Multiple Escaping Improvements (#17551)

There are multiple places where Gitea does not properly escape URLs that it is building and there are multiple places where it builds urls when there is already a simpler function available to use this.
    
This is an extensive PR attempting to fix these issues.

1. The first commit in this PR looks through all href, src and links in the Gitea codebase and has attempted to catch all the places where there is potentially incomplete escaping.
2. Whilst doing this we will prefer to use functions that create URLs over recreating them by hand.
3. All uses of strings should be directly escaped - even if they are not currently expected to contain escaping characters. The main benefit to doing this will be that we can consider relaxing the constraints on user names and reponames in future. 
4. The next commit looks at escaping in the wiki and re-considers the urls that are used there. Using the improved escaping here wiki files containing '/'. (This implementation will currently still place all of the wiki files the root directory of the repo but this would not be difficult to change.)
5. The title generation in feeds is now properly escaped.
6. EscapePound is no longer needed - urls should be PathEscaped / QueryEscaped as necessary but then re-escaped with Escape when creating html with locales Signed-off-by: Andrew Thornton <art27@cantab.net>

Signed-off-by: Andrew Thornton <art27@cantab.net>

modules/context/context.go

@@ -70,6 +70,16 @@ type Context struct {
 	Org  *Organization
 }
 
+// TrHTMLEscapeArgs runs Tr but pre-escapes all arguments with html.EscapeString.
+// This is useful if the locale message is intended to only produce HTML content.
+func (ctx *Context) TrHTMLEscapeArgs(msg string, args ...string) string {
+	trArgs := make([]interface{}, len(args))
+	for i, arg := range args {
+		trArgs[i] = html.EscapeString(arg)
+	}
+	return ctx.Tr(msg, trArgs...)
+}
+
 // GetData returns the data
 func (ctx *Context) GetData() map[string]interface{} {
 	return ctx.Data
@@ -120,9 +130,9 @@ func RedirectToUser(ctx *Context, userName string, redirectUserID int64) {
 	}
 
 	redirectPath := strings.Replace(
-		ctx.Req.URL.Path,
-		userName,
-		user.Name,
+		ctx.Req.URL.EscapedPath(),
+		url.PathEscape(userName),
+		url.PathEscape(user.Name),
 		1,
 	)
 	if ctx.Req.URL.RawQuery != "" {

modules/context/repo.go

@@ -41,10 +41,10 @@ var IssueTemplateDirCandidates = []string{
 
 // PullRequest contains information to make a pull request
 type PullRequest struct {
-	BaseRepo *models.Repository
-	Allowed  bool
-	SameRepo bool
-	HeadInfo string // [<user>:]<branch>
+	BaseRepo       *models.Repository
+	Allowed        bool
+	SameRepo       bool
+	HeadInfoSubURL string // [<user>:]<branch> url segment
 }
 
 // Repository contains information to operate a repository
@@ -189,11 +189,11 @@ func (r *Repository) GetCommitGraphsCount(hidePRRefs bool, branches []string, fi
 func (r *Repository) BranchNameSubURL() string {
 	switch {
 	case r.IsViewBranch:
-		return "branch/" + r.BranchName
+		return "branch/" + util.PathEscapeSegments(r.BranchName)
 	case r.IsViewTag:
-		return "tag/" + r.BranchName
+		return "tag/" + util.PathEscapeSegments(r.BranchName)
 	case r.IsViewCommit:
-		return "commit/" + r.BranchName
+		return "commit/" + util.PathEscapeSegments(r.BranchName)
 	}
 	log.Error("Unknown view type for repo: %v", r)
 	return ""
@@ -321,9 +321,9 @@ func RedirectToRepo(ctx *Context, redirectRepoID int64) {
 	}
 
 	redirectPath := strings.Replace(
-		ctx.Req.URL.Path,
-		fmt.Sprintf("%s/%s", ownerName, previousRepoName),
-		repo.FullName(),
+		ctx.Req.URL.EscapedPath(),
+		url.PathEscape(ownerName)+"/"+url.PathEscape(previousRepoName),
+		url.PathEscape(repo.OwnerName)+"/"+url.PathEscape(repo.Name),
 		1,
 	)
 	if ctx.Req.URL.RawQuery != "" {
@@ -588,7 +588,7 @@ func RepoAssignment(ctx *Context) (cancel context.CancelFunc) {
 		ctx.Data["BaseRepo"] = repo.BaseRepo
 		ctx.Repo.PullRequest.BaseRepo = repo.BaseRepo
 		ctx.Repo.PullRequest.Allowed = canPush
-		ctx.Repo.PullRequest.HeadInfo = ctx.Repo.Owner.Name + ":" + ctx.Repo.BranchName
+		ctx.Repo.PullRequest.HeadInfoSubURL = url.PathEscape(ctx.Repo.Owner.Name) + ":" + util.PathEscapeSegments(ctx.Repo.BranchName)
 	} else if repo.AllowsPulls() {
 		// Or, this is repository accepts pull requests between branches.
 		canCompare = true
@@ -596,7 +596,7 @@ func RepoAssignment(ctx *Context) (cancel context.CancelFunc) {
 		ctx.Repo.PullRequest.BaseRepo = repo
 		ctx.Repo.PullRequest.Allowed = canPush
 		ctx.Repo.PullRequest.SameRepo = true
-		ctx.Repo.PullRequest.HeadInfo = ctx.Repo.BranchName
+		ctx.Repo.PullRequest.HeadInfoSubURL = util.PathEscapeSegments(ctx.Repo.BranchName)
 	}
 	ctx.Data["CanCompareOrPull"] = canCompare
 	ctx.Data["PullRequestCtx"] = ctx.Repo.PullRequest
@@ -621,7 +621,7 @@ func RepoAssignment(ctx *Context) (cancel context.CancelFunc) {
 
 	if ctx.FormString("go-get") == "1" {
 		ctx.Data["GoGetImport"] = ComposeGoGetImport(owner.Name, repo.Name)
-		prefix := setting.AppURL + path.Join(owner.Name, repo.Name, "src", "branch", ctx.Repo.BranchName)
+		prefix := repo.HTMLURL() + "/src/branch/" + util.PathEscapeSegments(ctx.Repo.BranchName)
 		ctx.Data["GoDocDirectory"] = prefix + "{/dir}"
 		ctx.Data["GoDocFile"] = prefix + "{/dir}/{file}#L{line}"
 	}
@@ -810,7 +810,7 @@ func RepoRefByType(refType RepoRefType, ignoreNotExistErr ...bool) func(*Context
 			if isRenamedBranch && has {
 				renamedBranchName := ctx.Data["RenamedBranchName"].(string)
 				ctx.Flash.Info(ctx.Tr("repo.branch.renamed", refName, renamedBranchName))
-				link := strings.Replace(ctx.Req.RequestURI, refName, renamedBranchName, 1)
+				link := setting.AppSubURL + strings.Replace(ctx.Req.URL.EscapedPath(), util.PathEscapeSegments(refName), util.PathEscapeSegments(renamedBranchName), 1)
 				ctx.Redirect(link)
 				return
 			}
@@ -845,7 +845,7 @@ func RepoRefByType(refType RepoRefType, ignoreNotExistErr ...bool) func(*Context
 				// If short commit ID add canonical link header
 				if len(refName) < 40 {
 					ctx.Header().Set("Link", fmt.Sprintf("<%s>; rel=\"canonical\"",
-						util.URLJoin(setting.AppURL, strings.Replace(ctx.Req.URL.RequestURI(), refName, ctx.Repo.Commit.ID.String(), 1))))
+						util.URLJoin(setting.AppURL, strings.Replace(ctx.Req.URL.RequestURI(), util.PathEscapeSegments(refName), url.PathEscape(ctx.Repo.Commit.ID.String()), 1))))
 				}
 			} else {
 				if len(ignoreNotExistErr) > 0 && ignoreNotExistErr[0] {
@@ -857,11 +857,13 @@ func RepoRefByType(refType RepoRefType, ignoreNotExistErr ...bool) func(*Context
 
 			if refType == RepoRefLegacy {
 				// redirect from old URL scheme to new URL scheme
+				prefix := strings.TrimPrefix(setting.AppSubURL+strings.TrimSuffix(ctx.Req.URL.Path, ctx.Params("*")), ctx.Repo.RepoLink)
+
 				ctx.Redirect(path.Join(
-					setting.AppSubURL,
-					strings.TrimSuffix(ctx.Req.URL.Path, ctx.Params("*")),
+					ctx.Repo.RepoLink,
+					util.PathEscapeSegments(prefix),
 					ctx.Repo.BranchNameSubURL(),
-					ctx.Repo.TreePath))
+					util.PathEscapeSegments(ctx.Repo.TreePath)))
 				return
 			}
 		}

modules/convert/git_commit.go

@@ -5,6 +5,7 @@
 package convert
 
 import (
+	"net/url"
 	"time"
 
 	"code.gitea.io/gitea/models"
@@ -126,7 +127,7 @@ func ToCommit(repo *models.Repository, commit *git.Commit, userCache map[string]
 	for i := 0; i < commit.ParentCount(); i++ {
 		sha, _ := commit.ParentID(i)
 		apiParents[i] = &api.CommitMeta{
-			URL: repo.APIURL() + "/git/commits/" + sha.String(),
+			URL: repo.APIURL() + "/git/commits/" + url.PathEscape(sha.String()),
 			SHA: sha.String(),
 		}
 	}
@@ -147,13 +148,13 @@ func ToCommit(repo *models.Repository, commit *git.Commit, userCache map[string]
 
 	return &api.Commit{
 		CommitMeta: &api.CommitMeta{
-			URL:     repo.APIURL() + "/git/commits/" + commit.ID.String(),
+			URL:     repo.APIURL() + "/git/commits/" + url.PathEscape(commit.ID.String()),
 			SHA:     commit.ID.String(),
 			Created: commit.Committer.When,
 		},
-		HTMLURL: repo.HTMLURL() + "/commit/" + commit.ID.String(),
+		HTMLURL: repo.HTMLURL() + "/commit/" + url.PathEscape(commit.ID.String()),
 		RepoCommit: &api.RepoCommit{
-			URL: repo.APIURL() + "/git/commits/" + commit.ID.String(),
+			URL: repo.APIURL() + "/git/commits/" + url.PathEscape(commit.ID.String()),
 			Author: &api.CommitUser{
 				Identity: api.Identity{
 					Name:  commit.Author.Name,
@@ -170,7 +171,7 @@ func ToCommit(repo *models.Repository, commit *git.Commit, userCache map[string]
 			},
 			Message: commit.Message(),
 			Tree: &api.CommitMeta{
-				URL:     repo.APIURL() + "/git/trees/" + commit.ID.String(),
+				URL:     repo.APIURL() + "/git/trees/" + url.PathEscape(commit.ID.String()),
 				SHA:     commit.ID.String(),
 				Created: commit.Committer.When,
 			},

modules/convert/issue.go

@@ -6,6 +6,7 @@ package convert
 
 import (
 	"fmt"
+	"net/url"
 	"strings"
 
 	"code.gitea.io/gitea/models"
@@ -191,7 +192,7 @@ func ToLabel(label *models.Label, repo *models.Repository, org *models.User) *ap
 		}
 	} else { // BelongsToOrg
 		if org != nil {
-			result.URL = fmt.Sprintf("%sapi/v1/orgs/%s/labels/%d", setting.AppURL, org.Name, label.ID)
+			result.URL = fmt.Sprintf("%sapi/v1/orgs/%s/labels/%d", setting.AppURL, url.PathEscape(org.Name), label.ID)
 		} else {
 			log.Error("ToLabel did not get org to calculate url for label with id '%d'", label.ID)
 		}

modules/convert/notification.go

@@ -5,6 +5,8 @@
 package convert
 
 import (
+	"net/url"
+
 	"code.gitea.io/gitea/models"
 	api "code.gitea.io/gitea/modules/structs"
 )
@@ -58,7 +60,7 @@ func ToNotificationThread(n *models.Notification) *api.NotificationThread {
 			}
 		}
 	case models.NotificationSourceCommit:
-		url := n.Repository.HTMLURL() + "/commit/" + n.CommitID
+		url := n.Repository.HTMLURL() + "/commit/" + url.PathEscape(n.CommitID)
 		result.Subject = &api.NotificationSubject{
 			Type:    api.NotifySubjectCommit,
 			Title:   n.CommitID,

modules/git/utils.go

@@ -11,6 +11,8 @@ import (
 	"strconv"
 	"strings"
 	"sync"
+
+	"code.gitea.io/gitea/modules/util"
 )
 
 // ObjectCache provides thread-safe cache operations.
@@ -92,7 +94,7 @@ func RefEndName(refStr string) string {
 
 // RefURL returns the absolute URL for a ref in a repository
 func RefURL(repoURL, ref string) string {
-	refName := RefEndName(ref)
+	refName := util.PathEscapeSegments(RefEndName(ref))
 	switch {
 	case strings.HasPrefix(ref, BranchPrefix):
 		return repoURL + "/src/branch/" + refName

modules/repofiles/action.go

@@ -7,6 +7,7 @@ package repofiles
 import (
 	"fmt"
 	"html"
+	"net/url"
 	"regexp"
 	"strconv"
 	"strings"
@@ -175,7 +176,7 @@ func UpdateIssuesCommit(doer *models.User, repo *models.Repository, commits []*r
 				continue
 			}
 
-			message := fmt.Sprintf(`<a href="%s/commit/%s">%s</a>`, repo.Link(), c.Sha1, html.EscapeString(strings.SplitN(c.Message, "\n", 2)[0]))
+			message := fmt.Sprintf(`<a href="%s/commit/%s">%s</a>`, html.EscapeString(repo.Link()), html.EscapeString(url.PathEscape(c.Sha1)), html.EscapeString(strings.SplitN(c.Message, "\n", 2)[0]))
 			if err = models.CreateRefComment(doer, refRepo, refIssue, message, c.Sha1); err != nil {
 				return err
 			}

modules/repofiles/blob.go

@@ -5,6 +5,8 @@
 package repofiles
 
 import (
+	"net/url"
+
 	"code.gitea.io/gitea/models"
 	"code.gitea.io/gitea/modules/git"
 	"code.gitea.io/gitea/modules/setting"
@@ -31,7 +33,7 @@ func GetBlobBySHA(repo *models.Repository, sha string) (*api.GitBlobResponse, er
 	}
 	return &api.GitBlobResponse{
 		SHA:      gitBlob.ID.String(),
-		URL:      repo.APIURL() + "/git/blobs/" + gitBlob.ID.String(),
+		URL:      repo.APIURL() + "/git/blobs/" + url.PathEscape(gitBlob.ID.String()),
 		Size:     gitBlob.Size(),
 		Encoding: "base64",
 		Content:  content,

modules/repofiles/file.go

@@ -36,19 +36,19 @@ func GetFileCommitResponse(repo *models.Repository, commit *git.Commit) (*api.Fi
 	if commit == nil {
 		return nil, fmt.Errorf("commit cannot be nil")
 	}
-	commitURL, _ := url.Parse(repo.APIURL() + "/git/commits/" + commit.ID.String())
-	commitTreeURL, _ := url.Parse(repo.APIURL() + "/git/trees/" + commit.Tree.ID.String())
+	commitURL, _ := url.Parse(repo.APIURL() + "/git/commits/" + url.PathEscape(commit.ID.String()))
+	commitTreeURL, _ := url.Parse(repo.APIURL() + "/git/trees/" + url.PathEscape(commit.Tree.ID.String()))
 	parents := make([]*api.CommitMeta, commit.ParentCount())
 	for i := 0; i <= commit.ParentCount(); i++ {
 		if parent, err := commit.Parent(i); err == nil && parent != nil {
-			parentCommitURL, _ := url.Parse(repo.APIURL() + "/git/commits/" + parent.ID.String())
+			parentCommitURL, _ := url.Parse(repo.APIURL() + "/git/commits/" + url.PathEscape(parent.ID.String()))
 			parents[i] = &api.CommitMeta{
 				SHA: parent.ID.String(),
 				URL: parentCommitURL.String(),
 			}
 		}
 	}
-	commitHTMLURL, _ := url.Parse(repo.HTMLURL() + "/commit/" + commit.ID.String())
+	commitHTMLURL, _ := url.Parse(repo.HTMLURL() + "/commit/" + url.PathEscape(commit.ID.String()))
 	fileCommit := &api.FileCommitResponse{
 		CommitMeta: api.CommitMeta{
 			SHA: commit.ID.String(),

modules/repofiles/tree.go

@@ -6,6 +6,7 @@ package repofiles
 
 import (
 	"fmt"
+	"net/url"
 
 	"code.gitea.io/gitea/models"
 	"code.gitea.io/gitea/modules/git"
@@ -28,7 +29,7 @@ func GetTreeBySHA(repo *models.Repository, sha string, page, perPage int, recurs
 	}
 	tree := new(api.GitTreeResponse)
 	tree.SHA = gitTree.ResolvedID.String()
-	tree.URL = repo.APIURL() + "/git/trees/" + tree.SHA
+	tree.URL = repo.APIURL() + "/git/trees/" + url.PathEscape(tree.SHA)
 	var entries git.Entries
 	if recursive {
 		entries, err = gitTree.ListEntriesRecursive()

modules/repository/commits.go

@@ -6,6 +6,7 @@ package repository
 
 import (
 	"fmt"
+	"net/url"
 	"time"
 
 	"code.gitea.io/gitea/models"
@@ -81,7 +82,7 @@ func (pc *PushCommits) toAPIPayloadCommit(repoPath, repoLink string, commit *Pus
 	return &api.PayloadCommit{
 		ID:      commit.Sha1,
 		Message: commit.Message,
-		URL:     fmt.Sprintf("%s/commit/%s", repoLink, commit.Sha1),
+		URL:     fmt.Sprintf("%s/commit/%s", repoLink, url.PathEscape(commit.Sha1)),
 		Author: &api.PayloadUser{
 			Name:     commit.AuthorName,
 			Email:    commit.AuthorEmail,

modules/templates/helper.go

@@ -139,17 +139,14 @@ func NewFuncMap() []template.FuncMap {
 			}
 			return str[start:end]
 		},
-		"EllipsisString":        base.EllipsisString,
-		"DiffTypeToStr":         DiffTypeToStr,
-		"DiffLineTypeToStr":     DiffLineTypeToStr,
-		"Sha1":                  Sha1,
-		"ShortSha":              base.ShortSha,
-		"MD5":                   base.EncodeMD5,
-		"ActionContent2Commits": ActionContent2Commits,
-		"PathEscape":            url.PathEscape,
-		"EscapePound": func(str string) string {
-			return strings.NewReplacer("%", "%25", "#", "%23", " ", "%20", "?", "%3F").Replace(str)
-		},
+		"EllipsisString":                 base.EllipsisString,
+		"DiffTypeToStr":                  DiffTypeToStr,
+		"DiffLineTypeToStr":              DiffLineTypeToStr,
+		"Sha1":                           Sha1,
+		"ShortSha":                       base.ShortSha,
+		"MD5":                            base.EncodeMD5,
+		"ActionContent2Commits":          ActionContent2Commits,
+		"PathEscape":                     url.PathEscape,
 		"PathEscapeSegments":             util.PathEscapeSegments,
 		"URLJoin":                        util.URLJoin,
 		"RenderCommitMessage":            RenderCommitMessage,
@@ -742,7 +739,7 @@ func ReactionToEmoji(reaction string) template.HTML {
 	if val != nil {
 		return template.HTML(val.Emoji)
 	}
-	return template.HTML(fmt.Sprintf(`<img alt=":%s:" src="%s/assets/img/emoji/%s.png"></img>`, reaction, setting.StaticURLPrefix, reaction))
+	return template.HTML(fmt.Sprintf(`<img alt=":%s:" src="%s/assets/img/emoji/%s.png"></img>`, reaction, setting.StaticURLPrefix, url.PathEscape(reaction)))
 }
 
 // RenderNote renders the contents of a git-notes file as a commit message.

modules/upload/upload.go

@@ -6,6 +6,7 @@ package upload
 
 import (
 	"net/http"
+	"net/url"
 	"path"
 	"regexp"
 	"strings"
@@ -83,7 +84,7 @@ func AddUploadContext(ctx *context.Context, uploadType string) {
 		ctx.Data["UploadUrl"] = ctx.Repo.RepoLink + "/issues/attachments"
 		ctx.Data["UploadRemoveUrl"] = ctx.Repo.RepoLink + "/issues/attachments/remove"
 		if len(ctx.Params(":index")) > 0 {
-			ctx.Data["UploadLinkUrl"] = ctx.Repo.RepoLink + "/issues/" + ctx.Params(":index") + "/attachments"
+			ctx.Data["UploadLinkUrl"] = ctx.Repo.RepoLink + "/issues/" + url.PathEscape(ctx.Params(":index")) + "/attachments"
 		} else {
 			ctx.Data["UploadLinkUrl"] = ctx.Repo.RepoLink + "/issues/attachments"
 		}