// SPDX-FileCopyrightText: 2004-2023 Ryan Parman, Sam Sneddon, Ryan McCue
// SPDX-FileCopyrightText: 2008 Steve Minutillo
// SPDX-License-Identifier: BSD-3-Clause
* IRI parser/serialiser/normaliser
* @property ?string $scheme
* @property ?string $userinfo
* @property ?string $host
* @property-write int|string|null $port
* @property ?string $authority
* @property ?string $query
* @property ?string $fragment
protected $scheme = null;
protected $iuserinfo = null;
protected $iquery = null;
protected $ifragment = null;
* Each key is the scheme, each value is an array with each key as the IRI
* part and value as the default value for that part.
* @var array<string, array<string, mixed>>
protected $normalization = [
* Return the entire IRI when you try and read the object as a string
public function __toString()
return (string) $this->get_iri();
* Overload __set() to provide access via properties
* @param string $name Property name
* @param mixed $value Property value
public function __set(string $name, $value)
$callable = [$this, 'set_' . $name];
if (is_callable($callable)) {
call_user_func($callable, $value);
call_user_func([$this, 'set_' . substr($name, 1)], $value);
* Overload __get() to provide access via properties
* @param string $name Property name
public function __get(string $name)
// isset() returns false for null, we don't want to do that
// Also why we use array_key_exists below instead of isset()
$props = get_object_vars($this);
$name === 'iauthority' ||
$return = $this->{"get_$name"}();
} elseif (array_key_exists($name, $props)) {
elseif (array_key_exists($prop = 'i' . $name, $props)) {
elseif (($prop = substr($name, 1)) && array_key_exists($prop, $props)) {
trigger_error('Undefined property: ' . get_class($this) . '::' . $name, E_USER_NOTICE);
if ($return === null && isset($this->scheme, $this->normalization[$this->scheme][$name])) {
return $this->normalization[$this->scheme][$name];
* Overload __isset() to provide access via properties
* @param string $name Property name
public function __isset(string $name)
return method_exists($this, 'get_' . $name) || isset($this->$name);
* Overload __unset() to provide access via properties
* @param string $name Property name
public function __unset(string $name)
$callable = [$this, 'set_' . $name];
if (is_callable($callable)) {
call_user_func($callable, '');
* Create a new IRI object, from a specified string
* @param string|null $iri
public function __construct(?string $iri = null)
public function __destruct()
$this->set_iri(null, true);
$this->set_path(null, true);
$this->set_authority(null, true);
* Create a new IRI object by resolving a relative IRI
* Returns false if $base is not absolute, otherwise an IRI.
* @param IRI|string $base (Absolute) Base IRI
* @param IRI|string $relative Relative IRI
public static function absolutize($base, $relative)
if (!($relative instanceof IRI)) {
$relative = new IRI($relative);
if (!$relative->is_valid()) {
} elseif ($relative->scheme !== null) {
if (!($base instanceof IRI)) {
if ($base->scheme !== null && $base->is_valid()) {
if ($relative->get_iri() !== '') {
if ($relative->iuserinfo !== null || $relative->ihost !== null || $relative->port !== null) {
$target = clone $relative;
$target->scheme = $base->scheme;
$target->scheme = $base->scheme;
$target->iuserinfo = $base->iuserinfo;
$target->ihost = $base->ihost;
$target->port = $base->port;
if ($relative->ipath !== '') {
if ($relative->ipath[0] === '/') {
$target->ipath = $relative->ipath;
} elseif (($base->iuserinfo !== null || $base->ihost !== null || $base->port !== null) && $base->ipath === '') {
$target->ipath = '/' . $relative->ipath;
} elseif (($last_segment = strrpos($base->ipath, '/')) !== false) {
$target->ipath = substr($base->ipath, 0, $last_segment + 1) . $relative->ipath;
$target->ipath = $relative->ipath;
$target->ipath = $target->remove_dot_segments($target->ipath);
$target->iquery = $relative->iquery;
$target->ipath = $base->ipath;
if ($relative->iquery !== null) {
$target->iquery = $relative->iquery;
} elseif ($base->iquery !== null) {
$target->iquery = $base->iquery;
$target->ifragment = $relative->ifragment;
$target->ifragment = null;
$target->scheme_normalization();
* Parse an IRI into scheme/authority/path/query/fragment segments
* authority: string|null,
protected function parse_iri(string $iri)
$iri = trim($iri, "\x20\x09\x0A\x0C\x0D");
if (preg_match('/^(?:(?P<scheme>[^:\/?#]+):)?(:?\/\/(?P<authority>[^\/?#]*))?(?P<path>[^?#]*)(?:\?(?P<query>[^#]*))?(?:#(?P<fragment>.*))?$/', $iri, $match, \PREG_UNMATCHED_AS_NULL)) {
// TODO: Remove once we require PHP ≥ 7.4.
$match['query'] = $match['query'] ?? null;
$match['fragment'] = $match['fragment'] ?? null;
// This can occur when a paragraph is accidentally parsed as a URI
* Remove dot segments from a path
protected function remove_dot_segments(string $input)
while (strpos($input, './') !== false || strpos($input, '/.') !== false || $input === '.' || $input === '..') {
// A: If the input buffer begins with a prefix of "../" or "./", then remove that prefix from the input buffer; otherwise,
if (strpos($input, '../') === 0) {
$input = substr($input, 3);
} elseif (strpos($input, './') === 0) {
$input = substr($input, 2);
// B: if the input buffer begins with a prefix of "/./" or "/.", where "." is a complete path segment, then replace that prefix with "/" in the input buffer; otherwise,
elseif (strpos($input, '/./') === 0) {
$input = substr($input, 2);
} elseif ($input === '/.') {
// C: if the input buffer begins with a prefix of "/../" or "/..", where ".." is a complete path segment, then replace that prefix with "/" in the input buffer and remove the last segment and its preceding "/" (if any) from the output buffer; otherwise,
elseif (strpos($input, '/../') === 0) {
$input = substr($input, 3);
$output = substr_replace($output, '', intval(strrpos($output, '/')));
} elseif ($input === '/..') {
$output = substr_replace($output, '', intval(strrpos($output, '/')));
// D: if the input buffer consists only of "." or "..", then remove that from the input buffer; otherwise,
elseif ($input === '.' || $input === '..') {
// E: move the first path segment in the input buffer to the end of the output buffer, including the initial "/" character (if any) and any subsequent characters up to, but not including, the next "/" character or the end of the input buffer
elseif (($pos = strpos($input, '/', 1)) !== false) {
$output .= substr($input, 0, $pos);
$input = substr_replace($input, '', 0, $pos);
* Replace invalid character with percent encoding
* @param string $string Input string
* @param string $extra_chars Valid characters not in iunreserved or
* iprivate (this is ASCII-only)
* @param bool $iprivate Allow iprivate
protected function replace_invalid_with_pct_encoding(string $string, string $extra_chars, bool $iprivate = false)
// Normalize as many pct-encoded sections as possible
$string = preg_replace_callback('/(?:%[A-Fa-f0-9]{2})+/', [$this, 'remove_iunreserved_percent_encoded'], $string);
\assert(\is_string($string), "For PHPStan: Should not occur, the regex is valid");
// Replace invalid percent characters
$string = preg_replace('/%(?![A-Fa-f0-9]{2})/', '%25', $string);
\assert(\is_string($string), "For PHPStan: Should not occur, the regex is valid");
// Add unreserved and % to $extra_chars (the latter is safe because all
// pct-encoded sections are now valid).
$extra_chars .= 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-._~%';
// Now replace any bytes that aren't allowed with their pct-encoded versions
$strlen = strlen($string);
while (($position += strspn($string, $extra_chars, $position)) < $strlen) {
$value = ord($string[$position]);
// By default we are valid
// No one byte sequences are valid due to the while.
if (($value & 0xE0) === 0xC0) {
$character = ($value & 0x1F) << 6;
elseif (($value & 0xF0) === 0xE0) {
$character = ($value & 0x0F) << 12;
elseif (($value & 0xF8) === 0xF0) {
$character = ($value & 0x07) << 18;
if ($position + $length <= $strlen) {
for ($position++; $remaining; $position++) {
$value = ord($string[$position]);
// Check that the byte is valid, then add it to the character:
if (($value & 0xC0) === 0x80) {
$character |= ($value & 0x3F) << (--$remaining * 6);
// If it is invalid, count the sequence as invalid and reprocess the current byte:
// Percent encode anything invalid or not in ucschar
// Non-shortest form sequences are invalid
|| $length > 1 && $character <= 0x7F
|| $length > 2 && $character <= 0x7FF
|| $length > 3 && $character <= 0xFFFF
// Outside of range of ucschar codepoints
|| ($character & 0xFFFE) === 0xFFFE
|| $character >= 0xFDD0 && $character <= 0xFDEF
// Everything else not in ucschar
$character > 0xD7FF && $character < 0xF900
// Everything not in iprivate, if it applies
// If we were a character, pretend we weren't, but rather an error.
for ($j = $start; $j <= $position; $j++) {
$string = substr_replace($string, sprintf('%%%02X', ord($string[$j])), $j, 1);
* Callback function for preg_replace_callback.
* Removes sequences of percent encoded bytes that represent UTF-8
* encoded characters in iunreserved
* @param array{string} $match PCRE match, a capture group #0 consisting of a sequence of valid percent-encoded bytes
* @return string Replacement
protected function remove_iunreserved_percent_encoded(array $match)
It is recommended that you Edit text format, this type of Fix handles quite a lot in one request
