Add codeowners feature (#24910)

Hello.
This PR adds a github like configuration for the CODEOWNERS file.

Resolves: #10161

6 files changed, 329 insertions, 0 deletions

diff --git a/docs/content/doc/usage/code-owners.en-us.md b/docs/content/doc/usage/code-owners.en-us.md
new file mode 100644
index 0000000000..94f81eeae1
--- /dev/null
+++ b/docs/content/doc/usage/code-owners.en-us.md
@@ -0,0 +1,65 @@
+---
+date: "2023-05-24T16:00:00+00:00"
+title: "Code Owners"
+slug: "code-owners"
+weight: 30
+toc: false
+draft: false
+aliases:
+  - /en-us/code-owners
+menu:
+  sidebar:
+    parent: "usage"
+    name: "Code Owners"
+    weight: 30
+    identifier: "code-owners"
+---
+
+# Code Owners
+
+Gitea maintains code owner files. It looks for it in the following locations in this order:
+
+- `./CODEOWNERS`
+- `./docs/CODEOWNERS`
+- `./.gitea/CODEOWNERS`
+
+And stops at the first found file.
+
+File format: `<regexp rule> <@user or @org/team> [@user or @org/team]...`
+
+Regexp specified in golang Regex format.
+Regexp can start with `!` for negative rules - match all files except specified.
+
+Example file:
+
+```
+.*\\.go @user1 @user2 # This is comment
+
+# Comment too
+# You can assigning code owning for users or teams
+frontend/src/.*\\.js @org1/team1 @org1/team2 @user3
+
+# You can use negative pattern
+!frontend/src/.* @org1/team3 @user5
+
+# You can use power of go regexp
+docs/(aws|google|azure)/[^/]*\\.(md|txt) @user8 @org1/team4
+!/assets/.*\\.(bin|exe|msi) @user9
+```
+
+### Escaping
+
+You can escape characters `#`, ` ` (space) and `\` with `\`, like:
+
+```
+dir/with\#hashtag @user1
+path\ with\ space @user2
+path/with\\backslash @user3
+```
+
+Some character (`.+*?()|[]{}^$\`) should be escaped with `\\` inside regexp, like:
+
+```
+path/\\.with\\.dots
+path/with\\+plus
+```
diff --git a/models/issues/pull.go b/models/issues/pull.go
index 3f37d8d243..63bc258d2f 100644
--- a/models/issues/pull.go
+++ b/models/issues/pull.go
@@ -8,11 +8,13 @@ import (
 	"context"
 	"fmt"
 	"io"
+	"regexp"
 	"strconv"
 	"strings"
 
 	"code.gitea.io/gitea/models/db"
 	git_model "code.gitea.io/gitea/models/git"
+	org_model "code.gitea.io/gitea/models/organization"
 	pull_model "code.gitea.io/gitea/models/pull"
 	repo_model "code.gitea.io/gitea/models/repo"
 	user_model "code.gitea.io/gitea/models/user"
@@ -887,3 +889,222 @@ func MergeBlockedByOfficialReviewRequests(ctx context.Context, protectBranch *gi
 func MergeBlockedByOutdatedBranch(protectBranch *git_model.ProtectedBranch, pr *PullRequest) bool {
 	return protectBranch.BlockOnOutdatedBranch && pr.CommitsBehind > 0
 }
+
+func PullRequestCodeOwnersReview(ctx context.Context, pull *Issue, pr *PullRequest) error {
+	files := []string{"CODEOWNERS", "docs/CODEOWNERS", ".gitea/CODEOWNERS"}
+
+	if pr.IsWorkInProgress() {
+		return nil
+	}
+
+	if err := pr.LoadBaseRepo(ctx); err != nil {
+		return err
+	}
+
+	repo, err := git.OpenRepository(ctx, pr.BaseRepo.RepoPath())
+	if err != nil {
+		return err
+	}
+	defer repo.Close()
+
+	branch, err := repo.GetDefaultBranch()
+	if err != nil {
+		return err
+	}
+
+	commit, err := repo.GetBranchCommit(branch)
+	if err != nil {
+		return err
+	}
+
+	var data string
+	for _, file := range files {
+		if blob, err := commit.GetBlobByPath(file); err == nil {
+			data, err = blob.GetBlobContent()
+			if err == nil {
+				break
+			}
+		}
+	}
+
+	rules, _ := GetCodeOwnersFromContent(ctx, data)
+	changedFiles, err := repo.GetFilesChangedBetween(git.BranchPrefix+pr.BaseBranch, pr.GetGitRefName())
+	if err != nil {
+		return err
+	}
+
+	uniqUsers := make(map[int64]*user_model.User)
+	uniqTeams := make(map[string]*org_model.Team)
+	for _, rule := range rules {
+		for _, f := range changedFiles {
+			if (rule.Rule.MatchString(f) && !rule.Negative) || (!rule.Rule.MatchString(f) && rule.Negative) {
+				for _, u := range rule.Users {
+					uniqUsers[u.ID] = u
+				}
+				for _, t := range rule.Teams {
+					uniqTeams[fmt.Sprintf("%d/%d", t.OrgID, t.ID)] = t
+				}
+			}
+		}
+	}
+
+	for _, u := range uniqUsers {
+		if u.ID != pull.Poster.ID {
+			if _, err := AddReviewRequest(pull, u, pull.Poster); err != nil {
+				log.Warn("Failed add assignee user: %s to PR review: %s#%d, error: %s", u.Name, pr.BaseRepo.Name, pr.ID, err)
+				return err
+			}
+		}
+	}
+	for _, t := range uniqTeams {
+		if _, err := AddTeamReviewRequest(pull, t, pull.Poster); err != nil {
+			log.Warn("Failed add assignee team: %s to PR review: %s#%d, error: %s", t.Name, pr.BaseRepo.Name, pr.ID, err)
+			return err
+		}
+	}
+
+	return nil
+}
+
+// GetCodeOwnersFromContent returns the code owners configuration
+// Return empty slice if files missing
+// Return warning messages on parsing errors
+// We're trying to do the best we can when parsing a file.
+// Invalid lines are skipped. Non-existent users and teams too.
+func GetCodeOwnersFromContent(ctx context.Context, data string) ([]*CodeOwnerRule, []string) {
+	if len(data) == 0 {
+		return nil, nil
+	}
+
+	rules := make([]*CodeOwnerRule, 0)
+	lines := strings.Split(data, "\n")
+	warnings := make([]string, 0)
+
+	for i, line := range lines {
+		tokens := TokenizeCodeOwnersLine(line)
+		if len(tokens) == 0 {
+			continue
+		} else if len(tokens) < 2 {
+			warnings = append(warnings, fmt.Sprintf("Line: %d: incorrect format", i+1))
+			continue
+		}
+		rule, wr := ParseCodeOwnersLine(ctx, tokens)
+		for _, w := range wr {
+			warnings = append(warnings, fmt.Sprintf("Line: %d: %s", i+1, w))
+		}
+		if rule == nil {
+			continue
+		}
+
+		rules = append(rules, rule)
+	}
+
+	return rules, warnings
+}
+
+type CodeOwnerRule struct {
+	Rule     *regexp.Regexp
+	Negative bool
+	Users    []*user_model.User
+	Teams    []*org_model.Team
+}
+
+func ParseCodeOwnersLine(ctx context.Context, tokens []string) (*CodeOwnerRule, []string) {
+	var err error
+	rule := &CodeOwnerRule{
+		Users:    make([]*user_model.User, 0),
+		Teams:    make([]*org_model.Team, 0),
+		Negative: strings.HasPrefix(tokens[0], "!"),
+	}
+
+	warnings := make([]string, 0)
+
+	rule.Rule, err = regexp.Compile(fmt.Sprintf("^%s$", strings.TrimPrefix(tokens[0], "!")))
+	if err != nil {
+		warnings = append(warnings, fmt.Sprintf("incorrect codeowner regexp: %s", err))
+		return nil, warnings
+	}
+
+	for _, user := range tokens[1:] {
+		user = strings.TrimPrefix(user, "@")
+
+		// Only @org/team can contain slashes
+		if strings.Contains(user, "/") {
+			s := strings.Split(user, "/")
+			if len(s) != 2 {
+				warnings = append(warnings, fmt.Sprintf("incorrect codeowner group: %s", user))
+				continue
+			}
+			orgName := s[0]
+			teamName := s[1]
+
+			org, err := org_model.GetOrgByName(ctx, orgName)
+			if err != nil {
+				warnings = append(warnings, fmt.Sprintf("incorrect codeowner organization: %s", user))
+				continue
+			}
+			teams, err := org.LoadTeams()
+			if err != nil {
+				warnings = append(warnings, fmt.Sprintf("incorrect codeowner team: %s", user))
+				continue
+			}
+
+			for _, team := range teams {
+				if team.Name == teamName {
+					rule.Teams = append(rule.Teams, team)
+				}
+			}
+		} else {
+			u, err := user_model.GetUserByName(ctx, user)
+			if err != nil {
+				warnings = append(warnings, fmt.Sprintf("incorrect codeowner user: %s", user))
+				continue
+			}
+			rule.Users = append(rule.Users, u)
+		}
+	}
+
+	if (len(rule.Users) == 0) && (len(rule.Teams) == 0) {
+		warnings = append(warnings, "no users/groups matched")
+		return nil, warnings
+	}
+
+	return rule, warnings
+}
+
+func TokenizeCodeOwnersLine(line string) []string {
+	if len(line) == 0 {
+		return nil
+	}
+
+	line = strings.TrimSpace(line)
+	line = strings.ReplaceAll(line, "\t", " ")
+
+	tokens := make([]string, 0)
+
+	escape := false
+	token := ""
+	for _, char := range line {
+		if escape {
+			token += string(char)
+			escape = false
+		} else if string(char) == "\\" {
+			escape = true
+		} else if string(char) == "#" {
+			break
+		} else if string(char) == " " {
+			if len(token) > 0 {
+				tokens = append(tokens, token)
+				token = ""
+			}
+		} else {
+			token += string(char)
+		}
+	}
+
+	if len(token) > 0 {
+		tokens = append(tokens, token)
+	}
+
+	return tokens
+}
diff --git a/models/issues/pull_test.go b/models/issues/pull_test.go
index 7a183ac312..1eb106047c 100644
--- a/models/issues/pull_test.go
+++ b/models/issues/pull_test.go
@@ -303,3 +303,25 @@ func TestDeleteOrphanedObjects(t *testing.T) {
 	assert.NoError(t, err)
 	assert.EqualValues(t, countBefore, countAfter)
 }
+
+func TestParseCodeOwnersLine(t *testing.T) {
+	type CodeOwnerTest struct {
+		Line   string
+		Tokens []string
+	}
+
+	given := []CodeOwnerTest{
+		{Line: "", Tokens: nil},
+		{Line: "# comment", Tokens: []string{}},
+		{Line: "!.* @user1 @org1/team1", Tokens: []string{"!.*", "@user1", "@org1/team1"}},
+		{Line: `.*\\.js @user2 #comment`, Tokens: []string{`.*\.js`, "@user2"}},
+		{Line: `docs/(aws|google|azure)/[^/]*\\.(md|txt) @user3 @org2/team2`, Tokens: []string{`docs/(aws|google|azure)/[^/]*\.(md|txt)`, "@user3", "@org2/team2"}},
+		{Line: `\#path @user3`, Tokens: []string{`#path`, "@user3"}},
+		{Line: `path\ with\ spaces/ @user3`, Tokens: []string{`path with spaces/`, "@user3"}},
+	}
+
+	for _, g := range given {
+		tokens := issues_model.TokenizeCodeOwnersLine(g.Line)
+		assert.Equal(t, g.Tokens, tokens, "Codeowners tokenizer failed")
+	}
+}
diff --git a/routers/web/repo/view.go b/routers/web/repo/view.go
index 2fd893f91c..1d54f25884 100644
--- a/routers/web/repo/view.go
+++ b/routers/web/repo/view.go
@@ -21,6 +21,7 @@ import (
 	asymkey_model "code.gitea.io/gitea/models/asymkey"
 	"code.gitea.io/gitea/models/db"
 	git_model "code.gitea.io/gitea/models/git"
+	issue_model "code.gitea.io/gitea/models/issues"
 	repo_model "code.gitea.io/gitea/models/repo"
 	unit_model "code.gitea.io/gitea/models/unit"
 	user_model "code.gitea.io/gitea/models/user"
@@ -361,6 +362,13 @@ func renderFile(ctx *context.Context, entry *git.TreeEntry, treeLink, rawLink st
 		if workFlowErr != nil {
 			ctx.Data["FileError"] = ctx.Locale.Tr("actions.runs.invalid_workflow_helper", workFlowErr.Error())
 		}
+	} else if util.SliceContains([]string{"CODEOWNERS", "docs/CODEOWNERS", ".gitea/CODEOWNERS"}, ctx.Repo.TreePath) {
+		if data, err := blob.GetBlobContent(); err == nil {
+			_, warnings := issue_model.GetCodeOwnersFromContent(ctx, data)
+			if len(warnings) > 0 {
+				ctx.Data["FileWarning"] = strings.Join(warnings, "\n")
+			}
+		}
 	}
 
 	isDisplayingSource := ctx.FormString("display") == "source"
diff --git a/services/issue/issue.go b/services/issue/issue.go
index ce2b1c0976..61890c75de 100644
--- a/services/issue/issue.go
+++ b/services/issue/issue.go
@@ -57,6 +57,12 @@ func ChangeTitle(ctx context.Context, issue *issues_model.Issue, doer *user_mode
 		return
 	}
 
+	if issue.IsPull && issues_model.HasWorkInProgressPrefix(oldTitle) && !issues_model.HasWorkInProgressPrefix(title) {
+		if err = issues_model.PullRequestCodeOwnersReview(ctx, issue, issue.PullRequest); err != nil {
+			return
+		}
+	}
+
 	notification.NotifyIssueChangeTitle(ctx, doer, issue, oldTitle)
 
 	return nil
diff --git a/services/pull/pull.go b/services/pull/pull.go
index 8f2befa8ff..f44e690ab7 100644
--- a/services/pull/pull.go
+++ b/services/pull/pull.go
@@ -123,6 +123,13 @@ func NewPullRequest(ctx context.Context, repo *repo_model.Repository, pull *issu
 		}
 
 		_, _ = issue_service.CreateComment(ctx, ops)
+
+		if !pr.IsWorkInProgress() {
+			if err := issues_model.PullRequestCodeOwnersReview(ctx, pull, pr); err != nil {
+				return err
+			}
+		}
+
 	}
 
 	return nil