(j3.2006) IEEE_SET_FLAG

Bill Long longb
Wed Nov 7 15:15:30 EST 2012



On 11/7/12 1:49 PM, Van Snyder wrote:
> There is a requirement that the routine IEEE_SET_FLAG behave in a
> peculiar way. The argument FLAG is an array of type(IEEE_FLAG_TYPE).
>   But the entries are required to be unique.
>
This is a "shall" requirement on the program.

> If they are not unique, then what should happen?  Does an error

If they are not unique, the program is not conforming.   I don't see 
where we explain what happens with these functions if bad arguments are 
supplied.

> condition exist?  Should it be handled similarly to SQRT(-1.0)?  Is the
> result processor dependent?  Annex A.2 doesn't say so.  Can the elements

I don't think we want to occupy Annex A.2 with descriptions of what 
happens in all the cases of non-conforming programs.   The Annex is long 
enough as it is.

> in FLAG and FLAG_VALUE be processed in array-element order, ignoring
> duplicates?

The processor can do whatever it wants.  Your suggestion is one 
possibility.  But what is a "duplicate" - do both the flag and value 
have to match an earlier pair, or is the duplication only on the flag? 
A reasonable algorithm would be to process the flag / value pairs in 
array element order, with the value corresponding to the last instance 
of a particular flag being the one used.  On the other hand, the 
implementation could limit the length of the search to the total number 
of supported flags, and ignore the array entries that come after.

Cheers,
Bill

>
> --
> Van Snyder                    |  What fraction of Americans believe
> Van.Snyder at jpl.nasa.gov       |  Wrestling is real and NASA is fake?
> Any alleged opinions are my own and have not been approved or
> disapproved by JPL, CalTech, NASA, the President, or anybody else.
>
> _______________________________________________
> J3 mailing list
> J3 at mailman.j3-fortran.org
> http://mailman.j3-fortran.org/mailman/listinfo/j3
>

-- 
Bill Long                                           longb at cray.com
Fortran Technical Support    &                 voice: 651-605-9024
Bioinformatics Software Development            fax:   651-605-9142
Cray Inc./Cray Plaza, Suite 210/380 Jackson St./St. Paul, MN 55101





More information about the J3 mailing list