Search                        Top                                  Index
HELP CHAIN                                         R. Barrett et al 1984
                                         Updated: Adrian Howard Mar 1992


The procedure -chain- takes one argument, a procedure. When a procedure
C calls -chain-, control is immediately transferred to P. Any remaining
instructions in C are never executed. Local variable values are not
passed on from C to the procedure called by -chain-. When the procedure
given as argument to -chain- is run, all variable values are set to what
they were when C inherited them. When P finishes, control returns to the
procedure which originally called C.

Procedure -b-, below, has been altered to "chain" -c- rather than
calling it normally:

    vars x;

    define c;
        npr('running C');
        x * 2 -> x; x =>
        npr('C finished');

    define b;
        dlocal x;
        npr('running B');
        x * 2 -> x; x =>
        chain(c);           ;;; the call to -chain- to procedure -c-
        x =>
        npr('B finished');

    define a;
        dlocal x;
        npr('running A');
        3 -> x; x =>
        npr('continuing to run A');
        x =>
        npr('A finished');

The effect of the non-standard control structure can be seen when we run

    running A
    ** 3                ;;; -x- is 3 in -a-

    running B
    ** 6                ;;; -x- is 6 in -b-. This value for -x- will not
                        ;;; be passed on to -c-, as it is local to -b-.
                        ;;; At this point -c- is "chained" to.

    running C           ;;; -x- is reset to what it was before entry to
                        ;;; -b-, i.e. 3.

    ** 6                ;;; -c- immediately doubles the value of -x- and
                        ;;; prints it out

    C finished continuing to run A
    ** 6                ;;; -x- is global to -c-, and is therefore not
                        ;;; reset when control returns from -c- to -a-
                        ;;; (not to -b-).
    A finished

The caller of -chain-, procedure -b-, is exited from, and a -chain- to
-c- is made. The value of -x- in -c- is set as it was when -b- inherited
it (to what it was in -a-.) When -c- finishes, control returns to -a-,
not to -b-. The instructions in -b- after the call of -c- are never
executed. Since -x- is not local to -c-, its value is not reset in -a-.

For other non-standard control structures, see *CHAINFROM, *CHAINTO,

Also see:

        Procedure identifying a single caller at a specified level

        On POP-11 interrupt procedures

        A summary of control structures in POP-11

        Full details of POP-11 procedure calling mechanisms

--- C.all/help/chain
--- Copyright University of Sussex 1992. All rights reserved. ----------