A connection block declaration can contain two special types of procedures: a PROLOG procedure, an EPILOG procedure, or both. These procedures are available in both ALGOL and NEWP.

PROLOG Procedures

For each connection in any connection library based on that connection block TYPE, the PROLOG procedure is executed the first time that connection is used in any way. The first use of a connection might occur in any of the following ways:

EPILOG Procedures

By contrast, an EPILOG procedure specifies the actions performed for each connection just before the connection block is deleted. The connection block is deleted when a program exits the procedure (or outer block) in which the connection block is declared. A separate instance of the EPILOG procedure is executed for each active connection of each connection library that is based on that connection block. In this context, a connection is considered active if it is currently linked or was previously linked and then delinked. (If a particular connection was declared, but never used, then the EPILOG procedure is not executed for that connection.)

Uses Outside Connection Blocks

PROLOG and EPILOG procedures can also be declared in structure blocks, where they behave similarly to the way they behave for connection blocks. Furthermore, EPILOG procedures can be declared in any procedure, and specify actions to be performed when the procedure is exited (refer to Using EPILOG and EXCEPTION Procedures in Using Events and Interlocks.

Order of Execution for EPILOG Procedures

A connection block EPILOG procedure is executed before any EPILOG procedure for the procedure that declares the connection block.

PROLOG and EPILOG Example

The following ALGOL example includes various combinations of PROLOG and EPILOG procedures:

100 PROCEDURE OUTERPROC;
110    BEGIN
115    EBCDIC ARRAY LIBF[0:23];
120
130    EPILOG PROCEDURE OUTER_EPILOG;
140       BEGIN
150       % Procedure body statements
160       END;
170
180    TYPE CONNECTION BLOCK CL_TYPE;
190       BEGIN
200       INTEGER CONN_INT;
210
220       PROLOG PROCEDURE CL_PROLOG;
230         BEGIN
240         % Procedure body statements
250         END;
260
270       EPILOG PROCEDURE CL_EPILOG;
280         BEGIN
290         % Procedure body statements
300         END;
310
320       PROCEDURE CL_INIT(I);
330         INTEGER I;
340         BEGIN
350         CONN_INT:= I;
360         END;
370
380       PROCEDURE CL_EXP;
390         BEGIN
400         % Procedure body statements
410         END;
420       EXPORT CLEXP;
430       END;
440
450    CL_TYPE LIBRARY CL1 (LIBACCESS=BYFUNCTION, FUNCTIONNAME=“F1.”,
455                         CONNECTIONS=3);
460
470    % End declarations, begin statements of procedure OUTERPROC
480    CL1[0].CL_INIT(5);
485    REPLACE LIBF BY “CLTEST1.”;
490    LINKLIBRARY (CL1[0], WAITFORFILE, INTERFACENAME = LIBF);
500    REPLACE LIBF BY “CLTEST2.”;
510    LINKLIBRARY (CL1[1], WAITFORFILE, INTERFACENAME = LIBF);
520
530 END OUTERPROC;

In the preceding example, the statement at line 480 is the first statement to use connection 0, and therefore causes the CL_PROLOG procedure to be executed for connection 0. The statement at line 510 causes the CL_PROLOG procedure to be executed for connection 1. The CL_PROLOG procedure is never executed for connection 2, because connection 2 is never actually used in this example.

When the OUTERPROC procedure is being exited, the system invokes separate instances of the CL_EPILOG procedure for connections 0 and 1. CL_EPILOG is not invoked for connection 2, because connection 2 is never used in this example. Then the system invokes the OUTER_EPILOG procedure. OUTER_EPILOG is executed only once, because OUTER_EPILOG is associated with the procedure OUTERPROC rather than with the connection block CL_TYPE.