better commit and restore tests (almost complete)

added some file path management function for prfixes and trailing slash

6 files changed, 99 insertions, 27 deletions

diff --git a/TODO.md b/TODO.md
index bb95b39..9d2c75f 100644
--- a/TODO.md
+++ b/TODO.md
@@ -24,12 +24,13 @@ priority 1
         - [ ] uncompress before loading
     - [ ] store compressed chunks into tracks of trackSize (1024o)
 - [x] add chunk cache... what was it for again ??
+- [x] better tests for `(*Repo).Commit`
 
 priority 2
 ----------
 - [x] use more the `Reader` API (which is analoguous to the `IOStream` in Java)
 - [ ] refactor matchStream as right now it is quite complex
-- [ ] better test for `Repo.matchStream`
+- [x] better test for `(*Repo).matchStream`
 - [ ] tail packing of PartialChunks (this Struct does not exist yet as it is in fact just `TempChunks` for now)
 - [ ] option to commit without deltas to save new base chunks
 
diff --git a/repo.go b/repo.go
index 585e505..ce8d890 100644
--- a/repo.go
+++ b/repo.go
@@ -38,10 +38,12 @@ import (
 	"path"
 	"path/filepath"
 	"reflect"
+	"strings"
 
 	"github.com/chmduquesne/rollinghash/rabinkarp64"
 	"github.com/n-peugnet/dna-backup/cache"
 	"github.com/n-peugnet/dna-backup/sketch"
+	"github.com/n-peugnet/dna-backup/utils"
 )
 
 type FingerprintMap map[uint64]*ChunkId
@@ -100,6 +102,7 @@ func (r *Repo) Patcher() Patcher {
 }
 
 func (r *Repo) Commit(source string) {
+	source = utils.TrimTrailingSeparator(source)
 	versions := r.loadVersions()
 	newVersion := len(versions) // TODO: add newVersion functino
 	newPath := path.Join(r.path, fmt.Sprintf(versionFmt, newVersion))
@@ -116,7 +119,7 @@ func (r *Repo) Commit(source string) {
 	r.hashChunks(oldChunks)
 	recipe := r.matchStream(reader, newVersion)
 	storeRecipe(newRecipePath, recipe)
-	storeFileList(newFilesPath, files)
+	storeFileList(newFilesPath, unprefixFiles(files, source))
 	fmt.Println(files)
 }
 
@@ -177,6 +180,20 @@ func listFiles(path string) []File {
 	return files
 }
 
+func unprefixFiles(files []File, prefix string) (ret []File) {
+	ret = make([]File, len(files))
+	preSize := len(prefix)
+	for i, f := range files {
+		if !strings.HasPrefix(f.Path, prefix) {
+			log.Println("Warning", f.Path, "is not prefixed by", prefix)
+		} else {
+			f.Path = f.Path[preSize:]
+		}
+		ret[i] = f
+	}
+	return
+}
+
 func concatFiles(files []File, stream io.WriteCloser) {
 	for _, f := range files {
 		file, err := os.Open(f.Path)
@@ -189,10 +206,6 @@ func concatFiles(files []File, stream io.WriteCloser) {
 	stream.Close()
 }
 
-func (r *Repo) chunkMinLen() int {
-	return sketch.SuperFeatureSize(r.chunkSize, r.sketchSfCount, r.sketchFCount)
-}
-
 func storeFileList(dest string, files []File) {
 	file, err := os.Create(dest)
 	if err == nil {
@@ -288,6 +301,10 @@ func (r *Repo) loadChunks(versions []string, chunks chan<- IdentifiedChunk) {
 	close(chunks)
 }
 
+func (r *Repo) chunkMinLen() int {
+	return sketch.SuperFeatureSize(r.chunkSize, r.sketchSfCount, r.sketchFCount)
+}
+
 // hashChunks calculates the hashes for a channel of chunks.
 //
 // For each chunk, both a fingerprint (hash over the full content) and a sketch
diff --git a/repo_test.go b/repo_test.go
index 5127263..cd8a88e 100644
--- a/repo_test.go
+++ b/repo_test.go
@@ -231,41 +231,70 @@ func TestBsdiff(t *testing.T) {
 func TestCommit(t *testing.T) {
 	dest := t.TempDir()
 	source := path.Join("testdata", "logs")
+	expected := path.Join("testdata", "repo_8k")
 	repo := NewRepo(dest)
 	repo.Commit(source)
-	recipe := loadRecipe(path.Join(dest, "00000", recipeName))
-	log.Println(recipe)
+	assertSameTree(t, assertCompatibleRepoFile, expected, dest, "Commit")
 }
 
 func TestRestore(t *testing.T) {
 	dest := t.TempDir()
 	source := path.Join("testdata", "repo_8k")
+	expected := path.Join("testdata", "logs")
 	repo := NewRepo(source)
 	repo.Restore(dest)
-	destFiles := listFiles(dest)
-	sourceFiles := listFiles(path.Join("testdata", "logs"))
-	sfCount := len(sourceFiles)
-	if sfCount <= 0 {
-		t.Fatalf("No source files: %d", sfCount)
+	assertSameTree(t, assertSameFile, expected, dest, "Restore")
+}
+
+func assertSameTree(t *testing.T, apply func(t *testing.T, expected string, actual string, prefix string), expected string, actual string, prefix string) {
+	actualFiles := listFiles(actual)
+	expectedFiles := listFiles(expected)
+	efCount := len(expectedFiles)
+	if efCount <= 0 {
+		t.Fatalf("No expected files: %d", efCount)
 	}
-	dfCount := len(destFiles)
-	if sfCount != dfCount {
-		t.Fatalf("Incorrect number for destination files: %d, should be %d", dfCount, sfCount)
+	afCount := len(actualFiles)
+	if efCount != afCount {
+		t.Fatalf("Incorrect number of files: %d, should be %d", afCount, efCount)
 	}
-	for i, sf := range sourceFiles {
-		sfContent, err := os.ReadFile(sf.Path)
-		if err != nil {
-			t.Fatalf("Error reading from source file '%s': %s", sf.Path, err)
+	for i, ef := range expectedFiles {
+		af := actualFiles[i]
+		efRelPath := ef.Path[len(expected):]
+		afRelPath := af.Path[len(actual):]
+		if efRelPath != afRelPath {
+			t.Fatalf("File path '%s' does not match '%s'", afRelPath, efRelPath)
 		}
-		df := destFiles[i]
-		dfContent, err := os.ReadFile(df.Path)
-		if err != nil {
-			t.Fatalf("Error reading from source file '%s': %s", df.Path, err)
-		}
-		if bytes.Compare(sfContent, dfContent) != 0 {
-			t.Errorf("File content of '%s' does not match '%s'", df.Path, sf.Path)
+		apply(t, ef.Path, af.Path, prefix)
+	}
+}
+
+func assertCompatibleRepoFile(t *testing.T, expected string, actual string, prefix string) {
+	if path.Base(expected) == filesName {
+		eFiles := loadFileList(expected)
+		aFiles := loadFileList(actual)
+		assertLen(t, len(eFiles), aFiles, prefix)
+		for i := 0; i < len(eFiles); i++ {
+			if eFiles[i] != aFiles[i] {
+				t.Fatal(prefix, "file entry do not match:", aFiles[i], ", expected:", eFiles[i])
+			}
 		}
+	} else if path.Base(expected) == recipeName {
+		// TODO: check recipies equality
+	} else {
+		assertSameFile(t, expected, actual, prefix)
+	}
+}
+
+func assertSameFile(t *testing.T, expected string, actual string, prefix string) {
+	efContent, err := os.ReadFile(expected)
+	if err != nil {
+		t.Fatalf("%s Error reading from expected file '%s': %s", prefix, expected, err)
+	}
+	afContent, err := os.ReadFile(actual)
+	if err != nil {
+		t.Fatalf("%s Error reading from expected file '%s': %s", prefix, actual, err)
 	}
+	assertSameSlice(t, efContent, afContent, prefix+" files")
 }
 
 func assertLen(t *testing.T, expected int, actual interface{}, prefix string) {
diff --git a/testdata/repo_8k/00000/files b/testdata/repo_8k/00000/files
index 20d5905..5a7b1da 100644
--- a/testdata/repo_8k/00000/files
+++ b/testdata/repo_8k/00000/files
diff --git a/utils/fileutils.go b/utils/fileutils.go
new file mode 100644
index 0000000..6c269ca
--- /dev/null
+++ b/utils/fileutils.go
@@ -0,0 +1,10 @@
+package utils
+
+import (
+	"path/filepath"
+	"strings"
+)
+
+func TrimTrailingSeparator(path string) string {
+	return strings.TrimSuffix(path, string(filepath.Separator))
+}
diff --git a/utils/fileutils_test.go b/utils/fileutils_test.go
new file mode 100644
index 0000000..176d856
--- /dev/null
+++ b/utils/fileutils_test.go
@@ -0,0 +1,15 @@
+package utils
+
+import (
+	"path/filepath"
+	"testing"
+)
+
+func TestTrimTrailingSeparator(t *testing.T) {
+	if TrimTrailingSeparator("test"+string(filepath.Separator)) != "test" {
+		t.Error("Seprator should have been trimmed")
+	}
+	if TrimTrailingSeparator("test") != "test" {
+		t.Error("Path should not have changed")
+	}
+}