locked
App Crash

    Question

  • Hi,

    My Metro application crash, and don't show up the typical dialog to retry with the debugger, i just see this error report in event viewer.

    Fault bucket -1252793025, type 5
    Event Name: MoAppCrash
    Response: Not available
    Cab Id: 0
    
    Problem signature:
    P1: 88ef9e9d-456e-4e3c-8e01-148590965a8b_1.0.0.0_x86__zyjr9fbq46ng4
    P2: praid:App
    P3: 0.0.0.0
    P4: 4fd8eb16
    P5: ice34d.dll
    P6: 0.0.0.0
    P7: 4fd8e9fc
    P8: c0000005
    P9: 0003e28b
    P10:
    
    Attached files:
    C:\Users\pepe\AppData\Local\Temp\WER6335.tmp.WERInternalMetadata.xml
    
    These files may be available here:
    C:\Users\pepe\AppData\Local\Microsoft\Windows\WER\ReportArchive\AppCrash_88ef9e9d-456e-4e_72ba7b9c73e60852aead9d519c0dd73983fd16a_08ba691e
    
    Analysis symbol:
    Rechecking for solution: 0
    Report Id: fc330090-b590-11e1-a5df-000c29b8af73
    Report Status: 0
    Hashed bucket: cbf4c4de9767f0ccad94f100099bf4b0
    any ideas how can i track this down, i can reproduce the problem in my test suite, but is part of a large project and seems to me that will be difficult to isolate the issue in a small test case.
    Wednesday, June 13, 2012 8:39 PM

All replies

  • Hi,

    The crash signature information above is unique to your application, and without a stack trace, it isn't enough to find the root cause of the crash.  So, could you please tell me a little more about what you do to reproduce the problem and how you have been debugging the app?

    • Does the repro happen consistently--like every time you try to repro, you can, and it repros exactly the same way?
    • When you repro the problem, are you running the app under the debugger?
    • What operation does the app do when the app crashes?
    • Does your code handle exceptions thrown by the WinRT objects that it uses?
    • Was a dump file for the app created?
    • Anything else that's related to the repro?

    I have seen times that when a Win32 app silently disappears without it's exception handlers getting called and the post-mortem debugger doesn't run, the problem was because the thread's stack was badly messed up.  As you know, exception handling is stack-based so overwriting portions of the stack or overflowing the stack would be things to look out for.

    Sincerely,

    Dan Ruder [MSFT] 

    Thursday, June 14, 2012 6:26 PM
    Moderator
  • > Does the repro happen consistently--like every time you try to repro, you can, and it repros exactly the same way?

    In my test suite with one of the test cases this happens almost all times, but some times the test just run fine. This test is quite large and create/destroy lot of network sockets to test Bind operations.

    > When you repro the problem, are you running the app under the debugger?

    No is not and i cannot run my app in the Debugger because when i run the app inside the Debugger it hangs in _NLG_Notify, see this other issue http://social.msdn.microsoft.com/Forums/en-US/winappswithnativecode/thread/4dc5146e-a98f-4198-8646-bd5450cb9e7c

    > What operation does the app do when the app crashes?

    As i cannot run the app in the debugger is difficult to know, this test case is quite large.

    > I have seen times that when a Win32 app silently disappears without it's exception handlers getting called and the post-mortem debugger doesn't > run, the problem was because the thread's stack was badly messed up. As you know, exception handling is stack-based so overwriting portions of > the stack or overflowing the stack would be things to look out for.

    I use a custom class for start my threads, and in calls std::terminate in case of unhalted exceptions, this result in open the abort/retry dialog in Windows.

    Thursday, June 14, 2012 9:30 PM