Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
97.55% |
159 / 163 |
|
66.67% |
8 / 12 |
CRAP | |
0.00% |
0 / 1 |
| UrlUtils | |
97.55% |
159 / 163 |
|
66.67% |
8 / 12 |
90 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
15 / 15 |
|
100.00% |
1 / 1 |
14 | |||
| expand | |
97.22% |
35 / 36 |
|
0.00% |
0 / 1 |
20 | |||
| getServer | |
75.00% |
3 / 4 |
|
0.00% |
0 / 1 |
2.06 | |||
| getCanonicalServer | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| assemble | |
100.00% |
21 / 21 |
|
100.00% |
1 / 1 |
11 | |||
| removeDotSegments | |
100.00% |
39 / 39 |
|
100.00% |
1 / 1 |
18 | |||
| validProtocols | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
1 | |||
| validAbsoluteProtocols | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
1 | |||
| validProtocolsInternal | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
5 | |||
| parse | |
100.00% |
19 / 19 |
|
100.00% |
1 / 1 |
10 | |||
| expandIRI | |
87.50% |
7 / 8 |
|
0.00% |
0 / 1 |
2.01 | |||
| matchesDomainList | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
5 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace MediaWiki\Utils; |
| 4 | |
| 5 | use BadMethodCallException; |
| 6 | use InvalidArgumentException; |
| 7 | use MediaWiki\Debug\MWDebug; |
| 8 | use MediaWiki\MainConfigSchema; |
| 9 | |
| 10 | /** |
| 11 | * A service to expand, parse, and otherwise manipulate URLs. |
| 12 | * |
| 13 | * @since 1.39 |
| 14 | * @newable |
| 15 | */ |
| 16 | class UrlUtils { |
| 17 | public const SERVER = 'server'; |
| 18 | public const CANONICAL_SERVER = 'canonicalServer'; |
| 19 | public const INTERNAL_SERVER = 'internalServer'; |
| 20 | public const FALLBACK_PROTOCOL = 'fallbackProtocol'; |
| 21 | public const HTTPS_PORT = 'httpsPort'; |
| 22 | public const VALID_PROTOCOLS = 'validProtocols'; |
| 23 | |
| 24 | /** @var ?string */ |
| 25 | private $server = null; |
| 26 | |
| 27 | /** @var ?string */ |
| 28 | private $canonicalServer = null; |
| 29 | |
| 30 | /** @var ?string */ |
| 31 | private $internalServer = null; |
| 32 | /** @var string */ |
| 33 | private $fallbackProtocol = 'http'; |
| 34 | |
| 35 | /** @var int */ |
| 36 | private $httpsPort = 443; |
| 37 | |
| 38 | /** @var array */ |
| 39 | private $validProtocols = MainConfigSchema::UrlProtocols['default']; |
| 40 | |
| 41 | /** @var ?string */ |
| 42 | private $validProtocolsCache = null; |
| 43 | |
| 44 | /** @var ?string */ |
| 45 | private $validAbsoluteProtocolsCache = null; |
| 46 | |
| 47 | /** |
| 48 | * @stable to call |
| 49 | * @param array $options All keys are optional, but if you omit SERVER then calling expand() |
| 50 | * (and getServer(), expandIRI(), and matchesDomainList()) will throw. Recognized keys: |
| 51 | * * self::SERVER: The protocol and server portion of the URLs to expand, with no other parts |
| 52 | * (port, path, etc.). Example: 'https://example.com'. Protocol-relative URLs are |
| 53 | * allowed. |
| 54 | * * self::CANONICAL_SERVER: If SERVER is protocol-relative, this can be set to a |
| 55 | * fully-qualified version for use when PROTO_CANONICAL is passed to expand(). Defaults |
| 56 | * to SERVER, with 'http:' prepended if SERVER is protocol-relative. |
| 57 | * * self::INTERNAL_SERVER: An alternative to SERVER that's used when PROTO_INTERNAL is |
| 58 | * passed to expand(). It's intended for sites that have a different server name exposed |
| 59 | * to CDNs. Defaults to SERVER. |
| 60 | * * self::FALLBACK_PROTOCOL: Used by expand() when no $defaultProto parameter is provided. |
| 61 | * Defaults to 'http'. The instance created by ServiceWiring sets this to 'https' if the |
| 62 | * current request is detected to be via HTTPS, and 'http' otherwise. |
| 63 | * * self::HTTPS_PORT: Defaults to 443. Used when a protocol-relative URL is expanded to |
| 64 | * https. |
| 65 | * * self::VALID_PROTOCOLS: An array of recognized URL protocols. The default can be found |
| 66 | * in MainConfigSchema::UrlProtocols['default']. |
| 67 | */ |
| 68 | public function __construct( array $options = [] ) { |
| 69 | foreach ( $options as $key => $value ) { |
| 70 | switch ( $key ) { |
| 71 | case self::SERVER: |
| 72 | case self::CANONICAL_SERVER: |
| 73 | case self::INTERNAL_SERVER: |
| 74 | case self::FALLBACK_PROTOCOL: |
| 75 | case self::HTTPS_PORT: |
| 76 | case self::VALID_PROTOCOLS: |
| 77 | $this->$key = $value; |
| 78 | break; |
| 79 | |
| 80 | default: |
| 81 | throw new InvalidArgumentException( "Unrecognized option \"$key\"" ); |
| 82 | } |
| 83 | } |
| 84 | |
| 85 | if ( $this->server !== null ) { |
| 86 | if ( $this->canonicalServer === null || $this->canonicalServer === false ) { |
| 87 | $this->canonicalServer = $this->expand( $this->server, PROTO_HTTP ); |
| 88 | } |
| 89 | if ( $this->internalServer === null || $this->internalServer === false ) { |
| 90 | $this->internalServer = $this->server; |
| 91 | } |
| 92 | } |
| 93 | } |
| 94 | |
| 95 | /** |
| 96 | * Expand a potentially local URL to a fully-qualified URL using $wgServer |
| 97 | * (or one of its alternatives). |
| 98 | * |
| 99 | * The meaning of the PROTO_* constants is as follows: |
| 100 | * PROTO_HTTP: Output a URL starting with http:// |
| 101 | * PROTO_HTTPS: Output a URL starting with https:// |
| 102 | * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL) |
| 103 | * PROTO_FALLBACK: Output a URL starting with the FALLBACK_PROTOCOL option |
| 104 | * PROTO_CURRENT: Legacy alias for PROTO_FALLBACK |
| 105 | * PROTO_CANONICAL: For URLs without a domain, like /w/index.php, use CANONICAL_SERVER. For |
| 106 | * protocol-relative URLs, use the protocol of CANONICAL_SERVER |
| 107 | * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses INTERNAL_SERVER instead of CANONICAL_SERVER |
| 108 | * |
| 109 | * If $url specifies a protocol, or $url is domain-relative and $wgServer |
| 110 | * specifies a protocol, PROTO_HTTP, PROTO_HTTPS, PROTO_RELATIVE and |
| 111 | * PROTO_CURRENT do not change that. |
| 112 | * |
| 113 | * Parent references (/../) in the path are resolved (as in ::removeDotSegments). |
| 114 | * |
| 115 | * @todo this won't work with current-path-relative URLs like "subdir/foo.html", etc. |
| 116 | * |
| 117 | * @throws BadMethodCallException if no server was passed to the constructor |
| 118 | * @param string $url An URL; can be absolute (e.g. http://example.com/foo/bar), |
| 119 | * protocol-relative (//example.com/foo/bar) or domain-relative (/foo/bar). |
| 120 | * @param string|int|null $defaultProto One of the PROTO_* constants, as described above. |
| 121 | * @return ?string Fully-qualified URL, current-path-relative URL or null if |
| 122 | * no valid URL can be constructed |
| 123 | */ |
| 124 | public function expand( string $url, $defaultProto = PROTO_FALLBACK ): ?string { |
| 125 | if ( $defaultProto === PROTO_CANONICAL ) { |
| 126 | $serverUrl = $this->canonicalServer; |
| 127 | } elseif ( $defaultProto === PROTO_INTERNAL ) { |
| 128 | $serverUrl = $this->internalServer; |
| 129 | } else { |
| 130 | $serverUrl = $this->server; |
| 131 | if ( $defaultProto === PROTO_FALLBACK ) { |
| 132 | $defaultProto = $this->fallbackProtocol . '://'; |
| 133 | } |
| 134 | } |
| 135 | |
| 136 | if ( str_starts_with( $url, '/' ) ) { |
| 137 | if ( $serverUrl === null ) { |
| 138 | throw new BadMethodCallException( 'Cannot call expand() if the appropriate ' . |
| 139 | 'SERVER/CANONICAL_SERVER/INTERNAL_SERVER option was not passed to the ' . |
| 140 | 'constructor' ); |
| 141 | } |
| 142 | |
| 143 | // Analyze $serverUrl to obtain its protocol |
| 144 | $bits = $this->parse( $serverUrl ); |
| 145 | $serverProto = $bits && $bits['scheme'] != '' ? $bits['scheme'] . '://' : null; |
| 146 | |
| 147 | if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) { |
| 148 | // Fall back to HTTP in the ridiculous case that CanonicalServer or InternalServer |
| 149 | // doesn't have a protocol |
| 150 | $defaultProto = $serverProto ?? PROTO_HTTP; |
| 151 | } |
| 152 | |
| 153 | // @phan-suppress-next-line PhanTypeMismatchArgumentNullableInternal T308355 |
| 154 | $defaultProtoWithoutSlashes = $defaultProto === PROTO_FALLBACK ? '' : substr( $defaultProto, 0, -2 ); |
| 155 | |
| 156 | if ( str_starts_with( $url, '//' ) ) { |
| 157 | $url = $defaultProtoWithoutSlashes . $url; |
| 158 | } else { |
| 159 | // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes, |
| 160 | // otherwise leave it alone. |
| 161 | if ( $serverProto ) { |
| 162 | $url = $serverUrl . $url; |
| 163 | } else { |
| 164 | // If an HTTPS URL is synthesized from a protocol-relative Server, allow the |
| 165 | // user to override the port number (T67184) |
| 166 | if ( $defaultProto === PROTO_HTTPS && $this->httpsPort != 443 ) { |
| 167 | if ( isset( $bits['port'] ) ) { |
| 168 | throw new InvalidArgumentException( |
| 169 | 'A protocol-relative server may not contain a port number' ); |
| 170 | } |
| 171 | $url = "$defaultProtoWithoutSlashes$serverUrl:{$this->httpsPort}$url"; |
| 172 | } else { |
| 173 | $url = "$defaultProtoWithoutSlashes$serverUrl$url"; |
| 174 | } |
| 175 | } |
| 176 | } |
| 177 | } |
| 178 | |
| 179 | $bits = $this->parse( $url ); |
| 180 | |
| 181 | if ( $bits && isset( $bits['path'] ) ) { |
| 182 | $bits['path'] = self::removeDotSegments( $bits['path'] ); |
| 183 | return self::assemble( $bits ); |
| 184 | } elseif ( $bits ) { |
| 185 | # No path to expand |
| 186 | return $url; |
| 187 | } elseif ( !str_starts_with( $url, '/' ) ) { |
| 188 | # URL is a relative path |
| 189 | return self::removeDotSegments( $url ); |
| 190 | } |
| 191 | |
| 192 | # Expanded URL is not valid. |
| 193 | return null; |
| 194 | } |
| 195 | |
| 196 | /** |
| 197 | * Get the wiki's "server", i.e. the protocol and host part of the URL, with a |
| 198 | * protocol specified using a PROTO_* constant as in expand() |
| 199 | * |
| 200 | * @throws BadMethodCallException if no server was passed to the constructor |
| 201 | * @param string|int|null $proto One of the PROTO_* constants. |
| 202 | * @return ?string The URL, or null on failure |
| 203 | */ |
| 204 | public function getServer( $proto ): ?string { |
| 205 | $url = $this->expand( '/', $proto ); |
| 206 | if ( $url === null ) { |
| 207 | return null; |
| 208 | } |
| 209 | return substr( $url, 0, -1 ); |
| 210 | } |
| 211 | |
| 212 | /** |
| 213 | * Get the canonical server, i.e. the canonical protocol and host part of |
| 214 | * the wiki's URL. |
| 215 | * @return string |
| 216 | */ |
| 217 | public function getCanonicalServer(): string { |
| 218 | // @phan-suppress-next-line PhanTypeMismatchReturnNullable -- throw if unconfigured |
| 219 | return $this->canonicalServer; |
| 220 | } |
| 221 | |
| 222 | /** |
| 223 | * This function will reassemble a URL parsed with parse(). This is useful if you need to edit |
| 224 | * part of a URL and put it back together. |
| 225 | * |
| 226 | * This is the basic structure used (brackets contain keys for $urlParts): |
| 227 | * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment] |
| 228 | * |
| 229 | * @since 1.41 |
| 230 | * @param array $urlParts URL parts, as output from parse() |
| 231 | * @return string URL assembled from its component parts |
| 232 | */ |
| 233 | public static function assemble( array $urlParts ): string { |
| 234 | $result = ''; |
| 235 | |
| 236 | if ( isset( $urlParts['delimiter'] ) ) { |
| 237 | if ( isset( $urlParts['scheme'] ) ) { |
| 238 | $result .= $urlParts['scheme']; |
| 239 | } |
| 240 | |
| 241 | $result .= $urlParts['delimiter']; |
| 242 | } |
| 243 | |
| 244 | if ( isset( $urlParts['host'] ) ) { |
| 245 | if ( isset( $urlParts['user'] ) ) { |
| 246 | $result .= $urlParts['user']; |
| 247 | if ( isset( $urlParts['pass'] ) ) { |
| 248 | $result .= ':' . $urlParts['pass']; |
| 249 | } |
| 250 | $result .= '@'; |
| 251 | } |
| 252 | |
| 253 | $result .= $urlParts['host']; |
| 254 | |
| 255 | if ( isset( $urlParts['port'] ) ) { |
| 256 | $result .= ':' . $urlParts['port']; |
| 257 | } |
| 258 | } |
| 259 | |
| 260 | if ( isset( $urlParts['path'] ) ) { |
| 261 | $result .= $urlParts['path']; |
| 262 | } |
| 263 | |
| 264 | if ( isset( $urlParts['query'] ) && $urlParts['query'] !== '' ) { |
| 265 | $result .= '?' . $urlParts['query']; |
| 266 | } |
| 267 | |
| 268 | if ( isset( $urlParts['fragment'] ) ) { |
| 269 | $result .= '#' . $urlParts['fragment']; |
| 270 | } |
| 271 | |
| 272 | return $result; |
| 273 | } |
| 274 | |
| 275 | /** |
| 276 | * Remove all dot-segments in the provided URL path. For example, '/a/./b/../c/' becomes |
| 277 | * '/a/c/'. For details on the algorithm, please see RFC3986 section 5.2.4. |
| 278 | * |
| 279 | * @since 1.41 |
| 280 | * @param string $urlPath URL path, potentially containing dot-segments |
| 281 | * @return string URL path with all dot-segments removed |
| 282 | */ |
| 283 | public static function removeDotSegments( string $urlPath ): string { |
| 284 | $output = ''; |
| 285 | $inputOffset = 0; |
| 286 | $inputLength = strlen( $urlPath ); |
| 287 | |
| 288 | while ( $inputOffset < $inputLength ) { |
| 289 | $trimOutput = false; |
| 290 | if ( substr_compare( $urlPath, './', $inputOffset, 2 ) === 0 ) { |
| 291 | # Step A, remove leading "./" |
| 292 | $inputOffset += 2; |
| 293 | } elseif ( substr_compare( $urlPath, '../', $inputOffset, 3 ) === 0 ) { |
| 294 | # Step A, remove leading "../" |
| 295 | $inputOffset += 3; |
| 296 | } elseif ( $inputOffset + 2 === $inputLength && str_ends_with( $urlPath, '/.' ) ) { |
| 297 | # Step B, replace leading "/.$" with "/" |
| 298 | $inputOffset++; |
| 299 | $urlPath[$inputOffset] = '/'; |
| 300 | } elseif ( substr_compare( $urlPath, '/./', $inputOffset, 3 ) === 0 ) { |
| 301 | # Step B, replace leading "/./" with "/" |
| 302 | $inputOffset += 2; |
| 303 | } elseif ( $inputOffset + 3 === $inputLength && str_ends_with( $urlPath, '/..' ) ) { |
| 304 | # Step C, replace leading "/..$" with "/" and |
| 305 | # remove last path component in output |
| 306 | $inputOffset += 2; |
| 307 | $urlPath[$inputOffset] = '/'; |
| 308 | $trimOutput = true; |
| 309 | } elseif ( substr_compare( $urlPath, '/../', $inputOffset, 4 ) === 0 ) { |
| 310 | # Step C, replace leading "/../" with "/" and |
| 311 | # remove last path component in output |
| 312 | $inputOffset += 3; |
| 313 | $trimOutput = true; |
| 314 | } elseif ( $inputOffset + 1 === $inputLength && str_ends_with( $urlPath, '.' ) ) { |
| 315 | # Step D, remove "^.$" |
| 316 | $inputOffset++; |
| 317 | } elseif ( $inputOffset + 2 === $inputLength && str_ends_with( $urlPath, '..' ) ) { |
| 318 | # Step D, remove "^..$" |
| 319 | $inputOffset += 2; |
| 320 | } else { |
| 321 | # Step E, move leading path segment to output |
| 322 | if ( $urlPath[$inputOffset] === '/' ) { |
| 323 | $slashPos = strpos( $urlPath, '/', $inputOffset + 1 ); |
| 324 | } else { |
| 325 | $slashPos = strpos( $urlPath, '/', $inputOffset ); |
| 326 | } |
| 327 | if ( $slashPos === false ) { |
| 328 | $output .= substr( $urlPath, $inputOffset ); |
| 329 | $inputOffset = $inputLength; |
| 330 | } else { |
| 331 | $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset ); |
| 332 | $inputOffset += $slashPos - $inputOffset; |
| 333 | } |
| 334 | } |
| 335 | |
| 336 | if ( $trimOutput ) { |
| 337 | $slashPos = strrpos( $output, '/' ); |
| 338 | if ( $slashPos === false ) { |
| 339 | $output = ''; |
| 340 | } else { |
| 341 | $output = substr( $output, 0, $slashPos ); |
| 342 | } |
| 343 | } |
| 344 | } |
| 345 | |
| 346 | return $output; |
| 347 | } |
| 348 | |
| 349 | /** |
| 350 | * Returns a partial regular expression of recognized URL protocols, e.g. "http:\/\/|https:\/\/" |
| 351 | * |
| 352 | * @return string |
| 353 | */ |
| 354 | public function validProtocols(): string { |
| 355 | $this->validProtocolsCache ??= $this->validProtocolsInternal( true ); |
| 356 | return $this->validProtocolsCache; |
| 357 | } |
| 358 | |
| 359 | /** |
| 360 | * Like validProtocols(), but excludes '//' from the protocol list. Use this if you need a |
| 361 | * regex that matches all URL protocols but does not match protocol-relative URLs |
| 362 | * |
| 363 | * @return string |
| 364 | */ |
| 365 | public function validAbsoluteProtocols(): string { |
| 366 | $this->validAbsoluteProtocolsCache ??= $this->validProtocolsInternal( false ); |
| 367 | return $this->validAbsoluteProtocolsCache; |
| 368 | } |
| 369 | |
| 370 | /** |
| 371 | * Returns a partial regular expression of URL protocols, e.g. "http:\/\/|https:\/\/" |
| 372 | * |
| 373 | * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list. |
| 374 | * @return string |
| 375 | */ |
| 376 | private function validProtocolsInternal( bool $includeProtocolRelative ): string { |
| 377 | if ( !is_array( $this->validProtocols ) ) { |
| 378 | MWDebug::deprecated( '$wgUrlProtocols that is not an array', '1.39' ); |
| 379 | return (string)$this->validProtocols; |
| 380 | } |
| 381 | |
| 382 | $protocols = []; |
| 383 | foreach ( $this->validProtocols as $protocol ) { |
| 384 | // Filter out '//' if !$includeProtocolRelative |
| 385 | if ( $includeProtocolRelative || $protocol !== '//' ) { |
| 386 | $protocols[] = preg_quote( $protocol, '/' ); |
| 387 | } |
| 388 | } |
| 389 | |
| 390 | return implode( '|', $protocols ); |
| 391 | } |
| 392 | |
| 393 | /** |
| 394 | * Advanced and configurable version of parse_url(). |
| 395 | * |
| 396 | * 1) Add a "delimiter" element to the array, which helps permits to blindly re-assemble |
| 397 | * any URL regardless of protocol, including those that don't use `://`, |
| 398 | * such as "mailto:" and "news:". |
| 399 | * 2) Reject URLs with protocols not in $wgUrlProtocols. |
| 400 | * 3) Reject relative or incomplete URLs that parse_url would return a partial array for. |
| 401 | * |
| 402 | * If all you need is to extract parts of an HTTP or HTTPS URL (i.e. not specific to |
| 403 | * site-configurable extra protocols, or user input) then `parse_url()` can be used |
| 404 | * directly instead. |
| 405 | * |
| 406 | * @param string $url A URL to parse |
| 407 | * @return ?string[] Bits of the URL in an associative array, or null on failure. |
| 408 | * Possible fields: |
| 409 | * - scheme: URI scheme (protocol), e.g. 'http', 'mailto'. Lowercase, always present, but can |
| 410 | * be an empty string for protocol-relative URLs. |
| 411 | * - delimiter: either '://', ':' or '//'. Always present. |
| 412 | * - host: domain name / IP. Always present, but could be an empty string, e.g. for file: URLs. |
| 413 | * - port: port number. Will be missing when port is not explicitly specified. |
| 414 | * - user: user name, e.g. for HTTP Basic auth URLs such as http://user:pass@example.com/ |
| 415 | * Missing when there is no username. |
| 416 | * - pass: password, same as above. |
| 417 | * - path: path including the leading /. Will be missing when empty (e.g. 'http://example.com') |
| 418 | * - query: query string (as a string; see wfCgiToArray() for parsing it), can be missing. |
| 419 | * - fragment: the part after #, can be missing. |
| 420 | */ |
| 421 | public function parse( string $url ): ?array { |
| 422 | // Protocol-relative URLs are handled really badly by parse_url(). It's so bad that the |
| 423 | // easiest way to handle them is to just prepend 'http:' and strip the protocol out later. |
| 424 | $wasRelative = str_starts_with( $url, '//' ); |
| 425 | if ( $wasRelative ) { |
| 426 | $url = "http:$url"; |
| 427 | } |
| 428 | $bits = parse_url( $url ); |
| 429 | // parse_url() returns an array without scheme for invalid URLs, e.g. |
| 430 | // parse_url("something bad://example") == [ 'path' => 'something bad://example' ] |
| 431 | if ( !$bits || !isset( $bits['scheme'] ) ) { |
| 432 | return null; |
| 433 | } |
| 434 | |
| 435 | // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase. |
| 436 | $bits['scheme'] = strtolower( $bits['scheme'] ); |
| 437 | $bits['host'] ??= ''; |
| 438 | |
| 439 | // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it |
| 440 | if ( in_array( $bits['scheme'] . '://', $this->validProtocols ) ) { |
| 441 | $bits['delimiter'] = '://'; |
| 442 | } elseif ( in_array( $bits['scheme'] . ':', $this->validProtocols ) ) { |
| 443 | $bits['delimiter'] = ':'; |
| 444 | } else { |
| 445 | return null; |
| 446 | } |
| 447 | |
| 448 | /* parse_url loses the third / for file:///c:/ urls */ |
| 449 | if ( $bits['scheme'] === 'file' && isset( $bits['path'] ) && !str_starts_with( $bits['path'], '/' ) ) { |
| 450 | $bits['path'] = '/' . $bits['path']; |
| 451 | } |
| 452 | |
| 453 | // If the URL was protocol-relative, fix scheme and delimiter |
| 454 | if ( $wasRelative ) { |
| 455 | $bits['scheme'] = ''; |
| 456 | $bits['delimiter'] = '//'; |
| 457 | } |
| 458 | return $bits; |
| 459 | } |
| 460 | |
| 461 | /** |
| 462 | * Take a URL, make sure it's expanded to fully qualified, and replace any encoded non-ASCII |
| 463 | * Unicode characters with their UTF-8 original forms for more compact display and legibility |
| 464 | * for local audiences. |
| 465 | * |
| 466 | * @todo handle punycode domains too |
| 467 | * |
| 468 | * @throws BadMethodCallException if no server was passed to the constructor |
| 469 | * @param string $url |
| 470 | * @return ?string |
| 471 | */ |
| 472 | public function expandIRI( string $url ): ?string { |
| 473 | $expanded = $this->expand( $url ); |
| 474 | if ( $expanded === null ) { |
| 475 | return null; |
| 476 | } |
| 477 | return preg_replace_callback( |
| 478 | '/(?:%[89A-F][0-9A-F])+/i', |
| 479 | static fn ( $m ) => urldecode( $m[0] ), |
| 480 | $expanded |
| 481 | ); |
| 482 | } |
| 483 | |
| 484 | /** |
| 485 | * Check whether a given URL has a domain that occurs in a given set of domains |
| 486 | * |
| 487 | * @throws BadMethodCallException if no server was passed to the constructor |
| 488 | * @param string $url |
| 489 | * @param array $domains Array of domains (strings) |
| 490 | * @return bool True if the host part of $url ends in one of the strings in $domains |
| 491 | */ |
| 492 | public function matchesDomainList( string $url, array $domains ): bool { |
| 493 | $bits = $this->parse( $url ); |
| 494 | if ( is_array( $bits ) && isset( $bits['host'] ) ) { |
| 495 | $host = '.' . $bits['host']; |
| 496 | foreach ( $domains as $domain ) { |
| 497 | $domain = '.' . $domain; |
| 498 | if ( str_ends_with( $host, $domain ) ) { |
| 499 | return true; |
| 500 | } |
| 501 | } |
| 502 | } |
| 503 | return false; |
| 504 | } |
| 505 | } |