Escape story fields and clean story HTML with HTML Purifier.

Add HTML Purifier 4.19.1. Stories are purified when they are saved and
when they are shown.

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/data.php

@@ -0,0 +1,136 @@
+<?php
+
+/**
+ * Implements data: URI for base64 encoded images supported by GD.
+ */
+class HTMLPurifier_URIScheme_data extends HTMLPurifier_URIScheme
+{
+    /**
+     * @type bool
+     */
+    public $browsable = true;
+
+    /**
+     * @type array
+     */
+    public $allowed_types = array(
+        // you better write validation code for other types if you
+        // decide to allow them
+        'image/jpeg' => true,
+        'image/gif' => true,
+        'image/png' => true,
+    );
+    // this is actually irrelevant since we only write out the path
+    // component
+    /**
+     * @type bool
+     */
+    public $may_omit_host = true;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        $result = explode(',', $uri->path, 2);
+        $is_base64 = false;
+        $charset = null;
+        $content_type = null;
+        if (count($result) == 2) {
+            list($metadata, $data) = $result;
+            // do some legwork on the metadata
+            $metas = explode(';', $metadata);
+            while (!empty($metas)) {
+                $cur = array_shift($metas);
+                if ($cur == 'base64') {
+                    $is_base64 = true;
+                    break;
+                }
+                if (substr($cur, 0, 8) == 'charset=') {
+                    // doesn't match if there are arbitrary spaces, but
+                    // whatever dude
+                    if ($charset !== null) {
+                        continue;
+                    } // garbage
+                    $charset = substr($cur, 8); // not used
+                } else {
+                    if ($content_type !== null) {
+                        continue;
+                    } // garbage
+                    $content_type = $cur;
+                }
+            }
+        } else {
+            $data = $result[0];
+        }
+        if ($content_type !== null && empty($this->allowed_types[$content_type])) {
+            return false;
+        }
+        if ($charset !== null) {
+            // error; we don't allow plaintext stuff
+            $charset = null;
+        }
+        $data = rawurldecode($data);
+        if ($is_base64) {
+            $raw_data = base64_decode($data);
+        } else {
+            $raw_data = $data;
+        }
+        if ( strlen($raw_data) < 12 ) {
+            // error; exif_imagetype throws exception with small files,
+            // and this likely indicates a corrupt URI/failed parse anyway
+            return false;
+        }
+        // XXX probably want to refactor this into a general mechanism
+        // for filtering arbitrary content types
+        if (function_exists('sys_get_temp_dir')) {
+            $file = tempnam(sys_get_temp_dir(), "");
+        } else {
+            $file = tempnam("/tmp", "");
+        }
+        file_put_contents($file, $raw_data);
+        if (function_exists('exif_imagetype')) {
+            $image_code = exif_imagetype($file);
+            unlink($file);
+        } elseif (function_exists('getimagesize')) {
+            set_error_handler(array($this, 'muteErrorHandler'));
+            $info = getimagesize($file);
+            restore_error_handler();
+            unlink($file);
+            if ($info == false) {
+                return false;
+            }
+            $image_code = $info[2];
+        } else {
+            throw new Exception("could not find exif_imagetype or getimagesize functions");
+        }
+        $real_content_type = image_type_to_mime_type($image_code);
+        if ($real_content_type != $content_type) {
+            // we're nice guys; if the content type is something else we
+            // support, change it over
+            if (empty($this->allowed_types[$real_content_type])) {
+                return false;
+            }
+            $content_type = $real_content_type;
+        }
+        // ok, it's kosher, rewrite what we need
+        $uri->userinfo = null;
+        $uri->host = null;
+        $uri->port = null;
+        $uri->fragment = null;
+        $uri->query = null;
+        $uri->path = "$content_type;base64," . base64_encode($raw_data);
+        return true;
+    }
+
+    /**
+     * @param int $errno
+     * @param string $errstr
+     */
+    public function muteErrorHandler($errno, $errstr)
+    {
+    }
+}

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/file.php

@@ -0,0 +1,44 @@
+<?php
+
+/**
+ * Validates file as defined by RFC 1630 and RFC 1738.
+ */
+class HTMLPurifier_URIScheme_file extends HTMLPurifier_URIScheme
+{
+    /**
+     * Generally file:// URLs are not accessible from most
+     * machines, so placing them as an img src is incorrect.
+     * @type bool
+     */
+    public $browsable = false;
+
+    /**
+     * Basically the *only* URI scheme for which this is true, since
+     * accessing files on the local machine is very common.  In fact,
+     * browsers on some operating systems don't understand the
+     * authority, though I hear it is used on Windows to refer to
+     * network shares.
+     * @type bool
+     */
+    public $may_omit_host = true;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        // Authentication method is not supported
+        $uri->userinfo = null;
+        // file:// makes no provisions for accessing the resource
+        $uri->port = null;
+        // While it seems to work on Firefox, the querystring has
+        // no possible effect and is thus stripped.
+        $uri->query = null;
+        return true;
+    }
+}
+
+// vim: et sw=4 sts=4

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/ftp.php

@@ -0,0 +1,58 @@
+<?php
+
+/**
+ * Validates ftp (File Transfer Protocol) URIs as defined by generic RFC 1738.
+ */
+class HTMLPurifier_URIScheme_ftp extends HTMLPurifier_URIScheme
+{
+    /**
+     * @type int
+     */
+    public $default_port = 21;
+
+    /**
+     * @type bool
+     */
+    public $browsable = true; // usually
+
+    /**
+     * @type bool
+     */
+    public $hierarchical = true;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        $uri->query = null;
+
+        // typecode check
+        $semicolon_pos = strrpos($uri->path, ';'); // reverse
+        if ($semicolon_pos !== false) {
+            $type = substr($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
+                    $uri->path .= '%3B' . $type;
+                } elseif ($typecode === 'a' || $typecode === 'i' || $typecode === 'd') {
+                    $type_ret = ";type=$typecode";
+                }
+            } else {
+                $uri->path .= '%3B' . $type;
+            }
+            $uri->path = str_replace(';', '%3B', $uri->path);
+            $uri->path .= $type_ret;
+        }
+        return true;
+    }
+}
+
+// vim: et sw=4 sts=4

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/http.php

@@ -0,0 +1,36 @@
+<?php
+
+/**
+ * Validates http (HyperText Transfer Protocol) as defined by RFC 2616
+ */
+class HTMLPurifier_URIScheme_http extends HTMLPurifier_URIScheme
+{
+    /**
+     * @type int
+     */
+    public $default_port = 80;
+
+    /**
+     * @type bool
+     */
+    public $browsable = true;
+
+    /**
+     * @type bool
+     */
+    public $hierarchical = true;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        $uri->userinfo = null;
+        return true;
+    }
+}
+
+// vim: et sw=4 sts=4

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/https.php

@@ -0,0 +1,18 @@
+<?php
+
+/**
+ * Validates https (Secure HTTP) according to http scheme.
+ */
+class HTMLPurifier_URIScheme_https extends HTMLPurifier_URIScheme_http
+{
+    /**
+     * @type int
+     */
+    public $default_port = 443;
+    /**
+     * @type bool
+     */
+    public $secure = true;
+}
+
+// vim: et sw=4 sts=4

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/mailto.php

@@ -0,0 +1,40 @@
+<?php
+
+// VERY RELAXED! Shouldn't cause problems, not even Firefox checks if the
+// email is valid, but be careful!
+
+/**
+ * Validates mailto (for E-mail) according to RFC 2368
+ * @todo Validate the email address
+ * @todo Filter allowed query parameters
+ */
+
+class HTMLPurifier_URIScheme_mailto extends HTMLPurifier_URIScheme
+{
+    /**
+     * @type bool
+     */
+    public $browsable = false;
+
+    /**
+     * @type bool
+     */
+    public $may_omit_host = true;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        $uri->userinfo = null;
+        $uri->host     = null;
+        $uri->port     = null;
+        // we need to validate path against RFC 2368's addr-spec
+        return true;
+    }
+}
+
+// vim: et sw=4 sts=4

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/news.php

@@ -0,0 +1,35 @@
+<?php
+
+/**
+ * Validates news (Usenet) as defined by generic RFC 1738
+ */
+class HTMLPurifier_URIScheme_news extends HTMLPurifier_URIScheme
+{
+    /**
+     * @type bool
+     */
+    public $browsable = false;
+
+    /**
+     * @type bool
+     */
+    public $may_omit_host = true;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        $uri->userinfo = null;
+        $uri->host = null;
+        $uri->port = null;
+        $uri->query = null;
+        // typecode check needed on path
+        return true;
+    }
+}
+
+// vim: et sw=4 sts=4

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/nntp.php

@@ -0,0 +1,32 @@
+<?php
+
+/**
+ * Validates nntp (Network News Transfer Protocol) as defined by generic RFC 1738
+ */
+class HTMLPurifier_URIScheme_nntp extends HTMLPurifier_URIScheme
+{
+    /**
+     * @type int
+     */
+    public $default_port = 119;
+
+    /**
+     * @type bool
+     */
+    public $browsable = false;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        $uri->userinfo = null;
+        $uri->query = null;
+        return true;
+    }
+}
+
+// vim: et sw=4 sts=4

lib/htmlpurifier-4.19.1/library/HTMLPurifier/URIScheme/tel.php

@@ -0,0 +1,46 @@
+<?php
+
+/**
+ * Validates tel (for phone numbers).
+ *
+ * The relevant specifications for this protocol are RFC 3966 and RFC 5341,
+ * but this class takes a much simpler approach: we normalize phone
+ * numbers so that they only include (possibly) a leading plus,
+ * and then any number of digits and x'es.
+ */
+
+class HTMLPurifier_URIScheme_tel extends HTMLPurifier_URIScheme
+{
+    /**
+     * @type bool
+     */
+    public $browsable = false;
+
+    /**
+     * @type bool
+     */
+    public $may_omit_host = true;
+
+    /**
+     * @param HTMLPurifier_URI $uri
+     * @param HTMLPurifier_Config $config
+     * @param HTMLPurifier_Context $context
+     * @return bool
+     */
+    public function doValidate(&$uri, $config, $context)
+    {
+        $uri->userinfo = null;
+        $uri->host     = null;
+        $uri->port     = null;
+
+        // Delete all non-numeric characters, commas, and non-x characters
+        // from phone number, EXCEPT for a leading plus sign.
+        $uri->path = preg_replace('/(?!^\+)[^\dx,]/', '',
+                     // Normalize e(x)tension to lower-case
+                     str_replace('X', 'x', rawurldecode($uri->path)));
+
+        return true;
+    }
+}
+
+// vim: et sw=4 sts=4