DECLARE-TYPE-FREETYPE type var1 var2 ...) "... affects only variable bindings". Since declarations can occur in contexts other than establishing "variable bindings", most people interpret this statement to mean that type declarations not in such context are either (1) completely to be ignored, or (2) invalid CL syntax. Thus both of the following forms would be suspect in that the type declarations could not have any effect:
(if (and (typep x 'fixnum) (typep y 'fixnum))
(locally (declare (fixnum x y)) ;LOCALLY does not bind
...algorithm using x and y...) ; any variables.
...similar algorithm using x and y...)
(let ((y 'foo))
(setq y 10)
(let ((x 5)) ;'y' is not being bound in
(declare (fixnum y)) ; this particular context.
(incf y)
...random algorithm...))
In this proposal, a type declaration affects not only variable references within its scope, but also affects variable references that are outside the scope of the declaration but dynamically inside the execution of a form that is itself inside the scope of the declaration. Such references can exist when the variable is SPECIAL or when the declaration is not attached to the variable's binding, so that the scope of the declaration does not include the entire scope of the variable.
Clarify that if nested type declarations refer to the same variable, then the value of the variable must be a member of the intersection of the declared types.
SETQ of the declared variable within the scope of the declaration, it is an error for the newly assigned value of the declared variable not to be of the declared type; and at the moment the scope of the declaration is entered, it is an error for the value of the declared variable not to be of the declared type.
In this proposal, a type declaration affects only variable references within its scope, and the meaning of "free" and "variable-binding-associated" type declarations can be described identically.
This proposal is equivalent to saying that the meaning of a type declaration is equivalent to changing each reference to <var> within the scope of the declaration to (THE <type> <var>), changing each expression assigned to the variable within the scope of the declaration to (THE <type> <new-value>), and executing (THE <type> <var>) at the moment the scope of the declaration is entered.
Clarify that if nested type declarations refer to the same variable, then the value of the variable must be a member of the intersection of the declared types.
;; this is an error under DECLARE-TYPE-FREE:ALLOW: ;; the assertion that x is a fixnum is violated between the two ;; calls to (zap) ;; this is a valid program under DECLARE-TYPE-FREE:LEXICAL
(let ((x 12) (y 'foo))
(flet ((zap () (rotatef x y)))
(locally (declare (fixnum x))
(zap)
(zap)
x)))
;; this is an error under both proposals
(let ((x 12) (y 'foo))
(flet ((zap () (rotatef x y)))
(locally (declare (fixnum x))
(zap)
(print x)
(zap)
x)))
;; this is an error under DECLARE-TYPE-FREE:ALLOW, because ;; the assertion that x is a fixnum ;; is violated during the call to zap, even though few ;; implementations will be able to check: ;; this is a valid program under DECLARE-TYPE-FREE:LEXICAL
(let ((x 12) (y 'foo))
(flet ((zap ()
(rotatef x y)
(rotatef x y)))
(locally (declare (fixnum x))
(zap)
x)))
;; this is an error under both proposals, even though the ;; violation of the type constraint happens after the form ;; with the declaration is exited.
(let ((f (let ((x 3)) (declare (fixnum x)) #'(lambda (z) (incf x z))))) (funcall f 4.3))
DECLARE-TYPE-FREE:ALLOW is more restrictive on programs and hence allows more freedom for optimizing compilers. DECLARE-TYPE-FREE:LEXICAL is easier to understand but allows a specialized representation only where the scope of the variable is the same as the scope of the declaration or the compiler can prove that there are no relevant other references to the variable.
THE wrappers around every reference.DECLARATION-SCOPE, addresses the scope of declarations. This proposal carefully uses the phrase "within the scope of the declaration" to avoid confounding the two issues.
This issue has been discussed at the Fort Collins X3J13 meeting in November 1987, and at length on the various electronic mailing lists.
At least one current implementation is able to generate more efficient code when declarations are associated with a particular binding, since it then has the option to choose type-specific specialized storage for the runtime value of the variable. So, for example,
(let ((x v)) (declare (type float x)) (+ x x))
is sometimes more efficient than
(let ((x v)) (locally (declare (type float x)) (+ x x)))
However, the local type declarations allowed by this proposal do provide some useful information, even if it is not the *most* useful. It is possible for a sufficiently "smart" compiler to infer the equivalent of a "binding declaration" when it can ascertain that the type of the binding value -- 'v' above -- is commensurate with the type locally declared over the scope of usage of the variable.
It may be useful for a compiler to issue a warning whenever it finds nested type declarations referring to the same variable and the intersection of the declared types is null.
Documentation might want to discuss the style implications of nested declarations intersecting. The interesting cases are: - An inner declaration could be a subtype of an outer one. This is the most useful case and probably the only one to be encouraged in code written by humans. e.g., (locally (declare (type number x)) (locally (declare (type integer x)) ...use X as integer...)) - An outer declaration could be a subtype of an inner one. This is useless but harmless. It might happen as the result of certain macro situations. e.g., (locally (declare (type integer x)) (locally (declare (type number x)) ...use X as integer...)) - Two types may only partially overlap. This would presumably happen only as the result of a macro expansion. (locally (declare (type fixnum x)) (locally (declare (type (or bit package) x)) ...use X as BIT...))
*start* 05268 00024 USm GV-Info: X3J13-mailer@SAIL.Stanford.EDU at 7-Apr-89 15:28:47 from AG Return-Path: <X3J13-mailer@SAIL.Stanford.EDU> Received: from SAIL.Stanford.EDU ([36.86.0.194]) by Xerox.COM ; 07 APR 89 14:36:21 PDT Received: from STONY-BROOK.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 7 Apr 89 14:20:08 PDT Received: from EUPHRATES.SCRC.Symbolics.COM by STONY-BROOK.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 573609; Fri 7-Apr-89 17:19:10 EDT Date: Fri, 7 Apr 89 17:18 EDT From: David A. Moon <Moon@STONY-BROOK.SCRC.Symbolics.COM> Subject: Did you blow it? To: Guy Steele <gls@Think.COM>, rpg@lucid.com cc: x3j13@sail.stanford.edu In-Reply-To: <8904071720.AA18178@verdi.think.com> Message-ID: <19890407211851.9.MOON@EUPHRATES.SCRC.Symbolics.COM>
Date: Fri, 7 Apr 89 13:20:57 EDT From: Guy Steele <gls@Think.COM>
Date: Thu, 6 Apr 89 15:19:42 PDT From: Richard P. Gabriel <rpg@lucid.com>
When I read the effect of Issue: DECLARE-TYPE-FREE on page 223, I completely flipped. I think you might have gotten it wrong.
He did.
You say that in this code:
(defun f (x) (declare (type float x)) (let ((x 'a)) ...) ...)
The declaration affects both bindings of x. This cannot make any sense at all. I don't have marked which version of this issue passed, but I think neither implies this. The most that is implied is that if you say this:
(defun f (x) ... (let ((y 'a)) (declare (type float x)) ...) ...)
then the declaration applies to variable references within the let-y and not to within some larger scope.
I hope you are wedged, because otherwise the proposal is wedged.
I also hope I am wedged. I am taking the liberty of cc'ing this to X3J13 so that others can let me know what they think.
I believe I was confused by this paragraph from proposal DECLARE-TYPE-FREE:LEXICAL, passed January 1989:
This proposal is equivalent to saying that the meaning of a type declaration is equivalent to changing each reference to <var> within the scope of the declaration to (THE <type> <var>), changing each expression assigned to the variable within the scope of the declaration to (THE <type> <new-value>), and executing (THE <type> <var>) at the moment the scope of the declaration is entered.
The ambiguity concerns whether in
(defun f (x) (declare (type float x)) x ;reference 1 (let ((x 'a)) ...) x ;reference 2 ...)
the two references are construed to be to the same *variable*.
(I readily grant that they refer to different *bindings*.)
I assumed that they were contrued to be the same variable,
in which case I believe that what I wrote on page 223 of the
CLtL II draft is a correct conclusion.
The phrase "within the scope of the declaration" quoted above is supposed to be a precisely defined phrase. The passed cleanup issue DECLARATION-SCOPE was supposed to define that phrase. Unfortunately, there is a problem: the precise language in version 2 of the proposal was replaced with much less precise language in version 4, which was the version that was voted upon. The version 2 language was:
The scope of a `bound' declaration is exactly the scope of the associated lexical variable or function. If the declaration is associated with a special variable, the scope is the scope the variable would have had if it had not been special.
`Free' declarations are scoped as if they appeared in a new LOCALLY form which surrounded the entire special form at the beginning of whose body the declaration appears. This is the same as what CLtL p.155 defines to be the scope of `pervasive' declarations.
This answers your question about special variables. I think that for declarations that concern variable or function bindings, but are not actually attached to a binding (i.e. are used free), the correct scope is the same as the scope of a non-special binding of that name surrounding the form to which the binding is attached; the language about LOCALLY quoted above is out of date.
Gurg. Your example above is misindented. If you meant
(defun f (x) (declare (type float x)) x ;reference 1 (let ((x 'a)) x ;reference 2 ...)
then the scope of the declaration does not include (let ((x 'a)) x ...) because type declarations are not "pervasive". Thus what you wrote on p.223 of CLtL is wrong.
But if you meant
(defun f (x) (declare (type float x)) x ;reference 1 (let ((x 'a)) ...) x ;reference 2 ...)
then the example does not address the question since reference 1 and reference 2 are clearly both in the scope of the type declaration.
Now that you have pointed it out, I agree that a more likely and more desirable interpretation is that the references are considered to be to different variables. Question: what if x had been proclaimed SPECIAL? Then what would the interpretation be?
A special declaration should not affect the scope of a type declaration. Now, the version of DECLARATION-SCOPING that actually passed does not actually say that. But I don't think any other position is arguable.