-
Bug
-
Resolution: Fixed
-
P4
-
1.3.1
-
hopper
-
x86
-
windows_2000
Test case is in /net/cores.east/cores/62653346
I stripped it down to the smallest demonstration and have attached the
files. Launch "frame.htm" and you should hit the Dr. Watson. Now, if you
edit "navbar.htm" to use an <applet> tag, you will not get the Dr. Watson.
Here's the flow of what is happening. frame.htm loads a frameset with the
applet in the top frame and a blank page (blank.htm) in the lower. The
applet then loads "splash.htm" into the lower frame. When "splash.htm"
loads, it calls the applet's playPage() method, and it is in attempting to
make this call that the crash occurs. If you edit "splash.htm" to comment
out the call and uncomment the call with the 5 second delay, you won't get
the crash.
Note that the "splash.htm" page is being loaded in the applet's init()
method. If you change init() to start() and recompile the applet, no crash
occurs. This works fine with the Microsoft JVM (<applet> tag). Apparently
the plugin can't handle calls into an applet during it init() processing.
However, the javadoc for Applet doesn't say that init() is anything other
than a normal method, so this would appear to be a plugin bug.
I stripped it down to the smallest demonstration and have attached the
files. Launch "frame.htm" and you should hit the Dr. Watson. Now, if you
edit "navbar.htm" to use an <applet> tag, you will not get the Dr. Watson.
Here's the flow of what is happening. frame.htm loads a frameset with the
applet in the top frame and a blank page (blank.htm) in the lower. The
applet then loads "splash.htm" into the lower frame. When "splash.htm"
loads, it calls the applet's playPage() method, and it is in attempting to
make this call that the crash occurs. If you edit "splash.htm" to comment
out the call and uncomment the call with the 5 second delay, you won't get
the crash.
Note that the "splash.htm" page is being loaded in the applet's init()
method. If you change init() to start() and recompile the applet, no crash
occurs. This works fine with the Microsoft JVM (<applet> tag). Apparently
the plugin can't handle calls into an applet during it init() processing.
However, the javadoc for Applet doesn't say that init() is anything other
than a normal method, so this would appear to be a plugin bug.