Filter and account final glob candidates

internal/engine/glob_selection_test.go

@@ -0,0 +1,203 @@
+package engine
+
+import (
+	"io/fs"
+	"os"
+	"path/filepath"
+	"testing"
+	"testing/fstest"
+)
+
+func TestSelectGlobFiltersOnlyWildcardDiscoveredEntries(t *testing.T) {
+	t.Parallel()
+
+	root := t.TempDir()
+	for name, content := range map[string]string{
+		".hidden/secret.yaml": "value: hidden\n",
+		"visible/item.yaml":   "value: visible\n",
+		"visible/source.txt":  "value: source\n",
+	} {
+		filename := filepath.Join(root, filepath.FromSlash(name))
+		if err := os.MkdirAll(filepath.Dir(filename), 0o700); err != nil {
+			t.Fatal(err)
+		}
+		if err := os.WriteFile(filename, []byte(content), 0o600); err != nil {
+			t.Fatal(err)
+		}
+	}
+	for alias, target := range map[string]string{
+		"visible-alias":  "visible",
+		"selected.yaml":  "visible/source.txt",
+		"unselected.txt": "visible/item.yaml",
+		".secret.yaml":   "visible/source.txt",
+	} {
+		if err := os.Symlink(target, filepath.Join(root, filepath.FromSlash(alias))); err != nil {
+			t.Skipf("symlinks are unavailable: %v", err)
+		}
+	}
+	access, err := openHostAccess(root)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer access.close()
+
+	selection := selectGlobForTest(t, access, root, `    - glob: '*/*.yaml'
+      follow: false
+      hidden: false
+`)
+	if got, want := selectedPathsRelativeTo(selection, root), []string{"visible/item.yaml"}; !equalStrings(got, want) {
+		t.Fatalf("filtered glob candidates = %#v, want %#v", got, want)
+	}
+
+	selection = selectGlobForTest(t, access, root, `    - glob: '.hidden/*.yaml'
+      follow: false
+      hidden: false
+`)
+	if got, want := selectedPathsRelativeTo(selection, root), []string{".hidden/secret.yaml"}; !equalStrings(got, want) {
+		t.Fatalf("literal hidden component candidates = %#v, want %#v", got, want)
+	}
+
+	selection = selectGlobForTest(t, access, root, `    - glob: 'visible-alias/*.yaml'
+      follow: false
+      hidden: false
+`)
+	if got, want := selectedPathsRelativeTo(selection, root), []string{"visible-alias/item.yaml"}; !equalStrings(got, want) {
+		t.Fatalf("literal symlink component candidates = %#v, want %#v", got, want)
+	}
+
+	selection = selectGlobForTest(t, access, root, `    - glob: '*.yaml'
+      follow: false
+      hidden: true
+`)
+	if len(selection.candidates) != 0 {
+		t.Fatalf("follow:false selected a discovered file symlink: %#v", selection.candidates)
+	}
+
+	selection = selectGlobForTest(t, access, root, `    - glob: '*.yaml'
+      follow: true
+      hidden: false
+`)
+	if got, want := selectedPathsRelativeTo(selection, root), []string{"selected.yaml"}; !equalStrings(got, want) {
+		t.Fatalf("filename aliases were not filtered by candidate spelling: %#v, want %#v", got, want)
+	}
+}
+
+func TestSelectGlobAppliesExclusionsBeforeSourceAdmissionWithoutOpeningFiles(t *testing.T) {
+	t.Parallel()
+
+	filesystem := &observedGlobFilesystem{files: fstest.MapFS{
+		"parts":                 &fstest.MapFile{Mode: fs.ModeDir | 0o700},
+		"parts/keep.yaml":       &fstest.MapFile{Data: []byte("not parsed: [\n")},
+		"parts/skip-one.yaml":   &fstest.MapFile{Data: []byte("not parsed: [\n")},
+		"parts/skip-two.yaml":   &fstest.MapFile{Data: []byte("not parsed: [\n")},
+		"parts/skip-three.yaml": &fstest.MapFile{Data: []byte("not parsed: [\n")},
+	}}
+	access := newNativeAccess(filesystem, ".", &namespaceToken{})
+	specs, diagnostics := parseImportSpecsYAML(t, `    - glob: 'parts/*.yaml'
+      exclude: 'skip-*.yaml'
+`)
+	if len(diagnostics) != 0 {
+		t.Fatal(diagnostics)
+	}
+	policy, diagnostics := resolveImportPolicy(specs[0], importDefaults{})
+	if len(diagnostics) != 0 {
+		t.Fatal(diagnostics)
+	}
+	limits := defaultImportLimits()
+	limits.sources = 2
+	accounting := &importAccounting{sources: 1}
+	selection, diagnostics := selectGlobImportCandidates(access, ".", specs[0], policy, limits, accounting)
+	if len(diagnostics) != 0 || !equalStrings(selectedPaths(selection), []string{"parts/keep.yaml"}) {
+		t.Fatalf("post-exclusion selection, diagnostics = %#v, %#v", selection, diagnostics)
+	}
+	if len(filesystem.opened) != 1 || filesystem.opened[0] != "parts" {
+		t.Fatalf("glob opened paths = %#v, want only the parts directory", filesystem.opened)
+	}
+	if accounting.sources+len(selection.candidates) > limits.sources {
+		t.Fatalf("final candidate set exceeds the private source limit: accounting=%#v selection=%#v", accounting, selection.candidates)
+	}
+}
+
+func TestSelectGlobRejectsEscapingWildcardLinks(t *testing.T) {
+	t.Parallel()
+
+	root := t.TempDir()
+	outside := t.TempDir()
+	if err := os.WriteFile(filepath.Join(outside, "file.yaml"), []byte("value: outside\n"), 0o600); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.Symlink(outside, filepath.Join(root, "escape")); err != nil {
+		t.Skipf("symlinks are unavailable: %v", err)
+	}
+	access, err := openHostAccess(root)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer access.close()
+
+	_, diagnostics := selectGlobForTestWithDiagnostics(t, access, root, `    - glob: '*/file.yaml'
+      follow: true
+`)
+	if len(diagnostics) != 1 || diagnostics[0].Code != CodeImportOutsideRoot {
+		t.Fatalf("escaping glob diagnostics = %#v, want import.outside_root", diagnostics)
+	}
+	selection := selectGlobForTest(t, access, root, `    - glob: '*/file.yaml'
+      follow: false
+`)
+	if len(selection.candidates) != 0 {
+		t.Fatalf("follow:false selected an escaping alias: %#v", selection.candidates)
+	}
+}
+
+func TestSelectGlobPreservesMatchedDirectoryValidityBeforeExclusion(t *testing.T) {
+	t.Parallel()
+
+	filesystem := fstest.MapFS{
+		"parts":                &fstest.MapFile{Mode: fs.ModeDir | 0o700},
+		"parts/matched.yaml":   &fstest.MapFile{Mode: fs.ModeDir | 0o700},
+		"parts/matched.yaml/x": &fstest.MapFile{Data: []byte("value: x\n")},
+	}
+	access := newNativeAccess(filesystem, ".", &namespaceToken{})
+	_, diagnostics := selectGlobForTestWithDiagnostics(t, access, ".", `    - glob: 'parts/*.yaml'
+      exclude: '*.yaml'
+`)
+	if len(diagnostics) != 1 || diagnostics[0].Code != CodeSourceNotRegular {
+		t.Fatalf("excluded directory diagnostics = %#v, want source.not_regular", diagnostics)
+	}
+}
+
+func selectGlobForTest(t *testing.T, access *filesystemAccess, base, declaration string) importSelection {
+	t.Helper()
+	selection, diagnostics := selectGlobForTestWithDiagnostics(t, access, base, declaration)
+	if len(diagnostics) != 0 {
+		t.Fatalf("glob selection diagnostics = %#v", diagnostics)
+	}
+	return selection
+}
+
+func selectGlobForTestWithDiagnostics(t *testing.T, access *filesystemAccess, base, declaration string) (importSelection, []Diagnostic) {
+	t.Helper()
+	specs, diagnostics := parseImportSpecsYAML(t, declaration)
+	if len(diagnostics) != 0 {
+		t.Fatalf("parse diagnostics = %#v", diagnostics)
+	}
+	policy, diagnostics := resolveImportPolicy(specs[0], importDefaults{})
+	if len(diagnostics) != 0 {
+		t.Fatalf("policy diagnostics = %#v", diagnostics)
+	}
+	return selectGlobImportCandidates(access, base, specs[0], policy, defaultImportLimits(), &importAccounting{})
+}
+
+type observedGlobFilesystem struct {
+	files  fstest.MapFS
+	opened []string
+}
+
+func (filesystem *observedGlobFilesystem) Open(name string) (fs.File, error) {
+	filesystem.opened = append(filesystem.opened, name)
+	return filesystem.files.Open(name)
+}
+
+func (filesystem *observedGlobFilesystem) Stat(name string) (fs.FileInfo, error) {
+	return fs.Stat(filesystem.files, name)
+}