config: remove fields not documented in config/README.md

Drops RepositoryConfig.FullHashRepos/ExcludedFiles/ExcludeExternalTargets/
StreamBazelLogs and ServiceConfig's settable WorkerRootPath override — none
are documented in config/README.md, and WorkerRootPath is now always
computed by Parse as workspaces_root_path/.workers rather than optionally
overridden. graphrunner/native.go's Bzlmod query construction is adjusted
so external-target querying no longer depends on the removed
ExcludeExternalTargets field.

Also fixes make test-integration on macOS: excludes @bazel_tools's
platform-specific files (e.g. a Windows-only test wrapper) from disk-based
hashing via RequestOptions.ExtraExcludeFilesRegex, since a prior config-side
exclusion (excluded_files) was just removed as undocumented.

Author: yushan8

config/config.go

@@ -72,12 +72,10 @@ func Parse(configFilePath string) (*Config, error) {
 	if config.Service.WorkspacesRootPath == "" {
 		return nil, fmt.Errorf("service.workspaces_root_path must be set")
 	}
-	if config.Service.WorkerRootPath == "" {
-		config.Service.WorkerRootPath = filepath.Join(config.Service.WorkspacesRootPath, ".workers")
-	}
 	if config.Service.MaxWorkerPoolSize <= 0 {
 		return nil, fmt.Errorf("service.max_worker_pool_size must be > 0, got %d", config.Service.MaxWorkerPoolSize)
 	}
+	config.Service.WorkerRootPath = filepath.Join(config.Service.WorkspacesRootPath, ".workers")
 	if config.Service.Streaming.MaxNumTargets <= 0 {
 		config.Service.Streaming.MaxNumTargets = _defaultMaxNumTargets
 	}

config/config_test.go

@@ -80,7 +80,7 @@ func TestParse_ServiceDefaults(t *testing.T) {
 		wantMaxNumMetadataEntries int
 	}{
 		{
-			name: "worker_root_path and streaming default when unset",
+			name: "streaming defaults when unset",
 			give: _baseServiceYAML + `
 repository:
   - remote: "r1"
@@ -90,21 +90,6 @@ repository:
 			wantMaxNumChangedTargets:  _defaultMaxNumChangedTargets,
 			wantMaxNumMetadataEntries: _defaultMaxNumMetadataEntries,
 		},
-		{
-			name: "worker_root_path explicit value preserved",
-			give: `
-service:
-  workspaces_root_path: "/tmp/x"
-  worker_root_path: "/tmp/custom-workers"
-  max_worker_pool_size: 1
-repository:
-  - remote: "r1"
-`,
-			wantWorkerRootPath:        "/tmp/custom-workers",
-			wantMaxNumTargets:         _defaultMaxNumTargets,
-			wantMaxNumChangedTargets:  _defaultMaxNumChangedTargets,
-			wantMaxNumMetadataEntries: _defaultMaxNumMetadataEntries,
-		},
 		{
 			name: "streaming explicit values preserved",
 			give: _baseServiceYAML + `

config/repository_config.go

@@ -21,27 +21,18 @@ type RepositoryConfig struct {
 	// unique across all entries and match exactly what clients send in
 	// BuildDescription.remote.
 	Remote string `yaml:"remote"`
-	// TODO: FullHashRepos, ExcludedFiles, and ExcludeExternalTargets are not
-	// documented in config/README.md. Delete them if they turn out to be unneeded,
-	// otherwise document them there.
-	FullHashRepos          []string `yaml:"full_hash_repos"`
-	ExcludedFiles          []string `yaml:"excluded_files"`
-	ExcludeExternalTargets bool     `yaml:"exclude_external_targets"`
 	// BzlmodEnabled indicates whether this repository uses Bzlmod for external
 	// dependency management. Defaults to true if unset. Set to false only for
 	// repositories still using WORKSPACE.
 	BzlmodEnabled *bool `yaml:"bzlmod_enabled"`
 	// BazelCommandPath overrides the Bazel binary path. When empty, Tango
 	// automatically downloads and caches Bazelisk from GitHub.
 	BazelCommandPath string `yaml:"bazel_command_path"`
-	// QueryTimeoutSeconds is the Bazel query timeout in seconds. Defaults to 600.
-	QueryTimeoutSeconds int64 `yaml:"query_timeout_seconds"`
 	// BazelExtraArgs are extra arguments passed to `bazel query` invocations,
 	// inserted between the `query` subcommand and the query expression.
 	BazelExtraArgs []string `yaml:"bazel_extra_args"`
-	// TODO: StreamBazelLogs is not documented in config/README.md. Delete it if
-	// it turns out to be unneeded, otherwise document it there.
-	StreamBazelLogs bool `yaml:"stream_bazel_logs"`
+	// QueryTimeoutSeconds is the Bazel query timeout in seconds. Defaults to 600.
+	QueryTimeoutSeconds int64 `yaml:"query_timeout_seconds"`
 }
 
 // RepositoryConfigProvider looks up per-repository configuration by remote.

config/service_config.go

@@ -24,11 +24,12 @@ type ServiceConfig struct {
 	// and worker checkouts. Required. Layout: <workspaces_root_path>/<repo>/ for
 	// origin clones and <workspaces_root_path>/.workers/<repo>/worker-{1..N}/ for
 	// worker checkouts.
-	WorkspacesRootPath string `yaml:"workspaces_root_path"`
-	// TODO: WorkerRootPath is not documented in config/README.md. Delete it if
-	// it turns out to be unneeded, otherwise document it there.
-	WorkerRootPath string      `yaml:"worker_root_path"` // root directory for worker workspace checkouts; defaults to workspaces_root_path/.workers
-	Streaming      ChunkConfig `yaml:"streaming"`        // streaming chunk sizes; zero values fall back to package defaults
+	WorkspacesRootPath string      `yaml:"workspaces_root_path"`
+	Streaming          ChunkConfig `yaml:"streaming"` // streaming chunk sizes; zero values fall back to package defaults
+
+	// WorkerRootPath is the root directory for worker workspace checkouts,
+	// set by Parse to <workspaces_root_path>/.workers. Not settable via YAML.
+	WorkerRootPath string `yaml:"-"`
 }
 
 // ChunkConfig controls the number of entries per gRPC stream message.

example/tango-config.yaml

@@ -9,11 +9,6 @@ storage:
 # Repository configuration
 repository:
   - remote: "https://github.com/uber/tango.git"
-    full_hash_repos:
-      - ""
-    excluded_files:
-      - "^@@?bazel_tools/"
-    exclude_external_targets: true
     bzlmod_enabled: true
     query_timeout_seconds: 300
 
@@ -22,5 +17,3 @@ service:
   max_worker_pool_size: 5
   # root for origin clones; required
   workspaces_root_path: "/tmp/tango-repo-manager"
-  # root for worker checkouts; defaults to workspaces_root_path/.workers
-  worker_root_path: "/tmp/tango/workers"

graphrunner/native.go

@@ -58,9 +58,12 @@ func NewNativeGraphRunner(p NativeGraphRunnerParams) GraphRunner {
 }
 
 func (g *nativeGraphRunner) Compute(ctx context.Context, ws workspace.Workspace) (targethasher.Result, error) {
-	query := "//external:all-targets + deps(//...:all-targets)"
-	if g.config.ExcludeExternalTargets {
-		query = "deps(//...:all-targets)"
+	bzlmodEnabled := g.config.BzlmodEnabled == nil || *g.config.BzlmodEnabled
+	query := "deps(//...:all-targets)"
+	if !bzlmodEnabled {
+		// //external is only queryable under legacy WORKSPACE resolution;
+		// Bzlmod repos resolve external deps under @@<module> instead.
+		query = "//external:all-targets + " + query
 	}
 	additionalArgs := append(
 		[]string{"--order_output=no", "--proto:locations", "--noproto:default_values"},
@@ -74,8 +77,6 @@ func (g *nativeGraphRunner) Compute(ctx context.Context, ws workspace.Workspace)
 		// --proto:locations: we need to get external file location to make CTC more accurate
 		// --noproto: parameters exclude fields from the output that are not used for hashing anyways, making
 		// proto blob smaller and serialization/deserialization faster
-		// TODO: pass in --enable_workspace or --enable_bzlmod based on the config
-
 		AdditionalArgs: additionalArgs,
 	})
 	g.scope.Timer("bazel_query_duration").Record(time.Since(bazelStart))
@@ -92,9 +93,8 @@ func (g *nativeGraphRunner) Compute(ctx context.Context, ws workspace.Workspace)
 
 	hashConfig := targethasher.HashConfig{
 		KnownSourceHashes: knownSourceHashes,
-		FullHashRepos:     g.config.FullHashRepos,
-		ExcludedRegex:     append(g.config.ExcludedFiles, g.extraExcludedFiles...),
-		UseBzlmod:         g.config.BzlmodEnabled == nil || *g.config.BzlmodEnabled,
+		ExcludedRegex:     g.extraExcludedFiles,
+		UseBzlmod:         bzlmodEnabled,
 	}
 
 	hashStart := time.Now()

integration/integration_test.go

@@ -48,6 +48,11 @@ const (
 	configTemplateFile = "testdata/tango-config.yaml.tmpl"
 )
 
+// bazelToolsExcludeRegex excludes @bazel_tools's source files from disk-based hashing.
+// It ships platform-specific files (e.g. tools/test/tw.exe, a Windows-only test wrapper)
+// that don't exist on every OS, so stat-ing them to hash fails on machines missing them.
+var bazelToolsExcludeRegex = []string{"^@@?bazel_tools/"}
+
 func repoRemote(t *testing.T) string {
 	t.Helper()
 	remote := os.Getenv("TANGO_REPO_REMOTE")
@@ -266,6 +271,9 @@ func getChangedTargets(t *testing.T, client pb.TangoYARPCClient, remote, firstSH
 			Remote:  remote,
 			BaseSha: secondSHA,
 		},
+		RequestOptions: &pb.RequestOptions{
+			ExtraExcludeFilesRegex: bazelToolsExcludeRegex,
+		},
 	})
 	require.NoError(t, err, "failed to initiate GetChangedTargets stream")
 
@@ -361,6 +369,9 @@ func TestIntegration_GetTargetGraph(t *testing.T) {
 			Remote:  remote,
 			BaseSha: pinnedSHA,
 		},
+		RequestOptions: &pb.RequestOptions{
+			ExtraExcludeFilesRegex: bazelToolsExcludeRegex,
+		},
 	})
 	require.NoError(t, err, "failed to initiate GetTargetGraph stream")
 

integration/testdata/tango-config.yaml.tmpl

@@ -7,13 +7,9 @@ repository:
 {{- if .BazelCommand}}
     bazel_command_path: {{.BazelCommand}}
 {{- end}}
-    full_hash_repos: [""]
-    excluded_files: ["^@@?bazel_tools/"]
-    exclude_external_targets: true
     bzlmod_enabled: true
     query_timeout_seconds: 600
 
 service:
   max_worker_pool_size: 2
   workspaces_root_path: {{.ClonePath}}
-  worker_root_path: {{.WorkerPath}}

orchestrator/native_orchestrator.go

@@ -191,7 +191,6 @@ func (b *nativeOrchestrator) GetTargetGraph(ctx context.Context, req entity.GetT
 			Logger:        b.logger,
 			BazelCommand:  repoCfg.BazelCommandPath,
 			QueryTimeout:  time.Duration(repoCfg.QueryTimeoutSeconds) * time.Second,
-			StreamLogs:    repoCfg.StreamBazelLogs,
 		})
 		if err != nil {
 			return nil, fmt.Errorf("create bazel client: %w", err)

orchestrator/testdata/config.yaml

@@ -1,11 +1,5 @@
 repository:
   - remote: "git@github:uber/tango"
-    full_hash_repos:
-      - "//"
-      - "//external"
-    excluded_files:
-      - "*.gen.go"
-    exclude_external_targets: true
     bzlmod_enabled: true
     query_timeout_seconds: 15