Update of /cvsroot/sbcl/sbcl/src/compiler
In directory sc8-pr-cvs1:/tmp/cvs-serv947/src/compiler
Modified Files:
ir1tran.lisp ir1util.lisp
Log Message:
0.8.1.3:
> 13:17:03 <Xhosa> dan`b how does sbcl compile closures?
Usually good enough :-(
* Fixed bugcase 258: deny inline-expanding when expansion
references a deleted or let-converted function.
Index: ir1tran.lisp
===================================================================
RCS file: /cvsroot/sbcl/sbcl/src/compiler/ir1tran.lisp,v
retrieving revision 1.111
retrieving revision 1.112
diff -u -d -r1.111 -r1.112
--- ir1tran.lisp 16 Jun 2003 14:18:17 -0000 1.111
+++ ir1tran.lisp 25 Jun 2003 04:43:51 -0000 1.112
@@ -557,6 +557,8 @@
;;; functional instead.
(defun reference-leaf (start cont leaf)
(declare (type continuation start cont) (type leaf leaf))
+ (when (functional-p leaf)
+ (assure-functional-live-p leaf))
(let* ((type (lexenv-find leaf type-restrictions))
(leaf (or (and (defined-fun-p leaf)
(not (eq (defined-fun-inlinep leaf)
@@ -848,30 +850,7 @@
;;; are converting inline expansions for local functions during
;;; optimization.
(defun ir1-convert-local-combination (start cont form functional)
-
- ;; The test here is for "when LET converted", as a translation of
- ;; the old CMU CL comments into code. Unfortunately, the old CMU CL
- ;; comments aren't specific enough to tell whether the correct
- ;; translation is FUNCTIONAL-SOMEWHAT-LETLIKE-P or
- ;; FUNCTIONAL-LETLIKE-P or what. The old CMU CL code assumed that
- ;; any non-null FUNCTIONAL-KIND meant that the function "had been
- ;; LET converted", which might even be right, but seems fragile, so
- ;; we try to be pickier.
- (when (or
- ;; looks LET-converted
- (functional-somewhat-letlike-p functional)
- ;; It's possible for a LET-converted function to end up
- ;; deleted later. In that case, for the purposes of this
- ;; analysis, it is LET-converted: LET-converted functionals
- ;; are too badly trashed to expand them inline, and deleted
- ;; LET-converted functionals are even worse.
- (eql (functional-kind functional) :deleted))
- (throw 'locall-already-let-converted functional))
- ;; Any other non-NIL KIND value is a case we haven't found a
- ;; justification for, and at least some such values (e.g. :EXTERNAL
- ;; and :TOPLEVEL) seem obviously wrong.
- (aver (null (functional-kind functional)))
-
+ (assure-functional-live-p functional)
(ir1-convert-combination start
cont
form
Index: ir1util.lisp
===================================================================
RCS file: /cvsroot/sbcl/sbcl/src/compiler/ir1util.lisp,v
retrieving revision 1.67
retrieving revision 1.68
diff -u -d -r1.67 -r1.68
--- ir1util.lisp 24 Jun 2003 04:36:38 -0000 1.67
+++ ir1util.lisp 25 Jun 2003 04:43:51 -0000 1.68
@@ -1602,6 +1602,20 @@
*inline-expansion-limit*))
nil)
(t t))))
+
+;;; Make sure that FUNCTIONAL is not let-converted or deleted.
+(defun assure-functional-live-p (functional)
+ (declare (type functional functional))
+ (when (and (or
+ ;; looks LET-converted
+ (functional-somewhat-letlike-p functional)
+ ;; It's possible for a LET-converted function to end up
+ ;; deleted later. In that case, for the purposes of this
+ ;; analysis, it is LET-converted: LET-converted functionals
+ ;; are too badly trashed to expand them inline, and deleted
+ ;; LET-converted functionals are even worse.
+ (eql (functional-kind functional) :deleted)))
+ (throw 'locall-already-let-converted functional)))
;;;; careful call
-------------------------------------------------------
This SF.Net email is sponsored by: INetU
Attention Web Developers & Consultants: Become An INetU Hosting Partner.
Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission!
INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php
|
