soft_quay/core/catalog/canonical.go

package catalog

import (
	"bytes"
	"encoding/json"
	"fmt"
	"io"
	"regexp"
	"sort"
	"unicode/utf8"
)

var integerJSONNumber = regexp.MustCompile(`^(0|[1-9][0-9]*|-[1-9][0-9]*)$`)

func parseRestrictedJSON(data []byte) (any, error) {
	if !utf8.Valid(data) {
		return nil, fmt.Errorf("%w: input is not valid UTF-8", ErrInvalidDocument)
	}
	if err := validateJSONStringSurrogates(data); err != nil {
		return nil, err
	}
	decoder := json.NewDecoder(bytes.NewReader(data))
	decoder.UseNumber()

	value, err := decodeJSONValue(decoder)
	if err != nil {
		return nil, err
	}

	if _, err := decoder.Token(); err != io.EOF {
		if err == nil {
			return nil, fmt.Errorf("%w: trailing JSON value", ErrInvalidDocument)
		}
		return nil, fmt.Errorf("%w: trailing data: %v", ErrInvalidDocument, err)
	}
	return value, nil
}

func validateJSONStringSurrogates(data []byte) error {
	for index := 0; index < len(data); index++ {
		if data[index] != '"' {
			continue
		}
		next, err := scanJSONStringSurrogates(data, index)
		if err != nil {
			return err
		}
		index = next - 1
	}
	return nil
}

func scanJSONStringSurrogates(data []byte, start int) (int, error) {
	for index := start + 1; index < len(data); index++ {
		switch data[index] {
		case '"':
			return index + 1, nil
		case '\\':
			if index+1 >= len(data) {
				return 0, fmt.Errorf("%w: incomplete string escape", ErrInvalidDocument)
			}
			if data[index+1] != 'u' {
				index++
				continue
			}
			codeUnit, ok := decodeJSONHexCodeUnit(data, index+2)
			if !ok {
				return 0, fmt.Errorf("%w: invalid unicode escape", ErrInvalidDocument)
			}
			switch {
			case codeUnit >= 0xd800 && codeUnit <= 0xdbff:
				if index+7 >= len(data) || data[index+6] != '\\' || data[index+7] != 'u' {
					return 0, fmt.Errorf("%w: high surrogate is not paired", ErrInvalidDocument)
				}
				lowSurrogate, ok := decodeJSONHexCodeUnit(data, index+8)
				if !ok || lowSurrogate < 0xdc00 || lowSurrogate > 0xdfff {
					return 0, fmt.Errorf("%w: high surrogate is not followed by a low surrogate", ErrInvalidDocument)
				}
				index += 11
			case codeUnit >= 0xdc00 && codeUnit <= 0xdfff:
				return 0, fmt.Errorf("%w: low surrogate has no high surrogate", ErrInvalidDocument)
			default:
				index += 5
			}
		}
	}
	return 0, fmt.Errorf("%w: unterminated string", ErrInvalidDocument)
}

func decodeJSONHexCodeUnit(data []byte, start int) (uint16, bool) {
	if start+4 > len(data) {
		return 0, false
	}
	var value uint16
	for _, digit := range data[start : start+4] {
		value <<= 4
		switch {
		case digit >= '0' && digit <= '9':
			value |= uint16(digit - '0')
		case digit >= 'a' && digit <= 'f':
			value |= uint16(digit-'a') + 10
		case digit >= 'A' && digit <= 'F':
			value |= uint16(digit-'A') + 10
		default:
			return 0, false
		}
	}
	return value, true
}

func decodeJSONValue(decoder *json.Decoder) (any, error) {
	token, err := decoder.Token()
	if err != nil {
		return nil, fmt.Errorf("%w: %v", ErrInvalidDocument, err)
	}

	switch value := token.(type) {
	case json.Delim:
		switch value {
		case '{':
			object := make(map[string]any)
			for decoder.More() {
				keyToken, err := decoder.Token()
				if err != nil {
					return nil, fmt.Errorf("%w: object key: %v", ErrInvalidDocument, err)
				}
				key, ok := keyToken.(string)
				if !ok {
					return nil, fmt.Errorf("%w: object key is not a string", ErrInvalidDocument)
				}
				if _, exists := object[key]; exists {
					return nil, fmt.Errorf("%w: %q", ErrDuplicateField, key)
				}
				child, err := decodeJSONValue(decoder)
				if err != nil {
					return nil, err
				}
				object[key] = child
			}
			end, err := decoder.Token()
			if err != nil || end != json.Delim('}') {
				return nil, fmt.Errorf("%w: unterminated object", ErrInvalidDocument)
			}
			return object, nil
		case '[':
			var array []any
			for decoder.More() {
				child, err := decodeJSONValue(decoder)
				if err != nil {
					return nil, err
				}
				array = append(array, child)
			}
			end, err := decoder.Token()
			if err != nil || end != json.Delim(']') {
				return nil, fmt.Errorf("%w: unterminated array", ErrInvalidDocument)
			}
			return array, nil
		default:
			return nil, fmt.Errorf("%w: unexpected delimiter %q", ErrInvalidDocument, value)
		}
	case json.Number:
		if !integerJSONNumber.MatchString(string(value)) {
			return nil, fmt.Errorf("%w: %q", ErrUnsupportedNumber, value)
		}
		return value, nil
	case string, bool, nil:
		return value, nil
	default:
		return nil, fmt.Errorf("%w: unsupported token %T", ErrInvalidDocument, token)
	}
}

func canonicalJSON(value any) ([]byte, error) {
	var buffer bytes.Buffer
	if err := appendCanonicalJSON(&buffer, value); err != nil {
		return nil, err
	}
	return buffer.Bytes(), nil
}

func appendCanonicalJSON(buffer *bytes.Buffer, value any) error {
	switch value := value.(type) {
	case nil:
		buffer.WriteString("null")
	case bool:
		if value {
			buffer.WriteString("true")
		} else {
			buffer.WriteString("false")
		}
	case string:
		encoded, err := json.Marshal(value)
		if err != nil {
			return fmt.Errorf("%w: encode string: %v", ErrInvalidDocument, err)
		}
		buffer.Write(encoded)
	case json.Number:
		if !integerJSONNumber.MatchString(string(value)) {
			return fmt.Errorf("%w: %q", ErrUnsupportedNumber, value)
		}
		buffer.WriteString(string(value))
	case []any:
		buffer.WriteByte('[')
		for index, child := range value {
			if index > 0 {
				buffer.WriteByte(',')
			}
			if err := appendCanonicalJSON(buffer, child); err != nil {
				return err
			}
		}
		buffer.WriteByte(']')
	case map[string]any:
		keys := make([]string, 0, len(value))
		for key := range value {
			keys = append(keys, key)
		}
		sort.Strings(keys)

		buffer.WriteByte('{')
		for index, key := range keys {
			if index > 0 {
				buffer.WriteByte(',')
			}
			encodedKey, err := json.Marshal(key)
			if err != nil {
				return fmt.Errorf("%w: encode key: %v", ErrInvalidDocument, err)
			}
			buffer.Write(encodedKey)
			buffer.WriteByte(':')
			if err := appendCanonicalJSON(buffer, value[key]); err != nil {
				return err
			}
		}
		buffer.WriteByte('}')
	default:
		return fmt.Errorf("%w: unsupported value %T", ErrInvalidDocument, value)
	}
	return nil
}