| blob | 296a455628092e1f58c9231e14e5e6a42b6ee486 |
1 ;;;; This software is part of the SBCL system. See the README file for
2 ;;;; more information.
4 ;;;; This software is derived from software originally released by Xerox
5 ;;;; Corporation. Copyright and release statements follow. Later modifications
6 ;;;; to the software are in the public domain and are provided with
7 ;;;; absolutely no warranty. See the COPYING and CREDITS files for more
8 ;;;; information.
10 ;;;; copyright information from original PCL sources:
11 ;;;;
12 ;;;; Copyright (c) 1985, 1986, 1987, 1988, 1989, 1990 Xerox Corporation.
13 ;;;; All rights reserved.
14 ;;;;
15 ;;;; Use and copying of this software and preparation of derivative works based
16 ;;;; upon this software are permitted. Any distribution of this software or
17 ;;;; derivative works must comply with all applicable United States export
18 ;;;; control laws.
19 ;;;;
20 ;;;; This software is made available AS IS, and Xerox Corporation makes no
21 ;;;; warranty about the software, its performance or its conformity to any
22 ;;;; specification.
25 \f
26 #|
28 The CommonLoops evaluator is meta-circular.
30 Most of the code in PCL is methods on generic functions, including
31 most of the code that actually implements generic functions and method
32 lookup.
34 So, we have a classic bootstrapping problem. The solution to this is
35 to first get a cheap implementation of generic functions running,
36 these are called early generic functions. These early generic
37 functions and the corresponding early methods and early method lookup
38 are used to get enough of the system running that it is possible to
39 create real generic functions and methods and implement real method
41 !FIX-EARLY-GENERIC-FUNCTIONS is called to convert all the early generic
42 functions to real generic functions.
44 The cheap generic functions are built using the same
45 FUNCALLABLE-INSTANCE objects that real generic functions are made out of.
46 This means that as PCL is being bootstrapped, the cheap generic
47 function objects which are being created are the same objects which
48 will later be real generic functions. This is good because:
50 - we can keep pointers to the cheap generic function objects
51 during booting because those pointers will still point to
52 the right object after the generic functions are all fixed up.
54 This file defines the DEFMETHOD macro and the mechanism used to expand
55 it. This includes the mechanism for processing the body of a method.
56 DEFMETHOD basically expands into a call to LOAD-DEFMETHOD, which
57 basically calls ADD-METHOD to add the method to the generic function.
58 These expansions can be loaded either during bootstrapping or when PCL
59 is fully up and running.
61 An important effect of this arrangement is it means we can compile
62 files with DEFMETHOD forms in them in a completely running PCL, but
63 then load those files back in during bootstrapping. This makes
64 development easier. It also means there is only one set of code for
65 processing DEFMETHOD. Bootstrapping works by being sure to have
66 LOAD-METHOD be careful to call only primitives which work during
67 bootstrapping.
69 |#
72 ensure-generic-function-using-class
73 add-method remove-method))
79 ;;; For each of the early functions, arrange to have it point to its
80 ;;; early definition. Do this in a way that makes sure that if we
81 ;;; redefine one of the early definitions the redefinition will take
82 ;;; effect. This makes development easier.
89 name))))
91 ;;; *!GENERIC-FUNCTION-FIXUPS* is used by !FIX-EARLY-GENERIC-FUNCTIONS
92 ;;; to convert the few functions in the bootstrap which are supposed
93 ;;; to be generic functions but can't be early on.
94 ;;;
95 ;;; each entry is a list of the form
96 ;;;
97 ;;; (GENERIC-FUNCTION-NAME METHOD-COMBINATION-NAME METHODS)
98 ;;;
99 ;;; where methods is a list of lists of the form
100 ;;;
101 ;;; (LAMBDA-LIST SPECIALIZERS QUALIFIERS METHOD-BODY-FUNCTION-NAME)
102 ;;;
103 ;;;,where SPECIALIZERS is a list of class names.
106 standard
109 ()
110 real-add-method))
113 standard
116 ()
117 real-remove-method))
120 standard
123 ()
124 real-get-method))
127 standard
129 &key generic-function-class environment
132 ()
133 real-ensure-gf-using-class--generic-function)
135 &key generic-function-class environment
138 ()
139 real-ensure-gf-using-class--null))
142 standard
145 ()
146 real-make-method-lambda))
149 standard
151 lambda-expression environment)
153 ()
154 real-make-method-lambda-using-specializers))
157 standard
160 ()
161 real-make-method-specializers-form))
164 or
183 standard
186 ()
190 ()
194 ()
198 ()
202 ()
206 ()
210 ()
214 ()
218 standard
221 ()
222 real-parse-specializer-using-class))
225 standard
228 ()
229 real-unparse-specializer-using-class))
232 standard
234 lambda-expression
235 lambda-list environment)
237 ()
238 real-make-method-initargs-form))
241 standard
244 ()
245 standard-compute-effective-method)
248 ()
249 short-compute-effective-method))))
250 \f
280 DEFGENERIC: ~S~:@>"
283 ;; FIXME: this list is slightly weird.
284 ;; ANSI (on the DEFGENERIC page) in one
285 ;; place allows only OPTIMIZE; in
286 ;; another place gives this list of
287 ;; disallowed declaration specifiers.
288 ;; This seems to be the only place where
289 ;; the FUNCTION declaration is
290 ;; mentioned; TYPE seems to be missing.
291 ;; Very strange. -- CSR, 2002-10-21
293 inline notinline special))
296 is not allowed inside DEFGENERIC."
302 spec))))
303 (:method-combination
309 symbol: ~S"
313 (:argument-precedence-order
319 :ARGUMENT-PRECEDENCE-ORDER clause."
324 :ARGUMENT-PRECEDENCE-ORDER clause: ~
325 ~S and ~S"
335 (:method
338 ;; ANSI requires that unsupported things must get a
339 ;; PROGRAM-ERROR.
361 ;; Hmm. This is similar to BECOME-DEFINED-FUN-NAME
362 ;; except that it doesn't clear an :ASSUMED-TYPE. Should it?
370 :name fun-name
378 fun-name
379 :lambda-list lambda-list
380 :definition-source source-location
381 initargs))
385 ()
390 ;; PARSE-LAMBDA-LIST validates the skeleton,
391 ;; so just check for incorrect use of defaults.
392 ;; This works for both &OPTIONAL and &KEY.
397 :format-control
398 "~@<invalid ~A argument specifier ~S ~_in the ~
399 generic function lambda list ~S~:>"
403 lambda-list
409 ;; no defaults or supplied-p vars allowed for &OPTIONAL or &KEY
416 ~/sb-impl:print-symbol-with-prefix/, ~S, is not a lambda ~
417 form.~@:>"
418 context method-lambda))
419 method-lambda)
420 \f
423 ;;; As per CLHS -
424 ;;; "defmethod is not required to perform any compile-time side effects."
425 ;;; and we don't do much other than to make the function be defined,
426 ;;; which means that checking of callers' arglists can only occur after called
427 ;;; methods are actually loaded.
433 ;; :compile-toplevel is needed for subsequent forms
434 ;; :execute is needed for references to itself inside the body
436 ;; KLUDGE: this double expansion is quite a monumental
437 ;; workaround: it comes about because of a fantastic interaction
438 ;; between the processing rules of CLHS 3.2.3.1 and the
439 ;; bizarreness of MAKE-METHOD-LAMBDA.
440 ;;
441 ;; MAKE-METHOD-LAMBDA can be called by the user, and if the
442 ;; lambda itself doesn't refer to outside bindings the return
443 ;; value must be compileable in the null lexical environment.
444 ;; However, the function must also refer somehow to the
445 ;; associated method object, so that it can call NO-NEXT-METHOD
446 ;; with the appropriate arguments if there is no next method --
447 ;; but when the function is generated, the method object doesn't
448 ;; exist yet.
449 ;;
450 ;; In order to resolve this issue, we insert a literal cons cell
451 ;; into the body of the method lambda, return the same cons cell
452 ;; as part of the second (initargs) return value of
453 ;; MAKE-METHOD-LAMBDA, and a method on INITIALIZE-INSTANCE fills
454 ;; in the cell when the method is created. However, this
455 ;; strategy depends on having a fresh cons cell for every method
456 ;; lambda, which (without the workaround below) is skewered by
457 ;; the processing in CLHS 3.2.3.1, which permits implementations
458 ;; to macroexpand the bodies of EVAL-WHEN forms with both
459 ;; :COMPILE-TOPLEVEL and :LOAD-TOPLEVEL only once. The
460 ;; expansion below forces the double expansion in those cases,
461 ;; while expanding only once in the common case.
472 lambda-list body env)))
488 ;;; Take a name which is either a generic function name or a list specifying
489 ;;; a SETF generic function (like: (SETF <generic-function-name>)). Return
490 ;;; the prototype instance of the method-class for that generic function.
491 ;;;
492 ;;; If there is no generic function by that name, this returns the
493 ;;; default value, the prototype instance of the class
494 ;;; STANDARD-METHOD. This default value is also returned if the spec
495 ;;; names an ordinary function or even a macro. In effect, this leaves
496 ;;; the signalling of the appropriate error until load time.
497 ;;;
498 ;;; Note: During bootstrapping, this function is allowed to return NIL.
505 ; error at load time.
510 \f
511 ;;; These are used to communicate the method name and lambda-list to
512 ;;; MAKE-METHOD-LAMBDA-INTERNAL.
517 body env)
519 ;; don't care to figure out what parts of PCL would have
520 ;; to change to accept that, so coerce.
528 proto-gf proto-method qualifiers specializers method-lambda env))
531 proto-gf proto-method method-function-lambda initargs env))
534 proto-gf proto-method specializers env)))
539 specializers lambda-list)
540 ;; Note: We could DECLAIM the ftype of the generic function here,
541 ;; since ANSI specifies that we create it if it does not
542 ;; exist. However, I chose not to, because I think it's more
543 ;; useful to support a style of programming where every generic
544 ;; function has an explicit DEFGENERIC and any typos in DEFMETHODs
545 ;; are warned about. Otherwise
546 ;;
547 ;; (DEFGENERIC FOO-BAR-BLETCH (X))
548 ;; (DEFMETHOD FOO-BAR-BLETCH ((X HASH-TABLE)) ..)
549 ;; (DEFMETHOD FOO-BRA-BLETCH ((X SIMPLE-VECTOR)) ..)
550 ;; (DEFMETHOD FOO-BAR-BLETCH ((X VECTOR)) ..)
551 ;; (DEFMETHOD FOO-BAR-BLETCH ((X ARRAY)) ..)
552 ;; (DEFMETHOD FOO-BAR-BLETCH ((X LIST)) ..)
553 ;;
554 ;; compiles without raising an error and runs without raising an
555 ;; error (since SIMPLE-VECTOR cases fall through to VECTOR) but
556 ;; still doesn't do what was intended. I hate that kind of bug
557 ;; (code which silently gives the wrong answer), so we don't do a
558 ;; DECLAIM here. -- WHN 20000229
564 initargs-form)))
571 method-class-name initargs-form)
574 fn-lambda)
587 specializers)
598 ;; CONSTANT-FORM-VALUE? What I
599 ;; kind of want to know, though,
600 ;; is what happens if we don't do
601 ;; this for some slow-method
602 ;; function because of a hairy
603 ;; lexenv -- is the only bad
604 ;; effect that the method
605 ;; function ends up unnamed? If
606 ;; so, couldn't we arrange to
607 ;; name it later?
609 specl))
610 specializers))
618 name qualifiers `',specls
619 unspecialized-lambda-list method-class-name
621 #',mname
624 name qualifiers
625 specializers
626 #+nil
632 specializers))
633 unspecialized-lambda-list
634 method-class-name
635 initargs-form))))
639 method-class-name initargs-form)
641 ',method-class-name
642 ',name
643 ',qualifiers
644 ,specializers-form
645 ',unspecialized-lambda-list
646 ,initargs-form
655 proto-gf proto-method method-function-lambda initargs env)))
658 method-lambda initargs env)
670 (check-method-lambda method-lambda 'make-method-lambda) ; TODO remove check in make-method-lambda
671 ;; Default behavior: delegate to MAKE-METHOD-LAMBDA.
683 ;;; When bootstrapping PCL MAKE-METHOD-LAMBDA starts out as a regular
684 ;;; function: REAL-MAKE-METHOD-LAMBDA set to the fdefinition of
685 ;;; MAKE-METHOD-LAMBDA. Once generic functions are born,
686 ;;; REAL-MAKE-METHOD-LAMBDA is used to implement the default method.
687 ;;; MAKE-METHOD-LAMBDA-INTERNAL is split out into a separate function
688 ;;; so that changing it in a live image is easy, and changes actually
689 ;;; take effect.
703 ;;; A helper function for creating Python-friendly type declarations
704 ;;; in DEFMETHOD forms.
705 ;;;
706 ;;; This function operates on
707 ;;; * non-parsed specializers, i.e. class names and extended
708 ;;; specializer syntaxes
709 ;;; * parsed specializers, i.e. CLASSes, EQL-SPECIALIZERs,
710 ;;; CLASS-EQ-SPECIALIZERs and generic SPECIALIZERs
711 ;;;
712 ;;; We're too lazy to cons up a new environment for this, so we just
713 ;;; pass in the list of locally declared specials in addition to the
714 ;;; old environment.
724 ;; KLUDGE: PCL, in its wisdom, sometimes calls methods with
725 ;; types which don't match their specializers. (Specifically,
726 ;; it calls ENSURE-CLASS-USING-CLASS (T NULL) with a non-NULL
727 ;; second argument.) Hopefully it only does this kind of
728 ;; weirdness when bootstrapping.. -- WHN 20000610
731 ;; Independent of SPECIALIZER, bail out if the PARAMETER is
732 ;; known to be a special variable. Our rebinding magic for SETQ
733 ;; cases doesn't work right there as SET, (SETF SYMBOL-VALUE),
734 ;; etc. make things undecidable.
738 ;; Bail out on SLOT-OBJECT special case.
739 ;;
740 ;; KLUDGE: For some low-level implementation classes, perhaps
741 ;; because of some problems related to the incomplete
742 ;; integration of PCL into SBCL's type system, some specializer
743 ;; classes can't be declared as argument types. E.g.
744 ;; (DEFMETHOD FOO ((X SLOT-OBJECT))
745 ;; (DECLARE (TYPE SLOT-OBJECT X))
746 ;; ..)
747 ;; loses when
748 ;; (DEFSTRUCT BAR A B)
749 ;; (FOO (MAKE-BAR))
750 ;; perhaps because of the way that STRUCTURE-OBJECT inherits
751 ;; both from SLOT-OBJECT and from SB-KERNEL:INSTANCE. In an
752 ;; effort to sweep such problems under the rug, we exclude these
753 ;; problem cases by blacklisting them here. -- WHN 2001-01-19
757 ;; Bail out on unparsed EQL-specializers.
758 ;;
759 ;; KLUDGE: ANSI, in its wisdom, says that EQL-SPECIALIZER-FORMs
760 ;; in EQL specializers are evaluated at DEFMETHOD expansion
761 ;; time. Thus, although one might think that in
762 ;; (DEFMETHOD FOO ((X PACKAGE)
763 ;; (Y (EQL 12))
764 ;; ..))
765 ;; the PACKAGE and (EQL 12) forms are both parallel type names,
766 ;; they're not, as is made clear when you do
767 ;; (DEFMETHOD FOO ((X PACKAGE)
768 ;; (Y (EQL 'BAR)))
769 ;; ..)
770 ;; where Y needs to be a symbol named "BAR", not some cons made
771 ;; by (CONS 'QUOTE 'BAR). I.e. when the EQL-SPECIALIZER-FORM is
772 ;; (EQL 'X), it requires an argument to be of type (EQL X). It'd
773 ;; be easy to transform one to the other, but it'd be somewhat
774 ;; messier to do so while ensuring that the EQL-SPECIALIZER-FORM
775 ;; is only EVAL'd once. (The new code wouldn't be messy, but
776 ;; it'd require a big transformation of the old code.) So
777 ;; instead we punt. -- WHN 20000610
781 ;; Parsed specializer objects, i.e. CLASS, EQL-SPECIALIZER,
782 ;; CLASS-EQ-SPECIALIZER and generic SPECIALIZER.
783 ;;
784 ;; Also unparsed specializers other than EQL: these have to be
785 ;; either class names or extended specializers.
786 ;;
787 ;; For these, we can usually make Python very happy.
788 ;;
789 ;; KLUDGE: Since INFO doesn't work right for class objects here,
790 ;; and they are valid specializers, see if the specializer is
791 ;; a named class, and use the name in that case -- otherwise
792 ;; the class instance is ok, since info will just return NIL, NIL.
793 ;;
794 ;; We still need to deal with the class case too, but at
795 ;; least #.(find-class 'integer) and integer as equivalent
796 ;; specializers with this.
799 proto-generic-function proto-method specializer))))))
806 ;; We have the %METHOD-NAME declaration in the place
807 ;; where we expect it only if there is are no
808 ;; non-standard prior MAKE-METHOD-LAMBDA methods -- or
809 ;; unless they're fantastically unintrusive.
812 ;; Macroexpansion caused by code-walking may call
813 ;; make-method-lambda and end up with wrong values
817 ;; the method-cell is a way of communicating what method
818 ;; a method-function implements, for the purpose of
819 ;; NO-NEXT-METHOD. We need something that can be shared
820 ;; between function and initargs, but not something that
821 ;; will be coalesced as a constant (because we are
822 ;; naughty, oh yes) with the expansion of any other
823 ;; methods in the same file. -- CSR, 2007-05-30
835 ;; These declarations seem to be used by PCL to pass
836 ;; information to itself; when I tried to delete 'em
837 ;; ca. 0.6.10 it didn't work. I'm not sure how they
838 ;; work, but note the (VAR-DECLARATION '%CLASS ..)
839 ;; expression in CAN-OPTIMIZE-ACCESS1. -- WHN
840 ;; 2000-12-30
844 parameters specializers)
845 ;; These TYPE declarations weren't in the original PCL
846 ;; code, but the Python compiler likes them a
847 ;; lot. (We're telling the compiler about our
848 ;; knowledge of specialized argument types so that it
849 ;; can avoid run-time type dispatch overhead, which
850 ;; can be a huge win for Python.)
854 proto-gf proto-method par spec specials env))
855 parameters specializers))))
857 ;; Remove the documentation string and insert the
858 ;; appropriate class declarations. The documentation
859 ;; string is removed to make it easy for us to insert
860 ;; new declarations later, they will just go after the
861 ;; CADR of the method lambda. The class declarations
862 ;; are inserted to communicate the class of the method's
863 ;; arguments to the code walk.
865 ;; The default ignorability of method parameters
866 ;; doesn't seem to be specified by ANSI. PCL had
867 ;; them basically ignorable but was a little
868 ;; inconsistent. E.g. even though the two
869 ;; method definitions
870 ;; (DEFMETHOD FOO ((X T) (Y T)) "Z")
871 ;; (DEFMETHOD FOO ((X T) Y) "Z")
872 ;; are otherwise equivalent, PCL treated Y as
873 ;; ignorable in the first definition but not in the
874 ;; second definition. We make all required
875 ;; parameters ignorable as a way of systematizing
876 ;; the old PCL behavior. -- WHN 2000-11-24
878 ,class-declarations
879 ,@declarations
898 method-lambda required-parameters env slots))
907 walked-lambda-body
909 ,slot-name-lists
919 aux))))
922 (,lambda-list .method-args. .next-methods.
923 :call-next-method-p
929 ,@walked-declarations
931 %parameter-binding-modified))
935 %parameter-binding-modified))
947 proto-generic-function proto-method name environment)))
958 :format-control
959 "~@<~S is not a valid parameter specializer name.~@:>"
968 specializer-name))
978 ;; In case of unknown specializer or known specializer with syntax
979 ;; error, TYPECASE may fall through to default method with error
980 ;; signaling.
997 `(,type
998 (,function-name
999 proto-generic-function proto-method specializer-name environment))))
1001 ; ^ apparently not needed during bootstrapping
1002 symbol cons t)))))
1012 ;; TODO later protocol-unimplemented-error?
1015 nil)
1019 "~@<Cannot parse specializer ~S in ~S~@[: ~A~].~@:>"
1023 :format-control
1024 "~@<Cannot find type for specializer ~
1025 ~/sb-ext:print-symbol-with-prefix/ when executing ~S ~
1026 for a ~/sb-ext:print-type-specifier/ of a ~
1027 ~/sb-ext:print-type-specifier/.~@:>"
1028 :format-arguments
1035 (:primitive
1036 name)
1037 (:defined
1038 ;; This can happen if NAME is a DEFTYPE.
1040 name proto-generic-function proto-method))
1042 ;; CLOS classes are too expensive to check (as opposed
1043 ;; to STRUCTURE-CLASS and SYSTEM-CLASS).
1044 nil)
1046 ;; TODO proper warning condition?
1048 name proto-generic-function proto-method)
1049 nil)))))
1051 ;;; Non-parsed class specializers, i.e. class names
1052 ;;;
1053 ;;; Extended generic function classes with specializers which are
1054 ;;; designated by symbols have to install their own methods
1055 ;;; specialized on symbol to replace this logic.
1061 ;; Usually tries to find the class named
1062 ;; SPECIALIZER. Can do something different when there
1063 ;; is a non-default method on
1064 ;; PARSE-SPECIALIZER-USING-CLASS.
1066 proto-generic-function specializer)
1068 ;; SPECIALIZER does not name a class, but maybe it is
1069 ;; known to name a :forthcoming-defclass-type.
1070 ;; CLASS-NAME-TYPE-SPECIFIER will emit the warning and
1071 ;; return nil if not.
1073 specializer proto-generic-function proto-method))
1075 ;; This can only happen if there is an EQL-specialized
1076 ;; method on PARSE-SPECIALIZER-USING-CLASS matching
1077 ;; SPECIALIZER that signals an error.
1079 nil))))
1082 proto-generic-function proto-method specializer))))
1084 ;;; Non-parsed extended specializer with default syntax
1085 ;;; i.e. (SPECIALIZER-KIND &rest SPECIFIC-SYNTAX)
1092 proto-generic-function specializer)
1094 ;; This can happen, for example, if SPECIALIZER does
1095 ;; not designate any extended specializer or if it does
1096 ;; but then does not conform to the respective extended
1097 ;; specializer syntax.
1099 nil))))
1102 proto-generic-function proto-method specializer))))
1104 ;;; Parsed EQL and CLASS-EQ specializers
1116 ;;; Parsed class specializers
1131 ;; Make sure SPECIALIZER has a proper class name and that name
1132 ;; designates the class SPECIALIZER in the global environment.
1136 name proto-generic-function proto-method)))))
1148 `(,type
1149 (,function-name
1150 proto-generic-function proto-method specializer))))
1152 structure-class system-class class)))))
1162 result
1164 specializer generic-function))))
1172 ;; FIXME: this HANDLER-CASE is a bit of a hammer to crack a nut:
1173 ;; the idea is that we want to unparse permissively, so that the
1174 ;; lazy (or rather the "portable") specializer extender (who
1175 ;; does not define methods on these new SBCL-specific MOP
1176 ;; functions) can still subclass specializer and define methods
1177 ;; without everything going wrong. Making it cleaner and
1178 ;; clearer that that is what we are defending against would be
1179 ;; nice. -- CSR, 2007-06-01
1186 class-name
1187 type))
1188 type))
1191 specializer generic-function)))
1197 ;;; For passing a list (groveled by the walker) of the required
1198 ;;; parameters whose bindings are modified in the method body to the
1199 ;;; optimized-slot-value* macros.
1203 method-args
1204 next-methods
1215 next-method-call
1216 args
1217 rest-arg
1226 parameters-setqd applyp method-cell))
1227 &body body
1231 ;; always provide the lexical function NEXT-METHOD-P.
1232 ;; I would think this to be a good candidate for declaring INLINE
1233 ;; but that's not the way it was done before.
1244 `((%check-cnm-args cnm-args
1245 ,method-args
1247 nil)
1255 ',method-cell
1264 ;; Can't easily provide a RETRY restart here, as the return value here is
1265 ;; for the method, not the generic function.
1267 method args)))
1285 call-method-args)
1287 value)
1308 pv
1309 next-method-call
1310 arg-info)
1313 value)
1317 ;; The two variants of INVOKE-FAST-METHOD-CALL differ in how REST-ARGs
1318 ;; are handled. The first one will get REST-ARG as a single list (as
1319 ;; the last argument), and will thus need to use APPLY. The second one
1320 ;; will get them as a &MORE argument, so we can pass the arguments
1321 ;; directly with MULTIPLE-VALUE-CALL and %MORE-ARG-VALUES.
1330 more-context
1331 more-count
1337 ,@required-args
1340 ;; The cases with only small amounts of required arguments passed
1341 ;; are probably very common, and special-casing speeds them up by
1342 ;; a factor of 2 with very little effect on the other
1343 ;; cases. Though it'd be nice to have the generic case be equally
1344 ;; fast.
1351 ,@required-args
1362 \f
1363 ;;;; effective method functions
1369 ;;; This function was in the CMU CL version of PCL (ca Debian 2.4.8)
1370 ;;; without explanation. It appears to be intended for debugging, so
1371 ;;; it might be useful someday, so I haven't deleted it.
1372 ;;; But it isn't documented and isn't used for anything now, so
1373 ;;; I've conditionalized it out of the base system. -- WHN 19991213
1374 #+sb-show
1407 ,@more-arg
1410 ,restp
1411 ,@required-args
1416 ;; "What," you may wonder, "do these next two clauses do?" In that
1417 ;; case, you are not a PCL implementor, for they considered this to
1418 ;; be self-documenting.:-| Or CSR, for that matter, since he can
1419 ;; also figure it out by looking at it without breaking stride. For
1420 ;; the rest of us, though: From what the code is doing with .SLOTS.
1421 ;; and whatnot, evidently it's implementing SLOT-VALUEish and
1422 ;; GET-SLOT-VALUEish things. Then we can reason backwards and
1423 ;; conclude that setting EMF to a FIXNUM is an optimized way to
1424 ;; represent these slot access operations.
1435 value)))))
1443 ;; (In cmucl-2.4.8 there was a commented-out third ,@(WHEN
1444 ;; ...) clause here to handle SLOT-BOUNDish stuff. Since
1445 ;; there was no explanation and presumably the code is 10+
1446 ;; years stale, I simply deleted it. -- WHN)
1447 )))
1449 ;;; Before SBCL 0.9.16.7 instead of
1450 ;;; INVOKE-NARROW-EFFECTIVE-METHOD-FUNCTION we passed a (THE (OR
1451 ;;; FUNCTION METHOD-CALL FAST-METHOD-CALL) EMF) form as the EMF. Now,
1452 ;;; to make less work for the compiler we take a path that doesn't
1453 ;;; involve the slot-accessor clause (where EMF is a FIXNUM) at all.
1468 ,@more-arg
1471 ,restp
1472 ,@required-args
1476 emf-n restp required-args))
1499 args)
1522 args))))))
1525 args
1537 value)))
1554 \f
1557 method-cell
1558 cnm-args)
1561 ,next-method-call
1573 ,@args
1579 call-next-method-p
1580 setq-p
1581 parameters-setqd
1582 method-cell
1583 applyp))
1584 &body body
1604 nil)
1606 ,next-method-call
1608 ,method-cell
1609 cnm-args))))
1615 ;;; CMUCL comment (Gerd Moellmann):
1616 ;;;
1617 ;;; The standard says it's an error if CALL-NEXT-METHOD is called with
1618 ;;; arguments, and the set of methods applicable to those arguments is
1619 ;;; different from the set of methods applicable to the original
1620 ;;; method arguments. (According to Barry Margolin, this rule was
1621 ;;; probably added to ensure that before and around methods are always
1622 ;;; run before primary methods.)
1623 ;;;
1624 ;;; This could be optimized for the case that the generic function
1625 ;;; doesn't have hairy methods, does have standard method combination,
1626 ;;; is a standard generic function, there are no methods defined on it
1627 ;;; for COMPUTE-APPLICABLE-METHODS and probably a lot more of such
1628 ;;; preconditions. That looks hairy and is probably not worth it,
1629 ;;; because this check will never be fast.
1631 ;; 1. Check for no arguments.
1636 ;; 2. Requirement arguments pairwise: if all are EQL, the applicable
1637 ;; methods must be the same. This takes care of the relatively common
1638 ;; case of twiddling with &KEY arguments without being horribly
1639 ;; expensive.
1646 ;; 3. Only then do the full check.
1651 ~{~S~^, ~} to call-next-method is different from ~
1652 the set of methods ~S applicable to the original ~
1653 method argument~P ~{~S~^, ~}.~@:>"
1657 ;; FIXME: replacing this entire mess with DESTRUCTURING-BIND would correct
1658 ;; problems similar to those already solved by a correct implementation
1659 ;; of DESTRUCTURING-BIND, such as incorrect binding order:
1660 ;; e.g. (macroexpand-1 '(bind-args ((&optional (x nil xsp)) args) (form)))
1661 ;; -> (LET* ((.ARGS-TAIL. ARGS) (XSP (NOT (NULL .ARGS-TAIL.))) (X ...)))
1662 ;; It's mostly irrelevant unless a method uses CALL-NEXT-METHOD though.
1678 "encountered the non-standard lambda list keyword ~S"
1679 var)))
1680 nil)
1728 ,@bindings
1729 (.dummy0.
1741 ;; FIXME: Do we want to export this symbol? Or maybe use an
1742 ;; (ERROR 'SIMPLE-PROGRAM-ERROR) form?
1745 return tail))
1749 ;; method definition
1751 ;; a list of all required parameters whose bindings might be
1752 ;; modified in the method body.
1759 ;; hierarchy: nil -> :simple -> T.
1762 form)
1764 ;; The walker will split (SETQ A 1 B 2) to
1765 ;; separate (SETQ A 1) and (SETQ B 2) forms, so we
1766 ;; only need to handle the simple case of SETQ
1767 ;; here.
1772 ;; Note that we don't need to check for
1773 ;; %VARIABLE-REBINDING declarations like is
1774 ;; done in CAN-OPTIMIZE-ACCESS1, since the
1775 ;; bindings that will have that declation will
1776 ;; never be SETQd.
1778 ;; If a parameter binding is shadowed by
1779 ;; another binding it won't have a %CLASS
1780 ;; declaration anymore, and this won't get
1781 ;; executed.
1783 form)
1787 form)
1795 form))
1799 ;;; FIXME: the walker's rewriting of the source code causes
1800 ;;; trouble when doing code coverage. The rewrites should be
1801 ;;; removed, and the same operations done using
1802 ;;; compiler-macros or tranforms.
1804 walked-lambda
1805 method-lambda)
1806 call-next-method-p
1808 parameters-setqd)))))
1816 \f
1826 new-value)
1828 \f
1834 #+nil
1838 ll initargs source-location)))
1842 initargs source-location)
1851 :name gf-spec
1852 :new-location source-location
1853 :old-method method
1856 gf-spec qualifiers specializers lambda-list
1857 :definition-source source-location
1858 initargs)))
1861 ;; FIXME: should be STYLE-WARNING?
1863 "~&At the time the method with qualifiers ~:S and~%~
1864 specializers ~:S on the generic function ~S~%~
1865 was compiled, the method-class for that generic function was~%~
1866 ~S. But, the method class is now ~S, this~%~
1867 may mean that this method was compiled improperly.~%"
1868 qualifiers specializers gf-spec
1870 method))
1898 \f
1901 ;; We say "&MUMBLE is not allowed in a generic function lambda list"
1902 ;; whether this is called by DEFMETHOD or DEFGENERIC.
1903 ;; [It is used for either. Why else recognize and silently ignore &AUX?]
1907 :silent t
1913 ;; FIXME: this does more than return an FTYPE from a lambda list -
1914 ;; it unions the type with an existing ctype object. It needs a better name,
1915 ;; and to be reimplemented as "union and call sb-c::ftype-from-lambda-list".
1926 old-ftype))))
1941 keywords)
1944 *))))
1945 \f
1946 ;;;; early generic function support
1950 ;; CLHS doesn't specify &allow-other-keys here but I guess the supposition
1951 ;; is that they'll be checked by ENSURE-GENERIC-FUNCTION-USING-CLASS.
1952 ;; Except we don't do that either, so I think the blame, if any, lies there
1953 ;; for not catching errant keywords.
1965 existing fun-name all-keys)))))
1969 'simple-program-error
1971 already names an ordinary function or a ~
1972 macro.~@:>"
2022 arg-info-precedence
2023 arg-info-metatypes
2024 arg-info-number-optional
2025 arg-info-key/rest-p
2026 arg-info-keys ;nil no &KEY or &REST allowed
2027 ;(k1 k2 ..) Each method must accept these &KEY arguments.
2028 ;T must have &KEY or &REST
2030 gf-info-simple-accessor-type ; nil, reader, writer, boundp
2033 gf-info-static-c-a-m-emf
2035 gf-info-fast-mf-p)
2053 ;;; Create a gf lambda list from a method lambda list
2066 argument-precedence-order)
2091 existing methods of ~S.~%~
2092 Old lambda-list ~s"
2099 lambda-list
2115 arg-info))
2125 to the generic function~2I~_~S;~I~_~
2126 but ~?~:>"
2136 "the method has ~A required arguments than the generic function."
2140 "the method has ~A optional arguments than the generic function."
2144 "the method and generic function differ in whether they accept~_~
2151 ~S."
2152 gf-keywords)))))))
2161 ;;; FIXME: we don't actually need this; we could test for the exact
2162 ;;; class and deal with it as appropriate. In fact we probably don't
2163 ;;; need it anyway because we only use this for METHOD-SPECIALIZERS on
2164 ;;; the standard reader method for METHOD-SPECIALIZERS. Probably.
2176 standard-writer-method standard-boundp-method
2177 global-reader-method global-writer-method
2178 global-boundp-method))
2208 nil)))
2256 nil)
2262 ;; FIXME: this test will eventually be
2263 ;; superseded by the *internal-pcl...* test,
2264 ;; above. While we are in a process of
2265 ;; transition, however, it should probably
2266 ;; remain.
2272 #'make-method-lambda
2283 arg-info)
2285 ;;; This is the early definition of ENSURE-GENERIC-FUNCTION-USING-CLASS.
2286 ;;;
2287 ;;; The STATIC-SLOTS field of the funcallable instances used as early
2288 ;;; generic functions is used to store the early methods and early
2289 ;;; discriminator code for the early generic function. The static
2290 ;;; slots field of the fins contains a list whose:
2291 ;;; CAR - a list of the early methods on this early gf
2292 ;;; CADR - the early discriminator code for this method
2295 lambda-list-p)
2296 argument-precedence-order
2297 definition-source
2298 documentation
2304 existing)
2308 argument-precedence-order definition-source
2309 documentation)
2317 argument-precedence-order definition-source
2318 documentation))))
2321 function argument-precedence-order source-location
2322 documentation)
2342 :lambda-list lambda-list
2345 fin))
2355 new-value)
2361 dfun)))
2364 ;; Check that we are under the lock.
2365 #+sb-thread
2370 new-state))))
2371 dfun)
2406 nil)
2416 ;; Use the existing lambda-list, if any. It is reasonable to do eg.
2417 ;;
2418 ;; (if (fboundp name)
2419 ;; (ensure-generic-function name)
2420 ;; (ensure-generic-function name :lambda-list '(foo)))
2421 ;;
2422 ;; in which case we end up here with no lambda-list in the first leg.
2426 t)
2435 ;; FIXME: Ideally we would like to not clobber it, but because generic
2436 ;; functions assert their FTYPEs callers believing the FTYPE are left with
2437 ;; unsafe assumptions. Hence the clobbering. Be quiet when the new type
2438 ;; is a subtype of the old one, though -- even though the type is not
2439 ;; trusted anymore, the warning is still not quite as interesting.
2443 for the same name with ~S.~:@>"
2444 fun-name 'ftype
2449 fun-name))
2455 class-or-name)
2458 symbol that names a class.~@:>"
2459 context class-or-name))))
2462 class-or-name environment)))
2464 ;; FIXME: reference MOP documentation -- this is an
2465 ;; additional requirement on our users
2467 finalizeable~@:>"
2468 class)
2471 environment
2476 generic-function-class environment)))
2481 (:method-combination
2483 key
2490 value))))
2491 (:method-class
2493 value environment)))
2520 \f
2527 ;;; FIXME: this function took on a slightly greater role than it
2528 ;;; previously had around 2005-11-02, when CSR fixed the bug whereby
2529 ;;; having more than one subclass of standard-generic-function caused
2530 ;;; the whole system to die horribly through a metacircle in
2531 ;;; GF-ARG-INFO. The fix is to be slightly more disciplined about
2532 ;;; calling accessor methods -- we call GET-GENERIC-FUN-INFO when
2533 ;;; computing discriminating functions, so we need to be careful about
2534 ;;; having a base case for the recursion, and we provide that with the
2535 ;;; STANDARD-GENERIC-FUNCTION case below. However, we are not (yet)
2536 ;;; as disciplined as CLISP's CLOS/MOP, and it would be nice to get to
2537 ;;; that stage, where all potentially dangerous cases are enumerated
2538 ;;; and stopped. -- CSR, 2005-11-02.
2540 ;; values nreq applyp metatypes nkeys arg-info
2547 metatypes
2548 arg-info))
2557 nkeys
2558 arg-info))))
2569 &key slot-name object-class method-class-function
2570 definition-source)
2573 ;; Figure out whether we got class objects or class names as the
2574 ;; specializers and set parsed and unparsed appropriately. If we
2575 ;; got class objects, then we can compute unparsed, but if we
2576 ;; got class names we don't try to compute parsed.
2582 specializers))
2593 ;; the parsed specializers. This is used by
2594 ;; EARLY-METHOD-SPECIALIZERS to cache the parse.
2595 ;; Note that this only comes into play when there is
2596 ;; more than one early method on an early gf.
2597 parsed
2599 ;; A list to which REAL-MAKE-A-METHOD can be applied
2600 ;; to make a real method corresponding to this early
2601 ;; one.
2604 initargs doc)
2610 result))
2615 definition-source)
2626 :slot-definition slot-definition
2630 initargs)
2631 :definition-source definition-source
2632 initargs)))
2652 ;;; Fetch the specializers of an early method. This is basically just
2653 ;;; a simple accessor except that when the second argument is t, this
2654 ;;; converts the specializers from symbols into class objects. The
2655 ;;; class objects are cached in the early method, this makes
2656 ;;; bootstrapping faster because the class objects only have to be
2657 ;;; computed once.
2658 ;;;
2659 ;;; NOTE:
2660 ;;; The second argument should only be passed as T by
2661 ;;; early-lookup-method. This is to implement the rule that only when
2662 ;;; there is more than one early method on a generic function is the
2663 ;;; conversion from class names to class objects done. This
2664 ;;; corresponds to the fact that we are only allowed to have one
2665 ;;; method on any generic function up until the time classes exist.
2690 specializers arglist &rest initargs
2691 &key documentation definition-source
2694 ;; argument here because the default,
2695 ;; STANDARD-GENERIC-FUNCTION, is right for all early generic
2696 ;; functions. (See REAL-ADD-NAMED-METHOD)
2706 specializers initargs documentation
2711 ;;; This is the early version of ADD-METHOD. Later this will become a
2712 ;;; generic function. See !FIX-EARLY-GENERIC-FUNCTIONS which has
2713 ;;; special knowledge about ADD-METHOD.
2726 ;;; This is the early version of REMOVE-METHOD. See comments on
2727 ;;; the early version of ADD-METHOD.
2741 ;;; This is the early version of GET-METHOD. See comments on the early
2742 ;;; version of ADD-METHOD.
2748 specializers)
2750 specializers))
2755 nil))
2758 ;; minor KLUDGE: a separate code component for this function allows GCing
2759 ;; a few symbols and their associated code that would otherwise be retained:
2760 ;; *!EARLY-{GENERIC-}FUNCTIONS*, *!GENERIC-FUNCTION-FIXUPS*
2768 ;; Rearrange *!EARLY-GENERIC-FUNCTIONS* to speed up
2769 ;; FIX-EARLY-GENERIC-FUNCTIONS.
2776 generic-function-methods
2777 method-specializers
2778 specializer-type
2779 specializer-class
2780 slot-definition-location
2781 slot-definition-name
2782 class-slots
2783 gf-arg-info
2784 class-precedence-list
2785 slot-boundp-using-class
2787 slot-value-using-class)))
2798 early-method))))
2801 early-method t))
2816 ;; This is lifted from sb-introspect, OAOO and all that.
2835 map)
2849 'standard-method
2850 qualifiers lambda-list specializers initargs nil
2861 \f
2862 ;;; PARSE-DEFMETHOD is used by DEFMETHOD to parse the &REST argument
2863 ;;; into the 'real' arguments. This is where the syntax of DEFMETHOD
2864 ;;; is really implemented.
2886 \f
2890 ;; We don't need these, but according to the unit tests,
2891 ;; they're mandated by AMOP.
2897 ()
2900 ;; Return 3 values:
2901 ;; - the bound variables, without defaults, supplied-p vars, or &AUX vars.
2902 ;; - the lambda list without specializers.
2903 ;; - just the specializers
2907 arglist
2916 rest
2917 ;; Preserve keyword-names when given as (:KEYWORD var)
2923 \f
2925 \f
2926 ;;; FIXME: In here there was a #-CMU definition of SYMBOL-MACROLET
2927 ;;; which used %WALKER stuff. That suggests to me that maybe the code
2928 ;;; walker stuff was only used for implementing stuff like that; maybe
2929 ;;; it's not needed any more? Hunt down what it was used for and see.
2936 form)))
2945 ,in
2949 slot-entry
2953 slot-entry
2955 `(,var-name
2957 slots)
2967 ,in
2972 slots)