Implement md5sum

README.md

@@ -45,7 +45,7 @@ Utilities:
  * [ ] `logger`
  * [ ] `logname`
  * [ ] `ls`
- * [ ] `md5sum`
+ * [x] `md5sum`
  * [ ] `mkdir`
  * [ ] `mkfifo`
  * [ ] `mktemp`

error.go

@@ -1,4 +1,4 @@
-package common
+package coreutils
 
 import (
 	"fmt"

go.sum

@@ -4,6 +4,7 @@ github.com/alexflint/go-scalar v1.0.0 h1:NGupf1XV/Xb04wXskDFzS0KWOLH632W/EO4fAFi
 github.com/alexflint/go-scalar v1.0.0/go.mod h1:GpHzbCOZXEKMEcygYQ5n/aa4Aq84zbxjy3MxYW0gjYw=
 github.com/davecgh/go-spew v1.1.0 h1:ZDRjVQ15GmhC3fiQ8ni8+OwkZQO4DARzQgrnXU1Liz8=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
 github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=

hashsum.go

@@ -0,0 +1,199 @@
+package coreutils
+
+import (
+	"bufio"
+	"crypto/md5"
+	"errors"
+	"fmt"
+	"hash"
+	"io"
+	"os"
+	"regexp"
+
+	"github.com/alexflint/go-arg"
+)
+
+var MD5Regex = regexp.MustCompile("^(?P<hash>[0-9a-f]{32})  (?P<filename>.*)$")
+
+func MD5Sum(r io.Reader) (hash.Hash, error) {
+	h := md5.New()
+	if _, err := io.Copy(h, r); err != nil {
+		return nil, err
+	}
+	return h, nil
+}
+
+type CheckingResults struct {
+	ImproperlyFormattedCount uint
+	InvalidChecksumCount     uint
+	FilesNotRead             uint
+}
+
+// SumFunc is a type of function that computes a hash.Hash for data in io.Reader
+type SumFunc func(io.Reader) (hash.Hash, error)
+
+// ImproperlyFormattedErr is an error return when a line for checking
+// has an incorrect number or set of characters
+var ImproperlyFormattedErr = errors.New("improperly formatted line")
+
+func printHash(h hash.Hash, filename string) {
+	fmt.Printf("%x  %v\n", h.Sum(nil), filename)
+}
+
+// PrintSumForFile prints checksum computed by f for file with filename
+func PrintSumForFile(filename string, f SumFunc) error {
+	file, err := os.Open(filename)
+	if err != nil {
+		return err
+	}
+	defer file.Close()
+
+	h, err := f(file)
+	if err != nil {
+		return err
+	}
+
+	printHash(h, filename)
+	return nil
+}
+
+// PrintSumForStdin prints checksum computed by f for data from stdin
+func PrintSumForStdin(f SumFunc) {
+	h, err := f(os.Stdin)
+	if err != nil {
+		PrintToStderr(err)
+		os.Exit(1)
+	}
+	printHash(h, "<stdin>")
+	os.Exit(0)
+}
+
+func checkLine(line string, re *regexp.Regexp, f SumFunc) (valid bool, err error) {
+	matches := re.MatchString(line)
+	if !matches {
+		err = ImproperlyFormattedErr
+		valid = false
+		return
+	}
+	submatches := re.FindStringSubmatch(line)
+	hash := submatches[1]
+	filename := submatches[2]
+
+	file, err := os.Open(filename)
+	if err != nil {
+		valid = false
+		return
+	}
+	defer file.Close()
+	h, err := f(file)
+	if err != nil {
+		valid = false
+		return
+	}
+
+	result := "OK"
+	valid = true
+	if fmt.Sprintf("%x", h.Sum(nil)) != hash {
+		result = "FAILED"
+		valid = false
+	}
+	fmt.Printf("%s: %s\n", filename, result)
+
+	return
+}
+
+// CheckSumsInReader checks checksums in r using f
+func CheckSumsInReader(r io.Reader, re *regexp.Regexp, f SumFunc) (*CheckingResults, error) {
+	scanner := bufio.NewScanner(r)
+
+	results := CheckingResults{}
+	for scanner.Scan() {
+		valid, err := checkLine(scanner.Text(), re, f)
+		if err == ImproperlyFormattedErr {
+			results.ImproperlyFormattedCount++
+		} else if err != nil {
+			PrintToStderr(err)
+			results.FilesNotRead++
+		} else if !valid {
+			results.InvalidChecksumCount++
+		}
+	}
+	return &results, scanner.Err()
+}
+
+// CheckSumsInFile checks checksums in file with filename using f
+func CheckSumsInFile(filename string, re *regexp.Regexp, f SumFunc) (*CheckingResults, error) {
+	file, err := os.Open(filename)
+	if err != nil {
+		return nil, err
+	}
+	defer file.Close()
+
+	return CheckSumsInReader(file, re, f)
+}
+
+// PrintCheckingResults prints the number of mismatched checksums and improperly formatted lines to os.Stderr
+func PrintCheckingResults(results CheckingResults) {
+	if results.InvalidChecksumCount > 0 {
+		PrintToStderr(
+			fmt.Sprintf("%v computed checksums did NOT match", results.InvalidChecksumCount),
+		)
+	}
+	if results.ImproperlyFormattedCount > 0 {
+		PrintToStderr(
+			fmt.Sprintf("%v lines are improperly formatted", results.ImproperlyFormattedCount),
+		)
+	}
+	if results.FilesNotRead > 0 {
+		PrintToStderr(
+			fmt.Sprintf("%v listed files could not be read", results.FilesNotRead),
+		)
+	}
+}
+
+// SumMain is a function you're supposed to run in main of a hash sum utility
+func SumMain(re *regexp.Regexp, f SumFunc) {
+	var args struct {
+		Check bool     `arg:"-c"`
+		Files []string `arg:"positional"`
+	}
+
+	arg.MustParse(&args)
+
+	exitCode := 0
+	if len(args.Files) == 0 {
+		if !args.Check {
+			PrintSumForStdin(f)
+		} else {
+			results, err := CheckSumsInReader(os.Stdin, re, f)
+			if err != nil {
+				PrintToStderr(err)
+				exitCode = 1
+			}
+			PrintCheckingResults(*results)
+		}
+	}
+
+	totalResults := CheckingResults{}
+	for _, filename := range args.Files {
+		if !args.Check {
+			err := PrintSumForFile(filename, f)
+			if err != nil {
+				PrintToStderr(err)
+				exitCode = 1
+			}
+		} else {
+			results, err := CheckSumsInFile(filename, re, f)
+			if err != nil {
+				PrintToStderr(err)
+				exitCode = 1
+			}
+			if results != nil {
+				totalResults.ImproperlyFormattedCount += results.ImproperlyFormattedCount
+				totalResults.InvalidChecksumCount += results.InvalidChecksumCount
+			}
+		}
+	}
+	PrintCheckingResults(totalResults)
+	os.Exit(exitCode)
+}

hashsum_test.go

@@ -0,0 +1,69 @@
+package coreutils
+
+import (
+	"bytes"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/suite"
+	"path"
+	"strings"
+	"testing"
+)
+
+type CheckLineTestSuite struct {
+	suite.Suite
+}
+
+func (suite *CheckLineTestSuite) TestImproperlyFormattedLine() {
+	t := suite.T()
+	valid, err := checkLine("name chef", MD5Regex, MD5Sum)
+	assert.Equal(t, false, valid)
+	assert.Error(t, err)
+	assert.Equal(t, ImproperlyFormattedErr, err)
+}
+
+func (suite *CheckLineTestSuite) TestMismatchedChecksum() {
+	t := suite.T()
+	line := "4a5fb9ebd6c8ea7efb53d071053ef778  " + path.Join("test_data", "some_file.txt")
+	valid, err := checkLine(line, MD5Regex, MD5Sum)
+	assert.Equal(t, false, valid)
+	assert.NoError(t, err)
+}
+
+func (suite *CheckLineTestSuite) TestValid() {
+	t := suite.T()
+	line := "6a5fb9ebd6c8ea7efb53d071053ef778  " + path.Join("test_data", "some_file.txt")
+	valid, err := checkLine(line, MD5Regex, MD5Sum)
+	assert.Equal(t, true, valid)
+	assert.NoError(t, err)
+}
+
+func TestCheckLineTestSuite(t *testing.T) {
+	suite.Run(t, new(CheckLineTestSuite))
+}
+
+type CheckSumsInReaderTestSuite struct {
+	suite.Suite
+}
+
+func (suite *CheckSumsInReaderTestSuite) Test() {
+	t := suite.T()
+	buf := bytes.NewBufferString(strings.Join(
+		[]string{
+			"name chef",
+			"4a5fb9ebd6c8ea7efb53d071053ef778  " + path.Join("test_data", "some_file.txt"),
+			"6a5fb9ebd6c8ea7efb53d071053ef778  " + path.Join("test_data", "some_file.txt"),
+			"4a5fb9ebd6c8ea7efb53d071053ef778  nonexistant_file",
+		},
+		"\n",
+	))
+	results, err := CheckSumsInReader(buf, MD5Regex, MD5Sum)
+	assert.NotNil(t, results)
+	assert.Equal(t, uint(1), results.ImproperlyFormattedCount)
+	assert.Equal(t, uint(1), results.FilesNotRead)
+	assert.Equal(t, uint(1), results.InvalidChecksumCount)
+	assert.NoError(t, err)
+}
+
+func TestCheckSumsInReaderTestSuite(t *testing.T) {
+	suite.Run(t, new(CheckSumsInReaderTestSuite))
+}

md5sum/md5sum.go

@@ -1,141 +1,9 @@
 package main
 
 import (
-	"bufio"
-	"crypto/md5"
-	"errors"
-	"fmt"
-	"hash"
-	"io"
-	"os"
-	"strings"
-
-	"github.com/alexflint/go-arg"
 	common "source.heropunch.io/tomo/go-coreutils"
 )
 
-var ImproperlyFormattedErr = errors.New("improperly formatted line")
-
-func printSumForFile(filename string, sumFunc func(io.Reader) (hash.Hash, error)) error {
-	file, err := os.Open(filename)
-	if err != nil {
-		return err
-	}
-	defer file.Close()
-
-	h, err := sumFunc(file)
-	if err != nil {
-		return err
-	}
-
-	printHash(h, filename)
-	return nil
-}
-
-func printSumForStdin(sumFunc func(io.Reader) (hash.Hash, error)) {
-	h, err := sumFunc(os.Stdin)
-	if err != nil {
-		common.PrintToStderr(err)
-		os.Exit(1)
-	}
-	printHash(h, "<stdin>")
-	os.Exit(0)
-}
-
-func md5sum(r io.Reader) (hash.Hash, error) {
-	h := md5.New()
-	if _, err := io.Copy(h, r); err != nil {
-		return nil, err
-	}
-	return h, nil
-}
-
-func printHash(h hash.Hash, filename string) {
-	fmt.Printf("%x  %v\n", h.Sum(nil), filename)
-}
-
-func checkSumForFile(filename string) error {
-	file, err := os.Open(filename)
-	if err != nil {
-		return err
-	}
-	defer file.Close()
-
-	return checkSums(file)
-}
-
-func checkSums(r io.Reader) error {
-	scanner := bufio.NewScanner(r)
-
-	for scanner.Scan() {
-		err := checkMD5Sum(scanner.Text())
-		if err != nil {
-			return err
-		}
-	}
-	return scanner.Err()
-}
-
-func checkMD5Sum(line string) error {
-	s := strings.SplitN(line, "  ", 2)
-	hash := s[0]
-	filename := s[1]
-
-	if filename == "" {
-		return ImproperlyFormattedErr
-	}
-
-	file, err := os.Open(filename)
-	if err != nil {
-		return err
-	}
-	defer file.Close()
-
-	h, err := md5sum(file)
-	if err != nil {
-		return err
-	}
-	result := "OK"
-	if fmt.Sprintf("%x", h.Sum(nil)) != hash {
-		result = "FAILED"
-	}
-	fmt.Printf("%s: %s\n", filename, result)
-
-	return nil
-}
-
 func main() {
-	var args struct {
-		Check bool     `arg:"-c"`
-		Files []string `arg:"positional"`
-	}
-
-	arg.MustParse(&args)
-
-	exitCode := 0
-	if len(args.Files) == 0 {
-		if !args.Check {
-			printSumForStdin(md5sum)
-		} else {
-			fmt.Println("Checking from stdin")
-		}
-	}
-
-	for _, filename := range args.Files {
-		if !args.Check {
-			err := printSumForFile(filename, md5sum)
-			if err != nil {
-				common.PrintToStderr(err)
-				exitCode = 1
-			}
-		} else {
-			err := checkSumForFile(filename)
-			if err != nil {
-				common.PrintToStderr(err)
-				exitCode = 1
-			}
-		}
-	}
-
-	os.Exit(exitCode)
+	common.SumMain(common.MD5Regex, common.MD5Sum)
 }

test_data/some_file.txt

@@ -0,0 +1 @@
+blah blah