Unhandled exception at 0x100116cb

[pecan204]

Need some help on old f file with an integer function.

I get a break "Unhandled exception at 0x100116cb" at

IF

(ZENTRY.EQ. 1) THEN

ZIMIN=IMIN

When this function is called. Does any know why this is so. 

I am using IA32 and v9.

Perhaps compiler rules changed from old to new coding?

Thanks

Ken

The whole function looks like:

  INTEGER FUNCTION limit(ZIMIN,ZI,ZIMAX)
 USE IFPORT
 INTEGER ZN
 REAL ZARG
 INTEGER ZIMAX
 INTEGER ZI
 INTEGER ZIMIN
 INTEGER ZENTRY
 REAL TABLE(ZN)
 
!C.....INTEGER FUNCTION LIMITS I BETWEEN IMIN, AND IMAX.

 LIM(MIN,MAX)=MAX0(MIN,MIN0(I,MAX))
 ZENTRY= 1
 IMIN=ZIMIN
 I=ZI
 IMAX=ZIMAX

 LIMIT=LIM(IMIN,IMAX)
 GOTO 3

 ENTRY TLU(TABLE,ZARG,ZN,ZI)
 TLU=0.
 I=ZI
 N=ZN
 ARG=ZARG
 ZENTRY= 2

 I=LIM(1,N)
 IF (ARG.GE.TABLE(I)) GO TO 2

!C.....DESCEND IN TABLE.

1 I=I-1
 IF (I.LE.0) GOTO 3
 IF (ARG.GE.TABLE(I)) GOTO 3
 GO TO 1

!C.....ASCEND IN TABLE.

2 IF (I.GE.N) GOTO 3
 IF (ARG.LT.TABLE(I+1)) GOTO 3
 I=I+1
 GO TO 2
3 CONTINUE

 IF (ZENTRY.EQ. 1) THEN
 ZIMIN=IMIN
 ZI=I
 ZIMAX=IMAX
 elseif(ZENTRY.EQ. 2) THEN
 ZARG=ARG
 ZN=N
 ZI=I
 ENDIF
 END FUNCTION

[Steven_L_Intel1]

We'd need to see a sample program that calls this function.

[pecan204]

This is the only sub that calls it.

SUBROUTINE fntrpa(A,ZX,ZNA)INTEGER ZNAREAL ZXEXTERNAL LIMIT
!C.....COMMON SUBROUTINE EVALUATES A COEFFICIENTS IN DOUBLE
!C.....3-POINT INTERPOLATIONS.
!C L=NO. OF POINTS IN THE FIT
!C I=INDEX TO FIRST POINTREAL A(ZNA)COMMON /NTRPC1/ L,I,A11,A12,A13,A14,A22,A23,A24,A33,A34,A44
X = ZX
NA = ZNA
!C.....GET I AND L BY TABLE LOOK-UP.
L = LIMIT (1,NA,3)
M = MAX0(1,NA-2)
!C CALL TLU (A,X,NA,J)
DUMMY = TLU(A,X,NA,J)
!C
!!CJJC Changed following line to correct possible divide by zero condition
!C
!CJJC IF (J .EQ. LIMIT(2,J,M)) L=4IF (J .EQ. LIMIT(2,J,M) .AND. NA .GT. 3) L=4
I = LIMIT(1,J-1,M)
!C.....CALCULATE A-ARRAY.
A11 = A(I)
A22 = A(I+1)IF(I+2 .GT. ZNA)THEN
A33 = A(ZNA)ELSE
A33 = A(I+2)END IFIF (L .NE. 4) IF (L-2) 3,2,1IF(I+3 .GT. ZNA)THEN
A44 = A(ZNA)ELSE
A44 = A(I+3)END IF
A14 = A11-A44
A24 = A22-A44
A34 = A33-A44
A44 = X-A441 A13 = A11-A33
A23 = A22-A33
A33 = X-A332 A12 = A11-A22
A22 = X-A22
A11 = X-A113 GOTO 44 CONTINUE
ZX = X
ZNA = NA5 RETURNEND

[Steven_L_Intel1]

It's what I thought. You call LIMIT with 1 as the first argument, ZIMIN, and then assign to ZIMIN in the function. This is not legal. You should create local variables with the min and max to be updated, setting them initially at the beginning of the routine.