[feature/workers] handling directory recursively and computing hash w…

…ith workers (#6)

cmd/hash.go

@@ -1,40 +1,55 @@
 package cmd
 
 import (
-	"fmt"
 	"hash"
+	"io"
+	"os"
 
 	"github.com/spf13/cobra"
 )
 
+type HashResult struct {
+	res  []byte
+	path string
+	err  error
+}
+
+type JobsParam struct {
+	path string
+	h    hash.Hash
+}
+
 // hashCmd represents the hash command
 var hashCmd = &cobra.Command{
-	Use:           "hash",
-	Short:         "",
-	Long:          ``,
+	Use:   "hash",
+	Short: "The hash command computes the hash of a given FILE.",
+	Long: `The hash command computes the hash of a given FILE.
+
+Without FILE or when FILE is '-', read the standard input.
+If the list of FILE contains a directory, it will be proceed recursively.
+If the list of FILE contains './...' it will proceed directories recursively from the current directory.`,
 	SilenceErrors: true,
 }
 
 func init() {
 	rootCmd.AddCommand(hashCmd)
 }
 
-func getFilesToCompute(args []string) []string {
-	if len(args) == 0 || (len(args) == 1 && args[0] == "-") {
-		return []string{"-"}
-	}
-	return args
-}
-
-func computeFiles(filesToCheck []string, h hash.Hash) (string, error) {
-	var res string
-	for _, filePath := range filesToCheck {
-		hashedValue, err := computeHash(filePath, h)
+func computeHash(path string, h hash.Hash) ([]byte, error) {
+	var r io.Reader
+	if path == "-" {
+		r = os.Stdin
+	} else {
+		file, err := os.Open(path)
 		if err != nil {
-			return "", err
+			return nil, err
 		}
-		res += fmt.Sprintf("%x %s\n", hashedValue, filePath)
-		h.Reset()
+		defer file.Close()
+		r = file
+	}
+	_, err := io.Copy(h, r)
+	if err != nil {
+		return nil, err
 	}
-	return res, nil
+	return h.Sum(nil), nil
 }

cmd/md5Hash.go

@@ -2,7 +2,6 @@ package cmd
 
 import (
 	"crypto/md5"
-	"fmt"
 
 	"github.com/spf13/cobra"
 )
@@ -13,16 +12,28 @@ var md5HashCmd = &cobra.Command{
 	Short: "Display MD5 checksums (128 bits).",
 	Long: `Display MD5 checksums (128 bits).
 
-Without FILE or when FILE is '-', read the standard input.`,
+Without FILE or when FILE is '-', read the standard input.
+If the list of FILE contains a directory, it will be proceed recursively.
+If the list of FILE contains './...' it will proceed directories recursively from the current directory.`,
 	RunE: func(cmd *cobra.Command, args []string) error {
-		filesToCheck := getFilesToCompute(args)
-		h := md5.New()
-		res, err := computeFiles(filesToCheck, h)
+		filesToCheck, err := getFilesToCompute(args)
 		if err != nil {
 			return err
 		}
-		fmt.Print(res)
-		return nil
+
+		numJobs := len(filesToCheck)
+		jobs := make(chan JobsParam, numJobs)
+		results := make(chan HashResult, numJobs)
+
+		initWorkers(jobs, results)
+
+		for _, filePath := range filesToCheck {
+			h := md5.New()
+			jobs <- JobsParam{filePath, h}
+		}
+		close(jobs)
+
+		return waitForResult(numJobs, results)
 	},
 }
 

cmd/root.go

@@ -1,11 +1,13 @@
 package cmd
 
 import (
-	"bufio"
+	"encoding/hex"
 	"fmt"
-	"hash"
-	"io"
+	"io/fs"
 	"os"
+	"path/filepath"
+	"runtime"
+	"slices"
 
 	"github.com/spf13/cobra"
 )
@@ -29,25 +31,85 @@ func Execute() {
 	}
 }
 
-func computeHash(path string, h hash.Hash) ([]byte, error) {
-	if path == "-" {
-		reader := bufio.NewReader(os.Stdin)
-		buffer := make([]byte, 1024)
-		for {
-			n, err := reader.Read(buffer)
-			h.Write(buffer[:n])
-			if err == io.EOF {
-				break
-			}
+// https://stackoverflow.com/questions/37334119/how-to-delete-an-element-from-a-slice-in-golang
+func remove(s []string, i int) []string {
+	s[i] = s[len(s)-1]
+	return s[:len(s)-1]
+}
+
+func getFilesToCompute(args []string) ([]string, error) {
+	var err error
+	if len(args) == 0 || (len(args) == 1 && args[0] == "-") {
+		return []string{"-"}, nil
+	}
+
+	// Here ./ tells to start from the current folder, ... tells to go down recursively.
+	if slices.Contains(args, "./...") {
+		idx := slices.Index(args, "./...")
+		args = remove(args, idx)
+		args, err = computeDirRecursively(args, ".")
+		if err != nil {
+			return nil, err
 		}
-	} else {
-		file, err := os.Open(path)
+	}
+
+	// check for dirs and all files exist
+	for idx, elem := range args {
+		f, err := os.Open(elem)
 		if err != nil {
 			return nil, err
 		}
-		if _, err := io.Copy(h, file); err != nil {
+		defer f.Close()
+		fInfo, err := f.Stat()
+		if err != nil {
 			return nil, err
 		}
+		if fInfo.IsDir() {
+			args = remove(args, idx)
+			args, err = computeDirRecursively(args, elem)
+			if err != nil {
+				return nil, err
+			}
+		}
+	}
+	fmt.Printf("Computing hash for: %d files.", len(args))
+
+	return args, nil
+}
+
+func computeDirRecursively(l []string, rootFolder string) ([]string, error) {
+	err := filepath.WalkDir(rootFolder, func(path string, d fs.DirEntry, err error) error {
+		if err != nil {
+			return err
+		}
+		if !d.IsDir() {
+			l = append(l, path)
+		}
+		return nil
+	})
+	return l, err
+}
+
+func initWorkers(jobs chan JobsParam, results chan HashResult) {
+	for i := 0; i <= runtime.NumCPU(); i++ {
+		go worker(jobs, results)
+	}
+}
+
+func waitForResult(nbJobs int, results chan HashResult) error {
+	for a := 1; a <= nbJobs; a++ {
+		res := <-results
+		if res.err != nil {
+			return res.err
+		}
+		fmt.Printf("%s %s\n", hex.EncodeToString(res.res), res.path)
+	}
+	return nil
+}
+
+func worker(jobs <-chan JobsParam, results chan<- HashResult) {
+	for j := range jobs {
+		hashedValue, err := computeHash(j.path, j.h)
+		results <- HashResult{hashedValue, j.path, err}
 	}
-	return h.Sum(nil), nil
 }

cmd/sha1Hash.go

@@ -2,7 +2,6 @@ package cmd
 
 import (
 	"crypto/sha1"
-	"fmt"
 
 	"github.com/spf13/cobra"
 )
@@ -13,16 +12,28 @@ var sha1HashCmd = &cobra.Command{
 	Short: "Display SHA-1 checksums (160 bits).",
 	Long: `Display SHA-1 checksums (160 bits).
 
-Without FILE or when FILE is '-', read the standard input.`,
+Without FILE or when FILE is '-', read the standard input.
+If the list of FILE contains a directory, it will be proceed recursively.
+If the list of FILE contains './...' it will proceed directories recursively from the current directory.`,
 	RunE: func(cmd *cobra.Command, args []string) error {
-		filesToCheck := getFilesToCompute(args)
-		h := sha1.New()
-		res, err := computeFiles(filesToCheck, h)
+		filesToCheck, err := getFilesToCompute(args)
 		if err != nil {
 			return err
 		}
-		fmt.Print(res)
-		return nil
+
+		numJobs := len(filesToCheck)
+		jobs := make(chan JobsParam, numJobs)
+		results := make(chan HashResult, numJobs)
+
+		initWorkers(jobs, results)
+
+		for _, filePath := range filesToCheck {
+			h := sha1.New()
+			jobs <- JobsParam{filePath, h}
+		}
+		close(jobs)
+
+		return waitForResult(numJobs, results)
 	},
 }
 

cmd/sha224Hash.go

@@ -2,7 +2,6 @@ package cmd
 
 import (
 	"crypto/sha256"
-	"fmt"
 
 	"github.com/spf13/cobra"
 )
@@ -13,16 +12,28 @@ var sha224HashCmd = &cobra.Command{
 	Short: "Display SHA-224 checksums (224 bits).",
 	Long: `Display SHA-224 checksums (224 bits).
 
-Without FILE or when FILE is '-', read the standard input.`,
+Without FILE or when FILE is '-', read the standard input.
+If the list of FILE contains a directory, it will be proceed recursively.
+If the list of FILE contains './...' it will proceed directories recursively from the current directory.`,
 	RunE: func(cmd *cobra.Command, args []string) error {
-		filesToCheck := getFilesToCompute(args)
-		h := sha256.New224()
-		res, err := computeFiles(filesToCheck, h)
+		filesToCheck, err := getFilesToCompute(args)
 		if err != nil {
 			return err
 		}
-		fmt.Print(res)
-		return nil
+
+		numJobs := len(filesToCheck)
+		jobs := make(chan JobsParam, numJobs)
+		results := make(chan HashResult, numJobs)
+
+		initWorkers(jobs, results)
+
+		for _, filePath := range filesToCheck {
+			h := sha256.New224()
+			jobs <- JobsParam{filePath, h}
+		}
+		close(jobs)
+
+		return waitForResult(numJobs, results)
 	},
 }
 

cmd/sha256Hash.go

@@ -2,7 +2,6 @@ package cmd
 
 import (
 	"crypto/sha256"
-	"fmt"
 
 	"github.com/spf13/cobra"
 )
@@ -13,16 +12,28 @@ var sha256HashCmd = &cobra.Command{
 	Short: "Display SHA256 checksums (256 bits).",
 	Long: `Display SHA256 checksums (256 bits).
 
-Without FILE or when FILE is '-', read the standard input.`,
+Without FILE or when FILE is '-', read the standard input.
+If the list of FILE contains a directory, it will be proceed recursively.
+If the list of FILE contains './...' it will proceed directories recursively from the current directory.`,
 	RunE: func(cmd *cobra.Command, args []string) error {
-		filesToCheck := getFilesToCompute(args)
-		h := sha256.New()
-		res, err := computeFiles(filesToCheck, h)
+		filesToCheck, err := getFilesToCompute(args)
 		if err != nil {
 			return err
 		}
-		fmt.Print(res)
-		return nil
+
+		numJobs := len(filesToCheck)
+		jobs := make(chan JobsParam, numJobs)
+		results := make(chan HashResult, numJobs)
+
+		initWorkers(jobs, results)
+
+		for _, filePath := range filesToCheck {
+			h := sha256.New()
+			jobs <- JobsParam{filePath, h}
+		}
+		close(jobs)
+
+		return waitForResult(numJobs, results)
 	},
 }