[Patch, fortran] PR30887 and PR30888 - overdone constraints on %VAL

[Jerry DeLisle]

Tobias Burnus wrote:
> *ping*
> 
> Fixes:
> - %VAL( ) with non-default kinds now supported
> - The following program gave a compile-time error:
> 
> +      CALL DOIT( %VAL( P ) )
> +      CALL DOIT( I )
> +      CALL DOIT( %VAL( P ) )
> 
> 
> Tobias Burnus wrote:
>> Paul Thomas wrote:
>>   
>>> The attached are completely self explanatory fixes for the above to
>>> %VAL problems.
>>> 

Tobias,

With the patch of 2/22/2007 I am getting the following failures:

Running /home/jerry/gcc/gcc43/gcc/testsuite/gfortran.dg/dg.exp ...
FAIL: gfortran.dg/size_optional_dim_1.f90  -O0  execution test
FAIL: gfortran.dg/size_optional_dim_1.f90  -O1  execution test
FAIL: gfortran.dg/size_optional_dim_1.f90  -O2  execution test
FAIL: gfortran.dg/size_optional_dim_1.f90  -O3 -fomit-frame-pointer  execution test
FAIL: gfortran.dg/size_optional_dim_1.f90  -O3 -fomit-frame-pointer 
-funroll-loops  execution test
FAIL: gfortran.dg/size_optional_dim_1.f90  -O3 -fomit-frame-pointer 
-funroll-all-loops -finline-functions  execution test
FAIL: gfortran.dg/size_optional_dim_1.f90  -O3 -g  execution test
FAIL: gfortran.dg/size_optional_dim_1.f90  -Os  execution test


Do I have the right patch? or is this some other regression?

Jerry