PerformanceTests/Parser/html-parser is only testing parsing of the head element

https://bugs.webkit.org/show_bug.cgi?id=69283

Reviewed by Adam Barth.

While investigating https://bugs.webkit.org/show_bug.cgi?id=68944
I found that the html-parser benchmark was only parsing up to the
first script tag per loop!  We've fixed this by adding the sandbox
tag which will deny all external loads and allow the parser to
synchronously continue parsing the entire document to completion
(as we had expected it was doing).

This changes the profile somewhat.  Line number counting is much
hotter, since we're actually accounting for the parse of the entire
document in our sample.

Total sample time only about doubles, from 1800ms to 3800ms on my machine
which is less than I would have expected.

* Parser/html-parser.html:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@96531 268f45cc-cd09-0410-ab3c-d52691b4dbfc

diff --git a/PerformanceTests/ChangeLog b/PerformanceTests/ChangeLog
index 89c6302..3eec52e 100644 (file)
--- a/PerformanceTests/ChangeLog
+++ b/PerformanceTests/ChangeLog
@@ -1,3 +1,26 @@
+2011-10-03  Eric Seidel  <eric@webkit.org>
+
+        PerformanceTests/Parser/html-parser is only testing parsing of the head element
+        https://bugs.webkit.org/show_bug.cgi?id=69283
+
+        Reviewed by Adam Barth.
+
+        While investigating https://bugs.webkit.org/show_bug.cgi?id=68944
+        I found that the html-parser benchmark was only parsing up to the
+        first script tag per loop!  We've fixed this by adding the sandbox
+        tag which will deny all external loads and allow the parser to
+        synchronously continue parsing the entire document to completion
+        (as we had expected it was doing).
+
+        This changes the profile somewhat.  Line number counting is much
+        hotter, since we're actually accounting for the parse of the entire
+        document in our sample.
+
+        Total sample time only about doubles, from 1800ms to 3800ms on my machine
+        which is less than I would have expected.
+
+        * Parser/html-parser.html:
+
 2011-08-09  Alexandru Chiculita  <achicu@adobe.com>
 
         Fixing 65868 REGRESSION(r92610) caused by 65668 - Optimize floating elements lookup

diff --git a/PerformanceTests/Parser/html-parser.html b/PerformanceTests/Parser/html-parser.html
index bb2c73c..bf71302 100644 (file)
--- a/PerformanceTests/Parser/html-parser.html
+++ b/PerformanceTests/Parser/html-parser.html
@@ -8,7 +8,8 @@ var spec = loadFile("resources/html5.html");
 start(20, function() {
     var iframe = document.createElement("iframe");
     iframe.src = "about:blank";
-    iframe.style.display = "none";
+    iframe.style.display = "none";  // Prevent creation of the rendering tree, so we only test HTML parsing.
+    iframe.sandbox = '';  // Prevents loading of external scripts which would otherwise pause the parser.
     document.body.appendChild(iframe);
     iframe.contentDocument.open();
     iframe.contentDocument.write(spec);