(declaim (inline wrapper-class*))
(defun wrapper-class* (wrapper)
(or (wrapper-class wrapper)
- (ensure-non-standard-class
- (classoid-name (layout-classoid wrapper)))))
+ (let ((classoid (layout-classoid wrapper)))
+ (ensure-non-standard-class
+ (classoid-name classoid)
+ classoid))))
;;; The wrapper cache machinery provides general mechanism for
;;; trapping on the next access to any instance of a given class. This
;; FIXME: We are here inside PCL lock, but might someone be
;; accessing the wrapper at the same time from outside the lock?
- ;; Can it matter that they get 0 from one slot and a valid value
- ;; from another?
- (dotimes (i layout-clos-hash-length)
- (setf (layout-clos-hash owrapper i) 0))
+ (setf (layout-clos-hash owrapper) 0)
;; FIXME: We could save a whopping cons by using (STATE . WRAPPER)
;; instead
(remhash owrapper *previous-nwrappers*)
(setf (gethash nwrapper *previous-nwrappers*) new-previous)))
+;;; FIXME: This is not a good name: part of the constract here is that
+;;; we return the valid wrapper, which is not obvious from the name
+;;; (or the names of our callees.)
(defun check-wrapper-validity (instance)
(let* ((owrapper (wrapper-of instance))
(state (layout-invalid owrapper)))
(aver (not (eq state :uninitialized)))
- (etypecase state
- (null owrapper)
- ;; FIXME: I can't help thinking that, while this does cure the
- ;; symptoms observed from some class redefinitions, this isn't
- ;; the place to be doing this flushing. Nevertheless... --
- ;; CSR, 2003-05-31
- ;;
- ;; CMUCL comment:
- ;; We assume in this case, that the :INVALID is from a
- ;; previous call to REGISTER-LAYOUT for a superclass of
- ;; INSTANCE's class. See also the comment above
- ;; FORCE-CACHE-FLUSHES. Paul Dietz has test cases for this.
- ((member t)
- (force-cache-flushes (class-of instance))
- (check-wrapper-validity instance))
- (cons
- (ecase (car state)
- (:flush
- (flush-cache-trap owrapper (cadr state) instance))
- (:obsolete
- (obsolete-instance-trap owrapper (cadr state) instance)))))))
+ (cond ((not state)
+ owrapper)
+ ((not (layout-for-std-class-p owrapper))
+ ;; Obsolete structure trap.
+ (obsolete-instance-trap owrapper nil instance))
+ ((eq t state)
+ ;; FIXME: I can't help thinking that, while this does cure
+ ;; the symptoms observed from some class redefinitions,
+ ;; this isn't the place to be doing this flushing.
+ ;; Nevertheless... -- CSR, 2003-05-31
+ ;;
+ ;; CMUCL comment:
+ ;; We assume in this case, that the :INVALID is from a
+ ;; previous call to REGISTER-LAYOUT for a superclass of
+ ;; INSTANCE's class. See also the comment above
+ ;; FORCE-CACHE-FLUSHES. Paul Dietz has test cases for this.
+ (force-cache-flushes (class-of instance))
+ (check-wrapper-validity instance))
+ ((consp state)
+ (ecase (car state)
+ (:flush
+ (flush-cache-trap owrapper (cadr state) instance))
+ (:obsolete
+ (obsolete-instance-trap owrapper (cadr state) instance))))
+ (t
+ (bug "Invalid LAYOUT-INVALID: ~S" state)))))
(declaim (inline check-obsolete-instance))
(defun check-obsolete-instance (instance)
(when (invalid-wrapper-p (layout-of instance))
(check-wrapper-validity instance)))
+
+(defun valid-wrapper-of (instance)
+ (let ((wrapper (wrapper-of instance)))
+ (if (invalid-wrapper-p wrapper)
+ (check-wrapper-validity instance)
+ wrapper)))
\f
;;; NIL: means nothing so far, no actual arg info has NILs in the
-;;; metatype
+;;; metatype.
;;;
;;; CLASS: seen all sorts of metaclasses (specifically, more than one
;;; of the next 5 values) or else have seen something which doesn't
-;;; fall into a single category (SLOT-INSTANCE, FORWARD).
+;;; fall into a single category (SLOT-INSTANCE, FORWARD). Also used
+;;; when seen a non-standard specializer.
+;;;
+;;; T: means everything so far is the class T.
+;;;
+;;; The above three are the really important ones, as they affect how
+;;; discriminating functions are computed. There are some other
+;;; possible metatypes:
+;;;
+;;; * STANDARD-INSTANCE: seen only standard classes
+;;; * BUILT-IN-INSTANCE: seen only built in classes
+;;; * STRUCTURE-INSTANCE: seen only structure classes
+;;; * CONDITION-INSTANCE: seen only condition classes
;;;
-;;; T: means everything so far is the class T
-;;; STANDARD-INSTANCE: seen only standard classes
-;;; BUILT-IN-INSTANCE: seen only built in classes
-;;; STRUCTURE-INSTANCE: seen only structure classes
-;;; CONDITION-INSTANCE: seen only condition classes
+;;; but these are largely unexploited as of 2007-05-10. The
+;;; distinction between STANDARD-INSTANCE and the others is used in
+;;; emitting wrapper/slot-getting code in accessor discriminating
+;;; functions (see EMIT-FETCH-WRAPPER and EMIT-READER/WRITER); it is
+;;; possible that there was an intention to use these metatypes to
+;;; specialize cache implementation or discrimination nets, but this
+;;; has not occurred as yet.
(defun raise-metatype (metatype new-specializer)
(let ((slot (find-class 'slot-class))
(standard (find-class 'standard-class))
(built-in (find-class 'built-in-class))
(frc (find-class 'forward-referenced-class)))
(flet ((specializer->metatype (x)
- (let ((meta-specializer
- (if (eq *boot-state* 'complete)
- (class-of (specializer-class x))
- (class-of x))))
+ (let* ((specializer-class (if (eq *boot-state* 'complete)
+ (specializer-class-or-nil x)
+ x))
+ (meta-specializer (class-of specializer-class)))
(cond
((eq x *the-class-t*) t)
+ ((not specializer-class) 'non-standard)
((*subtypep meta-specializer standard) 'standard-instance)
((*subtypep meta-specializer fsc) 'standard-instance)
((*subtypep meta-specializer condition) 'condition-instance)
(let ((new-metatype (specializer->metatype new-specializer)))
(cond ((eq new-metatype 'slot-instance) 'class)
((eq new-metatype 'forward) 'class)
+ ((eq new-metatype 'non-standard) 'class)
((null metatype) new-metatype)
((eq metatype new-metatype) new-metatype)
(t 'class))))))