sow-tools/internal/topdata/editorconfig_format.go

package topdata

import (
	"bufio"
	"bytes"
	"fmt"
	"os"
	"path/filepath"
	"regexp"
	"strconv"
	"strings"
)

type jsonFileFormatting struct {
	Indent       string
	LineEnding   string
	FinalNewline bool
}

type editorConfigDocument struct {
	Dir      string
	Root     bool
	Sections []editorConfigSection
}

type editorConfigSection struct {
	Pattern    string
	Properties map[string]string
}

func defaultJSONFileFormatting() jsonFileFormatting {
	return jsonFileFormatting{
		Indent:       "  ",
		LineEnding:   "\n",
		FinalNewline: true,
	}
}

func resolveJSONFileFormatting(path string) (jsonFileFormatting, error) {
	formatting := defaultJSONFileFormatting()
	documents, err := editorConfigDocumentsForPath(path)
	if err != nil {
		return jsonFileFormatting{}, err
	}
	for _, document := range documents {
		relative, err := filepath.Rel(document.Dir, path)
		if err != nil {
			return jsonFileFormatting{}, fmt.Errorf("resolve editorconfig path for %s relative to %s: %w", path, document.Dir, err)
		}
		relative = filepath.ToSlash(relative)
		if strings.HasPrefix(relative, "../") || relative == ".." {
			continue
		}
		for _, section := range document.Sections {
			matches, err := editorConfigPatternMatches(section.Pattern, relative)
			if err != nil {
				return jsonFileFormatting{}, fmt.Errorf("%s/.editorconfig section [%s]: %w", document.Dir, section.Pattern, err)
			}
			if !matches {
				continue
			}
			next, err := applyEditorConfigProperties(formatting, section.Properties)
			if err != nil {
				return jsonFileFormatting{}, fmt.Errorf("%s/.editorconfig section [%s]: %w", document.Dir, section.Pattern, err)
			}
			formatting = next
		}
	}
	return formatting, nil
}

func editorConfigDocumentsForPath(path string) ([]editorConfigDocument, error) {
	dir := filepath.Dir(path)
	documents := []editorConfigDocument{}
	for {
		configPath := filepath.Join(dir, ".editorconfig")
		raw, err := os.ReadFile(configPath)
		if err == nil {
			document, err := parseEditorConfig(configPath, raw)
			if err != nil {
				return nil, err
			}
			document.Dir = dir
			documents = append(documents, document)
			if document.Root {
				break
			}
		} else if !os.IsNotExist(err) {
			return nil, fmt.Errorf("read %s: %w", configPath, err)
		}

		parent := filepath.Dir(dir)
		if parent == dir {
			break
		}
		dir = parent
	}
	for left, right := 0, len(documents)-1; left < right; left, right = left+1, right-1 {
		documents[left], documents[right] = documents[right], documents[left]
	}
	return documents, nil
}

func parseEditorConfig(path string, raw []byte) (editorConfigDocument, error) {
	document := editorConfigDocument{}
	scanner := bufio.NewScanner(bytes.NewReader(raw))
	var current *editorConfigSection
	for lineNumber := 1; scanner.Scan(); lineNumber++ {
		line := strings.TrimSpace(scanner.Text())
		if line == "" || strings.HasPrefix(line, "#") || strings.HasPrefix(line, ";") {
			continue
		}
		if strings.HasPrefix(line, "[") {
			end := strings.LastIndex(line, "]")
			if end < 0 {
				return editorConfigDocument{}, fmt.Errorf("%s:%d: malformed section header", path, lineNumber)
			}
			section := editorConfigSection{
				Pattern:    strings.TrimSpace(line[1:end]),
				Properties: map[string]string{},
			}
			document.Sections = append(document.Sections, section)
			current = &document.Sections[len(document.Sections)-1]
			continue
		}

		index := strings.IndexAny(line, "=:")
		if index < 0 {
			return editorConfigDocument{}, fmt.Errorf("%s:%d: expected key/value property", path, lineNumber)
		}
		key := strings.ToLower(strings.TrimSpace(line[:index]))
		value := strings.TrimSpace(line[index+1:])
		if current == nil {
			if key == "root" {
				document.Root = strings.EqualFold(value, "true")
			}
			continue
		}
		current.Properties[key] = value
	}
	if err := scanner.Err(); err != nil {
		return editorConfigDocument{}, fmt.Errorf("scan %s: %w", path, err)
	}
	return document, nil
}

func applyEditorConfigProperties(formatting jsonFileFormatting, properties map[string]string) (jsonFileFormatting, error) {
	style := strings.ToLower(strings.TrimSpace(properties["indent_style"]))
	sizeText := strings.ToLower(strings.TrimSpace(properties["indent_size"]))
	if style == "tab" {
		formatting.Indent = "\t"
	} else if style != "" && style != "space" && style != "unset" {
		return jsonFileFormatting{}, fmt.Errorf("unsupported indent_style %q", properties["indent_style"])
	}
	if sizeText != "" && sizeText != "unset" && sizeText != "tab" {
		size, err := strconv.Atoi(sizeText)
		if err != nil || size < 1 {
			return jsonFileFormatting{}, fmt.Errorf("indent_size must be a positive integer, got %q", properties["indent_size"])
		}
		if style != "tab" {
			formatting.Indent = strings.Repeat(" ", size)
		}
	}

	switch value := strings.ToLower(strings.TrimSpace(properties["end_of_line"])); value {
	case "", "unset":
	case "lf":
		formatting.LineEnding = "\n"
	case "crlf":
		formatting.LineEnding = "\r\n"
	case "cr":
		formatting.LineEnding = "\r"
	default:
		return jsonFileFormatting{}, fmt.Errorf("unsupported end_of_line %q", properties["end_of_line"])
	}

	switch value := strings.ToLower(strings.TrimSpace(properties["insert_final_newline"])); value {
	case "", "unset":
	case "true":
		formatting.FinalNewline = true
	case "false":
		formatting.FinalNewline = false
	default:
		return jsonFileFormatting{}, fmt.Errorf("insert_final_newline must be true or false, got %q", properties["insert_final_newline"])
	}
	return formatting, nil
}

func editorConfigPatternMatches(pattern, relativePath string) (bool, error) {
	pattern = filepath.ToSlash(strings.TrimSpace(pattern))
	pattern = strings.TrimPrefix(pattern, "/")
	if pattern == "" {
		return false, nil
	}
	if !strings.Contains(pattern, "/") {
		relativePath = pathBase(relativePath)
	}
	expression, err := editorConfigGlobRegexp(pattern)
	if err != nil {
		return false, err
	}
	return regexp.MatchString(expression, relativePath)
}

func editorConfigGlobRegexp(pattern string) (string, error) {
	body, err := translateEditorConfigGlobSegment(pattern)
	if err != nil {
		return "", err
	}
	expression := "^" + body + "$"
	if _, err := regexp.Compile(expression); err != nil {
		return "", err
	}
	return expression, nil
}

// translateEditorConfigGlobSegment converts an editorconfig glob into a regular
// expression body. In addition to *, **, and ?, it implements brace expansion:
// comma lists ({a,b,c}) become alternations and numeric ranges ({n..m}) expand
// to the integers in the range, matching the editorconfig specification.
func translateEditorConfigGlobSegment(pattern string) (string, error) {
	var builder strings.Builder
	for index := 0; index < len(pattern); {
		char := pattern[index]
		switch char {
		case '*':
			if index+1 < len(pattern) && pattern[index+1] == '*' {
				if index+2 < len(pattern) && pattern[index+2] == '/' {
					builder.WriteString("(?:.*/)?")
					index += 3
				} else {
					builder.WriteString(".*")
					index += 2
				}
				continue
			}
			builder.WriteString("[^/]*")
			index++
		case '?':
			builder.WriteString("[^/]")
			index++
		case '{':
			end := matchingEditorConfigBrace(pattern, index)
			if end < 0 {
				// An unbalanced brace is matched literally.
				builder.WriteString("\\{")
				index++
				continue
			}
			expansion, err := expandEditorConfigBrace(pattern[index+1 : end])
			if err != nil {
				return "", err
			}
			builder.WriteString(expansion)
			index = end + 1
		case '.', '+', '(', ')', '|', '^', '$', '}', '[', ']', '\\':
			builder.WriteByte('\\')
			builder.WriteByte(char)
			index++
		default:
			builder.WriteByte(char)
			index++
		}
	}
	return builder.String(), nil
}

// matchingEditorConfigBrace returns the index of the '}' that closes the '{' at
// open, accounting for nesting, or -1 when the braces are unbalanced.
func matchingEditorConfigBrace(pattern string, open int) int {
	depth := 0
	for index := open; index < len(pattern); index++ {
		switch pattern[index] {
		case '{':
			depth++
		case '}':
			depth--
			if depth == 0 {
				return index
			}
		}
	}
	return -1
}

// expandEditorConfigBrace turns the body of a brace expression into a regex
// fragment. A comma list becomes an alternation; a lone "n..m" body becomes a
// numeric range. A body with neither comma nor range is treated as literal
// braces, matching the editorconfig reference implementation.
func expandEditorConfigBrace(inner string) (string, error) {
	options := splitTopLevelBraceCommas(inner)
	if len(options) == 1 {
		if rangeExpr, ok, err := editorConfigNumericRange(inner); err != nil {
			return "", err
		} else if ok {
			return rangeExpr, nil
		}
		segment, err := translateEditorConfigGlobSegment(inner)
		if err != nil {
			return "", err
		}
		return "\\{" + segment + "\\}", nil
	}
	parts := make([]string, 0, len(options))
	for _, option := range options {
		segment, err := translateEditorConfigGlobSegment(option)
		if err != nil {
			return "", err
		}
		parts = append(parts, segment)
	}
	return "(?:" + strings.Join(parts, "|") + ")", nil
}

// splitTopLevelBraceCommas splits a brace body on commas that are not nested
// inside an inner brace group.
func splitTopLevelBraceCommas(inner string) []string {
	parts := []string{}
	depth := 0
	start := 0
	for index := 0; index < len(inner); index++ {
		switch inner[index] {
		case '{':
			depth++
		case '}':
			depth--
		case ',':
			if depth == 0 {
				parts = append(parts, inner[start:index])
				start = index + 1
			}
		}
	}
	return append(parts, inner[start:])
}

// editorConfigNumericRange expands "n..m" into an alternation of the integers
// in [n, m]. It returns ok=false when the body is not a numeric range.
func editorConfigNumericRange(inner string) (string, bool, error) {
	separator := strings.Index(inner, "..")
	if separator < 0 {
		return "", false, nil
	}
	low, err := strconv.Atoi(strings.TrimSpace(inner[:separator]))
	if err != nil {
		return "", false, nil
	}
	high, err := strconv.Atoi(strings.TrimSpace(inner[separator+2:]))
	if err != nil {
		return "", false, nil
	}
	if low > high {
		low, high = high, low
	}
	const maxRangeSize = 65536
	if high-low >= maxRangeSize {
		return "", false, fmt.Errorf("editorconfig numeric range {%s} is too large", inner)
	}
	parts := make([]string, 0, high-low+1)
	for value := low; value <= high; value++ {
		parts = append(parts, strconv.Itoa(value))
	}
	return "(?:" + strings.Join(parts, "|") + ")", true, nil
}

func pathBase(path string) string {
	if index := strings.LastIndex(path, "/"); index >= 0 {
		return path[index+1:]
	}
	return path
}