add checksum based status check, alternative to timestamp based

README.md

@@ -244,6 +244,25 @@ Task will compare the modification date/time of the files to determine if it's
 necessary to run the task. If not, it will just print a message like
 `Task "js" is up to date`.
 
+If you prefer this check to be made by the content of the files, instead of
+its timestamp, just set the `method` property to `checksum`.
+You will probably want to ignore the `.task` folder in your `.gitignore` file
+(It's there that Task stores the last checksum).
+This feature is still experimental and can change until it's stable.
+
+```yml
+build:
+  cmds:
+    - go build .
+  sources:
+    - ./*.go
+  generates:
+    - app{{exeExt}}
+  method: checksum
+```
+
+> TIP: method `none` skips any validation and always run the task.
+
 Alternatively, you can inform a sequence of tests as `status`. If no error
 is returned (exit status 0), the task is considered up-to-date:
 

Taskvars.yml

@@ -6,3 +6,4 @@ GO_PACKAGES:
   ./args
   ./cmd/task
   ./execext
+  ./status

status.go

@@ -2,19 +2,52 @@ package task
 
 import (
 	"context"
-	"os"
-	"path/filepath"
-	"time"
+	"fmt"
 
 	"github.com/go-task/task/execext"
-	"github.com/mattn/go-zglob"
+	"github.com/go-task/task/status"
 )
 
 func (t *Task) isUpToDate(ctx context.Context) (bool, error) {
 	if len(t.Status) > 0 {
 		return t.isUpToDateStatus(ctx)
 	}
-	return t.isUpToDateTimestamp(ctx)
+
+	checker, err := t.getStatusChecker()
+	if err != nil {
+		return false, err
+	}
+
+	return checker.IsUpToDate()
+}
+
+func (t *Task) statusOnError() error {
+	checker, err := t.getStatusChecker()
+	if err != nil {
+		return err
+	}
+	return checker.OnError()
+}
+
+func (t *Task) getStatusChecker() (status.Checker, error) {
+	switch t.Method {
+	case "", "timestamp":
+		return &status.Timestamp{
+			Dir:       t.Dir,
+			Sources:   t.Sources,
+			Generates: t.Generates,
+		}, nil
+	case "checksum":
+		return &status.Checksum{
+			Dir:     t.Dir,
+			Task:    t.Task,
+			Sources: t.Sources,
+		}, nil
+	case "none":
+		return status.None{}, nil
+	default:
+		return nil, fmt.Errorf(`task: invalid method "%s"`, t.Method)
+	}
 }
 
 func (t *Task) isUpToDateStatus(ctx context.Context) (bool, error) {
@@ -31,93 +64,3 @@ func (t *Task) isUpToDateStatus(ctx context.Context) (bool, error) {
 	}
 	return true, nil
 }
-
-func (t *Task) isUpToDateTimestamp(ctx context.Context) (bool, error) {
-	if len(t.Sources) == 0 || len(t.Generates) == 0 {
-		return false, nil
-	}
-
-	sourcesMaxTime, err := getPatternsMaxTime(t.Dir, t.Sources)
-	if err != nil || sourcesMaxTime.IsZero() {
-		return false, nil
-	}
-
-	generatesMinTime, err := getPatternsMinTime(t.Dir, t.Generates)
-	if err != nil || generatesMinTime.IsZero() {
-		return false, nil
-	}
-	return !generatesMinTime.Before(sourcesMaxTime), nil
-}
-
-func getPatternsMinTime(dir string, patterns []string) (m time.Time, err error) {
-	for _, p := range patterns {
-		if !filepath.IsAbs(p) {
-			p = filepath.Join(dir, p)
-		}
-		mp, err := getPatternMinTime(p)
-		if err != nil {
-			return time.Time{}, err
-		}
-		m = minTime(m, mp)
-	}
-	return
-}
-func getPatternsMaxTime(dir string, patterns []string) (m time.Time, err error) {
-	for _, p := range patterns {
-		if !filepath.IsAbs(p) {
-			p = filepath.Join(dir, p)
-		}
-		mp, err := getPatternMaxTime(p)
-		if err != nil {
-			return time.Time{}, err
-		}
-		m = maxTime(m, mp)
-	}
-	return
-}
-
-func getPatternMinTime(pattern string) (m time.Time, err error) {
-	files, err := zglob.Glob(pattern)
-	if err != nil {
-		return time.Time{}, err
-	}
-
-	for _, f := range files {
-		info, err := os.Stat(f)
-		if err != nil {
-			return time.Time{}, err
-		}
-		m = minTime(m, info.ModTime())
-	}
-	return
-}
-
-func getPatternMaxTime(pattern string) (m time.Time, err error) {
-	files, err := zglob.Glob(pattern)
-	if err != nil {
-		return time.Time{}, err
-	}
-
-	for _, f := range files {
-		info, err := os.Stat(f)
-		if err != nil {
-			return time.Time{}, err
-		}
-		m = maxTime(m, info.ModTime())
-	}
-	return
-}
-
-func minTime(a, b time.Time) time.Time {
-	if !a.IsZero() && a.Before(b) {
-		return a
-	}
-	return b
-}
-
-func maxTime(a, b time.Time) time.Time {
-	if a.After(b) {
-		return a
-	}
-	return b
-}

status/checksum.go

@@ -0,0 +1,64 @@
+package status
+
+import (
+	"crypto/md5"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"os"
+	"path/filepath"
+	"strings"
+)
+
+// Checksum validades if a task is up to date by calculating its source
+// files checksum
+type Checksum struct {
+	Dir     string
+	Task    string
+	Sources []string
+}
+
+// IsUpToDate implements the Checker interface
+func (c *Checksum) IsUpToDate() (bool, error) {
+	checksumFile := filepath.Join(c.Dir, ".task", c.Task)
+
+	data, _ := ioutil.ReadFile(checksumFile)
+	oldMd5 := strings.TrimSpace(string(data))
+
+	sources, err := glob(c.Dir, c.Sources)
+	if err != nil {
+		return false, err
+	}
+
+	newMd5, err := c.checksum(sources...)
+	if err != nil {
+		return false, nil
+	}
+
+	_ = os.MkdirAll(filepath.Join(c.Dir, ".task"), 0755)
+	if err = ioutil.WriteFile(checksumFile, []byte(newMd5), 0644); err != nil {
+		return false, err
+	}
+	return oldMd5 == newMd5, nil
+}
+
+func (c *Checksum) checksum(files ...string) (string, error) {
+	h := md5.New()
+
+	for _, f := range files {
+		f, err := os.Open(f)
+		if err != nil {
+			return "", err
+		}
+		if _, err := io.Copy(h, f); err != nil {
+			return "", err
+		}
+	}
+
+	return fmt.Sprintf("%x", h.Sum(nil)), nil
+}
+
+// OnError implements the Checker interface
+func (c *Checksum) OnError() error {
+	return os.Remove(filepath.Join(c.Dir, ".task", c.Task))
+}

status/glob.go

@@ -0,0 +1,23 @@
+package status
+
+import (
+	"path/filepath"
+	"sort"
+
+	"github.com/mattn/go-zglob"
+)
+
+func glob(dir string, globs []string) (files []string, err error) {
+	for _, g := range globs {
+		if !filepath.IsAbs(g) {
+			g = filepath.Join(dir, g)
+		}
+		f, err := zglob.Glob(g)
+		if err != nil {
+			return nil, err
+		}
+		files = append(files, f...)
+	}
+	sort.Strings(files)
+	return
+}

status/none.go

@@ -0,0 +1,14 @@
+package status
+
+// None is a no-op Checker
+type None struct{}
+
+// IsUpToDate implements the Checker interface
+func (None) IsUpToDate() (bool, error) {
+	return false, nil
+}
+
+// OnError implements the Checker interface
+func (None) OnError() error {
+	return nil
+}

status/status.go

@@ -0,0 +1,13 @@
+package status
+
+var (
+	_ Checker = &Timestamp{}
+	_ Checker = &Checksum{}
+	_ Checker = None{}
+)
+
+// Checker is an interface that checks if the status is up-to-date
+type Checker interface {
+	IsUpToDate() (bool, error)
+	OnError() error
+}

status/timestamp.go

@@ -0,0 +1,85 @@
+package status
+
+import (
+	"os"
+	"time"
+)
+
+// Timestamp checks if any source change compared with the generated files,
+// using file modifications timestamps.
+type Timestamp struct {
+	Dir       string
+	Sources   []string
+	Generates []string
+}
+
+// IsUpToDate implements the Checker interface
+func (t *Timestamp) IsUpToDate() (bool, error) {
+	if len(t.Sources) == 0 || len(t.Generates) == 0 {
+		return false, nil
+	}
+
+	sources, err := glob(t.Dir, t.Sources)
+	if err != nil {
+		return false, nil
+	}
+	generates, err := glob(t.Dir, t.Generates)
+	if err != nil {
+		return false, nil
+	}
+
+	sourcesMaxTime, err := getMaxTime(sources...)
+	if err != nil || sourcesMaxTime.IsZero() {
+		return false, nil
+	}
+
+	generatesMinTime, err := getMinTime(generates...)
+	if err != nil || generatesMinTime.IsZero() {
+		return false, nil
+	}
+
+	return !generatesMinTime.Before(sourcesMaxTime), nil
+}
+
+func getMinTime(files ...string) (time.Time, error) {
+	var t time.Time
+	for _, f := range files {
+		info, err := os.Stat(f)
+		if err != nil {
+			return time.Time{}, err
+		}
+		t = minTime(t, info.ModTime())
+	}
+	return t, nil
+}
+
+func getMaxTime(files ...string) (time.Time, error) {
+	var t time.Time
+	for _, f := range files {
+		info, err := os.Stat(f)
+		if err != nil {
+			return time.Time{}, err
+		}
+		t = maxTime(t, info.ModTime())
+	}
+	return t, nil
+}
+
+func minTime(a, b time.Time) time.Time {
+	if !a.IsZero() && a.Before(b) {
+		return a
+	}
+	return b
+}
+
+func maxTime(a, b time.Time) time.Time {
+	if a.After(b) {
+		return a
+	}
+	return b
+}
+
+// OnError implements the Checker interface
+func (*Timestamp) OnError() error {
+	return nil
+}

task.go

@@ -49,6 +49,7 @@ type Tasks map[string]*Task
 
 // Task represents a task
 type Task struct {
+	Task      string
 	Cmds      []*Cmd
 	Deps      []*Dep
 	Desc      string
@@ -60,6 +61,7 @@ type Task struct {
 	Set       string
 	Env       Vars
 	Silent    bool
+	Method    string
 }
 
 // Run runs Task
@@ -135,14 +137,17 @@ func (e *Executor) RunTask(ctx context.Context, call Call) error {
 			return err
 		}
 		if upToDate {
-			e.printfln(`task: Task "%s" is up to date`, call.Task)
+			e.printfln(`task: Task "%s" is up to date`, t.Task)
 			return nil
 		}
 	}
 
 	for i := range t.Cmds {
 		if err := e.runCommand(ctx, t, call, i); err != nil {
-			return &taskRunError{call.Task, err}
+			if err2 := t.statusOnError(); err2 != nil {
+				e.verbosePrintfln("task: error cleaning status on error: %v", err2)
+			}
+			return &taskRunError{t.Task, err}
 		}
 	}
 	return nil

task_test.go

@@ -312,6 +312,40 @@ func TestGenerates(t *testing.T) {
 	}
 }
 
+func TestStatusChecksum(t *testing.T) {
+	const dir = "testdata/checksum"
+
+	files := []string{
+		"generated.txt",
+		".task/build",
+	}
+
+	for _, f := range files {
+		_ = os.Remove(filepath.Join(dir, f))
+
+		_, err := os.Stat(filepath.Join(dir, f))
+		assert.Error(t, err)
+	}
+
+	var buff bytes.Buffer
+	e := task.Executor{
+		Dir:    dir,
+		Stdout: &buff,
+		Stderr: &buff,
+	}
+	assert.NoError(t, e.ReadTaskfile())
+
+	assert.NoError(t, e.Run(task.Call{Task: "build"}))
+	for _, f := range files {
+		_, err := os.Stat(filepath.Join(dir, f))
+		assert.NoError(t, err)
+	}
+
+	buff.Reset()
+	assert.NoError(t, e.Run(task.Call{Task: "build"}))
+	assert.Equal(t, `task: Task "build" is up to date`+"\n", buff.String())
+}
+
 func TestInit(t *testing.T) {
 	const dir = "testdata/init"
 	var file = filepath.Join(dir, "Taskfile.yml")

taskfile.go

@@ -31,6 +31,10 @@ func (e *Executor) ReadTaskfile() error {
 	if err := mergo.MapWithOverwrite(&e.Tasks, osTasks); err != nil {
 		return err
 	}
+	for name, task := range e.Tasks {
+		task.Task = name
+	}
+
 	return e.readTaskvars()
 }
 

testdata/checksum/.gitignore

@@ -0,0 +1,2 @@
+.task/
+generated.txt

testdata/checksum/Taskfile.yml

@@ -0,0 +1,8 @@
+build:
+  cmds:
+    - cp ./source.txt ./generated.txt
+  sources:
+    - ./source.txt
+  generates:
+    - ./generated.txt
+  method: checksum

testdata/checksum/source.txt

@@ -0,0 +1 @@
+Hello, World!

variables.go

@@ -210,6 +210,7 @@ func (e *Executor) CompiledTask(call Call) (*Task, error) {
 	}
 
 	new := Task{
+		Task:      origTask.Task,
 		Desc:      r.replace(origTask.Desc),
 		Sources:   r.replaceSlice(origTask.Sources),
 		Generates: r.replaceSlice(origTask.Generates),
@@ -219,6 +220,7 @@ func (e *Executor) CompiledTask(call Call) (*Task, error) {
 		Set:       r.replace(origTask.Set),
 		Env:       r.replaceVars(origTask.Env),
 		Silent:    origTask.Silent,
+		Method:    r.replace(origTask.Method),
 	}
 	if e.Dir != "" && !filepath.IsAbs(new.Dir) {
 		new.Dir = filepath.Join(e.Dir, new.Dir)