Loading ...
Sorry, an error occurred while loading the content.

Re: [eiffel_software] Runtime panic

Expand Messages
  • Colin Paul Adams
    ... Emmanuel Usually it means a crash during a GC Cycle due to some Emmanuel object corruptions, in this case, rescue clause are not Emmanuel operational.
    Message 1 of 5 , Jul 1, 2006
    • 0 Attachment
      >>>>> "Emmanuel" == Emmanuel Stapf [ES] <Emmanuel> writes:

      >> Looking at the failing routine - make_area - suggests that
      >> maybe memory was exhausted, but I had presumed I would get an
      >> exception in that case.

      Emmanuel> Usually it means a crash during a GC Cycle due to some
      Emmanuel> object corruptions, in this case, rescue clause are not
      Emmanuel> operational.

      This is terrible for my application, which is a driver for the XSLT
      2.0 test suite. There are over 4000 tests, so the driver parses the
      test catalog (an XML file), and for each test case, runs an XSLT
      transformation. It includes a rescue clause in case the XSLT processor
      crashes, so it can simply report a failure for that test, and carry
      on.

      Instead, the whole test program falls over.

      I was able to isolate the problem - a bug in the XSLT library is
      causing an infinite recursion, which will cause the stack to be
      exhausted (running it in EiffelStudio triggers stack overflow
      warnings). Of course, i shall have to fix my bug, but shouldn't we be
      able to get an out-of-memory exception in the rescue clause for such a situation?
      --
      Colin Adams
      Preston Lancashire
    • Colin Paul Adams
      ... Colin This is terrible for my application, which is a driver for Colin the XSLT 2.0 test suite. There are over 4000 tests, so the Colin driver parses
      Message 2 of 5 , Jul 2, 2006
      • 0 Attachment
        >>>>> "Colin" == Colin Paul Adams <colin@...> writes:

        >>>>> "Emmanuel" == Emmanuel Stapf [ES] <Emmanuel> writes:

        Colin> This is terrible for my application, which is a driver for
        Colin> the XSLT 2.0 test suite. There are over 4000 tests, so the
        Colin> driver parses the test catalog (an XML file), and for each
        Colin> test case, runs an XSLT transformation. It includes a
        Colin> rescue clause in case the XSLT processor crashes, so it can
        Colin> simply report a failure for that test, and carry on.

        Colin> Instead, the whole test program falls over.

        Having fixed my bug, and so side-stepped the panic, it now occurs
        later on - this time for a floating point exception.
        WHY???
        --
        Colin Adams
        Preston Lancashire
      • Emmanuel Stapf [ES]
        ... Hard to say. Can you send a bug report with the steps required to reproduce the problem? Regards, Manu
        Message 3 of 5 , Jul 2, 2006
        • 0 Attachment
          > Having fixed my bug, and so side-stepped the panic, it now
          > occurs later on - this time for a floating point exception.
          > WHY???

          Hard to say. Can you send a bug report with the steps required to reproduce the
          problem?

          Regards,
          Manu
        Your message has been successfully submitted and would be delivered to recipients shortly.