Leave a narrowed build only its named paths, manifest and locks, in one function for both build modules (issue 1758, ADR 0499 §4)

go/buildstep/narrow.go

@@ -0,0 +1,69 @@
+package buildstep
+
+import (
+	"fmt"
+	"os"
+	"path"
+	"path/filepath"
+	"strings"
+)
+
+// narrowKept are the files of a module's directory a narrowed build keeps beside its named paths: its manifest and
+// its lock files.
+var narrowKept = []string{"module.json", "package.json", "package-lock.json", "go.mod", "go.sum"}
+
+// Narrow leaves in the fetched tree only what a build of an artifact that names its paths may see (novox/hq ADR 0499
+// clause 4): r.Dir's named paths, its manifest and lock files, and .git; everything else is removed, so a file outside
+// them cannot change what is built without changing its key. A named path absent at the tree is an error naming it,
+// and nothing is removed then. With no paths in r it changes nothing. Both build modules' fetch steps call it.
+func Narrow(tree string, r FetchRequest) error {
+	if len(r.Paths) == 0 {
+		return nil
+	}
+	keep := map[string]bool{".git": true}
+	for _, p := range r.Paths {
+		rel := path.Join(r.Dir, p)
+		if strings.HasPrefix(rel, "../") || rel == ".." || path.IsAbs(p) {
+			return fmt.Errorf("the named path %q leaves the module's directory", p)
+		}
+		if _, err := os.Lstat(filepath.Join(tree, filepath.FromSlash(rel))); err != nil {
+			return fmt.Errorf("the named path %q is not in the fetched tree: %w", p, err)
+		}
+		keep[rel] = true
+	}
+	for _, f := range narrowKept {
+		keep[path.Join(r.Dir, f)] = true
+	}
+	return prune(tree, "", keep)
+}
+
+// prune removes from dir (rel within the tree) every entry that is neither kept nor on the way to one.
+func prune(root, rel string, keep map[string]bool) error {
+	entries, err := os.ReadDir(filepath.Join(root, filepath.FromSlash(rel)))
+	if err != nil {
+		return err
+	}
+	for _, e := range entries {
+		p := path.Join(rel, e.Name())
+		if keep[p] {
+			continue
+		}
+		ancestor := false
+		for k := range keep {
+			if strings.HasPrefix(k, p+"/") {
+				ancestor = true
+				break
+			}
+		}
+		if ancestor && e.IsDir() {
+			if err := prune(root, p, keep); err != nil {
+				return err
+			}
+			continue
+		}
+		if err := os.RemoveAll(filepath.Join(root, filepath.FromSlash(p))); err != nil {
+			return err
+		}
+	}
+	return nil
+}

go/buildstep/narrow_test.go

@@ -0,0 +1,67 @@
+package buildstep
+
+import (
+	"os"
+	"path/filepath"
+	"sort"
+	"strings"
+	"testing"
+)
+
+func filesUnder(t *testing.T, root string) []string {
+	t.Helper()
+	var out []string
+	_ = filepath.Walk(root, func(p string, info os.FileInfo, err error) error {
+		if err == nil && !info.IsDir() {
+			rel, _ := filepath.Rel(root, p)
+			out = append(out, filepath.ToSlash(rel))
+		}
+		return nil
+	})
+	sort.Strings(out)
+	return out
+}
+
+// A build of an artifact that names its paths sees only those paths, its manifest and its lock files (novox/hq ADR 0499
+// clause 4): Narrow removes everything else of the fetched tree but .git; with no paths it changes nothing.
+func TestNarrowLeavesOnlyTheNamedPathsTheManifestAndTheLocks(t *testing.T) {
+	tree := t.TempDir()
+	for _, f := range []string{".git/HEAD", "module.json", "package.json", "package-lock.json", "go.sum", "src/index.ts",
+		"src/deep/a.ts", "test/a.test.ts", "tsconfig.json", "go/delivery/x.go", "README.md", "other/module.json"} {
+		writeTreeFile(t, tree, f, "x")
+	}
+	if err := Narrow(tree, FetchRequest{}); err != nil {
+		t.Fatal(err)
+	}
+	if got := len(filesUnder(t, tree)); got != 12 {
+		t.Fatalf("a request without paths changed the tree: %d files", got)
+	}
+	if err := Narrow(tree, FetchRequest{Dir: ".", Paths: []string{"src", "test", "tsconfig.json"}}); err != nil {
+		t.Fatal(err)
+	}
+	want := ".git/HEAD,go.sum,module.json,package-lock.json,package.json,src/deep/a.ts,src/index.ts,test/a.test.ts,tsconfig.json"
+	if got := strings.Join(filesUnder(t, tree), ","); got != want {
+		t.Fatalf("the narrowed tree holds %s\nwant %s", got, want)
+	}
+}
+
+func TestNarrowKeepsAModuleDirectorysPathsAndFailsOnAnAbsentOne(t *testing.T) {
+	tree := t.TempDir()
+	for _, f := range []string{".git/config", "modules/a/module.json", "modules/a/go.mod", "modules/a/src/x.ts",
+		"modules/a/notes.md", "modules/b/module.json", "root.txt"} {
+		writeTreeFile(t, tree, f, "x")
+	}
+	err := Narrow(tree, FetchRequest{Dir: "modules/a", Paths: []string{"src", "gone"}})
+	if err == nil || !strings.Contains(err.Error(), `"gone"`) {
+		t.Fatalf("an absent named path gave %v", err)
+	}
+	if len(filesUnder(t, tree)) != 7 {
+		t.Fatal("a refused narrowing removed files")
+	}
+	if err := Narrow(tree, FetchRequest{Dir: "modules/a", Paths: []string{"src"}}); err != nil {
+		t.Fatal(err)
+	}
+	if got := strings.Join(filesUnder(t, tree), ","); got != ".git/config,modules/a/go.mod,modules/a/module.json,modules/a/src/x.ts" {
+		t.Fatalf("the narrowed tree holds %s", got)
+	}
+}