Re: debug problems on a dual G5 2GHZ

From: David Teran (davidtera..ac.com)
Date: Mon Feb 16 2004 - 08:04:55 EST

  • Next message: Harald Niesche: "Re: Classpath problems"

    Hi Christian,

    thanks for the hint. OK, it seems a 'fast computer machine problem'
    because Frank working on a dual G5 867 could not reproduce the problem
    at all. Therefore maybe a dual CPU and fast machine related (java)
    bug... not the first one we found.

    regards David

    On 16.02.2004, at 13:57, Christian Mittendorf wrote:

    > Hi,
    >
    > I've seen the described problem as well, but on a dual G4 1.25 GHZ. It
    > does occur from time to time - though not all the time. It's therefore
    > difficult to reproduce. However, I will try to use your "turn off a
    > processor" tip and see if that does solve the problem.
    >
    > cu
    > cm
    >
    >
    > On 16.02.2004, at 13:40, David Teran wrote:
    >
    >> Hi,
    >>
    >> i would like to ask everyone who uses a dual G5 with 2 GHZ (maybe 1.8
    >> is also OK) to check if debugging works fine or not with the 3.x
    >> version of eclipse. I am currently using 3M7 but it does also occur
    >> with older integration builds and maybe also with 2.1.x.
    >>
    >> Problem description:
    >> The debugger stopps the application at a specified breakpoint but the
    >> java code window does not reflect this. I cannot step until with the
    >> debugger because all threads are running, at least that whats the gui
    >> tells me. But the debugger stopped at the specified breakpoint the
    >> thread. When i 'pause' the current worker thread i can see the
    >> stacktrace to point to the code line where i defined the breakpoint
    >> but stepping is not possible.
    >>
    >> Workaround (no kidding!):
    >> Simply install CHUD dev tool (common hardware understanding ?) from
    >> xcode distribution and turn off one CPU with the system prefence pane
    >> named CPU. Then debugging works as expected and very stable.
    >>
    >> I would like to know if anyone else can verify this problem because
    >> then apple can reproduce it, too and then i will file a bug.
    >>
    >> regards David
    >
    >





    This archive was generated by hypermail 2.0.0 : Mon Feb 16 2004 - 08:05:02 EST