Debugger completely broken in 4.0

I have three coworkers who all fell in love with the 3.0.x debugger but find it completely unusable in 4.0.x. The problem is with socket debugging with Weblogic - haven't tried Weblogic integration yet. But anyway, what happens is that when the execution reaches a breakpoint and the button is hit to skip to the next statement, the debugger hangs indefinitely. In other words, once a breakpoint is hit that debugging session is OVER.

Does anybody else have problems with this or is there anything we can try to see if it fixes the problem?

Thanks,

Justin

8 comments
Comment actions Permalink

Justin Ashworth wrote:

I have three coworkers who all fell in love with the 3.0.x debugger but find it completely unusable in 4.0.x. The problem is with socket debugging with Weblogic - haven't tried Weblogic integration yet. But anyway, what happens is that when the execution reaches a breakpoint and the button is hit to skip to the next statement, the debugger hangs indefinitely. In other words, once a breakpoint is hit that debugging session is OVER.

Does anybody else have problems with this or is there anything we can try to see if it fixes the problem?


I've seen the problem on occasion but haven't isolated it.

Jon

0
Comment actions Permalink

I have seen where the debugger takes a while to get it's bearing when it
hits a breakpoint. However, it always comes back and continues to work.

W2KPro + Tomcat 4.1.18 (don't blame me - IT won't let me upgrade).

Norris Shelton
Sun Certified Java Programmer




Justin Ashworth wrote:

>I have three coworkers who all fell in love with the 3.0.x debugger but find it completely unusable in 4.0.x. The problem is with socket debugging with Weblogic - haven't tried Weblogic integration yet. But anyway, what happens is that when the execution reaches a breakpoint and the button is hit to skip to the next statement, the debugger hangs indefinitely. In other words, once a breakpoint is hit that debugging session is OVER.
>
>Does anybody else have problems with this or is there anything we can try to see if it fixes the problem?
>
>Thanks,
>
>Justin

>

0
Comment actions Permalink

I've seen this too. It does come back to life....eventually.

W2KPro + JDK1.4.2_04 + WebLogic 7.0.4.0 (ditto on the upgrade comment).

Tim

"Norris Shelton" <i.hate.spam@yahoo.com> wrote in message
news:40AE158E.7060600@yahoo.com...

I have seen where the debugger takes a while to get it's bearing when it
hits a breakpoint. However, it always comes back and continues to work.

>

W2KPro + Tomcat 4.1.18 (don't blame me - IT won't let me upgrade).

>

Norris Shelton
Sun Certified Java Programmer

>
>
>
>

Justin Ashworth wrote:

>

>I have three coworkers who all fell in love with the 3.0.x debugger but

find it completely unusable in 4.0.x. The problem is with socket debugging
with Weblogic - haven't tried Weblogic integration yet. But anyway, what
happens is that when the execution reaches a breakpoint and the button is
hit to skip to the next statement, the debugger hangs indefinitely. In
other words, once a breakpoint is hit that debugging session is OVER.

>
>Does anybody else have problems with this or is there anything we can try

to see if it fixes the problem?

>
>Thanks,
>
>Justin
>
>



0
Comment actions Permalink

Yes, the 4.0 debugger is ridiculously bad. We had a serious problem getting Rational Robot to load test our app. Being that Robot costs 6 figures, our development team was strongly encouraged to get it to work. We couldn't figure out what was going wrong so we initiated a remote debugging session on the host server and had the Robot tool proceed to where the problem was. This was done with about half a dozen of us from different departments conferencing in -- thus if we have embarassing problems debugging the app, it wastes the time of a lot of people.

Well, debugging this scenario is painfully slow, but I hardly fault Idea for that. It was going over the wire, afterall. If it were simply slow, that would be acceptable. Unfortunately...

1) When stopped at a breakpoint, I was inspecting some variables using Alt-F8. Occasionally for reasons unknown to me, Idea will take an eternity evaluating something. When this happens, the breakpoint is dead and we must restart the test from the beginning. Failure to resume the debugger and repeat the test results in a seemingly infinite wait trying to evaluate anything else.

2) Since this is a webserver we had a few glitches with other connections setting off my breakpoints and ruining the iteration and forcing a repeat of the test. This is not so much a bug as just an inherrent limitation of the current Idea debugger. The button "Suspend all threads when stepping" does nothing for me here and even if it did, it seems completely broken when I am trying to step.

3) This is the kicker. Sometimes Idea will just completely freeze when remote debugging. I mean, it doesn't even repaint the screen. I cannot fathom what aspecet of debugging would work on the UI thread and I gave it a LONG time to recover and it didn't. So, we killed Idea and tried to reconnect. Unfortunately when Idea crashed it also hosed our server and we had to phone in to some sys admins to get it restarted which took another half hour while we all waited. This is consistent: if you are remote debugging and kill idea you also kill whatever process you were debugging. I presume this is an issue with the JDPA but in any event, Idea should not get into the state that requires me to halt its task.

0
Comment actions Permalink

Jon Steelman wrote:

I've seen the problem on occasion but haven't isolated it.


Debugger never detatches from JPDA/JBoss for me, and I'm even started to
get NullPointerExceptions inside the debugger occasionally now ( 2026 )

--
Discouragement is a dissatisfaction with the past, a distaste for the
present, and a distrust of the future - Maree De Jong, CLCA.

Mark Derricutt --- mark@ talios.com --- http://www.talios.com

0
Comment actions Permalink

Could you please send the stacktrace when IDEA frozes?

0
Comment actions Permalink

I meant thread dump. (Ctrl+Break in console)

0
Comment actions Permalink

I'll see what I can do to reproduce it.

0

Please sign in to leave a comment.