Sanitize logs for mirror sync (#3057, #3082) (#3078)

* Sanitize logs for mirror sync

* Fix error message sanitiziation (#3082)

Author: ethantkoenig

models/repo.go

@@ -605,9 +605,14 @@ func (repo *Repository) RepoPath() string {
 	return repo.repoPath(x)
 }
 
+// GitConfigPath returns the path to a repository's git config/ directory
+func GitConfigPath(repoPath string) string {
+	return filepath.Join(repoPath, "config")
+}
+
 // GitConfigPath returns the repository git config path
 func (repo *Repository) GitConfigPath() string {
-	return filepath.Join(repo.RepoPath(), "config")
+	return GitConfigPath(repo.RepoPath())
 }
 
 // RelLink returns the repository relative link

models/repo_mirror.go

@@ -6,18 +6,18 @@ package models
 
 import (
 	"fmt"
-	"strings"
 	"time"
 
-	"github.com/Unknwon/com"
-	"github.com/go-xorm/xorm"
-	"gopkg.in/ini.v1"
-
 	"code.gitea.io/git"
 	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/process"
 	"code.gitea.io/gitea/modules/setting"
 	"code.gitea.io/gitea/modules/sync"
+	"code.gitea.io/gitea/modules/util"
+
+	"github.com/Unknwon/com"
+	"github.com/go-xorm/xorm"
+	"gopkg.in/ini.v1"
 )
 
 // MirrorQueue holds an UniqueQueue object of the mirror
@@ -76,41 +76,41 @@ func (m *Mirror) ScheduleNextUpdate() {
 	m.NextUpdate = time.Now().Add(m.Interval)
 }
 
+func remoteAddress(repoPath string) (string, error) {
+	cfg, err := ini.Load(GitConfigPath(repoPath))
+	if err != nil {
+		return "", err
+	}
+	return cfg.Section("remote \"origin\"").Key("url").Value(), nil
+}
+
 func (m *Mirror) readAddress() {
 	if len(m.address) > 0 {
 		return
 	}
-
-	cfg, err := ini.Load(m.Repo.GitConfigPath())
+	var err error
+	m.address, err = remoteAddress(m.Repo.RepoPath())
 	if err != nil {
-		log.Error(4, "Load: %v", err)
-		return
+		log.Error(4, "remoteAddress: %v", err)
 	}
-	m.address = cfg.Section("remote \"origin\"").Key("url").Value()
 }
 
-// HandleCloneUserCredentials replaces user credentials from HTTP/HTTPS URL
-// with placeholder <credentials>.
-// It will fail for any other forms of clone addresses.
-func HandleCloneUserCredentials(url string, mosaics bool) string {
-	i := strings.Index(url, "@")
-	if i == -1 {
-		return url
-	}
-	start := strings.Index(url, "://")
-	if start == -1 {
-		return url
-	}
-	if mosaics {
-		return url[:start+3] + "<credentials>" + url[i:]
+// sanitizeOutput sanitizes output of a command, replacing occurrences of the
+// repository's remote address with a sanitized version.
+func sanitizeOutput(output, repoPath string) (string, error) {
+	remoteAddr, err := remoteAddress(repoPath)
+	if err != nil {
+		// if we're unable to load the remote address, then we're unable to
+		// sanitize.
+		return "", err
 	}
-	return url[:start+3] + url[i+1:]
+	return util.SanitizeMessage(output, remoteAddr), nil
 }
 
 // Address returns mirror address from Git repository config without credentials.
 func (m *Mirror) Address() string {
 	m.readAddress()
-	return HandleCloneUserCredentials(m.address, false)
+	return util.SanitizeURLCredentials(m.address, false)
 }
 
 // FullAddress returns mirror address from Git repository config.
@@ -145,7 +145,14 @@ func (m *Mirror) runSync() bool {
 	if _, stderr, err := process.GetManager().ExecDir(
 		timeout, repoPath, fmt.Sprintf("Mirror.runSync: %s", repoPath),
 		"git", gitArgs...); err != nil {
-		desc := fmt.Sprintf("Failed to update mirror repository '%s': %s", repoPath, stderr)
+		// sanitize the output, since it may contain the remote address, which may
+		// contain a password
+		message, err := sanitizeOutput(stderr, repoPath)
+		if err != nil {
+			log.Error(4, "sanitizeOutput: %v", err)
+			return false
+		}
+		desc := fmt.Sprintf("Failed to update mirror repository '%s': %s", repoPath, message)
 		log.Error(4, desc)
 		if err = CreateRepositoryNotice(desc); err != nil {
 			log.Error(4, "CreateRepositoryNotice: %v", err)
@@ -170,7 +177,14 @@ func (m *Mirror) runSync() bool {
 		if _, stderr, err := process.GetManager().ExecDir(
 			timeout, wikiPath, fmt.Sprintf("Mirror.runSync: %s", wikiPath),
 			"git", "remote", "update", "--prune"); err != nil {
-			desc := fmt.Sprintf("Failed to update mirror wiki repository '%s': %s", wikiPath, stderr)
+			// sanitize the output, since it may contain the remote address, which may
+			// contain a password
+			message, err := sanitizeOutput(stderr, wikiPath)
+			if err != nil {
+				log.Error(4, "sanitizeOutput: %v", err)
+				return false
+			}
+			desc := fmt.Sprintf("Failed to update mirror wiki repository '%s': %s", wikiPath, message)
 			log.Error(4, desc)
 			if err = CreateRepositoryNotice(desc); err != nil {
 				log.Error(4, "CreateRepositoryNotice: %v", err)

modules/util/sanitize.go

@@ -0,0 +1,48 @@
+// Copyright 2017 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package util
+
+import (
+	"net/url"
+	"strings"
+)
+
+// urlSafeError wraps an error whose message may contain a sensitive URL
+type urlSafeError struct {
+	err            error
+	unsanitizedURL string
+}
+
+func (err urlSafeError) Error() string {
+	return SanitizeMessage(err.err.Error(), err.unsanitizedURL)
+}
+
+// URLSanitizedError returns the sanitized version an error whose message may
+// contain a sensitive URL
+func URLSanitizedError(err error, unsanitizedURL string) error {
+	return urlSafeError{err: err, unsanitizedURL: unsanitizedURL}
+}
+
+// SanitizeMessage sanitizes a message which may contains a sensitive URL
+func SanitizeMessage(message, unsanitizedURL string) string {
+	sanitizedURL := SanitizeURLCredentials(unsanitizedURL, true)
+	return strings.Replace(message, unsanitizedURL, sanitizedURL, -1)
+}
+
+// SanitizeURLCredentials sanitizes a url, either removing user credentials
+// or replacing them with a placeholder.
+func SanitizeURLCredentials(unsanitizedURL string, usePlaceholder bool) string {
+	u, err := url.Parse(unsanitizedURL)
+	if err != nil {
+		// don't log the error, since it might contain unsanitized URL.
+		return "(unparsable url)"
+	}
+	if u.User != nil && usePlaceholder {
+		u.User = url.User("<credentials>")
+	} else {
+		u.User = nil
+	}
+	return u.String()
+}

routers/api/v1/repo/repo.go

@@ -9,15 +9,14 @@ import (
 	"net/http"
 	"strings"
 
-	api "code.gitea.io/sdk/gitea"
-
 	"code.gitea.io/gitea/models"
 	"code.gitea.io/gitea/modules/auth"
 	"code.gitea.io/gitea/modules/context"
 	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/setting"
 	"code.gitea.io/gitea/modules/util"
 	"code.gitea.io/gitea/routers/api/v1/convert"
+	api "code.gitea.io/sdk/gitea"
 )
 
 // Search repositories via options
@@ -322,12 +321,13 @@ func Migrate(ctx *context.APIContext, form auth.MigrateRepoForm) {
 		RemoteAddr:  remoteAddr,
 	})
 	if err != nil {
+		err = util.URLSanitizedError(err, remoteAddr)
 		if repo != nil {
 			if errDelete := models.DeleteRepository(ctx.User, ctxUser.ID, repo.ID); errDelete != nil {
 				log.Error(4, "DeleteRepository: %v", errDelete)
 			}
 		}
-		ctx.Error(500, "MigrateRepository", models.HandleCloneUserCredentials(err.Error(), true))
+		ctx.Error(500, "MigrateRepository", err)
 		return
 	}
 

routers/repo/repo.go

@@ -20,6 +20,7 @@ import (
 	"code.gitea.io/gitea/modules/context"
 	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/setting"
+	"code.gitea.io/gitea/modules/util"
 )
 
 const (
@@ -232,6 +233,9 @@ func MigratePost(ctx *context.Context, form auth.MigrateRepoForm) {
 		return
 	}
 
+	// remoteAddr may contain credentials, so we sanitize it
+	err = util.URLSanitizedError(err, remoteAddr)
+
 	if repo != nil {
 		if errDelete := models.DeleteRepository(ctx.User, ctxUser.ID, repo.ID); errDelete != nil {
 			log.Error(4, "DeleteRepository: %v", errDelete)
@@ -241,11 +245,11 @@ func MigratePost(ctx *context.Context, form auth.MigrateRepoForm) {
 	if strings.Contains(err.Error(), "Authentication failed") ||
 		strings.Contains(err.Error(), "could not read Username") {
 		ctx.Data["Err_Auth"] = true
-		ctx.RenderWithErr(ctx.Tr("form.auth_failed", models.HandleCloneUserCredentials(err.Error(), true)), tplMigrate, &form)
+		ctx.RenderWithErr(ctx.Tr("form.auth_failed", err.Error()), tplMigrate, &form)
 		return
 	} else if strings.Contains(err.Error(), "fatal:") {
 		ctx.Data["Err_CloneAddr"] = true
-		ctx.RenderWithErr(ctx.Tr("repo.migrate.failed", models.HandleCloneUserCredentials(err.Error(), true)), tplMigrate, &form)
+		ctx.RenderWithErr(ctx.Tr("repo.migrate.failed", err.Error()), tplMigrate, &form)
 		return
 	}