From 0756833f53c50c35a467a32b67aa0a0a80943199 2014-03-04 00:10:19
From: MinRK <benjaminrk@gmail.com>
Date: 2014-03-04 00:10:19
Subject: [PATCH] sanitize CSS

rather than removing it entirely

---

diff --git a/IPython/html/static/base/js/security.js b/IPython/html/static/base/js/security.js
index 3ccdeb6..e311146 100644
--- a/IPython/html/static/base/js/security.js
+++ b/IPython/html/static/base/js/security.js
@@ -42,6 +42,7 @@ IPython.security = (function (IPython) {
     if (window && window.html) {
         caja = window.html;
         caja.html4 = window.html4;
+        caja.sanitizeStylesheet = window.sanitizeStylesheet;
     }
     
     var sanitizeAttribs = function (tagName, attribs, opt_naiveUriRewriter, opt_nmTokenPolicy, opt_logger) {
@@ -59,8 +60,37 @@ IPython.security = (function (IPython) {
         return caja.sanitizeAttribs(tagName, attribs, opt_naiveUriRewriter, opt_nmTokenPolicy, opt_logger);
     };
     
-    var sanitize = function (html, log) {
+    var sanitize_css = function (css, tagPolicy) {
+        return caja.sanitizeStylesheet(
+            window.location.pathname,
+            css,
+            {
+                containerClass: null,
+                idSuffix: '',
+                tagPolicy: tagPolicy,
+                virtualizeAttrName: noop
+            },
+            noop
+        );
+    };
+    
+    var sanitize_stylesheets = function (html, tagPolicy) {
+        var h = $("<div/>").append(html);
+        var style_tags = h.find("style");
+        if (!style_tags.length) {
+            // no style tags to sanitize
+            return html;
+        }
+        style_tags.each(function(i, style) {
+            style.innerHTML = sanitize_css(style.innerHTML, tagPolicy);
+        });
+        return h.html();
+    };
+    
+    var sanitize = function (html, allow_css) {
         // sanitize HTML
+        // if allow_css is true (default), CSS is sanitized as well.
+        // otherwise, CSS elements and attributes are simply removed.
         // returns a struct of
         // {
         //   src: original_html,
@@ -69,6 +99,20 @@ IPython.security = (function (IPython) {
         //                        This is an incomplete indication,
         //                        only used to indicate whether further verification is necessary.
         // }
+        var html4 = caja.html4;
+
+        if (allow_css === undefined) allow_css = true;
+        if (allow_css) {
+            // allow sanitization of style tags,
+            // not just scrubbing
+            html4.ELEMENTS.style &= ~html4.eflags.UNSAFE;
+            html4.ATTRIBS.style = html4.atype.STYLE;
+        } else {
+            // scrub all CSS
+            html4.ELEMENTS.style |= html4.eflags.UNSAFE;
+            html4.ATTRIBS.style = html4.atype.SCRIPT;
+        }
+        
         var result = {
             src : html,
             _maybe_safe : true
@@ -78,7 +122,6 @@ IPython.security = (function (IPython) {
             result._maybe_safe = false;
         };
         
-        var html4 = caja.html4;
         var policy = function (tagName, attribs) {
             if (!(html4.ELEMENTS[tagName] & html4.eflags.UNSAFE)) {
                 return {
@@ -92,8 +135,14 @@ IPython.security = (function (IPython) {
                 });
             }
         };
-
+        
         result.sanitized = caja.sanitizeWithPolicy(html, policy);
+        
+        if (allow_css) {
+            // sanitize style tags as stylesheets
+            result.sanitized = sanitize_stylesheets(result.sanitized, policy);
+        }
+        
         return result;
     };
     
@@ -104,6 +153,7 @@ IPython.security = (function (IPython) {
     
     var is_safe = function (html) {
         // just return bool for whether an HTML string is safe
+        // this is not currently used for anything other than tests.
         var result = sanitize(html);
         
         // caja can strip whole elements without logging,
@@ -117,6 +167,7 @@ IPython.security = (function (IPython) {
     };
     
     return {
+        caja: caja,
         is_safe: is_safe,
         sanitize: sanitize,
         sanitize_html: sanitize_html
diff --git a/IPython/html/templates/notebook.html b/IPython/html/templates/notebook.html
index 42f5424..2252496 100644
--- a/IPython/html/templates/notebook.html
+++ b/IPython/html/templates/notebook.html
@@ -291,7 +291,7 @@ class="notebook_app"
 
 {{super()}}
 
-<script src="{{ static_url("components/google-caja/google-caja/html-sanitizer-minified.js") }}" charset="utf-8"></script>
+<script src="{{ static_url("components/google-caja/html-css-sanitizer-minified.js") }}" charset="utf-8"></script>
 <script src="{{ static_url("components/codemirror/lib/codemirror.js") }}" charset="utf-8"></script>
 <script type="text/javascript">
     CodeMirror.modeURL = "{{ static_url("components/codemirror/mode/%N/%N.js", include_version=False) }}";
diff --git a/IPython/html/tests/base/security.js b/IPython/html/tests/base/security.js
index 35edb42..04d69a2 100644
--- a/IPython/html/tests/base/security.js
+++ b/IPython/html/tests/base/security.js
@@ -3,6 +3,7 @@ safe_tests = [
     '<h1 class="foo">Hi There!</h1>',
     '<a data-cite="foo">citation</a>',
     '<div><span>Hi There</span></div>',
+    '<style>div.foo { background: #ffff; }</style>',
 ];
 
 unsafe_tests = [
@@ -17,7 +18,7 @@ unsafe_tests = [
     '<META HTTP-EQUIV="refresh" CONTENT="0; URL=http://;URL=javascript:alert(999);">',
     '<IFRAME SRC="javascript:alert(999);"></IFRAME>',
     '<IFRAME SRC=# onmouseover="alert(document.cookie)"></IFRAME>',
-    '<style type="text/css">div.foo { background: #ffff; }</style>',
+    '<style src="http://untrusted/style.css"></style>',
     '<EMBED SRC="data:image/svg+xml;base64,PHN2ZyB4bWxuczpzdmc9Imh0dH A6Ly93d3cudzMub3JnLzIwMDAvc3ZnIiB4bWxucz0iaHR0cDovL3d3dy53My5vcmcv MjAwMC9zdmciIHhtbG5zOnhsaW5rPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5L3hs aW5rIiB2ZXJzaW9uPSIxLjAiIHg9IjAiIHk9IjAiIHdpZHRoPSIxOTQiIGhlaWdodD0iMjAw IiBpZD0ieHNzIj48c2NyaXB0IHR5cGU9InRleHQvZWNtYXNjcmlwdCI+YWxlcnQoIlh TUyIpOzwvc2NyaXB0Pjwvc3ZnPg==" type="image/svg+xml" AllowScriptAccess="always"></EMBED>',
 ];
 
@@ -26,6 +27,10 @@ casper.notebook_test(function () {
         var is_safe = self.evaluate(function (item) {
             return IPython.security.is_safe(item);
         }, item);
+        var sanitized = self.evaluate(function (item) {
+            return IPython.security.sanitize_html(item);
+        }, item);
+        
         this.test.assert(is_safe, "Safe: " + item);
     });
     this.each(unsafe_tests, function (self, item) {