X-Git-Url: https://git.heureux-cyclage.org/?a=blobdiff_plain;f=includes%2FIP.php;h=1b40f4bc1abd6257c7a74daea101d80a88d7b8e5;hb=d26f07888b8002fd1a7f5101fe5f6a70bbb85e2a;hp=9aeba39f2c0165e9cc7a76d88d4dff7167eb3576;hpb=9a301f1636a962ff8cd5d4447f06679fdc890b5a;p=lhc%2Fweb%2Fwiklou.git diff --git a/includes/IP.php b/includes/IP.php index 9aeba39f2c..1b40f4bc1a 100644 --- a/includes/IP.php +++ b/includes/IP.php @@ -18,7 +18,7 @@ * http://www.gnu.org/copyleft/gpl.html * * @file - * @author Ashar Voultoiz , Aaron Schulz + * @author Antoine Musso "", Aaron Schulz */ // Some regex definition to "play" with IP address and IP address blocks @@ -35,19 +35,22 @@ define( 'RE_IP_BLOCK', RE_IP_ADD . '\/' . RE_IP_PREFIX ); define( 'RE_IPV6_WORD', '([0-9A-Fa-f]{1,4})' ); define( 'RE_IPV6_PREFIX', '(12[0-8]|1[01][0-9]|[1-9]?\d)'); define( 'RE_IPV6_ADD', - '(?:' . // starts with "::" (includes the address "::") - '::|:(?::' . RE_IPV6_WORD . '){1,7}' . - '|' . // ends with "::" (not including the address "::") + '(?:' . // starts with "::" (including "::") + ':(?::|(?::' . RE_IPV6_WORD . '){1,7})' . + '|' . // ends with "::" (except "::") RE_IPV6_WORD . '(?::' . RE_IPV6_WORD . '){0,6}::' . - '|' . // has no "::" - RE_IPV6_WORD . '(?::' . RE_IPV6_WORD . '){7}' . - '|' . // contains one "::" in the middle (awkward regex for PCRE 4.0+ compatibility) - RE_IPV6_WORD . '(?::(?!(?P=abn))(?P:(?P))?' . RE_IPV6_WORD . '){1,6}(?P=iabn)' . + '|' . // contains one "::" in the middle, ending in "::WORD" + RE_IPV6_WORD . '(?::' . RE_IPV6_WORD . '){0,5}' . '::' . RE_IPV6_WORD . + '|' . // contains one "::" in the middle, not ending in "::WORD" (regex for PCRE 4.0+) + RE_IPV6_WORD . '(?::(?P:(?P))?' . RE_IPV6_WORD . '(?!:(?P=abn))){1,5}' . + ':' . RE_IPV6_WORD . '(?P=iabn)' . // NOTE: (?!(?P=abn)) fails iff "::" used twice; (?P=iabn) passes iff a "::" was found. - - // Better regexp (PCRE 7.2+ only), allows intuitive regex concatenation - #RE_IPV6_WORD . '(?::((?(-1)|:))?' . RE_IPV6_WORD . '){1,6}(?(-2)|^)' . + '|' . // contains no "::" + RE_IPV6_WORD . '(?::' . RE_IPV6_WORD . '){7}' . ')' + // NOTE: With PCRE 7.2+, we can combine the two '"::" in the middle' cases into: + // RE_IPV6_WORD . '(?::((?(-1)|:))?' . RE_IPV6_WORD . '){1,6}(?(-2)|^)' + // This also improves regex concatenation by using relative references. ); // An IPv6 block is an IP address and a prefix (d1 to d128) define( 'RE_IPV6_BLOCK', RE_IPV6_ADD . '\/' . RE_IPV6_PREFIX ); @@ -130,7 +133,7 @@ class IP { } /** - * Convert an IP into a nice standard form. + * Convert an IP into a verbose, uppercase, normalized form. * IPv6 addresses in octet notation are expanded to 8 words. * IPv4 addresses are just trimmed. * @@ -182,6 +185,124 @@ class IP { return $ip; } + /** + * Prettify an IP for display to end users. + * This will make it more compact and lower-case. + * + * @param $ip string + * @return string + */ + public static function prettifyIP( $ip ) { + $ip = self::sanitizeIP( $ip ); // normalize (removes '::') + if ( self::isIPv6( $ip ) ) { + // Split IP into an address and a CIDR + if ( strpos( $ip, '/' ) !== false ) { + list( $ip, $cidr ) = explode( '/', $ip, 2 ); + } else { + list( $ip, $cidr ) = array( $ip, '' ); + } + // Get the largest slice of words with multiple zeros + $offset = 0; + $longest = $longestPos = false; + while ( preg_match( + '!(?:^|:)0(?::0)+(?:$|:)!', $ip, $m, PREG_OFFSET_CAPTURE, $offset + ) ) { + list( $match, $pos ) = $m[0]; // full match + if ( strlen( $match ) > strlen( $longest ) ) { + $longest = $match; + $longestPos = $pos; + } + $offset += ( $pos + strlen( $match ) ); // advance + } + if ( $longest !== false ) { + // Replace this portion of the string with the '::' abbreviation + $ip = substr_replace( $ip, '::', $longestPos, strlen( $longest ) ); + } + // Add any CIDR back on + if ( $cidr !== '' ) { + $ip = "{$ip}/{$cidr}"; + } + // Convert to lower case to make it more readable + $ip = strtolower( $ip ); + } + return $ip; + } + + /** + * Given a host/port string, like one might find in the host part of a URL + * per RFC 2732, split the hostname part and the port part and return an + * array with an element for each. If there is no port part, the array will + * have false in place of the port. If the string was invalid in some way, + * false is returned. + * + * This was easy with IPv4 and was generally done in an ad-hoc way, but + * with IPv6 it's somewhat more complicated due to the need to parse the + * square brackets and colons. + * + * A bare IPv6 address is accepted despite the lack of square brackets. + * + * @param $both string The string with the host and port + * @return array + */ + public static function splitHostAndPort( $both ) { + if ( substr( $both, 0, 1 ) === '[' ) { + if ( preg_match( '/^\[(' . RE_IPV6_ADD . ')\](?::(?P\d+))?$/', $both, $m ) ) { + if ( isset( $m['port'] ) ) { + return array( $m[1], intval( $m['port'] ) ); + } else { + return array( $m[1], false ); + } + } else { + // Square bracket found but no IPv6 + return false; + } + } + $numColons = substr_count( $both, ':' ); + if ( $numColons >= 2 ) { + // Is it a bare IPv6 address? + if ( preg_match( '/^' . RE_IPV6_ADD . '$/', $both ) ) { + return array( $both, false ); + } else { + // Not valid IPv6, but too many colons for anything else + return false; + } + } + if ( $numColons >= 1 ) { + // Host:port? + $bits = explode( ':', $both ); + if ( preg_match( '/^\d+/', $bits[1] ) ) { + return array( $bits[0], intval( $bits[1] ) ); + } else { + // Not a valid port + return false; + } + } + // Plain hostname + return array( $both, false ); + } + + /** + * Given a host name and a port, combine them into host/port string like + * you might find in a URL. If the host contains a colon, wrap it in square + * brackets like in RFC 2732. If the port matches the default port, omit + * the port specification + * + * @param $host string + * @param $port int + * @param $defaultPort bool|int + * @return string + */ + public static function combineHostAndPort( $host, $port, $defaultPort = false ) { + if ( strpos( $host, ':' ) !== false ) { + $host = "[$host]"; + } + if ( $defaultPort !== false && $port == $defaultPort ) { + return $host; + } else { + return "$host:$port"; + } + } + /** * Given an unsigned integer, returns an IPv6 address in octet notation * @@ -300,7 +421,7 @@ class IP { static $privateRanges = false; if ( !$privateRanges ) { $privateRanges = array( - array( 'fc::', 'fdff:ffff:ffff:ffff:ffff:ffff:ffff:ffff' ), # RFC 4193 (local) + array( 'fc00::', 'fdff:ffff:ffff:ffff:ffff:ffff:ffff:ffff' ), # RFC 4193 (local) array( '0:0:0:0:0:0:0:1', '0:0:0:0:0:0:0:1' ), # loopback ); } @@ -376,6 +497,10 @@ class IP { return $n; } + /** + * @param $ip + * @return String + */ private static function toUnsigned6( $ip ) { return wfBaseConvert( self::IPv6ToRawHex( $ip ), 16, 10 ); } @@ -475,6 +600,8 @@ class IP { * Convert a network specification in IPv6 CIDR notation to an * integer network and a number of bits * + * @param $range + * * @return array(string, int) */ private static function parseCIDR6( $range ) { @@ -512,6 +639,9 @@ class IP { * 2001:0db8:85a3::7344/96 CIDR * 2001:0db8:85a3::7344 - 2001:0db8:85a3::7344 Explicit range * 2001:0db8:85a3::7344/96 Single IP + * + * @param $range + * * @return array(string, string) */ private static function parseRange6( $range ) { @@ -584,6 +714,7 @@ class IP { * @return String: valid dotted quad IPv4 address or null */ public static function canonicalize( $addr ) { + $addr = preg_replace( '/\%.*/','', $addr ); // remove zone info (bug 35738) if ( self::isValid( $addr ) ) { return $addr; } @@ -597,7 +728,7 @@ class IP { // IPv6 loopback address $m = array(); if ( preg_match( '/^0*' . RE_IPV6_GAP . '1$/', $addr, $m ) ) { - return '127.0.0.1'; + return '127.0.0.1'; } // IPv4-mapped and IPv4-compatible IPv6 addresses if ( preg_match( '/^' . RE_IPV6_V4_PREFIX . '(' . RE_IP_ADD . ')$/i', $addr, $m ) ) { @@ -611,4 +742,20 @@ class IP { return null; // give up } + + /** + * Gets rid of uneeded numbers in quad-dotted/octet IP strings + * For example, 127.111.113.151/24 -> 127.111.113.0/24 + * @param $range String: IP address to normalize + * @return string + */ + public static function sanitizeRange( $range ) { + list( /*...*/, $bits ) = self::parseCIDR( $range ); + list( $start, /*...*/ ) = self::parseRange( $range ); + $start = self::formatHex( $start ); + if ( $bits === false ) { + return $start; // wasn't actually a range + } + return "$start/$bits"; + } }