1
0
mirror of https://github.com/ezyang/htmlpurifier.git synced 2025-08-08 15:16:54 +02:00

Release 2.1.0, merged in 1255 to HEAD.

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/branches/strict@1368 48356398-32a2-884e-a903-53898d9a118a
This commit is contained in:
Edward Z. Yang
2007-08-05 02:02:46 +00:00
parent 503e76081b
commit 80c60bb9b5
141 changed files with 4250 additions and 1155 deletions

View File

@@ -38,19 +38,24 @@ class HTMLPurifier_AttrDef_CSS_FontFamily extends HTMLPurifier_AttrDef
$quote = $font[0];
if ($font[$length - 1] !== $quote) continue;
$font = substr($font, 1, $length - 2);
// double-backslash processing is buggy
$font = str_replace("\\$quote", $quote, $font); // de-escape quote
$font = str_replace("\\\n", "\n", $font); // de-escape newlines
}
// process font
// $font is a pure representation of the font name
if (ctype_alnum($font)) {
// very simple font, allow it in unharmed
$final .= $font . ', ';
continue;
}
$nospace = str_replace(array(' ', '.', '!'), '', $font);
if (ctype_alnum($nospace)) {
// font with spaces in it
$final .= "'$font', ";
continue;
}
// complicated font, requires quoting
// armor single quotes and new lines
$font = str_replace("'", "\\'", $font);
$font = str_replace("\n", "\\\n", $font);
$final .= "'$font', ";
}
$final = rtrim($final, ', ');
if ($final === '') return false;

View File

@@ -15,7 +15,7 @@ class HTMLPurifier_AttrDef_CSS_URI extends HTMLPurifier_AttrDef_URI
{
function HTMLPurifier_AttrDef_CSS_URI() {
$this->HTMLPurifier_AttrDef_URI(true); // always embedded
parent::HTMLPurifier_AttrDef_URI(true); // always embedded
}
function validate($uri_string, $config, &$context) {

View File

@@ -1,90 +1,65 @@
<?php
require_once 'HTMLPurifier/AttrDef.php';
require_once 'HTMLPurifier/URIParser.php';
require_once 'HTMLPurifier/URIScheme.php';
require_once 'HTMLPurifier/URISchemeRegistry.php';
require_once 'HTMLPurifier/AttrDef/URI/Host.php';
require_once 'HTMLPurifier/PercentEncoder.php';
HTMLPurifier_ConfigSchema::define(
'URI', 'DefaultScheme', 'http', 'string',
'Defines through what scheme the output will be served, in order to '.
'select the proper object validator when no scheme information is present.'
);
// special case filtering directives
HTMLPurifier_ConfigSchema::define(
'URI', 'Host', null, 'string/null',
'Defines the domain name of the server, so we can determine whether or '.
'an absolute URI is from your website or not. Not strictly necessary, '.
'as users should be using relative URIs to reference resources on your '.
'website. It will, however, let you use absolute URIs to link to '.
'subdomains of the domain you post here: i.e. example.com will allow '.
'sub.example.com. However, higher up domains will still be excluded: '.
'if you set %URI.Host to sub.example.com, example.com will be blocked. '.
'This directive has been available since 1.2.0.'
);
'URI', 'Munge', null, 'string/null', '
<p>
Munges all browsable (usually http, https and ftp)
absolute URI\'s into another URI, usually a URI redirection service.
This directive accepts a URI, formatted with a <code>%s</code> where
the url-encoded original URI should be inserted (sample:
<code>http://www.google.com/url?q=%s</code>).
</p>
<p>
Uses for this directive:
</p>
<ul>
<li>
Prevent PageRank leaks, while being fairly transparent
to users (you may also want to add some client side JavaScript to
override the text in the statusbar). <strong>Notice</strong>:
Many security experts believe that this form of protection does not deter spam-bots.
</li>
<li>
Redirect users to a splash page telling them they are leaving your
website. While this is poor usability practice, it is often mandated
in corporate environments.
</li>
</ul>
<p>
This directive has been available since 1.3.0.
</p>
');
// disabling directives
HTMLPurifier_ConfigSchema::define(
'URI', 'DisableExternal', false, 'bool',
'Disables links to external websites. This is a highly effective '.
'anti-spam and anti-pagerank-leech measure, but comes at a hefty price: no'.
'links or images outside of your domain will be allowed. Non-linkified '.
'URIs will still be preserved. If you want to be able to link to '.
'subdomains or use absolute URIs, specify %URI.Host for your website. '.
'This directive has been available since 1.2.0.'
);
HTMLPurifier_ConfigSchema::define(
'URI', 'DisableExternalResources', false, 'bool',
'Disables the embedding of external resources, preventing users from '.
'embedding things like images from other hosts. This prevents '.
'access tracking (good for email viewers), bandwidth leeching, '.
'cross-site request forging, goatse.cx posting, and '.
'other nasties, but also results in '.
'a loss of end-user functionality (they can\'t directly post a pic '.
'they posted from Flickr anymore). Use it if you don\'t have a '.
'robust user-content moderation team. This directive has been '.
'available since 1.3.0.'
);
HTMLPurifier_ConfigSchema::define(
'URI', 'DisableResources', false, 'bool',
'Disables embedding resources, essentially meaning no pictures. You can '.
'still link to them though. See %URI.DisableExternalResources for why '.
'this might be a good idea. This directive has been available since 1.3.0.'
);
HTMLPurifier_ConfigSchema::define(
'URI', 'Munge', null, 'string/null',
'Munges all browsable (usually http, https and ftp) URI\'s into some URL '.
'redirection service. Pass this directive a URI, with %s inserted where '.
'the url-encoded original URI should be inserted (sample: '.
'<code>http://www.google.com/url?q=%s</code>). '.
'This prevents PageRank leaks, while being as transparent as possible '.
'to users (you may also want to add some client side JavaScript to '.
'override the text in the statusbar). Warning: many security experts '.
'believe that this form of protection does not deter spam-bots. '.
'You can also use this directive to redirect users to a splash page '.
'telling them they are leaving your website. '.
'This directive has been available since 1.3.0.'
);
HTMLPurifier_ConfigSchema::define(
'URI', 'HostBlacklist', array(), 'list',
'List of strings that are forbidden in the host of any URI. Use it to '.
'kill domain names of spam, etc. Note that it will catch anything in '.
'the domain, so <tt>moo.com</tt> will catch <tt>moo.com.example.com</tt>. '.
'This directive has been available since 1.3.0.'
);
HTMLPurifier_ConfigSchema::define(
'URI', 'Disable', false, 'bool',
'Disables all URIs in all forms. Not sure why you\'d want to do that '.
'(after all, the Internet\'s founded on the notion of a hyperlink). '.
'This directive has been available since 1.3.0.'
);
'URI', 'Disable', false, 'bool', '
<p>
Disables all URIs in all forms. Not sure why you\'d want to do that
(after all, the Internet\'s founded on the notion of a hyperlink).
This directive has been available since 1.3.0.
</p>
');
HTMLPurifier_ConfigSchema::defineAlias('Attr', 'DisableURI', 'URI', 'Disable');
HTMLPurifier_ConfigSchema::define(
'URI', 'DisableResources', false, 'bool', '
<p>
Disables embedding resources, essentially meaning no pictures. You can
still link to them though. See %URI.DisableExternalResources for why
this might be a good idea. This directive has been available since 1.3.0.
</p>
');
/**
* Validates a URI as defined by RFC 3986.
* @note Scheme-specific mechanics deferred to HTMLPurifier_URIScheme
@@ -92,214 +67,83 @@ HTMLPurifier_ConfigSchema::defineAlias('Attr', 'DisableURI', 'URI', 'Disable');
class HTMLPurifier_AttrDef_URI extends HTMLPurifier_AttrDef
{
var $host;
var $embeds_resource;
var $parser, $percentEncoder;
var $embedsResource;
/**
* @param $embeds_resource_resource Does the URI here result in an extra HTTP request?
*/
function HTMLPurifier_AttrDef_URI($embeds_resource = false) {
$this->host = new HTMLPurifier_AttrDef_URI_Host();
$this->embeds_resource = (bool) $embeds_resource;
$this->parser = new HTMLPurifier_URIParser();
$this->percentEncoder = new HTMLPurifier_PercentEncoder();
$this->embedsResource = (bool) $embeds_resource;
}
function validate($uri, $config, &$context) {
static $PercentEncoder = null;
if ($PercentEncoder === null) $PercentEncoder = new HTMLPurifier_PercentEncoder();
// We'll write stack-based parsers later, for now, use regexps to
// get things working as fast as possible (irony)
if ($config->get('URI', 'Disable')) return false;
// parse as CDATA
// initial operations
$uri = $this->parseCDATA($uri);
$uri = $this->percentEncoder->normalize($uri);
// fix up percent-encoding
$uri = $PercentEncoder->normalize($uri);
// parse the URI
$uri = $this->parser->parse($uri);
if ($uri === false) return false;
// while it would be nice to use parse_url(), that's specifically
// for HTTP and thus won't work for our generic URI parsing
// add embedded flag to context for validators
$context->register('EmbeddedURI', $this->embedsResource);
// according to the RFC... (but this cuts corners, i.e. non-validating)
$r_URI = '!'.
'(([^:/?#<>\'"]+):)?'. // 2. Scheme
'(//([^/?#<>\'"]*))?'. // 4. Authority
'([^?#<>\'"]*)'. // 5. Path
'(\?([^#<>\'"]*))?'. // 7. Query
'(#([^<>\'"]*))?'. // 8. Fragment
'!';
$matches = array();
$result = preg_match($r_URI, $uri, $matches);
if (!$result) return false; // invalid URI
// seperate out parts
$scheme = !empty($matches[1]) ? $matches[2] : null;
$authority = !empty($matches[3]) ? $matches[4] : null;
$path = $matches[5]; // always present, can be empty
$query = !empty($matches[6]) ? $matches[7] : null;
$fragment = !empty($matches[8]) ? $matches[9] : null;
$registry =& HTMLPurifier_URISchemeRegistry::instance();
if ($scheme !== null) {
// no need to validate the scheme's fmt since we do that when we
// retrieve the specific scheme object from the registry
$scheme = ctype_lower($scheme) ? $scheme : strtolower($scheme);
$scheme_obj = $registry->getScheme($scheme, $config, $context);
if (!$scheme_obj) return false; // invalid scheme, clean it out
} else {
$scheme_obj = $registry->getScheme(
$config->get('URI', 'DefaultScheme'), $config, $context
);
}
// something funky weird happened in the registry, abort!
if (!$scheme_obj) {
trigger_error(
'Default scheme object "' . $config->get('URI', 'DefaultScheme') . '" was not readable',
E_USER_WARNING
);
return false;
}
// the URI we're processing embeds_resource a resource in the page, but the URI
// it references cannot be located
if ($this->embeds_resource && !$scheme_obj->browsable) {
return false;
}
if ($authority !== null) {
$ok = false;
do {
// remove URI if it's absolute and we disabled externals or
// if it's absolute and embedded and we disabled external resources
unset($our_host);
if (
$config->get('URI', 'DisableExternal') ||
(
$config->get('URI', 'DisableExternalResources') &&
$this->embeds_resource
)
) {
$our_host = $config->get('URI', 'Host');
if ($our_host === null) return false;
// generic validation
$result = $uri->validate($config, $context);
if (!$result) break;
// chained validation
$uri_def =& $config->getDefinition('URI');
$result = $uri_def->filter($uri, $config, $context);
if (!$result) break;
// scheme-specific validation
$scheme_obj = $uri->getSchemeObj($config, $context);
if (!$scheme_obj) break;
if ($this->embedsResource && !$scheme_obj->browsable) break;
$result = $scheme_obj->validate($uri, $config, $context);
if (!$result) break;
// survived gauntlet
$ok = true;
} while (false);
$context->destroy('EmbeddedURI');
if (!$ok) return false;
// munge scheme off if necessary (this must be last)
if (!is_null($uri->scheme) && is_null($uri->host)) {
if ($uri_def->defaultScheme == $uri->scheme) {
$uri->scheme = null;
}
$HEXDIG = '[A-Fa-f0-9]';
$unreserved = 'A-Za-z0-9-._~'; // make sure you wrap with []
$sub_delims = '!$&\'()'; // needs []
$pct_encoded = "%$HEXDIG$HEXDIG";
$r_userinfo = "(?:[$unreserved$sub_delims:]|$pct_encoded)*";
$r_authority = "/^(($r_userinfo)@)?(\[[^\]]+\]|[^:]*)(:(\d*))?/";
$matches = array();
preg_match($r_authority, $authority, $matches);
// overloads regexp!
$userinfo = !empty($matches[1]) ? $matches[2] : null;
$host = !empty($matches[3]) ? $matches[3] : null;
$port = !empty($matches[4]) ? $matches[5] : null;
// validate port
if ($port !== null) {
$port = (int) $port;
if ($port < 1 || $port > 65535) $port = null;
}
$host = $this->host->validate($host, $config, $context);
if ($host === false) $host = null;
if ($this->checkBlacklist($host, $config, $context)) return false;
// more lenient absolute checking
if (isset($our_host)) {
$host_parts = array_reverse(explode('.', $host));
// could be cached
$our_host_parts = array_reverse(explode('.', $our_host));
foreach ($our_host_parts as $i => $discard) {
if (!isset($host_parts[$i])) return false;
if ($host_parts[$i] != $our_host_parts[$i]) return false;
}
}
// userinfo and host are validated within the regexp
} else {
$port = $host = $userinfo = null;
}
// back to string
$result = $uri->toString();
// query and fragment are quite simple in terms of definition:
// *( pchar / "/" / "?" ), so define their validation routines
// when we start fixing percent encoding
// path gets to be validated against a hodge-podge of rules depending
// on the status of authority and scheme, but it's not that important,
// esp. since it won't be applicable to everyone
// okay, now we defer execution to the subobject for more processing
// note that $fragment is omitted
list($userinfo, $host, $port, $path, $query) =
$scheme_obj->validateComponents(
$userinfo, $host, $port, $path, $query, $config, $context
);
// reconstruct authority
$authority = null;
if (!is_null($userinfo) || !is_null($host) || !is_null($port)) {
$authority = '';
if($userinfo !== null) $authority .= $userinfo . '@';
$authority .= $host;
if($port !== null) $authority .= ':' . $port;
}
// reconstruct the result
$result = '';
if ($scheme !== null) $result .= "$scheme:";
if ($authority !== null) $result .= "//$authority";
$result .= $path;
if ($query !== null) $result .= "?$query";
if ($fragment !== null) $result .= "#$fragment";
// munge if necessary
$munge = $config->get('URI', 'Munge');
if (!empty($scheme_obj->browsable) && $munge !== null) {
if ($authority !== null) {
$result = str_replace('%s', rawurlencode($result), $munge);
}
// munge entire URI if necessary
if (
!is_null($uri->host) && // indicator for authority
!empty($scheme_obj->browsable) &&
!is_null($munge = $config->get('URI', 'Munge'))
) {
$result = str_replace('%s', rawurlencode($result), $munge);
}
return $result;
}
/**
* Checks a host against an array blacklist
* @param $host Host to check
* @param $config HTMLPurifier_Config instance
* @param $context HTMLPurifier_Context instance
* @return bool Is spam?
*/
function checkBlacklist($host, &$config, &$context) {
$blacklist = $config->get('URI', 'HostBlacklist');
if (!empty($blacklist)) {
foreach($blacklist as $blacklisted_host_fragment) {
if (strpos($host, $blacklisted_host_fragment) !== false) {
return true;
}
}
}
return false;
}
}

View File

@@ -5,6 +5,7 @@ require_once 'HTMLPurifier/ConfigSchema.php';
// member variables
require_once 'HTMLPurifier/HTMLDefinition.php';
require_once 'HTMLPurifier/CSSDefinition.php';
require_once 'HTMLPurifier/URIDefinition.php';
require_once 'HTMLPurifier/Doctype.php';
require_once 'HTMLPurifier/DefinitionCacheFactory.php';
@@ -41,7 +42,7 @@ class HTMLPurifier_Config
/**
* HTML Purifier's version
*/
var $version = '2.0.1';
var $version = '2.1.1';
/**
* Two-level associative array of configuration directives
@@ -75,6 +76,11 @@ class HTMLPurifier_Config
*/
var $serials = array();
/**
* Serial for entire configuration object
*/
var $serial;
/**
* @param $definition HTMLPurifier_ConfigSchema that defines what directives
* are allowed.
@@ -98,7 +104,6 @@ class HTMLPurifier_Config
$ret = HTMLPurifier_Config::createDefault();
if (is_string($config)) $ret->loadIni($config);
elseif (is_array($config)) $ret->loadArray($config);
if (isset($revision)) $ret->revision = $revision;
return $ret;
}
@@ -165,6 +170,17 @@ class HTMLPurifier_Config
return $this->serials[$namespace];
}
/**
* Returns a md5 signature for the entire configuration object
* that uniquely identifies that particular configuration
*/
function getSerial() {
if (empty($this->serial)) {
$this->serial = md5(serialize($this->getAll()));
}
return $this->serial;
}
/**
* Retrieves all directives, organized by namespace
*/
@@ -295,6 +311,8 @@ class HTMLPurifier_Config
$this->definitions[$type] = new HTMLPurifier_HTMLDefinition();
} elseif ($type == 'CSS') {
$this->definitions[$type] = new HTMLPurifier_CSSDefinition();
} elseif ($type == 'URI') {
$this->definitions[$type] = new HTMLPurifier_URIDefinition();
} else {
trigger_error("Definition of $type type not supported");
$false = false;
@@ -393,6 +411,26 @@ class HTMLPurifier_Config
* @static
*/
static function loadArrayFromForm($array, $index, $allowed = true, $mq_fix = true) {
$ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix);
$config = HTMLPurifier_Config::create($ret);
return $config;
}
/**
* Merges in configuration values from $_GET/$_POST to object. NOT STATIC.
* @note Same parameters as loadArrayFromForm
*/
function mergeArrayFromForm($array, $index, $allowed = true, $mq_fix = true) {
$ret = HTMLPurifier_Config::prepareArrayFromForm($array, $index, $allowed, $mq_fix);
$this->loadArray($ret);
}
/**
* Prepares an array from a form into something usable for the more
* strict parts of HTMLPurifier_Config
* @static
*/
static function prepareArrayFromForm($array, $index, $allowed = true, $mq_fix = true) {
$array = (isset($array[$index]) && is_array($array[$index])) ? $array[$index] : array();
$mq = get_magic_quotes_gpc() && $mq_fix;
@@ -409,9 +447,7 @@ class HTMLPurifier_Config
$value = $mq ? stripslashes($array[$skey]) : $array[$skey];
$ret[$ns][$directive] = $value;
}
$config = HTMLPurifier_Config::create($ret);
return $config;
return $ret;
}
/**

View File

@@ -6,6 +6,8 @@ require_once 'HTMLPurifier/ConfigDef/Namespace.php';
require_once 'HTMLPurifier/ConfigDef/Directive.php';
require_once 'HTMLPurifier/ConfigDef/DirectiveAlias.php';
if (!defined('HTMLPURIFIER_SCHEMA_STRICT')) define('HTMLPURIFIER_SCHEMA_STRICT', false);
/**
* Configuration definition, defines directives and their defaults.
* @note If you update this, please update Printer_ConfigForm
@@ -49,6 +51,8 @@ class HTMLPurifier_ConfigSchema {
var $types = array(
'string' => 'String',
'istring' => 'Case-insensitive string',
'text' => 'Text',
'itext' => 'Case-insensitive text',
'int' => 'Integer',
'float' => 'Float',
'bool' => 'Boolean',
@@ -100,27 +104,30 @@ class HTMLPurifier_ConfigSchema {
* HTMLPurifier_DirectiveDef::$type for allowed values
* @param $description Description of directive for documentation
*/
static function define(
$namespace, $name, $default, $type,
$description
) {
static function define($namespace, $name, $default, $type, $description) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (!isset($def->info[$namespace])) {
trigger_error('Cannot define directive for undefined namespace',
E_USER_ERROR);
return;
}
if (!ctype_alnum($name)) {
trigger_error('Directive name must be alphanumeric',
E_USER_ERROR);
return;
}
if (empty($description)) {
trigger_error('Description must be non-empty',
E_USER_ERROR);
return;
// basic sanity checks
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!isset($def->info[$namespace])) {
trigger_error('Cannot define directive for undefined namespace',
E_USER_ERROR);
return;
}
if (!ctype_alnum($name)) {
trigger_error('Directive name must be alphanumeric',
E_USER_ERROR);
return;
}
if (empty($description)) {
trigger_error('Description must be non-empty',
E_USER_ERROR);
return;
}
}
if (isset($def->info[$namespace][$name])) {
// already defined
if (
$def->info[$namespace][$name]->type !== $type ||
$def->defaults[$namespace][$name] !== $default
@@ -129,29 +136,35 @@ class HTMLPurifier_ConfigSchema {
return;
}
} else {
// process modifiers
// needs defining
// process modifiers (OPTIMIZE!)
$type_values = explode('/', $type, 2);
$type = $type_values[0];
$modifier = isset($type_values[1]) ? $type_values[1] : false;
$allow_null = ($modifier === 'null');
if (!isset($def->types[$type])) {
trigger_error('Invalid type for configuration directive',
E_USER_ERROR);
return;
}
$default = $def->validate($default, $type, $allow_null);
if ($def->isError($default)) {
trigger_error('Default value does not match directive type',
E_USER_ERROR);
return;
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!isset($def->types[$type])) {
trigger_error('Invalid type for configuration directive',
E_USER_ERROR);
return;
}
$default = $def->validate($default, $type, $allow_null);
if ($def->isError($default)) {
trigger_error('Default value does not match directive type',
E_USER_ERROR);
return;
}
}
$def->info[$namespace][$name] =
new HTMLPurifier_ConfigDef_Directive();
$def->info[$namespace][$name]->type = $type;
$def->info[$namespace][$name]->allow_null = $allow_null;
$def->defaults[$namespace][$name] = $default;
}
if (!HTMLPURIFIER_SCHEMA_STRICT) return;
$backtrace = debug_backtrace();
$file = $def->mungeFilename($backtrace[0]['file']);
$line = $backtrace[0]['line'];
@@ -166,19 +179,21 @@ class HTMLPurifier_ConfigSchema {
*/
static function defineNamespace($namespace, $description) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (isset($def->info[$namespace])) {
trigger_error('Cannot redefine namespace', E_USER_ERROR);
return;
}
if (!ctype_alnum($namespace)) {
trigger_error('Namespace name must be alphanumeric',
E_USER_ERROR);
return;
}
if (empty($description)) {
trigger_error('Description must be non-empty',
E_USER_ERROR);
return;
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (isset($def->info[$namespace])) {
trigger_error('Cannot redefine namespace', E_USER_ERROR);
return;
}
if (!ctype_alnum($namespace)) {
trigger_error('Namespace name must be alphanumeric',
E_USER_ERROR);
return;
}
if (empty($description)) {
trigger_error('Description must be non-empty',
E_USER_ERROR);
return;
}
}
$def->info[$namespace] = array();
$def->info_namespace[$namespace] = new HTMLPurifier_ConfigDef_Namespace();
@@ -199,23 +214,25 @@ class HTMLPurifier_ConfigSchema {
*/
static function defineValueAliases($namespace, $name, $aliases) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (!isset($def->info[$namespace][$name])) {
if (HTMLPURIFIER_SCHEMA_STRICT && !isset($def->info[$namespace][$name])) {
trigger_error('Cannot set value alias for non-existant directive',
E_USER_ERROR);
return;
}
foreach ($aliases as $alias => $real) {
if (!$def->info[$namespace][$name] !== true &&
!isset($def->info[$namespace][$name]->allowed[$real])
) {
trigger_error('Cannot define alias to value that is not allowed',
E_USER_ERROR);
return;
}
if (isset($def->info[$namespace][$name]->allowed[$alias])) {
trigger_error('Cannot define alias over allowed value',
E_USER_ERROR);
return;
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!$def->info[$namespace][$name] !== true &&
!isset($def->info[$namespace][$name]->allowed[$real])
) {
trigger_error('Cannot define alias to value that is not allowed',
E_USER_ERROR);
return;
}
if (isset($def->info[$namespace][$name]->allowed[$alias])) {
trigger_error('Cannot define alias over allowed value',
E_USER_ERROR);
return;
}
}
$def->info[$namespace][$name]->aliases[$alias] = $real;
}
@@ -230,14 +247,14 @@ class HTMLPurifier_ConfigSchema {
*/
static function defineAllowedValues($namespace, $name, $allowed_values) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (!isset($def->info[$namespace][$name])) {
if (HTMLPURIFIER_SCHEMA_STRICT && !isset($def->info[$namespace][$name])) {
trigger_error('Cannot define allowed values for undefined directive',
E_USER_ERROR);
return;
}
$directive =& $def->info[$namespace][$name];
$type = $directive->type;
if ($type != 'string' && $type != 'istring') {
if (HTMLPURIFIER_SCHEMA_STRICT && $type != 'string' && $type != 'istring') {
trigger_error('Cannot define allowed values for directive whose type is not string',
E_USER_ERROR);
return;
@@ -248,8 +265,11 @@ class HTMLPurifier_ConfigSchema {
foreach ($allowed_values as $value) {
$directive->allowed[$value] = true;
}
if ($def->defaults[$namespace][$name] !== null &&
!isset($directive->allowed[$def->defaults[$namespace][$name]])) {
if (
HTMLPURIFIER_SCHEMA_STRICT &&
$def->defaults[$namespace][$name] !== null &&
!isset($directive->allowed[$def->defaults[$namespace][$name]])
) {
trigger_error('Default value must be in allowed range of variables',
E_USER_ERROR);
$directive->allowed = true; // undo undo!
@@ -267,30 +287,32 @@ class HTMLPurifier_ConfigSchema {
*/
static function defineAlias($namespace, $name, $new_namespace, $new_name) {
$def =& HTMLPurifier_ConfigSchema::instance();
if (!isset($def->info[$namespace])) {
trigger_error('Cannot define directive alias in undefined namespace',
E_USER_ERROR);
return;
}
if (!ctype_alnum($name)) {
trigger_error('Directive name must be alphanumeric',
E_USER_ERROR);
return;
}
if (isset($def->info[$namespace][$name])) {
trigger_error('Cannot define alias over directive',
E_USER_ERROR);
return;
}
if (!isset($def->info[$new_namespace][$new_name])) {
trigger_error('Cannot define alias to undefined directive',
E_USER_ERROR);
return;
}
if ($def->info[$new_namespace][$new_name]->class == 'alias') {
trigger_error('Cannot define alias to alias',
E_USER_ERROR);
return;
if (HTMLPURIFIER_SCHEMA_STRICT) {
if (!isset($def->info[$namespace])) {
trigger_error('Cannot define directive alias in undefined namespace',
E_USER_ERROR);
return;
}
if (!ctype_alnum($name)) {
trigger_error('Directive name must be alphanumeric',
E_USER_ERROR);
return;
}
if (isset($def->info[$namespace][$name])) {
trigger_error('Cannot define alias over directive',
E_USER_ERROR);
return;
}
if (!isset($def->info[$new_namespace][$new_name])) {
trigger_error('Cannot define alias to undefined directive',
E_USER_ERROR);
return;
}
if ($def->info[$new_namespace][$new_name]->class == 'alias') {
trigger_error('Cannot define alias to alias',
E_USER_ERROR);
return;
}
}
$def->info[$namespace][$name] =
new HTMLPurifier_ConfigDef_DirectiveAlias(
@@ -313,8 +335,10 @@ class HTMLPurifier_ConfigSchema {
return $var;
case 'istring':
case 'string':
case 'text': // no difference, just is longer/multiple line string
case 'itext':
if (!is_string($var)) break;
if ($type === 'istring') $var = strtolower($var);
if ($type === 'istring' || $type === 'itext') $var = strtolower($var);
return $var;
case 'int':
if (is_string($var) && ctype_digit($var)) $var = (int) $var;
@@ -345,9 +369,13 @@ class HTMLPurifier_ConfigSchema {
// a single empty string item, but having an empty
// array is more intuitive
if ($var == '') return array();
// simplistic string to array method that only works
// for simple lists of tag names or alphanumeric characters
$var = explode(',',$var);
if (strpos($var, "\n") === false && strpos($var, "\r") === false) {
// simplistic string to array method that only works
// for simple lists of tag names or alphanumeric characters
$var = explode(',',$var);
} else {
$var = preg_split('/(,|[\n\r]+)/', $var);
}
// remove spaces
foreach ($var as $i => $j) $var[$i] = trim($j);
if ($type === 'hash') {
@@ -388,6 +416,7 @@ class HTMLPurifier_ConfigSchema {
* Takes an absolute path and munges it into a more manageable relative path
*/
function mungeFilename($filename) {
if (!HTMLPURIFIER_SCHEMA_STRICT) return $filename;
$offset = strrpos($filename, 'HTMLPurifier');
$filename = substr($filename, $offset);
$filename = str_replace('\\', '/', $filename);

View File

@@ -5,6 +5,7 @@ require_once 'HTMLPurifier/ChildDef.php';
require_once 'HTMLPurifier/ChildDef/Empty.php';
require_once 'HTMLPurifier/ChildDef/Required.php';
require_once 'HTMLPurifier/ChildDef/Optional.php';
require_once 'HTMLPurifier/ChildDef/Custom.php';
// NOT UNIT TESTED!!!

View File

@@ -99,7 +99,7 @@ class HTMLPurifier_DefinitionCache_Serializer extends
*/
function generateBaseDirectoryPath($config) {
$base = $config->get('Cache', 'SerializerPath');
$base = is_null($base) ? dirname(__FILE__) . '/Serializer' : $base;
$base = is_null($base) ? HTMLPURIFIER_PREFIX . '/HTMLPurifier/DefinitionCache/Serializer' : $base;
return $base;
}

View File

@@ -19,7 +19,7 @@ class HTMLPurifier_EntityLookup {
*/
function setup($file = false) {
if (!$file) {
$file = dirname(__FILE__) . '/EntityLookup/entities.ser';
$file = HTMLPURIFIER_PREFIX . '/HTMLPurifier/EntityLookup/entities.ser';
}
$this->table = unserialize(file_get_contents($file));
}

View File

@@ -110,12 +110,13 @@ HTMLPurifier_ConfigSchema::define(
');
HTMLPurifier_ConfigSchema::define(
'HTML', 'Allowed', null, 'string/null', '
'HTML', 'Allowed', null, 'itext/null', '
<p>
This is a convenience directive that rolls the functionality of
%HTML.AllowedElements and %HTML.AllowedAttributes into one directive.
Specify elements and attributes that are allowed using:
<code>element1[attr1|attr2],element2...</code>.
<code>element1[attr1|attr2],element2...</code>. You can also use
newlines instead of commas to separate elements.
</p>
<p>
<strong>Warning</strong>:
@@ -426,8 +427,9 @@ class HTMLPurifier_HTMLDefinition extends HTMLPurifier_Definition
$elements = array();
$attributes = array();
$chunks = explode(',', $list);
$chunks = preg_split('/(,|[\n\r]+)/', $list);
foreach ($chunks as $chunk) {
if (empty($chunk)) continue;
// remove TinyMCE element control characters
if (!strpos($chunk, '[')) {
$element = $chunk;

View File

@@ -0,0 +1,28 @@
<?php
require_once 'HTMLPurifier/HTMLModule.php';
/**
* XHTML 1.1 Ruby Annotation Module, defines elements that indicate
* short runs of text alongside base text for annotation or pronounciation.
*/
class HTMLPurifier_HTMLModule_Ruby extends HTMLPurifier_HTMLModule
{
var $name = 'Ruby';
function HTMLPurifier_HTMLModule_Ruby() {
$this->addElement('ruby', true, 'Inline',
'Custom: ((rb, (rt | (rp, rt, rp))) | (rbc, rtc, rtc?))',
'Common');
$this->addElement('rbc', true, false, 'Required: rb', 'Common');
$this->addElement('rtc', true, false, 'Required: rt', 'Common');
$rb =& $this->addElement('rb', true, false, 'Inline', 'Common');
$rb->excludes = array('ruby' => true);
$rt =& $this->addElement('rt', true, false, 'Inline', 'Common', array('rbspan' => 'Number'));
$rt->excludes = array('ruby' => true);
$this->addElement('rp', true, false, 'Optional: #PCDATA', 'Common');
}
}

View File

@@ -28,6 +28,7 @@ require_once 'HTMLPurifier/HTMLModule/Target.php';
require_once 'HTMLPurifier/HTMLModule/Scripting.php';
require_once 'HTMLPurifier/HTMLModule/XMLCommonAttributes.php';
require_once 'HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php';
require_once 'HTMLPurifier/HTMLModule/Ruby.php';
// tidy modules
require_once 'HTMLPurifier/HTMLModule/Tidy.php';
@@ -215,8 +216,8 @@ class HTMLPurifier_HTMLModuleManager
$this->doctypes->register(
'XHTML 1.1', true,
array_merge($common, $xml),
array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Proprietary'), // Tidy_XHTML1_1
array_merge($common, $xml, array('Ruby')),
array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_Proprietary', 'Tidy_XHTMLStrict'), // Tidy_XHTML1_1
array(),
'-//W3C//DTD XHTML 1.1//EN',
'http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd'

View File

@@ -8,6 +8,11 @@
class HTMLPurifier_Injector
{
/**
* Advisory name of injector, this is for friendly error messages
*/
var $name;
/**
* Amount of tokens the injector needs to skip + 1. Because
* the decrement is the first thing that happens, this needs to
@@ -40,16 +45,37 @@ class HTMLPurifier_Injector
var $inputIndex;
/**
* Prepares the injector by giving it the config and context objects,
* so that important variables can be extracted and not passed via
* parameter constantly. Remember: always instantiate a new injector
* when handling a set of HTML.
* Array of elements and attributes this injector creates and therefore
* need to be allowed by the definition. Takes form of
* array('element' => array('attr', 'attr2'), 'element2')
*/
var $needed = array();
/**
* Prepares the injector by giving it the config and context objects:
* this allows references to important variables to be made within
* the injector. This function also checks if the HTML environment
* will work with the Injector: if p tags are not allowed, the
* Auto-Paragraphing injector should not be enabled.
* @param $config Instance of HTMLPurifier_Config
* @param $context Instance of HTMLPurifier_Context
* @return Boolean false if success, string of missing needed element/attribute if failure
*/
function prepare($config, &$context) {
$this->htmlDefinition = $config->getHTMLDefinition();
// perform $needed checks
foreach ($this->needed as $element => $attributes) {
if (is_int($element)) $element = $attributes;
if (!isset($this->htmlDefinition->info[$element])) return $element;
if (!is_array($attributes)) continue;
foreach ($attributes as $name) {
if (!isset($this->htmlDefinition->info[$element]->attr[$name])) return "$element.$name";
}
}
$this->currentNesting =& $context->get('CurrentNesting');
$this->inputTokens =& $context->get('InputTokens');
$this->inputIndex =& $context->get('InputIndex');
return false;
}
/**
@@ -74,12 +100,12 @@ class HTMLPurifier_Injector
/**
* Handler that is called when a text token is processed
*/
function handleText(&$token, $config, &$context) {}
function handleText(&$token) {}
/**
* Handler that is called when a start token is processed
* Handler that is called when a start or empty token is processed
*/
function handleStart(&$token, $config, &$context) {}
function handleElement(&$token) {}
}

View File

@@ -15,6 +15,11 @@ HTMLPurifier_ConfigSchema::define(
block elements in nodes that allow paragraph tags</li>
<li>There are double newlines in paragraph tags</li>
</ul>
<p>
<code>p</code> tags must be allowed for this directive to take effect.
We do not use <code>br</code> tags for paragraphing, as that is
semantically incorrect.
</p>
<p>
This directive has been available since 2.0.1.
</p>
@@ -27,13 +32,16 @@ HTMLPurifier_ConfigSchema::define(
class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
{
var $name = 'AutoParagraph';
var $needed = array('p');
function _pStart() {
$par = new HTMLPurifier_Token_Start('p');
$par->armor['MakeWellFormed_TagClosedError'] = true;
return $par;
}
function handleText(&$token, $config, &$context) {
function handleText(&$token) {
$text = $token->data;
if (empty($this->currentNesting)) {
if (!$this->allowsElement('p')) return;
@@ -79,7 +87,7 @@ class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
}
function handleStart(&$token, $config, &$context) {
function handleElement(&$token) {
// check if we're inside a tag already
if (!empty($this->currentNesting)) {
if ($this->allowsElement('p')) {
@@ -88,11 +96,19 @@ class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
// this token is already paragraph, abort
if ($token->name == 'p') return;
// check if this token is adjacent to the parent
if ($this->inputTokens[$this->inputIndex - 1]->type != 'start') {
// this token is a block level, abort
if (!$this->_isInline($token)) return;
// check if this token is adjacent to the parent token
$prev = $this->inputTokens[$this->inputIndex - 1];
if ($prev->type != 'start') {
// not adjacent, we can abort early
// add lead paragraph tag if our token is inline
if ($this->_isInline($token)) {
// and the previous tag was an end paragraph
if (
$prev->name == 'p' && $prev->type == 'end' &&
$this->_isInline($token)
) {
$token = array($this->_pStart(), $token);
}
return;
@@ -105,8 +121,8 @@ class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
$ok = false;
// maintain a mini-nesting counter, this lets us bail out
// early if possible
$j = 2; // current nesting, is two due to parent and this start
for ($i = $this->inputIndex + 1; isset($this->inputTokens[$i]); $i++) {
$j = 1; // current nesting, one is due to parent (we recalculate current token)
for ($i = $this->inputIndex; isset($this->inputTokens[$i]); $i++) {
if ($this->inputTokens[$i]->type == 'start') $j++;
if ($this->inputTokens[$i]->type == 'end') $j--;
if ($this->inputTokens[$i]->type == 'text') {
@@ -150,7 +166,14 @@ class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
$needs_start = false;
$needs_end = false;
for ($i = 0, $c = count($raw_paragraphs); $i < $c; $i++) {
$c = count($raw_paragraphs);
if ($c == 1) {
// there were no double-newlines, abort quickly
$result[] = new HTMLPurifier_Token_Text($data);
return;
}
for ($i = 0; $i < $c; $i++) {
$par = $raw_paragraphs[$i];
if (trim($par) !== '') {
$paragraphs[] = $par;

View File

@@ -6,7 +6,8 @@ HTMLPurifier_ConfigSchema::define(
'AutoFormat', 'Linkify', false, 'bool', '
<p>
This directive turns on linkification, auto-linking http, ftp and
https URLs. This directive has been available since 2.0.1.
https URLs. <code>a</code> tags with the <code>href</code> attribute
must be allowed. This directive has been available since 2.0.1.
</p>
');
@@ -16,7 +17,10 @@ HTMLPurifier_ConfigSchema::define(
class HTMLPurifier_Injector_Linkify extends HTMLPurifier_Injector
{
function handleText(&$token, $config, &$context) {
var $name = 'Linkify';
var $needed = array('a' => array('href'));
function handleText(&$token) {
if (!$this->allowsElement('a')) return;
if (strpos($token->data, '://') === false) {

View File

@@ -6,8 +6,9 @@ HTMLPurifier_ConfigSchema::define(
'AutoFormat', 'PurifierLinkify', false, 'bool', '
<p>
Internal auto-formatter that converts configuration directives in
syntax <a>%Namespace.Directive</a> to links. This directive has been available
since 2.0.1.
syntax <a>%Namespace.Directive</a> to links. <code>a</code> tags
with the <code>href</code> attribute must be allowed.
This directive has been available since 2.0.1.
</p>
');
@@ -27,14 +28,16 @@ HTMLPurifier_ConfigSchema::define(
class HTMLPurifier_Injector_PurifierLinkify extends HTMLPurifier_Injector
{
var $name = 'PurifierLinkify';
var $docURL;
var $needed = array('a' => array('href'));
function prepare($config, &$context) {
parent::prepare($config, $context);
$this->docURL = $config->get('AutoFormatParam', 'PurifierLinkifyDocURL');
return parent::prepare($config, $context);
}
function handleText(&$token, $config, &$context) {
function handleText(&$token) {
if (!$this->allowsElement('a')) return;
if (strpos($token->data, '%') === false) return;

View File

@@ -28,7 +28,7 @@ $messages = array(
'Strategy_RemoveForeignElements: Foreign element to text' => 'Unrecognized $CurrentToken.Serialized tag converted to text',
'Strategy_RemoveForeignElements: Foreign element removed' => 'Unrecognized $CurrentToken.Serialized tag removed',
'Strategy_RemoveForeignElements: Comment removed' => 'Comment containing "$CurrentToken.Data" removed',
'Strategy_RemoveForeignElements: Script removed' => 'Script removed',
'Strategy_RemoveForeignElements: Foreign meta element removed' => 'Unrecognized $CurrentToken.Serialized meta tag and all descendants removed',
'Strategy_RemoveForeignElements: Token removed to end' => 'Tags and text starting from $1 element where removed to end',
'Strategy_MakeWellFormed: Unnecessary end tag removed' => 'Unnecessary $CurrentToken.Serialized tag removed',

View File

@@ -82,7 +82,7 @@ class HTMLPurifier_LanguageFactory
*/
function setup() {
$this->validator = new HTMLPurifier_AttrDef_Lang();
$this->dir = dirname(__FILE__);
$this->dir = HTMLPURIFIER_PREFIX . '/HTMLPurifier';
}
/**

View File

@@ -66,6 +66,16 @@ HTMLPurifier_ConfigSchema::define(
</p>
');
HTMLPurifier_ConfigSchema::define(
'Core', 'AggressivelyFixLt', false, 'bool', '
This directive enables aggressive pre-filter fixes HTML Purifier can
perform in order to ensure that open angled-brackets do not get killed
during parsing stage. Enabling this will result in two preg_replace_callback
calls and one preg_replace call for every bit of HTML passed through here.
It is not necessary and will have no effect for PHP 4.
This directive has been available since 2.1.0.
');
/**
* Forgivingly lexes HTML (SGML-style) markup into tokens.
*

View File

@@ -42,6 +42,16 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
$html = $this->normalize($html, $config, $context);
// attempt to armor stray angled brackets that cannot possibly
// form tags and thus are probably being used as emoticons
if ($config->get('Core', 'AggressivelyFixLt')) {
$char = '[^a-z!\/]';
$comment = "/<!--(.*?)(-->|\z)/is";
$html = preg_replace_callback($comment, array('HTMLPurifier_Lexer_DOMLex', 'callbackArmorCommentEntities'), $html);
$html = preg_replace("/<($char)/i", '&lt;\\1', $html);
$html = preg_replace_callback($comment, array('HTMLPurifier_Lexer_DOMLex', 'callbackUndoCommentSubst'), $html); // fix comments
}
// preprocess html, essential for UTF-8
$html =
'<!DOCTYPE html '.
@@ -151,5 +161,21 @@ class HTMLPurifier_Lexer_DOMLex extends HTMLPurifier_Lexer
*/
public function muteErrorHandler($errno, $errstr) {}
/**
* Callback function for undoing escaping of stray angled brackets
* in comments
*/
static public function callbackUndoCommentSubst($matches) {
return '<!--' . strtr($matches[1], array('&amp;'=>'&','&lt;'=>'<')) . $matches[2];
}
/**
* Callback function that entity-izes ampersands in comments so that
* callbackUndoCommentSubst doesn't clobber them
*/
static public function callbackArmorCommentEntities($matches) {
return '<!--' . str_replace('&', '&amp;', $matches[1]) . $matches[2];
}
}

View File

@@ -150,6 +150,14 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
// We are in tag and it is well formed
// Grab the internals of the tag
$strlen_segment = $position_next_gt - $cursor;
if ($strlen_segment < 1) {
// there's nothing to process!
$token = new HTMLPurifier_Token_Text('<');
$cursor++;
continue;
}
$segment = substr($html, $cursor, $strlen_segment);
// Check if it's a comment
@@ -204,7 +212,8 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
// Check leading character is alnum, if not, we may
// have accidently grabbed an emoticon. Translate into
// text and go our merry way
if (!ctype_alnum($segment[0])) {
if (!ctype_alpha($segment[0])) {
// XML: $segment[0] !== '_' && $segment[0] !== ':'
if ($e) $e->send(E_NOTICE, 'Lexer: Unescaped lt');
$token = new
HTMLPurifier_Token_Text(
@@ -371,6 +380,7 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
$value = $quoted_value;
}
}
if ($value === false) $value = '';
return array($key => $value);
}
@@ -385,7 +395,6 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
// infinite loop protection
$loops = 0;
while(true) {
// infinite loop protection
@@ -399,7 +408,6 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
}
$cursor += ($value = strspn($string, $this->_whitespace, $cursor));
// grab the key
$key_begin = $cursor; //we're currently at the start of the key
@@ -435,6 +443,11 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
$cursor++;
$cursor += strspn($string, $this->_whitespace, $cursor);
if ($cursor === false) {
$array[$key] = '';
break;
}
// we might be in front of a quote right now
$char = @$string[$cursor];
@@ -452,7 +465,14 @@ class HTMLPurifier_Lexer_DirectLex extends HTMLPurifier_Lexer
$value_end = $cursor;
}
// we reached a premature end
if ($cursor === false) {
$cursor = $size;
$value_end = $cursor;
}
$value = substr($string, $value_begin, $value_end - $value_begin);
if ($value === false) $value = '';
$array[$key] = $this->parseData($value);
$cursor++;

View File

@@ -1,7 +1,7 @@
.hp-config {}
.hp-config tbody th {text-align:right;}
.hp-config tbody th {text-align:right; padding-right:0.5em;}
.hp-config thead, .hp-config .namespace {background:#3C578C; color:#FFF;}
.hp-config .namespace th {text-align:center;}
.hp-config .verbose {display:none;}

View File

@@ -23,18 +23,52 @@ class HTMLPurifier_Printer_ConfigForm extends HTMLPurifier_Printer
*/
var $name;
/**
* Whether or not to compress directive names, clipping them off
* after a certain amount of letters
*/
var $compress = false;
/**
* @param $name Form element name for directives to be stuffed into
* @param $doc_url String documentation URL, will have fragment tagged on
* @param $compress Integer max length before compressing a directive name, set to false to turn off
*/
function HTMLPurifier_Printer_ConfigForm($name, $doc_url = null) {
function HTMLPurifier_Printer_ConfigForm(
$name, $doc_url = null, $compress = false
) {
parent::HTMLPurifier_Printer();
$this->docURL = $doc_url;
$this->name = $name;
$this->compress = $compress;
$this->fields['default'] = new HTMLPurifier_Printer_ConfigForm_default();
$this->fields['bool'] = new HTMLPurifier_Printer_ConfigForm_bool();
}
/**
* @param $cols Integer columns of textarea, null to use default
* @param $rows Integer rows of textarea, null to use default
*/
function setTextareaDimensions($cols = null, $rows = null) {
if ($cols) $this->fields['default']->cols = $cols;
if ($rows) $this->fields['default']->rows = $rows;
}
/**
* Retrieves styling, in case the directory it's in is not publically
* available
*/
function getCSS() {
return file_get_contents(HTMLPURIFIER_PREFIX . '/HTMLPurifier/Printer/ConfigForm.css');
}
/**
* Retrieves JavaScript, in case directory is not public
*/
function getJavaScript() {
return file_get_contents(HTMLPURIFIER_PREFIX . '/HTMLPurifier/Printer/ConfigForm.js');
}
/**
* Returns HTML output for a configuration form
* @param $config Configuration object of current form state
@@ -98,11 +132,12 @@ class HTMLPurifier_Printer_ConfigForm extends HTMLPurifier_Printer
$ret .= $this->start('a', array('href' => $url));
}
$attr = array('for' => "{$this->name}:$ns.$directive");
// crop directive name if it's too long
if (strlen($directive) < 14) {
if (!$this->compress || (strlen($directive) < $this->compress)) {
$directive_disp = $directive;
} else {
$directive_disp = substr($directive, 0, 12) . '...';
$directive_disp = substr($directive, 0, $this->compress - 2) . '...';
$attr['title'] = $directive;
}
@@ -176,6 +211,8 @@ class HTMLPurifier_Printer_ConfigForm_NullDecorator extends HTMLPurifier_Printer
* Swiss-army knife configuration form field printer
*/
class HTMLPurifier_Printer_ConfigForm_default extends HTMLPurifier_Printer {
var $cols = 18;
var $rows = 5;
function render($ns, $directive, $value, $name, $config) {
$this->prepareGenerator($config);
// this should probably be split up a little
@@ -190,12 +227,12 @@ class HTMLPurifier_Printer_ConfigForm_default extends HTMLPurifier_Printer {
$value[] = $val;
}
case 'list':
$value = implode(',', $value);
$value = implode(PHP_EOL, $value);
break;
case 'hash':
$nvalue = '';
foreach ($value as $i => $v) {
$nvalue .= "$i:$v,";
$nvalue .= "$i:$v" . PHP_EOL;
}
$value = $nvalue;
break;
@@ -220,6 +257,15 @@ class HTMLPurifier_Printer_ConfigForm_default extends HTMLPurifier_Printer {
$ret .= $this->element('option', $val, $attr);
}
$ret .= $this->end('select');
} elseif (
$def->type == 'text' || $def->type == 'itext' ||
$def->type == 'list' || $def->type == 'hash' || $def->type == 'lookup'
) {
$attr['cols'] = $this->cols;
$attr['rows'] = $this->rows;
$ret .= $this->start('textarea', $attr);
$ret .= $this->text($value);
$ret .= $this->end('textarea');
} else {
$attr['value'] = $value;
$attr['type'] = 'text';

View File

@@ -67,7 +67,8 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
unset($injectors['Custom']); // special case
foreach ($injectors as $injector => $b) {
$injector = "HTMLPurifier_Injector_$injector";
if ($b) $this->injectors[] = new $injector;
if (!$b) continue;
$this->injectors[] = new $injector;
}
foreach ($custom_injectors as $injector) {
if (is_string($injector)) {
@@ -87,7 +88,11 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
// give the injectors references to the definition and context
// variables for performance reasons
foreach ($this->injectors as $i => $x) {
$this->injectors[$i]->prepare($config, $context);
$error = $this->injectors[$i]->prepare($config, $context);
if (!$error) continue;
list($injector) = array_splice($this->injectors, $i, 1);
$name = $injector->name;
trigger_error("Cannot enable $name injector because $error is not allowed", E_USER_WARNING);
}
// -- end INJECTOR --
@@ -109,7 +114,7 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
if ($token->type === 'text') {
// injector handler code; duplicated for performance reasons
foreach ($this->injectors as $i => $x) {
if (!$x->skip) $x->handleText($token, $config, $context);
if (!$x->skip) $x->handleText($token);
if (is_array($token)) {
$this->currentInjector = $i;
break;
@@ -122,26 +127,24 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
$info = $definition->info[$token->name]->child;
// quick checks:
// test if it claims to be a start tag but is empty
// quick tag checks: anything that's *not* an end tag
$ok = false;
if ($info->type == 'empty' && $token->type == 'start') {
$result[] = new HTMLPurifier_Token_Empty($token->name, $token->attr);
continue;
}
// test if it claims to be empty but really is a start tag
if ($info->type != 'empty' && $token->type == 'empty' ) {
$result[] = new HTMLPurifier_Token_Start($token->name, $token->attr);
$result[] = new HTMLPurifier_Token_End($token->name);
continue;
}
// automatically insert empty tags
if ($token->type == 'empty') {
$result[] = $token;
continue;
}
// start tags have precedence, so they get passed through...
if ($token->type == 'start') {
// test if it claims to be a start tag but is empty
$token = new HTMLPurifier_Token_Empty($token->name, $token->attr);
$ok = true;
} elseif ($info->type != 'empty' && $token->type == 'empty' ) {
// claims to be empty but really is a start tag
$token = array(
new HTMLPurifier_Token_Start($token->name, $token->attr),
new HTMLPurifier_Token_End($token->name)
);
$ok = true;
} elseif ($token->type == 'empty') {
// real empty token
$ok = true;
} elseif ($token->type == 'start') {
// start tag
// ...unless they also have to close their parent
if (!empty($this->currentNesting)) {
@@ -163,16 +166,18 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
$this->currentNesting[] = $parent; // undo the pop
}
// injector handler code; duplicated for performance reasons
$ok = true;
}
// injector handler code; duplicated for performance reasons
if ($ok) {
foreach ($this->injectors as $i => $x) {
if (!$x->skip) $x->handleStart($token, $config, $context);
if (!$x->skip) $x->handleElement($token);
if (is_array($token)) {
$this->currentInjector = $i;
break;
}
}
$this->processToken($token, $config, $context);
continue;
}
@@ -280,9 +285,11 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
array_splice($this->inputTokens, $this->inputIndex--, 1, $token);
// adjust the injector skips based on the array substitution
$offset = count($token) + 1;
for ($i = 0; $i <= $this->currentInjector; $i++) {
$this->injectors[$i]->skip += $offset;
if ($this->injectors) {
$offset = count($token) + 1;
for ($i = 0; $i <= $this->currentInjector; $i++) {
$this->injectors[$i]->skip += $offset;
}
}
} elseif ($token) {
// regular case

View File

@@ -8,19 +8,38 @@ require_once 'HTMLPurifier/TagTransform.php';
require_once 'HTMLPurifier/AttrValidator.php';
HTMLPurifier_ConfigSchema::define(
'Core', 'RemoveInvalidImg', true, 'bool',
'This directive enables pre-emptive URI checking in <code>img</code> '.
'tags, as the attribute validation strategy is not authorized to '.
'remove elements from the document. This directive has been available '.
'since 1.3.0, revert to pre-1.3.0 behavior by setting to false.'
'Core', 'RemoveInvalidImg', true, 'bool', '
<p>
This directive enables pre-emptive URI checking in <code>img</code>
tags, as the attribute validation strategy is not authorized to
remove elements from the document. This directive has been available
since 1.3.0, revert to pre-1.3.0 behavior by setting to false.
</p>
'
);
HTMLPurifier_ConfigSchema::define(
'Core', 'RemoveScriptContents', true, 'bool', '
'Core', 'RemoveScriptContents', null, 'bool/null', '
<p>
This directive enables HTML Purifier to remove not only script tags
but all of their contents. This directive has been available since 2.0.0,
revert to pre-2.0.0 behavior by setting to false.
but all of their contents. This directive has been deprecated since 2.1.0,
and when not set the value of %Core.HiddenElements will take
precedence. This directive has been available since 2.0.0, and can be used to
revert to pre-2.0.0 behavior by setting it to false.
</p>
'
);
HTMLPurifier_ConfigSchema::define(
'Core', 'HiddenElements', array('script' => true, 'style' => true), 'lookup', '
<p>
This directive is a lookup array of elements which should have their
contents removed when they are not allowed by the HTML definition.
For example, the contents of a <code>script</code> tag are not
normally shown in a document, so if script tags are to be removed,
their contents should be removed to. This is opposed to a <code>b</code>
tag, which defines some presentational changes but does not hide its
contents.
</p>
'
);
@@ -43,7 +62,16 @@ class HTMLPurifier_Strategy_RemoveForeignElements extends HTMLPurifier_Strategy
$escape_invalid_tags = $config->get('Core', 'EscapeInvalidTags');
$remove_invalid_img = $config->get('Core', 'RemoveInvalidImg');
$remove_script_contents = $config->get('Core', 'RemoveScriptContents');
$hidden_elements = $config->get('Core', 'HiddenElements');
// remove script contents compatibility
if ($remove_script_contents === true) {
$hidden_elements['script'] = true;
} elseif ($remove_script_contents === false && isset($hidden_elements['script'])) {
unset($hidden_elements['script']);
}
$attr_validator = new HTMLPurifier_AttrValidator();
@@ -107,7 +135,7 @@ class HTMLPurifier_Strategy_RemoveForeignElements extends HTMLPurifier_Strategy
}
// CAN BE GENERICIZED
if ($token->name == 'script' && $token->type == 'start') {
if (isset($hidden_elements[$token->name]) && $token->type == 'start') {
$textify_comments = $token->name;
} elseif ($token->name === $textify_comments && $token->type == 'end') {
$textify_comments = false;
@@ -122,7 +150,7 @@ class HTMLPurifier_Strategy_RemoveForeignElements extends HTMLPurifier_Strategy
} else {
// check if we need to destroy all of the tag's children
// CAN BE GENERICIZED
if ($token->name == 'script' && $remove_script_contents) {
if (isset($hidden_elements[$token->name])) {
if ($token->type == 'start') {
$remove_until = $token->name;
} elseif ($token->type == 'empty') {
@@ -130,7 +158,7 @@ class HTMLPurifier_Strategy_RemoveForeignElements extends HTMLPurifier_Strategy
} else {
$remove_until = false;
}
if ($e) $e->send(E_ERROR, 'Strategy_RemoveForeignElements: Script removed');
if ($e) $e->send(E_ERROR, 'Strategy_RemoveForeignElements: Foreign meta element removed');
} else {
if ($e) $e->send(E_ERROR, 'Strategy_RemoveForeignElements: Foreign element removed');
}

View File

@@ -46,6 +46,7 @@ class HTMLPurifier_Strategy_ValidateAttributes extends HTMLPurifier_Strategy
}
$context->destroy('IDAccumulator');
$context->destroy('CurrentToken');
return $tokens;
}

View File

@@ -0,0 +1,119 @@
<?php
require_once 'HTMLPurifier/URIParser.php';
require_once 'HTMLPurifier/URIFilter.php';
/**
* HTML Purifier's internal representation of a URI
*/
class HTMLPurifier_URI
{
var $scheme, $userinfo, $host, $port, $path, $query, $fragment;
/**
* @note Automatically normalizes scheme and port
*/
function HTMLPurifier_URI($scheme, $userinfo, $host, $port, $path, $query, $fragment) {
$this->scheme = is_null($scheme) || ctype_lower($scheme) ? $scheme : strtolower($scheme);
$this->userinfo = $userinfo;
$this->host = $host;
$this->port = is_null($port) ? $port : (int) $port;
$this->path = $path;
$this->query = $query;
$this->fragment = $fragment;
}
/**
* Retrieves a scheme object corresponding to the URI's scheme/default
* @param $config Instance of HTMLPurifier_Config
* @param $context Instance of HTMLPurifier_Context
* @return Scheme object appropriate for validating this URI
*/
function getSchemeObj($config, &$context) {
$registry =& HTMLPurifier_URISchemeRegistry::instance();
if ($this->scheme !== null) {
$scheme_obj = $registry->getScheme($this->scheme, $config, $context);
if (!$scheme_obj) return false; // invalid scheme, clean it out
} else {
// no scheme: retrieve the default one
$def = $config->getDefinition('URI');
$scheme_obj = $registry->getScheme($def->defaultScheme, $config, $context);
if (!$scheme_obj) {
// something funky happened to the default scheme object
trigger_error(
'Default scheme object "' . $def->defaultScheme . '" was not readable',
E_USER_WARNING
);
return false;
}
}
return $scheme_obj;
}
/**
* Generic validation method applicable for all schemes
* @param $config Instance of HTMLPurifier_Config
* @param $context Instance of HTMLPurifier_Context
* @return True if validation/filtering succeeds, false if failure
*/
function validate($config, &$context) {
// validate host
if (!is_null($this->host)) {
$host_def = new HTMLPurifier_AttrDef_URI_Host();
$this->host = $host_def->validate($this->host, $config, $context);
if ($this->host === false) $this->host = null;
}
// validate port
if (!is_null($this->port)) {
if ($this->port < 1 || $this->port > 65535) $this->port = null;
}
// query and fragment are quite simple in terms of definition:
// *( pchar / "/" / "?" ), so define their validation routines
// when we start fixing percent encoding
// path gets to be validated against a hodge-podge of rules depending
// on the status of authority and scheme, but it's not that important,
// esp. since it won't be applicable to everyone
return true;
}
/**
* Convert URI back to string
* @return String URI appropriate for output
*/
function toString() {
// reconstruct authority
$authority = null;
if (!is_null($this->host)) {
$authority = '';
if(!is_null($this->userinfo)) $authority .= $this->userinfo . '@';
$authority .= $this->host;
if(!is_null($this->port)) $authority .= ':' . $this->port;
}
// reconstruct the result
$result = '';
if (!is_null($this->scheme)) $result .= $this->scheme . ':';
if (!is_null($authority)) $result .= '//' . $authority;
$result .= $this->path;
if (!is_null($this->query)) $result .= '?' . $this->query;
if (!is_null($this->fragment)) $result .= '#' . $this->fragment;
return $result;
}
/**
* Returns a copy of the URI object
*/
function copy() {
return unserialize(serialize($this));
}
}

View File

@@ -0,0 +1,145 @@
<?php
require_once 'HTMLPurifier/Definition.php';
require_once 'HTMLPurifier/URIFilter.php';
require_once 'HTMLPurifier/URIParser.php';
require_once 'HTMLPurifier/URIFilter/DisableExternal.php';
require_once 'HTMLPurifier/URIFilter/DisableExternalResources.php';
require_once 'HTMLPurifier/URIFilter/HostBlacklist.php';
require_once 'HTMLPurifier/URIFilter/MakeAbsolute.php';
HTMLPurifier_ConfigSchema::define(
'URI', 'DefinitionID', null, 'string/null', '
<p>
Unique identifier for a custom-built URI definition. If you want
to add custom URIFilters, you must specify this value.
This directive has been available since 2.1.0.
</p>
');
HTMLPurifier_ConfigSchema::define(
'URI', 'DefinitionRev', 1, 'int', '
<p>
Revision identifier for your custom definition. See
%HTML.DefinitionRev for details. This directive has been available
since 2.1.0.
</p>
');
// informative URI directives
HTMLPurifier_ConfigSchema::define(
'URI', 'DefaultScheme', 'http', 'string', '
<p>
Defines through what scheme the output will be served, in order to
select the proper object validator when no scheme information is present.
</p>
');
HTMLPurifier_ConfigSchema::define(
'URI', 'Host', null, 'string/null', '
<p>
Defines the domain name of the server, so we can determine whether or
an absolute URI is from your website or not. Not strictly necessary,
as users should be using relative URIs to reference resources on your
website. It will, however, let you use absolute URIs to link to
subdomains of the domain you post here: i.e. example.com will allow
sub.example.com. However, higher up domains will still be excluded:
if you set %URI.Host to sub.example.com, example.com will be blocked.
<strong>Note:</strong> This directive overrides %URI.Base because
a given page may be on a sub-domain, but you wish HTML Purifier to be
more relaxed and allow some of the parent domains too.
This directive has been available since 1.2.0.
</p>
');
HTMLPurifier_ConfigSchema::define(
'URI', 'Base', null, 'string/null', '
<p>
The base URI is the URI of the document this purified HTML will be
inserted into. This information is important if HTML Purifier needs
to calculate absolute URIs from relative URIs, such as when %URI.MakeAbsolute
is on. You may use a non-absolute URI for this value, but behavior
may vary (%URI.MakeAbsolute deals nicely with both absolute and
relative paths, but forwards-compatibility is not guaranteed).
<strong>Warning:</strong> If set, the scheme on this URI
overrides the one specified by %URI.DefaultScheme. This directive has
been available since 2.1.0.
</p>
');
class HTMLPurifier_URIDefinition extends HTMLPurifier_Definition
{
var $type = 'URI';
var $filters = array();
var $registeredFilters = array();
/**
* HTMLPurifier_URI object of the base specified at %URI.Base
*/
var $base;
/**
* String host to consider "home" base
*/
var $host;
/**
* Name of default scheme based on %URI.DefaultScheme and %URI.Base
*/
var $defaultScheme;
function HTMLPurifier_URIDefinition() {
$this->registerFilter(new HTMLPurifier_URIFilter_DisableExternal());
$this->registerFilter(new HTMLPurifier_URIFilter_DisableExternalResources());
$this->registerFilter(new HTMLPurifier_URIFilter_HostBlacklist());
$this->registerFilter(new HTMLPurifier_URIFilter_MakeAbsolute());
}
function registerFilter($filter) {
$this->registeredFilters[$filter->name] = $filter;
}
function addFilter($filter, $config) {
$filter->prepare($config);
$this->filters[$filter->name] = $filter;
}
function doSetup($config) {
$this->setupMemberVariables($config);
$this->setupFilters($config);
}
function setupFilters($config) {
foreach ($this->registeredFilters as $name => $filter) {
$conf = $config->get('URI', $name);
if ($conf !== false && $conf !== null) {
$this->addFilter($filter, $config);
}
}
unset($this->registeredFilters);
}
function setupMemberVariables($config) {
$this->host = $config->get('URI', 'Host');
$base_uri = $config->get('URI', 'Base');
if (!is_null($base_uri)) {
$parser = new HTMLPurifier_URIParser();
$this->base = $parser->parse($base_uri);
$this->defaultScheme = $this->base->scheme;
if (is_null($this->host)) $this->host = $this->base->host;
}
if (is_null($this->defaultScheme)) $this->defaultScheme = $config->get('URI', 'DefaultScheme');
}
function filter(&$uri, $config, &$context) {
foreach ($this->filters as $name => $x) {
$result = $this->filters[$name]->filter($uri, $config, $context);
if (!$result) return false;
}
return true;
}
}

View File

@@ -0,0 +1,24 @@
<?php
/**
* Chainable filters for custom URI processing
*/
class HTMLPurifier_URIFilter
{
var $name;
/**
* Performs initialization for the filter
*/
function prepare($config) {}
/**
* Filter a URI object
* @param &$uri Reference to URI object
* @param $config Instance of HTMLPurifier_Config
* @param &$context Instance of HTMLPurifier_Context
*/
function filter(&$uri, $config, &$context) {
trigger_error('Cannot call abstract function', E_USER_ERROR);
}
}

View File

@@ -0,0 +1,34 @@
<?php
require_once 'HTMLPurifier/URIFilter.php';
HTMLPurifier_ConfigSchema::define(
'URI', 'DisableExternal', false, 'bool',
'Disables links to external websites. This is a highly effective '.
'anti-spam and anti-pagerank-leech measure, but comes at a hefty price: no'.
'links or images outside of your domain will be allowed. Non-linkified '.
'URIs will still be preserved. If you want to be able to link to '.
'subdomains or use absolute URIs, specify %URI.Host for your website. '.
'This directive has been available since 1.2.0.'
);
class HTMLPurifier_URIFilter_DisableExternal extends HTMLPurifier_URIFilter
{
var $name = 'DisableExternal';
var $ourHostParts = false;
function prepare($config) {
$our_host = $config->get('URI', 'Host');
if ($our_host !== null) $this->ourHostParts = array_reverse(explode('.', $our_host));
}
function filter(&$uri, $config, &$context) {
if (is_null($uri->host)) return true;
if ($this->ourHostParts === false) return false;
$host_parts = array_reverse(explode('.', $uri->host));
foreach ($this->ourHostParts as $i => $x) {
if (!isset($host_parts[$i])) return false;
if ($host_parts[$i] != $this->ourHostParts[$i]) return false;
}
return true;
}
}

View File

@@ -0,0 +1,26 @@
<?php
require_once 'HTMLPurifier/URIFilter/DisableExternal.php';
HTMLPurifier_ConfigSchema::define(
'URI', 'DisableExternalResources', false, 'bool',
'Disables the embedding of external resources, preventing users from '.
'embedding things like images from other hosts. This prevents '.
'access tracking (good for email viewers), bandwidth leeching, '.
'cross-site request forging, goatse.cx posting, and '.
'other nasties, but also results in '.
'a loss of end-user functionality (they can\'t directly post a pic '.
'they posted from Flickr anymore). Use it if you don\'t have a '.
'robust user-content moderation team. This directive has been '.
'available since 1.3.0.'
);
class HTMLPurifier_URIFilter_DisableExternalResources extends HTMLPurifier_URIFilter_DisableExternal
{
var $name = 'DisableExternalResources';
function filter(&$uri, $config, &$context) {
if (!$context->get('EmbeddedURI', true)) return true;
return parent::filter($uri, $config, $context);
}
}

View File

@@ -0,0 +1,28 @@
<?php
require_once 'HTMLPurifier/URIFilter.php';
HTMLPurifier_ConfigSchema::define(
'URI', 'HostBlacklist', array(), 'list',
'List of strings that are forbidden in the host of any URI. Use it to '.
'kill domain names of spam, etc. Note that it will catch anything in '.
'the domain, so <tt>moo.com</tt> will catch <tt>moo.com.example.com</tt>. '.
'This directive has been available since 1.3.0.'
);
class HTMLPurifier_URIFilter_HostBlacklist extends HTMLPurifier_URIFilter
{
var $name = 'HostBlacklist';
var $blacklist = array();
function prepare($config) {
$this->blacklist = $config->get('URI', 'HostBlacklist');
}
function filter(&$uri, $config, &$context) {
foreach($this->blacklist as $blacklisted_host_fragment) {
if (strpos($uri->host, $blacklisted_host_fragment) !== false) {
return false;
}
}
return true;
}
}

View File

@@ -0,0 +1,115 @@
<?php
// does not support network paths
require_once 'HTMLPurifier/URIFilter.php';
HTMLPurifier_ConfigSchema::define(
'URI', 'MakeAbsolute', false, 'bool', '
<p>
Converts all URIs into absolute forms. This is useful when the HTML
being filtered assumes a specific base path, but will actually be
viewed in a different context (and setting an alternate base URI is
not possible). %URI.Base must be set for this directive to work.
This directive has been available since 2.1.0.
</p>
');
class HTMLPurifier_URIFilter_MakeAbsolute extends HTMLPurifier_URIFilter
{
var $name = 'MakeAbsolute';
var $base;
var $basePathStack = array();
function prepare($config) {
$def = $config->getDefinition('URI');
$this->base = $def->base;
if (is_null($this->base)) {
trigger_error('URI.MakeAbsolute is being ignored due to lack of value for URI.Base configuration', E_USER_ERROR);
return;
}
$this->base->fragment = null; // fragment is invalid for base URI
$stack = explode('/', $this->base->path);
array_pop($stack); // discard last segment
$stack = $this->_collapseStack($stack); // do pre-parsing
$this->basePathStack = $stack;
}
function filter(&$uri, $config, &$context) {
if (is_null($this->base)) return true; // abort early
if (
$uri->path === '' && is_null($uri->scheme) &&
is_null($uri->host) && is_null($uri->query) && is_null($uri->fragment)
) {
// reference to current document
$uri = $this->base->copy();
return true;
}
if (!is_null($uri->scheme)) {
// absolute URI already: don't change
if (!is_null($uri->host)) return true;
$scheme_obj = $uri->getSchemeObj($config, $context);
if (!$scheme_obj->hierarchical) {
// non-hierarchal URI with explicit scheme, don't change
return true;
}
// special case: had a scheme but always is hierarchical and had no authority
}
if (!is_null($uri->host)) {
// network path, don't bother
return true;
}
if ($uri->path === '') {
$uri->path = $this->base->path;
}elseif ($uri->path[0] !== '/') {
// relative path, needs more complicated processing
$stack = explode('/', $uri->path);
$new_stack = array_merge($this->basePathStack, $stack);
$new_stack = $this->_collapseStack($new_stack);
$uri->path = implode('/', $new_stack);
}
// re-combine
$uri->scheme = $this->base->scheme;
if (is_null($uri->userinfo)) $uri->userinfo = $this->base->userinfo;
if (is_null($uri->host)) $uri->host = $this->base->host;
if (is_null($uri->port)) $uri->port = $this->base->port;
return true;
}
/**
* Resolve dots and double-dots in a path stack
* @private
*/
function _collapseStack($stack) {
$result = array();
for ($i = 0; isset($stack[$i]); $i++) {
$is_folder = false;
// absorb an internally duplicated slash
if ($stack[$i] == '' && $i && isset($stack[$i+1])) continue;
if ($stack[$i] == '..') {
if (!empty($result)) {
$segment = array_pop($result);
if ($segment === '' && empty($result)) {
// error case: attempted to back out too far:
// restore the leading slash
$result[] = '';
} elseif ($segment === '..') {
$result[] = '..'; // cannot remove .. with ..
}
} else {
// relative path, preserve the double-dots
$result[] = '..';
}
$is_folder = true;
continue;
}
if ($stack[$i] == '.') {
// silently absorb
$is_folder = true;
continue;
}
$result[] = $stack[$i];
}
if ($is_folder) $result[] = '';
return $result;
}
}

View File

@@ -0,0 +1,62 @@
<?php
require_once 'HTMLPurifier/URI.php';
/**
* Parses a URI into the components and fragment identifier as specified
* by RFC 2396.
* @todo Replace regexps with a native PHP parser
*/
class HTMLPurifier_URIParser
{
/**
* Parses a URI
* @param $uri string URI to parse
* @return HTMLPurifier_URI representation of URI
*/
function parse($uri) {
$r_URI = '!'.
'(([^:/?#<>\'"]+):)?'. // 2. Scheme
'(//([^/?#<>\'"]*))?'. // 4. Authority
'([^?#<>\'"]*)'. // 5. Path
'(\?([^#<>\'"]*))?'. // 7. Query
'(#([^<>\'"]*))?'. // 8. Fragment
'!';
$matches = array();
$result = preg_match($r_URI, $uri, $matches);
if (!$result) return false; // *really* invalid URI
// seperate out parts
$scheme = !empty($matches[1]) ? $matches[2] : null;
$authority = !empty($matches[3]) ? $matches[4] : null;
$path = $matches[5]; // always present, can be empty
$query = !empty($matches[6]) ? $matches[7] : null;
$fragment = !empty($matches[8]) ? $matches[9] : null;
// further parse authority
if ($authority !== null) {
// ridiculously inefficient: it's a stacked regex!
$HEXDIG = '[A-Fa-f0-9]';
$unreserved = 'A-Za-z0-9-._~'; // make sure you wrap with []
$sub_delims = '!$&\'()'; // needs []
$pct_encoded = "%$HEXDIG$HEXDIG";
$r_userinfo = "(?:[$unreserved$sub_delims:]|$pct_encoded)*";
$r_authority = "/^(($r_userinfo)@)?(\[[^\]]+\]|[^:]*)(:(\d*))?/";
$matches = array();
preg_match($r_authority, $authority, $matches);
$userinfo = !empty($matches[1]) ? $matches[2] : null;
$host = !empty($matches[3]) ? $matches[3] : '';
$port = !empty($matches[4]) ? (int) $matches[5] : null;
} else {
$port = $host = $userinfo = null;
}
return new HTMLPurifier_URI(
$scheme, $userinfo, $host, $port, $path, $query, $fragment);
}
}

View File

@@ -19,24 +19,24 @@ class HTMLPurifier_URIScheme
*/
var $browsable = false;
/**
* Whether or not the URI always uses <hier_part>, resolves edge cases
* with making relative URIs absolute
*/
var $hierarchical = false;
/**
* Validates the components of a URI
* @note This implementation should be called by children if they define
* a default port, as it does port processing.
* @note Fragment is omitted as that is scheme independent
* @param $userinfo User info found before at sign in authority
* @param $host Hostname in authority
* @param $port Port found after colon in authority
* @param $path Path of URI
* @param $query Query of URI, found after question mark
* @param $uri Instance of HTMLPurifier_URI
* @param $config HTMLPurifier_Config object
* @param $context HTMLPurifier_Context object
* @return Bool success or failure
*/
function validateComponents(
$userinfo, $host, $port, $path, $query, $config, &$context
) {
if ($this->default_port == $port) $port = null;
return array($userinfo, $host, $port, $path, $query);
function validate(&$uri, $config, &$context) {
if ($this->default_port == $uri->port) $uri->port = null;
return true;
}
}

View File

@@ -9,35 +9,35 @@ class HTMLPurifier_URIScheme_ftp extends HTMLPurifier_URIScheme {
var $default_port = 21;
var $browsable = true; // usually
var $hierarchical = true;
function validateComponents(
$userinfo, $host, $port, $path, $query, $config, &$context
) {
list($userinfo, $host, $port, $path, $query) =
parent::validateComponents(
$userinfo, $host, $port, $path, $query, $config, $context );
$semicolon_pos = strrpos($path, ';'); // reverse
function validate(&$uri, $config, &$context) {
parent::validate($uri, $config, $context);
$uri->query = null;
// typecode check
$semicolon_pos = strrpos($uri->path, ';'); // reverse
if ($semicolon_pos !== false) {
// typecode check
$type = substr($path, $semicolon_pos + 1); // no semicolon
$path = substr($path, 0, $semicolon_pos);
$type = substr($uri->path, $semicolon_pos + 1); // no semicolon
$uri->path = substr($uri->path, 0, $semicolon_pos);
$type_ret = '';
if (strpos($type, '=') !== false) {
// figure out whether or not the declaration is correct
list($key, $typecode) = explode('=', $type, 2);
if ($key !== 'type') {
// invalid key, tack it back on encoded
$path .= '%3B' . $type;
$uri->path .= '%3B' . $type;
} elseif ($typecode === 'a' || $typecode === 'i' || $typecode === 'd') {
$type_ret = ";type=$typecode";
}
} else {
$path .= '%3B' . $type;
$uri->path .= '%3B' . $type;
}
$path = str_replace(';', '%3B', $path);
$path .= $type_ret;
$uri->path = str_replace(';', '%3B', $uri->path);
$uri->path .= $type_ret;
}
return array($userinfo, $host, $port, $path, null);
return true;
}
}

View File

@@ -9,14 +9,12 @@ class HTMLPurifier_URIScheme_http extends HTMLPurifier_URIScheme {
var $default_port = 80;
var $browsable = true;
var $hierarchical = true;
function validateComponents(
$userinfo, $host, $port, $path, $query, $config, &$context
) {
list($userinfo, $host, $port, $path, $query) =
parent::validateComponents(
$userinfo, $host, $port, $path, $query, $config, $context );
return array(null, $host, $port, $path, $query);
function validate(&$uri, $config, &$context) {
parent::validate($uri, $config, $context);
$uri->userinfo = null;
return true;
}
}

View File

@@ -15,14 +15,13 @@ class HTMLPurifier_URIScheme_mailto extends HTMLPurifier_URIScheme {
var $browsable = false;
function validateComponents(
$userinfo, $host, $port, $path, $query, $config, &$context
) {
list($userinfo, $host, $port, $path, $query) =
parent::validateComponents(
$userinfo, $host, $port, $path, $query, $config, $context );
function validate(&$uri, $config, &$context) {
parent::validate($uri, $config, $context);
$uri->userinfo = null;
$uri->host = null;
$uri->port = null;
// we need to validate path against RFC 2368's addr-spec
return array(null, null, null, $path, $query);
return true;
}
}

View File

@@ -9,14 +9,14 @@ class HTMLPurifier_URIScheme_news extends HTMLPurifier_URIScheme {
var $browsable = false;
function validateComponents(
$userinfo, $host, $port, $path, $query, $config, &$context
) {
list($userinfo, $host, $port, $path, $query) =
parent::validateComponents(
$userinfo, $host, $port, $path, $query, $config, $context );
function validate(&$uri, $config, &$context) {
parent::validate($uri, $config, $context);
$uri->userinfo = null;
$uri->host = null;
$uri->port = null;
$uri->query = null;
// typecode check needed on path
return array(null, null, null, $path, null);
return true;
}
}

View File

@@ -10,13 +10,11 @@ class HTMLPurifier_URIScheme_nntp extends HTMLPurifier_URIScheme {
var $default_port = 119;
var $browsable = false;
function validateComponents(
$userinfo, $host, $port, $path, $query, $config, &$context
) {
list($userinfo, $host, $port, $path, $query) =
parent::validateComponents(
$userinfo, $host, $port, $path, $query, $config, $context );
return array(null, $host, $port, $path, null);
function validate(&$uri, $config, &$context) {
parent::validate($uri, $config, $context);
$uri->userinfo = null;
$uri->query = null;
return true;
}
}

View File

@@ -79,12 +79,14 @@ class HTMLPurifier_URISchemeRegistry
}
if (isset($this->schemes[$scheme])) return $this->schemes[$scheme];
if (empty($this->_dir)) $this->_dir = dirname(__FILE__) . '/URIScheme/';
if (empty($this->_dir)) $this->_dir = HTMLPURIFIER_PREFIX . '/HTMLPurifier/URIScheme/';
if (!isset($allowed_schemes[$scheme])) return $null;
@include_once $this->_dir . $scheme . '.php';
// this bit of reflection is not very efficient, and a bit
// hacky too
$class = 'HTMLPurifier_URIScheme_' . $scheme;
if (!class_exists($class)) include_once $this->_dir . $scheme . '.php';
if (!class_exists($class)) return $null;
$this->schemes[$scheme] = new $class();
return $this->schemes[$scheme];