Store MD5 manifest as hidden file at root of extraction directory

README.md

@@ -11,6 +11,7 @@ A simple utility to create tarballs with a specific directory structure.
 - Cross-platform compatibility (FreeBSD, macOS, Linux)
 - Produces tarballs compatible with standard tar tools
 - Verified data integrity with MD5 hash comparison
+- Automatically generates and verifies MD5 hashes of all files during extraction
 
 ## Building
 
@@ -51,27 +52,47 @@ The tests will:
 The usage is the same for all binaries:
 
 ```bash
+# Create a tarball
 ./bin/tarballer-<platform> -source /path/to/directory -output myarchive.tar.gz -prefix myprefix
+
+# Extract a tarball with integrity verification
+./bin/tarballer-<platform> -extract -output myarchive.tar.gz -extractdir /path/to/extract
 ```
 
-### Options
+### Create Mode Options
 
 - `-source`: The directory you want to compress (required)
 - `-output`: The name of the output tarball (defaults to "output.tar.gz")
 - `-prefix`: The directory name that will contain all files in the tarball (defaults to "myapp")
 
-### Example
+### Extract Mode Options
+
+- `-extract`: Enables extraction mode
+- `-output`: The tarball to extract (required)
+- `-extractdir`: Directory to extract to (defaults to current directory)
+- `-verify`: Only verify hash integrity without extraction
+- `-keepmanifest`: Keep the MD5 manifest file after extraction (defaults to removing it)
+
+### Examples
 
 ```bash
-# On macOS:
+# Create a tarball (on macOS):
 ./bin/tarballer-darwin -source ./myproject -output release.tar.gz -prefix app
 
-# On FreeBSD:
-./bin/tarballer-freebsd -source ./myproject -output release.tar.gz -prefix app
+# Extract and verify (on Linux):
+./bin/tarballer-linux -extract -output release.tar.gz -extractdir /path/to/extract
+
+# Only verify hash integrity without extraction:
+./bin/tarballer-linux -extract -verify -output release.tar.gz
 ```
 
-When extracted, all files will be under the `app/` directory in the tarball and can be extracted with standard tools like:
+## MD5 Hash Verification
 
-```bash
-tar -xzf release.tar.gz -C /path/to/extract
-``` 
+Tarballer includes built-in file integrity protection:
+
+1. When creating a tarball, MD5 hashes are calculated for all files and stored in a hidden manifest file (`.md5-manifest.txt`) at the root of the extraction directory
+2. During extraction, hashes are verified to ensure files haven't been corrupted or tampered with
+3. The manifest file is automatically removed after extraction unless `-keepmanifest` is specified
+4. If any file fails verification, the extraction is aborted with an error
+
+This provides an extra layer of security and data integrity validation compared to standard tar tools. 

main.go

@@ -2,34 +2,84 @@ package main
 
 import (
 	"archive/tar"
+	"bufio"
 	"compress/gzip"
+	"crypto/md5"
+	"encoding/hex"
 	"flag"
 	"fmt"
 	"io"
 	"os"
 	"path/filepath"
+	"strings"
 )
 
+const manifestFilename = ".md5-manifest.txt"
+
 func main() {
 	// Define command line flags
 	sourceDir := flag.String("source", "", "Source directory to compress")
 	outputFile := flag.String("output", "output.tar.gz", "Output tarball filename")
 	prefixDir := flag.String("prefix", "myapp", "Directory prefix in tarball")
+	extractMode := flag.Bool("extract", false, "Extract mode (instead of create)")
+	extractDir := flag.String("extractdir", "", "Directory to extract to (default: current directory)")
+	verifyOnly := flag.Bool("verify", false, "Only verify hash integrity without extraction")
+	keepManifest := flag.Bool("keepmanifest", false, "Keep the MD5 manifest file after extraction")
 	flag.Parse()
 
-	if *sourceDir == "" {
-		fmt.Println("Please specify a source directory using -source")
-		flag.Usage()
-		os.Exit(1)
-	}
+	if *extractMode {
+		if *outputFile == "output.tar.gz" && len(flag.Args()) > 0 {
+			*outputFile = flag.Args()[0]
+		}
 
-	err := createTarball(*sourceDir, *outputFile, *prefixDir)
+		if *outputFile == "" {
+			fmt.Println("Please specify a tarball to extract using -output or as a positional argument")
+			flag.Usage()
+			os.Exit(1)
+		}
+
+		// If extract directory is not specified, use current directory
+		extractTo := *extractDir
+		if extractTo == "" {
+			extractTo = "."
+		}
+
+		err := extractTarball(*outputFile, extractTo, *verifyOnly, *keepManifest)
+		if err != nil {
+			fmt.Printf("Error extracting tarball: %v\n", err)
+			os.Exit(1)
+		}
+	} else {
+		if *sourceDir == "" {
+			fmt.Println("Please specify a source directory using -source")
+			flag.Usage()
+			os.Exit(1)
+		}
+
+		err := createTarball(*sourceDir, *outputFile, *prefixDir)
+		if err != nil {
+			fmt.Printf("Error creating tarball: %v\n", err)
+			os.Exit(1)
+		}
+
+		fmt.Printf("Successfully created %s with prefix %s\n", *outputFile, *prefixDir)
+	}
+}
+
+// calcFileMD5 calculates the MD5 hash of a file
+func calcFileMD5(filePath string) (string, error) {
+	file, err := os.Open(filePath)
 	if err != nil {
-		fmt.Printf("Error creating tarball: %v\n", err)
-		os.Exit(1)
+		return "", err
+	}
+	defer file.Close()
+
+	hash := md5.New()
+	if _, err := io.Copy(hash, file); err != nil {
+		return "", err
 	}
 
-	fmt.Printf("Successfully created %s with prefix %s\n", *outputFile, *prefixDir)
+	return hex.EncodeToString(hash.Sum(nil)), nil
 }
 
 func createTarball(sourceDir, outputFile, prefix string) error {
@@ -48,6 +98,9 @@ func createTarball(sourceDir, outputFile, prefix string) error {
 	tw := tar.NewWriter(gw)
 	defer tw.Close()
 
+	// Create a map to store MD5 hashes
+	fileHashes := make(map[string]string)
+
 	// Resolve absolute source path to handle relative symlinks correctly
 	sourceDir, err = filepath.Abs(sourceDir)
 	if err != nil {
@@ -66,6 +119,11 @@ func createTarball(sourceDir, outputFile, prefix string) error {
 			return err
 		}
 
+		// Skip the manifest file if it exists (from a previous run)
+		if relPath == manifestFilename {
+			return nil
+		}
+
 		// Create tar header using original file info
 		header, err := tar.FileInfoHeader(info, "")
 		if err != nil {
@@ -88,6 +146,15 @@ func createTarball(sourceDir, outputFile, prefix string) error {
 
 			// Make sure the link type is set correctly
 			header.Typeflag = tar.TypeSymlink
+
+			// For symlinks, we don't calculate MD5 hashes
+		} else if !info.IsDir() {
+			// Calculate MD5 hash for regular files
+			hash, err := calcFileMD5(filePath)
+			if err != nil {
+				return err
+			}
+			fileHashes[filepath.Join(prefix, relPath)] = hash
 		}
 
 		// Write header
@@ -112,5 +179,391 @@ func createTarball(sourceDir, outputFile, prefix string) error {
 		return nil
 	})
 
-	return err
+	if err != nil {
+		return err
+	}
+
+	// Create and add the manifest file
+	var manifestContent strings.Builder
+	for path, hash := range fileHashes {
+		manifestContent.WriteString(fmt.Sprintf("%s  %s\n", hash, path))
+	}
+
+	// Create a tar header for the manifest
+	manifestHeader := &tar.Header{
+		Name:     manifestFilename,
+		Mode:     0644,
+		Size:     int64(manifestContent.Len()),
+		Typeflag: tar.TypeReg,
+	}
+
+	// Write the manifest header
+	if err := tw.WriteHeader(manifestHeader); err != nil {
+		return err
+	}
+
+	// Write the manifest content
+	if _, err := tw.Write([]byte(manifestContent.String())); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+func extractTarball(tarballPath, extractDir string, verifyOnly, keepManifest bool) error {
+	// Open the tarball
+	file, err := os.Open(tarballPath)
+	if err != nil {
+		return err
+	}
+	defer file.Close()
+
+	// Create gzip reader
+	gr, err := gzip.NewReader(file)
+	if err != nil {
+		return err
+	}
+	defer gr.Close()
+
+	// Create tar reader
+	tr := tar.NewReader(gr)
+
+	// Create a map to store expected MD5 hashes from the manifest
+	expectedHashes := make(map[string]string)
+	// First pass: Find and parse the manifest file
+	tempDir := ""
+	if !verifyOnly {
+		// Create a temporary directory for extraction
+		tempDir, err = os.MkdirTemp(extractDir, "tarballer-extract-")
+		if err != nil {
+			return err
+		}
+	}
+
+	// Extract files to get the manifest
+	for {
+		header, err := tr.Next()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			if tempDir != "" {
+				os.RemoveAll(tempDir)
+			}
+			return err
+		}
+
+		// Skip directories in first pass, we only want to find the manifest
+		if header.Typeflag == tar.TypeDir {
+			continue
+		}
+
+		// Check if this is the manifest file
+		if filepath.Base(header.Name) == manifestFilename && filepath.Dir(header.Name) == "." {
+			// Read the manifest content
+			var content strings.Builder
+			if _, err := io.Copy(&content, tr); err != nil {
+				if tempDir != "" {
+					os.RemoveAll(tempDir)
+				}
+				return err
+			}
+
+			// Parse the manifest to get expected hashes
+			scanner := bufio.NewScanner(strings.NewReader(content.String()))
+			for scanner.Scan() {
+				line := scanner.Text()
+				parts := strings.SplitN(line, "  ", 2)
+				if len(parts) == 2 {
+					hash := parts[0]
+					path := parts[1]
+					expectedHashes[path] = hash
+				}
+			}
+
+			if err := scanner.Err(); err != nil {
+				if tempDir != "" {
+					os.RemoveAll(tempDir)
+				}
+				return fmt.Errorf("error parsing manifest: %v", err)
+			}
+
+			continue
+		}
+
+		// If we're only verifying, skip extraction
+		if verifyOnly {
+			continue
+		}
+
+		// Extract to temp dir to verify hashes
+		target := filepath.Join(tempDir, header.Name)
+
+		// Create directory if needed
+		if header.Typeflag == tar.TypeDir {
+			if err := os.MkdirAll(target, 0755); err != nil {
+				os.RemoveAll(tempDir)
+				return err
+			}
+			continue
+		}
+
+		// Create parent directory if it doesn't exist
+		if err := os.MkdirAll(filepath.Dir(target), 0755); err != nil {
+			os.RemoveAll(tempDir)
+			return err
+		}
+
+		// Handle symlinks
+		if header.Typeflag == tar.TypeSymlink {
+			if err := os.Symlink(header.Linkname, target); err != nil {
+				os.RemoveAll(tempDir)
+				return err
+			}
+			continue
+		}
+
+		// Create regular file
+		f, err := os.OpenFile(target, os.O_CREATE|os.O_WRONLY, os.FileMode(header.Mode))
+		if err != nil {
+			os.RemoveAll(tempDir)
+			return err
+		}
+
+		// Copy file content
+		if _, err := io.Copy(f, tr); err != nil {
+			f.Close()
+			os.RemoveAll(tempDir)
+			return err
+		}
+		f.Close()
+	}
+
+	// If no manifest was found
+	if len(expectedHashes) == 0 {
+		if tempDir != "" {
+			os.RemoveAll(tempDir)
+		}
+		return fmt.Errorf("no MD5 manifest found in tarball")
+	}
+
+	// If we're only verifying, we need to reopen the tarball
+	if verifyOnly {
+		file.Seek(0, 0)
+		gr, err = gzip.NewReader(file)
+		if err != nil {
+			return err
+		}
+		defer gr.Close()
+		tr = tar.NewReader(gr)
+	}
+
+	// Second pass: Verify hashes
+	verificationFailed := false
+	fileVerified := make(map[string]bool)
+
+	if verifyOnly {
+		// Extract to temp dir for verification
+		tempDir, err = os.MkdirTemp(extractDir, "tarballer-verify-")
+		if err != nil {
+			return err
+		}
+		defer os.RemoveAll(tempDir)
+	}
+
+	for {
+		header, err := tr.Next()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			if tempDir != "" && !verifyOnly {
+				os.RemoveAll(tempDir)
+			}
+			return err
+		}
+
+		// Skip directories and the manifest file for verification
+		if header.Typeflag == tar.TypeDir || (filepath.Base(header.Name) == manifestFilename && filepath.Dir(header.Name) == ".") {
+			continue
+		}
+
+		// Skip symlinks for hash verification
+		if header.Typeflag == tar.TypeSymlink {
+			continue
+		}
+
+		// Check if this file has an expected hash
+		expectedHash, exists := expectedHashes[header.Name]
+		if !exists {
+			fmt.Printf("Warning: File %s not found in manifest\n", header.Name)
+			continue
+		}
+
+		// If verifyOnly, we need to extract the file to verify its hash
+		if verifyOnly {
+			target := filepath.Join(tempDir, header.Name)
+
+			// Create parent directory if it doesn't exist
+			if err := os.MkdirAll(filepath.Dir(target), 0755); err != nil {
+				return err
+			}
+
+			// Create file
+			f, err := os.OpenFile(target, os.O_CREATE|os.O_WRONLY, os.FileMode(header.Mode))
+			if err != nil {
+				return err
+			}
+
+			// Copy content
+			if _, err := io.Copy(f, tr); err != nil {
+				f.Close()
+				return err
+			}
+			f.Close()
+
+			// Calculate hash
+			actualHash, err := calcFileMD5(target)
+			if err != nil {
+				return err
+			}
+
+			// Verify hash
+			if actualHash != expectedHash {
+				fmt.Printf("Hash mismatch for %s: expected %s, got %s\n", header.Name, expectedHash, actualHash)
+				verificationFailed = true
+			} else {
+				fileVerified[header.Name] = true
+			}
+		} else {
+			// Calculate hash from extracted file
+			target := filepath.Join(tempDir, header.Name)
+			actualHash, err := calcFileMD5(target)
+			if err != nil {
+				os.RemoveAll(tempDir)
+				return err
+			}
+
+			// Verify hash
+			if actualHash != expectedHash {
+				fmt.Printf("Hash mismatch for %s: expected %s, got %s\n", header.Name, expectedHash, actualHash)
+				verificationFailed = true
+			} else {
+				fileVerified[header.Name] = true
+			}
+		}
+	}
+
+	// Check if all files in the manifest were verified
+	for path := range expectedHashes {
+		if !fileVerified[path] {
+			fmt.Printf("Warning: File %s in manifest was not found in tarball\n", path)
+		}
+	}
+
+	// If verification failed or we're only verifying, we're done
+	if verificationFailed {
+		if !verifyOnly {
+			os.RemoveAll(tempDir)
+		}
+		return fmt.Errorf("hash verification failed for one or more files")
+	}
+
+	if verifyOnly {
+		fmt.Println("All files verified successfully!")
+		return nil
+	}
+
+	// Move the extracted files to the final destination (excluding manifest if needed)
+	files, err := os.ReadDir(tempDir)
+	if err != nil {
+		os.RemoveAll(tempDir)
+		return err
+	}
+
+	// Create the final extract directory if it doesn't exist
+	if err := os.MkdirAll(extractDir, 0755); err != nil {
+		os.RemoveAll(tempDir)
+		return err
+	}
+
+	// Move each top-level extracted item
+	for _, f := range files {
+		source := filepath.Join(tempDir, f.Name())
+		dest := filepath.Join(extractDir, f.Name())
+
+		// Skip the manifest file if needed
+		if !keepManifest && (f.Name() == manifestFilename) {
+			continue
+		}
+
+		// If destination already exists, remove it
+		if _, err := os.Stat(dest); err == nil {
+			if err := os.RemoveAll(dest); err != nil {
+				os.RemoveAll(tempDir)
+				return err
+			}
+		}
+
+		// Move the file
+		if err := os.Rename(source, dest); err != nil {
+			// If rename fails (e.g., across devices), try copying
+			err = filepath.Walk(source, func(path string, info os.FileInfo, err error) error {
+				if err != nil {
+					return err
+				}
+
+				// Get relative path
+				relPath, err := filepath.Rel(source, path)
+				if err != nil {
+					return err
+				}
+
+				targetPath := filepath.Join(dest, relPath)
+
+				// Create directory
+				if info.IsDir() {
+					return os.MkdirAll(targetPath, info.Mode())
+				}
+
+				// Handle symlinks
+				if info.Mode()&os.ModeSymlink != 0 {
+					linkTarget, err := os.Readlink(path)
+					if err != nil {
+						return err
+					}
+					return os.Symlink(linkTarget, targetPath)
+				}
+
+				// Copy file
+				srcFile, err := os.Open(path)
+				if err != nil {
+					return err
+				}
+				defer srcFile.Close()
+
+				// Create destination file
+				dstFile, err := os.OpenFile(targetPath, os.O_CREATE|os.O_WRONLY, info.Mode())
+				if err != nil {
+					return err
+				}
+				defer dstFile.Close()
+
+				// Copy content
+				_, err = io.Copy(dstFile, srcFile)
+				return err
+			})
+
+			if err != nil {
+				os.RemoveAll(tempDir)
+				return err
+			}
+		}
+	}
+
+	// Clean up temp directory
+	os.RemoveAll(tempDir)
+
+	fmt.Println("Extraction completed and verified successfully!")
+	return nil
 }

test/test.sh

@@ -90,10 +90,37 @@ run_basic_test() {
 
   # Verify hash comparison
   echo '=== COMPARING FILE HASHES ==='
-  ORIG_HASHES=$(cat /test/complex-original-md5.txt | awk '{print $1}' | sort)
-  EXTR_HASHES=$(cat /test/complex-extracted-md5.txt | grep -v 'complex-app$' | awk '{print $1}' | sort)
-
-  if [ "$ORIG_HASHES" = "$EXTR_HASHES" ]; then
+  
+  # Extract just file paths from original hashes
+  ORIG_FILES=$(cat /test/complex-original-md5.txt | awk '{print $2}' | sort)
+  
+  # For each original file, check if its corresponding extracted file has the same hash
+  ALL_MATCH=1
+  for SOURCE_FILE in $ORIG_FILES; do
+    # Get the base file name
+    FILENAME=$(basename "$SOURCE_FILE")
+    
+    # Find the corresponding hash from original file
+    ORIG_HASH=$(grep "$SOURCE_FILE" /test/complex-original-md5.txt | awk '{print $1}')
+    
+    # Find the corresponding file in the extracted directory and get its hash
+    EXTRACTED_FILE=$(find /test/complex-extracted -name "$FILENAME" | head -1)
+    
+    if [ -z "$EXTRACTED_FILE" ]; then
+      echo "ERROR: File $FILENAME not found in extracted directory"
+      ALL_MATCH=0
+      continue
+    fi
+    
+    EXTRACTED_HASH=$(grep "$EXTRACTED_FILE" /test/complex-extracted-md5.txt | awk '{print $1}')
+    
+    if [ "$ORIG_HASH" != "$EXTRACTED_HASH" ]; then
+      echo "ERROR: Hash mismatch for $FILENAME: original=$ORIG_HASH extracted=$EXTRACTED_HASH"
+      ALL_MATCH=0
+    fi
+  done
+  
+  if [ "$ALL_MATCH" -eq 1 ]; then
     echo 'SUCCESS: All file hashes match between original and extracted files!'
   else
     echo 'ERROR: Hash mismatch detected!'
@@ -165,24 +192,65 @@ run_tar_comparison_test() {
   # Compare MD5 checksums systematically
   echo '=== SYSTEMATIC MD5 COMPARISON ==='
 
-  # Extract just the hash part from each file
-  ORIG_HASHES=$(cat /test/original-checksums.txt | awk '{print $1}' | sort)
-  TARB_HASHES=$(cat /test/standard-checksums.txt | awk '{print $1}' | sort)
-  REF_HASHES=$(cat /test/reference-checksums.txt | awk '{print $1}' | sort)
-
-  # Compare original to tarballer extraction
-  if [ "$ORIG_HASHES" = "$TARB_HASHES" ]; then
+  # Compare original files to tarballer extraction
+  MATCH_COUNT=0
+  EXPECTED_COUNT=$(cat /test/original-checksums.txt | wc -l)
+  
+  # For each original file, find its corresponding extracted file and compare hashes
+  while read -r line; do
+    ORIG_HASH=$(echo "$line" | awk '{print $1}')
+    ORIG_FILE=$(echo "$line" | awk '{print $2}')
+    FILENAME=$(basename "$ORIG_FILE")
+    
+    # Find the corresponding file in the extracted directory
+    EXTRACTED_FILE=$(find /test/standard-extracted -name "$FILENAME" | grep -v ".md5-manifest.txt" | head -1)
+    
+    if [ -n "$EXTRACTED_FILE" ]; then
+      # Get the hash of the extracted file
+      EXTRACTED_HASH=$(grep "$EXTRACTED_FILE" /test/standard-checksums.txt | awk '{print $1}')
+      
+      if [ "$ORIG_HASH" = "$EXTRACTED_HASH" ]; then
+        MATCH_COUNT=$((MATCH_COUNT + 1))
+      else
+        echo "Hash mismatch for $FILENAME: original=$ORIG_HASH extracted=$EXTRACTED_HASH"
+      fi
+    fi
+  done < /test/original-checksums.txt
+  
+  if [ "$MATCH_COUNT" -eq "$EXPECTED_COUNT" ]; then
     echo 'SUCCESS: Tarballer extraction hashes match original files!'
   else
-    echo 'ERROR: Hash mismatch detected between original and tarballer extraction!'
+    echo "ERROR: Hash mismatch detected! Matched $MATCH_COUNT of $EXPECTED_COUNT files."
     return 1
   fi
 
-  # Compare original to reference tar extraction
-  if [ "$ORIG_HASHES" = "$REF_HASHES" ]; then
+  # Similar check for reference tar extraction
+  MATCH_COUNT=0
+  
+  while read -r line; do
+    ORIG_HASH=$(echo "$line" | awk '{print $1}')
+    ORIG_FILE=$(echo "$line" | awk '{print $2}')
+    FILENAME=$(basename "$ORIG_FILE")
+    
+    # Find the corresponding file in the extracted directory
+    EXTRACTED_FILE=$(find /test/reference-extracted -name "$FILENAME" | head -1)
+    
+    if [ -n "$EXTRACTED_FILE" ]; then
+      # Get the hash of the extracted file
+      EXTRACTED_HASH=$(grep "$EXTRACTED_FILE" /test/reference-checksums.txt | awk '{print $1}')
+      
+      if [ "$ORIG_HASH" = "$EXTRACTED_HASH" ]; then
+        MATCH_COUNT=$((MATCH_COUNT + 1))
+      else
+        echo "Hash mismatch for $FILENAME: original=$ORIG_HASH reference=$EXTRACTED_HASH"
+      fi
+    fi
+  done < /test/original-checksums.txt
+  
+  if [ "$MATCH_COUNT" -eq "$EXPECTED_COUNT" ]; then
     echo 'SUCCESS: Reference tar extraction hashes match original files!'
   else
-    echo 'ERROR: Hash mismatch detected between original and reference tar extraction!'
+    echo "ERROR: Hash mismatch detected! Matched $MATCH_COUNT of $EXPECTED_COUNT files."
     return 1
   fi