TEST_BUG: test scaffold error output detection is flakey

The test scaffold has some logic to detect various error output
strings and fails the test if one of the magic strings appears
in the output. It appears that logic is somehow issue false
positives.

Since these are essentially 'drive by shootings', I'm prefixing
the test names with 'XXX/' so they don't show up on the fail_list.

So far these failures have only been seen on Win* machines. In
fact, just on machine vm-x86-3 and machine vmnightly2. Is there
something special about these two machines? Do they have the same
MKS version as other machines?

Here are the entries from my nightly analysis report:

New JDI_REGRESSION failures (from 2007.12.12)
    XXX/com/sun/jdi/RedefineFinal.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vm-x86-3). However, the dump of the file doesn't show the magic
        string.

New JDI_REGRESSION failures (from 2007.12.06)
    XXX/com/sun/jdi/JdbLockTest.sh
        This test failed due to "Fail: main: Error output found: "An
        unexpected error" in ..." on Win32 Client VM (machine
        vm-x86-3). However, the dump of the file doesn't show the magic
        string.

New JDI_REGRESSION failures (from 2007.12.03)
    XXX/com/sun/jdi/GetLocalVariables3Test.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Client VM (machine
        vmnightly2). However, the dump of the file doesn't show the
        magic string.

New JDI_REGRESSION failures (from 2007.11.07)
    XXX/com/sun/jdi/Redefine-g.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vm-x86-3). However, the dump of the file doesn't show the
        magic string.

New nsk.jdb failures (from 2007.10.30)
    XXX/com/sun/jdi/JdbMissStep.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vmnightly2). However, the dump of the file doesn't show the
        magic string.

New JDI_REGRESSION failures (from 2007.10.29)
    XXX/com/sun/jdi/MixedSuspendTest.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vmnightly2). However, the dump of the file doesn't show the
        magic string.

New JDI_REGRESSION failures (from 2007.09.29)
    XXX/com/sun/jdi/RedefineClearBreakpoint.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Client VM (machine
        vmnightly2). However, the dump of the file doesn't show the
        magic string.

New JDI_REGRESSION failures (from 2007.09.26)
    XXX/com/sun/jdi/JdbLockTest.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vmnightly2). However, the dump of the file doesn't show the
        magic string.

New JDI_REGRESSION failures (from 2007.08.30)
    XXX/com/sun/jdi/WatchFramePop.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Client VM (machine
        vmnightly2). However, the dump of the file doesn't show the
        magic string.

New JDI_REGRESSION failures (from 2007.08.30)
    XXX/com/sun/jdi/CommandCommentDelimiter.sh
    XXX/com/sun/jdi/JdbMissStep.sh
        These tests failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vm-x86-3). However, the dump of the file doesn't show the
        magic string.

New JDI_REGRESSION failures (from 2007.06.30)
    XXX/com/sun/jdi/RedefineStep.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Client VM (machine
        vmnightly2). However, the dump of the file doesn't show the
        magic string.

        Update: In the 2007.12.15 nightly, the failure was due to
        "Fail: main: Error output found: "An unexpected error" in ..."
        on Win32 Client VM (machine vm-x86-3).

New JDI_REGRESSION failures (from 2007.06.27)
    XXX/com/sun/jdi/ArrayLengthDumpTest.sh
    XXX/com/sun/jdi/RedefineAnnotation.sh
        These two tests failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Client VM (machine
        vm-x86-3). However, the dump of the file doesn't show the magic
        string.
JdbMissStep.sh again...

New JDI_REGRESSION failures (from 2007.12.22)
    XXX/com/sun/jdi/JdbMissStep.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Client VM (machine
        vm-x86-3). However, the dump of the file doesn't show the magic
        string. This is another occurrence of the following bug:

            6644919 4/4 test scaffold error output detection is flakey

        I will add this entry to 6644919.
RedefineStep.sh again...

New JDI_REGRESSION failures (from 2008.01.11)
    XXX/com/sun/jdi/RedefineStep.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vm-x86-3). However, the dump of the file doesn't show the magic
        string. This is another occurrence of the following bug:

            6644919 4/4 test scaffold error output detection is flakey

        I will add this entry to 6644919.
RedefineFinal again with Jim new debugging output:

New JDI_REGRESSION failures (from 2008.01.25)
* XXX/com/sun/jdi/RedefineFinal.sh
        This test failed due to "Fail: main: Error output found:
        "Virtual Machine Error" in ..." on Win32 Server VM (machine
        vm-x86-3). However, the dump of the file doesn't show the magic
        string. This is another occurrence of the following bug:

            6644919 4/4 test scaffold error output detection is flakey

        Jim's debugging output shows the following:
            JJX DEBUG: string found ================================================
              cmd = cat C:/local/4225.JDK7.NIGHTLY.VM+windows-i586_server_comp_JDI_REGRESSION/results/classes/com/sun/jdi/aa3916/debuggeeOutput.txt | sed -e 's@> @@g' -e 's@>@@g' | egrep /Virtual Machine Error/
              stat = 0
              result = //

        I will add this entry to 6644919.

http://sqeweb.sfbay/nfs/results/vm/gtee/JDK7/NIGHTLY/VM/2008-01-25/Serv_Baseline/vm/windows-i586/server/comp/vm-windows-i586_server_comp_JDI_REGRESSION2008-01-25-19-58-17/com/sun/jdi/RedefineFinal.jtr
Another sighting in nightly:

New JDI_REGRESSION failures (from 2008.09.04)
* XXX/com/sun/jdi/RedefineIntConstantToLong.sh
        This test failed due to 'Fail: main: Error output found:
        "Virtual Machine Error" in C:/local/.../debuggeeOutput.txt'
        on Win32 Client VM (machine vmnightly2). The following bug
        is used to track this issue:

            6644919 4/4 test scaffold error output detection is flakey

        I will add this entry to the bug.
Another sighting in nightly:

New JDI_REGRESSION failures (from 2008.09.06)
* com/sun/jdi/JdbVarargsTest.sh
* com/sun/jdi/RedefineTTYLineNumber.sh
        These tests failed due to 'Fail: main: Error output found:
        "Virtual Machine Error" in C:/local/.../debuggeeOutput.txt'
        on Win32 Client VM (machine vmnightly2). The following bug
        is used to track this issue:

            6644919 4/4 test scaffold error output detection is flakey

        I will add this entry to the bug.
Another sighting in nightly:

New JDI_REGRESSION failures (from 2008.09.10)
* com/sun/jdi/RedefineIntConstantToLong.sh
        This test failed due to 'Fail: main: Error output found:
        "Virtual Machine Error" in C:/local/.../debuggeeOutput.txt'
        on Win32 Client VM (machine vmnightly2). The following bug
        is used to track this issue:

            6644919 4/4 test scaffold error output detection is flakey

        I have added this entry to the bug.