Title: [106600] trunk
- Revision
- 106600
- Author
- [email protected]
- Date
- 2012-02-02 15:42:26 -0800 (Thu, 02 Feb 2012)
Log Message
Running a Web Worker on about:blank crashes the interpreter
https://bugs.webkit.org/show_bug.cgi?id=77593
Patch by Benjamin Poulain <[email protected]> on 2012-02-02
Reviewed by Michael Saboff.
Source/_javascript_Core:
The method Interpreter::execute() was crashing on empty programs because
the assumption is made the source is not null.
This patch shortcut the execution when the String is null to avoid invalid
memory access.
* interpreter/Interpreter.cpp:
(JSC::Interpreter::execute):
LayoutTests:
The Worker thread should not crash if running on an empty page.
The timer is necessary to give a chance for the thread to start.
* fast/workers/empty-worker-nocrash-expected.txt: Added.
* fast/workers/empty-worker-nocrash.html: Added.
Modified Paths
Added Paths
Diff
Modified: trunk/LayoutTests/ChangeLog (106599 => 106600)
--- trunk/LayoutTests/ChangeLog 2012-02-02 23:41:11 UTC (rev 106599)
+++ trunk/LayoutTests/ChangeLog 2012-02-02 23:42:26 UTC (rev 106600)
@@ -1,3 +1,16 @@
+2012-02-02 Benjamin Poulain <[email protected]>
+
+ Running a Web Worker on about:blank crashes the interpreter
+ https://bugs.webkit.org/show_bug.cgi?id=77593
+
+ Reviewed by Michael Saboff.
+
+ The Worker thread should not crash if running on an empty page.
+ The timer is necessary to give a chance for the thread to start.
+
+ * fast/workers/empty-worker-nocrash-expected.txt: Added.
+ * fast/workers/empty-worker-nocrash.html: Added.
+
2012-02-02 Filip Pizlo <[email protected]>
retrieveCallerFromVMCode should call trueCallerFrame
Added: trunk/LayoutTests/fast/workers/empty-worker-nocrash-expected.txt (0 => 106600)
--- trunk/LayoutTests/fast/workers/empty-worker-nocrash-expected.txt (rev 0)
+++ trunk/LayoutTests/fast/workers/empty-worker-nocrash-expected.txt 2012-02-02 23:42:26 UTC (rev 106600)
@@ -0,0 +1 @@
+PASS
Added: trunk/LayoutTests/fast/workers/empty-worker-nocrash.html (0 => 106600)
--- trunk/LayoutTests/fast/workers/empty-worker-nocrash.html (rev 0)
+++ trunk/LayoutTests/fast/workers/empty-worker-nocrash.html 2012-02-02 23:42:26 UTC (rev 106600)
@@ -0,0 +1,11 @@
+<html>
+<script>
+var worker = new Worker('about:blank');
+if (window.layoutTestController) {
+ layoutTestController.dumpAsText();
+ layoutTestController.waitUntilDone();
+}
+setTimeout('window.layoutTestController.notifyDone()', 20);
+</script>
+<p>PASS</p>
+</html>
Modified: trunk/Source/_javascript_Core/ChangeLog (106599 => 106600)
--- trunk/Source/_javascript_Core/ChangeLog 2012-02-02 23:41:11 UTC (rev 106599)
+++ trunk/Source/_javascript_Core/ChangeLog 2012-02-02 23:42:26 UTC (rev 106600)
@@ -1,3 +1,19 @@
+2012-02-02 Benjamin Poulain <[email protected]>
+
+ Running a Web Worker on about:blank crashes the interpreter
+ https://bugs.webkit.org/show_bug.cgi?id=77593
+
+ Reviewed by Michael Saboff.
+
+ The method Interpreter::execute() was crashing on empty programs because
+ the assumption is made the source is not null.
+
+ This patch shortcut the execution when the String is null to avoid invalid
+ memory access.
+
+ * interpreter/Interpreter.cpp:
+ (JSC::Interpreter::execute):
+
2012-02-02 Kalev Lember <[email protected]>
[GTK] Use win32 native threading
Modified: trunk/Source/_javascript_Core/interpreter/Interpreter.cpp (106599 => 106600)
--- trunk/Source/_javascript_Core/interpreter/Interpreter.cpp 2012-02-02 23:41:11 UTC (rev 106599)
+++ trunk/Source/_javascript_Core/interpreter/Interpreter.cpp 2012-02-02 23:42:26 UTC (rev 106600)
@@ -885,6 +885,8 @@
Vector<JSONPData> JSONPData;
bool parseResult;
const UString programSource = program->source().toString();
+ if (programSource.isNull())
+ return jsUndefined();
if (programSource.is8Bit()) {
LiteralParser<LChar> literalParser(callFrame, programSource.characters8(), programSource.length(), JSONP);
parseResult = literalParser.tryJSONPParse(JSONPData, scopeChain->globalObject->globalObjectMethodTable()->supportsRichSourceInfo(scopeChain->globalObject.get()));
_______________________________________________
webkit-changes mailing list
[email protected]
http://lists.webkit.org/mailman/listinfo.cgi/webkit-changes
