3 Bugs can be reported on the help mailing list
4 sbcl-help@lists.sourceforge.net
5 or on the development mailing list
6 sbcl-devel@lists.sourceforge.net
8 Please include enough information in a bug report that someone reading
9 it can reproduce the problem, i.e. don't write
10 Subject: apparent bug in PRINT-OBJECT (or *PRINT-LENGTH*?)
11 PRINT-OBJECT doesn't seem to work with *PRINT-LENGTH*. Is this a bug?
13 Subject: apparent bug in PRINT-OBJECT (or *PRINT-LENGTH*?)
14 In sbcl-1.2.3 running under OpenBSD 4.5 on my Alpha box, when
15 I compile and load the file
16 (DEFSTRUCT (FOO (:PRINT-OBJECT (LAMBDA (X Y)
17 (LET ((*PRINT-LENGTH* 4))
20 then at the command line type
22 the program loops endlessly instead of printing the object.
27 There is also some information on bugs in the manual page and
28 in the TODO file. Eventually more such information may move here.
30 The gaps in the number sequence belong to old bug descriptions which
31 have gone away (typically because they were fixed, but sometimes for
32 other reasons, e.g. because they were moved elsewhere).
36 DEFSTRUCT almost certainly should overwrite the old LAYOUT information
37 instead of just punting when a contradictory structure definition
38 is loaded. As it is, if you redefine DEFSTRUCTs in a way which
39 changes their layout, you probably have to rebuild your entire
40 program, even if you know or guess enough about the internals of
41 SBCL to wager that this (undefined in ANSI) operation would be safe.
43 3: "type checking of structure slots"
45 ANSI specifies that a type mismatch in a structure slot
46 initialization value should not cause a warning.
48 This one might not be fixed for a while because while we're big
49 believers in ANSI compatibility and all, (1) there's no obvious
50 simple way to do it (short of disabling all warnings for type
51 mismatches everywhere), and (2) there's a good portable
52 workaround, and (3) by their own reasoning, it looks as though
53 ANSI may have gotten it wrong. ANSI justifies this specification
55 The restriction against issuing a warning for type mismatches
56 between a slot-initform and the corresponding slot's :TYPE
57 option is necessary because a slot-initform must be specified
58 in order to specify slot options; in some cases, no suitable
60 However, in SBCL (as in CMU CL or, for that matter, any compiler
61 which really understands Common Lisp types) a suitable default
62 does exist, in all cases, because the compiler understands the
63 concept of functions which never return (i.e. has return type NIL).
64 Thus, as a portable workaround, you can use a call to some
65 known-never-to-return function as the default. E.g.
67 (BAR (ERROR "missing :BAR argument")
68 :TYPE SOME-TYPE-TOO-HAIRY-TO-CONSTRUCT-AN-INSTANCE-OF))
70 (DECLAIM (FTYPE (FUNCTION () NIL) MISSING-ARG))
71 (DEFUN REQUIRED-ARG () ; workaround for SBCL non-ANSI slot init typing
72 (ERROR "missing required argument"))
74 (BAR (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
75 (BLETCH (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
76 (N-REFS-SO-FAR 0 :TYPE (INTEGER 0)))
77 Such code should compile without complaint and work correctly either
78 on SBCL or on any other completely compliant Common Lisp system.
80 b: &AUX argument in a boa-constructor without a default value means
81 "do not initilize this slot" and does not cause type error. But
82 an error may be signalled at read time and it would be good if
88 The "compiling top-level form:" output ought to be condensed.
89 Perhaps any number of such consecutive lines ought to turn into a
90 single "compiling top-level forms:" line.
93 (I *think* this is a bug. It certainly seems like strange behavior. But
94 the ANSI spec is scary, dark, and deep.. -- WHN)
95 (FORMAT NIL "~,1G" 1.4) => "1. "
96 (FORMAT NIL "~3,1G" 1.4) => "1. "
99 Sometimes (SB-EXT:QUIT) fails with
100 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
101 Process inferior-lisp exited abnormally with code 1
102 I haven't noticed a repeatable case of this yet.
105 The printer doesn't report closures very well. This is true in
109 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
110 It would be nice to make closures have a settable name slot,
111 and make things like DEFSTRUCT and FLET, which create closures,
112 set helpful values into this slot.
115 And as long as we're wishing, it would be awfully nice if INSPECT could
116 also report on closures, telling about the values of the bound variables.
119 The compiler assumes that any time a function of declared FTYPE
120 doesn't signal an error, its arguments were of the declared type.
121 E.g. compiling and loading
122 (DECLAIM (OPTIMIZE (SAFETY 3)))
123 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
125 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE)) FACTORIAL))
127 (COND ((> (FACTORIAL X) 1.0E6)
128 (FORMAT T "too big~%"))
130 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
132 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
135 will cause the INTEGERP case to be selected, giving bogus output a la
137 This violates the "declarations are assertions" principle.
138 According to the ANSI spec, in the section "System Class FUNCTION",
139 this is a case of "lying to the compiler", but the lying is done
140 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
141 not by the unexpectedly general definition of FACTORIAL. In any case,
142 "declarations are assertions" means that lying to the compiler should
143 cause an error to be signalled, and should not cause a bogus
144 result to be returned. Thus, the compiler should not assume
145 that arbitrary functions check their argument types. (It might
146 make sense to add another flag (CHECKED?) to DEFKNOWN to
147 identify functions which *do* check their argument types.)
148 (Also, verify that the compiler handles declared function
149 return types as assertions.)
152 The definitions of SIGCONTEXT-FLOAT-REGISTER and
153 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
154 supported on FreeBSD because the floating point state is not saved,
155 but at least as of FreeBSD 4.0, the floating point state *is* saved,
156 so they could be supported after all. Very likely
157 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
160 The debugger LIST-LOCATIONS command doesn't work properly.
161 (How should it work properly?)
164 Compiling and loading
165 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
167 then requesting a BACKTRACE at the debugger prompt gives no information
168 about where in the user program the problem occurred.
171 Using the pretty-printer from the command prompt gives funny
172 results, apparently because the pretty-printer doesn't know
173 about user's command input, including the user's carriage return
174 that the user, and therefore the pretty-printer thinks that
175 the new output block should start indented 2 or more characters
176 rightward of the correct location.
179 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
180 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
181 crashes SBCL. In general tracing anything which is used in the
182 implementation of TRACE is likely to have the same problem.
185 ANSI says in one place that type declarations can be abbreviated even
186 when the type name is not a symbol, e.g.
187 (DECLAIM ((VECTOR T) *FOOVECTOR*))
188 SBCL doesn't support this. But ANSI says in another place that this
189 isn't allowed. So it's not clear this is a bug after all. (See the
190 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
194 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
195 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
196 an easily guessable temporary filename in a way which might open
197 applications using LOAD-FOREIGN to hijacking by malicious users
198 on the same machine. Incantations for doing this safely are
199 floating around the net in various "how to write secure programs
200 despite Unix" documents, and it would be good to (1) fix this in
201 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
202 files and make it share the same new safe logic.
204 (partially alleviated in sbcl-0.7.9.32 by a fix by Matthew Danish to
205 make the temporary filename less easily guessable)
208 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
209 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
210 the mean of the distribution can be systematically O(0.1%) wrong.
211 Just increasing R-I-E-B is probably not a good solution, since
212 it would decrease efficiency more than is probably necessary. Perhaps
213 using some sort of accept/reject method would be better.
216 Internally the compiler sometimes evaluates
217 (sb-kernel:type/= (specifier-type '*) (specifier-type t))
218 (I stumbled across this when I added an
219 (assert (not (eq type1 *wild-type*)))
220 in the NAMED :SIMPLE-= type method.) '* isn't really a type, and
221 in a type context should probably be translated to T, and so it's
222 probably wrong to ask whether it's equal to the T type and then (using
223 the EQ type comparison in the NAMED :SIMPLE-= type method) return NIL.
224 (I haven't tried to investigate this bug enough to guess whether
225 there might be any user-level symptoms.)
227 In fact, the type system is likely to depend on this inequality not
228 holding... * is not equivalent to T in many cases, such as
229 (VECTOR *) /= (VECTOR T).
232 The facility for dumping a running Lisp image to disk gets confused
233 when run without the PURIFY option, and creates an unnecessarily large
234 core file (apparently representing memory usage up to the previous
235 high-water mark). Moreover, when the file is loaded, it confuses the
236 GC, so that thereafter memory usage can never be reduced below that
239 (As of 0.8.7.3 it's likely that the latter half of this bug is fixed.
240 The interaction between gencgc and the variables used by
241 save-lisp-and-die is still nonoptimal, though, so no respite from
245 In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
246 CL), out-of-line structure slot setters are horribly inefficient
247 whenever the type of the slot is declared, because out-of-line
248 structure slot setters are implemented as closures to save space,
249 so the compiler doesn't compile the type test into code, but
250 instead just saves the type in a lexical closure and interprets it
252 To exercise the problem, compile and load
253 (cl:in-package :cl-user)
255 (bar (error "missing") :type bar))
258 (loop (setf (foo-bar *foo*) x)))
260 (defvar *bar* (make-bar))
261 (defvar *foo* (make-foo :bar *bar*))
262 (defvar *setf-foo-bar* #'(setf foo-bar))
264 (loop (funcall *setf-foo-bar* x *foo*)))
265 then run (WASTREL1 *BAR*) or (WASTREL2 *BAR*), hit Ctrl-C, and
266 use BACKTRACE, to see it's spending all essentially all its time
267 in %TYPEP and VALUES-SPECIFIER-TYPE and so forth.
268 One possible solution would be simply to give up on
269 representing structure slot accessors as functions, and represent
270 them as macroexpansions instead. This can be inconvenient for users,
271 but it's not clear that it's worse than trying to help by expanding
272 into a horribly inefficient implementation.
273 As a workaround for the problem, #'(SETF FOO) expressions
274 can be replaced with (EFFICIENT-SETF-FUNCTION FOO), where
275 (defmacro efficient-setf-function (place-function-name)
276 (or #+sbcl (and (sb-int:info :function :accessor-for place-function-name)
277 ;; a workaround for the problem, encouraging the
278 ;; inline expansion of the structure accessor, so
279 ;; that the compiler can optimize its type test
280 (let ((new-value (gensym "NEW-VALUE-"))
281 (structure-value (gensym "STRUCTURE-VALUE-")))
282 `(lambda (,new-value ,structure-value)
283 (setf (,place-function-name ,structure-value)
285 ;; no problem, can just use the ordinary expansion
286 `(function (setf ,place-function-name))))
289 There's apparently a bug in CEILING optimization which caused
290 Douglas Crosher to patch the CMU CL version. Martin Atzmueller
291 applied the patches to SBCL and they didn't seem to cause problems
292 (as reported sbcl-devel 2001-05-04). However, since the patches
293 modify nontrivial code which was apparently written incorrectly
294 the first time around, until regression tests are written I'm not
295 comfortable merging the patches in the CVS version of SBCL.
298 (TIME (ROOM T)) reports more than 200 Mbytes consed even for
299 a clean, just-started SBCL system. And it seems to be right:
300 (ROOM T) can bring a small computer to its knees for a *long*
301 time trying to GC afterwards. Surely there's some more economical
302 way to implement (ROOM T).
304 Daniel Barlow doesn't know what fixed this, but observes that it
305 doesn't seem to be the case in 0.8.7.3 any more. Instead, (ROOM T)
306 in a fresh SBCL causes
308 debugger invoked on a SB-INT:BUG in thread 5911:
309 failed AVER: "(SAP= CURRENT END)"
311 unless a GC has happened beforehand.
314 When the compiler inline expands functions, it may be that different
315 kinds of return values are generated from different code branches.
316 E.g. an inline expansion of POSITION generates integer results
317 from one branch, and NIL results from another. When that inline
318 expansion is used in a context where only one of those results
321 (aref *a1* (position x *a2*)))
322 and the compiler can't prove that the unacceptable branch is
323 never taken, then bogus type mismatch warnings can be generated.
324 If you need to suppress the type mismatch warnings, you can
325 suppress the inline expansion,
327 #+sbcl (declare (notinline position)) ; to suppress bug 117 bogowarnings
328 (aref *a1* (position x *a2*)))
329 or, sometimes, suppress them by declaring the result to be of an
332 (aref *a1* (the integer (position x *a2*))))
334 This is not a new compiler problem in 0.7.0, but the new compiler
335 transforms for FIND, POSITION, FIND-IF, and POSITION-IF make it
336 more conspicuous. If you don't need performance from these functions,
337 and the bogus warnings are a nuisance for you, you can return to
338 your pre-0.7.0 state of grace with
339 #+sbcl (declaim (notinline find position find-if position-if)) ; bug 117..
344 As of version 0.pre7.14, SBCL's implementation of MACROLET makes
345 the entire lexical environment at the point of MACROLET available
346 in the bodies of the macroexpander functions. In particular, it
347 allows the function bodies (which run at compile time) to try to
348 access lexical variables (which are only defined at runtime).
349 It doesn't even issue a warning, which is bad.
351 The SBCL behavior arguably conforms to the ANSI spec (since the
352 spec says that the behavior is undefined, ergo anything conforms).
353 However, it would be better to issue a compile-time error.
354 Unfortunately I (WHN) don't see any simple way to detect this
355 condition in order to issue such an error, so for the meantime
356 SBCL just does this weird broken "conforming" thing.
358 The ANSI standard says, in the definition of the special operator
360 The macro-expansion functions defined by MACROLET are defined
361 in the lexical environment in which the MACROLET form appears.
362 Declarations and MACROLET and SYMBOL-MACROLET definitions affect
363 the local macro definitions in a MACROLET, but the consequences
364 are undefined if the local macro definitions reference any
365 local variable or function bindings that are visible in that
367 Then it seems to contradict itself by giving the example
369 (macrolet ((fudge (z)
370 ;The parameters x and flag are not accessible
371 ; at this point; a reference to flag would be to
372 ; the global variable of that name.
373 ` (if flag (* ,z ,z) ,z)))
374 ;The parameters x and flag are accessible here.
378 The comment "a reference to flag would be to the global variable
379 of the same name" sounds like good behavior for the system to have.
380 but actual specification quoted above says that the actual behavior
383 (Since 0.7.8.23 macroexpanders are defined in a restricted version
384 of the lexical environment, containing no lexical variables and
385 functions, which seems to conform to ANSI and CLtL2, but signalling
386 a STYLE-WARNING for references to variables similar to locals might
390 (as reported by Gabe Garza on cmucl-help 2001-09-21)
392 (defun test-pred (x y)
396 (func (lambda () x)))
397 (print (eq func func))
398 (print (test-pred func func))
399 (delete func (list func))))
400 Now calling (TEST-CASE) gives output
403 (#<FUNCTION {500A9EF9}>)
404 Evidently Python thinks of the lambda as a code transformation so
405 much that it forgets that it's also an object.
408 Ideally, uninterning a symbol would allow it, and its associated
409 FDEFINITION and PROCLAIM data, to be reclaimed by the GC. However,
410 at least as of sbcl-0.7.0, this isn't the case. Information about
411 FDEFINITIONs and PROCLAIMed properties is stored in globaldb.lisp
412 essentially in ordinary (non-weak) hash tables keyed by symbols.
413 Thus, once a system has an entry in this system, it tends to live
414 forever, even when it is uninterned and all other references to it
418 (reported by Jesse Bouwman 2001-10-24 through the unfortunately
419 prominent SourceForge web/db bug tracking system, which is
420 unfortunately not a reliable way to get a timely response from
421 the SBCL maintainers)
422 In the course of trying to build a test case for an
423 application error, I encountered this behavior:
424 If you start up sbcl, and then lay on CTRL-C for a
425 minute or two, the lisp process will eventually say:
426 %PRIMITIVE HALT called; the party is over.
427 and throw you into the monitor. If I start up lisp,
428 attach to the process with strace, and then do the same
429 (abusive) thing, I get instead:
430 access failure in heap page not marked as write-protected
431 and the monitor again. I don't know enough to have the
432 faintest idea of what is going on here.
433 This is with sbcl 6.12, uname -a reports:
434 Linux prep 2.2.19 #4 SMP Tue Apr 24 13:59:52 CDT 2001 i686 unknown
435 I (WHN) have verified that the same thing occurs on sbcl-0.pre7.141
436 under OpenBSD 2.9 on my X86 laptop. Do be patient when you try it:
437 it took more than two minutes (but less than five) for me.
441 ANSI allows types `(COMPLEX ,FOO) to use very hairy values for
442 FOO, e.g. (COMPLEX (AND REAL (SATISFIES ODDP))). The old CMU CL
443 COMPLEX implementation didn't deal with this, and hasn't been
444 upgraded to do so. (This doesn't seem to be a high priority
445 conformance problem, since seems hard to construct useful code
448 b. (fixed in 0.8.3.43)
451 Floating point errors are reported poorly. E.g. on x86 OpenBSD
454 debugger invoked on condition of type SB-KERNEL:FLOATING-POINT-EXCEPTION:
455 An arithmetic error SB-KERNEL:FLOATING-POINT-EXCEPTION was signalled.
456 No traps are enabled? How can this be?
457 It should be possible to be much more specific (overflow, division
458 by zero, etc.) and of course the "How can this be?" should be fixable.
460 See also bugs #45.c and #183
463 (reported by Robert E. Brown 2002-04-16)
464 When a function is called with too few arguments, causing the
465 debugger to be entered, the uninitialized slots in the bad call frame
466 seem to cause GCish problems, being interpreted as tagged data even
467 though they're not. In particular, executing ROOM in the
468 debugger at that point causes AVER failures:
471 * (lisp-implementation-version)
477 failed AVER: "(SAP= CURRENT END)"
478 (Christophe Rhodes reports that this doesn't occur on the SPARC, which
479 isn't too surprising since there are many differences in stack
480 implementation and GC conservatism between the X86 and other ports.)
482 This is probably the same bug as 216
485 In sbcl-0.7.3.11, compiling the (illegal) code
486 (in-package :cl-user)
487 (defmethod prove ((uustk uustk))
490 gives the (not terribly clear) error message
492 ; (during macroexpansion of (DEFMETHOD PROVE ...))
493 ; can't get template for (FROB NIL NIL)
494 The problem seems to be that the code walker used by the DEFMETHOD
495 macro is unhappy with the illegal syntax in the method body, and
496 is giving an unclear error message.
499 The compiler sometimes tries to constant-fold expressions before
500 it checks to see whether they can be reached. This can lead to
501 bogus warnings about errors in the constant folding, e.g. in code
504 (WRITE-STRING (> X 0) "+" "0"))
505 compiled in a context where the compiler can prove that X is NIL,
506 and the compiler complains that (> X 0) causes a type error because
507 NIL isn't a valid argument to #'>. Until sbcl-0.7.4.10 or so this
508 caused a full WARNING, which made the bug really annoying because then
509 COMPILE and COMPILE-FILE returned FAILURE-P=T for perfectly legal
510 code. Since then the warning has been downgraded to STYLE-WARNING,
511 so it's still a bug but at least it's a little less annoying.
513 183: "IEEE floating point issues"
514 Even where floating point handling is being dealt with relatively
515 well (as of sbcl-0.7.5, on sparc/sunos and alpha; see bug #146), the
516 accrued-exceptions and current-exceptions part of the fp control
517 word don't seem to bear much relation to reality. E.g. on
521 debugger invoked on condition of type DIVISION-BY-ZERO:
522 arithmetic error DIVISION-BY-ZERO signalled
523 0] (sb-vm::get-floating-point-modes)
525 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
526 :ROUNDING-MODE :NEAREST
527 :CURRENT-EXCEPTIONS NIL
528 :ACCRUED-EXCEPTIONS (:INEXACT)
531 * (sb-vm::get-floating-point-modes)
532 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
533 :ROUNDING-MODE :NEAREST
534 :CURRENT-EXCEPTIONS (:INEXACT)
535 :ACCRUED-EXCEPTIONS (:INEXACT)
538 188: "compiler performance fiasco involving type inference and UNION-TYPE"
542 (declare (optimize (safety 3)))
543 (declare (optimize (compilation-speed 2)))
544 (declare (optimize (speed 1) (debug 1) (space 1)))
546 (declare (type (integer 0) start))
547 (print (incf start 22))
548 (print (incf start 26))
549 (print (incf start 28)))
551 (declare (type (integer 0) start))
552 (print (incf start 22))
553 (print (incf start 26)))
555 (declare (type (integer 0) start))
556 (print (incf start 22))
557 (print (incf start 26))))))
559 This example could be solved with clever enough constraint
560 propagation or with SSA, but consider
565 The careful type of X is {2k} :-(. Is it really important to be
566 able to work with unions of many intervals?
568 191: "Miscellaneous PCL deficiencies"
569 (reported by Alexey Dejneka sbcl-devel 2002-08-04)
570 a. DEFCLASS does not inform the compiler about generated
571 functions. Compiling a file with
575 (WITH-SLOTS (A-CLASS-X) A
577 results in a STYLE-WARNING:
579 SB-SLOT-ACCESSOR-NAME::|COMMON-LISP-USER A-CLASS-X slot READER|
581 APD's fix for this was checked in to sbcl-0.7.6.20, but Pierre
582 Mai points out that the declamation of functions is in fact
583 incorrect in some cases (most notably for structure
584 classes). This means that at present erroneous attempts to use
585 WITH-SLOTS and the like on classes with metaclass STRUCTURE-CLASS
586 won't get the corresponding STYLE-WARNING.
587 c. (fixed in 0.8.4.23)
589 201: "Incautious type inference from compound types"
590 a. (reported by APD sbcl-devel 2002-09-17)
592 (LET ((Y (CAR (THE (CONS INTEGER *) X))))
594 (FORMAT NIL "~S IS ~S, Y = ~S"
601 (FOO ' (1 . 2)) => "NIL IS INTEGER, Y = 1"
605 (declare (type (array * (4 4)) x))
607 (setq x (make-array '(4 4)))
608 (adjust-array y '(3 5))
609 (= (array-dimension y 0) (eval `(array-dimension ,y 0)))))
611 * (foo (make-array '(4 4) :adjustable t))
614 205: "environment issues in cross compiler"
615 (These bugs have no impact on user code, but should be fixed or
617 a. Macroexpanders introduced with MACROLET are defined in the null
619 b. The body of (EVAL-WHEN (:COMPILE-TOPLEVEL) ...) is evaluated in
620 the null lexical environment.
621 c. The cross-compiler cannot inline functions defined in a non-null
624 206: ":SB-FLUID feature broken"
625 (reported by Antonio Martinez-Shotton sbcl-devel 2002-10-07)
626 Enabling :SB-FLUID in the target-features list in sbcl-0.7.8 breaks
629 207: "poorly distributed SXHASH results for compound data"
630 SBCL's SXHASH could probably try a little harder. ANSI: "the
631 intent is that an implementation should make a good-faith
632 effort to produce hash-codes that are well distributed
633 within the range of non-negative fixnums". But
634 (let ((hits (make-hash-table)))
637 (let* ((ij (cons i j))
638 (newlist (push ij (gethash (sxhash ij) hits))))
640 (format t "~&collision: ~S~%" newlist))))))
641 reports lots of collisions in sbcl-0.7.8. A stronger MIX function
642 would be an obvious way of fix. Maybe it would be acceptably efficient
643 to redo MIX using a lookup into a 256-entry s-box containing
644 29-bit pseudorandom numbers?
646 211: "keywords processing"
647 a. :ALLOW-OTHER-KEYS T should allow a function to receive an odd
648 number of keyword arguments.
651 (flet ((foo (&key y) (list y)))
652 (list (foo :y 1 :y 2)))
654 issues confusing message
659 ; caught STYLE-WARNING:
660 ; The variable #:G15 is defined but never used.
662 212: "Sequence functions and circular arguments"
663 COERCE, MERGE and CONCATENATE go into an infinite loop when given
664 circular arguments; it would be good for the user if they could be
665 given an error instead (ANSI 17.1.1 allows this behaviour on the part
666 of the implementation, as conforming code cannot give non-proper
667 sequences to these functions. MAP also has this problem (and
668 solution), though arguably the convenience of being able to do
669 (MAP 'LIST '+ FOO '#1=(1 . #1#))
670 might be classed as more important (though signalling an error when
671 all of the arguments are circular is probably desireable).
673 213: "Sequence functions and type checking"
674 b. MAP, when given a type argument that is SUBTYPEP LIST, does not
675 check that it will return a sequence of the given type. Fixing
676 it along the same lines as the others (cf. work done around
677 sbcl-0.7.8.45) is possible, but doing so efficiently didn't look
678 entirely straightforward.
679 c. All of these functions will silently accept a type of the form
681 whether or not the return value is of this type. This is
682 probably permitted by ANSI (see "Exceptional Situations" under
683 ANSI MAKE-SEQUENCE), but the DERIVE-TYPE mechanism does not
684 know about this escape clause, so code of the form
685 (INTEGERP (CAR (MAKE-SEQUENCE '(CONS INTEGER *) 2)))
686 can erroneously return T.
688 215: ":TEST-NOT handling by functions"
689 a. FIND and POSITION currently signal errors when given non-NIL for
690 both their :TEST and (deprecated) :TEST-NOT arguments, but by
691 ANSI 17.2 "the consequences are unspecified", which by ANSI 1.4.2
692 means that the effect is "unpredictable but harmless". It's not
693 clear what that actually means; it may preclude conforming
694 implementations from signalling errors.
695 b. COUNT, REMOVE and the like give priority to a :TEST-NOT argument
696 when conflict occurs. As a quality of implementation issue, it
697 might be preferable to treat :TEST and :TEST-NOT as being in some
698 sense the same &KEY, and effectively take the first test function in
700 c. Again, a quality of implementation issue: it would be good to issue a
701 STYLE-WARNING at compile-time for calls with :TEST-NOT, and a
702 WARNING for calls with both :TEST and :TEST-NOT; possibly this
703 latter should be WARNed about at execute-time too.
705 216: "debugger confused by frames with invalid number of arguments"
706 In sbcl-0.7.8.51, executing e.g. (VECTOR-PUSH-EXTEND T), BACKTRACE, Q
707 leaves the system confused, enough so that (QUIT) no longer works.
708 It's as though the process of working with the uninitialized slot in
709 the bad VECTOR-PUSH-EXTEND frame causes GC problems, though that may
710 not be the actual problem. (CMU CL 18c doesn't have problems with this.)
712 This is probably the same bug as 162
714 217: "Bad type operations with FUNCTION types"
717 * (values-type-union (specifier-type '(function (base-char)))
718 (specifier-type '(function (integer))))
720 #<FUN-TYPE (FUNCTION (BASE-CHAR) *)>
722 It causes insertion of wrong type assertions into generated
726 (let ((f (etypecase x
727 (character #'write-char)
728 (integer #'write-byte))))
731 (character (write-char x s))
732 (integer (write-byte x s)))))
734 Then (FOO #\1 *STANDARD-OUTPUT*) signals type error.
736 (In 0.7.9.1 the result type is (FUNCTION * *), so Python does not
737 produce invalid code, but type checking is not accurate.)
739 233: bugs in constraint propagation
741 (declaim (optimize (speed 2) (safety 3)))
743 (if (typep (prog1 x (setq x y)) 'double-float)
746 (foo 1d0 5) => segmentation violation
748 235: "type system and inline expansion"
750 (declaim (ftype (function (cons) number) acc))
751 (declaim (inline acc))
753 (the number (car c)))
756 (values (locally (declare (optimize (safety 0)))
758 (locally (declare (optimize (safety 3)))
761 (foo '(nil) '(t)) => NIL, T.
763 237: "Environment arguments to type functions"
764 a. Functions SUBTYPEP, TYPEP, UPGRADED-ARRAY-ELEMENT-TYPE, and
765 UPGRADED-COMPLEX-PART-TYPE now have an optional environment
766 argument, but they ignore it completely. This is almost
767 certainly not correct.
768 b. Also, the compiler's optimizers for TYPEP have not been informed
769 about the new argument; consequently, they will not transform
770 calls of the form (TYPEP 1 'INTEGER NIL), even though this is
771 just as optimizeable as (TYPEP 1 'INTEGER).
773 238: "REPL compiler overenthusiasm for CLOS code"
775 * (defclass foo () ())
776 * (defmethod bar ((x foo) (foo foo)) (call-next-method))
777 causes approximately 100 lines of code deletion notes. Some
778 discussion on this issue happened under the title 'Three "interesting"
779 bugs in PCL', resulting in a fix for this oververbosity from the
780 compiler proper; however, the problem persists in the interactor
781 because the notion of original source is not preserved: for the
782 compiler, the original source of the above expression is (DEFMETHOD
783 BAR ((X FOO) (FOO FOO)) (CALL-NEXT-METHOD)), while by the time the
784 compiler gets its hands on the code needing compilation from the REPL,
785 it has been macroexpanded several times.
787 A symptom of the same underlying problem, reported by Tony Martinez:
789 (with-input-from-string (*query-io* " no")
791 (simple-type-error () 'error))
793 ; (SB-KERNEL:FLOAT-WAIT)
795 ; note: deleting unreachable code
796 ; compilation unit finished
799 242: "WRITE-SEQUENCE suboptimality"
800 (observed from clx performance)
801 In sbcl-0.7.13, WRITE-SEQUENCE of a sequence of type
802 (SIMPLE-ARRAY (UNSIGNED-BYTE 8) (*)) on a stream with element-type
803 (UNSIGNED-BYTE 8) will write to the stream one byte at a time,
804 rather than writing the sequence in one go, leading to severe
805 performance degradation.
807 243: "STYLE-WARNING overenthusiasm for unused variables"
808 (observed from clx compilation)
809 In sbcl-0.7.14, in the presence of the macros
810 (DEFMACRO FOO (X) `(BAR ,X))
811 (DEFMACRO BAR (X) (DECLARE (IGNORABLE X)) 'NIL)
812 somewhat surprising style warnings are emitted for
813 (COMPILE NIL '(LAMBDA (Y) (FOO Y))):
815 ; (LAMBDA (Y) (FOO Y))
817 ; caught STYLE-WARNING:
818 ; The variable Y is defined but never used.
820 245: bugs in disassembler
821 b. On X86 operand size prefix is not recognized.
824 (defun foo (&key (a :x))
828 does not cause a warning. (BTW: old SBCL issued a warning, but for a
829 function, which was never called!)
832 Compiler does not emit warnings for
834 a. (lambda () (svref (make-array 8 :adjustable t) 1))
837 (list (let ((y (the real x)))
838 (unless (floatp y) (error ""))
843 (declare (optimize (debug 0)))
844 (declare (type vector x))
845 (list (fill-pointer x)
849 Complex array type does not have corresponding type specifier.
851 This is a problem because the compiler emits optimization notes when
852 you use a non-simple array, and without a type specifier for hairy
853 array types, there's no good way to tell it you're doing it
854 intentionally so that it should shut up and just compile the code.
856 Another problem is confusing error message "asserted type ARRAY
857 conflicts with derived type (VALUES SIMPLE-VECTOR &OPTIONAL)" during
858 compiling (LAMBDA (V) (VALUES (SVREF V 0) (VECTOR-POP V))).
860 The last problem is that when type assertions are converted to type
861 checks, types are represented with type specifiers, so we could lose
862 complex attribute. (Now this is probably not important, because
863 currently checks for complex arrays seem to be performed by
867 (compile nil '(lambda () (aref (make-array 0) 0))) compiles without
868 warning. Analogous cases with the index and length being equal and
869 greater than 0 are warned for; the problem here seems to be that the
870 type required for an array reference of this type is (INTEGER 0 (0))
871 which is canonicalized to NIL.
876 (t1 (specifier-type s)))
877 (eval `(defstruct ,s))
878 (type= t1 (specifier-type s)))
883 b. The same for CSUBTYPEP.
885 262: "yet another bug in inline expansion of local functions"
889 (declare (integer x y))
892 (declare (integer u))
893 (if (> (1+ (the unsigned-byte u)) 0)
895 (return (+ 38 (cos (/ u 78)))))))
896 (declare (inline xyz))
898 (* (funcall (eval #'xyz) x)
900 (funcall (if (> x 5) #'xyz #'identity)
905 Urgh... It's time to write IR1-copier.
908 David Lichteblau provided (sbcl-devel 2003-06-01) a patch to fix
909 behaviour of streams with element-type (SIGNED-BYTE 8). The patch
910 looks reasonable, if not obviously correct; however, it caused the
911 PPC/Linux port to segfault during warm-init while loading
912 src/pcl/std-class.fasl. A workaround patch was made, but it would
913 be nice to understand why the first patch caused problems, and to
914 fix the cause if possible.
916 268: "wrong free declaration scope"
917 The following code must signal type error:
919 (locally (declare (optimize (safety 3)))
920 (flet ((foo (x &optional (y (car x)))
921 (declare (optimize (safety 0)))
923 (funcall (eval #'foo) 1)))
926 SCALE-FLOAT should accept any integer for its second argument.
929 In the following function constraint propagator optimizes nothing:
932 (declare (integer x))
933 (declare (optimize speed))
941 Compilation of the following two forms causes "X is unbound" error:
943 (symbol-macrolet ((x pi))
944 (macrolet ((foo (y) (+ x y)))
945 (declaim (inline bar))
951 (See (COERCE (CDR X) 'FUNCTION) in IR1-CONVERT-INLINE-LAMBDA.)
954 CLHS says that type declaration of a symbol macro should not affect
955 its expansion, but in SBCL it does. (If you like magic and want to
956 fix it, don't forget to change all uses of MACROEXPAND to
960 The following code (taken from CLOCC) takes a lot of time to compile:
963 (declare (type (integer 0 #.large-constant) n))
966 (fixed in 0.8.2.51, but a test case would be good)
971 (declare (optimize speed))
972 (loop for i of-type (integer 0) from 0 by 2 below 10
975 uses generic arithmetic.
977 b. (fixed in 0.8.3.6)
979 279: type propagation error -- correctly inferred type goes astray?
980 In sbcl-0.8.3 and sbcl-0.8.1.47, the warning
981 The binding of ABS-FOO is a (VALUES (INTEGER 0 0)
982 &OPTIONAL), not a (INTEGER 1 536870911)
983 is emitted when compiling this file:
984 (declaim (ftype (function ((integer 0 #.most-positive-fixnum))
985 (integer #.most-negative-fixnum 0))
990 (let* (;; Uncomment this for a type mismatch warning indicating
991 ;; that the type of (FOO X) is correctly understood.
992 #+nil (fs-foo (float-sign (foo x)))
993 ;; Uncomment this for a type mismatch warning
994 ;; indicating that the type of (ABS (FOO X)) is
995 ;; correctly understood.
996 #+nil (fs-abs-foo (float-sign (abs (foo x))))
997 ;; something wrong with this one though
998 (abs-foo (abs (foo x))))
999 (declare (type (integer 1 100) abs-foo))
1004 281: COMPUTE-EFFECTIVE-METHOD error signalling.
1005 (slightly obscured by a non-0 default value for
1006 SB-PCL::*MAX-EMF-PRECOMPUTE-METHODS*)
1007 It would be natural for COMPUTE-EFFECTIVE-METHOD to signal errors
1008 when it finds a method with invalid qualifiers. However, it
1009 shouldn't signal errors when any such methods are not applicable to
1010 the particular call being evaluated, and certainly it shouldn't when
1011 simply precomputing effective methods that may never be called.
1012 (setf sb-pcl::*max-emf-precompute-methods* 0)
1014 (:method-combination +)
1015 (:method ((x symbol)) 1)
1016 (:method + ((x number)) x))
1017 (foo 1) -> ERROR, but should simply return 1
1019 The issue seems to be that construction of a discriminating function
1020 calls COMPUTE-EFFECTIVE-METHOD with methods that are not all applicable.
1022 283: Thread safety: libc functions
1023 There are places that we call unsafe-for-threading libc functions
1024 that we should find alternatives for, or put locks around. Known or
1025 strongly suspected problems, as of 0.8.3.10: please update this
1026 bug instead of creating new ones
1028 localtime() - called for timezone calculations in code/time.lisp
1030 284: Thread safety: special variables
1031 There are lots of special variables in SBCL, and I feel sure that at
1032 least some of them are indicative of potentially thread-unsafe
1033 parts of the system. See doc/internals/notes/threading-specials
1035 286: "recursive known functions"
1036 Self-call recognition conflicts with known function
1037 recognition. Currently cross compiler and target COMPILE do not
1038 recognize recursion, and in target compiler it can be disabled. We
1039 can always disable it for known functions with RECURSIVE attribute,
1040 but there remains a possibility of a function with a
1041 (tail)-recursive simplification pass and transforms/VOPs for base
1044 287: PPC/Linux miscompilation or corruption in first GC
1045 When the runtime is compiled with -O3 on certain PPC/Linux machines, a
1046 segmentation fault is reported at the point of first triggered GC,
1047 during the compilation of DEFSTRUCT WRAPPER. As a temporary workaround,
1048 the runtime is no longer compiled with -O3 on PPC/Linux, but it is likely
1049 that this merely obscures, not solves, the underlying problem; as and when
1050 underlying problems are fixed, it would be worth trying again to provoke
1053 288: fundamental cross-compilation issues (from old UGLINESS file)
1054 Using host floating point numbers to represent target floating point
1055 numbers, or host characters to represent target characters, is
1056 theoretically shaky. (The characters are OK as long as the characters
1057 are in the ANSI-guaranteed character set, though, so they aren't a
1058 real problem as long as the sources don't need anything but that;
1059 the floats are a real problem.)
1061 289: "type checking and source-transforms"
1063 (block nil (let () (funcall #'+ (eval 'nil) (eval '1) (return :good))))
1066 Our policy is to check argument types at the moment of a call. It
1067 disagrees with ANSI, which says that type assertions are put
1068 immediately onto argument expressions, but is easier to implement in
1069 IR1 and is more compatible to type inference, inline expansion,
1070 etc. IR1-transforms automatically keep this policy, but source
1071 transforms for associative functions (such as +), being applied
1072 during IR1-convertion, do not. It may be tolerable for direct calls
1073 (+ x y z), but for (FUNCALL #'+ x y z) it is non-conformant.
1075 b. Another aspect of this problem is efficiency. [x y + z +]
1076 requires less registers than [x y z + +]. This transformation is
1077 currently performed with source transforms, but it would be good to
1078 also perform it in IR1 optimization phase.
1080 290: Alpha floating point and denormalized traps
1081 In SBCL 0.8.3.6x on the alpha, we work around what appears to be a
1082 hardware or kernel deficiency: the status of the enable/disable
1083 denormalized-float traps bit seems to be ambiguous; by the time we
1084 get to os_restore_fp_control after a trap, denormalized traps seem
1085 to be enabled. Since we don't want a trap every time someone uses a
1086 denormalized float, in general, we mask out that bit when we restore
1087 the control word; however, this clobbers any change the user might
1091 (reported by Adam Warner, sbcl-devel 2003-09-23)
1093 The --load toplevel argument does not perform any sanitization of its
1094 argument. As a result, files with Lisp pathname pattern characters
1095 (#\* or #\?, for instance) or quotation marks can cause the system
1096 to perform arbitrary behaviour.
1099 LOOP with non-constant arithmetic step clauses suffers from overzealous
1100 type constraint: code of the form
1101 (loop for d of-type double-float from 0d0 to 10d0 by x collect d)
1102 compiles to a type restriction on X of (AND DOUBLE-FLOAT (REAL
1103 (0))). However, an integral value of X should be legal, because
1104 successive adds of integers to double-floats produces double-floats,
1105 so none of the type restrictions in the code is violated.
1107 300: (reported by Peter Graves) Function PEEK-CHAR checks PEEK-TYPE
1108 argument type only after having read a character. This is caused
1109 with EXPLICIT-CHECK attribute in DEFKNOWN. The similar problem
1110 exists with =, /=, <, >, <=, >=. They were fixed, but it is probably
1111 less error prone to have EXPLICIT-CHECK be a local declaration,
1112 being put into the definition, instead of an attribute being kept in
1113 a separate file; maybe also put it into SB-EXT?
1115 301: ARRAY-SIMPLE-=-TYPE-METHOD breaks on corner cases which can arise
1116 in NOTE-ASSUMED-TYPES
1117 In sbcl-0.8.7.32, compiling the file
1119 (declare (type integer x))
1120 (declare (type (vector (or hash-table bit)) y))
1123 (declare (type integer x))
1124 (declare (type (simple-array base (2)) y))
1127 failed AVER: "(NOT (AND (NOT EQUALP) CERTAINP))"
1129 302: Undefined type messes up DATA-VECTOR-REF expansion.
1131 (defun dis (s ei x y)
1132 (declare (type (simple-array function (2)) s) (type ei ei))
1133 (funcall (aref s ei) x y))
1134 on sbcl-0.8.7.36/X86/Linux causes a BUG to be signalled:
1135 full call to SB-KERNEL:DATA-VECTOR-REF
1137 303: "nonlinear LVARs" (aka MISC.293)
1139 (multiple-value-call #'list
1141 (multiple-value-prog1
1142 (eval '(values :a :b :c))
1148 (throw 'bar (values 3 4)))))))))))
1150 (BUU 1) returns garbage.
1152 The problem is that both EVALs sequentially write to the same LVAR.
1155 (Reported by Dave Roberts.)
1156 Local INLINE/NOTINLINE declaration removes local FTYPE declaration:
1159 (declare (ftype (function () (integer 0 10)) fee)
1163 uses generic arithmetic with INLINE and fixnum without.
1165 306: "Imprecise unions of array types"
1167 (declare (optimize speed)
1168 (type (or (array cons) (array vector)) x))
1170 (foo #((0))) => TYPE-ERROR
1177 ,@(loop for x across sb-vm:*specialized-array-element-type-properties*
1178 collect `(array ,(sb-vm:saetp-specifier x)))))
1179 => NIL, T (when it should be T, T)
1181 308: "Characters without names"
1182 (reported by Bruno Haible sbcl-devel "character names are missing"
1184 (graphic-char-p (code-char 255))
1186 (char-name (code-char 255))
1189 SBCL is unsure of what to do about characters with codes in the
1190 range 128-255. Currently they are treated as non-graphic, but don't
1191 have names, which is not compliant with the standard. Various fixes
1192 are possible, such as
1193 * giving them names such as NON-ASCII-128;
1194 * reducing CHAR-CODE-LIMIT to 127 (almost certainly unpopular);
1195 * making the characters graphic (makes a certain amount of sense);
1196 * biting the bullet and implementing Unicode (probably quite hard).
1198 309: "Dubious values for implementation limits"
1199 (reported by Bruno Haible sbcl-devel "Incorrect value of
1200 multiple-values-limit" 2004-04-19)
1201 (values-list (make-list 1000000)), on x86/linux, signals a stack
1202 exhaustion condition, despite MULTIPLE-VALUES-LIMIT being
1203 significantly larger than 1000000. There are probably similar
1204 dubious values for CALL-ARGUMENTS-LIMIT (see cmucl-help/cmucl-imp
1205 around the same time regarding a call to LIST on sparc with 1000
1206 arguments) and other implementation limit constants.
1208 311: "Tokeniser not thread-safe"
1209 (see also Robert Marlow sbcl-help "Multi threaded read chucking a
1211 The tokenizer's use of *read-buffer* and *read-buffer-length* causes
1212 spurious errors should two threads attempt to tokenise at the same
1215 314: "LOOP :INITIALLY clauses and scope of initializers"
1216 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1217 test suite, originally by Thomas F. Burdick.
1218 ;; <http://www.lisp.org/HyperSpec/Body/sec_6-1-7-2.html>
1219 ;; According to the HyperSpec 6.1.2.1.4, in for-as-equals-then, var is
1220 ;; initialized to the result of evaluating form1. 6.1.7.2 says that
1221 ;; initially clauses are evaluated in the loop prologue, which precedes all
1222 ;; loop code except for the initial settings provided by with, for, or as.
1223 (loop :for x = 0 :then (1+ x)
1224 :for y = (1+ x) :then (ash y 1)
1225 :for z :across #(1 3 9 27 81 243)
1227 :initially (assert (zerop x)) :initially (assert (= 2 w))
1228 :until (>= w 100) :collect w)
1229 Expected: (2 6 15 38)
1232 317: "FORMAT of floating point numbers"
1233 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1235 (format nil "~1F" 10) => "0." ; "10." expected
1236 (format nil "~0F" 10) => "0." ; "10." expected
1237 (format nil "~2F" 1234567.1) => "1000000." ; "1234567." expected
1238 it would be nice if whatever fixed this also untangled the two
1239 competing implementations of floating point printing (Steele and
1240 White, and Burger and Dybvig) present in src/code/print.lisp
1242 318: "stack overflow in compiler warning with redefined class"
1243 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1245 (setq *print-pretty* nil)
1247 (setf (find-class 'foo) nil)
1248 (defstruct foo slot-1)
1250 ...#<SB-KERNEL:STRUCTURE-CLASSOID #<SB-KERNEL:STRUCTURE-CLASSOID #<SB-KERNEL:STRUCTURE-CLASSOID #<SB-KERNEL:STRUCTUREControl stack guard page temporarily disabled: proceed with caution
1251 (it's not really clear what it should give: is (SETF FIND-CLASS)
1252 meant to be enough to delete structure classes from the system?
1253 Giving a stack overflow is definitely suboptimal, though.)
1255 319: "backquote with comma inside array"
1256 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1258 (read-from-string "`#1A(1 2 ,(+ 2 2) 4)")
1260 #(1 2 ((SB-IMPL::|,|) + 2 2) 4)
1261 which probably isn't intentional.
1263 323: "REPLACE, BIT-BASH and large strings"
1264 The transform for REPLACE on simple-base-strings uses BIT-BASH, which
1265 at present has an upper limit in size. Consequently, in sbcl-0.8.10
1267 (declare (optimize speed (safety 1)))
1268 (let ((x (make-string 140000000))
1269 (y (make-string 140000000)))
1270 (length (replace x y))))
1273 debugger invoked on a TYPE-ERROR in thread 2412:
1274 The value 1120000000 is not of type (MOD 536870911).
1275 (see also "more and better sequence transforms" sbcl-devel 2004-05-10)
1277 324: "STREAMs and :ELEMENT-TYPE with large bytesize"
1278 In theory, (open foo :element-type '(unsigned-byte <x>)) should work
1279 for all positive integral <x>. At present, it only works for <x> up
1280 to about 1024 (and similarly for signed-byte), so
1281 (open "/dev/zero" :element-type '(unsigned-byte 1025))
1282 gives an error in sbcl-0.8.10.
1284 325: "CLOSE :ABORT T on supeseding streams"
1285 Closing a stream opened with :IF-EXISTS :SUPERSEDE with :ABORT T leaves no
1286 file on disk, even if one existed before opening.
1288 The illegality of this is not crystal clear, as the ANSI dictionary
1289 entry for CLOSE says that when :ABORT is T superseded files are not
1290 superseded (ie. the original should be restored), whereas the OPEN
1291 entry says about :IF-EXISTS :SUPERSEDE "If possible, the
1292 implementation should not destroy the old file until the new stream
1293 is closed." -- implying that even though undesirable, early deletion
1294 is legal. Restoring the original would none the less be the polite
1297 326: "*PRINT-CIRCLE* crosstalk between streams"
1298 In sbcl-0.8.10.48 it's possible for *PRINT-CIRCLE* references to be
1299 mixed between streams when output operations are intermingled closely
1300 enough (as by doing output on S2 from within (PRINT-OBJECT X S1) in the
1301 test case below), so that e.g. the references #2# appears on a stream
1302 with no preceding #2= on that stream to define it (because the #2= was
1303 sent to another stream).
1304 (cl:in-package :cl-user)
1305 (defstruct foo index)
1306 (defparameter *foo* (make-foo :index 4))
1308 (defparameter *bar* (make-bar))
1309 (defparameter *tangle* (list *foo* *bar* *foo*))
1310 (defmethod print-object ((foo foo) stream)
1311 (let ((index (foo-index foo)))
1312 (format *trace-output*
1313 "~&-$- emitting FOO ~D, ambient *BAR*=~S~%"
1315 (format stream "[FOO ~D]" index))
1317 (let ((tsos (make-string-output-stream))
1318 (ssos (make-string-output-stream)))
1319 (let ((*print-circle* t)
1320 (*trace-output* tsos)
1321 (*standard-output* ssos))
1322 (prin1 *tangle* *standard-output*))
1323 (let ((string (get-output-stream-string ssos)))
1324 (unless (string= string "(#1=[FOO 4] #S(BAR) #1#)")
1325 ;; In sbcl-0.8.10.48 STRING was "(#1=[FOO 4] #2# #1#)".:-(
1326 (error "oops: ~S" string)))))
1327 It might be straightforward to fix this by turning the
1328 *CIRCULARITY-HASH-TABLE* and *CIRCULARITY-COUNTER* variables into
1329 per-stream slots, but (1) it would probably be sort of messy faking
1330 up the special variable binding semantics using UNWIND-PROTECT and
1331 (2) it might be sort of a pain to test that no other bugs had been
1334 328: "Profiling generic functions", transplanted from #241
1335 (from tonyms on #lisp IRC 2003-02-25)
1336 In sbcl-0.7.12.55, typing
1337 (defclass foo () ((bar :accessor foo-bar)))
1340 (defclass foo () ((bar :accessor foo-bar)))
1341 gives the error message
1342 "#:FOO-BAR already names an ordinary function or a macro."
1344 Problem: when a generic function is profiled, it appears as an ordinary
1345 function to PCL. (Remembering the uninterned accessor is OK, as the
1346 redefinition must be able to remove old accessors from their generic
1349 329: "Sequential class redefinition"
1350 reported by Bruno Haible:
1351 (defclass reactor () ((max-temp :initform 10000000)))
1352 (defvar *r1* (make-instance 'reactor))
1353 (defvar *r2* (make-instance 'reactor))
1354 (slot-value *r1* 'max-temp)
1355 (slot-value *r2* 'max-temp)
1356 (defclass reactor () ((uptime :initform 0)))
1357 (slot-value *r1* 'uptime)
1358 (defclass reactor () ((uptime :initform 0) (max-temp :initform 10000)))
1359 (slot-value *r1* 'max-temp) ; => 10000
1360 (slot-value *r2* 'max-temp) ; => 10000000 oops...
1363 The method effective when the wrapper is obsoleted can be saved
1364 in the wrapper, and then to update the instance just run through
1365 all the old wrappers in order from oldest to newest.
1367 331: "lazy creation of CLOS classes for user-defined conditions"
1369 (defstruct (bar (:include foo)))
1370 (sb-mop:class-direct-subclasses (find-class 'foo))
1371 returns NIL, rather than a singleton list containing the BAR class.
1373 332: "fasl stack inconsistency in structure redefinition"
1374 (reported by Tim Daly Jr sbcl-devel 2004-05-06)
1375 Even though structure redefinition is undefined by the standard, the
1376 following behaviour is suboptimal: running
1377 (defun stimulate-sbcl ()
1378 (let ((filename (format nil "/tmp/~A.lisp" (gensym))))
1379 ;;create a file which redefines a structure incompatibly
1380 (with-open-file (f filename :direction :output :if-exists :supersede)
1381 (print '(defstruct astruct foo) f)
1382 (print '(defstruct astruct foo bar) f))
1383 ;;compile and load the file, then invoke the continue restart on
1384 ;;the structure redefinition error
1385 (handler-bind ((error (lambda (c) (continue c))))
1386 (load (compile-file filename)))))
1388 and choosing the CONTINUE restart yields the message
1389 debugger invoked on a SB-INT:BUG in thread 27726:
1390 fasl stack not empty when it should be
1392 333: "CHECK-TYPE TYPE-ERROR-DATUM place"
1393 (reported by Tony Martinez sbcl-devel 2004-05-23)
1394 When CHECK-TYPE signals a TYPE-ERROR, the TYPE-ERROR-DATUM holds the
1395 lisp symbolic place in question rather than the place's value. This
1398 334: "COMPUTE-SLOTS used to add slots to classes"
1399 (reported by Bruno Haible sbcl-devel 2004-06-01)
1400 a. Adding a local slot does not work:
1401 (use-package "SB-PCL")
1403 (defmethod compute-slots ((class (eql (find-class 'b))))
1404 (append (call-next-method)
1405 (list (make-instance 'standard-effective-slot-definition
1407 :allocation :instance))))
1408 (defclass a () ((x :allocation :class)))
1409 ;; A should now have a shared slot, X, and a local slot, Y.
1410 (mapcar #'slot-definition-location (class-slots (find-class 'b)))
1412 There is no applicable method for the generic function
1413 #<STANDARD-GENERIC-FUNCTION CLASS-SLOTS (3)>
1414 when called with arguments
1417 b. Adding a class slot does not work:
1418 (use-package "SB-PCL")
1420 (defmethod compute-slots ((class (eql (find-class 'b))))
1421 (append (call-next-method)
1422 (list (make-instance 'standard-effective-slot-definition
1424 :allocation :class))))
1425 (defclass a () ((x :allocation :class)))
1426 ;; A should now have two shared slots, X and Y.
1427 (mapcar #'slot-definition-location (class-slots (find-class 'b)))
1429 There is no applicable method for the generic function
1430 #<STANDARD-GENERIC-FUNCTION SB-PCL::CLASS-SLOT-CELLS (1)>
1431 when called with arguments
1434 336: "slot-definitions must retain the generic functions of accessors"
1435 reported by Tony Martinez:
1436 (defclass foo () ((bar :reader foo-bar)))
1437 (defun foo-bar (x) x)
1438 (defclass foo () ((bar :reader get-bar))) ; => error, should work
1440 Note: just punting the accessor removal if the fdefinition
1441 is not a generic function is not enough:
1443 (defclass foo () ((bar :reader foo-bar)))
1444 (defvar *reader* #'foo-bar)
1445 (defun foo-bar (x) x)
1446 (defclass foo () ((bar :initform 'ok :reader get-bar)))
1447 (funcall *reader* (make-instance 'foo)) ; should be an error, since
1448 ; the method must be removed
1449 ; by the class redefinition
1451 Fixing this should also fix a subset of #328 -- update the
1452 description with a new test-case then.
1454 337: MAKE-METHOD and user-defined method classes
1455 (reported by Bruno Haible sbcl-devel 2004-06-11)
1459 (defclass user-method (standard-method) (myslot))
1460 (defmacro def-user-method (name &rest rest)
1461 (let* ((lambdalist-position (position-if #'listp rest))
1462 (qualifiers (subseq rest 0 lambdalist-position))
1463 (lambdalist (elt rest lambdalist-position))
1464 (body (subseq rest (+ lambdalist-position 1)))
1466 (subseq lambdalist 0 (or
1468 (lambda (x) (member x lambda-list-keywords))
1470 (length lambdalist))))
1471 (specializers (mapcar #'find-class
1472 (mapcar (lambda (x) (if (consp x) (second x) t))
1474 (unspecialized-required-part
1475 (mapcar (lambda (x) (if (consp x) (first x) x)) required-part))
1476 (unspecialized-lambdalist
1477 (append unspecialized-required-part
1478 (subseq lambdalist (length required-part)))))
1481 (MAKE-INSTANCE 'USER-METHOD
1482 :QUALIFIERS ',qualifiers
1483 :LAMBDA-LIST ',unspecialized-lambdalist
1484 :SPECIALIZERS ',specializers
1486 (LAMBDA (ARGUMENTS NEXT-METHODS-LIST)
1487 (FLET ((NEXT-METHOD-P () NEXT-METHODS-LIST)
1488 (CALL-NEXT-METHOD (&REST NEW-ARGUMENTS)
1489 (UNLESS NEW-ARGUMENTS (SETQ NEW-ARGUMENTS ARGUMENTS))
1490 (IF (NULL NEXT-METHODS-LIST)
1491 (ERROR "no next method for arguments ~:S" ARGUMENTS)
1492 (FUNCALL (SB-PCL:METHOD-FUNCTION
1493 (FIRST NEXT-METHODS-LIST))
1494 NEW-ARGUMENTS (REST NEXT-METHODS-LIST)))))
1495 (APPLY #'(LAMBDA ,unspecialized-lambdalist ,@body) ARGUMENTS)))))
1499 (defgeneric test-um03 (x))
1500 (defmethod test-um03 ((x integer))
1501 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1502 (def-user-method test-um03 ((x rational))
1503 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1504 (defmethod test-um03 ((x real))
1505 (list 'real x (not (null (next-method-p)))))
1510 (defgeneric test-um10 (x))
1511 (defmethod test-um10 ((x integer))
1512 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1513 (defmethod test-um10 ((x rational))
1514 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1515 (defmethod test-um10 ((x real))
1516 (list 'real x (not (null (next-method-p)))))
1517 (defmethod test-um10 :after ((x real)))
1518 (def-user-method test-um10 :around ((x integer))
1519 (list* 'around-integer x
1520 (not (null (next-method-p))) (call-next-method)))
1521 (defmethod test-um10 :around ((x rational))
1522 (list* 'around-rational x
1523 (not (null (next-method-p))) (call-next-method)))
1524 (defmethod test-um10 :around ((x real))
1525 (list* 'around-real x (not (null (next-method-p))) (call-next-method)))
1527 fails with a type error, and
1530 (defgeneric test-um12 (x))
1531 (defmethod test-um12 ((x integer))
1532 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1533 (defmethod test-um12 ((x rational))
1534 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1535 (defmethod test-um12 ((x real))
1536 (list 'real x (not (null (next-method-p)))))
1537 (defmethod test-um12 :after ((x real)))
1538 (defmethod test-um12 :around ((x integer))
1539 (list* 'around-integer x
1540 (not (null (next-method-p))) (call-next-method)))
1541 (defmethod test-um12 :around ((x rational))
1542 (list* 'around-rational x
1543 (not (null (next-method-p))) (call-next-method)))
1544 (def-user-method test-um12 :around ((x real))
1545 (list* 'around-real x (not (null (next-method-p))) (call-next-method)))
1547 fails with NO-APPLICABLE-METHOD.
1549 338: "MOP specializers as type specifiers"
1550 (reported by Bruno Haible sbcl-devel 2004-06-11)
1553 Because every valid parameter specializer is also a valid type
1554 specifier, the function typep can be used during method selection
1555 to determine whether an argument satisfies a parameter
1558 however, SBCL's EQL specializers are not type specifiers:
1559 (defmethod foo ((x (eql 4.0))) 3.0)
1560 (typep 1 (first (sb-pcl:method-specializers *)))