projects / github / wallabag / wallabag.git / blame
| Commit | Line | Data |
|---|---|---|
| d4949327 NL |
1 | <?php\r |
| 2 | \r | |
| 3 | /**\r | |
| 4 | * HTML Purifier's internal representation of a URI.\r | |
| 5 | * @note\r | |
| 6 | * Internal data-structures are completely escaped. If the data needs\r | |
| 7 | * to be used in a non-URI context (which is very unlikely), be sure\r | |
| 8 | * to decode it first. The URI may not necessarily be well-formed until\r | |
| 9 | * validate() is called.\r | |
| 10 | */\r | |
| 11 | class HTMLPurifier_URI\r | |
| 12 | {\r | |
| 13 | /**\r | |
| 14 | * @type string\r | |
| 15 | */\r | |
| 16 | public $scheme;\r | |
| 17 | \r | |
| 18 | /**\r | |
| 19 | * @type string\r | |
| 20 | */\r | |
| 21 | public $userinfo;\r | |
| 22 | \r | |
| 23 | /**\r | |
| 24 | * @type string\r | |
| 25 | */\r | |
| 26 | public $host;\r | |
| 27 | \r | |
| 28 | /**\r | |
| 29 | * @type int\r | |
| 30 | */\r | |
| 31 | public $port;\r | |
| 32 | \r | |
| 33 | /**\r | |
| 34 | * @type string\r | |
| 35 | */\r | |
| 36 | public $path;\r | |
| 37 | \r | |
| 38 | /**\r | |
| 39 | * @type string\r | |
| 40 | */\r | |
| 41 | public $query;\r | |
| 42 | \r | |
| 43 | /**\r | |
| 44 | * @type string\r | |
| 45 | */\r | |
| 46 | public $fragment;\r | |
| 47 | \r | |
| 48 | /**\r | |
| 49 | * @param string $scheme\r | |
| 50 | * @param string $userinfo\r | |
| 51 | * @param string $host\r | |
| 52 | * @param int $port\r | |
| 53 | * @param string $path\r | |
| 54 | * @param string $query\r | |
| 55 | * @param string $fragment\r | |
| 56 | * @note Automatically normalizes scheme and port\r | |
| 57 | */\r | |
| 58 | public function __construct($scheme, $userinfo, $host, $port, $path, $query, $fragment)\r | |
| 59 | {\r | |
| 60 | $this->scheme = is_null($scheme) || ctype_lower($scheme) ? $scheme : strtolower($scheme);\r | |
| 61 | $this->userinfo = $userinfo;\r | |
| 62 | $this->host = $host;\r | |
| 63 | $this->port = is_null($port) ? $port : (int)$port;\r | |
| 64 | $this->path = $path;\r | |
| 65 | $this->query = $query;\r | |
| 66 | $this->fragment = $fragment;\r | |
| 67 | }\r | |
| 68 | \r | |
| 69 | /**\r | |
| 70 | * Retrieves a scheme object corresponding to the URI's scheme/default\r | |
| 71 | * @param HTMLPurifier_Config $config\r | |
| 72 | * @param HTMLPurifier_Context $context\r | |
| 73 | * @return HTMLPurifier_URIScheme Scheme object appropriate for validating this URI\r | |
| 74 | */\r | |
| 75 | public function getSchemeObj($config, $context)\r | |
| 76 | {\r | |
| 77 | $registry = HTMLPurifier_URISchemeRegistry::instance();\r | |
| 78 | if ($this->scheme !== null) {\r | |
| 79 | $scheme_obj = $registry->getScheme($this->scheme, $config, $context);\r | |
| 80 | if (!$scheme_obj) {\r | |
| 81 | return false;\r | |
| 82 | } // invalid scheme, clean it out\r | |
| 83 | } else {\r | |
| 84 | // no scheme: retrieve the default one\r | |
| 85 | $def = $config->getDefinition('URI');\r | |
| 86 | $scheme_obj = $def->getDefaultScheme($config, $context);\r | |
| 87 | if (!$scheme_obj) {\r | |
| 88 | // something funky happened to the default scheme object\r | |
| 89 | trigger_error(\r | |
| 90 | 'Default scheme object "' . $def->defaultScheme . '" was not readable',\r | |
| 91 | E_USER_WARNING\r | |
| 92 | );\r | |
| 93 | return false;\r | |
| 94 | }\r | |
| 95 | }\r | |
| 96 | return $scheme_obj;\r | |
| 97 | }\r | |
| 98 | \r | |
| 99 | /**\r | |
| 100 | * Generic validation method applicable for all schemes. May modify\r | |
| 101 | * this URI in order to get it into a compliant form.\r | |
| 102 | * @param HTMLPurifier_Config $config\r | |
| 103 | * @param HTMLPurifier_Context $context\r | |
| 104 | * @return bool True if validation/filtering succeeds, false if failure\r | |
| 105 | */\r | |
| 106 | public function validate($config, $context)\r | |
| 107 | {\r | |
| 108 | // ABNF definitions from RFC 3986\r | |
| 109 | $chars_sub_delims = '!$&\'()*+,;=';\r | |
| 110 | $chars_gen_delims = ':/?#[]@';\r | |
| 111 | $chars_pchar = $chars_sub_delims . ':@';\r | |
| 112 | \r | |
| 113 | // validate host\r | |
| 114 | if (!is_null($this->host)) {\r | |
| 115 | $host_def = new HTMLPurifier_AttrDef_URI_Host();\r | |
| 116 | $this->host = $host_def->validate($this->host, $config, $context);\r | |
| 117 | if ($this->host === false) {\r | |
| 118 | $this->host = null;\r | |
| 119 | }\r | |
| 120 | }\r | |
| 121 | \r | |
| 122 | // validate scheme\r | |
| 123 | // NOTE: It's not appropriate to check whether or not this\r | |
| 124 | // scheme is in our registry, since a URIFilter may convert a\r | |
| 125 | // URI that we don't allow into one we do. So instead, we just\r | |
| 126 | // check if the scheme can be dropped because there is no host\r | |
| 127 | // and it is our default scheme.\r | |
| 128 | if (!is_null($this->scheme) && is_null($this->host) || $this->host === '') {\r | |
| 129 | // support for relative paths is pretty abysmal when the\r | |
| 130 | // scheme is present, so axe it when possible\r | |
| 131 | $def = $config->getDefinition('URI');\r | |
| 132 | if ($def->defaultScheme === $this->scheme) {\r | |
| 133 | $this->scheme = null;\r | |
| 134 | }\r | |
| 135 | }\r | |
| 136 | \r | |
| 137 | // validate username\r | |
| 138 | if (!is_null($this->userinfo)) {\r | |
| 139 | $encoder = new HTMLPurifier_PercentEncoder($chars_sub_delims . ':');\r | |
| 140 | $this->userinfo = $encoder->encode($this->userinfo);\r | |
| 141 | }\r | |
| 142 | \r | |
| 143 | // validate port\r | |
| 144 | if (!is_null($this->port)) {\r | |
| 145 | if ($this->port < 1 || $this->port > 65535) {\r | |
| 146 | $this->port = null;\r | |
| 147 | }\r | |
| 148 | }\r | |
| 149 | \r | |
| 150 | // validate path\r | |
| 151 | $segments_encoder = new HTMLPurifier_PercentEncoder($chars_pchar . '/');\r | |
| 152 | if (!is_null($this->host)) { // this catches $this->host === ''\r | |
| 153 | // path-abempty (hier and relative)\r | |
| 154 | // http://www.example.com/my/path\r | |
| 155 | // //www.example.com/my/path (looks odd, but works, and\r | |
| 156 | // recognized by most browsers)\r | |
| 157 | // (this set is valid or invalid on a scheme by scheme\r | |
| 158 | // basis, so we'll deal with it later)\r | |
| 159 | // file:///my/path\r | |
| 160 | // ///my/path\r | |
| 161 | $this->path = $segments_encoder->encode($this->path);\r | |
| 162 | } elseif ($this->path !== '') {\r | |
| 163 | if ($this->path[0] === '/') {\r | |
| 164 | // path-absolute (hier and relative)\r | |
| 165 | // http:/my/path\r | |
| 166 | // /my/path\r | |
| 167 | if (strlen($this->path) >= 2 && $this->path[1] === '/') {\r | |
| 168 | // This could happen if both the host gets stripped\r | |
| 169 | // out\r | |
| 170 | // http://my/path\r | |
| 171 | // //my/path\r | |
| 172 | $this->path = '';\r | |
| 173 | } else {\r | |
| 174 | $this->path = $segments_encoder->encode($this->path);\r | |
| 175 | }\r | |
| 176 | } elseif (!is_null($this->scheme)) {\r | |
| 177 | // path-rootless (hier)\r | |
| 178 | // http:my/path\r | |
| 179 | // Short circuit evaluation means we don't need to check nz\r | |
| 180 | $this->path = $segments_encoder->encode($this->path);\r | |
| 181 | } else {\r | |
| 182 | // path-noscheme (relative)\r | |
| 183 | // my/path\r | |
| 184 | // (once again, not checking nz)\r | |
| 185 | $segment_nc_encoder = new HTMLPurifier_PercentEncoder($chars_sub_delims . '@');\r | |
| 186 | $c = strpos($this->path, '/');\r | |
| 187 | if ($c !== false) {\r | |
| 188 | $this->path =\r | |
| 189 | $segment_nc_encoder->encode(substr($this->path, 0, $c)) .\r | |
| 190 | $segments_encoder->encode(substr($this->path, $c));\r | |
| 191 | } else {\r | |
| 192 | $this->path = $segment_nc_encoder->encode($this->path);\r | |
| 193 | }\r | |
| 194 | }\r | |
| 195 | } else {\r | |
| 196 | // path-empty (hier and relative)\r | |
| 197 | $this->path = ''; // just to be safe\r | |
| 198 | }\r | |
| 199 | \r | |
| 200 | // qf = query and fragment\r | |
| 201 | $qf_encoder = new HTMLPurifier_PercentEncoder($chars_pchar . '/?');\r | |
| 202 | \r | |
| 203 | if (!is_null($this->query)) {\r | |
| 204 | $this->query = $qf_encoder->encode($this->query);\r | |
| 205 | }\r | |
| 206 | \r | |
| 207 | if (!is_null($this->fragment)) {\r | |
| 208 | $this->fragment = $qf_encoder->encode($this->fragment);\r | |
| 209 | }\r | |
| 210 | return true;\r | |
| 211 | }\r | |
| 212 | \r | |
| 213 | /**\r | |
| 214 | * Convert URI back to string\r | |
| 215 | * @return string URI appropriate for output\r | |
| 216 | */\r | |
| 217 | public function toString()\r | |
| 218 | {\r | |
| 219 | // reconstruct authority\r | |
| 220 | $authority = null;\r | |
| 221 | // there is a rendering difference between a null authority\r | |
| 222 | // (http:foo-bar) and an empty string authority\r | |
| 223 | // (http:///foo-bar).\r | |
| 224 | if (!is_null($this->host)) {\r | |
| 225 | $authority = '';\r | |
| 226 | if (!is_null($this->userinfo)) {\r | |
| 227 | $authority .= $this->userinfo . '@';\r | |
| 228 | }\r | |
| 229 | $authority .= $this->host;\r | |
| 230 | if (!is_null($this->port)) {\r | |
| 231 | $authority .= ':' . $this->port;\r | |
| 232 | }\r | |
| 233 | }\r | |
| 234 | \r | |
| 235 | // Reconstruct the result\r | |
| 236 | // One might wonder about parsing quirks from browsers after\r | |
| 237 | // this reconstruction. Unfortunately, parsing behavior depends\r | |
| 238 | // on what *scheme* was employed (file:///foo is handled *very*\r | |
| 239 | // differently than http:///foo), so unfortunately we have to\r | |
| 240 | // defer to the schemes to do the right thing.\r | |
| 241 | $result = '';\r | |
| 242 | if (!is_null($this->scheme)) {\r | |
| 243 | $result .= $this->scheme . ':';\r | |
| 244 | }\r | |
| 245 | if (!is_null($authority)) {\r | |
| 246 | $result .= '//' . $authority;\r | |
| 247 | }\r | |
| 248 | $result .= $this->path;\r | |
| 249 | if (!is_null($this->query)) {\r | |
| 250 | $result .= '?' . $this->query;\r | |
| 251 | }\r | |
| 252 | if (!is_null($this->fragment)) {\r | |
| 253 | $result .= '#' . $this->fragment;\r | |
| 254 | }\r | |
| 255 | \r | |
| 256 | return $result;\r | |
| 257 | }\r | |
| 258 | \r | |
| 259 | /**\r | |
| 260 | * Returns true if this URL might be considered a 'local' URL given\r | |
| 261 | * the current context. This is true when the host is null, or\r | |
| 262 | * when it matches the host supplied to the configuration.\r | |
| 263 | *\r | |
| 264 | * Note that this does not do any scheme checking, so it is mostly\r | |
| 265 | * only appropriate for metadata that doesn't care about protocol\r | |
| 266 | * security. isBenign is probably what you actually want.\r | |
| 267 | * @param HTMLPurifier_Config $config\r | |
| 268 | * @param HTMLPurifier_Context $context\r | |
| 269 | * @return bool\r | |
| 270 | */\r | |
| 271 | public function isLocal($config, $context)\r | |
| 272 | {\r | |
| 273 | if ($this->host === null) {\r | |
| 274 | return true;\r | |
| 275 | }\r | |
| 276 | $uri_def = $config->getDefinition('URI');\r | |
| 277 | if ($uri_def->host === $this->host) {\r | |
| 278 | return true;\r | |
| 279 | }\r | |
| 280 | return false;\r | |
| 281 | }\r | |
| 282 | \r | |
| 283 | /**\r | |
| 284 | * Returns true if this URL should be considered a 'benign' URL,\r | |
| 285 | * that is:\r | |
| 286 | *\r | |
| 287 | * - It is a local URL (isLocal), and\r | |
| 288 | * - It has a equal or better level of security\r | |
| 289 | * @param HTMLPurifier_Config $config\r | |
| 290 | * @param HTMLPurifier_Context $context\r | |
| 291 | * @return bool\r | |
| 292 | */\r | |
| 293 | public function isBenign($config, $context)\r | |
| 294 | {\r | |
| 295 | if (!$this->isLocal($config, $context)) {\r | |
| 296 | return false;\r | |
| 297 | }\r | |
| 298 | \r | |
| 299 | $scheme_obj = $this->getSchemeObj($config, $context);\r | |
| 300 | if (!$scheme_obj) {\r | |
| 301 | return false;\r | |
| 302 | } // conservative approach\r | |
| 303 | \r | |
| 304 | $current_scheme_obj = $config->getDefinition('URI')->getDefaultScheme($config, $context);\r | |
| 305 | if ($current_scheme_obj->secure) {\r | |
| 306 | if (!$scheme_obj->secure) {\r | |
| 307 | return false;\r | |
| 308 | }\r | |
| 309 | }\r | |
| 310 | return true;\r | |
| 311 | }\r | |
| 312 | }\r | |
| 313 | \r | |
| 314 | // vim: et sw=4 sts=4\r |