Community
Participate
Working Groups
build 20020430 My log file has a lot of InvalidStackFrameExceptions. A while ago I entered a different PR with a log file that contained InvalidStackFrameExceptions. You asked if I had seen this again. I have now. I also got the strange "break in loadLibrary" behavior a couple of times today. The first time a breakpoint is hit in my target Eclipse the debugger stops at some SWT loadLibrary call. Hitting resume then stops at the real breakpoint. The InvalidStackFrameException is definitely related to the weird breakpoint behavior. I just reproduced the breakpoint behavior by restarting my debug session and the log contained a fresh InvalidStackFrameException.
Created attachment 740 [details] log file
My target Eclipse is running on J9 1.5 20020206, the runtime Eclipse is running on J9 2.0 20020426. The bug I mention above is bug 13948.
Where exactly are you putting your real breakpoint? Do you have any other breakpoints set? What about options? Do you have the option to "Suspend on uncaught exceptions" or the option to "Suspend on compilation errors" turned on?
The breakpoint is in org.eclipse.swt.custom.StyledText.handleTextChanging on the line if (event.replaceCharCount < 0) { This is the only breakpoint. "Suspend on uncaught exceptions" is off "Suspend on compilation errors" is on The Java/Debug preference page is set to the defaults. This problem is similar to the one described in bug 12740 except that I only get it when debugging Eclipse and the bogus stack trace is slightly different. I'm going to attach the complete stack in the debugger.
Created attachment 744 [details] debugger stack for bogus breakpoint
The debugger actually highlights the loadLibraryWithPath line in java.lang.Runtime.loadLibraryWithClassLoader when it stops.
*** This bug has been marked as a duplicate of 12740 ***
