Move codeformat folder to tools (#35758)

Followup to https://github.com/go-gitea/gitea/pull/35734.

- Move `codeformat` folder to `tools`
- Add `tools` to `GO_DIRS`
- Move `misspellings.csv` to `assets` so we can lint the whole `tools`
directory without filter shenanigans.

---------

Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>

tools/code-batch-process.go

@@ -16,7 +16,7 @@ import (
 	"strconv"
 	"strings"
 
-	"code.gitea.io/gitea/build/codeformat"
+	"code.gitea.io/gitea/tools/codeformat"
 )
 
 // Windows has a limitation for command line arguments, the size can not exceed 32KB.

tools/codeformat/formatimports.go

@@ -0,0 +1,195 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package codeformat
+
+import (
+	"bytes"
+	"errors"
+	"io"
+	"os"
+	"sort"
+	"strings"
+)
+
+var importPackageGroupOrders = map[string]int{
+	"":                     1, // internal
+	"code.gitea.io/gitea/": 2,
+}
+
+var errInvalidCommentBetweenImports = errors.New("comments between imported packages are invalid, please move comments to the end of the package line")
+
+var (
+	importBlockBegin = []byte("\nimport (\n")
+	importBlockEnd   = []byte("\n)")
+)
+
+type importLineParsed struct {
+	group   string
+	pkg     string
+	content string
+}
+
+func parseImportLine(line string) (*importLineParsed, error) {
+	il := &importLineParsed{content: line}
+	p1 := strings.IndexRune(line, '"')
+	if p1 == -1 {
+		return nil, errors.New("invalid import line: " + line)
+	}
+	p1++
+	p := strings.IndexRune(line[p1:], '"')
+	if p == -1 {
+		return nil, errors.New("invalid import line: " + line)
+	}
+	p2 := p1 + p
+	il.pkg = line[p1:p2]
+
+	pDot := strings.IndexRune(il.pkg, '.')
+	pSlash := strings.IndexRune(il.pkg, '/')
+	if pDot != -1 && pDot < pSlash {
+		il.group = "domain-package"
+	}
+	for groupName := range importPackageGroupOrders {
+		if groupName == "" {
+			continue // skip internal
+		}
+		if strings.HasPrefix(il.pkg, groupName) {
+			il.group = groupName
+		}
+	}
+	return il, nil
+}
+
+type (
+	importLineGroup    []*importLineParsed
+	importLineGroupMap map[string]importLineGroup
+)
+
+func formatGoImports(contentBytes []byte) ([]byte, error) {
+	p1 := bytes.Index(contentBytes, importBlockBegin)
+	if p1 == -1 {
+		return nil, nil
+	}
+	p1 += len(importBlockBegin)
+	p := bytes.Index(contentBytes[p1:], importBlockEnd)
+	if p == -1 {
+		return nil, nil
+	}
+	p2 := p1 + p
+
+	importGroups := importLineGroupMap{}
+	r := bytes.NewBuffer(contentBytes[p1:p2])
+	eof := false
+	for !eof {
+		line, err := r.ReadString('\n')
+		eof = err == io.EOF
+		if err != nil && !eof {
+			return nil, err
+		}
+		line = strings.TrimSpace(line)
+		if line != "" {
+			if strings.HasPrefix(line, "//") || strings.HasPrefix(line, "/*") {
+				return nil, errInvalidCommentBetweenImports
+			}
+			importLine, err := parseImportLine(line)
+			if err != nil {
+				return nil, err
+			}
+			importGroups[importLine.group] = append(importGroups[importLine.group], importLine)
+		}
+	}
+
+	var groupNames []string
+	for groupName, importLines := range importGroups {
+		groupNames = append(groupNames, groupName)
+		sort.Slice(importLines, func(i, j int) bool {
+			return strings.Compare(importLines[i].pkg, importLines[j].pkg) < 0
+		})
+	}
+
+	sort.Slice(groupNames, func(i, j int) bool {
+		n1 := groupNames[i]
+		n2 := groupNames[j]
+		o1 := importPackageGroupOrders[n1]
+		o2 := importPackageGroupOrders[n2]
+		if o1 != 0 && o2 != 0 {
+			return o1 < o2
+		}
+		if o1 == 0 && o2 == 0 {
+			return strings.Compare(n1, n2) < 0
+		}
+		return o1 != 0
+	})
+
+	formattedBlock := bytes.Buffer{}
+	for _, groupName := range groupNames {
+		hasNormalImports := false
+		hasDummyImports := false
+		// non-dummy import comes first
+		for _, importLine := range importGroups[groupName] {
+			if strings.HasPrefix(importLine.content, "_") {
+				hasDummyImports = true
+			} else {
+				formattedBlock.WriteString("\t" + importLine.content + "\n")
+				hasNormalImports = true
+			}
+		}
+		// dummy (_ "pkg") comes later
+		if hasDummyImports {
+			if hasNormalImports {
+				formattedBlock.WriteString("\n")
+			}
+			for _, importLine := range importGroups[groupName] {
+				if strings.HasPrefix(importLine.content, "_") {
+					formattedBlock.WriteString("\t" + importLine.content + "\n")
+				}
+			}
+		}
+		formattedBlock.WriteString("\n")
+	}
+	formattedBlockBytes := bytes.TrimRight(formattedBlock.Bytes(), "\n")
+
+	var formattedBytes []byte
+	formattedBytes = append(formattedBytes, contentBytes[:p1]...)
+	formattedBytes = append(formattedBytes, formattedBlockBytes...)
+	formattedBytes = append(formattedBytes, contentBytes[p2:]...)
+	return formattedBytes, nil
+}
+
+// FormatGoImports format the imports by our rules (see unit tests)
+func FormatGoImports(file string, doWriteFile bool) error {
+	f, err := os.Open(file)
+	if err != nil {
+		return err
+	}
+	var contentBytes []byte
+	{
+		defer f.Close()
+		contentBytes, err = io.ReadAll(f)
+		if err != nil {
+			return err
+		}
+	}
+	formattedBytes, err := formatGoImports(contentBytes)
+	if err != nil {
+		return err
+	}
+	if formattedBytes == nil {
+		return nil
+	}
+	if bytes.Equal(contentBytes, formattedBytes) {
+		return nil
+	}
+
+	if doWriteFile {
+		f, err = os.OpenFile(file, os.O_TRUNC|os.O_WRONLY, 0o644)
+		if err != nil {
+			return err
+		}
+		defer f.Close()
+		_, err = f.Write(formattedBytes)
+		return err
+	}
+
+	return err
+}

tools/codeformat/formatimports_test.go

@@ -0,0 +1,124 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package codeformat
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestFormatImportsSimple(t *testing.T) {
+	formatted, err := formatGoImports([]byte(`
+package codeformat
+
+import (
+	"github.com/stretchr/testify/assert"
+	"testing"
+)
+`))
+
+	expected := `
+package codeformat
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+`
+
+	assert.NoError(t, err)
+	assert.Equal(t, expected, string(formatted))
+}
+
+func TestFormatImportsGroup(t *testing.T) {
+	// gofmt/goimports won't group the packages, for example, they produce such code:
+	//     "bytes"
+	//     "image"
+	//        (a blank line)
+	//     "fmt"
+	//     "image/color/palette"
+	// our formatter does better, and these packages are grouped into one.
+
+	formatted, err := formatGoImports([]byte(`
+package test
+
+import (
+	"bytes"
+	"fmt"
+	"image"
+	"image/color"
+
+	_ "image/gif"  // for processing gif images
+	_ "image/jpeg" // for processing jpeg images
+	_ "image/png"  // for processing png images
+
+	"code.gitea.io/other/package"
+
+	"code.gitea.io/gitea/modules/setting"
+	"code.gitea.io/gitea/modules/util"
+
+  "xorm.io/the/package"
+
+	"github.com/issue9/identicon"
+	"github.com/nfnt/resize"
+	"github.com/oliamb/cutter"
+)
+`))
+
+	expected := `
+package test
+
+import (
+	"bytes"
+	"fmt"
+	"image"
+	"image/color"
+
+	_ "image/gif"  // for processing gif images
+	_ "image/jpeg" // for processing jpeg images
+	_ "image/png"  // for processing png images
+
+	"code.gitea.io/gitea/modules/setting"
+	"code.gitea.io/gitea/modules/util"
+
+	"code.gitea.io/other/package"
+	"github.com/issue9/identicon"
+	"github.com/nfnt/resize"
+	"github.com/oliamb/cutter"
+	"xorm.io/the/package"
+)
+`
+
+	assert.NoError(t, err)
+	assert.Equal(t, expected, string(formatted))
+}
+
+func TestFormatImportsInvalidComment(t *testing.T) {
+	// why we shouldn't write comments between imports: it breaks the grouping of imports
+	// for example:
+	//    "pkg1"
+	//    "pkg2"
+	//    // a comment
+	//    "pkgA"
+	//    "pkgB"
+	// the comment splits the packages into two groups, pkg1/2 are sorted separately, pkgA/B are sorted separately
+	// we don't want such code, so the code should be:
+	//    "pkg1"
+	//    "pkg2"
+	//    "pkgA" // a comment
+	//    "pkgB"
+
+	_, err := formatGoImports([]byte(`
+package test
+
+import (
+  "image/jpeg"
+	// for processing gif images
+	"image/gif"
+)
+`))
+	assert.ErrorIs(t, err, errInvalidCommentBetweenImports)
+}

tools/misspellings.csv

@@ -1,21 +0,0 @@
-acounts,accounts
-canidate,candidate
-comfirm,confirm
-converage,coverage
-currrently,currently
-delimeter,delimiter
-differrent,different
-exclusing,excluding
-finshed,finished
-formated,formatted
-inderect,indirect
-insuficient,insufficient
-likly,likely
-mergable,mergeable
-overrided,overridden
-priortized,prioritized
-registeration,registration
-reuqest,request
-reviwer,reviewer
-superceded,superseded
-underlaying,underlying