Re: [Gm2] latest cvs changes

[Dragiša Durić]

Gaius,

If LGPL suits you, regard all my testcases as LGPL. If not, at your
discretion, then GPL is also ok. My preference is LGPL, but better gm2
comes first.

Thanks for fixes
dd

On Fri, 2010-08-27 at 10:47 -0400, address@hidden wrote:
> Hi,
> 
> I've just checked in quite a number of changes (largely to fix the bugs 
> reported by
> Dragiša Durić).  These certainly make the compiler more strict - many fixes to
> typed constants.  To try and keep track of these bug fixes I've added 
> Dragiša's
> tests into a non-free testsuite.  Currently of the 13 reported, these are 
> fixed:
> 
> Running 
> /home/gaius/GM2/graft-4.1.2/gcc-4.1.2/gcc/testsuite/gm2/non-free/dragisa-duric/testcase01/pass/gm2.exp
>  ...
> Running 
> /home/gaius/GM2/graft-4.1.2/gcc-4.1.2/gcc/testsuite/gm2/non-free/dragisa-duric/testcase02/pass/gm2.exp
>  ...
> Running 
> /home/gaius/GM2/graft-4.1.2/gcc-4.1.2/gcc/testsuite/gm2/non-free/dragisa-duric/testcase03/pass/gm2.exp
>  ...
> Running 
> /home/gaius/GM2/graft-4.1.2/gcc-4.1.2/gcc/testsuite/gm2/non-free/dragisa-duric/testcase04/pass/gm2.exp
>  ...
> Running 
> /home/gaius/GM2/graft-4.1.2/gcc-4.1.2/gcc/testsuite/gm2/non-free/dragisa-duric/testcase12/pass/gm2.exp
>  ...
> Running 
> /home/gaius/GM2/graft-4.1.2/gcc-4.1.2/gcc/testsuite/gm2/non-free/dragisa-duric/testcase13/pass/gm2.exp
> 
> regards,
> Gaius
> 
> 
> The change log is here:
> 
>    * gm2/testsuite/gm2/pim/pass/subrange7.mod:  corrected testcase
>      in light of new changes.
>    * gm2/testsuite/gm2/pim/pass/wordconst.mod:  corrected testcase
>      in light of new changes which make gm2 stricter.
>    * gm2/testsuite/gm2/pim/run/pass/str6.mod:  use HALT rather than
>      Terminate, exit.
>    * The following changes are bug fixes to a series of bugs reported
>      by DragiÅ¡a DuriÄ\207, many of which relate to constants, their
>      types, agregates and qualidents in constants.
>    * gm2/gm2/bnf/m2-3.bnf:  introduce a production for
>      SilentConstructor which removes a constructor from the fifo queue.
>    * gm2/gm2/bnf/m2-c.bnf:  create constant symbols and fill in aggregates
>      with their associated types before pass 3.
>    * gm2/gm2/bnf/m2-h.bnf:  introduce a production for
>      SilentConstructor which removes a constructor from the fifo queue.
>    * gm2/gm2/gm2-compiler/M2ALU.mod:  initialise aggregates and test for
>      the special case of string.
>    * gm2/gm2/gm2-compiler/M2Code.mod:  make a call to ResolveConstTypes
>      and ResolveConstructorTypes at the end of DoPassC.
>    * gm2/gm2/gm2-compiler/M2GenGCC.mod:  improve debugging of PrintVerbose.
>    * gm2/gm2/gm2-compiler/M2Quads.def:  (New procedures) 
> SilentBuildConstructor,
>      SilentBuildConstructorStart, BuildConstructor.
>    * gm2/gm2/gm2-compiler/M2Quads.mod:  (implement new procedures)
>      SilentBuildConstructor, SilentBuildConstructorStart, BuildConstructor.
>        Change to BuildDesignator array now that the aggregate type will be 
> known.
>    * gm2/gm2/gm2-compiler/P2SymBuild.def:  (New procedures) 
> FixupConstAsString,
>      FixupConstType, FixupConstExpr.
>    * gm2/gm2/gm2-compiler/P2SymBuild.mod:  (New procedures) findConstMeta,
>      findConstType, fixupConstCast, fixupConstMeta, addToConstList.
>    * gm2/gm2/gm2-compiler/P3SymBuild.mod:  remove import of
>      GetConstructorFromFifoQueue.
>    * gm2/gm2/gm2-compiler/PCSymBuild.mod:  many changes.  New procedures
>      StartCurrentConst, EndCurrentConst, SetCurrentConstToString,
>      SetCurrentConstType, SetConstTypeOrExpr.
> 
> once all outstanding bugs have been fixed and pass on 32 bit i386 and
> 64 bit x86 then I think release 1.0 should occur.
> 
> regards,
> Gaius
> 
> _______________________________________________
> gm2 mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/gm2

-- 
Dragiša Durić <address@hidden>