From 010efc9b814b433bc60353caec185d905688a32b Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@madoka.volgo-balt.ru>
Date: Tue, 5 Jun 2012 21:52:21 +0400
Subject: Revert "remove htmlpurifier"

This reverts commit c21a462d52bd32737c32c29b060da03b38f1c2e6.
---
 .../library/HTMLPurifier/URIScheme/data.php        | 96 ++++++++++++++++++++++
 .../library/HTMLPurifier/URIScheme/file.php        | 32 ++++++++
 .../library/HTMLPurifier/URIScheme/ftp.php         | 42 ++++++++++
 .../library/HTMLPurifier/URIScheme/http.php        | 19 +++++
 .../library/HTMLPurifier/URIScheme/https.php       | 13 +++
 .../library/HTMLPurifier/URIScheme/mailto.php      | 27 ++++++
 .../library/HTMLPurifier/URIScheme/news.php        | 22 +++++
 .../library/HTMLPurifier/URIScheme/nntp.php        | 19 +++++
 8 files changed, 270 insertions(+)
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/data.php
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/file.php
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/ftp.php
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/http.php
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/https.php
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/mailto.php
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/news.php
 create mode 100644 lib/htmlpurifier/library/HTMLPurifier/URIScheme/nntp.php

(limited to 'lib/htmlpurifier/library/HTMLPurifier/URIScheme')

diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/data.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/data.php
new file mode 100644
index 000000000..a5c43989e
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/data.php
@@ -0,0 +1,96 @@
+<?php
+
+/**
+ * Implements data: URI for base64 encoded images supported by GD.
+ */
+class HTMLPurifier_URIScheme_data extends HTMLPurifier_URIScheme {
+
+    public $browsable = true;
+    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
+    public $may_omit_host = true;
+
+    public function doValidate(&$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;
+        }
+        // XXX probably want to refactor this into a general mechanism
+        // for filtering arbitrary content types
+        $file = tempnam("/tmp", "");
+        file_put_contents($file, $raw_data);
+        if (function_exists('exif_imagetype')) {
+            $image_code = exif_imagetype($file);
+        } elseif (function_exists('getimagesize')) {
+            set_error_handler(array($this, 'muteErrorHandler'));
+            $info = getimagesize($file);
+            restore_error_handler();
+            if ($info == false) return false;
+            $image_code = $info[2];
+        } else {
+            trigger_error("could not find exif_imagetype or getimagesize functions", E_USER_ERROR);
+        }
+        $real_content_type = image_type_to_mime_type($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($raw_data);
+        return true;
+    }
+
+    public function muteErrorHandler($errno, $errstr) {}
+
+}
+
diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/file.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/file.php
new file mode 100644
index 000000000..d74a3f198
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/file.php
@@ -0,0 +1,32 @@
+<?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.
+    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.
+    public $may_omit_host = true;
+
+    public function doValidate(&$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
diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/ftp.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/ftp.php
new file mode 100644
index 000000000..0fb2abf64
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/ftp.php
@@ -0,0 +1,42 @@
+<?php
+
+/**
+ * Validates ftp (File Transfer Protocol) URIs as defined by generic RFC 1738.
+ */
+class HTMLPurifier_URIScheme_ftp extends HTMLPurifier_URIScheme {
+
+    public $default_port = 21;
+    public $browsable = true; // usually
+    public $hierarchical = true;
+
+    public function doValidate(&$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
diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/http.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/http.php
new file mode 100644
index 000000000..959b8daff
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/http.php
@@ -0,0 +1,19 @@
+<?php
+
+/**
+ * Validates http (HyperText Transfer Protocol) as defined by RFC 2616
+ */
+class HTMLPurifier_URIScheme_http extends HTMLPurifier_URIScheme {
+
+    public $default_port = 80;
+    public $browsable = true;
+    public $hierarchical = true;
+
+    public function doValidate(&$uri, $config, $context) {
+        $uri->userinfo = null;
+        return true;
+    }
+
+}
+
+// vim: et sw=4 sts=4
diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/https.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/https.php
new file mode 100644
index 000000000..159c2874e
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/https.php
@@ -0,0 +1,13 @@
+<?php
+
+/**
+ * Validates https (Secure HTTP) according to http scheme.
+ */
+class HTMLPurifier_URIScheme_https extends HTMLPurifier_URIScheme_http {
+
+    public $default_port = 443;
+    public $secure = true;
+
+}
+
+// vim: et sw=4 sts=4
diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/mailto.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/mailto.php
new file mode 100644
index 000000000..9db4cb23f
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/mailto.php
@@ -0,0 +1,27 @@
+<?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 {
+
+    public $browsable = false;
+    public $may_omit_host = true;
+
+    public function doValidate(&$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
diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/news.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/news.php
new file mode 100644
index 000000000..84a6748d8
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/news.php
@@ -0,0 +1,22 @@
+<?php
+
+/**
+ * Validates news (Usenet) as defined by generic RFC 1738
+ */
+class HTMLPurifier_URIScheme_news extends HTMLPurifier_URIScheme {
+
+    public $browsable = false;
+    public $may_omit_host = true;
+
+    public function doValidate(&$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
diff --git a/lib/htmlpurifier/library/HTMLPurifier/URIScheme/nntp.php b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/nntp.php
new file mode 100644
index 000000000..4ccea0dfc
--- /dev/null
+++ b/lib/htmlpurifier/library/HTMLPurifier/URIScheme/nntp.php
@@ -0,0 +1,19 @@
+<?php
+
+/**
+ * Validates nntp (Network News Transfer Protocol) as defined by generic RFC 1738
+ */
+class HTMLPurifier_URIScheme_nntp extends HTMLPurifier_URIScheme {
+
+    public $default_port = 119;
+    public $browsable = false;
+
+    public function doValidate(&$uri, $config, $context) {
+        $uri->userinfo = null;
+        $uri->query    = null;
+        return true;
+    }
+
+}
+
+// vim: et sw=4 sts=4
-- 
cgit v1.2.3