perf(git): speed up activity top authors and subdirectory listings (#39526)

Speeds up two slow git paths. Results match `main` apart
from the `log.follow` fix.

- Activity top authors skip `--numstat` and an unused `rev-list
--count`, they only need names and emails
- Subdirectory listings pass only the directory as pathspec, which
already matches all its entries
- Directory listings pass `--no-follow` to `git log`, a configured
`log.follow` disabled parent rewriting and gave wrong last commits

| Benchmark | main | PR | Change |
|---|--:|--:|--:|
| Top authors, gitea, 1 month | 632 ms | 21 ms | -97% |
| Top authors, gitea, 1 year | 3216 ms | 68 ms | -98% |
| Top authors, tea | 62 ms | 12 ms | -80% |
| Listing, git `Documentation/technical` | 189 ms | 91 ms | -52% |
| Listing, gitea `options/license` | 329 ms | 208 ms | -37% |
| Listing, gitea `templates/repo` | 294 ms | 223 ms | -24% |

Tested with unit and sqlite integration tests in default and `gogit`
builds on git 2.25 and 2.56, and by comparing listing results and cache
writes with `main` on randomized histories and the gitea, tea and git
repos. Benchmarks are medians of 8 interleaved macOS runs.

AGENTS.md

@@ -12,6 +12,7 @@
 - In `options/locale`, only edit `locale_en-US.json`, other locales are synced automatically
 - In TS, use `!` instead of `?.`/`??` when a value always exists
 - In Go, prefer to use modern language features wherever possible
+- In Go, function-name prefixes in errors like `fmt.Errorf("Foo: %w", err)` must always name the function they are in
 - Write sizes as multiplications like `64 * 1024`, not bit shifts like `64 << 10`
 - Prefer `tw-*` utilities over inline `style` and `flex-*` helpers over per-child `tw-ml-*`/`tw-mr-*` margins, falling back to `tw-*` where specificity requires `!important`
 - Run `make fmt` after `.go` edits, `make tidy` after `go.mod` edits, `make generate-swagger` after API changes, and lint what changed with `make lint-go`, `lint-js`, `lint-css` or `lint-templates`

models/activities/repo_activity.go

@@ -83,23 +83,17 @@ func GetActivityStats(ctx context.Context, repo *repo_model.Repository, timeFrom
 
 // GetActivityStatsTopAuthors returns top author stats for git commits for all branches
 func GetActivityStatsTopAuthors(ctx context.Context, repo *repo_model.Repository, timeFrom time.Time, count int) ([]*ActivityAuthorData, error) {
-	gitRepo, closer, err := git.RepositoryFromContextOrOpen(ctx, repo)
-	if err != nil {
-		return nil, fmt.Errorf("OpenRepository: %w", err)
-	}
-	defer closer.Close()
-
-	code, err := gitRepo.GetCodeActivityStats(ctx, timeFrom, "")
+	authors, err := git.GetCodeActivityAuthors(ctx, repo, timeFrom)
 	if err != nil {
 		return nil, fmt.Errorf("FillFromGit: %w", err)
 	}
-	if code.Authors == nil {
+	if authors == nil {
 		return nil, nil
 	}
 	users := make(map[int64]*ActivityAuthorData)
 	var unknownUserID int64
 	unknownUserAvatarLink := user_model.NewGhostUser().AvatarLink(ctx)
-	for _, v := range code.Authors {
+	for _, v := range authors {
 		if len(v.Email) == 0 {
 			continue
 		}

modules/git/log_name_status_nogogit.go

@@ -23,26 +23,17 @@ import (
 // logNameStatusRepo opens git log --raw in the provided repo and returns a parser
 func logNameStatusRepo(ctx context.Context, repo RepositoryFacade, head, treepath string, paths ...string) *logNameStatusRepoParser {
 	cmd := gitcmd.NewCommand()
-	cmd.AddArguments("log", "--name-status", "-c", "--format=commit%x00%H %P%x00", "--parents", "--no-renames", "-t", "-z").AddDynamicArguments(head)
+	cmd.AddArguments("log", "--name-status", "-c", "--format=commit%x00%H %P%x00", "--parents", "--no-renames", "--no-follow", "-t", "-z").AddDynamicArguments(head)
 
 	var files []string
-	if len(paths) < 70 {
-		if treepath != "" {
-			files = append(files, treepath)
-			for _, pth := range paths {
-				if pth != "" {
-					files = append(files, path.Join(treepath, pth))
-				}
-			}
-		} else {
-			for _, pth := range paths {
-				if pth != "" {
-					files = append(files, pth)
-				}
+	if treepath != "" {
+		files = append(files, treepath) // matches all entries, adding them would not narrow the log
+	} else if len(paths) < 70 {
+		for _, pth := range paths {
+			if pth != "" {
+				files = append(files, pth)
 			}
 		}
-	} else if treepath != "" {
-		files = append(files, treepath)
 	}
 	// Use the :(literal) pathspec magic to handle edge cases with files named like ":file.txt" or "*.jpg"
 	for i, file := range files {

modules/git/repo_stats.go

@@ -61,10 +61,28 @@ func (repo *Repository) GetCodeActivityStats(ctx context.Context, fromTime time.
 	} else {
 		gitCmd.AddArguments("--first-parent").AddDynamicArguments(branch)
 	}
+	if err := parseCodeActivityLog(ctx, repo, gitCmd, stats); err != nil {
+		return nil, err
+	}
 
+	return stats, nil
+}
+
+// GetCodeActivityAuthors returns the authors of non-merge commits in all branches since fromTime, sorted by commit count
+func GetCodeActivityAuthors(ctx context.Context, repo RepositoryFacade, fromTime time.Time) ([]*CodeActivityAuthor, error) {
+	stats := &CodeActivityStats{}
+	gitCmd := gitcmd.NewCommand("log", "--no-merges", "--pretty=format:---%n%H%n%aN%n%aE%n", "--branches=*").
+		AddOptionFormat("--since=%s", fromTime.Format(time.RFC3339))
+	if err := parseCodeActivityLog(ctx, repo, gitCmd, stats); err != nil {
+		return nil, err
+	}
+	return stats.Authors, nil
+}
+
+func parseCodeActivityLog(ctx context.Context, repo RepositoryFacade, gitCmd *gitcmd.Command, stats *CodeActivityStats) error {
 	stdoutReader, stdoutReaderClose := gitCmd.MakeStdoutPipe()
 	defer stdoutReaderClose()
-	err = gitCmd.
+	err := gitCmd.
 		WithRepo(repo).
 		WithPipelineFunc(func(ctx gitcmd.Context) error {
 			scanner := bufio.NewScanner(stdoutReader)
@@ -116,8 +134,8 @@ func (repo *Repository) GetCodeActivityStats(ctx context.Context, fromTime time.
 					}
 				}
 			}
-			if err = scanner.Err(); err != nil {
-				return fmt.Errorf("GetCodeActivityStats scan: %w", err)
+			if err := scanner.Err(); err != nil {
+				return fmt.Errorf("parseCodeActivityLog scan: %w", err)
 			}
 			a := make([]*CodeActivityAuthor, 0, len(authors))
 			for _, v := range authors {
@@ -134,8 +152,7 @@ func (repo *Repository) GetCodeActivityStats(ctx context.Context, fromTime time.
 		}).
 		RunWithStderr(ctx)
 	if err != nil {
-		return nil, fmt.Errorf("GetCodeActivityStats: %w", err)
+		return fmt.Errorf("parseCodeActivityLog: %w", err)
 	}
-
-	return stats, nil
+	return nil
 }

modules/git/repo_stats_test.go

@@ -33,4 +33,8 @@ func TestRepository_GetCodeActivityStats(t *testing.T) {
 	assert.Equal(t, "tris.git@shoddynet.org", code.Authors[1].Email)
 	assert.EqualValues(t, 3, code.Authors[1].Commits)
 	assert.EqualValues(t, 5, code.Authors[0].Commits)
+
+	authors, err := GetCodeActivityAuthors(t.Context(), bareRepo1, timeFrom)
+	assert.NoError(t, err)
+	assert.Equal(t, code.Authors, authors)
 }