                                // if true, the repo pattern lacks a scheme
   549  }
   550  
   551  // FromDir inspects dir and its parents to determine the
   552  // version control system and code repository to use.
   553  // On return, root is the import path
   554  // corresponding to the root of the repository.
   555  func FromDir(dir, srcRoot string) (vcs *Cmd, root string, err error) {
   556  	// Clean and double-check that dir is in (a subdirectory of) srcRoot.
   557  	dir = filepath.Clean(dir)
   558  	srcRoot = filepath.Clean(srcRoot)
   559  	if len(dir) <= len(srcRoot) || dir[len(srcRoot)] != filepath.Separator {
   560  		return nil, "", fmt.Errorf("directory %q is outside source root %q", dir, srcRoot)
   561  	}
   562  
   563  	var vcsRet *Cmd
   564  	var rootRet string
   565  
   566  	origDir := dir
   567  	for len(dir) > len(srcRoot) {
   568  		for _, vcs := range vcsList {
   569  			if _, err := os.Stat(filepath.Join(dir, "."+vcs.Cmd)); err == nil {
   570  				root := filepath.ToSlash(dir[len(srcRoot)+1:])
   571  				// Record first VCS we find, but keep looking,
   572  				// to detect mistakes like one kind of VCS inside another.
   573  				if vcsRet == nil {
   574  					vcsRet = vcs
   575  					rootRet = root
   576  					continue
   577  				}
   578  				// Allow .git inside .git, which can arise due to submodules.
   579  				if vcsRet == vcs && vcs.Cmd == "git" {
   580  					continue
   581  				}
   582  				// Otherwise, we have one VCS inside a different VCS.
   583  				return nil, "", fmt.Errorf("directory %q uses %s, but parent %q uses %s",
   584  					filepath.Join(srcRoot, rootRet), vcsRet.Cmd, filepath.Join(srcRoot, root), vcs.Cmd)
   585  			}
   586  		}
   587  
   588  		// Move to parent.
   589  		ndir := filepath.Dir(dir)
   590  		if len(ndir) >= len(dir) {
   591  			// Shouldn't happen, but just in case, stop.
   592  			break
   593  		}
   594  		dir = ndir
   595  	}
   596  
   597  	if vcsRet != nil {
   598  		if err := checkGOVCS(vcsRet, rootRet); err != nil {
   599  			return nil, "", err
   600  		}
   601  		return vcsRet, rootRet, nil
   602  	}
   603  
   604  	return nil, "", fmt.Errorf("directory %q is not using a known version control system", origDir)
   605  }
   606  
   607  // A govcsRule is a single GOVCS rule like private:hg|svn.
   608  type govcsRule struct {
   609  	pattern string
   610  	allowed []string
   611  }
   612  
   613  // A govcsConfig is a full GOVCS configuration.
   614  type govcsConfig []govcsRule
   615  
   616  func parseGOVCS(s string) (govcsConfig, error) {
   617  	s = strings.TrimSpace(s)
   618  	if s == "" {
   619  		return nil, nil
   620  	}
   621  	var cfg govcsConfig
   622  	have := make(map[string]string)
   623  	for _, item := range strings.Split(s, ",") {
   624  		item = strings.TrimSpace(item)
   625  		if item == "" {
   626  			return nil, fmt.Errorf("empty entry in GOVCS")
   627  		}
   628  		i := strings.Index(item, ":")
   629  		if i < 0 {
   630  			return nil, fmt.Errorf("malformed entry in GOVCS (missing colon): %q", item)
   631  		}
   632  		pattern, list := strings.TrimSpace(item[:i]), strings.TrimSpace(item[i+1:])
   633  		if pattern == "" {
   634  			return nil, fmt.Errorf("empty pattern in GOVCS: %q", item)
   635  		}
   636  		if list == "" {
   637  			return nil, fmt.Errorf("empty VCS list in GOVCS: %q", item)
   638  		}
   639  		if search.IsRelativePath(pattern) {
   640  			return nil, fmt.Errorf("relative pattern not allowed in GOVCS: %q", pattern)
   641  		}
   642  		if old := have[pattern]; old != "" {
   643  			return nil, fmt.Errorf("unreachable pattern in GOVCS: %q after %q", item, old)
   644  		}
   645  		have[pattern] = item
   646  		allowed := strings.Split(list, "|")
   647  		for i, a := range allowed {
   648  			a = strings.TrimSpace(a)
   649  			if a == "" {
   650  				return nil, fmt.Errorf("empty VCS name in GOVCS: %q", item)
   651  			}
   652  			allowed[i] = a
   653  		}
   654  		cfg = append(cfg, govcsRule{pattern, allowed})
   655  	}
   656  	return cfg, nil
   657  }
   658  
   659  func (c *govcsConfig) allow(path string, private bool, vcs string) bool {
   660  	for _, rule := range *c {
   661  		match := false
   662  		switch rule.pattern {
   663  		case "private":
   664  			match = private
   665  		case "public":
   666  			match = !private
   667  		default:
   668  			// Note: rule.pattern is known to be comma-free,
   669  			// so MatchPrefixPatterns is only matching a single pattern for us.
   670  			match = module.MatchPrefixPatterns(rule.pattern, path)
   671  		}
   672  		if !match {
   673  			continue
   674  		}
   675  		for _, allow := range rule.allowed {
   676  			if allow == vcs || allow == "all" {
   677  				return true
   678  			}
   679  		}
   680  		return false
   681  	}
   682  
   683  	// By default, nothing is allowed.
   684  	return false
   685  }
   686  
   687  var (
   688  	govcs     govcsConfig
   689  	govcsErr  error
   690  	govcsOnce sync.Once
   691  )
   692  
   693  // defaultGOVCS is the default setting for GOVCS.
   694  // Setting GOVCS adds entries ahead of these but does not remove them.
   695  // (They are appended to the parsed GOVCS setting.)
   696  //
   697  // The rationale behind allowing only Git and Mercurial is that
   698  // these two systems have had the most attention to issues
   699  // of being run as clients of untrusted servers. In contrast,
   700  // Bazaar, Fossil, and Subversion have primarily been used
   701  // in trusted, authenticated environments and are not as well
   702  // scrutinized as attack surfaces.
   703  //
   704  // See golang.org/issue/41730 for details.
   705  var defaultGOVCS = govcsConfig{
   706  	{"private", []string{"all"}},
   707  	{"public", []string{"git", "hg"}},
   708  }
   709  
   710  func checkGOVCS(vcs *Cmd, root string) error {
   711  	if vcs == vcsMod {
   712  		// Direct module (proxy protocol) fetches don't
   713  		// involve an external version control system
   714  		// and are always allowed.
   715  		return nil
   716  	}
   717  
   718  	govcsOnce.Do(func() {
   719  		govcs, govcsErr = parseGOVCS(os.Getenv("GOVCS"))
   720  		govcs = append(govcs, defaultGOVCS...)
   721  	})
   722  	if govcsErr != nil {
   723  		return govcsErr
   724  	}
   725  
   726  	private := module.MatchPrefixPatterns(cfg.GOPRIVATE, root)
   727  	if !govcs.allow(root, private, vcs.Cmd) {
   728  		what := "public"
   729  		if private {
   730  			what = "private"
   731  		}
   732  		return fmt.Errorf("GOVCS disallows using %s for %s %s; see 'go help vcs'", vcs.Cmd, what, root)
   733  	}
   734  
   735  	return nil
   736  }
   737  
   738  // CheckNested checks for an incorrectly-nested VCS-inside-VCS
   739  // situation for dir, checking parents up until srcRoot.
   740  func CheckNested(vcs *Cmd, dir, srcRoot string) error {
   741  	if len(dir) <= len(srcRoot) || dir[len(srcRoot)] != filepath.Separator {
   742  		return fmt.Errorf("directory %q is outside source root %q", dir, srcRoot)
   743  	}
   744  
   745  	otherDir := dir
   746  	for len(otherDir) > len(srcRoot) {
   747  		for _, otherVCS := range vcsList {
   748  			if _, err := os.Stat(filepath.Join(otherDir, "."+otherVCS.Cmd)); err == nil {
   749  				// Allow expected vcs in original dir.
   750  				if otherDir == dir && otherVCS == vcs {
   751  					continue
   752  				}
   753  				// Allow .git inside .git, which can arise due to submodules.
   754  				if otherVCS == vcs && vcs.Cmd == "git" {
   755  					continue
   756  				}
   757  				// Otherwise, we have one VCS inside a different VCS.
   758  				return fmt.Errorf("directory %q uses %s, but parent %q uses %s", dir, vcs.Cmd, otherDir, otherVCS.Cmd)
   759  			}
   760  		}
   761  		// Move to parent.
   762  		newDir := filepath.Dir(otherDir)
   763  		if len(newDir) >= len(otherDir) {
   764  			// Shouldn't happen, but just in case, stop.
   765  			break
   766  		}
   767  		otherDir = newDir
   768  	}
   769  
   770  	return nil
   771  }
   772  
   773  // RepoRoot describes the repository root for a tree of source code.
   774  type RepoRoot struct {
   775  	Repo     string // repository URL, including scheme
   776  	Root     string // import path corresponding to root of repo
   777  	IsCustom bool   // defined by served <meta> tags (as opposed to hard-coded pattern)
   778  	VCS      *Cmd
   779  }
   780  
   781  func httpPrefix(s string) string {
   782  	for _, prefix := range [...]string{"http:", "https:"} {
   783  		if strings.HasPrefix(s, prefix) {
   784  			return prefix
   785  		}
   786  	}
   787  	return ""
   788  }
   789  
   790  // ModuleMode specifies whether to prefer modules when looking up code sources.
   791  type ModuleMode int
   792  
   793  const (
   794  	IgnoreMod ModuleMode = iota
   795  	PreferMod
   796  )
   797  
   798  // RepoRootForImportPath analyzes importPath to determine the
   799  // version control system, and code repository to use.
   800  func RepoRootForImportPath(importPath string, mod ModuleMode, security web.SecurityMode) (*RepoRoot, error) {
   801  	rr, err := repoRootFromVCSPaths(importPath, security, vcsPaths)
   802  	if err == errUnknownSite {
   803  		rr, err = repoRootForImportDynamic(importPath, mod, security)
   804  		if err != nil {
   805  			err = importErrorf(importPath, "unrecognized import path %q: %v", importPath, err)
   806  		}
   807  	}
   808  	if err != nil {
   809  		rr1, err1 := repoRootFromVCSPaths(importPath, security, vcsPathsAfterDynamic)
   810  		if err1 == nil {
   811  			rr = rr1
   812  			err = nil
   813  		}
   814  	}
   815  
   816  	// Should have been taken care of above, but make sure.
   817  	if err == nil && strings.Contains(importPath, "...") && strings.Contains(rr.Root, "...") {
   818  		// Do not allow wildcards in the repo root.
   819  		rr = nil
   820  		err = importErrorf(importPath, "cannot expand ... in %q", importPath)
   821  	}
   822  	return rr, err
   823  }
   824  
   825  var errUnknownSite = errors.New("dynamic lookup required to find mapping")
   826  
   827  // repoRootFromVCSPaths attempts to map importPath to a repoRoot
   828  // using the mappings defined in vcsPaths.
   829  func repoRootFromVCSPaths(importPath string, security web.SecurityMode, vcsPaths []*vcsPath) (*RepoRoot, error) {
   830  	if str.HasPathPrefix(importPath, "example.net") {
   831  		// TODO(rsc): This should not be necessary, but it's required to keep
   832  		// tests like ../../testdata/script/mod_get_extra.txt from using the network.
   833  		// That script has everything it needs in the replacement set, but it is still
   834  		// doing network calls.
   835  		return nil, fmt.Errorf("no modules on example.net")
   836  	}
   837  	if importPath == "rsc.io" {
   838  		// This special case allows tests like ../../testdata/script/govcs.txt
   839  		// to avoid making any network calls. The module lookup for a path
   840  		// like rsc.io/nonexist.svn/foo needs to not make a network call for
   841  		// a lookup on rsc.io.
   842  		return nil, fmt.Errorf("rsc.io is not a module")
   843  	}
   844  	// A common error is to use https://packagepath because that's what
   845  	// hg and git require. Diagnose this helpfully.
   846  	if prefix := httpPrefix(importPath); prefix != "" {
   847  		// The importPath has been cleaned, so has only one slash. The pattern
   848  		// ignores the slashes; the error message puts them back on the RHS at least.
   849  		return nil, fmt.Errorf("%q not allowed in import path", prefix+"//")
   850  	}
   851  	for _, srv := range vcsPaths {
   852  		if !str.HasPathPrefix(importPath, srv.pathPrefix) {
   853  			continue
   854  		}
   855  		m := srv.regexp.FindStringSubmatch(importPath)
   856  		if m == nil {
   857  			if srv.pathPrefix != "" {
   858  				return nil, importErrorf(importPath, "invalid %s import path %q", srv.pathPrefix, importPath)
   859  			}
   860  			continue
   861  		}
   862  
   863  		// Build map of named subexpression matches for expand.
   864  		match := map[string]string{
   865  			"prefix": srv.pathPrefix + "/",
   866  			"import": importPath,
   867  		}
   868  		for i, name := range srv.regexp.SubexpNames() {
   869  			if name != "" && match[name] == "" {
   870  				match[name] = m[i]
   871  			}
   872  		}
   873  		if srv.vcs != "" {
   874  			match["vcs"] = expand(match, srv.vcs)
   875  		}
   876  		if srv.repo != "" {
   877  			match["repo"] = expand(match, srv.repo)
   878  		}
   879  		if srv.check != nil {
   880  			if err := srv.check(match); err != nil {
   881  				return nil, err
   882  			}
   883  		}
   884  		vcs := vcsByCmd(match["vcs"])
   885  		if vcs == nil {
   886  			return nil, fmt.Errorf("unknown version control system %q", match["vcs"])
   887  		}
   888  		if err := checkGOVCS(vcs, match["root"]); err != nil {
   889  			return nil, err
   890  		}
   891  		var repoURL string
   892  		if !srv.schemelessRepo {
   893  			repoURL = match["repo"]
   894  		} else {
   895  			scheme := vcs.Scheme[0] // default to first scheme
   896  			repo := match["repo"]
   897  			if vcs.PingCmd != "" {
   898  				// If we know how to test schemes, scan to find one.
   899  				for _, s := range vcs.Scheme {
   900  					if security == web.SecureOnly && !vcs.isSecureScheme(s) {
   901  						continue
   902  					}
   903  					if vcs.Ping(s, repo) == nil {
   904  						scheme = s
   905  						break
   906  					}
   907  				}
   908  			}
   909  			repoURL = scheme + "://" + repo
   910  		}
   911  		rr := &RepoRoot{
   912  			Repo: repoURL,
   913  			Root: match["root"],
   914  			VCS:  vcs,
   915  		}
   916  		return rr, nil
   917  	}
   918  	return nil, errUnknownSite
   919  }
   920  
   921  // urlForImportPath returns a partially-populated URL for the given Go import path.
   922  //
   923  // The URL leaves the Scheme field blank so that web.Get will try any scheme
   924  // allowed by the selected security mode.
   925  func urlForImportPath(importPath string) (*urlpkg.URL, error) {
   926  	slash := strings.Index(importPath, "/")
   927  	if slash < 0 {
   928  		slash = len(importPath)
   929  	}
   930  	host, path := importPath[:slash], importPath[slash:]
   931  	if !strings.Contains(host, ".") {
   932  		return nil, errors.New("import path does not begin with hostname")
   933  	}
   934  	if len(path) == 0 {
   935  		path = "/"
   936  	}
   937  	return &urlpkg.URL{Host: host, Path: path, RawQuery: "go-get=1"}, nil
   938  }
   939  
   940  // repoRootForImportDynamic finds a *RepoRoot for a custom domain that's not
   941  // statically known by repoRootForImportPathStatic.
   942  //
   943  // This handles custom import paths like "name.tld/pkg/foo" or just "name.tld".
   944  func repoRootForImportDynamic(importPath string, mod ModuleMode, security web.SecurityMode) (*RepoRoot, error) {
   945  	url, err := urlForImportPath(importPath)
   946  	if err != nil {
   947  		return nil, err
   948  	}
   949  	resp, err := web.Get(security, url)
   950  	if err != nil {
   951  		msg := "https fetch: %v"
   952  		if security == web.Insecure {
   953  			msg = "http/" + msg
   954  		}
   955  		return nil, fmt.Errorf(msg, err)
   956  	}
   957  	body := resp.Body
   958  	defer body.Close()
   959  	imports, err := parseMetaGoImports(body, mod)
   960  	if len(imports) == 0 {
   961  		if respErr := resp.Err(); respErr != nil {
   962  			// If the server's status was not OK, prefer to report that instead of
   963  			// an XML parse error.
   964  			return nil, respErr
   965  		}
   966  	}
   967  	if err != nil {
   968  		return nil, fmt.Errorf("parsing %s: %v", importPath, err)
   969  	}
   970  	// Find the matched meta import.
   971  	mmi, err := matchGoImport(imports, importPath)
   972  	if err != nil {
   973  		if _, ok := err.(ImportMismatchError); !ok {
   974  			return nil, fmt.Errorf("parse %s: %v", url, err)
   975  		}
   976  		return nil, fmt.Errorf("parse %s: no go-import meta tags (%s)", resp.URL, err)
   977  	}
   978  	if cfg.BuildV {
   979  		log.Printf("get %q: found meta tag %#v at %s", importPath, mmi, url)
   980  	}
   981  	// If the import was "uni.edu/bob/project", which said the
   982  	// prefix was "uni.edu" and the RepoRoot was "evilroot.com",
   983  	// make sure we don't trust Bob and check out evilroot.com to
   984  	// "uni.edu" yet (possibly overwriting/preempting another
   985  	// non-evil student). Instead, first verify the root and see
   986  	// if it matches Bob's claim.
   987  	if mmi.Prefix != importPath {
   988  		if cfg.BuildV {
   989  			log.Printf("get %q: verifying non-authoritative meta tag", importPath)
   990  		}
   991  		var imports []metaImport
   992  		url, imports, err = metaImportsForPrefix(mmi.Prefix, mod, security)
   993  		if err != nil {
   994  			return nil, err
   995  		}
   996  		metaImport2, err := matchGoImport(imports, importPath)
   997  		if err != nil || mmi != metaImport2 {
   998  			return nil, fmt.Errorf("%s and %s disagree about go-import for %s", resp.URL, url, mmi.Prefix)
   999  		}
  1000  	}
  1001  
  1002  	if err := validateRepoRoot(mmi.RepoRoot); err != nil {
  1003  		return nil, fmt.Errorf("%s: invalid repo root %q: %v", resp.URL, mmi.RepoRoot, err)
  1004  	}
  1005  	var vcs *Cmd
  1006  	if mmi.VCS == "mod" {
  1007  		vcs = vcsMod
  1008  	} else {
  1009  		vcs = vcsByCmd(mmi.VCS)
  1010  		if vcs == nil {
  1011  			return nil, fmt.Errorf("%s: unknown vcs %q", resp.URL, mmi.VCS)
  1012  		}
  1013  	}
  1014  
  1015  	if err := checkGOVCS(vcs, mmi.Prefix); err != nil {
  1016  		return nil, err
  1017  	}
  1018  
  1019  	rr := &RepoRoot{
  1020  		Repo:     mmi.RepoRoot,
  1021  		Root:     mmi.Prefix,
  1022  		IsCustom: true,
  1023  		VCS:      vcs,
  1024  	}
  1025  	return rr, nil
  1026  }
  1027  
  1028  // validateRepoRoot returns an error if repoRoot does not seem to be
  1029  // a valid URL with scheme.
  1030  func validateRepoRoot(repoRoot string) error {
  1031  	url, err := urlpkg.Parse(repoRoot)
  1032  	if err != nil {
  1033  		return err
  1034  	}
  1035  	if url.Scheme == "" {
  1036  		return errors.New("no scheme")
  1037  	}
  1038  	if url.Scheme == "file" {
  1039  		return errors.New("file scheme disallowed")
  1040  	}
  1041  	return nil
  1042  }
  1043  
  1044  var fetchGroup singleflight.Group
  1045  var (
  1046  	fetchCacheMu sync.Mutex
  1047  	fetchCache   = map[string]fetchResult{} // key is metaImportsForPrefix's importPrefix
  1048  )
  1049  
  1050  // metaImportsForPrefix takes a package's root import path as declared in a <meta> tag
  1051  // and returns its HTML discovery URL and the parsed metaImport lines
  1052  // found on the page.
  1053  //
  1054  // The importPath is of the form "golang.org/x/tools".
  1055  // It is an error if no imports are found.
  1056  // url will still be valid if err != nil.
  1057  // The returned url will be of the form "https://golang.org/x/tools?go-get=1"
  1058  func metaImportsForPrefix(importPrefix string, mod ModuleMode, security web.SecurityMode) (*urlpkg.URL, []metaImport, error) {
  1059  	setCache := func(res fetchResult) (fetchResult, error) {
  1060  		fetchCacheMu.Lock()
  1061  		defer fetchCacheMu.Unlock()
  1062  		fetchCache[importPrefix] = res
  1063  		return res, nil
  1064  	}
  1065  
  1066  	resi, _, _ := fetchGroup.Do(importPrefix, func() (resi interface{}, err error) {
  1067  		fetchCacheMu.Lock()
  1068  		if res, ok := fetchCache[importPrefix]; ok {
  1069  			fetchCacheMu.Unlock()
  1070  			return res, nil
  1071  		}
  1072  		fetchCacheMu.Unlock()
  1073  
  1074  		url, err := urlForImportPath(importPrefix)
  1075  		if err != nil {
  1076  			return setCache(fetchResult{err: err})
  1077  		}
  1078  		resp, err := web.Get(security, url)
  1079  		if err != nil {
  1080  			return setCache(fetchResult{url: url, err: fmt.Errorf("fetching %s: %v", importPrefix, err)})
  1081  		}
  1082  		body := resp.Body
  1083  		defer body.Close()
  1084  		imports, err := parseMetaGoImports(body, mod)
  1085  		if len(imports) == 0 {
  1086  			if respErr := resp.Err(); respErr != nil {
  1087  				// If the server's status was not OK, prefer to report that instead of
  1088  				// an XML parse error.
  1089  				return setCache(fetchResult{url: url, err: respErr})
  1090  			}
  1091  		}
  1092  		if err != nil {
  1093  			return setCache(fetchResult{url: url, err: fmt.Errorf("parsing %s: %v", resp.URL, err)})
  1094  		}
  1095  		if len(imports) == 0 {
  1096  			err = fmt.Errorf("fetching %s: no go-import meta tag found in %s", importPrefix, resp.URL)
  1097  		}
  1098  		return setCache(fetchResult{url: url, imports: imports, err: err})
  1099  	})
  1100  	res := resi.(fetchResult)
  1101  	return res.url, res.imports, res.err
  1102  }
  1103  
  1104  type fetchResult struct {
  1105  	url     *urlpkg.URL
  1106  	imports []metaImport
  1107  	err     error
  1108  }
  1109  
  1110  // metaImport represents the parsed <meta name="go-import"
  1111  // content="prefix vcs reporoot" /> tags from HTML files.
  1112  type metaImport struct {
  1113  	Prefix, VCS, RepoRoot string
  1114  }
  1115  
  1116  // A ImportMismatchError is returned where metaImport/s are present
  1117  // but none match our import path.
  1118  type ImportMismatchError struct {
  1119  	importPath string
  1120  	mismatches []string // the meta imports that were discarded for not matching our importPath
  1121  }
  1122  
  1123  func (m ImportMismatchError) Error() string {
  1124  	formattedStrings := make([]string, len(m.mismatches))
  1125  	for i, pre := range m.mismatches {
  1126  		formattedStrings[i] = fmt.Sprintf("meta tag %s did not match import path %s", pre, m.importPath)
  1127  	}
  1128  	return strings.Join(formattedStrings, ", ")
  1129  }
  1130  
  1131  // matchGoImport returns the metaImport from imports matching importPath.
  1132  // An error is returned if there are multiple matches.
  1133  // An ImportMismatchError is returned if none match.
  1134  func matchGoImport(imports []metaImport, importPath string) (metaImport, error) {
  1135  	match := -1
  1136  
  1137  	errImportMismatch := ImportMismatchError{importPath: importPath}
  1138  	for i, im := range imports {
  1139  		if !str.HasPathPrefix(importPath, im.Prefix) {
  1140  			errImportMismatch.mismatches = append(errImportMismatch.mismatches, im.Prefix)
  1141  			continue
  1142  		}
  1143  
  1144  		if match >= 0 {
  1145  			if imports[match].VCS == "mod" && im.VCS != "mod" {
  1146  				// All the mod entries precede all the non-mod entries.
  1147  				// We have a mod entry and don't care about the rest,
  1148  				// matching or not.
  1149  				break
  1150  			}
  1151  			return metaImport{}, fmt.Errorf("multiple meta tags match import path %q", importPath)
  1152  		}
  1153  		match = i
  1154  	}
  1155  
  1156  	if match == -1 {
  1157  		return metaImport{}, errImportMismatch
  1158  	}
  1159  	return imports[match], nil
  1160  }
  1161  
  1162  // expand rewrites s to replace {k} with match[k] for each key k in match.
  1163  func expand(match map[string]string, s string) string {
  1164  	// We want to replace each match exactly once, and the result of expansion
  1165  	// must not depend on the iteration order through the map.
  1166  	// A strings.Replacer has exactly the properties we're looking for.
  1167  	oldNew := make([]string, 0, 2*len(match))
  1168  	for k, v := range match {
  1169  		oldNew = append(oldNew, "{"+k+"}", v)
  1170  	}
  1171  	return strings.NewReplacer(oldNew...).Replace(s)
  1172  }
  1173  
  1174  // vcsPaths defines the meaning of import paths referring to
  1175  // commonly-used VCS hosting sites (github.com/user/dir)
  1176  // and import paths referring to a fully-qualified importPath
  1177  // containing a VCS type (foo.com/repo.git/dir)
  1178  var vcsPaths = []*vcsPath{
  1179  	// GitHub
  1180  	{
  1181  		pathPrefix: "github.com",
  1182  		regexp:     lazyregexp.New(`^(?P<root>github\.com/[A-Za-z0-9_.\-]+/[A-Za-z0-9_.\-]+)(/[A-Za-z0-9_.\-]+)*$`),
  1183  		vcs:        "git",
  1184  		repo:       "https://{root}",
  1185  		check:      noVCSSuffix,
  1186  	},
  1187  
  1188  	// Bitbucket
  1189  	{
  1190  		pathPrefix: "bitbucket.org",
  1191  		regexp:     lazyregexp.New(`^(?P<root>bitbucket\.org/(?P<bitname>[A-Za-z0-9_.\-]+/[A-Za-z0-9_.\-]+))(/[A-Za-z0-9_.\-]+)*$`),
  1192  		repo:       "https://{root}",
  1193  		check:      bitbucketVCS,
  1194  	},
  1195  
  1196  	// IBM DevOps Services (JazzHub)
  1197  	{
  1198  		pathPrefix: "hub.jazz.net/git",
  1199  		regexp:     lazyregexp.New(`^(?P<root>hub\.jazz\.net/git/[a-z0-9]+/[A-Za-z0-9_.\-]+)(/[A-Za-z0-9_.\-]+)*$`),
  1200  		vcs:        "git",
  1201  		repo:       "https://{root}",
  1202  		check:      noVCSSuffix,
  1203  	},
  1204  
  1205  	// Git at Apache
  1206  	{
  1207  		pathPrefix: "git.apache.org",
  1208  		regexp:     lazyregexp.New(`^(?P<root>git\.apache\.org/[a-z0-9_.\-]+\.git)(/[A-Za-z0-9_.\-]+)*$`),
  1209  		vcs:        "git",
  1210  		repo:       "https://{root}",
  1211  	},
  1212  
  1213  	// Git at OpenStack
  1214  	{
  1215  		pathPrefix: "git.openstack.org",
  1216  		regexp:     lazyregexp.New(`^(?P<root>git\.openstack\.org/[A-Za-z0-9_.\-]+/[A-Za-z0-9_.\-]+)(\.git)?(/[A-Za-z0-9_.\-]+)*$`),
  1217  		vcs:        "git",
  1218  		repo:       "https://{root}",
  1219  	},
  1220  
  1221  	// chiselapp.com for fossil
  1222  	{
  1223  		pathPrefix: "chiselapp.com",
  1224  		regexp:     lazyregexp.New(`^(?P<root>chiselapp\.com/user/[A-Za-z0-9]+/repository/[A-Za-z0-9_.\-]+)$`),
  1225  		vcs:        "fossil",
  1226  		repo:       "https://{root}",
  1227  	},
  1228  
  1229  	// General syntax for any server.
  1230  	// Must be last.
  1231  	{
  1232  		regexp:         lazyregexp.New(`(?P<root>(?P<repo>([a-z0-9.\-]+\.)+[a-z0-9.\-]+(:[0-9]+)?(/~?[A-Za-z0-9_.\-]+)+?)\.(?P<vcs>bzr|fossil|git|hg|svn))(/~?[A-Za-z0-9_.\-]+)*$`),
  1233  		schemelessRepo: true,
  1234  	},
  1235  }
  1236  
  1237  // vcsPathsAfterDynamic gives additional vcsPaths entries
  1238  // to try after the dynamic HTML check.
  1239  // This gives those sites a chance to introduce <meta> tags
  1240  // as part of a graceful transition away from the hard-coded logic.
  1241  var vcsPathsAfterDynamic = []*vcsPath{
  1242  	// Launchpad. See golang.org/issue/11436.
  1243  	{
  1244  		pathPrefix: "launchpad.net",
  1245  		regexp:     lazyregexp.New(`^(?P<root>launchpad\.net/((?P<project>[A-Za-z0-9_.\-]+)(?P<series>/[A-Za-z0-9_.\-]+)?|~[A-Za-z0-9_.\-]+/(\+junk|[A-Za-z0-9_.\-]+)/[A-Za-z0-9_.\-]+))(/[A-Za-z0-9_.\-]+)*$`),
  1246  		vcs:        "bzr",
  1247  		repo:       "https://{root}",
  1248  		check:      launchpadVCS,
  1249  	},
  1250  }
  1251  
  1252  // noVCSSuffix checks that the repository name does not
  1253  // end in .foo for any version control system foo.
  1254  // The usual culprit is ".git".
  1255  func noVCSSuffix(match map[string]string) error {
  1256  	repo := match["repo"]
  1257  	for _, vcs := range vcsList {
  1258  		if strings.HasSuffix(repo, "."+vcs.Cmd) {
  1259  			return fmt.Errorf("invalid version control suffix in %s path", match["prefix"])
  1260  		}
  1261  	}
  1262  	return nil
  1263  }
  1264  
  1265  // bitbucketVCS determines the version control system for a
  1266  // Bitbucket repository, by using the Bitbucket API.
  1267  func bitbucketVCS(match map[string]string) error {
  1268  	if err := noVCSSuffix(match); err != nil {
  1269  		return err
  1270  	}
  1271  
  1272  	var resp struct {
  1273  		SCM string `json:"scm"`
  1274  	}
  1275  	url := &urlpkg.URL{
  1276  		Scheme:   "https",
  1277  		Host:     "api.bitbucket.org",
  1278  		Path:     expand(match, "/2.0/repositories/{bitname}"),
  1279  		RawQuery: "fields=scm",
  1280  	}
  1281  	data, err := web.GetBytes(url)
  1282  	if err != nil {
  1283  		if httpErr, ok := err.(*web.HTTPError); ok && httpErr.StatusCode == 403 {
  1284  			// this may be a private repository. If so, attempt to determine which
  1285  			// VCS it uses. See issue 5375.
  1286  			root := match["root"]
  1287  			for _, vcs := range []string{"git", "hg"} {
  1288  				if vcsByCmd(vcs).Ping("https", root) == nil {
  1289  					resp.SCM = vcs
  1290  					break
  1291  				}
  1292  			}
  1293  		}
  1294  
  1295  		if resp.SCM == "" {
  1296  			return err
  1297  		}
  1298  	} else {
  1299  		if err := json.Unmarshal(data, &resp); err != nil {
  1300  			return fmt.Errorf("decoding %s: %v", url, err)
  1301  		}
  1302  	}
  1303  
  1304  	if vcsByCmd(resp.SCM) != nil {
  1305  		match["vcs"] = resp.SCM
  1306  		if resp.SCM == "git" {
  1307  			match["repo"] += ".git"
  1308  		}
  1309  		return nil
  1310  	}
  1311  
  1312  	return fmt.Errorf("unable to detect version control system for bitbucket.org/ path")
  1313  }
  1314  
  1315  // launchpadVCS solves the ambiguity for "lp.net/project/foo". In this case,
  1316  // "foo" could be a series name registered in Launchpad with its own branch,
  1317  // and it could also be the name of a directory within the main project
  1318  // branch one level up.
  1319  func launchpadVCS(match map[string]string) error {
  1320  	if match["project"] == "" || match["series"] == "" {
  1321  		return nil
  1322  	}
  1323  	url := &urlpkg.URL{
  1324  		Scheme: "https",
  1325  		Host:   "code.launchpad.net",
  1326  		Path:   expand(match, "/{project}{series}/.bzr/branch-format"),
  1327  	}
  1328  	_, err := web.GetBytes(url)
  1329  	if err != nil {
  1330  		match["root"] = expand(match, "launchpad.net/{project}")
  1331  		match["repo"] = expand(match, "https://{root}")
  1332  	}
  1333  	return nil
  1334  }
  1335  
  1336  // importError is a copy of load.importError, made to avoid a dependency cycle
  1337  // on cmd/go/internal/load. It just needs to satisfy load.ImportPathError.
  1338  type importError struct {
  1339  	importPath string
  1340  	err        error
  1341  }
  1342  
  1343  func importErrorf(path, format string, args ...interface{}) error {
  1344  	err := &importError{importPath: path, err: fmt.Errorf(format, args...)}
  1345  	if errStr := err.Error(); !strings.Contains(errStr, path) {
  1346  		panic(fmt.Sprintf("path %q not in error %q", path, errStr))
  1347  	}
  1348  	return err
  1349  }
  1350  
  1351  func (e *importError) Error() string {
  1352  	return e.err.Error()
  1353  }
  1354  
  1355  func (e *importError) Unwrap() error {
  1356  	// Don't return e.err directly, since we're only wrapping an error if %w
  1357  	// was passed to ImportErrorf.
  1358  	return errors.Unwrap(e.err)
  1359  }
  1360  
  1361  func (e *importError) ImportPath() string {
  1362  	return e.importPath
  1363  }
  1364  

View as plain text