Unidentified routine malfunction

In most circumstances, an unidentified routine malfunction does not occur when using the compiler. If your routine terminates abnormally without an accompanying Language Environment runtime diagnostic message, the error causing the termination might also be inhibiting the production of a message. Check for the following:

• Your job control statements might be in error, particularly in defining data sets.
• Your routine might overwrite main storage areas containing executable instructions. This can happen if you have accidentally:
  • Assigned a value to a nonexistent array element. For example:

    DCL ARRAY(10);
         ⋮
    DO I = 1 TO 100;
    ARRAY(I) = VALUE;

    To detect this type of error in a compiled module, set the SUBSCRIPTRANGE condition so that each attempt to access an element outside the declared range of subscript values raises the SUBSCRIPTRANGE condition. If there is no ON-unit for this condition, a diagnostic message is printed and the ERROR condition is raised. This facility, though expensive in run time and storage space, is a valuable routine-testing aid.
  • Used an incorrect locator value for a locator (pointer or offset) variable. This type of error can occur if a locator value is obtained by means of record-oriented transmission. Ensure that locator values created in one routine, transmitted to a data set, and subsequently retrieved for use in another routine, are valid for use in the second routine.
  • Attempted to free a nonbased variable. This can happen when you free a based variable after its qualifying pointer value has been changed. For example:

    DCL A STATIC,B BASED (P);
    ALLOCATE B;
    P = ADDR(A);
    FREE B;

  • Used the SUBSTR pseudovariable to assign a string to a location beyond the end of the target string. For example:

    DCL X CHAR(3);
    I=3
    SUBSTR(X,2,I) = 'ABC';

    To detect this type of error, enable the STRINGRANGE condition during compilation.