Phantom Breakpoints [message #238110] |
Wed, 17 October 2007 22:20 |
Eclipse User |
|
|
|
Originally posted by: l2j1983.gmail.com
So I am working on an overhaul of an old Java project for work, I have
some multi-threaded code that I need to see what is going on and when I go
to debug, there is a single breakpoint that is hit that doesn't exist,
which really slows down my development process.
I have tried doing
Run->Remove all Breakpoints
Launching eclipse with the -clean argument
switching to the newest Eclipse IDE (europa)
still no luck, the breakpoint still shows up and sometimes causes the JVM
to crash.
The fact that it traveled from one version of the IDE to another makes me
think this breakpoint is somehow a remnant stored in a file somewhere.
What can I do to get rid of this for good.
Thanks in advance.
|
|
|
|
Re: Phantom Breakpoints [message #238114 is a reply to message #238112] |
Thu, 18 October 2007 02:13 |
Eclipse User |
|
|
|
Originally posted by: L2J1983.gmail.com
Nope, sure doesn't. Which really confuses me.
All of my fellow programmers/co-workers have no idea what is going on.
The breakpoint does not show in the Breakpoints View nor the Sidebar(where
the line numbers are). Remove all breakpoints does nothing, checking out
new files from our subversion repository/s does not make it go away. When
the breakpoint is hit, it doesn't show up even temporarily in the
breakpoints view.
|
|
|
Re: Phantom Breakpoints [message #238116 is a reply to message #238114] |
Thu, 18 October 2007 02:53 |
Eclipse User |
|
|
|
Originally posted by: wegener.cboenospam.com
L wrote:
> Nope, sure doesn't. Which really confuses me.
> All of my fellow programmers/co-workers have no idea what is going on.
>
> The breakpoint does not show in the Breakpoints View nor the
> Sidebar(where the line numbers are). Remove all breakpoints does
> nothing, checking out new files from our subversion repository/s does
> not make it go away. When the breakpoint is hit, it doesn't show up
> even temporarily in the breakpoints view.
>
>
Are you sure you aren't hitting an exception breakpoint?
|
|
|
Re: Phantom Breakpoints [message #238117 is a reply to message #238116] |
Thu, 18 October 2007 05:27 |
Eclipse User |
|
|
|
Originally posted by: sreed.spamcop.net
David Wegener's message received 10/17/2007 10:53 PM:
> L wrote:
>> Nope, sure doesn't. Which really confuses me.
>> All of my fellow programmers/co-workers have no idea what is going on.
>>
>> The breakpoint does not show in the Breakpoints View nor the
>> Sidebar(where the line numbers are). Remove all breakpoints does
>> nothing, checking out new files from our subversion repository/s does
>> not make it go away. When the breakpoint is hit, it doesn't show up
>> even temporarily in the breakpoints view.
>>
>>
> Are you sure you aren't hitting an exception breakpoint?
I have seen something like this when there is an exception breakpoint in
compiled library code. You will see the execution bar in the most
recently called method with available source code. Perhaps you can see
one or more threads in the call stack view that are suspended and if you
expand them you can see where/what the problem is?
|
|
|
Re: Phantom Breakpoints [message #238119 is a reply to message #238117] |
Thu, 18 October 2007 05:33 |
Eclipse User |
|
|
|
Originally posted by: sreed.spamcop.net
S Reed's message received 10/18/2007 1:27 AM:
> David Wegener's message received 10/17/2007 10:53 PM:
>> L wrote:
>>> Nope, sure doesn't. Which really confuses me.
>>> All of my fellow programmers/co-workers have no idea what is going on.
>>>
>>> The breakpoint does not show in the Breakpoints View nor the
>>> Sidebar(where the line numbers are). Remove all breakpoints does
>>> nothing, checking out new files from our subversion repository/s
>>> does not make it go away. When the breakpoint is hit, it doesn't
>>> show up even temporarily in the breakpoints view.
>>>
>>>
>> Are you sure you aren't hitting an exception breakpoint?
> I have seen something like this when there is an exception breakpoint
> in compiled library code. You will see the execution bar in the most
> recently called method with available source code. Perhaps you can see
> one or more threads in the call stack view that are suspended and if
> you expand them you can see where/what the problem is?
But then, Remove All Breakpoints removes the exception breakpoints, so
that's probably not the explanation. I hope the call stack reveals
something.
|
|
|
Re: Phantom Breakpoints [message #238206 is a reply to message #238119] |
Thu, 18 October 2007 18:17 |
Eclipse User |
|
|
|
Originally posted by: l2j1983.gmail.com
I was able to locate the issue today, see the program has 13 threads that
pop an object off the queue and do some tests on remote machines reporting
status to a jsp page, once all threads see an empty queue, the threads are
disposed of. The app then waits for two minutes and does it all again...
inside one of these threads, there was a null pointer being thrown which
caused the thread to enter a suspended state and threw an exception.
However the strange part is that the single line that was throwing the
exception was surrounded in a try catch (catching the generic Exception
e), and when the apparently unhanded exception was seen nothing was output
to the console. When I switched this to catch the null pointer itself, it
all works and the JVM doesn't crash, the thread doesn't get suspended, and
the threads can actually all return the empty queue status to the thread
controller.
Thanks for pointing me in the correct direction, I was thinking too
complexly but even still the problem ended up being strange.
|
|
|
Powered by
FUDForum. Page generated in 0.08465 seconds