Seldaek_monolog/src/Monolog/Formatter/NormalizerFormatter.php

<?php

/*
 * This file is part of the Monolog package.
 *
 * (c) Jordi Boggiano <j.boggiano@seld.be>
 *
 * For the full copyright and license information, please view the LICENSE
 * file that was distributed with this source code.
 */

namespace Monolog\Formatter;

use Throwable;

/**
 * Normalizes incoming records to remove objects/resources so it's easier to dump to various targets
 *
 * @author Jordi Boggiano <j.boggiano@seld.be>
 */
class NormalizerFormatter implements FormatterInterface
{
    const SIMPLE_DATE = "Y-m-d H:i:s";

    protected $dateFormat;

    /**
     * @param string $dateFormat The format of the timestamp: one supported by DateTime::format
     */
    public function __construct($dateFormat = null)
    {
        $this->dateFormat = $dateFormat ?: static::SIMPLE_DATE;
        if (!function_exists('json_encode')) {
            throw new \RuntimeException('PHP\'s json extension is required to use Monolog\'s NormalizerFormatter');
        }
    }

    /**
     * {@inheritdoc}
     */
    public function format(array $record)
    {
        return $this->normalize($record);
    }

    /**
     * {@inheritdoc}
     */
    public function formatBatch(array $records)
    {
        foreach ($records as $key => $record) {
            $records[$key] = $this->format($record);
        }

        return $records;
    }

    protected function normalize($data)
    {
        if (null === $data || is_scalar($data)) {
            if (is_float($data)) {
                if (is_infinite($data)) {
                    return ($data > 0 ? '' : '-') . 'INF';
                }
                if (is_nan($data)) {
                    return 'NaN';
                }
            }

            return $data;
        }

        if (is_array($data) || $data instanceof \Traversable) {
            $normalized = array();

            $count = 1;
            foreach ($data as $key => $value) {
                if ($count++ >= 1000) {
                    $normalized['...'] = 'Over 1000 items, aborting normalization';
                    break;
                }
                $normalized[$key] = $this->normalize($value);
            }

            return $normalized;
        }

        if ($data instanceof \DateTimeInterface) {
            return $data->format($this->dateFormat);
        }

        if (is_object($data)) {
            if ($data instanceof Throwable) {
                return $this->normalizeException($data);
            }

            // non-serializable objects that implement __toString stringified
            if (method_exists($data, '__toString') && !$data instanceof \JsonSerializable) {
                $value = $data->__toString();
            } else {
                // the rest is json-serialized in some way
                $value = $this->toJson($data, true);
            }

            return sprintf("[object] (%s: %s)", get_class($data), $value);
        }

        if (is_resource($data)) {
            return sprintf('[resource] (%s)', get_resource_type($data));
        }

        return '[unknown('.gettype($data).')]';
    }

    protected function normalizeException($e)
    {
        if (!$e instanceof Throwable) {
            throw new \InvalidArgumentException('Throwable expected, got '.gettype($e).' / '.get_class($e));
        }

        $data = array(
            'class' => get_class($e),
            'message' => $e->getMessage(),
            'code' => $e->getCode(),
            'file' => $e->getFile().':'.$e->getLine(),
        );

        $trace = $e->getTrace();
        foreach ($trace as $frame) {
            if (isset($frame['file'])) {
                $data['trace'][] = $frame['file'].':'.$frame['line'];
            } elseif (isset($frame['function']) && $frame['function'] === '{closure}') {
                // We should again normalize the frames, because it might contain invalid items
                $data['trace'][] = $frame['function'];
            } else {
                // We should again normalize the frames, because it might contain invalid items
                $data['trace'][] = $this->toJson($this->normalize($frame), true);
            }
        }

        if ($previous = $e->getPrevious()) {
            $data['previous'] = $this->normalizeException($previous);
        }

        return $data;
    }

    /**
     * Return the JSON representation of a value
     *
     * @param  mixed             $data
     * @param  bool              $ignoreErrors
     * @throws \RuntimeException if encoding fails and errors are not ignored
     * @return string
     */
    protected function toJson($data, $ignoreErrors = false)
    {
        // suppress json_encode errors since it's twitchy with some inputs
        if ($ignoreErrors) {
            return @$this->jsonEncode($data);
        }

        $json = $this->jsonEncode($data);

        if ($json === false) {
            $json = $this->handleJsonError(json_last_error(), $data);
        }

        return $json;
    }

    /**
     * @param  mixed  $data
     * @return string JSON encoded data or null on failure
     */
    private function jsonEncode($data)
    {
        return json_encode($data, JSON_UNESCAPED_SLASHES | JSON_UNESCAPED_UNICODE);
    }

    /**
     * Handle a json_encode failure.
     *
     * If the failure is due to invalid string encoding, try to clean the
     * input and encode again. If the second encoding iattempt fails, the
     * inital error is not encoding related or the input can't be cleaned then
     * raise a descriptive exception.
     *
     * @param  int               $code return code of json_last_error function
     * @param  mixed             $data data that was meant to be encoded
     * @throws \RuntimeException if failure can't be corrected
     * @return string            JSON encoded data after error correction
     */
    private function handleJsonError($code, $data)
    {
        if ($code !== JSON_ERROR_UTF8) {
            $this->throwEncodeError($code, $data);
        }

        if (is_string($data)) {
            $this->detectAndCleanUtf8($data);
        } elseif (is_array($data)) {
            array_walk_recursive($data, array($this, 'detectAndCleanUtf8'));
        } else {
            $this->throwEncodeError($code, $data);
        }

        $json = $this->jsonEncode($data);

        if ($json === false) {
            $this->throwEncodeError(json_last_error(), $data);
        }

        return $json;
    }

    /**
     * Throws an exception according to a given code with a customized message
     *
     * @param  int               $code return code of json_last_error function
     * @param  mixed             $data data that was meant to be encoded
     * @throws \RuntimeException
     */
    private function throwEncodeError($code, $data)
    {
        switch ($code) {
            case JSON_ERROR_DEPTH:
                $msg = 'Maximum stack depth exceeded';
                break;
            case JSON_ERROR_STATE_MISMATCH:
                $msg = 'Underflow or the modes mismatch';
                break;
            case JSON_ERROR_CTRL_CHAR:
                $msg = 'Unexpected control character found';
                break;
            case JSON_ERROR_UTF8:
                $msg = 'Malformed UTF-8 characters, possibly incorrectly encoded';
                break;
            default:
                $msg = 'Unknown error';
        }

        throw new \RuntimeException('JSON encoding failed: '.$msg.'. Encoding: '.var_export($data, true));
    }

    /**
     * Detect invalid UTF-8 string characters and convert to valid UTF-8.
     *
     * Valid UTF-8 input will be left unmodified, but strings containing
     * invalid UTF-8 codepoints will be reencoded as UTF-8 with an assumed
     * original encoding of ISO-8859-15. This conversion may result in
     * incorrect output if the actual encoding was not ISO-8859-15, but it
     * will be clean UTF-8 output and will not rely on expensive and fragile
     * detection algorithms.
     *
     * Function converts the input in place in the passed variable so that it
     * can be used as a callback for array_walk_recursive.
     *
     * @param mixed &$data Input to check and convert if needed
     * @private
     */
    public function detectAndCleanUtf8(&$data)
    {
        if (is_string($data) && !preg_match('//u', $data)) {
            $data = preg_replace_callback(
                '/[\x80-\xFF]+/',
                function ($m) { return utf8_encode($m[0]); },
                $data
            );
            $data = str_replace(
                array('¤', '¦', '¨', '´', '¸', '¼', '½', '¾'),
                array('€', 'Š', 'š', 'Ž', 'ž', 'Œ', 'œ', 'Ÿ'),
                $data
            );
        }
    }
}