Rfc822AddressesParser.php

Go to the documentation of this file.

<?php
namespace TYPO3\CMS\Core\Mail;

class Rfc822AddressesParser
{
    private $address = '';

    private $default_domain = 'localhost';

    private $validate = true;

    private $addresses = [];

    private $structure = [];

    private $error = null;

    private $index = null;

    private $num_groups = 0;

    private $limit = null;

    public function __construct($address = null, $default_domain = null, $validate = null, $limit = null)
    {
        if (isset($address)) {
            $this->address = $address;
        }
        if (isset($default_domain)) {
            $this->default_domain = $default_domain;
        }
        if (isset($validate)) {
            $this->validate = $validate;
        }
        if (isset($limit)) {
            $this->limit = $limit;
        }
    }

    public function parseAddressList($address = null, $default_domain = null, $validate = null, $limit = null)
    {
        if (isset($address)) {
            $this->address = $address;
        }
        if (isset($default_domain)) {
            $this->default_domain = $default_domain;
        }
        if (isset($validate)) {
            $this->validate = $validate;
        }
        if (isset($limit)) {
            $this->limit = $limit;
        }
        $this->structure = [];
        $this->addresses = [];
        $this->error = null;
        $this->index = null;
        // Unfold any long lines in $this->address.
        $this->address = preg_replace('/\\r?\\n/', '
', $this->address);
        $this->address = preg_replace('/\\r\\n(\\t| )+/', ' ', $this->address);
        while ($this->address = $this->_splitAddresses($this->address)) {
        }
        if ($this->address === false || isset($this->error)) {
            throw new \InvalidArgumentException($this->error, 1294681466);
        }
        // Validate each address individually.  If we encounter an invalid
        // address, stop iterating and return an error immediately.
        foreach ($this->addresses as $address) {
            $valid = $this->_validateAddress($address);
            if ($valid === false || isset($this->error)) {
                throw new \InvalidArgumentException($this->error, 1294681467);
            }
            $this->structure = array_merge($this->structure, $valid);
        }
        return $this->structure;
    }

    protected function _splitAddresses($address)
    {
        if (!empty($this->limit) && count($this->addresses) == $this->limit) {
            return '';
        }
        if ($this->_isGroup($address) && !isset($this->error)) {
            $split_char = ';';
            $is_group = true;
        } elseif (!isset($this->error)) {
            $split_char = ',';
            $is_group = false;
        } elseif (isset($this->error)) {
            return false;
        }
        // Split the string based on the above ten or so lines.
        $parts = explode($split_char, $address);
        $string = $this->_splitCheck($parts, $split_char);
        // If a group...
        if ($is_group) {
            // If $string does not contain a colon outside of
            // brackets/quotes etc then something's fubar.
            // First check there's a colon at all:
            if (strpos($string, ':') === false) {
                $this->error = 'Invalid address: ' . $string;
                return false;
            }
            // Now check it's outside of brackets/quotes:
            if (!$this->_splitCheck(explode(':', $string), ':')) {
                return false;
            }
            // We must have a group at this point, so increase the counter:
            $this->num_groups++;
        }
        // $string now contains the first full address/group.
        // Add to the addresses array.
        $this->addresses[] = [
            'address' => trim($string),
            'group' => $is_group
        ];
        // Remove the now stored address from the initial line, the +1
        // is to account for the explode character.
        $address = trim(substr($address, strlen($string) + 1));
        // If the next char is a comma and this was a group, then
        // there are more addresses, otherwise, if there are any more
        // chars, then there is another address.
        if ($is_group && $address[0] === ',') {
            $address = trim(substr($address, 1));
            return $address;
        }
        if ($address !== '') {
            return $address;
        }
        return '';
    }

    protected function _isGroup($address)
    {
        // First comma not in quotes, angles or escaped:
        $parts = explode(',', $address);
        $string = $this->_splitCheck($parts, ',');
        // Now we have the first address, we can reliably check for a
        // group by searching for a colon that's not escaped or in
        // quotes or angle brackets.
        if (count(($parts = explode(':', $string))) > 1) {
            $string2 = $this->_splitCheck($parts, ':');
            return $string2 !== $string;
        }
        return false;
    }

    protected function _splitCheck($parts, $char)
    {
        $string = $parts[0];
        $partsCounter = count($parts);
        for ($i = 0; $i < $partsCounter; $i++) {
            if ($this->_hasUnclosedQuotes($string) || $this->_hasUnclosedBrackets($string, '<>') || $this->_hasUnclosedBrackets($string, '[]') || $this->_hasUnclosedBrackets($string, '()') || substr($string, -1) === '\\') {
                if (isset($parts[$i + 1])) {
                    $string = $string . $char . $parts[$i + 1];
                } else {
                    $this->error = 'Invalid address spec. Unclosed bracket or quotes';
                    return false;
                }
            } else {
                $this->index = $i;
                break;
            }
        }
        return $string;
    }

    protected function _hasUnclosedQuotes($string)
    {
        $string = trim($string);
        $iMax = strlen($string);
        $in_quote = false;
        $i = ($slashes = 0);
        for (; $i < $iMax; ++$i) {
            switch ($string[$i]) {
                case '\\':
                    ++$slashes;
                    break;
                case '"':
                    if ($slashes % 2 == 0) {
                        $in_quote = !$in_quote;
                    }
                    // no break
                default:
                    $slashes = 0;
            }
        }
        return $in_quote;
    }

    protected function _hasUnclosedBrackets($string, $chars)
    {
        $num_angle_start = substr_count($string, $chars[0]);
        $num_angle_end = substr_count($string, $chars[1]);
        $this->_hasUnclosedBracketsSub($string, $num_angle_start, $chars[0]);
        $this->_hasUnclosedBracketsSub($string, $num_angle_end, $chars[1]);
        if ($num_angle_start < $num_angle_end) {
            $this->error = 'Invalid address spec. Unmatched quote or bracket (' . $chars . ')';
            return false;
        }
        return $num_angle_start > $num_angle_end;
    }

    protected function _hasUnclosedBracketsSub($string, &$num, $char)
    {
        $parts = explode($char, $string);
        $partsCounter = count($parts);
        for ($i = 0; $i < $partsCounter; $i++) {
            if (substr($parts[$i], -1) === '\\' || $this->_hasUnclosedQuotes($parts[$i])) {
                $num--;
            }
            if (isset($parts[$i + 1])) {
                $parts[$i + 1] = $parts[$i] . $char . $parts[$i + 1];
            }
        }
        return $num;
    }

    protected function _validateAddress($address)
    {
        $is_group = false;
        $addresses = [];
        if ($address['group']) {
            $is_group = true;
            // Get the group part of the name
            $parts = explode(':', $address['address']);
            $groupname = $this->_splitCheck($parts, ':');
            $structure = [];
            // And validate the group part of the name.
            if (!$this->_validatePhrase($groupname)) {
                $this->error = 'Group name did not validate.';
                return false;
            }
            $address['address'] = ltrim(substr($address['address'], strlen($groupname . ':')));
        }
        // If a group then split on comma and put into an array.
        // Otherwise, Just put the whole address in an array.
        if ($is_group) {
            while ($address['address'] !== '') {
                $parts = explode(',', $address['address']);
                $addresses[] = $this->_splitCheck($parts, ',');
                $address['address'] = trim(substr($address['address'], strlen(end($addresses) . ',')));
            }
        } else {
            $addresses[] = $address['address'];
        }
        // Check that $addresses is set, if address like this:
        // Groupname:;
        // Then errors were appearing.
        if (empty($addresses)) {
            $this->error = 'Empty group.';
            return false;
        }
        // Trim the whitespace from all of the address strings.
        array_map('trim', $addresses);
        // Validate each mailbox.
        // Format could be one of: name <geezer@domain.com>
        //                         geezer@domain.com
        //                         geezer
        // ... or any other format valid by RFC 822.
        $addressesCount = count($addresses);
        for ($i = 0; $i < $addressesCount; $i++) {
            if (!$this->validateMailbox($addresses[$i])) {
                if (empty($this->error)) {
                    $this->error = 'Validation failed for: ' . $addresses[$i];
                }
                return false;
            }
        }
        if ($is_group) {
            $structure = array_merge($structure, $addresses);
        } else {
            $structure = $addresses;
        }
        return $structure;
    }

    protected function _validatePhrase($phrase)
    {
        // Splits on one or more Tab or space.
        $parts = preg_split('/[ \\x09]+/', $phrase, -1, PREG_SPLIT_NO_EMPTY);
        $phrase_parts = [];
        while (!empty($parts)) {
            $phrase_parts[] = $this->_splitCheck($parts, ' ');
            for ($i = 0; $i < $this->index + 1; $i++) {
                array_shift($parts);
            }
        }
        foreach ($phrase_parts as $part) {
            // If quoted string:
            if ($part[0] === '"') {
                if (!$this->_validateQuotedString($part)) {
                    return false;
                }
                continue;
            }
            // Otherwise it's an atom:
            if (!$this->_validateAtom($part)) {
                return false;
            }
        }
        return true;
    }

    protected function _validateAtom($atom)
    {
        if (!$this->validate) {
            // Validation has been turned off; assume the atom is okay.
            return true;
        }
        // Check for any char from ASCII 0 - ASCII 127
        if (!preg_match('/^[\\x00-\\x7E]+$/i', $atom, $matches)) {
            return false;
        }
        // Check for specials:
        if (preg_match('/[][()<>@,;\\:". ]/', $atom)) {
            return false;
        }
        // Check for control characters (ASCII 0-31):
        if (preg_match('/[\\x00-\\x1F]+/', $atom)) {
            return false;
        }
        return true;
    }

    protected function _validateQuotedString($qstring)
    {
        // Leading and trailing "
        $qstring = substr($qstring, 1, -1);
        // Perform check, removing quoted characters first.
        return !preg_match('/[\\x0D\\\\"]/', preg_replace('/\\\\./', '', $qstring));
    }

    protected function validateMailbox(&$mailbox)
    {
        // A couple of defaults.
        $phrase = '';
        $comment = '';
        $comments = [];
        // Catch any RFC822 comments and store them separately.
        $_mailbox = $mailbox;
        while (trim($_mailbox) !== '') {
            $parts = explode('(', $_mailbox);
            $before_comment = $this->_splitCheck($parts, '(');
            if ($before_comment != $_mailbox) {
                // First char should be a (.
                $comment = substr(str_replace($before_comment, '', $_mailbox), 1);
                $parts = explode(')', $comment);
                $comment = $this->_splitCheck($parts, ')');
                $comments[] = $comment;
                // +2 is for the brackets
                $_mailbox = substr($_mailbox, strpos($_mailbox, ('(' . $comment)) + strlen($comment) + 2);
            } else {
                break;
            }
        }
        foreach ($comments as $comment) {
            $mailbox = str_replace('(' . $comment . ')', '', $mailbox);
        }
        $mailbox = trim($mailbox);
        // Check for name + route-addr
        if (substr($mailbox, -1) === '>' && $mailbox[0] !== '<') {
            $parts = explode('<', $mailbox);
            $name = $this->_splitCheck($parts, '<');
            $phrase = trim($name);
            $route_addr = trim(substr($mailbox, strlen($name . '<'), -1));
            if ($this->_validatePhrase($phrase) === false || ($route_addr = $this->_validateRouteAddr($route_addr)) === false) {
                return false;
            }
        } else {
            // First snip angle brackets if present.
            if ($mailbox[0] === '<' && substr($mailbox, -1) === '>') {
                $addr_spec = substr($mailbox, 1, -1);
            } else {
                $addr_spec = $mailbox;
            }
            if (($addr_spec = $this->_validateAddrSpec($addr_spec)) === false) {
                return false;
            }
        }
        // Construct the object that will be returned.
        $mbox = new \stdClass();
        // Add the phrase (even if empty) and comments
        $mbox->personal = $phrase;
        $mbox->comment = isset($comments) ? $comments : [];
        if (isset($route_addr)) {
            $mbox->mailbox = $route_addr['local_part'];
            $mbox->host = $route_addr['domain'];
            $route_addr['adl'] !== '' ? ($mbox->adl = $route_addr['adl']) : '';
        } else {
            $mbox->mailbox = $addr_spec['local_part'];
            $mbox->host = $addr_spec['domain'];
        }
        $mailbox = $mbox;
        return true;
    }

    protected function _validateRouteAddr($route_addr)
    {
        // Check for colon.
        if (strpos($route_addr, ':') !== false) {
            $parts = explode(':', $route_addr);
            $route = $this->_splitCheck($parts, ':');
        } else {
            $route = $route_addr;
        }
        // If $route is same as $route_addr then the colon was in
        // quotes or brackets or, of course, non existent.
        if ($route === $route_addr) {
            unset($route);
            $addr_spec = $route_addr;
            if (($addr_spec = $this->_validateAddrSpec($addr_spec)) === false) {
                return false;
            }
        } else {
            // Validate route part.
            if (($route = $this->_validateRoute($route)) === false) {
                return false;
            }
            $addr_spec = substr($route_addr, strlen($route . ':'));
            // Validate addr-spec part.
            if (($addr_spec = $this->_validateAddrSpec($addr_spec)) === false) {
                return false;
            }
        }
        if (isset($route)) {
            $return['adl'] = $route;
        } else {
            $return['adl'] = '';
        }
        $return = array_merge($return, $addr_spec);
        return $return;
    }

    protected function _validateRoute($route)
    {
        // Split on comma.
        $domains = explode(',', trim($route));
        foreach ($domains as $domain) {
            $domain = str_replace('@', '', trim($domain));
            if (!$this->_validateDomain($domain)) {
                return false;
            }
        }
        return $route;
    }

    protected function _validateDomain($domain)
    {
        // Note the different use of $subdomains and $sub_domains
        $subdomains = explode('.', $domain);
        while (!empty($subdomains)) {
            $sub_domains[] = $this->_splitCheck($subdomains, '.');
            for ($i = 0; $i < $this->index + 1; $i++) {
                array_shift($subdomains);
            }
        }
        foreach ($sub_domains as $sub_domain) {
            if (!$this->_validateSubdomain(trim($sub_domain))) {
                return false;
            }
        }
        // Managed to get here, so return input.
        return $domain;
    }

    protected function _validateSubdomain($subdomain)
    {
        if (preg_match('|^\\[(.*)]$|', $subdomain, $arr)) {
            if (!$this->_validateDliteral($arr[1])) {
                return false;
            }
        } else {
            if (!$this->_validateAtom($subdomain)) {
                return false;
            }
        }
        // Got here, so return successful.
        return true;
    }

    protected function _validateDliteral($dliteral)
    {
        return !preg_match('/(.)[][\\x0D\\\\]/', $dliteral, $matches) && $matches[1] !== '\\';
    }

    protected function _validateAddrSpec($addr_spec)
    {
        $addr_spec = trim($addr_spec);
        // Split on @ sign if there is one.
        if (strpos($addr_spec, '@') !== false) {
            $parts = explode('@', $addr_spec);
            $local_part = $this->_splitCheck($parts, '@');
            $domain = substr($addr_spec, strlen($local_part . '@'));
        } else {
            $local_part = $addr_spec;
            $domain = $this->default_domain;
        }
        if (($local_part = $this->_validateLocalPart($local_part)) === false) {
            return false;
        }
        if (($domain = $this->_validateDomain($domain)) === false) {
            return false;
        }
        // Got here so return successful.
        return ['local_part' => $local_part, 'domain' => $domain];
    }

    protected function _validateLocalPart($local_part)
    {
        $parts = explode('.', $local_part);
        $words = [];
        // Split the local_part into words.
        while (!empty($parts)) {
            $words[] = $this->_splitCheck($parts, '.');
            for ($i = 0; $i < $this->index + 1; $i++) {
                array_shift($parts);
            }
        }
        // Validate each word.
        foreach ($words as $word) {
            // If this word contains an unquoted space, it is invalid. (6.2.4)
            if (strpos($word, ' ') && $word[0] !== '"') {
                return false;
            }
            if ($this->_validatePhrase(trim($word)) === false) {
                return false;
            }
        }
        // Managed to get here, so return the input.
        return $local_part;
    }
}