| 1 |
// Copyright 2016 The Closure Library Authors. All Rights Reserved. |
| 2 |
// |
| 3 |
// Licensed under the Apache License, Version 2.0 (the "License"); |
| 4 |
// you may not use this file except in compliance with the License. |
| 5 |
// You may obtain a copy of the License at |
| 6 |
// |
| 7 |
// http://www.apache.org/licenses/LICENSE-2.0 |
| 8 |
// |
| 9 |
// Unless required by applicable law or agreed to in writing, software |
| 10 |
// distributed under the License is distributed on an "AS-IS" BASIS, |
| 11 |
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 12 |
// See the License for the specific language governing permissions and |
| 13 |
// limitations under the License. |
| 14 |
|
| 15 |
|
| 16 |
/** |
| 17 |
* @fileoverview An HTML sanitizer that can satisfy a variety of security |
| 18 |
* policies. |
| 19 |
* |
| 20 |
* This package provides html sanitizing functions. It does not enforce string |
| 21 |
* to string conversion, instead returning a dom-like element when possible. |
| 22 |
* |
| 23 |
* Examples of usage of the static {@code goog.goog.html.sanitizer.sanitize}: |
| 24 |
* <pre> |
| 25 |
* var safeHtml = goog.html.sanitizer.sanitize('<script src="xss.js" />'); |
| 26 |
* goog.dom.safe.setInnerHtml(el, safeHtml); |
| 27 |
* </pre> |
| 28 |
* |
| 29 |
* @supported IE 10+, Chrome 26+, Firefox 22+, Safari 7.1+, Opera 15+ |
| 30 |
*/ |
| 31 |
|
| 32 |
goog.provide('goog.html.sanitizer.HtmlSanitizer'); |
| 33 |
goog.provide('goog.html.sanitizer.HtmlSanitizer.Builder'); |
| 34 |
goog.provide('goog.html.sanitizer.HtmlSanitizerAttributePolicy'); |
| 35 |
goog.provide('goog.html.sanitizer.HtmlSanitizerPolicy'); |
| 36 |
goog.provide('goog.html.sanitizer.HtmlSanitizerPolicyContext'); |
| 37 |
goog.provide('goog.html.sanitizer.HtmlSanitizerPolicyHints'); |
| 38 |
goog.provide('goog.html.sanitizer.HtmlSanitizerUrlPolicy'); |
| 39 |
|
| 40 |
goog.require('goog.array'); |
| 41 |
goog.require('goog.asserts'); |
| 42 |
goog.require('goog.dom'); |
| 43 |
goog.require('goog.dom.NodeType'); |
| 44 |
goog.require('goog.dom.TagName'); |
| 45 |
goog.require('goog.functions'); |
| 46 |
goog.require('goog.html.SafeHtml'); |
| 47 |
goog.require('goog.html.SafeStyle'); |
| 48 |
goog.require('goog.html.SafeStyleSheet'); |
| 49 |
goog.require('goog.html.SafeUrl'); |
| 50 |
goog.require('goog.html.sanitizer.AttributeSanitizedWhitelist'); |
| 51 |
goog.require('goog.html.sanitizer.AttributeWhitelist'); |
| 52 |
goog.require('goog.html.sanitizer.CssSanitizer'); |
| 53 |
goog.require('goog.html.sanitizer.ElementWeakMap'); |
| 54 |
goog.require('goog.html.sanitizer.TagBlacklist'); |
| 55 |
goog.require('goog.html.sanitizer.TagWhitelist'); |
| 56 |
goog.require('goog.html.sanitizer.noclobber'); |
| 57 |
goog.require('goog.html.uncheckedconversions'); |
| 58 |
goog.require('goog.object'); |
| 59 |
goog.require('goog.string'); |
| 60 |
goog.require('goog.string.Const'); |
| 61 |
goog.require('goog.userAgent'); |
| 62 |
|
| 63 |
|
| 64 |
/** |
| 65 |
* Type for optional hints to policy handler functions. |
| 66 |
* @typedef {{ |
| 67 |
* tagName: (string|undefined), |
| 68 |
* attributeName: (string|undefined), |
| 69 |
* cssProperty: (string|undefined) |
| 70 |
* }} |
| 71 |
*/ |
| 72 |
goog.html.sanitizer.HtmlSanitizerPolicyHints; |
| 73 |
|
| 74 |
|
| 75 |
/** |
| 76 |
* Type for optional context objects to the policy handler functions. |
| 77 |
* @typedef {{ |
| 78 |
* cssStyle: (?CSSStyleDeclaration|undefined) |
| 79 |
* }} |
| 80 |
*/ |
| 81 |
goog.html.sanitizer.HtmlSanitizerPolicyContext; |
| 82 |
|
| 83 |
|
| 84 |
/** |
| 85 |
* Type for a policy function. |
| 86 |
* @typedef {function(string, goog.html.sanitizer.HtmlSanitizerPolicyHints=, |
| 87 |
* goog.html.sanitizer.HtmlSanitizerPolicyContext=, |
| 88 |
* (function(string, ?=, ?=, ?=):?string)=):?string} |
| 89 |
*/ |
| 90 |
goog.html.sanitizer.HtmlSanitizerPolicy; |
| 91 |
|
| 92 |
|
| 93 |
/** |
| 94 |
* Type for a URL policy function. |
| 95 |
* |
| 96 |
* @typedef {function(string, !goog.html.sanitizer.HtmlSanitizerPolicyHints=): |
| 97 |
* ?goog.html.SafeUrl} |
| 98 |
*/ |
| 99 |
goog.html.sanitizer.HtmlSanitizerUrlPolicy; |
| 100 |
|
| 101 |
|
| 102 |
/** |
| 103 |
* Type for attribute policy configuration. |
| 104 |
* @typedef {{ |
| 105 |
* tagName: string, |
| 106 |
* attributeName: string, |
| 107 |
* policy: ?goog.html.sanitizer.HtmlSanitizerPolicy |
| 108 |
* }} |
| 109 |
*/ |
| 110 |
goog.html.sanitizer.HtmlSanitizerAttributePolicy; |
| 111 |
|
| 112 |
|
| 113 |
/** |
| 114 |
* Whether the HTML sanitizer is supported. For now mainly exclude |
| 115 |
* IE9 or below where we know the sanitizer is insecure. |
| 116 |
* @private @const {boolean} |
| 117 |
*/ |
| 118 |
goog.html.sanitizer.HTML_SANITIZER_SUPPORTED_ = |
| 119 |
!goog.userAgent.IE || document.documentMode >= 10; |
| 120 |
|
| 121 |
|
| 122 |
/** |
| 123 |
* Whether the template tag is supported. |
| 124 |
* @package @const {boolean} |
| 125 |
*/ |
| 126 |
goog.html.sanitizer.HTML_SANITIZER_TEMPLATE_SUPPORTED = |
| 127 |
!goog.userAgent.IE || document.documentMode == null; |
| 128 |
|
| 129 |
|
| 130 |
/** |
| 131 |
* Prefix used by all internal html sanitizer booking properties. |
| 132 |
* @private @const {string} |
| 133 |
*/ |
| 134 |
goog.html.sanitizer.HTML_SANITIZER_BOOKKEEPING_PREFIX_ = 'data-sanitizer-'; |
| 135 |
|
| 136 |
|
| 137 |
/** |
| 138 |
* Temporary attribute name in which html sanitizer uses for bookkeeping. |
| 139 |
* @private @const {string} |
| 140 |
*/ |
| 141 |
goog.html.sanitizer.HTML_SANITIZER_BOOKKEEPING_ATTR_NAME_ = |
| 142 |
goog.html.sanitizer.HTML_SANITIZER_BOOKKEEPING_PREFIX_ + 'elem-num'; |
| 143 |
|
| 144 |
|
| 145 |
/** |
| 146 |
* Attribute name added to span tags that replace unknown tags. The value of |
| 147 |
* this attribute is the name of the tag before the sanitization occurred. |
| 148 |
* @private @const {string} |
| 149 |
*/ |
| 150 |
goog.html.sanitizer.HTML_SANITIZER_SANITIZED_ATTR_NAME_ = |
| 151 |
goog.html.sanitizer.HTML_SANITIZER_BOOKKEEPING_PREFIX_ + 'original-tag'; |
| 152 |
|
| 153 |
|
| 154 |
/** |
| 155 |
* Attribute name added to blacklisted tags to then filter them from the output. |
| 156 |
* @private @const {string} |
| 157 |
*/ |
| 158 |
goog.html.sanitizer.HTML_SANITIZER_BLACKLISTED_TAG_ = |
| 159 |
goog.html.sanitizer.HTML_SANITIZER_BOOKKEEPING_PREFIX_ + 'blacklisted-tag'; |
| 160 |
|
| 161 |
|
| 162 |
/** |
| 163 |
* Special value for the STYLE container ID, which makes the sanitizer choose |
| 164 |
* a new random ID on each call to {@link sanitize}. |
| 165 |
* @private @const {string} |
| 166 |
*/ |
| 167 |
goog.html.sanitizer.RANDOM_CONTAINER_ = '*'; |
| 168 |
|
| 169 |
|
| 170 |
/** |
| 171 |
* Creates an HTML sanitizer. |
| 172 |
* @param {!goog.html.sanitizer.HtmlSanitizer.Builder=} opt_builder |
| 173 |
* @final @constructor @struct |
| 174 |
*/ |
| 175 |
goog.html.sanitizer.HtmlSanitizer = function(opt_builder) { |
| 176 |
var builder = opt_builder || new goog.html.sanitizer.HtmlSanitizer.Builder(); |
| 177 |
|
| 178 |
builder.installPolicies_(); |
| 179 |
|
| 180 |
/** @private @const {boolean} */ |
| 181 |
this.shouldSanitizeTemplateContents_ = |
| 182 |
builder.shouldSanitizeTemplateContents_; |
| 183 |
|
| 184 |
/** |
| 185 |
* @private @const {!Object<string, !goog.html.sanitizer.HtmlSanitizerPolicy>} |
| 186 |
*/ |
| 187 |
this.attributeHandlers_ = goog.object.clone(builder.attributeWhitelist_); |
| 188 |
|
| 189 |
/** @private @const {!Object<string, boolean>} */ |
| 190 |
this.tagBlacklist_ = goog.object.clone(builder.tagBlacklist_); |
| 191 |
|
| 192 |
/** @private @const {!Object<string, boolean>} */ |
| 193 |
this.tagWhitelist_ = goog.object.clone(builder.tagWhitelist_); |
| 194 |
|
| 195 |
/** @private @const {boolean} */ |
| 196 |
this.shouldAddOriginalTagNames_ = builder.shouldAddOriginalTagNames_; |
| 197 |
|
| 198 |
// Add whitelist data-* attributes from the builder to the attributeHandlers |
| 199 |
// with a default cleanUpAttribute function. data-* attributes are inert as |
| 200 |
// per HTML5 specs, so not much sanitization needed. |
| 201 |
goog.array.forEach(builder.dataAttributeWhitelist_, function(dataAttr) { |
| 202 |
goog.asserts.assert(goog.string.startsWith(dataAttr, 'data-')); |
| 203 |
goog.asserts.assert(!goog.string.startsWith( |
| 204 |
dataAttr, goog.html.sanitizer.HTML_SANITIZER_BOOKKEEPING_PREFIX_)); |
| 205 |
|
| 206 |
this.attributeHandlers_['* ' + dataAttr.toUpperCase()] = |
| 207 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ ( |
| 208 |
goog.html.sanitizer.HtmlSanitizer.cleanUpAttribute_); |
| 209 |
}, this); |
| 210 |
|
| 211 |
/** @private @const {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} */ |
| 212 |
this.networkRequestUrlPolicy_ = builder.networkRequestUrlPolicy_; |
| 213 |
|
| 214 |
/** @private @const {?string} */ |
| 215 |
this.styleContainer_ = builder.styleContainer_; |
| 216 |
|
| 217 |
/** @private @const {boolean} */ |
| 218 |
this.inlineStyleRules_ = builder.inlineStyleRules_; |
| 219 |
}; |
| 220 |
|
| 221 |
|
| 222 |
/** |
| 223 |
* Transforms a {@link HtmlSanitizerUrlPolicy} into a |
| 224 |
* {@link HtmlSanitizerPolicy} by returning a wrapper that calls the {@link |
| 225 |
* HtmlSanitizerUrlPolicy} with the required arguments and unwraps the returned |
| 226 |
* {@link SafeUrl}. This is necessary because internally the sanitizer works |
| 227 |
* with {@HtmlSanitizerPolicy} to sanitize attributes, but its public API must |
| 228 |
* use {@HtmlSanitizerUrlPolicy} to ensure that callers do not violate SafeHtml |
| 229 |
* invariants in their custom handlers. |
| 230 |
* @param {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} urlPolicy |
| 231 |
* @return {!goog.html.sanitizer.HtmlSanitizerPolicy} |
| 232 |
* @private |
| 233 |
*/ |
| 234 |
goog.html.sanitizer.HtmlSanitizer.wrapUrlPolicy_ = function(urlPolicy) { |
| 235 |
return /** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ (function( |
| 236 |
url, policyHints) { |
| 237 |
var trimmed = |
| 238 |
goog.html.sanitizer.HtmlSanitizer.cleanUpAttribute_(url, policyHints); |
| 239 |
var safeUrl = urlPolicy(trimmed, policyHints); |
| 240 |
if (safeUrl && |
| 241 |
goog.html.SafeUrl.unwrap(safeUrl) != |
| 242 |
goog.html.SafeUrl.INNOCUOUS_STRING) { |
| 243 |
return goog.html.SafeUrl.unwrap(safeUrl); |
| 244 |
} else { |
| 245 |
return null; |
| 246 |
} |
| 247 |
}); |
| 248 |
}; |
| 249 |
|
| 250 |
|
| 251 |
|
| 252 |
/** |
| 253 |
* The builder for the HTML Sanitizer. All methods except build return this. |
| 254 |
* @final @constructor @struct |
| 255 |
*/ |
| 256 |
goog.html.sanitizer.HtmlSanitizer.Builder = function() { |
| 257 |
/** |
| 258 |
* A set of attribute sanitization functions. Default built-in handlers are |
| 259 |
* all tag-agnostic by design. Note that some attributes behave differently |
| 260 |
* when attached to different nodes (for example, the href attribute will |
| 261 |
* generally not make a network request, but <link href=""> does), and |
| 262 |
* so when necessary a tag-specific handler can be used to override a |
| 263 |
* tag-agnostic one. |
| 264 |
* @private {!Object<string, !goog.html.sanitizer.HtmlSanitizerPolicy>} |
| 265 |
*/ |
| 266 |
this.attributeWhitelist_ = {}; |
| 267 |
goog.array.forEach( |
| 268 |
[ |
| 269 |
goog.html.sanitizer.AttributeWhitelist, |
| 270 |
goog.html.sanitizer.AttributeSanitizedWhitelist |
| 271 |
], |
| 272 |
function(wl) { |
| 273 |
goog.array.forEach(goog.object.getKeys(wl), function(attr) { |
| 274 |
this.attributeWhitelist_[attr] = |
| 275 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ |
| 276 |
(goog.html.sanitizer.HtmlSanitizer.cleanUpAttribute_); |
| 277 |
}, this); |
| 278 |
}, |
| 279 |
this); |
| 280 |
|
| 281 |
/** |
| 282 |
* A set of attribute handlers that should not inherit their default policy |
| 283 |
* during build(). |
| 284 |
* @private @const {!Object<string, boolean>} |
| 285 |
*/ |
| 286 |
this.attributeOverrideList_ = {}; |
| 287 |
|
| 288 |
/** |
| 289 |
* Whether the content of TEMPLATE tags (assuming TEMPLATE is whitelisted) |
| 290 |
* should be sanitized or passed through. |
| 291 |
* @private {boolean} |
| 292 |
*/ |
| 293 |
this.shouldSanitizeTemplateContents_ = true; |
| 294 |
|
| 295 |
/** |
| 296 |
* List of data attributes to whitelist. Data-attributes are inert and don't |
| 297 |
* require sanitization. |
| 298 |
* @private @const {!Array<string>} |
| 299 |
*/ |
| 300 |
this.dataAttributeWhitelist_ = []; |
| 301 |
|
| 302 |
/** |
| 303 |
* A tag blacklist, to effectively remove an element and its children from the |
| 304 |
* dom. |
| 305 |
* @private @const {!Object<string, boolean>} |
| 306 |
*/ |
| 307 |
this.tagBlacklist_ = goog.object.clone(goog.html.sanitizer.TagBlacklist); |
| 308 |
|
| 309 |
/** |
| 310 |
* A tag whitelist, to effectively allow an element and its children from the |
| 311 |
* dom. |
| 312 |
* @private {!Object<string, boolean>} |
| 313 |
*/ |
| 314 |
this.tagWhitelist_ = goog.object.clone(goog.html.sanitizer.TagWhitelist); |
| 315 |
|
| 316 |
/** |
| 317 |
* Whether non-whitelisted and non-blacklisted tags that have been converted |
| 318 |
* to <span&rt; tags will contain the original tag in a data attribute. |
| 319 |
* @private {boolean} |
| 320 |
*/ |
| 321 |
this.shouldAddOriginalTagNames_ = false; |
| 322 |
|
| 323 |
/** |
| 324 |
* A function to be applied to URLs found on the parsing process which do not |
| 325 |
* trigger requests. |
| 326 |
* @private {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} |
| 327 |
*/ |
| 328 |
this.urlPolicy_ = goog.html.sanitizer.HtmlSanitizer.defaultUrlPolicy_; |
| 329 |
|
| 330 |
/** |
| 331 |
* A function to be applied to urls found on the parsing process which may |
| 332 |
* trigger requests. |
| 333 |
* @private {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} |
| 334 |
*/ |
| 335 |
this.networkRequestUrlPolicy_ = |
| 336 |
goog.html.sanitizer.HtmlSanitizer.defaultNetworkRequestUrlPolicy_; |
| 337 |
|
| 338 |
/** |
| 339 |
* A function to be applied to names found on the parsing process. |
| 340 |
* @private {!goog.html.sanitizer.HtmlSanitizerPolicy} |
| 341 |
*/ |
| 342 |
this.namePolicy_ = goog.html.sanitizer.HtmlSanitizer.defaultNamePolicy_; |
| 343 |
|
| 344 |
/** |
| 345 |
* A function to be applied to other tokens (i.e. classes and IDs) found on |
| 346 |
* the parsing process. |
| 347 |
* @private {!goog.html.sanitizer.HtmlSanitizerPolicy} |
| 348 |
*/ |
| 349 |
this.tokenPolicy_ = goog.html.sanitizer.HtmlSanitizer.defaultTokenPolicy_; |
| 350 |
|
| 351 |
/** |
| 352 |
* A function to sanitize inline CSS styles. Defaults to deny all. |
| 353 |
* @private {function( |
| 354 |
* !goog.html.sanitizer.HtmlSanitizerPolicy, |
| 355 |
* string, |
| 356 |
* !goog.html.sanitizer.HtmlSanitizerPolicyHints, |
| 357 |
* !goog.html.sanitizer.HtmlSanitizerPolicyContext):?string} |
| 358 |
*/ |
| 359 |
this.sanitizeInlineCssPolicy_ = goog.functions.NULL; |
| 360 |
|
| 361 |
/** |
| 362 |
* An optional ID to restrict the scope of CSS rules when STYLE tags are |
| 363 |
* allowed. |
| 364 |
* @private {?string} |
| 365 |
*/ |
| 366 |
this.styleContainer_ = null; |
| 367 |
|
| 368 |
/** |
| 369 |
* Whether rules in STYLE tags should be inlined into style attributes. |
| 370 |
* @private {boolean} |
| 371 |
*/ |
| 372 |
this.inlineStyleRules_ = false; |
| 373 |
|
| 374 |
/** |
| 375 |
* True iff policies have been installed for the instance. |
| 376 |
* @private {boolean} |
| 377 |
*/ |
| 378 |
this.policiesInstalled_ = false; |
| 379 |
}; |
| 380 |
|
| 381 |
|
| 382 |
/** |
| 383 |
* Extends the list of allowed data attributes. |
| 384 |
* @param {!Array<string>} dataAttributeWhitelist |
| 385 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 386 |
*/ |
| 387 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.allowDataAttributes = |
| 388 |
function(dataAttributeWhitelist) { |
| 389 |
goog.array.extend(this.dataAttributeWhitelist_, dataAttributeWhitelist); |
| 390 |
return this; |
| 391 |
}; |
| 392 |
|
| 393 |
|
| 394 |
/** |
| 395 |
* Allows form tags in the HTML. Without this all form tags and content will be |
| 396 |
* dropped. |
| 397 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 398 |
*/ |
| 399 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.allowFormTag = function() { |
| 400 |
delete this.tagBlacklist_['FORM']; |
| 401 |
return this; |
| 402 |
}; |
| 403 |
|
| 404 |
|
| 405 |
/** |
| 406 |
* Allows STYLE tags. Note that the sanitizer wraps the output of each call to |
| 407 |
* {@link sanitize} with a SPAN tag, give it a random ID unique across multiple |
| 408 |
* calls, and then restrict all CSS rules found inside STYLE tags to only apply |
| 409 |
* to children of the SPAN tag. This means that CSS rules in STYLE tags will |
| 410 |
* only apply to content provided in the same call to {@link sanitize}. This |
| 411 |
* feature is not compatible with {@link inlineStyleRules}. |
| 412 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 413 |
*/ |
| 414 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.allowStyleTag = function() { |
| 415 |
if (this.inlineStyleRules_) { |
| 416 |
throw new Error('Rules from STYLE tags are already being inlined.'); |
| 417 |
} |
| 418 |
delete this.tagBlacklist_['STYLE']; |
| 419 |
this.styleContainer_ = goog.html.sanitizer.RANDOM_CONTAINER_; |
| 420 |
return this; |
| 421 |
}; |
| 422 |
|
| 423 |
|
| 424 |
/** |
| 425 |
* Fixes the ID of the style container used for CSS rules found in STYLE tags, |
| 426 |
* and disables automatic wrapping with the container. This allows multiple |
| 427 |
* calls to {@link sanitize} to share STYLE rules. If opt_styleContainer is |
| 428 |
* missing, the sanitizer will stop restricting the scope of CSS rules |
| 429 |
* altogether. Requires {@link allowStyleTag} to be called first. |
| 430 |
* @param {string=} opt_styleContainer An optional container ID to restrict the |
| 431 |
* scope of any CSS rule found in STYLE tags. |
| 432 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 433 |
*/ |
| 434 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.withStyleContainer = |
| 435 |
function(opt_styleContainer) { |
| 436 |
if ('STYLE' in this.tagBlacklist_) { |
| 437 |
throw new Error('STYLE tags must first be allowed through allowStyleTag.'); |
| 438 |
} |
| 439 |
if (opt_styleContainer != undefined) { |
| 440 |
if (!/^[a-zA-Z][\w-:\.]*$/.test(opt_styleContainer)) { |
| 441 |
throw new Error('Invalid ID.'); |
| 442 |
} |
| 443 |
this.styleContainer_ = opt_styleContainer; |
| 444 |
} else { |
| 445 |
this.styleContainer_ = null; |
| 446 |
} |
| 447 |
return this; |
| 448 |
}; |
| 449 |
|
| 450 |
|
| 451 |
/** |
| 452 |
* Converts rules in STYLE tags into style attributes on the tags they apply to. |
| 453 |
* This feature is not compatible with {@link withStyleContainer} and {@link |
| 454 |
* allowStyleTag}. This method requires {@link allowCssStyles} (otherwise rules |
| 455 |
* would be deleted after being inlined), and is not compatible with {@link |
| 456 |
* allowStyleTag}. |
| 457 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 458 |
*/ |
| 459 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.inlineStyleRules = |
| 460 |
function() { |
| 461 |
if (this.sanitizeInlineCssPolicy_ == goog.functions.NULL) { |
| 462 |
throw new Error( |
| 463 |
'Inlining style rules requires allowing STYLE attributes ' + |
| 464 |
'first.'); |
| 465 |
} |
| 466 |
if (!('STYLE' in this.tagBlacklist_)) { |
| 467 |
throw new Error( |
| 468 |
'You have already configured the builder to allow STYLE tags in the ' + |
| 469 |
'output. Inlining style rules would prevent STYLE tags from ' + |
| 470 |
'appearing in the output and conflict with such directive.'); |
| 471 |
} |
| 472 |
this.inlineStyleRules_ = true; |
| 473 |
return this; |
| 474 |
}; |
| 475 |
|
| 476 |
|
| 477 |
/** |
| 478 |
* Allows inline CSS styles. |
| 479 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 480 |
*/ |
| 481 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.allowCssStyles = |
| 482 |
function() { |
| 483 |
this.sanitizeInlineCssPolicy_ = |
| 484 |
goog.html.sanitizer.HtmlSanitizer.sanitizeCssDeclarationList_; |
| 485 |
return this; |
| 486 |
}; |
| 487 |
|
| 488 |
|
| 489 |
/** |
| 490 |
* Extends the tag whitelist (Package-internal utility method only). |
| 491 |
* @param {!Array<string>} tags The list of tags to be added to the whitelist. |
| 492 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 493 |
* @package |
| 494 |
*/ |
| 495 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype |
| 496 |
.alsoAllowTagsPrivateDoNotAccessOrElse = function(tags) { |
| 497 |
goog.array.forEach(tags, function(tag) { |
| 498 |
this.tagWhitelist_[tag.toUpperCase()] = true; |
| 499 |
}, this); |
| 500 |
return this; |
| 501 |
}; |
| 502 |
|
| 503 |
|
| 504 |
/** |
| 505 |
* Extends the attribute whitelist (Package-internal utility method only). |
| 506 |
* @param {!Array<(string|!goog.html.sanitizer.HtmlSanitizerAttributePolicy)>} |
| 507 |
* attrs The list of attributes to be added to the whitelist. |
| 508 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 509 |
* @package |
| 510 |
*/ |
| 511 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype |
| 512 |
.alsoAllowAttributesPrivateDoNotAccessOrElse = function(attrs) { |
| 513 |
goog.array.forEach(attrs, function(attr) { |
| 514 |
if (goog.isString(attr)) { |
| 515 |
attr = {tagName: '*', attributeName: attr, policy: null}; |
| 516 |
} |
| 517 |
var handlerName = goog.html.sanitizer.HtmlSanitizer.attrIdentifier_( |
| 518 |
attr.tagName, attr.attributeName); |
| 519 |
this.attributeWhitelist_[handlerName] = attr.policy ? |
| 520 |
attr.policy : |
| 521 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ ( |
| 522 |
goog.html.sanitizer.HtmlSanitizer.cleanUpAttribute_); |
| 523 |
this.attributeOverrideList_[handlerName] = true; |
| 524 |
}, this); |
| 525 |
return this; |
| 526 |
}; |
| 527 |
|
| 528 |
|
| 529 |
/** |
| 530 |
* Turns off sanitization of template tag contents and pass them unmodified |
| 531 |
* (Package-internal utility method only). |
| 532 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 533 |
* @throws {!Error} |
| 534 |
* @package |
| 535 |
*/ |
| 536 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype |
| 537 |
.keepUnsanitizedTemplateContentsPrivateDoNotAccessOrElse = function() { |
| 538 |
if (!goog.html.sanitizer.HTML_SANITIZER_TEMPLATE_SUPPORTED) { |
| 539 |
throw new Error( |
| 540 |
'Cannot let unsanitized template contents through on ' + |
| 541 |
'browsers that do not support TEMPLATE.'); |
| 542 |
} |
| 543 |
this.shouldSanitizeTemplateContents_ = false; |
| 544 |
return this; |
| 545 |
}; |
| 546 |
|
| 547 |
|
| 548 |
/** |
| 549 |
* Allows only the provided whitelist of tags. Tags still need to be in the |
| 550 |
* TagWhitelist to be allowed. |
| 551 |
* <p> |
| 552 |
* SPAN tags are ALWAYS ALLOWED as part of the mechanism required to preserve |
| 553 |
* the HTML tree structure (when removing non-blacklisted tags and |
| 554 |
* non-whitelisted tags). |
| 555 |
* @param {!Array<string>} tagWhitelist |
| 556 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 557 |
* @throws {Error} Thrown if an attempt is made to allow a non-whitelisted tag. |
| 558 |
*/ |
| 559 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.onlyAllowTags = function( |
| 560 |
tagWhitelist) { |
| 561 |
this.tagWhitelist_ = {'SPAN': true}; |
| 562 |
goog.array.forEach(tagWhitelist, function(tag) { |
| 563 |
tag = tag.toUpperCase(); |
| 564 |
if (goog.html.sanitizer.TagWhitelist[tag]) { |
| 565 |
this.tagWhitelist_[tag] = true; |
| 566 |
} else { |
| 567 |
throw new Error( |
| 568 |
'Only whitelisted tags can be allowed. See ' + |
| 569 |
'goog.html.sanitizer.TagWhitelist.'); |
| 570 |
} |
| 571 |
}, this); |
| 572 |
return this; |
| 573 |
}; |
| 574 |
|
| 575 |
|
| 576 |
/** |
| 577 |
* Allows only the provided whitelist of attributes, possibly setting a custom |
| 578 |
* policy for them. The set of tag/attribute combinations need to be a subset of |
| 579 |
* the currently allowed combinations. |
| 580 |
* <p> |
| 581 |
* Note that you cannot define a generic handler for an attribute if only a |
| 582 |
* tag-specific one is present, and vice versa. To configure the sanitizer to |
| 583 |
* accept an attribute only for a specific tag when only a generic handler is |
| 584 |
* whitelisted, use the goog.html.sanitizer.HtmlSanitizerPolicyHints parameter |
| 585 |
* and simply reject the attribute in unwanted tags. |
| 586 |
* <p> |
| 587 |
* Also note that the sanitizer's policy is still called after the provided one, |
| 588 |
* to ensure that supplying misconfigured policy cannot introduce |
| 589 |
* vulnerabilities. To completely override an existing attribute policy or to |
| 590 |
* allow new attributes, see the goog.html.sanitizer.unsafe package. |
| 591 |
* @param {!Array<(string|!goog.html.sanitizer.HtmlSanitizerAttributePolicy)>} |
| 592 |
* attrWhitelist The subset of attributes that the sanitizer will accept. |
| 593 |
* Attributes can come in of two forms: |
| 594 |
* - string: allow all values for this attribute on all tags. |
| 595 |
* - HtmlSanitizerAttributePolicy: allows specifying a policy for a |
| 596 |
* particular tag. The tagName can be "*", which means all tags. If no |
| 597 |
* policy is passed, the default is to allow all values. |
| 598 |
* The tag and attribute names are case-insensitive. |
| 599 |
* Note that the policy for id, URLs, names etc is controlled separately |
| 600 |
* (using withCustom* methods). |
| 601 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 602 |
* @throws {Error} Thrown if an attempt is made to allow a non-whitelisted |
| 603 |
* attribute. |
| 604 |
*/ |
| 605 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.onlyAllowAttributes = |
| 606 |
function(attrWhitelist) { |
| 607 |
var oldWhitelist = this.attributeWhitelist_; |
| 608 |
this.attributeWhitelist_ = {}; |
| 609 |
goog.array.forEach(attrWhitelist, function(attr) { |
| 610 |
if (goog.typeOf(attr) === 'string') { |
| 611 |
attr = {tagName: '*', attributeName: attr.toUpperCase(), policy: null}; |
| 612 |
} |
| 613 |
var handlerName = goog.html.sanitizer.HtmlSanitizer.attrIdentifier_( |
| 614 |
attr.tagName, attr.attributeName); |
| 615 |
if (!oldWhitelist[handlerName]) { |
| 616 |
throw new Error('Only whitelisted attributes can be allowed.'); |
| 617 |
} |
| 618 |
this.attributeWhitelist_[handlerName] = attr.policy ? |
| 619 |
attr.policy : |
| 620 |
/** @type {goog.html.sanitizer.HtmlSanitizerPolicy} */ ( |
| 621 |
goog.html.sanitizer.HtmlSanitizer.cleanUpAttribute_); |
| 622 |
}, this); |
| 623 |
return this; |
| 624 |
}; |
| 625 |
|
| 626 |
|
| 627 |
/** |
| 628 |
* Adds the original tag name in the data attribute 'original-tag' when unknown |
| 629 |
* tags are sanitized to <span&rt;, so that caller can distinguish them from |
| 630 |
* actual <span&rt; tags. |
| 631 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 632 |
*/ |
| 633 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.addOriginalTagNames = |
| 634 |
function() { |
| 635 |
this.shouldAddOriginalTagNames_ = true; |
| 636 |
return this; |
| 637 |
}; |
| 638 |
|
| 639 |
|
| 640 |
/** |
| 641 |
* Sets a custom network URL policy. |
| 642 |
* @param {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} |
| 643 |
* customNetworkReqUrlPolicy |
| 644 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 645 |
*/ |
| 646 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype |
| 647 |
.withCustomNetworkRequestUrlPolicy = function(customNetworkReqUrlPolicy) { |
| 648 |
this.networkRequestUrlPolicy_ = customNetworkReqUrlPolicy; |
| 649 |
return this; |
| 650 |
}; |
| 651 |
|
| 652 |
|
| 653 |
/** |
| 654 |
* Sets a custom non-network URL policy. |
| 655 |
* @param {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} customUrlPolicy |
| 656 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 657 |
*/ |
| 658 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.withCustomUrlPolicy = |
| 659 |
function(customUrlPolicy) { |
| 660 |
this.urlPolicy_ = customUrlPolicy; |
| 661 |
return this; |
| 662 |
}; |
| 663 |
|
| 664 |
|
| 665 |
/** |
| 666 |
* Sets a custom name policy. |
| 667 |
* @param {!goog.html.sanitizer.HtmlSanitizerPolicy} customNamePolicy |
| 668 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 669 |
*/ |
| 670 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.withCustomNamePolicy = |
| 671 |
function(customNamePolicy) { |
| 672 |
this.namePolicy_ = customNamePolicy; |
| 673 |
return this; |
| 674 |
}; |
| 675 |
|
| 676 |
|
| 677 |
/** |
| 678 |
* Sets a custom token policy. |
| 679 |
* @param {!goog.html.sanitizer.HtmlSanitizerPolicy} customTokenPolicy |
| 680 |
* @return {!goog.html.sanitizer.HtmlSanitizer.Builder} |
| 681 |
*/ |
| 682 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.withCustomTokenPolicy = |
| 683 |
function(customTokenPolicy) { |
| 684 |
this.tokenPolicy_ = customTokenPolicy; |
| 685 |
return this; |
| 686 |
}; |
| 687 |
|
| 688 |
|
| 689 |
/** |
| 690 |
* Wraps a custom policy function with the sanitizer's default policy. |
| 691 |
* @param {?goog.html.sanitizer.HtmlSanitizerPolicy} customPolicy The custom |
| 692 |
* policy for the tag/attribute combination. |
| 693 |
* @param {!goog.html.sanitizer.HtmlSanitizerPolicy} defaultPolicy The |
| 694 |
* sanitizer's policy that is always called after the custom policy. |
| 695 |
* @return {!goog.html.sanitizer.HtmlSanitizerPolicy} |
| 696 |
* @private |
| 697 |
*/ |
| 698 |
goog.html.sanitizer.HtmlSanitizer.wrapPolicy_ = function( |
| 699 |
customPolicy, defaultPolicy) { |
| 700 |
return /** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ (function( |
| 701 |
value, hints, ctx, policy) { |
| 702 |
var result = customPolicy(value, hints, ctx, policy); |
| 703 |
return result == null ? null : defaultPolicy(result, hints, ctx, policy); |
| 704 |
}); |
| 705 |
}; |
| 706 |
|
| 707 |
|
| 708 |
/** |
| 709 |
* Installs the sanitizer's default policy for a specific tag/attribute |
| 710 |
* combination on the provided whitelist, but only if a policy already exists. |
| 711 |
* @param {!Object<string, !goog.html.sanitizer.HtmlSanitizerPolicy>} |
| 712 |
* whitelist The whitelist to modify. |
| 713 |
* @param {!Object<string, boolean>} overrideList The set of attributes handlers |
| 714 |
* that should not be wrapped with a default policy. |
| 715 |
* @param {string} key The tag/attribute combination |
| 716 |
* @param {!goog.html.sanitizer.HtmlSanitizerPolicy} defaultPolicy The |
| 717 |
* sanitizer's policy. |
| 718 |
* @private |
| 719 |
*/ |
| 720 |
goog.html.sanitizer.HtmlSanitizer.installDefaultPolicy_ = function( |
| 721 |
whitelist, overrideList, key, defaultPolicy) { |
| 722 |
if (whitelist[key] && !overrideList[key]) { |
| 723 |
whitelist[key] = goog.html.sanitizer.HtmlSanitizer.wrapPolicy_( |
| 724 |
whitelist[key], defaultPolicy); |
| 725 |
} |
| 726 |
}; |
| 727 |
|
| 728 |
|
| 729 |
/** |
| 730 |
* Builds and returns a goog.html.sanitizer.HtmlSanitizer object. |
| 731 |
* @return {!goog.html.sanitizer.HtmlSanitizer} |
| 732 |
*/ |
| 733 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.build = function() { |
| 734 |
return new goog.html.sanitizer.HtmlSanitizer(this); |
| 735 |
}; |
| 736 |
|
| 737 |
|
| 738 |
/** |
| 739 |
* Installs the sanitization policies for the attributes. |
| 740 |
* May only be called once. |
| 741 |
* @private |
| 742 |
*/ |
| 743 |
goog.html.sanitizer.HtmlSanitizer.Builder.prototype.installPolicies_ = |
| 744 |
function() { |
| 745 |
if (this.policiesInstalled_) { |
| 746 |
throw new Error('HtmlSanitizer.Builder.build() can only be used once.'); |
| 747 |
} |
| 748 |
|
| 749 |
var installPolicy = goog.html.sanitizer.HtmlSanitizer.installDefaultPolicy_; |
| 750 |
|
| 751 |
// Binding all the non-trivial attribute sanitizers to the appropriate, |
| 752 |
// potentially customizable, handling functions at build(). |
| 753 |
installPolicy( |
| 754 |
this.attributeWhitelist_, this.attributeOverrideList_, '* USEMAP', |
| 755 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ ( |
| 756 |
goog.html.sanitizer.HtmlSanitizer.sanitizeUrlFragment_)); |
| 757 |
|
| 758 |
var urlAttributes = ['* ACTION', '* CITE', '* HREF']; |
| 759 |
var urlPolicy = |
| 760 |
goog.html.sanitizer.HtmlSanitizer.wrapUrlPolicy_(this.urlPolicy_); |
| 761 |
goog.array.forEach(urlAttributes, function(attribute) { |
| 762 |
installPolicy( |
| 763 |
this.attributeWhitelist_, this.attributeOverrideList_, attribute, |
| 764 |
urlPolicy); |
| 765 |
}, this); |
| 766 |
|
| 767 |
var networkUrlAttributes = [ |
| 768 |
// LONGDESC can result in a network request. See b/23381636. |
| 769 |
'* LONGDESC', '* SRC', 'LINK HREF' |
| 770 |
]; |
| 771 |
var networkRequestUrlPolicy = |
| 772 |
goog.html.sanitizer.HtmlSanitizer.wrapUrlPolicy_( |
| 773 |
this.networkRequestUrlPolicy_); |
| 774 |
goog.array.forEach(networkUrlAttributes, function(attribute) { |
| 775 |
installPolicy( |
| 776 |
this.attributeWhitelist_, this.attributeOverrideList_, attribute, |
| 777 |
networkRequestUrlPolicy); |
| 778 |
}, this); |
| 779 |
|
| 780 |
var nameAttributes = ['* FOR', '* HEADERS', '* NAME']; |
| 781 |
goog.array.forEach(nameAttributes, function(attribute) { |
| 782 |
installPolicy( |
| 783 |
this.attributeWhitelist_, this.attributeOverrideList_, attribute, |
| 784 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ (goog.partial( |
| 785 |
goog.html.sanitizer.HtmlSanitizer.sanitizeName_, |
| 786 |
this.namePolicy_))); |
| 787 |
}, this); |
| 788 |
|
| 789 |
installPolicy( |
| 790 |
this.attributeWhitelist_, this.attributeOverrideList_, 'A TARGET', |
| 791 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ (goog.partial( |
| 792 |
goog.html.sanitizer.HtmlSanitizer.allowedAttributeValues_, |
| 793 |
['_blank', '_self']))); |
| 794 |
|
| 795 |
installPolicy( |
| 796 |
this.attributeWhitelist_, this.attributeOverrideList_, '* CLASS', |
| 797 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ (goog.partial( |
| 798 |
goog.html.sanitizer.HtmlSanitizer.sanitizeClasses_, |
| 799 |
this.tokenPolicy_))); |
| 800 |
|
| 801 |
installPolicy( |
| 802 |
this.attributeWhitelist_, this.attributeOverrideList_, '* ID', |
| 803 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ (goog.partial( |
| 804 |
goog.html.sanitizer.HtmlSanitizer.sanitizeId_, this.tokenPolicy_))); |
| 805 |
|
| 806 |
installPolicy( |
| 807 |
this.attributeWhitelist_, this.attributeOverrideList_, '* STYLE', |
| 808 |
/** @type {!goog.html.sanitizer.HtmlSanitizerPolicy} */ |
| 809 |
(goog.partial(this.sanitizeInlineCssPolicy_, networkRequestUrlPolicy))); |
| 810 |
|
| 811 |
this.policiesInstalled_ = true; |
| 812 |
}; |
| 813 |
|
| 814 |
|
| 815 |
/** |
| 816 |
* The default policy for URLs: allow any. |
| 817 |
* @private @const {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} |
| 818 |
*/ |
| 819 |
goog.html.sanitizer.HtmlSanitizer.defaultUrlPolicy_ = |
| 820 |
goog.html.SafeUrl.sanitize; |
| 821 |
|
| 822 |
|
| 823 |
/** |
| 824 |
* The default policy for URLs which cause network requests: drop all. |
| 825 |
* @private @const {!goog.html.sanitizer.HtmlSanitizerUrlPolicy} |
| 826 |
*/ |
| 827 |
goog.html.sanitizer.HtmlSanitizer.defaultNetworkRequestUrlPolicy_ = |
| 828 |
goog.functions.NULL; |
| 829 |
|
| 830 |
|
| 831 |
/** |
| 832 |
* The default policy for attribute names: drop all. |
| 833 |
* @private @const {!goog.html.sanitizer.HtmlSanitizerPolicy} |
| 834 |
*/ |
| 835 |
goog.html.sanitizer.HtmlSanitizer.defaultNamePolicy_ = goog.functions.NULL; |
| 836 |
|
| 837 |
|
| 838 |
/** |
| 839 |
* The default policy for other tokens (i.e. class names and IDs): drop all. |
| 840 |
* @private @const {!goog.html.sanitizer.HtmlSanitizerPolicy} |
| 841 |
*/ |
| 842 |
goog.html.sanitizer.HtmlSanitizer.defaultTokenPolicy_ = goog.functions.NULL; |
| 843 |
|
| 844 |
|
| 845 |
|
| 846 |
/** |
| 847 |
* Returns a key into the attribute handlers dictionary given a node name and |
| 848 |
* an attribute name. If no node name is given, returns a key applying to all |
| 849 |
* nodes. |
| 850 |
* @param {?string} nodeName |
| 851 |
* @param {string} attributeName |
| 852 |
* @return {string} key into attribute handlers dict |
| 853 |
* @private |
| 854 |
*/ |
| 855 |
goog.html.sanitizer.HtmlSanitizer.attrIdentifier_ = function( |
| 856 |
nodeName, attributeName) { |
| 857 |
if (!nodeName) { |
| 858 |
nodeName = '*'; |
| 859 |
} |
| 860 |
return (nodeName + ' ' + attributeName).toUpperCase(); |
| 861 |
}; |
| 862 |
|
| 863 |
|
| 864 |
/** |
| 865 |
* Sanitizes a list of CSS declarations. |
| 866 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicy} policySanitizeUrl |
| 867 |
* @param {string} attrValue |
| 868 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyHints} policyHints |
| 869 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyContext} policyContext |
| 870 |
* @return {?string} sanitizedCss from the policyContext |
| 871 |
* @private |
| 872 |
*/ |
| 873 |
goog.html.sanitizer.HtmlSanitizer.sanitizeCssDeclarationList_ = function( |
| 874 |
policySanitizeUrl, attrValue, policyHints, policyContext) { |
| 875 |
if (!policyContext.cssStyle) { |
| 876 |
return null; |
| 877 |
} |
| 878 |
var naiveUriRewriter = function(uri, prop) { |
| 879 |
policyHints.cssProperty = prop; |
| 880 |
var sanitizedUrl = policySanitizeUrl(uri, policyHints); |
| 881 |
if (sanitizedUrl == null) { |
| 882 |
return null; |
| 883 |
} |
| 884 |
return goog.html.uncheckedconversions |
| 885 |
.safeUrlFromStringKnownToSatisfyTypeContract( |
| 886 |
goog.string.Const.from( |
| 887 |
'HtmlSanitizerPolicy created with networkRequestUrlPolicy_ ' + |
| 888 |
'when installing \'* STYLE\' handler.'), |
| 889 |
sanitizedUrl); |
| 890 |
}; |
| 891 |
var sanitizedStyle = goog.html.SafeStyle.unwrap( |
| 892 |
goog.html.sanitizer.CssSanitizer.sanitizeInlineStyle( |
| 893 |
policyContext.cssStyle, naiveUriRewriter)); |
| 894 |
return sanitizedStyle == '' ? null : sanitizedStyle; |
| 895 |
}; |
| 896 |
|
| 897 |
|
| 898 |
/** |
| 899 |
* Cleans up an attribute value that we don't particularly want to do anything |
| 900 |
* to. At the moment we just trim the whitespace. |
| 901 |
* @param {string} attrValue |
| 902 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyHints} policyHints |
| 903 |
* @return {string} sanitizedAttrValue |
| 904 |
* @private |
| 905 |
*/ |
| 906 |
goog.html.sanitizer.HtmlSanitizer.cleanUpAttribute_ = function( |
| 907 |
attrValue, policyHints) { |
| 908 |
return goog.string.trim(attrValue); |
| 909 |
}; |
| 910 |
|
| 911 |
|
| 912 |
/** |
| 913 |
* Allows a set of attribute values. |
| 914 |
* @param {!Array<string>} allowedValues Set of allowed values lowercased. |
| 915 |
* @param {string} attrValue |
| 916 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyHints} policyHints |
| 917 |
* @return {?string} sanitizedAttrValue |
| 918 |
* @private |
| 919 |
*/ |
| 920 |
goog.html.sanitizer.HtmlSanitizer.allowedAttributeValues_ = function( |
| 921 |
allowedValues, attrValue, policyHints) { |
| 922 |
var trimmed = goog.string.trim(attrValue); |
| 923 |
return goog.array.contains(allowedValues, trimmed.toLowerCase()) ? trimmed : |
| 924 |
null; |
| 925 |
}; |
| 926 |
|
| 927 |
|
| 928 |
/** |
| 929 |
* Sanitizes URL fragments. |
| 930 |
* @param {string} urlFragment |
| 931 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyHints} policyHints |
| 932 |
* @return {?string} sanitizedAttrValue |
| 933 |
* @private |
| 934 |
*/ |
| 935 |
goog.html.sanitizer.HtmlSanitizer.sanitizeUrlFragment_ = function( |
| 936 |
urlFragment, policyHints) { |
| 937 |
var trimmed = goog.string.trim(urlFragment); |
| 938 |
if (trimmed && trimmed.charAt(0) == '#') { |
| 939 |
// We do not apply the name or token policy to Url Fragments by design. |
| 940 |
return trimmed; |
| 941 |
} |
| 942 |
return null; |
| 943 |
}; |
| 944 |
|
| 945 |
|
| 946 |
/** |
| 947 |
* Runs an attribute name through a name policy. |
| 948 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicy} namePolicy |
| 949 |
* @param {string} attrName |
| 950 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyHints} policyHints |
| 951 |
* @return {?string} sanitizedAttrValue |
| 952 |
* @private |
| 953 |
*/ |
| 954 |
goog.html.sanitizer.HtmlSanitizer.sanitizeName_ = function( |
| 955 |
namePolicy, attrName, policyHints) { |
| 956 |
var trimmed = goog.string.trim(attrName); |
| 957 |
/* TODO(user): fail on names which contain illegal characters. |
| 958 |
* NOTE(jasvir): |
| 959 |
* There are two cases to be concerned about - escaped quotes in attribute |
| 960 |
* values which is the responsibility of the serializer and illegal |
| 961 |
* characters. The latter does violate the spec but I do not believe it has |
| 962 |
* a security consequence. |
| 963 |
*/ |
| 964 |
return namePolicy(trimmed, policyHints); |
| 965 |
}; |
| 966 |
|
| 967 |
|
| 968 |
/** |
| 969 |
* Ensures that the class prefix is present on all space-separated tokens |
| 970 |
* (i.e. all class names). |
| 971 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicy} tokenPolicy |
| 972 |
* @param {string} attrValue |
| 973 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyHints} policyHints |
| 974 |
* @return {?string} sanitizedAttrValue |
| 975 |
* @private |
| 976 |
*/ |
| 977 |
goog.html.sanitizer.HtmlSanitizer.sanitizeClasses_ = function( |
| 978 |
tokenPolicy, attrValue, policyHints) { |
| 979 |
// TODO(user): use a browser-supplied class list instead of a string. |
| 980 |
var classes = attrValue.split(/(?:\s+)/); |
| 981 |
var sanitizedClasses = []; |
| 982 |
for (var i = 0; i < classes.length; i++) { |
| 983 |
// TODO(user): skip classes which contain illegal characters. |
| 984 |
var sanitizedClass = tokenPolicy(classes[i], policyHints); |
| 985 |
if (sanitizedClass) { |
| 986 |
sanitizedClasses.push(sanitizedClass); |
| 987 |
} |
| 988 |
} |
| 989 |
return sanitizedClasses.length == 0 ? null : sanitizedClasses.join(' '); |
| 990 |
}; |
| 991 |
|
| 992 |
|
| 993 |
/** |
| 994 |
* Ensures that the id prefix is present. |
| 995 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicy} tokenPolicy |
| 996 |
* @param {string} attrValue |
| 997 |
* @param {goog.html.sanitizer.HtmlSanitizerPolicyHints} policyHints |
| 998 |
* @return {?string} sanitizedAttrValue |
| 999 |
* @private |
| 1000 |
*/ |
| 1001 |
goog.html.sanitizer.HtmlSanitizer.sanitizeId_ = function( |
| 1002 |
tokenPolicy, attrValue, policyHints) { |
| 1003 |
var trimmed = goog.string.trim(attrValue); |
| 1004 |
// TODO(user): fail on IDs which contain illegal characters. |
| 1005 |
return tokenPolicy(trimmed, policyHints); |
| 1006 |
}; |
| 1007 |
|
| 1008 |
|
| 1009 |
/** |
| 1010 |
* Parses a string of unsanitized HTML and provides an iterator over the |
| 1011 |
* resulting DOM tree nodes. This DOM parsing must be wholly inert (that is, |
| 1012 |
* it does not cause execution of any active content or cause the browser to |
| 1013 |
* issue any requests). The returned iterator is guaranteed to iterate over a |
| 1014 |
* parent element before iterating over any of its children. |
| 1015 |
* @param {string} unsanitizedHtml |
| 1016 |
* @return {!TreeWalker} Dom tree iterator |
| 1017 |
* @private |
| 1018 |
*/ |
| 1019 |
goog.html.sanitizer.HtmlSanitizer.getDomTreeWalker_ = function( |
| 1020 |
unsanitizedHtml) { |
| 1021 |
var iteratorParent; |
| 1022 |
// Use a <template> element if possible. |
| 1023 |
var templateElement = document.createElement('template'); |
| 1024 |
if ('content' in templateElement) { |
| 1025 |
templateElement.innerHTML = unsanitizedHtml; |
| 1026 |
iteratorParent = templateElement.content; |
| 1027 |
} else { |
| 1028 |
// In browsers where <template> is not implemented, use an HTMLDocument. |
| 1029 |
var doc = document.implementation.createHTMLDocument('x'); |
| 1030 |
iteratorParent = doc.body; |
| 1031 |
doc.body.innerHTML = unsanitizedHtml; |
| 1032 |
} |
| 1033 |
return document.createTreeWalker( |
| 1034 |
iteratorParent, NodeFilter.SHOW_ELEMENT | NodeFilter.SHOW_TEXT, |
| 1035 |
null /* filter */, false /* entityReferenceExpansion */); |
| 1036 |
}; |
| 1037 |
|
| 1038 |
|
| 1039 |
/** |
| 1040 |
* Parses the DOM tree of a given HTML string, then walks the tree. For each |
| 1041 |
* element, it creates a new sanitized version, applies sanitized attributes, |
| 1042 |
* and returns a SafeHtml object representing the sanitized tree. |
| 1043 |
* @param {string} unsanitizedHtml |
| 1044 |
* @return {!goog.html.SafeHtml} Sanitized HTML |
| 1045 |
*/ |
| 1046 |
goog.html.sanitizer.HtmlSanitizer.prototype.sanitize = function( |
| 1047 |
unsanitizedHtml) { |
| 1048 |
var styleContainer = this.getStyleContainerId_(); |
| 1049 |
var sanitizedParent = |
| 1050 |
this.sanitizeToDomNode_(unsanitizedHtml, styleContainer); |
| 1051 |
var sanitizedString = new XMLSerializer().serializeToString(sanitizedParent); |
| 1052 |
|
| 1053 |
// Remove the outer span added in sanitizeToDomNode. We could create an |
| 1054 |
// element from it and then pull out the innerHTML, but this is more |
| 1055 |
// performant. |
| 1056 |
sanitizedString = sanitizedString.slice( |
| 1057 |
sanitizedString.indexOf('>') + 1, sanitizedString.lastIndexOf('</')); |
| 1058 |
if (sanitizedString == '') { |
| 1059 |
return goog.html.SafeHtml.EMPTY; |
| 1060 |
} |
| 1061 |
|
| 1062 |
var safeHtml = goog.html.uncheckedconversions |
| 1063 |
.safeHtmlFromStringKnownToSatisfyTypeContract( |
| 1064 |
goog.string.Const.from('Output of HTML sanitizer'), |
| 1065 |
sanitizedString); |
| 1066 |
// If appropriate, rewrap the output with a style container. |
| 1067 |
if (this.styleContainer_ == goog.html.sanitizer.RANDOM_CONTAINER_) { |
| 1068 |
safeHtml = goog.html.SafeHtml.create( |
| 1069 |
goog.dom.TagName.SPAN, {id: styleContainer}, safeHtml); |
| 1070 |
} |
| 1071 |
|
| 1072 |
return safeHtml; |
| 1073 |
}; |
| 1074 |
|
| 1075 |
|
| 1076 |
/** |
| 1077 |
* Parses the DOM tree of a given HTML string, then walks the tree. For each |
| 1078 |
* element, it creates a new sanitized version, applies sanitized attributes, |
| 1079 |
* and returns a span element containing the sanitized content. |
| 1080 |
* @param {string} unsanitizedHtml |
| 1081 |
* @return {!HTMLSpanElement} Sanitized HTML |
| 1082 |
*/ |
| 1083 |
goog.html.sanitizer.HtmlSanitizer.prototype.sanitizeToDomNode = function( |
| 1084 |
unsanitizedHtml) { |
| 1085 |
return this.sanitizeToDomNode_(unsanitizedHtml); |
| 1086 |
}; |
| 1087 |
|
| 1088 |
|
| 1089 |
/** |
| 1090 |
* Private variant of {@link sanitizeToDomNode} that allows passing a style |
| 1091 |
* container. If used with an unsafe style container, it can cause an XSS and |
| 1092 |
* thus we keep it private. Use the {@link withStyleContainer} builder function |
| 1093 |
* to set a name for the style container. |
| 1094 |
* @param {string} unsanitizedHtml |
| 1095 |
* @param {?string=} opt_styleContainer |
| 1096 |
* @return {!HTMLSpanElement} Sanitized HTML |
| 1097 |
* @private |
| 1098 |
*/ |
| 1099 |
goog.html.sanitizer.HtmlSanitizer.prototype.sanitizeToDomNode_ = function( |
| 1100 |
unsanitizedHtml, opt_styleContainer) { |
| 1101 |
var sanitizedParent = |
| 1102 |
/** @type {!HTMLSpanElement} */ (document.createElement('span')); |
| 1103 |
|
| 1104 |
if (!goog.html.sanitizer.HTML_SANITIZER_SUPPORTED_) { |
| 1105 |
return sanitizedParent; |
| 1106 |
} |
| 1107 |
|
| 1108 |
// Inline style rules on the unsanitized input, so that we don't have to |
| 1109 |
// worry about customTokenPolicy and customNamePolicy interferring with |
| 1110 |
// selectors. |
| 1111 |
if (this.inlineStyleRules_) { |
| 1112 |
// TODO(pelizzi): we are going to parse the document into a DOM tree twice, |
| 1113 |
// once with DOMParser here, and once with TEMPLATE in the main sanitization |
| 1114 |
// loop. It would be best if we used one technique consistently and parse it |
| 1115 |
// once, but the decision to use TEMPLATE in the main body predates the work |
| 1116 |
// on supporting STYLE tags, and we later found on that TEMPLATE inert |
| 1117 |
// documents do not have computed stylesheet information on STYLE tags. |
| 1118 |
var inertUnsanitizedDom = |
| 1119 |
goog.html.sanitizer.CssSanitizer.safeParseHtmlAndGetInertElement( |
| 1120 |
'<span>' + unsanitizedHtml + '</span>'); |
| 1121 |
goog.asserts.assert( |
| 1122 |
inertUnsanitizedDom, |
| 1123 |
'Older browsers that don\'t support inert ' + |
| 1124 |
'parsing should not get to this branch'); |
| 1125 |
goog.html.sanitizer.CssSanitizer.inlineStyleRules(inertUnsanitizedDom); |
| 1126 |
unsanitizedHtml = inertUnsanitizedDom.innerHTML; |
| 1127 |
} |
| 1128 |
|
| 1129 |
|
| 1130 |
var styleContainer = opt_styleContainer != undefined ? |
| 1131 |
opt_styleContainer : |
| 1132 |
this.getStyleContainerId_(); |
| 1133 |
if (styleContainer) { |
| 1134 |
sanitizedParent.id = styleContainer; |
| 1135 |
} |
| 1136 |
|
| 1137 |
// Get the treeWalker initialized. |
| 1138 |
try { |
| 1139 |
var treeWalker = |
| 1140 |
goog.html.sanitizer.HtmlSanitizer.getDomTreeWalker_(unsanitizedHtml); |
| 1141 |
} catch (e) { |
| 1142 |
return sanitizedParent; |
| 1143 |
} |
| 1144 |
|
| 1145 |
// Used in order to find the correct parent node in the sanitizedParent. |
| 1146 |
var elementMap = goog.html.sanitizer.ElementWeakMap.newWeakMap(); |
| 1147 |
// Used for iteration. |
| 1148 |
var dirtyNode; |
| 1149 |
while (dirtyNode = treeWalker.nextNode()) { |
| 1150 |
// Get a clean (sanitized) version of the dirty node. |
| 1151 |
var cleanNode = this.sanitizeNode_(dirtyNode, styleContainer); |
| 1152 |
if (goog.html.sanitizer.noclobber.isNodeElement(cleanNode)) { |
| 1153 |
// If sanitizeNode_ returned an element, then the original node was also |
| 1154 |
// an element. |
| 1155 |
dirtyNode = goog.html.sanitizer.noclobber.assertNodeIsElement(dirtyNode); |
| 1156 |
cleanNode = goog.html.sanitizer.noclobber.assertNodeIsElement(cleanNode); |
| 1157 |
this.sanitizeAttrs_(dirtyNode, cleanNode); |
| 1158 |
elementMap.set(dirtyNode, cleanNode); |
| 1159 |
|
| 1160 |
// Template tag contents require special handling as they are not |
| 1161 |
// traversed by the treewalker. |
| 1162 |
this.maybeProcessTemplateContents_(dirtyNode, cleanNode); |
| 1163 |
} |
| 1164 |
|
| 1165 |
// TODO(pelizzi): [IMPROVEMENT] type-checking against clobbering (e.g. |
| 1166 |
// ClobberedNode wrapper). Closure can unwrap these at compile time, see |
| 1167 |
// ClosureOptimizePrimitives.java, jakubvrana has created one for |
| 1168 |
// goog.dom.Tag. Alternatively, create two actual wrappers that expose |
| 1169 |
// clobber-safe functions, getters and setters for Node and Element. |
| 1170 |
|
| 1171 |
// TODO(pelizzi): [IMPROVEMENT] add an API to sanitize *from* DOM nodes so |
| 1172 |
// that we don't have to use innerHTML on template recursion but instead we |
| 1173 |
// can use importNode. The API could also be public as it is still a way to |
| 1174 |
// make a document fragment conform to a policy, somewhat useful. |
| 1175 |
|
| 1176 |
// Finds the parent to which cleanNode should be appended. |
| 1177 |
var dirtyParent = goog.html.sanitizer.noclobber.getParentNode(dirtyNode); |
| 1178 |
var isSanitizedParent = false; |
| 1179 |
if (dirtyParent) { |
| 1180 |
var dirtyParentNodeType = |
| 1181 |
goog.html.sanitizer.noclobber.getNodeType(dirtyParent); |
| 1182 |
var dirtyParentNodeName = |
| 1183 |
goog.html.sanitizer.noclobber.getNodeName(dirtyParent).toLowerCase(); |
| 1184 |
var dirtyGrandParent = |
| 1185 |
goog.html.sanitizer.noclobber.getParentNode(dirtyParent); |
| 1186 |
// The following checks if target is an immediate child of the inert |
| 1187 |
// parent template element |
| 1188 |
if (dirtyParentNodeType == goog.dom.NodeType.DOCUMENT_FRAGMENT && |
| 1189 |
!dirtyGrandParent) { |
| 1190 |
isSanitizedParent = true; |
| 1191 |
} else if (dirtyParentNodeName == 'body' && dirtyGrandParent) { |
| 1192 |
// The following checks if target is an immediate child of the inert |
| 1193 |
// parent HtmlDocument. |
| 1194 |
var dirtyGreatGrandParent = |
| 1195 |
goog.html.sanitizer.noclobber.getParentNode(dirtyGrandParent); |
| 1196 |
if (dirtyGreatGrandParent && |
| 1197 |
!goog.html.sanitizer.noclobber.getParentNode( |
| 1198 |
dirtyGreatGrandParent)) { |
| 1199 |
isSanitizedParent = true; |
| 1200 |
} |
| 1201 |
} |
| 1202 |
} |
| 1203 |
var target; |
| 1204 |
if (isSanitizedParent || !dirtyParent) { |
| 1205 |
target = sanitizedParent; |
| 1206 |
} else if (goog.html.sanitizer.noclobber.isNodeElement(dirtyParent)) { |
| 1207 |
target = elementMap.get(/** @type {!Element} */ (dirtyParent)); |
| 1208 |
} |
| 1209 |
if (target.content) { |
| 1210 |
target = target.content; |
| 1211 |
} |
| 1212 |
// Do not attach blacklisted tags that have been sanitized into templates. |
| 1213 |
if (!goog.html.sanitizer.noclobber.isNodeElement(cleanNode) || |
| 1214 |
!cleanNode.hasAttribute( |
| 1215 |
goog.html.sanitizer.HTML_SANITIZER_BLACKLISTED_TAG_)) { |
| 1216 |
target.appendChild(cleanNode); |
| 1217 |
} |
| 1218 |
} |
| 1219 |
if (elementMap.clear) { |
| 1220 |
elementMap.clear(); |
| 1221 |
} |
| 1222 |
return sanitizedParent; |
| 1223 |
}; |
| 1224 |
|
| 1225 |
|
| 1226 |
/** |
| 1227 |
* Gets the style container ID for the sanitized output, or creates a new random |
| 1228 |
* one. If no style container is necessary or style containment is disabled, |
| 1229 |
* returns null. |
| 1230 |
* @return {?string} |
| 1231 |
* @private |
| 1232 |
*/ |
| 1233 |
goog.html.sanitizer.HtmlSanitizer.prototype.getStyleContainerId_ = function() { |
| 1234 |
var randomStyleContainmentEnabled = |
| 1235 |
this.styleContainer_ == goog.html.sanitizer.RANDOM_CONTAINER_; |
| 1236 |
var randomStyleContainmentNecessary = |
| 1237 |
!('STYLE' in this.tagBlacklist_) && 'STYLE' in this.tagWhitelist_; |
| 1238 |
// If the builder was configured to create a random unique ID, create one, but |
| 1239 |
// do so only if STYLE is allowed to begin with. |
| 1240 |
return randomStyleContainmentEnabled && randomStyleContainmentNecessary ? |
| 1241 |
'sanitizer-' + goog.string.getRandomString() : |
| 1242 |
this.styleContainer_; |
| 1243 |
}; |
| 1244 |
|
| 1245 |
|
| 1246 |
/** |
| 1247 |
* Returns a sanitized version of a node, with no children or user-provided |
| 1248 |
* attributes. |
| 1249 |
* @param {!Node} dirtyNode |
| 1250 |
* @param {?string} styleContainer |
| 1251 |
* @return {!Node} |
| 1252 |
* @private |
| 1253 |
*/ |
| 1254 |
goog.html.sanitizer.HtmlSanitizer.prototype.sanitizeNode_ = function( |
| 1255 |
dirtyNode, styleContainer) { |
| 1256 |
// Text nodes don't need to be sanitized, unless they are children of STYLE |
| 1257 |
// and STYLE tags are allowed. |
| 1258 |
if (goog.html.sanitizer.noclobber.getNodeType(dirtyNode) == |
| 1259 |
goog.dom.NodeType.TEXT) { |
| 1260 |
var textContent = dirtyNode.data; |
| 1261 |
// If STYLE is allowed, apply a policy to its text content. Ideally |
| 1262 |
// sanitizing text content of tags shouldn't be hardcoded for STYLE, but we |
| 1263 |
// have no plans to support sanitizing the text content of other nodes for |
| 1264 |
// now. |
| 1265 |
var dirtyParent = goog.html.sanitizer.noclobber.getParentNode(dirtyNode); |
| 1266 |
if (dirtyParent && |
| 1267 |
goog.html.sanitizer.noclobber.getNodeName(dirtyParent).toLowerCase() == |
| 1268 |
'style' && |
| 1269 |
!('STYLE' in this.tagBlacklist_) && 'STYLE' in this.tagWhitelist_) { |
| 1270 |
// Note that we don't have access to the parsed CSS declarations inside a |
| 1271 |
// TEMPLATE tag, so the CSS sanitizer accepts a string and parses it |
| 1272 |
// on its own using DOMParser. |
| 1273 |
textContent = goog.html.SafeStyleSheet.unwrap( |
| 1274 |
goog.html.sanitizer.CssSanitizer.sanitizeStyleSheetString( |
| 1275 |
textContent, styleContainer, goog.bind(function(uri, propName) { |
| 1276 |
return this.networkRequestUrlPolicy_( |
| 1277 |
uri, {cssProperty: propName}); |
| 1278 |
}, this))); |
| 1279 |
} |
| 1280 |
return document.createTextNode(textContent); |
| 1281 |
} |
| 1282 |
|
| 1283 |
// Non text nodes get an empty node based on black/white lists. |
| 1284 |
var elemName = |
| 1285 |
goog.html.sanitizer.noclobber.getNodeName(dirtyNode).toUpperCase(); |
| 1286 |
var sanitized = false; |
| 1287 |
var blacklisted = false; |
| 1288 |
var cleanElemName; |
| 1289 |
if (elemName in this.tagBlacklist_) { |
| 1290 |
// If it's in the inert blacklist, replace with template (and then add a |
| 1291 |
// special data attribute to distinguish it from real template tags). |
| 1292 |
// Note that this node will not be added to the final output, i.e. the |
| 1293 |
// template tag is only an internal representation, and eventually will be |
| 1294 |
// deleted. |
| 1295 |
cleanElemName = 'template'; |
| 1296 |
blacklisted = true; |
| 1297 |
} else if (this.tagWhitelist_[elemName]) { |
| 1298 |
// If it's in the whitelist, keep as is. |
| 1299 |
cleanElemName = elemName; |
| 1300 |
} else { |
| 1301 |
// If it's not in any list, replace with span. If the relevant builder |
| 1302 |
// option is enabled, they will bear the original tag name in a data |
| 1303 |
// attribute. |
| 1304 |
cleanElemName = 'span'; |
| 1305 |
sanitized = true; |
| 1306 |
} |
| 1307 |
var cleanElem = document.createElement(cleanElemName); |
| 1308 |
if (this.shouldAddOriginalTagNames_ && sanitized) { |
| 1309 |
goog.html.sanitizer.noclobber.setElementAttribute( |
| 1310 |
cleanElem, goog.html.sanitizer.HTML_SANITIZER_SANITIZED_ATTR_NAME_, |
| 1311 |
elemName.toLowerCase()); |
| 1312 |
} |
| 1313 |
if (blacklisted) { |
| 1314 |
goog.html.sanitizer.noclobber.setElementAttribute( |
| 1315 |
cleanElem, goog.html.sanitizer.HTML_SANITIZER_BLACKLISTED_TAG_, ''); |
| 1316 |
} |
| 1317 |
return cleanElem; |
| 1318 |
}; |
| 1319 |
|
| 1320 |
|
| 1321 |
/** |
| 1322 |
* Applies sanitized versions of attributes from a dirtyElement to a |
| 1323 |
* corresponding cleanElement. |
| 1324 |
* @param {!Element} dirtyElement |
| 1325 |
* @param {!Element} cleanElement |
| 1326 |
* @private |
| 1327 |
*/ |
| 1328 |
goog.html.sanitizer.HtmlSanitizer.prototype.sanitizeAttrs_ = function( |
| 1329 |
dirtyElement, cleanElement) { |
| 1330 |
var attributes = |
| 1331 |
goog.html.sanitizer.noclobber.getElementAttributes(dirtyElement); |
| 1332 |
if (attributes == null) { |
| 1333 |
return; |
| 1334 |
} |
| 1335 |
for (var i = 0, attribute; attribute = attributes[i]; i++) { |
| 1336 |
if (attribute.specified) { |
| 1337 |
var cleanValue = this.sanitizeAttribute_(dirtyElement, attribute); |
| 1338 |
if (!goog.isNull(cleanValue)) { |
| 1339 |
goog.html.sanitizer.noclobber.setElementAttribute( |
| 1340 |
cleanElement, attribute.name, cleanValue); |
| 1341 |
} |
| 1342 |
} |
| 1343 |
} |
| 1344 |
}; |
| 1345 |
|
| 1346 |
|
| 1347 |
/** |
| 1348 |
* Sanitizes an attribute value by looking up an attribute handler for the given |
| 1349 |
* node and attribute names. |
| 1350 |
* @param {!Element} dirtyElement |
| 1351 |
* @param {!Attr} attribute |
| 1352 |
* @return {?string} sanitizedAttrValue |
| 1353 |
* @private |
| 1354 |
*/ |
| 1355 |
goog.html.sanitizer.HtmlSanitizer.prototype.sanitizeAttribute_ = function( |
| 1356 |
dirtyElement, attribute) { |
| 1357 |
var attributeName = attribute.name; |
| 1358 |
if (goog.string.startsWith( |
| 1359 |
attributeName, |
| 1360 |
goog.html.sanitizer.HTML_SANITIZER_BOOKKEEPING_PREFIX_)) { |
| 1361 |
return null; |
| 1362 |
} |
| 1363 |
|
| 1364 |
var elementName = goog.html.sanitizer.noclobber.getNodeName(dirtyElement); |
| 1365 |
var unsanitizedAttrValue = attribute.value; |
| 1366 |
|
| 1367 |
// Create policy hints object |
| 1368 |
var policyHints = { |
| 1369 |
tagName: goog.string.trim(elementName).toLowerCase(), |
| 1370 |
attributeName: goog.string.trim(attributeName).toLowerCase() |
| 1371 |
}; |
| 1372 |
var policyContext = goog.html.sanitizer.HtmlSanitizer.getContext_( |
| 1373 |
policyHints.attributeName, dirtyElement); |
| 1374 |
|
| 1375 |
// Prefer attribute handler for this specific tag. |
| 1376 |
var tagHandlerIndex = goog.html.sanitizer.HtmlSanitizer.attrIdentifier_( |
| 1377 |
elementName, attributeName); |
| 1378 |
if (tagHandlerIndex in this.attributeHandlers_) { |
| 1379 |
var handler = this.attributeHandlers_[tagHandlerIndex]; |
| 1380 |
return handler(unsanitizedAttrValue, policyHints, policyContext); |
| 1381 |
} |
| 1382 |
// Fall back on attribute handler for wildcard tag. |
| 1383 |
var genericHandlerIndex = |
| 1384 |
goog.html.sanitizer.HtmlSanitizer.attrIdentifier_(null, attributeName); |
| 1385 |
if (genericHandlerIndex in this.attributeHandlers_) { |
| 1386 |
var handler = this.attributeHandlers_[genericHandlerIndex]; |
| 1387 |
return handler(unsanitizedAttrValue, policyHints, policyContext); |
| 1388 |
} |
| 1389 |
return null; |
| 1390 |
}; |
| 1391 |
|
| 1392 |
|
| 1393 |
/** |
| 1394 |
* Processes the contents of a template tag. These are not traversed through the |
| 1395 |
* treewalker because they belong to a separate document, and thus require |
| 1396 |
* special handling. |
| 1397 |
* <p> |
| 1398 |
* If the relevant builder option is enabled and the template tag is allowed, |
| 1399 |
* this method copies the contents over to the output DOM tree without |
| 1400 |
* sanitization, otherwise the template contents are sanitized recursively. |
| 1401 |
* @param {!Element} dirtyElement |
| 1402 |
* @param {!Element} cleanElement |
| 1403 |
* @private |
| 1404 |
*/ |
| 1405 |
goog.html.sanitizer.HtmlSanitizer.prototype.maybeProcessTemplateContents_ = |
| 1406 |
function(dirtyElement, cleanElement) { |
| 1407 |
var dirtyElementName = |
| 1408 |
goog.html.sanitizer.noclobber.getNodeName(dirtyElement); |
| 1409 |
if (!goog.html.sanitizer.HTML_SANITIZER_TEMPLATE_SUPPORTED || |
| 1410 |
dirtyElementName.toLowerCase() != 'template' || |
| 1411 |
cleanElement.hasAttribute( |
| 1412 |
goog.html.sanitizer.HTML_SANITIZER_BLACKLISTED_TAG_)) { |
| 1413 |
return; |
| 1414 |
} |
| 1415 |
|
| 1416 |
// If the template element was sanitized into a span tag, do not insert |
| 1417 |
// unsanitized tags! |
| 1418 |
if (this.shouldSanitizeTemplateContents_ || |
| 1419 |
cleanElement.nodeName.toLowerCase() !== 'template') { |
| 1420 |
var dirtyElementHTML = |
| 1421 |
goog.html.sanitizer.noclobber.getElementInnerHTML(dirtyElement); |
| 1422 |
var templateSpan = this.sanitizeToDomNode(dirtyElementHTML); |
| 1423 |
// appendChild with a forEach instead of an innertHTML as the latter is |
| 1424 |
// slower. |
| 1425 |
goog.array.forEach(templateSpan.childNodes, function(node) { |
| 1426 |
cleanElement.appendChild(node); |
| 1427 |
}); |
| 1428 |
} else { |
| 1429 |
var templateDoc = |
| 1430 |
/** @type {!HTMLTemplateElement} */ (cleanElement) |
| 1431 |
.content.ownerDocument; |
| 1432 |
var dirtyCopy = |
| 1433 |
goog.asserts.assert(templateDoc.importNode(dirtyElement, true)); |
| 1434 |
var dirtyCopyChildren = |
| 1435 |
goog.html.sanitizer.noclobber.getChildNodes(dirtyCopy); |
| 1436 |
// appendChild with a forEach instead of an innerHTML as the latter is |
| 1437 |
// slower and vulnerable to mXSS. |
| 1438 |
goog.array.forEach(dirtyCopyChildren, function(node) { |
| 1439 |
cleanElement.appendChild(node); |
| 1440 |
}); |
| 1441 |
} |
| 1442 |
}; |
| 1443 |
|
| 1444 |
|
| 1445 |
/** |
| 1446 |
* Retrieves a HtmlSanitizerPolicyContext from a dirty node given an attribute |
| 1447 |
* name. |
| 1448 |
* @param {string} attributeName |
| 1449 |
* @param {!Element} dirtyElement |
| 1450 |
* @return {!goog.html.sanitizer.HtmlSanitizerPolicyContext} |
| 1451 |
* @private |
| 1452 |
*/ |
| 1453 |
goog.html.sanitizer.HtmlSanitizer.getContext_ = function( |
| 1454 |
attributeName, dirtyElement) { |
| 1455 |
var policyContext = {cssStyle: undefined}; |
| 1456 |
if (attributeName == 'style') { |
| 1457 |
policyContext.cssStyle = |
| 1458 |
goog.html.sanitizer.noclobber.getElementStyle(dirtyElement); |
| 1459 |
} |
| 1460 |
return policyContext; |
| 1461 |
}; |
| 1462 |
|
| 1463 |
|
| 1464 |
/** |
| 1465 |
* Sanitizes a HTML string using a sanitizer with default options. |
| 1466 |
* @param {string} unsanitizedHtml |
| 1467 |
* @return {!goog.html.SafeHtml} sanitizedHtml |
| 1468 |
*/ |
| 1469 |
goog.html.sanitizer.HtmlSanitizer.sanitize = function(unsanitizedHtml) { |
| 1470 |
var sanitizer = new goog.html.sanitizer.HtmlSanitizer.Builder().build(); |
| 1471 |
return sanitizer.sanitize(unsanitizedHtml); |
| 1472 |
}; |