Uploaded image for project: 'JDK'
  1. JDK
  2. JDK-4040354

Two dozen assorted JCK test failures on Windows NT with the JIT enabled.

    XMLWordPrintable

Details

    • jit
    • 1.1.1
    • generic
    • solaris_2.4
    • Verified

    Description


      allan.jacobs@Eng 1997-03-20

      The list of JCK tests that the NT version of the JIT fails is:

      api/java_io/PipedInputStream/EquivClass.html#PipedInputStream0019
      api/java_io/PipedInputStream/EquivClass.html#PipedInputStream2006
      api/java_lang/Double/DoubleTests.html#DoubleTests19
      api/java_lang/Float/FloatTests.html#FloatTests1
      api/java_lang/Float/FloatTests.html#FloatTests40
      api/java_lang/Math/MathTests.html#MathTests14
      api/java_lang/Math/MathTests.html#MathTests69
      api/java_lang/ThreadGroup/ThreadGroupTests.html#ThreadGroupTests51
      api/java_math/BigDecimal/index.html#BigDecimalDivision
      lang/BINC/binc001/binc00101/binc00101.html
      lang/BINC/binc017/binc01701/binc01701.html
      lang/BINC/binc020/binc02001/binc02001.html
      lang/BINC/binc020/binc02002/binc02002.html
      lang/BINC/binc051/binc05102/binc05102.html
      lang/EXEC/exec005/exec00503/exec00503.html
      vm/instr/areturn/areturn006/areturn00601/areturn00601.html
      vm/instr/athrow/athrow009/athrow00901/athrow00901.html
      vm/instr/athrow/athrow009/athrow00902/athrow00902.html
      vm/jni/AllocObject/alob001/alob00101/alob00101.html
      vm/jni/CallDoubleMethodA/cdma001/cdma00102/cdma00102.html
      vm/jni/CallNonvirtualCharMethod/cncm001/cncm00102/cncm00102.html
      vm/jni/CallNonvirtualObjectMethodA/cnoa001/cnoa00102/cnoa00102.html
      vm/jni/CallBooleanMethod/czmt001/czmt00102/czmt00102.html
      vm/jni/SetShortField/ssfl001/ssfl00101/ssfl00101.html

      The BINC tests have probably been fixed by modifications made to the
      Intel Solaris side. The same is true of the tests areturn00601,
      athrow00901, and athrow00902.


      allan.jacobs@Eng 1997-04-02
      The test ThreadGroupTests51 contains calls to suspend. Bug 4041469 says that
      this can cause a problem in a native threads implementation of the VM.
      Also, there is no reason to expect that any two calls to either of the
      ThreadGroup methods activeCount or activeGroupCount should remain
      constant. The test relies on this property.

      Attachments

        Activity

          People

            dspottsunw David Spott (Inactive)
            ajacobssunw Allan Jacobs (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:
              Imported:
              Indexed: