cyclone/docs/api/srfi/2.md
2016-12-09 23:49:42 -05:00

73 lines
4.2 KiB
Markdown

# SRFI 2 - `and-let*`
The `(srfi 2)` library provides the `and-let*` macro, an `and` with local bindings.
# Overview
Like an ordinary `and`, an `and-let*` special form evaluates its arguments -- expressions -- one after another in order, till the first one that yields `#f`. Unlike `and`, however, a non-#f result of one expression can be bound to a fresh variable and used in the subsequent expressions. `and-let*` is a cross-breed between `let*` and `and`.
See the [SRFI document](http://srfi.schemers.org/srfi-2/srfi-2.html) for more information.
- [`and-let*`](#and-let)
# Rationale
In case of an ordinary and formed of proper boolean expressions:
(and E1 E2 ...)
expression E2, if it gets to be evaluated, knows that E1 has returned non-#f. Moreover, E2 knows exactly what the result of E1 was -- #t -- which E2 can use to its advantage. If E1 however is an extended boolean expression, E2 can no longer tell which particular non-#f value E1 has returned. Chances are it took a lot of work to evaluate E1, and the produced result (a number, a vector, a string, etc) may be of value to E2. Alas, the `and` form merely checks that the result is not an #f, and throws it away. If E2 needs it, it has to compute that value anew. This proposed `and-let*` special form lets constituent expressions get hold of the results of already evaluated expressions, without re-doing their work.
`and-let*` can be thought of as a combination of `let*` and `and`, or a generalization of `cond`'s send operator `=>`. An `and-let*` form can also be considered a sequence of guarded expressions. In a regular program, forms may produce results, bind them to variables and let other forms use these results. `and-let*` differs in that it checks to make sure that every produced result "makes sense" (that is, not an `#f`). The first "failure" triggers the guard and aborts the rest of the sequence (which presumably would not make any sense to execute anyway). Examples:
(and-let* ((my-list (compute-list)) ((not (null? my-list))))
(do-something my-list))
(define (look-up key alist)
(and-let* ((x (assq key alist))) (cdr x)))
(or
(and-let* ((c (read-char))
((not (eof-object? c))))
(string-set! some-str i c)
(set! i (+ 1 i)))
(begin (do-process-eof)))
; A more realistic example
; Parse the 'timestamp' ::= 'token1' 'token2'
; token1 ::= 'YY' 'MM' 'J'
; token2 ::= 'GG' 'gg' "/"
(define (parse-full-timestamp token1 token2)
(and-let* (((= 5 (string-length token1)))
((= 5 (string-length token2)))
(timestamp
(OS:string->time "%m/%d/%y %H:%M"
(string
(string-ref token1 2) (string-ref token1 3) #\/
(string-ref token1 0) (string-ref token1 1) #\/
(case (string-ref token1 4)
((#\8 #\9) #\9) (else #\0))
(string-ref token1 4) #\space
(string-ref token2 0) (string-ref token2 1) #\:
(string-ref token2 2) (string-ref token2 3))))
((positive? timestamp)))
timestamp))
`and-let*` is also similar to an "anaphoric AND" LISP macro [Rob Warnock, comp.lang.scheme, 26 Feb 1998 09:06:43 GMT, Message-ID: 6d3bb3$3804h@fido.asd.sgi.com]. `and-let*` allows however more than one intermediate result, each of which continues to be bound through the rest of the form.
# and-let*
*Syntax*
(and-let* (claws) body)
claws ::= '() | (cons claw claws)
claw ::= (variable expression) | (expression) |
bound-variable
- The `claws` are evaluated in the strict left-to-right order
- For each `claw`, the `expression` part is evaluated first (or `bound-variable` is looked up)
- If the result is `#f`, `and-let*` immediately returns `#f`
- Otherwise, if the `claw` is of the form `(variable expression)` the `expression`'s value is bound to a freshly made `variable`
- The `variable` is available for the rest of the `claws` , and the `body`
- As usual, all `variable`s must be unique (like in `let*`)