Re: [Gm2] CASE: gm2-4.1.2 of 2011-11-08 bug (or possibly philosophical difference) (CVS fixes)

[Gaius Mulley]

Hi,

all fixed now in the latest cvs on the gcc-4.1.2 branch:

   * gm2/gm2-compiler/M2CaseList.def:  add TypeCaseBounds to the
     definition module.
   * gm2/gm2-compiler/M2CaseList.mod:  GetVariantTagType,
     checkTypes,
     inRange, TypeCaseBounds (New procedure functions).  Check to
     see that case bounds are of the correct type.
     * gm2/gm2-compiler/M2MetaError.mod:  fix internal error
     message.
     * gm2/gm2-compiler/M2Options.def:  new boolean
     VariantValueChecking
     which is set if -fiso is used.  FinaliseOptions new procedure.
     * gm2/gm2-compiler/M2Quads.mod:  reverse PopT(r1) and PopT(r2)
     to correct range bug.
     * gm2/gm2-compiler/M2Range.mod:  only check missing values
     in variant record case bounds if VariantValueChecking is
     enabled.
     We always check for overlapping bounds.
     * testsuite/gm2/pim/fail/variant.mod:  new file.
   * testsuite/gm2/pim/fail/variant2.mod:  new file.
   * testsuite/gm2/pim/fail/variant3.mod:  new file.

regards,
Gaius