[patch, fortran] PR43851 Add _gfortran_error_stop_numeric

[Jerry DeLisle]

On 05/19/2010 05:07 AM, Janne Blomqvist wrote:
> On Wed, May 19, 2010 at 04:33, Jerry DeLisle<jvdelisle@verizon.net>  wrote:
>> Hi all,
>>
>> I hope this is the final on this patch.  This patch avoids breaking ABI by
>> creating two new functions stop_numeric_f2008 and pause_numeric_f2008 along
>> with the error_stop_numeric.  This allows the STOP and PAUSE statements to
>> accept constant expressions and not use a value of -1 to signal an empty
>> statement. Existing stop_numeric and pause_numeric are untouched.
>
> Yes, this avoids breaking the ABI. Although I think the same could
> have been achieved by translating an empty stop code to
> (error_)stop_string with a NULL string argument, without having to
> create new library functions. But, this will get the job done as well,
> so I think it's ok, no need to bikeshed it to death.

I liked this idea better, so this is what I did. No duplicate functions created 
and no ABI breakage. Added check for default character kind as well.  I will 
commit a new test case or two next window of opportunity.  Seems to work fine 
with -fdefault-integer-8. Regression tested on x86-64.

Committed:

$ svn commit
Sending        gcc/fortran/ChangeLog
Sending        gcc/fortran/match.c
Sending        gcc/fortran/trans-decl.c
Sending        gcc/fortran/trans-stmt.c
Sending        gcc/fortran/trans.h
Sending        gcc/testsuite/ChangeLog
Sending        gcc/testsuite/gfortran.dg/label_1.f90
Sending        libgfortran/ChangeLog
Sending        libgfortran/gfortran.map
Sending        libgfortran/libgfortran.h
Sending        libgfortran/runtime/pause.c
Sending        libgfortran/runtime/stop.c
Transmitting file data ............
Committed revision 159609.

Regards,

Jerry