<?php

namespace Laminas\Json;

use Laminas\Json\Exception\InvalidArgumentException;
use Laminas\Json\Exception\RuntimeException;
use stdClass;

use function chr;
use function floatval;
use function function_exists;
use function hexdec;
use function in_array;
use function intval;
use function is_numeric;
use function is_string;
use function mb_convert_encoding;
use function ord;
use function preg_match;
use function sprintf;
use function str_replace;
use function strlen;
use function strpos;
use function substr;

use const PREG_OFFSET_CAPTURE;

/**
 * Decode JSON encoded string to PHP variable constructs
 */
class Decoder
{
    /**
     * Parse tokens used to decode the JSON object. These are not
     * for public consumption, they are just used internally to the
     * class.
     */
    public const EOF      = 0;
    public const DATUM    = 1;
    public const LBRACE   = 2;
    public const LBRACKET = 3;
    public const RBRACE   = 4;
    public const RBRACKET = 5;
    public const COMMA    = 6;
    public const COLON    = 7;

    /**
     * Use to maintain a "pointer" to the source being decoded
     *
     * @var string
     */
    protected $source;

    /**
     * Caches the source length
     *
     * @var int
     */
    protected $sourceLength;

    /**
     * The offset within the source being decoded
     *
     * @var int
     */
    protected $offset;

    /**
     * The current token being considered in the parser cycle
     *
     * @var int
     */
    protected $token;

    /**
     * Flag indicating how objects should be decoded
     *
     * @var int
     * @access protected
     */
    protected $decodeType;

    /** @var mixed */
    protected $tokenValue;

    /**
     * Decode Unicode Characters from \u0000 ASCII syntax.
     *
     * This algorithm was originally developed for the
     * Solar Framework by Paul M. Jones
     *
     * @link   http://solarphp.com/
     * @link   https://github.com/solarphp/core/blob/master/Solar/Json.php
     *
     * @param  string $chrs
     * @return string
     */
    public static function decodeUnicodeString($chrs)
    {
        $chrs       = (string) $chrs;
        $utf8       = '';
        $strlenChrs = strlen($chrs);

        for ($i = 0; $i < $strlenChrs; $i++) {
            $ordChrsC = ord($chrs[$i]);

            switch (true) {
                case preg_match('/\\\u[0-9A-F]{4}/i', substr($chrs, $i, 6)):
                    // single, escaped unicode character
                    $utf16    = chr(hexdec(substr($chrs, $i + 2, 2)))
                           . chr(hexdec(substr($chrs, $i + 4, 2)));
                    $utf8char = self::utf162utf8($utf16);
                    $search   = ['\\', "\n", "\t", "\r", chr(0x08), chr(0x0C), '"', '\'', '/'];
                    if (in_array($utf8char, $search)) {
                        $replace  = ['\\\\', '\\n', '\\t', '\\r', '\\b', '\\f', '\\"', '\\\'', '\\/'];
                        $utf8char = str_replace($search, $replace, $utf8char);
                    }
                    $utf8 .= $utf8char;
                    $i    += 5;
                    break;
                case ($ordChrsC >= 0x20) && ($ordChrsC <= 0x7F):
                    $utf8 .= $chrs[$i];
                    break;
                case ($ordChrsC & 0xE0) === 0xC0:
                    // characters U-00000080 - U-000007FF, mask 110XXXXX
                    //see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
                    $utf8 .= substr($chrs, $i, 2);
                    ++$i;
                    break;
                case ($ordChrsC & 0xF0) === 0xE0:
                    // characters U-00000800 - U-0000FFFF, mask 1110XXXX
                    // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
                    $utf8 .= substr($chrs, $i, 3);
                    $i    += 2;
                    break;
                case ($ordChrsC & 0xF8) === 0xF0:
                    // characters U-00010000 - U-001FFFFF, mask 11110XXX
                    // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
                    $utf8 .= substr($chrs, $i, 4);
                    $i    += 3;
                    break;
                case ($ordChrsC & 0xFC) === 0xF8:
                    // characters U-00200000 - U-03FFFFFF, mask 111110XX
                    // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
                    $utf8 .= substr($chrs, $i, 5);
                    $i    += 4;
                    break;
                case ($ordChrsC & 0xFE) === 0xFC:
                    // characters U-04000000 - U-7FFFFFFF, mask 1111110X
                    // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
                    $utf8 .= substr($chrs, $i, 6);
                    $i    += 5;
                    break;
            }
        }

        return $utf8;
    }

    /**
     * Constructor
     *
     * @param string $source     String source to decode
     * @param int    $decodeType How objects should be decoded -- see
     *     {@link Json::TYPE_ARRAY} and {@link Json::TYPE_OBJECT} for * valid
     *     values
     * @throws InvalidArgumentException
     */
    protected function __construct($source, $decodeType)
    {
        // Set defaults
        $this->source       = self::decodeUnicodeString($source);
        $this->sourceLength = strlen($this->source);
        $this->token        = self::EOF;
        $this->offset       = 0;

        $this->decodeType = match ($decodeType) {
            Json::TYPE_ARRAY, Json::TYPE_OBJECT => $decodeType,
            default => throw new InvalidArgumentException(sprintf(
                'Unknown decode type "%s", please use one of the Json::TYPE_* constants',
                $decodeType
            )),
        };

        // Set pointer at first token
        $this->getNextToken();
    }

    /**
     * Decode a JSON source string.
     *
     * Decodes a JSON encoded string; the value returned will be one of the
     * following:
     *
     * - integer
     * - float
     * - boolean
     * - null
     * - stdClass
     * - array
     * - array of one or more of the above types
     *
     * By default, decoded objects will be returned as a stdClass object;
     * to return associative arrays instead, pass {@link Json::TYPE_ARRAY}
     * to the $objectDecodeType parameter.
     *
     * @param string $source String to be decoded.
     * @param int $objectDecodeType How objects should be decoded; should be
     *     either or {@link Json::TYPE_ARRAY} or {@link Json::TYPE_OBJECT};
     *     defaults to Json::TYPE_OBJECT.
     * @return mixed
     */
    public static function decode($source, $objectDecodeType = Json::TYPE_OBJECT)
    {
        $decoder = new static($source, $objectDecodeType);
        return $decoder->decodeValue();
    }

    /**
     * Recursive routine for supported toplevel types.
     *
     * @return mixed
     */
    protected function decodeValue()
    {
        switch ($this->token) {
            case self::DATUM:
                $result = $this->tokenValue;
                $this->getNextToken();
                return $result;
            case self::LBRACE:
                return $this->decodeObject();
            case self::LBRACKET:
                return $this->decodeArray();
            default:
                return;
        }
    }

    /**
     * Decodes an object of the form { "attribute: value, "attribute2" : value, ... }
     *
     * If Laminas\Json\Encoder was used to encode the original object, then
     * a special attribute called __className will specify a class
     * name with which to wrap the data contained within the encoded source.
     *
     * Decodes to either an array or stdClass object, based on the value of
     * {@link $decodeType}. If invalid $decodeType present, returns as an
     * array.
     *
     * @return array|stdClass
     * @throws RuntimeException
     */
    protected function decodeObject()
    {
        $members = [];
        $tok     = $this->getNextToken();

        while ($tok && $tok !== self::RBRACE) {
            if ($tok !== self::DATUM || ! is_string($this->tokenValue)) {
                throw new RuntimeException(sprintf('Missing key in object encoding: %s', $this->source));
            }

            $key = $this->tokenValue;
            $tok = $this->getNextToken();

            if ($tok !== self::COLON) {
                throw new RuntimeException(sprintf('Missing ":" in object encoding: %s', $this->source));
            }

            $this->getNextToken();
            $members[$key] = $this->decodeValue();
            $tok           = $this->token;

            if ($tok === self::RBRACE) {
                break;
            }

            if ($tok !== self::COMMA) {
                throw new RuntimeException(sprintf('Missing "," in object encoding: %s', $this->source));
            }

            $tok = $this->getNextToken();
        }

        switch ($this->decodeType) {
            case Json::TYPE_OBJECT:
                // Create new stdClass and populate with $members
                $result = new stdClass();
                foreach ($members as $key => $value) {
                    if ($key === '') {
                        $key = '_empty_';
                    }
                    $result->$key = $value;
                }
                break;
            case Json::TYPE_ARRAY:
                // intentionally fall-through
            default:
                $result = $members;
                break;
        }

        $this->getNextToken();
        return $result;
    }

    /**
     * Decodes the JSON array format [element, element2, ..., elementN]
     *
     * @return array
     * @throws RuntimeException
     */
    protected function decodeArray()
    {
        $result = [];
        $tok    = $this->getNextToken(); // Move past the '['
        $index  = 0;

        while ($tok && $tok !== self::RBRACKET) {
            $result[$index++] = $this->decodeValue();

            $tok = $this->token;

            if ($tok === self::RBRACKET || ! $tok) {
                break;
            }

            if ($tok !== self::COMMA) {
                throw new RuntimeException(sprintf('Missing "," in array encoding: %s', $this->source));
            }

            $tok = $this->getNextToken();
        }

        $this->getNextToken();
        return $result;
    }

    /**
     * Removes whitespace characters from the source input.
     */
    protected function eatWhitespace()
    {
        if (
            preg_match('/([\t\b\f\n\r ])*/s', $this->source, $matches, PREG_OFFSET_CAPTURE, $this->offset)
            && $matches[0][1] === $this->offset
        ) {
            $this->offset += strlen($matches[0][0]);
        }
    }

    /**
     * Retrieves the next token from the source stream.
     *
     * @return int Token constant value specified in class definition.
     * @throws RuntimeException
     */
    protected function getNextToken()
    {
        $this->token      = self::EOF;
        $this->tokenValue = null;
        $this->eatWhitespace();

        if ($this->offset >= $this->sourceLength) {
            return self::EOF;
        }

        $str       = $this->source;
        $strLength = $this->sourceLength;
        $i         = $this->offset;
        $start     = $i;

        switch ($str[$i]) {
            case '{':
                $this->token = self::LBRACE;
                break;
            case '}':
                $this->token = self::RBRACE;
                break;
            case '[':
                $this->token = self::LBRACKET;
                break;
            case ']':
                $this->token = self::RBRACKET;
                break;
            case ',':
                $this->token = self::COMMA;
                break;
            case ':':
                $this->token = self::COLON;
                break;
            case '"':
                $result = '';
                do {
                    $i++;
                    if ($i >= $strLength) {
                        break;
                    }

                    $chr = $str[$i];

                    if ($chr === '"') {
                        break;
                    }

                    if ($chr !== '\\') {
                        $result .= $chr;
                        continue;
                    }

                    $i++;

                    if ($i >= $strLength) {
                        break;
                    }

                    $chr = $str[$i];
                    match ($chr) {
                        '"' => $result  .= '"',
                        '\\' => $result .= '\\',
                        '/' => $result  .= '/',
                        'b' => $result  .= "\x08",
                        'f' => $result  .= "\x0c",
                        'n' => $result  .= "\x0a",
                        'r' => $result  .= "\x0d",
                        't' => $result  .= "\x09",
                        '\'' => $result .= '\'',
                        default => throw new RuntimeException(sprintf('Illegal escape sequence "%s"', $chr)),
                    };
                } while ($i < $strLength);

                $this->token      = self::DATUM;
                $this->tokenValue = $result;
                break;
            case 't':
                if (($i + 3) < $strLength && $start === strpos($str, "true", $start)) {
                    $this->token = self::DATUM;
                }
                $this->tokenValue = true;
                $i               += 3;
                break;
            case 'f':
                if (($i + 4) < $strLength && $start === strpos($str, "false", $start)) {
                    $this->token = self::DATUM;
                }
                $this->tokenValue = false;
                $i               += 4;
                break;
            case 'n':
                if (($i + 3) < $strLength && $start === strpos($str, "null", $start)) {
                    $this->token = self::DATUM;
                }
                $this->tokenValue = null;
                $i               += 3;
                break;
        }

        if ($this->token !== self::EOF) {
            $this->offset = $i + 1; // Consume the last token character
            return $this->token;
        }

        $chr = $str[$i];

        if ($chr !== '-' && $chr !== '.' && ($chr < '0' || $chr > '9')) {
            throw new RuntimeException('Illegal Token');
        }

        if (
            preg_match('/-?([0-9])*(\.[0-9]*)?((e|E)((-|\+)?)[0-9]+)?/s', $str, $matches, PREG_OFFSET_CAPTURE, $start)
            && $matches[0][1] === $start
        ) {
            $datum = $matches[0][0];

            if (! is_numeric($datum)) {
                throw new RuntimeException(sprintf('Illegal number format: %s', $datum));
            }

            if (preg_match('/^0\d+$/', $datum)) {
                throw new RuntimeException(sprintf('Octal notation not supported by JSON (value: %o)', $datum));
            }

            $val  = intval($datum);
            $fVal = floatval($datum);

            // phpcs:ignore SlevomatCodingStandard.Operators.DisallowEqualOperators.DisallowedEqualOperator
            $this->tokenValue = $val == $fVal ? $val : $fVal;
            $this->token      = self::DATUM;
            $this->offset     = $start + strlen($datum);
        }

        return $this->token;
    }

    /**
     * Convert a string from one UTF-16 char to one UTF-8 char.
     *
     * Normally should be handled by mb_convert_encoding, but provides a slower
     * PHP-only method for installations that lack the multibyte string
     * extension.
     *
     * This method is from the Solar Framework by Paul M. Jones.
     *
     * @link   http://solarphp.com
     *
     * @param  string $utf16 UTF-16 character
     * @return string UTF-8 character
     */
    protected static function utf162utf8($utf16)
    {
        // Check for mb extension otherwise do by hand.
        if (function_exists('mb_convert_encoding')) {
            return mb_convert_encoding($utf16, 'UTF-8', 'UTF-16');
        }

        $bytes = (ord($utf16[0]) << 8) | ord($utf16[1]);
        return match (true) {
            // This case should never be reached, because we are in ASCII range;
            // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
            (0x7F & $bytes) === $bytes => chr(0x7F & $bytes),

            // Return a 2-byte UTF-8 character;
            // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
            (0x07FF & $bytes) === $bytes => chr(0xC0 | (($bytes >> 6) & 0x1F))
                . chr(0x80 | ($bytes & 0x3F)),

            // Return a 3-byte UTF-8 character;
            // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
            (0xFFFF & $bytes) === $bytes => chr(0xE0 | (($bytes >> 12) & 0x0F))
                . chr(0x80 | (($bytes >> 6) & 0x3F))
                . chr(0x80 | ($bytes & 0x3F)),

            // ignoring UTF-32 for now, sorry
            default => '',
        };
    }
}