🚀 go-pugleaf

In article <1992Oct12.112837.18130@monu6.cc.monash.edu.au>,
bena@yoyo.cc.monash.edu.au (Ben Aveling) writes:
>I'm sure anyone who has ever used Fortran has seen the ***** type output
>when a number is too big to fit in it's format statement.  My gripe is
>that every time my programs do that, it counts as a warning, and does a
>stack dump.  Can some kind soul please tell me how I can stop this
>happening, short of turning off all messages.

This probably isn't the answer you wanted, since it involves editting your
source (presumably a lot of it) but if there is an iostat=<xx> specifier in your
I/O statement, the stack dump won't appear.

I suppose the correct solution that doesn't involve changing your I/O statements
would be to provide a condition handler that checked for that specific condition
and resignalled all others. I'd imagine you could put such a condition handler
on your main program. I expect it is the default condition handler that produces
the stack dump. If it's one down in the FORTRAN RTL, then this won't work, and
you'll have to use IOSTAT.
--
Terry Poot <tp@mccall.com>                   The McCall Pattern Company
(uucp: ...!rutgers!depot!mccall!tp)          615 McCall Road
(800)255-2762, in KS (913)776-4041           Manhattan, KS 66502, USA

In article <1992Oct12.112837.18130@monu6.cc.monash.edu.au>, bena@yoyo.cc.monash.edu.au (Ben Aveling) writes:
> I'm sure anyone who has ever used Fortran has seen the ***** type output
> when a number is too big to fit in it's format statement.  My gripe is
> that every time my programs do that, it counts as a warning, and does a
> stack dump.  Can some kind soul please tell me how I can stop this
> happening, short of turning off all messages.

Simply add the IOSTAT parameter to your WRITE statements and ignore the
value returned:

	INTEGER*4 OK
	...
	WRITE (6,801,IOSTAT=OK) X,Y,Z
	...

		Cheers, Ken
-- 
 Dr. Kenneth H. Fairfield    |  Internet: Fairfield@Slacvx.Slac.Stanford.Edu
 SLAC, P.O.Box 4349, MS 98   |  DECnet:   45537::FAIRFIELD (45537=SLACVX)
 Stanford, CA   94309        |  BITNET    Fairfield@Slacvx
 ----------------------------------------------------------------------------
 These opinions are mine, not SLAC's, Stanford's, nor the DOE's...

In article <1992Oct12.112837.18130@monu6.cc.monash.edu.au>,
bena@yoyo.cc.monash.edu.au (Ben Aveling) writes:
> I'm sure anyone who has ever used Fortran has seen the ***** type output
> when a number is too big to fit in it's format statement.  My gripe is
> that every time my programs do that, it counts as a warning, and does a
> stack dump.  Can some kind soul please tell me how I can stop this
> happening, short of turning off all messages.
> 
> 	Ben

Yes, it's a pain isn't it? But you can use LIB$ESTABLISH to get rid of the
messages. Below my signature is a quick and (rather) dirty example that 
should point you in the right direction.

        --- Allan


Allan Engelhardt           | Internet:     Allan.Engelhardt@nbi.dk
Niels Bohr Institute       | HEPnet:       NBIVAX::ENGELHARDT
Blegdamsvej 17             | Bitnet/EARN:  ENGELH at CERNVM
DK-2100 Copenhagen         | Phone direct: (+45)31424284<dial-tone>355
DENMARK                    | Fax:          (+45)31421016

---
      PROGRAM TEST
      REAL R1,R2,R3                 ! These three lines are
      VOLATILE R1,R2,R3             ! used to generate an error
      DATA R2 /1./ , R3 /0./        ! that is not a format error.
      INTEGER*4 HANDLR
      EXTERNAL HANDLR
      CALL LIB$ESTABLISH(HANDLR)    ! This is the trick.
      I = 100
      DO 10 J = 1 , 4
        WRITE(*,1000) I
        I = I*10                    ! J=4 will give a format error.
   10 CONTINUE
      R1=R2/R3                      ! Floating divide zero.
 1000 FORMAT( 1X,'Number is ',I5 )
      END

      INTEGER*4 FUNCTION HANDLR(S,M)
      INTEGER*4 S(*),M(5)
      INCLUDE '($SSDEF)'
      IF ( S(2).EQ.1606138 ) THEN   ! I *said* dirty :-}
        HANDLR = SS$_CONTINUE
      ELSE
        HANDLR = SS$_RESIGNAL
      ENDIF
      RETURN
      END