I have created a test case that reproduces this problem in Webkit nightly builds (r24285) and will attach the source for a java applet and an HTML page to this report. LiveConnectTestApplet spawns a thread which invokes JSObject.getWindow(this).call(...) once every second. The invoked javascript function counts the number of times it has been invoked and displays this information. The applet meanwhile counts and displays the number of times it attempts to invoke the javascript function and the number of failures it detects (via the win.call() return code). On Firefox and Safari2, this runs seemingly forever without failure. On Webkit nightlys and Safari 3 beta, this works for about the first 15 seconds without failure and then starts to fail consistently for as long as I leave the applet run. I have tested this on both a PowerBook G4 and on an iMac (Intel Duo), both running Mac OS X 10.4.10.
Created attachment 15550 [details] Test case applet source code You need to have plugin.jar (for JSObject) in your class path to compile this applet.
Created attachment 15551 [details] Test case HTML with embedded javascript. Invoke this to run testcase once LiveConnectTestApplet.java has been compiled into a jar and is in the same folder as this HTML.
Created attachment 15552 [details] Ant build file for test case applet
(In reply to comment #0) My java runtime is 1.5: Java Plug-in 1.5.0 Using JRE version 1.5.0_07 Java HotSpot(TM) Client VM
Quick steps to build: 1. Download all three attachments into one directory. 2. Open Terminal and 'cd' to that directory. 3. ln -s /System/Library/Frameworks/JavaVM.framework/Versions/1.5.0/Home/lib/plugin.jar . 4. ant build
Confirmed with a local debug build of WebKit r24439 with Safari 3.0 (522.12) on Mac OS X 10.4.10 (8R218) with Java 1.5.0_07 on a PowerBook G4. $ java -version java version "1.5.0_07" Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_07-164) Java HotSpot(TM) Client VM (build 1.5.0_07-87, mixed mode, sharing)
<rdar://problem/5346330>
Per the description, this is a regression; marking accordingly.
This seems to have been fixed at some point between r24513 and r24594. <http://trac.webkit.org/projects/webkit/changeset/24559> is the most likely candidate.
(In reply to comment #9) > This seems to have been fixed at some point between r24513 and r24594. > <http://trac.webkit.org/projects/webkit/changeset/24559> is the most likely > candidate. > I just retested on both r25168 and r25328. Test works properly on both. Thanks, Glenn