Use git attributes to determine generated and vendored status for language stats and diffs (#16773)

Replaces #16262
Replaces #16250
Replaces #14833

This PR first implements a `git check-attr` pipe reader - using `git check-attr --stdin -z --cached` - taking account of the change in the output format in git 1.8.5 and creates a helper function to read a tree into a temporary index file for that pipe reader.

It then wires this in to the language stats helper and into the git diff generation.

Files which are marked generated will be folded by default.

Fixes #14786
Fixes #12653

modules/analyze/generated.go

@@ -0,0 +1,28 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package analyze
+
+import (
+	"path/filepath"
+	"strings"
+
+	"github.com/go-enry/go-enry/v2/data"
+)
+
+// IsGenerated returns whether or not path is a generated path.
+func IsGenerated(path string) bool {
+	ext := strings.ToLower(filepath.Ext(path))
+	if _, ok := data.GeneratedCodeExtensions[ext]; ok {
+		return true
+	}
+
+	for _, m := range data.GeneratedCodeNameMatchers {
+		if m(path) {
+			return true
+		}
+	}
+
+	return false
+}

modules/git/repo_attribute.go

@@ -6,7 +6,12 @@ package git
 
 import (
 	"bytes"
+	"context"
 	"fmt"
+	"io"
+	"os"
+	"strconv"
+	"strings"
 )
 
 // CheckAttributeOpts represents the possible options to CheckAttribute
@@ -21,7 +26,7 @@ type CheckAttributeOpts struct {
 func (repo *Repository) CheckAttribute(opts CheckAttributeOpts) (map[string]map[string]string, error) {
 	err := LoadGitVersion()
 	if err != nil {
-		return nil, fmt.Errorf("Git version missing: %v", err)
+		return nil, fmt.Errorf("git version missing: %v", err)
 	}
 
 	stdOut := new(bytes.Buffer)
@@ -55,13 +60,14 @@ func (repo *Repository) CheckAttribute(opts CheckAttributeOpts) (map[string]map[
 	cmd := NewCommand(cmdArgs...)
 
 	if err := cmd.RunInDirPipeline(repo.Path, stdOut, stdErr); err != nil {
-		return nil, fmt.Errorf("Failed to run check-attr: %v\n%s\n%s", err, stdOut.String(), stdErr.String())
+		return nil, fmt.Errorf("failed to run check-attr: %v\n%s\n%s", err, stdOut.String(), stdErr.String())
 	}
 
+	// FIXME: This is incorrect on versions < 1.8.5
 	fields := bytes.Split(stdOut.Bytes(), []byte{'\000'})
 
 	if len(fields)%3 != 1 {
-		return nil, fmt.Errorf("Wrong number of fields in return from check-attr")
+		return nil, fmt.Errorf("wrong number of fields in return from check-attr")
 	}
 
 	var name2attribute2info = make(map[string]map[string]string)
@@ -80,3 +86,276 @@ func (repo *Repository) CheckAttribute(opts CheckAttributeOpts) (map[string]map[
 
 	return name2attribute2info, nil
 }
+
+// CheckAttributeReader provides a reader for check-attribute content that can be long running
+type CheckAttributeReader struct {
+	// params
+	Attributes []string
+	Repo       *Repository
+	IndexFile  string
+	WorkTree   string
+
+	stdinReader io.ReadCloser
+	stdinWriter *os.File
+	stdOut      attributeWriter
+	cmd         *Command
+	env         []string
+	ctx         context.Context
+	cancel      context.CancelFunc
+	running     chan struct{}
+}
+
+// Init initializes the cmd
+func (c *CheckAttributeReader) Init(ctx context.Context) error {
+	c.running = make(chan struct{})
+	cmdArgs := []string{"check-attr", "--stdin", "-z"}
+
+	if len(c.IndexFile) > 0 && CheckGitVersionAtLeast("1.7.8") == nil {
+		cmdArgs = append(cmdArgs, "--cached")
+		c.env = []string{"GIT_INDEX_FILE=" + c.IndexFile}
+	}
+
+	if len(c.WorkTree) > 0 && CheckGitVersionAtLeast("1.7.8") == nil {
+		c.env = []string{"GIT_WORK_TREE=" + c.WorkTree}
+	}
+
+	if len(c.Attributes) > 0 {
+		cmdArgs = append(cmdArgs, c.Attributes...)
+		cmdArgs = append(cmdArgs, "--")
+	} else {
+		lw := new(nulSeparatedAttributeWriter)
+		lw.attributes = make(chan attributeTriple)
+
+		c.stdOut = lw
+		c.stdOut.Close()
+		return fmt.Errorf("no provided Attributes to check")
+	}
+
+	c.ctx, c.cancel = context.WithCancel(ctx)
+	c.cmd = NewCommandContext(c.ctx, cmdArgs...)
+	var err error
+	c.stdinReader, c.stdinWriter, err = os.Pipe()
+	if err != nil {
+		return err
+	}
+
+	if CheckGitVersionAtLeast("1.8.5") == nil {
+		lw := new(nulSeparatedAttributeWriter)
+		lw.attributes = make(chan attributeTriple, 5)
+
+		c.stdOut = lw
+	} else {
+		lw := new(lineSeparatedAttributeWriter)
+		lw.attributes = make(chan attributeTriple, 5)
+
+		c.stdOut = lw
+	}
+	return nil
+}
+
+// Run run cmd
+func (c *CheckAttributeReader) Run() error {
+	stdErr := new(bytes.Buffer)
+	err := c.cmd.RunInDirTimeoutEnvFullPipelineFunc(c.env, -1, c.Repo.Path, c.stdOut, stdErr, c.stdinReader, func(_ context.Context, _ context.CancelFunc) error {
+		close(c.running)
+		return nil
+	})
+	defer c.cancel()
+	_ = c.stdOut.Close()
+	if err != nil && c.ctx.Err() != nil && err.Error() != "signal: killed" {
+		return fmt.Errorf("failed to run attr-check. Error: %w\nStderr: %s", err, stdErr.String())
+	}
+
+	return nil
+}
+
+// CheckPath check attr for given path
+func (c *CheckAttributeReader) CheckPath(path string) (map[string]string, error) {
+	select {
+	case <-c.ctx.Done():
+		return nil, c.ctx.Err()
+	case <-c.running:
+	}
+
+	if _, err := c.stdinWriter.Write([]byte(path + "\x00")); err != nil {
+		defer c.cancel()
+		return nil, err
+	}
+
+	if err := c.stdinWriter.Sync(); err != nil {
+		defer c.cancel()
+		return nil, err
+	}
+
+	rs := make(map[string]string)
+	for range c.Attributes {
+		select {
+		case attr := <-c.stdOut.ReadAttribute():
+			rs[attr.Attribute] = attr.Value
+		case <-c.ctx.Done():
+			return nil, c.ctx.Err()
+		}
+	}
+	return rs, nil
+}
+
+// Close close pip after use
+func (c *CheckAttributeReader) Close() error {
+	select {
+	case <-c.running:
+	default:
+		close(c.running)
+	}
+	defer c.cancel()
+	return c.stdinWriter.Close()
+}
+
+type attributeWriter interface {
+	io.WriteCloser
+	ReadAttribute() <-chan attributeTriple
+}
+
+type attributeTriple struct {
+	Filename  string
+	Attribute string
+	Value     string
+}
+
+type nulSeparatedAttributeWriter struct {
+	tmp        []byte
+	attributes chan attributeTriple
+	working    attributeTriple
+	pos        int
+}
+
+func (wr *nulSeparatedAttributeWriter) Write(p []byte) (n int, err error) {
+	l, read := len(p), 0
+
+	nulIdx := bytes.IndexByte(p, '\x00')
+	for nulIdx >= 0 {
+		wr.tmp = append(wr.tmp, p[:nulIdx]...)
+		switch wr.pos {
+		case 0:
+			wr.working = attributeTriple{
+				Filename: string(wr.tmp),
+			}
+		case 1:
+			wr.working.Attribute = string(wr.tmp)
+		case 2:
+			wr.working.Value = string(wr.tmp)
+		}
+		wr.tmp = wr.tmp[:0]
+		wr.pos++
+		if wr.pos > 2 {
+			wr.attributes <- wr.working
+			wr.pos = 0
+		}
+		read += nulIdx + 1
+		if l > read {
+			p = p[nulIdx+1:]
+			nulIdx = bytes.IndexByte(p, '\x00')
+		} else {
+			return l, nil
+		}
+	}
+	wr.tmp = append(wr.tmp, p...)
+	return len(p), nil
+}
+
+func (wr *nulSeparatedAttributeWriter) ReadAttribute() <-chan attributeTriple {
+	return wr.attributes
+}
+
+func (wr *nulSeparatedAttributeWriter) Close() error {
+	close(wr.attributes)
+	return nil
+}
+
+type lineSeparatedAttributeWriter struct {
+	tmp        []byte
+	attributes chan attributeTriple
+}
+
+func (wr *lineSeparatedAttributeWriter) Write(p []byte) (n int, err error) {
+	l := len(p)
+
+	nlIdx := bytes.IndexByte(p, '\n')
+	for nlIdx >= 0 {
+		wr.tmp = append(wr.tmp, p[:nlIdx]...)
+
+		if len(wr.tmp) == 0 {
+			// This should not happen
+			if len(p) > nlIdx+1 {
+				wr.tmp = wr.tmp[:0]
+				p = p[nlIdx+1:]
+				nlIdx = bytes.IndexByte(p, '\n')
+				continue
+			} else {
+				return l, nil
+			}
+		}
+
+		working := attributeTriple{}
+		if wr.tmp[0] == '"' {
+			sb := new(strings.Builder)
+			remaining := string(wr.tmp[1:])
+			for len(remaining) > 0 {
+				rn, _, tail, err := strconv.UnquoteChar(remaining, '"')
+				if err != nil {
+					if len(remaining) > 2 && remaining[0] == '"' && remaining[1] == ':' && remaining[2] == ' ' {
+						working.Filename = sb.String()
+						wr.tmp = []byte(remaining[3:])
+						break
+					}
+					return l, fmt.Errorf("unexpected tail %s", string(remaining))
+				}
+				_, _ = sb.WriteRune(rn)
+				remaining = tail
+			}
+		} else {
+			idx := bytes.IndexByte(wr.tmp, ':')
+			if idx < 0 {
+				return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+			}
+			working.Filename = string(wr.tmp[:idx])
+			if len(wr.tmp) < idx+2 {
+				return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+			}
+			wr.tmp = wr.tmp[idx+2:]
+		}
+
+		idx := bytes.IndexByte(wr.tmp, ':')
+		if idx < 0 {
+			return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+		}
+
+		working.Attribute = string(wr.tmp[:idx])
+		if len(wr.tmp) < idx+2 {
+			return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+		}
+
+		working.Value = string(wr.tmp[idx+2:])
+
+		wr.attributes <- working
+		wr.tmp = wr.tmp[:0]
+		if len(p) > nlIdx+1 {
+			p = p[nlIdx+1:]
+			nlIdx = bytes.IndexByte(p, '\n')
+			continue
+		} else {
+			return l, nil
+		}
+	}
+
+	wr.tmp = append(wr.tmp, p...)
+	return l, nil
+}
+
+func (wr *lineSeparatedAttributeWriter) ReadAttribute() <-chan attributeTriple {
+	return wr.attributes
+}
+
+func (wr *lineSeparatedAttributeWriter) Close() error {
+	close(wr.attributes)
+	return nil
+}

modules/git/repo_attribute_test.go

@@ -0,0 +1,159 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package git
+
+import (
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func Test_nulSeparatedAttributeWriter_ReadAttribute(t *testing.T) {
+	wr := &nulSeparatedAttributeWriter{
+		attributes: make(chan attributeTriple, 5),
+	}
+
+	testStr := ".gitignore\"\n\x00linguist-vendored\x00unspecified\x00"
+
+	n, err := wr.Write([]byte(testStr))
+
+	assert.Equal(t, n, len(testStr))
+	assert.NoError(t, err)
+	select {
+	case attr := <-wr.ReadAttribute():
+		assert.Equal(t, ".gitignore\"\n", attr.Filename)
+		assert.Equal(t, "linguist-vendored", attr.Attribute)
+		assert.Equal(t, "unspecified", attr.Value)
+	case <-time.After(100 * time.Millisecond):
+		assert.Fail(t, "took too long to read an attribute from the list")
+	}
+	// Write a second attribute again
+	n, err = wr.Write([]byte(testStr))
+
+	assert.Equal(t, n, len(testStr))
+	assert.NoError(t, err)
+
+	select {
+	case attr := <-wr.ReadAttribute():
+		assert.Equal(t, ".gitignore\"\n", attr.Filename)
+		assert.Equal(t, "linguist-vendored", attr.Attribute)
+		assert.Equal(t, "unspecified", attr.Value)
+	case <-time.After(100 * time.Millisecond):
+		assert.Fail(t, "took too long to read an attribute from the list")
+	}
+
+	//Write a partial attribute
+	_, err = wr.Write([]byte("incomplete-file"))
+	assert.NoError(t, err)
+	_, err = wr.Write([]byte("name\x00"))
+	assert.NoError(t, err)
+
+	select {
+	case <-wr.ReadAttribute():
+		assert.Fail(t, "There should not be an attribute ready to read")
+	case <-time.After(100 * time.Millisecond):
+	}
+	_, err = wr.Write([]byte("attribute\x00"))
+	assert.NoError(t, err)
+	select {
+	case <-wr.ReadAttribute():
+		assert.Fail(t, "There should not be an attribute ready to read")
+	case <-time.After(100 * time.Millisecond):
+	}
+
+	_, err = wr.Write([]byte("value\x00"))
+	assert.NoError(t, err)
+
+	attr := <-wr.ReadAttribute()
+	assert.Equal(t, "incomplete-filename", attr.Filename)
+	assert.Equal(t, "attribute", attr.Attribute)
+	assert.Equal(t, "value", attr.Value)
+
+	_, err = wr.Write([]byte("shouldbe.vendor\x00linguist-vendored\x00set\x00shouldbe.vendor\x00linguist-generated\x00unspecified\x00shouldbe.vendor\x00linguist-language\x00unspecified\x00"))
+	assert.NoError(t, err)
+	attr = <-wr.ReadAttribute()
+	assert.NoError(t, err)
+	assert.EqualValues(t, attributeTriple{
+		Filename:  "shouldbe.vendor",
+		Attribute: "linguist-vendored",
+		Value:     "set",
+	}, attr)
+	attr = <-wr.ReadAttribute()
+	assert.NoError(t, err)
+	assert.EqualValues(t, attributeTriple{
+		Filename:  "shouldbe.vendor",
+		Attribute: "linguist-generated",
+		Value:     "unspecified",
+	}, attr)
+	attr = <-wr.ReadAttribute()
+	assert.NoError(t, err)
+	assert.EqualValues(t, attributeTriple{
+		Filename:  "shouldbe.vendor",
+		Attribute: "linguist-language",
+		Value:     "unspecified",
+	}, attr)
+}
+
+func Test_lineSeparatedAttributeWriter_ReadAttribute(t *testing.T) {
+	wr := &lineSeparatedAttributeWriter{
+		attributes: make(chan attributeTriple, 5),
+	}
+
+	testStr := `".gitignore\"\n": linguist-vendored: unspecified
+`
+	n, err := wr.Write([]byte(testStr))
+
+	assert.Equal(t, n, len(testStr))
+	assert.NoError(t, err)
+
+	select {
+	case attr := <-wr.ReadAttribute():
+		assert.Equal(t, ".gitignore\"\n", attr.Filename)
+		assert.Equal(t, "linguist-vendored", attr.Attribute)
+		assert.Equal(t, "unspecified", attr.Value)
+	case <-time.After(100 * time.Millisecond):
+		assert.Fail(t, "took too long to read an attribute from the list")
+	}
+
+	// Write a second attribute again
+	n, err = wr.Write([]byte(testStr))
+
+	assert.Equal(t, n, len(testStr))
+	assert.NoError(t, err)
+
+	select {
+	case attr := <-wr.ReadAttribute():
+		assert.Equal(t, ".gitignore\"\n", attr.Filename)
+		assert.Equal(t, "linguist-vendored", attr.Attribute)
+		assert.Equal(t, "unspecified", attr.Value)
+	case <-time.After(100 * time.Millisecond):
+		assert.Fail(t, "took too long to read an attribute from the list")
+	}
+
+	//Write a partial attribute
+	_, err = wr.Write([]byte("incomplete-file"))
+	assert.NoError(t, err)
+	_, err = wr.Write([]byte("name: "))
+	assert.NoError(t, err)
+	select {
+	case <-wr.ReadAttribute():
+		assert.Fail(t, "There should not be an attribute ready to read")
+	case <-time.After(100 * time.Millisecond):
+	}
+	_, err = wr.Write([]byte("attribute: "))
+	assert.NoError(t, err)
+	select {
+	case <-wr.ReadAttribute():
+		assert.Fail(t, "There should not be an attribute ready to read")
+	case <-time.After(100 * time.Millisecond):
+	}
+	_, err = wr.Write([]byte("value\n"))
+	assert.NoError(t, err)
+	attr := <-wr.ReadAttribute()
+	assert.Equal(t, "incomplete-filename", attr.Filename)
+	assert.Equal(t, "attribute", attr.Attribute)
+	assert.Equal(t, "value", attr.Value)
+}

modules/git/repo_index.go

@@ -6,11 +6,17 @@ package git
 
 import (
 	"bytes"
+	"context"
+	"io/ioutil"
+	"os"
 	"strings"
+
+	"code.gitea.io/gitea/modules/log"
+	"code.gitea.io/gitea/modules/util"
 )
 
 // ReadTreeToIndex reads a treeish to the index
-func (repo *Repository) ReadTreeToIndex(treeish string) error {
+func (repo *Repository) ReadTreeToIndex(treeish string, indexFilename ...string) error {
 	if len(treeish) != 40 {
 		res, err := NewCommand("rev-parse", "--verify", treeish).RunInDir(repo.Path)
 		if err != nil {
@@ -24,17 +30,42 @@ func (repo *Repository) ReadTreeToIndex(treeish string) error {
 	if err != nil {
 		return err
 	}
-	return repo.readTreeToIndex(id)
+	return repo.readTreeToIndex(id, indexFilename...)
 }
 
-func (repo *Repository) readTreeToIndex(id SHA1) error {
+func (repo *Repository) readTreeToIndex(id SHA1, indexFilename ...string) error {
-	_, err := NewCommand("read-tree", id.String()).RunInDir(repo.Path)
+	var env []string
+	if len(indexFilename) > 0 {
+		env = append(os.Environ(), "GIT_INDEX_FILE="+indexFilename[0])
+	}
+	_, err := NewCommand("read-tree", id.String()).RunInDirWithEnv(repo.Path, env)
 	if err != nil {
 		return err
 	}
 	return nil
 }
 
+// ReadTreeToTemporaryIndex reads a treeish to a temporary index file
+func (repo *Repository) ReadTreeToTemporaryIndex(treeish string) (filename string, cancel context.CancelFunc, err error) {
+	tmpIndex, err := ioutil.TempFile("", "index")
+	if err != nil {
+		return
+	}
+	filename = tmpIndex.Name()
+	cancel = func() {
+		err := util.Remove(filename)
+		if err != nil {
+			log.Error("failed to remove tmp index file: %v", err)
+		}
+	}
+	err = repo.ReadTreeToIndex(treeish, filename)
+	if err != nil {
+		defer cancel()
+		return "", func() {}, err
+	}
+	return
+}
+
 // EmptyIndex empties the index
 func (repo *Repository) EmptyIndex() error {
 	_, err := NewCommand("read-tree", "--empty").RunInDir(repo.Path)

modules/git/repo_language_stats_gogit.go

@@ -9,10 +9,12 @@ package git
 
 import (
 	"bytes"
+	"context"
 	"io"
 	"io/ioutil"
 
 	"code.gitea.io/gitea/modules/analyze"
+	"code.gitea.io/gitea/modules/log"
 
 	"github.com/go-enry/go-enry/v2"
 	"github.com/go-git/go-git/v5"
@@ -42,9 +44,73 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 		return nil, err
 	}
 
+	var checker *CheckAttributeReader
+
+	if CheckGitVersionAtLeast("1.7.8") == nil {
+		indexFilename, deleteTemporaryFile, err := repo.ReadTreeToTemporaryIndex(commitID)
+		if err == nil {
+			defer deleteTemporaryFile()
+
+			checker = &CheckAttributeReader{
+				Attributes: []string{"linguist-vendored", "linguist-generated", "linguist-language"},
+				Repo:       repo,
+				IndexFile:  indexFilename,
+			}
+			ctx, cancel := context.WithCancel(DefaultContext)
+			if err := checker.Init(ctx); err != nil {
+				log.Error("Unable to open checker for %s. Error: %v", commitID, err)
+			} else {
+				go func() {
+					err = checker.Run()
+					if err != nil {
+						log.Error("Unable to open checker for %s. Error: %v", commitID, err)
+						cancel()
+					}
+				}()
+			}
+			defer cancel()
+		}
+	}
+
 	sizes := make(map[string]int64)
 	err = tree.Files().ForEach(func(f *object.File) error {
-		if f.Size == 0 || analyze.IsVendor(f.Name) || enry.IsDotFile(f.Name) ||
+		if f.Size == 0 {
+			return nil
+		}
+
+		notVendored := false
+		notGenerated := false
+
+		if checker != nil {
+			attrs, err := checker.CheckPath(f.Name)
+			if err == nil {
+				if vendored, has := attrs["linguist-vendored"]; has {
+					if vendored == "set" || vendored == "true" {
+						return nil
+					}
+					notVendored = vendored == "false"
+				}
+				if generated, has := attrs["linguist-generated"]; has {
+					if generated == "set" || generated == "true" {
+						return nil
+					}
+					notGenerated = generated == "false"
+				}
+				if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
+					// group languages, such as Pug -> HTML; SCSS -> CSS
+					group := enry.GetLanguageGroup(language)
+					if len(group) == 0 {
+						language = group
+					}
+
+					sizes[language] += f.Size
+
+					return nil
+				}
+			}
+		}
+
+		if (!notVendored && analyze.IsVendor(f.Name)) || enry.IsDotFile(f.Name) ||
 			enry.IsDocumentation(f.Name) || enry.IsConfiguration(f.Name) {
 			return nil
 		}
@@ -54,7 +120,7 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 		if f.Size <= bigFileSize {
 			content, _ = readFile(f, fileSizeLimit)
 		}
-		if enry.IsGenerated(f.Name, content) {
+		if !notGenerated && enry.IsGenerated(f.Name, content) {
 			return nil
 		}
 

modules/git/repo_language_stats_nogogit.go

@@ -10,6 +10,7 @@ package git
 import (
 	"bufio"
 	"bytes"
+	"context"
 	"io"
 	"math"
 
@@ -62,13 +63,78 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 		return nil, err
 	}
 
+	var checker *CheckAttributeReader
+
+	if CheckGitVersionAtLeast("1.7.8") == nil {
+		indexFilename, deleteTemporaryFile, err := repo.ReadTreeToTemporaryIndex(commitID)
+		if err == nil {
+			defer deleteTemporaryFile()
+
+			checker = &CheckAttributeReader{
+				Attributes: []string{"linguist-vendored", "linguist-generated", "linguist-language"},
+				Repo:       repo,
+				IndexFile:  indexFilename,
+			}
+			ctx, cancel := context.WithCancel(DefaultContext)
+			if err := checker.Init(ctx); err != nil {
+				log.Error("Unable to open checker for %s. Error: %v", commitID, err)
+			} else {
+				go func() {
+					err = checker.Run()
+					if err != nil {
+						log.Error("Unable to open checker for %s. Error: %v", commitID, err)
+						cancel()
+					}
+				}()
+			}
+			defer cancel()
+		}
+	}
+
 	contentBuf := bytes.Buffer{}
 	var content []byte
 	sizes := make(map[string]int64)
 	for _, f := range entries {
 		contentBuf.Reset()
 		content = contentBuf.Bytes()
-		if f.Size() == 0 || analyze.IsVendor(f.Name()) || enry.IsDotFile(f.Name()) ||
+
+		if f.Size() == 0 {
+			continue
+		}
+
+		notVendored := false
+		notGenerated := false
+
+		if checker != nil {
+			attrs, err := checker.CheckPath(f.Name())
+			if err == nil {
+				if vendored, has := attrs["linguist-vendored"]; has {
+					if vendored == "set" || vendored == "true" {
+						continue
+					}
+					notVendored = vendored == "false"
+				}
+				if generated, has := attrs["linguist-generated"]; has {
+					if generated == "set" || generated == "true" {
+						continue
+					}
+					notGenerated = generated == "false"
+				}
+				if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
+					// group languages, such as Pug -> HTML; SCSS -> CSS
+					group := enry.GetLanguageGroup(language)
+					if len(group) == 0 {
+						language = group
+					}
+
+					sizes[language] += f.Size()
+
+					continue
+				}
+			}
+		}
+
+		if (!notVendored && analyze.IsVendor(f.Name())) || enry.IsDotFile(f.Name()) ||
 			enry.IsDocumentation(f.Name()) || enry.IsConfiguration(f.Name()) {
 			continue
 		}
@@ -102,11 +168,10 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 				return nil, err
 			}
 		}
-		if enry.IsGenerated(f.Name(), content) {
+		if !notGenerated && enry.IsGenerated(f.Name(), content) {
 			continue
 		}
 
-		// TODO: Use .gitattributes file for linguist overrides
 		// FIXME: Why can't we split this and the IsGenerated tests to avoid reading the blob unless absolutely necessary?
 		// - eg. do the all the detection tests using filename first before reading content.
 		language := analyze.GetCodeLanguage(f.Name(), content)

options/locale/locale_en-US.ini

@@ -2011,6 +2011,8 @@ diff.file_byte_size = Size
 diff.file_suppressed = File diff suppressed because it is too large
 diff.file_suppressed_line_too_long = File diff suppressed because one or more lines are too long
 diff.too_many_files = Some files were not shown because too many files changed in this diff
+diff.generated = generated
+diff.vendored = vendored
 diff.comment.placeholder = Leave a comment
 diff.comment.markdown_info = Styling with markdown is supported.
 diff.comment.add_single_comment = Add single comment

services/gitdiff/gitdiff.go

@@ -23,6 +23,7 @@ import (
 	"time"
 
 	"code.gitea.io/gitea/models"
+	"code.gitea.io/gitea/modules/analyze"
 	"code.gitea.io/gitea/modules/charset"
 	"code.gitea.io/gitea/modules/git"
 	"code.gitea.io/gitea/modules/highlight"
@@ -30,6 +31,7 @@ import (
 	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/process"
 	"code.gitea.io/gitea/modules/setting"
+	"code.gitea.io/gitea/modules/util"
 
 	"github.com/sergi/go-diff/diffmatchpatch"
 	stdcharset "golang.org/x/net/html/charset"
@@ -593,6 +595,8 @@ type DiffFile struct {
 	IsIncomplete            bool
 	IsIncompleteLineTooLong bool
 	IsProtected             bool
+	IsGenerated             bool
+	IsVendored              bool
 }
 
 // GetType returns type of diff file.
@@ -1268,7 +1272,81 @@ func GetDiffRangeWithWhitespaceBehavior(gitRepo *git.Repository, beforeCommitID,
 	if err != nil {
 		return nil, fmt.Errorf("ParsePatch: %v", err)
 	}
+
+	var checker *git.CheckAttributeReader
+
+	if git.CheckGitVersionAtLeast("1.7.8") == nil {
+		indexFilename, deleteTemporaryFile, err := gitRepo.ReadTreeToTemporaryIndex(afterCommitID)
+		if err == nil {
+			defer deleteTemporaryFile()
+			workdir, err := ioutil.TempDir("", "empty-work-dir")
+			if err != nil {
+				log.Error("Unable to create temporary directory: %v", err)
+				return nil, err
+			}
+			defer func() {
+				_ = util.RemoveAll(workdir)
+			}()
+
+			checker = &git.CheckAttributeReader{
+				Attributes: []string{"linguist-vendored", "linguist-generated"},
+				Repo:       gitRepo,
+				IndexFile:  indexFilename,
+				WorkTree:   workdir,
+			}
+			ctx, cancel := context.WithCancel(git.DefaultContext)
+			if err := checker.Init(ctx); err != nil {
+				log.Error("Unable to open checker for %s. Error: %v", afterCommitID, err)
+			} else {
+				go func() {
+					err = checker.Run()
+					if err != nil && err != ctx.Err() {
+						log.Error("Unable to open checker for %s. Error: %v", afterCommitID, err)
+					}
+					cancel()
+				}()
+			}
+			defer func() {
+				cancel()
+			}()
+		}
+	}
+
 	for _, diffFile := range diff.Files {
+
+		gotVendor := false
+		gotGenerated := false
+		if checker != nil {
+			attrs, err := checker.CheckPath(diffFile.Name)
+			if err == nil {
+				if vendored, has := attrs["linguist-vendored"]; has {
+					if vendored == "set" || vendored == "true" {
+						diffFile.IsVendored = true
+						gotVendor = true
+					} else {
+						gotVendor = vendored == "false"
+					}
+				}
+				if generated, has := attrs["linguist-generated"]; has {
+					if generated == "set" || generated == "true" {
+						diffFile.IsGenerated = true
+						gotGenerated = true
+					} else {
+						gotGenerated = generated == "false"
+					}
+				}
+			} else {
+				log.Error("Unexpected error: %v", err)
+			}
+		}
+
+		if !gotVendor {
+			diffFile.IsVendored = analyze.IsVendor(diffFile.Name)
+		}
+		if !gotGenerated {
+			diffFile.IsGenerated = analyze.IsGenerated(diffFile.Name)
+		}
+
 		tailSection := diffFile.GetTailSection(gitRepo, beforeCommitID, afterCommitID)
 		if tailSection != nil {
 			diffFile.Sections = append(diffFile.Sections, tailSection)

templates/repo/diff/box.tmpl

@@ -49,11 +49,15 @@
 			{{$isImage := or (call $.IsBlobAnImage $blobBase) (call $.IsBlobAnImage $blobHead)}}
 			{{$isCsv := (call $.IsCsvFile $file)}}
 			{{$showFileViewToggle := or $isImage (and (not $file.IsIncomplete) $isCsv)}}
-			<div class="diff-file-box diff-box file-content {{TabSizeClass $.Editorconfig $file.Name}} mt-3" id="diff-{{.Index}}">
+			<div class="diff-file-box diff-box file-content {{TabSizeClass $.Editorconfig $file.Name}} mt-3" id="diff-{{.Index}}" {{if $file.IsGenerated}}data-folded="true"{{end}}>
 				<h4 class="diff-file-header sticky-2nd-row ui top attached normal header df ac sb">
-					<div class="df ac">
+					<div class="fold-file df ac">
-						<a role="button" class="fold-file muted mr-2">
+						<a role="button" class="chevron muted mr-2">
+							{{if $file.IsGenerated}}
+								{{svg "octicon-chevron-right" 18}}
+							{{else}}
 								{{svg "octicon-chevron-down" 18}}
+							{{end}}
 						</a>
 						<div class="bold df ac">
 							{{if $file.IsBin}}
@@ -65,6 +69,12 @@
 							{{end}}
 						</div>
 						<span class="file mono">{{if $file.IsRenamed}}{{$file.OldName}} &rarr; {{end}}{{$file.Name}}{{if .IsLFSFile}} ({{$.i18n.Tr "repo.stored_lfs"}}){{end}}</span>
+						{{if $file.IsGenerated}}
+							<span class="ui label ml-3">{{$.i18n.Tr "repo.diff.generated"}}</span>
+						{{end}}
+						{{if $file.IsVendored}}
+							<span class="ui label ml-3">{{$.i18n.Tr "repo.diff.vendored"}}</span>
+						{{end}}
 					</div>
 					<div class="diff-file-header-actions df ac">
 						{{if $showFileViewToggle}}

web_src/js/index.js

@@ -2349,8 +2349,9 @@ function initCodeView() {
   }
   $(document).on('click', '.fold-file', ({currentTarget}) => {
     const box = currentTarget.closest('.file-content');
+    const chevron = currentTarget.querySelector('a.chevron');
     const folded = box.dataset.folded !== 'true';
-    currentTarget.innerHTML = svg(`octicon-chevron-${folded ? 'right' : 'down'}`, 18);
+    chevron.innerHTML = svg(`octicon-chevron-${folded ? 'right' : 'down'}`, 18);
     box.dataset.folded = String(folded);
   });
   $(document).on('click', '.blob-excerpt', async ({currentTarget}) => {