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 Sometimes (SB-EXT:QUIT) fails with
89 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
90 Process inferior-lisp exited abnormally with code 1
91 I haven't noticed a repeatable case of this yet.
94 And as long as we're wishing, it would be awfully nice if INSPECT could
95 also report on closures, telling about the values of the bound variables.
98 The compiler assumes that any time a function of declared FTYPE
99 doesn't signal an error, its arguments were of the declared type.
100 E.g. compiling and loading
101 (DECLAIM (OPTIMIZE (SAFETY 3)))
102 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
104 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE)) FACTORIAL))
106 (COND ((> (FACTORIAL X) 1.0E6)
107 (FORMAT T "too big~%"))
109 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
111 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
114 will cause the INTEGERP case to be selected, giving bogus output a la
116 This violates the "declarations are assertions" principle.
117 According to the ANSI spec, in the section "System Class FUNCTION",
118 this is a case of "lying to the compiler", but the lying is done
119 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
120 not by the unexpectedly general definition of FACTORIAL. In any case,
121 "declarations are assertions" means that lying to the compiler should
122 cause an error to be signalled, and should not cause a bogus
123 result to be returned. Thus, the compiler should not assume
124 that arbitrary functions check their argument types. (It might
125 make sense to add another flag (CHECKED?) to DEFKNOWN to
126 identify functions which *do* check their argument types.)
127 (Also, verify that the compiler handles declared function
128 return types as assertions.)
131 The definitions of SIGCONTEXT-FLOAT-REGISTER and
132 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
133 supported on FreeBSD because the floating point state is not saved,
134 but at least as of FreeBSD 4.0, the floating point state *is* saved,
135 so they could be supported after all. Very likely
136 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
139 Compiling and loading
140 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
142 then requesting a BACKTRACE at the debugger prompt gives no information
143 about where in the user program the problem occurred.
145 (this is apparently mostly fixed on the SPARC, PPC, and x86 architectures:
146 while giving the backtrace the non-x86 systems complains about "unknown
147 source location: using block start", but apart from that the
148 backtrace seems reasonable. On x86 this is masked by bug 353. See
149 tests/debug.impure.lisp for a test case)
152 Using the pretty-printer from the command prompt gives funny
153 results, apparently because the pretty-printer doesn't know
154 about user's command input, including the user's carriage return
155 that the user, and therefore the pretty-printer thinks that
156 the new output block should start indented 2 or more characters
157 rightward of the correct location.
160 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
161 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
162 crashes SBCL. In general tracing anything which is used in the
163 implementation of TRACE is likely to have the same problem.
166 ANSI says in one place that type declarations can be abbreviated even
167 when the type name is not a symbol, e.g.
168 (DECLAIM ((VECTOR T) *FOOVECTOR*))
169 SBCL doesn't support this. But ANSI says in another place that this
170 isn't allowed. So it's not clear this is a bug after all. (See the
171 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
175 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
176 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
177 the mean of the distribution can be systematically O(0.1%) wrong.
178 Just increasing R-I-E-B is probably not a good solution, since
179 it would decrease efficiency more than is probably necessary. Perhaps
180 using some sort of accept/reject method would be better.
183 Internally the compiler sometimes evaluates
184 (sb-kernel:type/= (specifier-type '*) (specifier-type t))
185 (I stumbled across this when I added an
186 (assert (not (eq type1 *wild-type*)))
187 in the NAMED :SIMPLE-= type method.) '* isn't really a type, and
188 in a type context should probably be translated to T, and so it's
189 probably wrong to ask whether it's equal to the T type and then (using
190 the EQ type comparison in the NAMED :SIMPLE-= type method) return NIL.
191 (I haven't tried to investigate this bug enough to guess whether
192 there might be any user-level symptoms.)
194 In fact, the type system is likely to depend on this inequality not
195 holding... * is not equivalent to T in many cases, such as
196 (VECTOR *) /= (VECTOR T).
199 The facility for dumping a running Lisp image to disk gets confused
200 when run without the PURIFY option, and creates an unnecessarily large
201 core file (apparently representing memory usage up to the previous
202 high-water mark). Moreover, when the file is loaded, it confuses the
203 GC, so that thereafter memory usage can never be reduced below that
206 (As of 0.8.7.3 it's likely that the latter half of this bug is fixed.
207 The interaction between gencgc and the variables used by
208 save-lisp-and-die is still nonoptimal, though, so no respite from
212 In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
213 CL), out-of-line structure slot setters are horribly inefficient
214 whenever the type of the slot is declared, because out-of-line
215 structure slot setters are implemented as closures to save space,
216 so the compiler doesn't compile the type test into code, but
217 instead just saves the type in a lexical closure and interprets it
219 To exercise the problem, compile and load
220 (cl:in-package :cl-user)
222 (bar (error "missing") :type bar))
225 (loop (setf (foo-bar *foo*) x)))
227 (defvar *bar* (make-bar))
228 (defvar *foo* (make-foo :bar *bar*))
229 (defvar *setf-foo-bar* #'(setf foo-bar))
231 (loop (funcall *setf-foo-bar* x *foo*)))
232 then run (WASTREL1 *BAR*) or (WASTREL2 *BAR*), hit Ctrl-C, and
233 use BACKTRACE, to see it's spending all essentially all its time
234 in %TYPEP and VALUES-SPECIFIER-TYPE and so forth.
235 One possible solution would be simply to give up on
236 representing structure slot accessors as functions, and represent
237 them as macroexpansions instead. This can be inconvenient for users,
238 but it's not clear that it's worse than trying to help by expanding
239 into a horribly inefficient implementation.
240 As a workaround for the problem, #'(SETF FOO) expressions
241 can be replaced with (EFFICIENT-SETF-FUNCTION FOO), where
242 (defmacro efficient-setf-function (place-function-name)
243 (or #+sbcl (and (sb-int:info :function :accessor-for place-function-name)
244 ;; a workaround for the problem, encouraging the
245 ;; inline expansion of the structure accessor, so
246 ;; that the compiler can optimize its type test
247 (let ((new-value (gensym "NEW-VALUE-"))
248 (structure-value (gensym "STRUCTURE-VALUE-")))
249 `(lambda (,new-value ,structure-value)
250 (setf (,place-function-name ,structure-value)
252 ;; no problem, can just use the ordinary expansion
253 `(function (setf ,place-function-name))))
256 There's apparently a bug in CEILING optimization which caused
257 Douglas Crosher to patch the CMU CL version. Martin Atzmueller
258 applied the patches to SBCL and they didn't seem to cause problems
259 (as reported sbcl-devel 2001-05-04). However, since the patches
260 modify nontrivial code which was apparently written incorrectly
261 the first time around, until regression tests are written I'm not
262 comfortable merging the patches in the CVS version of SBCL.
265 (TIME (ROOM T)) reports more than 200 Mbytes consed even for
266 a clean, just-started SBCL system. And it seems to be right:
267 (ROOM T) can bring a small computer to its knees for a *long*
268 time trying to GC afterwards. Surely there's some more economical
269 way to implement (ROOM T).
271 Daniel Barlow doesn't know what fixed this, but observes that it
272 doesn't seem to be the case in 0.8.7.3 any more. Instead, (ROOM T)
273 in a fresh SBCL causes
275 debugger invoked on a SB-INT:BUG in thread 5911:
276 failed AVER: "(SAP= CURRENT END)"
278 unless a GC has happened beforehand.
281 When the compiler inline expands functions, it may be that different
282 kinds of return values are generated from different code branches.
283 E.g. an inline expansion of POSITION generates integer results
284 from one branch, and NIL results from another. When that inline
285 expansion is used in a context where only one of those results
288 (aref *a1* (position x *a2*)))
289 and the compiler can't prove that the unacceptable branch is
290 never taken, then bogus type mismatch warnings can be generated.
291 If you need to suppress the type mismatch warnings, you can
292 suppress the inline expansion,
294 #+sbcl (declare (notinline position)) ; to suppress bug 117 bogowarnings
295 (aref *a1* (position x *a2*)))
296 or, sometimes, suppress them by declaring the result to be of an
299 (aref *a1* (the integer (position x *a2*))))
301 This is not a new compiler problem in 0.7.0, but the new compiler
302 transforms for FIND, POSITION, FIND-IF, and POSITION-IF make it
303 more conspicuous. If you don't need performance from these functions,
304 and the bogus warnings are a nuisance for you, you can return to
305 your pre-0.7.0 state of grace with
306 #+sbcl (declaim (notinline find position find-if position-if)) ; bug 117..
311 As of version 0.pre7.14, SBCL's implementation of MACROLET makes
312 the entire lexical environment at the point of MACROLET available
313 in the bodies of the macroexpander functions. In particular, it
314 allows the function bodies (which run at compile time) to try to
315 access lexical variables (which are only defined at runtime).
316 It doesn't even issue a warning, which is bad.
318 The SBCL behavior arguably conforms to the ANSI spec (since the
319 spec says that the behavior is undefined, ergo anything conforms).
320 However, it would be better to issue a compile-time error.
321 Unfortunately I (WHN) don't see any simple way to detect this
322 condition in order to issue such an error, so for the meantime
323 SBCL just does this weird broken "conforming" thing.
325 The ANSI standard says, in the definition of the special operator
327 The macro-expansion functions defined by MACROLET are defined
328 in the lexical environment in which the MACROLET form appears.
329 Declarations and MACROLET and SYMBOL-MACROLET definitions affect
330 the local macro definitions in a MACROLET, but the consequences
331 are undefined if the local macro definitions reference any
332 local variable or function bindings that are visible in that
334 Then it seems to contradict itself by giving the example
336 (macrolet ((fudge (z)
337 ;The parameters x and flag are not accessible
338 ; at this point; a reference to flag would be to
339 ; the global variable of that name.
340 ` (if flag (* ,z ,z) ,z)))
341 ;The parameters x and flag are accessible here.
345 The comment "a reference to flag would be to the global variable
346 of the same name" sounds like good behavior for the system to have.
347 but actual specification quoted above says that the actual behavior
350 (Since 0.7.8.23 macroexpanders are defined in a restricted version
351 of the lexical environment, containing no lexical variables and
352 functions, which seems to conform to ANSI and CLtL2, but signalling
353 a STYLE-WARNING for references to variables similar to locals might
357 Ideally, uninterning a symbol would allow it, and its associated
358 FDEFINITION and PROCLAIM data, to be reclaimed by the GC. However,
359 at least as of sbcl-0.7.0, this isn't the case. Information about
360 FDEFINITIONs and PROCLAIMed properties is stored in globaldb.lisp
361 essentially in ordinary (non-weak) hash tables keyed by symbols.
362 Thus, once a system has an entry in this system, it tends to live
363 forever, even when it is uninterned and all other references to it
367 (reported by Jesse Bouwman 2001-10-24 through the unfortunately
368 prominent SourceForge web/db bug tracking system, which is
369 unfortunately not a reliable way to get a timely response from
370 the SBCL maintainers)
371 In the course of trying to build a test case for an
372 application error, I encountered this behavior:
373 If you start up sbcl, and then lay on CTRL-C for a
374 minute or two, the lisp process will eventually say:
375 %PRIMITIVE HALT called; the party is over.
376 and throw you into the monitor. If I start up lisp,
377 attach to the process with strace, and then do the same
378 (abusive) thing, I get instead:
379 access failure in heap page not marked as write-protected
380 and the monitor again. I don't know enough to have the
381 faintest idea of what is going on here.
382 This is with sbcl 6.12, uname -a reports:
383 Linux prep 2.2.19 #4 SMP Tue Apr 24 13:59:52 CDT 2001 i686 unknown
384 I (WHN) have verified that the same thing occurs on sbcl-0.pre7.141
385 under OpenBSD 2.9 on my X86 laptop. Do be patient when you try it:
386 it took more than two minutes (but less than five) for me.
390 ANSI allows types `(COMPLEX ,FOO) to use very hairy values for
391 FOO, e.g. (COMPLEX (AND REAL (SATISFIES ODDP))). The old CMU CL
392 COMPLEX implementation didn't deal with this, and hasn't been
393 upgraded to do so. (This doesn't seem to be a high priority
394 conformance problem, since seems hard to construct useful code
397 [ partially fixed by CSR in 0.8.17.17 because of a PFD ansi-tests
398 report that (COMPLEX RATIO) was failing; still failing on types of
399 the form (AND NUMBER (SATISFIES REALP) (SATISFIES ZEROP)). ]
401 b. (fixed in 0.8.3.43)
404 Floating point errors are reported poorly. E.g. on x86 OpenBSD
407 debugger invoked on condition of type SB-KERNEL:FLOATING-POINT-EXCEPTION:
408 An arithmetic error SB-KERNEL:FLOATING-POINT-EXCEPTION was signalled.
409 No traps are enabled? How can this be?
410 It should be possible to be much more specific (overflow, division
411 by zero, etc.) and of course the "How can this be?" should be fixable.
413 See also bugs #45.c and #183
416 (reported by Robert E. Brown 2002-04-16)
417 When a function is called with too few arguments, causing the
418 debugger to be entered, the uninitialized slots in the bad call frame
419 seem to cause GCish problems, being interpreted as tagged data even
420 though they're not. In particular, executing ROOM in the
421 debugger at that point causes AVER failures:
424 * (lisp-implementation-version)
430 failed AVER: "(SAP= CURRENT END)"
431 (Christophe Rhodes reports that this doesn't occur on the SPARC, which
432 isn't too surprising since there are many differences in stack
433 implementation and GC conservatism between the X86 and other ports.)
435 This is probably the same bug as 216
438 The compiler sometimes tries to constant-fold expressions before
439 it checks to see whether they can be reached. This can lead to
440 bogus warnings about errors in the constant folding, e.g. in code
443 (WRITE-STRING (> X 0) "+" "0"))
444 compiled in a context where the compiler can prove that X is NIL,
445 and the compiler complains that (> X 0) causes a type error because
446 NIL isn't a valid argument to #'>. Until sbcl-0.7.4.10 or so this
447 caused a full WARNING, which made the bug really annoying because then
448 COMPILE and COMPILE-FILE returned FAILURE-P=T for perfectly legal
449 code. Since then the warning has been downgraded to STYLE-WARNING,
450 so it's still a bug but at least it's a little less annoying.
452 183: "IEEE floating point issues"
453 Even where floating point handling is being dealt with relatively
454 well (as of sbcl-0.7.5, on sparc/sunos and alpha; see bug #146), the
455 accrued-exceptions and current-exceptions part of the fp control
456 word don't seem to bear much relation to reality. E.g. on
460 debugger invoked on condition of type DIVISION-BY-ZERO:
461 arithmetic error DIVISION-BY-ZERO signalled
462 0] (sb-vm::get-floating-point-modes)
464 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
465 :ROUNDING-MODE :NEAREST
466 :CURRENT-EXCEPTIONS NIL
467 :ACCRUED-EXCEPTIONS (:INEXACT)
470 * (sb-vm::get-floating-point-modes)
471 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
472 :ROUNDING-MODE :NEAREST
473 :CURRENT-EXCEPTIONS (:INEXACT)
474 :ACCRUED-EXCEPTIONS (:INEXACT)
477 188: "compiler performance fiasco involving type inference and UNION-TYPE"
481 (declare (optimize (safety 3)))
482 (declare (optimize (compilation-speed 2)))
483 (declare (optimize (speed 1) (debug 1) (space 1)))
485 (declare (type (integer 0) start))
486 (print (incf start 22))
487 (print (incf start 26))
488 (print (incf start 28)))
490 (declare (type (integer 0) start))
491 (print (incf start 22))
492 (print (incf start 26)))
494 (declare (type (integer 0) start))
495 (print (incf start 22))
496 (print (incf start 26))))))
498 This example could be solved with clever enough constraint
499 propagation or with SSA, but consider
504 The careful type of X is {2k} :-(. Is it really important to be
505 able to work with unions of many intervals?
507 191: "Miscellaneous PCL deficiencies"
508 (reported by Alexey Dejneka sbcl-devel 2002-08-04)
509 a. DEFCLASS does not inform the compiler about generated
510 functions. Compiling a file with
514 (WITH-SLOTS (A-CLASS-X) A
516 results in a STYLE-WARNING:
518 SB-SLOT-ACCESSOR-NAME::|COMMON-LISP-USER A-CLASS-X slot READER|
520 APD's fix for this was checked in to sbcl-0.7.6.20, but Pierre
521 Mai points out that the declamation of functions is in fact
522 incorrect in some cases (most notably for structure
523 classes). This means that at present erroneous attempts to use
524 WITH-SLOTS and the like on classes with metaclass STRUCTURE-CLASS
525 won't get the corresponding STYLE-WARNING.
526 c. (fixed in 0.8.4.23)
528 201: "Incautious type inference from compound types"
529 a. (reported by APD sbcl-devel 2002-09-17)
531 (LET ((Y (CAR (THE (CONS INTEGER *) X))))
533 (FORMAT NIL "~S IS ~S, Y = ~S"
540 (FOO ' (1 . 2)) => "NIL IS INTEGER, Y = 1"
544 (declare (type (array * (4 4)) x))
546 (setq x (make-array '(4 4)))
547 (adjust-array y '(3 5))
548 (= (array-dimension y 0) (eval `(array-dimension ,y 0)))))
550 * (foo (make-array '(4 4) :adjustable t))
553 205: "environment issues in cross compiler"
554 (These bugs have no impact on user code, but should be fixed or
556 a. Macroexpanders introduced with MACROLET are defined in the null
558 b. The body of (EVAL-WHEN (:COMPILE-TOPLEVEL) ...) is evaluated in
559 the null lexical environment.
560 c. The cross-compiler cannot inline functions defined in a non-null
563 206: ":SB-FLUID feature broken"
564 (reported by Antonio Martinez-Shotton sbcl-devel 2002-10-07)
565 Enabling :SB-FLUID in the target-features list in sbcl-0.7.8 breaks
568 207: "poorly distributed SXHASH results for compound data"
569 SBCL's SXHASH could probably try a little harder. ANSI: "the
570 intent is that an implementation should make a good-faith
571 effort to produce hash-codes that are well distributed
572 within the range of non-negative fixnums". But
573 (let ((hits (make-hash-table)))
576 (let* ((ij (cons i j))
577 (newlist (push ij (gethash (sxhash ij) hits))))
579 (format t "~&collision: ~S~%" newlist))))))
580 reports lots of collisions in sbcl-0.7.8. A stronger MIX function
581 would be an obvious way of fix. Maybe it would be acceptably efficient
582 to redo MIX using a lookup into a 256-entry s-box containing
583 29-bit pseudorandom numbers?
585 211: "keywords processing"
586 a. :ALLOW-OTHER-KEYS T should allow a function to receive an odd
587 number of keyword arguments.
590 (flet ((foo (&key y) (list y)))
591 (list (foo :y 1 :y 2)))
593 issues confusing message
598 ; caught STYLE-WARNING:
599 ; The variable #:G15 is defined but never used.
601 212: "Sequence functions and circular arguments"
602 COERCE, MERGE and CONCATENATE go into an infinite loop when given
603 circular arguments; it would be good for the user if they could be
604 given an error instead (ANSI 17.1.1 allows this behaviour on the part
605 of the implementation, as conforming code cannot give non-proper
606 sequences to these functions. MAP also has this problem (and
607 solution), though arguably the convenience of being able to do
608 (MAP 'LIST '+ FOO '#1=(1 . #1#))
609 might be classed as more important (though signalling an error when
610 all of the arguments are circular is probably desireable).
612 213: "Sequence functions and type checking"
613 b. MAP, when given a type argument that is SUBTYPEP LIST, does not
614 check that it will return a sequence of the given type. Fixing
615 it along the same lines as the others (cf. work done around
616 sbcl-0.7.8.45) is possible, but doing so efficiently didn't look
617 entirely straightforward.
618 c. All of these functions will silently accept a type of the form
620 whether or not the return value is of this type. This is
621 probably permitted by ANSI (see "Exceptional Situations" under
622 ANSI MAKE-SEQUENCE), but the DERIVE-TYPE mechanism does not
623 know about this escape clause, so code of the form
624 (INTEGERP (CAR (MAKE-SEQUENCE '(CONS INTEGER *) 2)))
625 can erroneously return T.
627 215: ":TEST-NOT handling by functions"
628 a. FIND and POSITION currently signal errors when given non-NIL for
629 both their :TEST and (deprecated) :TEST-NOT arguments, but by
630 ANSI 17.2 "the consequences are unspecified", which by ANSI 1.4.2
631 means that the effect is "unpredictable but harmless". It's not
632 clear what that actually means; it may preclude conforming
633 implementations from signalling errors.
634 b. COUNT, REMOVE and the like give priority to a :TEST-NOT argument
635 when conflict occurs. As a quality of implementation issue, it
636 might be preferable to treat :TEST and :TEST-NOT as being in some
637 sense the same &KEY, and effectively take the first test function in
639 c. Again, a quality of implementation issue: it would be good to issue a
640 STYLE-WARNING at compile-time for calls with :TEST-NOT, and a
641 WARNING for calls with both :TEST and :TEST-NOT; possibly this
642 latter should be WARNed about at execute-time too.
644 216: "debugger confused by frames with invalid number of arguments"
645 In sbcl-0.7.8.51, executing e.g. (VECTOR-PUSH-EXTEND T), BACKTRACE, Q
646 leaves the system confused, enough so that (QUIT) no longer works.
647 It's as though the process of working with the uninitialized slot in
648 the bad VECTOR-PUSH-EXTEND frame causes GC problems, though that may
649 not be the actual problem. (CMU CL 18c doesn't have problems with this.)
651 This is probably the same bug as 162
653 217: "Bad type operations with FUNCTION types"
656 * (values-type-union (specifier-type '(function (base-char)))
657 (specifier-type '(function (integer))))
659 #<FUN-TYPE (FUNCTION (BASE-CHAR) *)>
661 It causes insertion of wrong type assertions into generated
665 (let ((f (etypecase x
666 (character #'write-char)
667 (integer #'write-byte))))
670 (character (write-char x s))
671 (integer (write-byte x s)))))
673 Then (FOO #\1 *STANDARD-OUTPUT*) signals type error.
675 (In 0.7.9.1 the result type is (FUNCTION * *), so Python does not
676 produce invalid code, but type checking is not accurate.)
678 233: bugs in constraint propagation
680 (declaim (optimize (speed 2) (safety 3)))
682 (if (typep (prog1 x (setq x y)) 'double-float)
685 (foo 1d0 5) => segmentation violation
687 235: "type system and inline expansion"
689 (declaim (ftype (function (cons) number) acc))
690 (declaim (inline acc))
692 (the number (car c)))
695 (values (locally (declare (optimize (safety 0)))
697 (locally (declare (optimize (safety 3)))
700 (foo '(nil) '(t)) => NIL, T.
702 237: "Environment arguments to type functions"
703 a. Functions SUBTYPEP, TYPEP, UPGRADED-ARRAY-ELEMENT-TYPE, and
704 UPGRADED-COMPLEX-PART-TYPE now have an optional environment
705 argument, but they ignore it completely. This is almost
706 certainly not correct.
707 b. Also, the compiler's optimizers for TYPEP have not been informed
708 about the new argument; consequently, they will not transform
709 calls of the form (TYPEP 1 'INTEGER NIL), even though this is
710 just as optimizeable as (TYPEP 1 'INTEGER).
712 238: "REPL compiler overenthusiasm for CLOS code"
714 * (defclass foo () ())
715 * (defmethod bar ((x foo) (foo foo)) (call-next-method))
716 causes approximately 100 lines of code deletion notes. Some
717 discussion on this issue happened under the title 'Three "interesting"
718 bugs in PCL', resulting in a fix for this oververbosity from the
719 compiler proper; however, the problem persists in the interactor
720 because the notion of original source is not preserved: for the
721 compiler, the original source of the above expression is (DEFMETHOD
722 BAR ((X FOO) (FOO FOO)) (CALL-NEXT-METHOD)), while by the time the
723 compiler gets its hands on the code needing compilation from the REPL,
724 it has been macroexpanded several times.
726 A symptom of the same underlying problem, reported by Tony Martinez:
728 (with-input-from-string (*query-io* " no")
730 (simple-type-error () 'error))
732 ; (SB-KERNEL:FLOAT-WAIT)
734 ; note: deleting unreachable code
735 ; compilation unit finished
738 242: "WRITE-SEQUENCE suboptimality"
739 (observed from clx performance)
740 In sbcl-0.7.13, WRITE-SEQUENCE of a sequence of type
741 (SIMPLE-ARRAY (UNSIGNED-BYTE 8) (*)) on a stream with element-type
742 (UNSIGNED-BYTE 8) will write to the stream one byte at a time,
743 rather than writing the sequence in one go, leading to severe
744 performance degradation.
746 243: "STYLE-WARNING overenthusiasm for unused variables"
747 (observed from clx compilation)
748 In sbcl-0.7.14, in the presence of the macros
749 (DEFMACRO FOO (X) `(BAR ,X))
750 (DEFMACRO BAR (X) (DECLARE (IGNORABLE X)) 'NIL)
751 somewhat surprising style warnings are emitted for
752 (COMPILE NIL '(LAMBDA (Y) (FOO Y))):
754 ; (LAMBDA (Y) (FOO Y))
756 ; caught STYLE-WARNING:
757 ; The variable Y is defined but never used.
759 245: bugs in disassembler
760 b. On X86 operand size prefix is not recognized.
763 (defun foo (&key (a :x))
767 does not cause a warning. (BTW: old SBCL issued a warning, but for a
768 function, which was never called!)
771 Compiler does not emit warnings for
773 a. (lambda () (svref (make-array 8 :adjustable t) 1))
776 (list (let ((y (the real x)))
777 (unless (floatp y) (error ""))
782 (declare (optimize (debug 0)))
783 (declare (type vector x))
784 (list (fill-pointer x)
788 Complex array type does not have corresponding type specifier.
790 This is a problem because the compiler emits optimization notes when
791 you use a non-simple array, and without a type specifier for hairy
792 array types, there's no good way to tell it you're doing it
793 intentionally so that it should shut up and just compile the code.
795 Another problem is confusing error message "asserted type ARRAY
796 conflicts with derived type (VALUES SIMPLE-VECTOR &OPTIONAL)" during
797 compiling (LAMBDA (V) (VALUES (SVREF V 0) (VECTOR-POP V))).
799 The last problem is that when type assertions are converted to type
800 checks, types are represented with type specifiers, so we could lose
801 complex attribute. (Now this is probably not important, because
802 currently checks for complex arrays seem to be performed by
806 (compile nil '(lambda () (aref (make-array 0) 0))) compiles without
807 warning. Analogous cases with the index and length being equal and
808 greater than 0 are warned for; the problem here seems to be that the
809 type required for an array reference of this type is (INTEGER 0 (0))
810 which is canonicalized to NIL.
815 (t1 (specifier-type s)))
816 (eval `(defstruct ,s))
817 (type= t1 (specifier-type s)))
822 b. The same for CSUBTYPEP.
824 262: "yet another bug in inline expansion of local functions"
825 During inline expansion of a local function Python can try to
826 reference optimized away objects (functions, variables, CTRANs from
827 tags and blocks), which later may lead to problems. Some of the
828 cases are worked around by forbidding expansion in such cases, but
829 the better way would be to reimplement inline expansion by copying
833 David Lichteblau provided (sbcl-devel 2003-06-01) a patch to fix
834 behaviour of streams with element-type (SIGNED-BYTE 8). The patch
835 looks reasonable, if not obviously correct; however, it caused the
836 PPC/Linux port to segfault during warm-init while loading
837 src/pcl/std-class.fasl. A workaround patch was made, but it would
838 be nice to understand why the first patch caused problems, and to
839 fix the cause if possible.
841 268: "wrong free declaration scope"
842 The following code must signal type error:
844 (locally (declare (optimize (safety 3)))
845 (flet ((foo (x &optional (y (car x)))
846 (declare (optimize (safety 0)))
848 (funcall (eval #'foo) 1)))
851 In the following function constraint propagator optimizes nothing:
854 (declare (integer x))
855 (declare (optimize speed))
863 Compilation of the following two forms causes "X is unbound" error:
865 (symbol-macrolet ((x pi))
866 (macrolet ((foo (y) (+ x y)))
867 (declaim (inline bar))
873 (See (COERCE (CDR X) 'FUNCTION) in IR1-CONVERT-INLINE-LAMBDA.)
876 CLHS says that type declaration of a symbol macro should not affect
877 its expansion, but in SBCL it does. (If you like magic and want to
878 fix it, don't forget to change all uses of MACROEXPAND to
882 The following code (taken from CLOCC) takes a lot of time to compile:
885 (declare (type (integer 0 #.large-constant) n))
888 (fixed in 0.8.2.51, but a test case would be good)
891 b. The same as in a., but using MULTIPLE-VALUE-SETQ instead of SETQ.
893 (defmethod faa ((*faa* double-float))
894 (set '*faa* (when (< *faa* 0) (- *faa*)))
896 (faa 1d0) => type error
901 (declare (optimize speed))
902 (loop for i of-type (integer 0) from 0 by 2 below 10
905 uses generic arithmetic.
907 b. (fixed in 0.8.3.6)
909 279: type propagation error -- correctly inferred type goes astray?
910 In sbcl-0.8.3 and sbcl-0.8.1.47, the warning
911 The binding of ABS-FOO is a (VALUES (INTEGER 0 0)
912 &OPTIONAL), not a (INTEGER 1 536870911)
913 is emitted when compiling this file:
914 (declaim (ftype (function ((integer 0 #.most-positive-fixnum))
915 (integer #.most-negative-fixnum 0))
920 (let* (;; Uncomment this for a type mismatch warning indicating
921 ;; that the type of (FOO X) is correctly understood.
922 #+nil (fs-foo (float-sign (foo x)))
923 ;; Uncomment this for a type mismatch warning
924 ;; indicating that the type of (ABS (FOO X)) is
925 ;; correctly understood.
926 #+nil (fs-abs-foo (float-sign (abs (foo x))))
927 ;; something wrong with this one though
928 (abs-foo (abs (foo x))))
929 (declare (type (integer 1 100) abs-foo))
934 281: COMPUTE-EFFECTIVE-METHOD error signalling.
935 (slightly obscured by a non-0 default value for
936 SB-PCL::*MAX-EMF-PRECOMPUTE-METHODS*)
937 It would be natural for COMPUTE-EFFECTIVE-METHOD to signal errors
938 when it finds a method with invalid qualifiers. However, it
939 shouldn't signal errors when any such methods are not applicable to
940 the particular call being evaluated, and certainly it shouldn't when
941 simply precomputing effective methods that may never be called.
942 (setf sb-pcl::*max-emf-precompute-methods* 0)
944 (:method-combination +)
945 (:method ((x symbol)) 1)
946 (:method + ((x number)) x))
947 (foo 1) -> ERROR, but should simply return 1
949 The issue seems to be that construction of a discriminating function
950 calls COMPUTE-EFFECTIVE-METHOD with methods that are not all applicable.
952 283: Thread safety: libc functions
953 There are places that we call unsafe-for-threading libc functions
954 that we should find alternatives for, or put locks around. Known or
955 strongly suspected problems, as of 0.8.3.10: please update this
956 bug instead of creating new ones
958 localtime() - called for timezone calculations in code/time.lisp
960 284: Thread safety: special variables
961 There are lots of special variables in SBCL, and I feel sure that at
962 least some of them are indicative of potentially thread-unsafe
963 parts of the system. See doc/internals/notes/threading-specials
965 286: "recursive known functions"
966 Self-call recognition conflicts with known function
967 recognition. Currently cross compiler and target COMPILE do not
968 recognize recursion, and in target compiler it can be disabled. We
969 can always disable it for known functions with RECURSIVE attribute,
970 but there remains a possibility of a function with a
971 (tail)-recursive simplification pass and transforms/VOPs for base
974 287: PPC/Linux miscompilation or corruption in first GC
975 When the runtime is compiled with -O3 on certain PPC/Linux machines, a
976 segmentation fault is reported at the point of first triggered GC,
977 during the compilation of DEFSTRUCT WRAPPER. As a temporary workaround,
978 the runtime is no longer compiled with -O3 on PPC/Linux, but it is likely
979 that this merely obscures, not solves, the underlying problem; as and when
980 underlying problems are fixed, it would be worth trying again to provoke
983 288: fundamental cross-compilation issues (from old UGLINESS file)
984 Using host floating point numbers to represent target floating point
985 numbers, or host characters to represent target characters, is
986 theoretically shaky. (The characters are OK as long as the characters
987 are in the ANSI-guaranteed character set, though, so they aren't a
988 real problem as long as the sources don't need anything but that;
989 the floats are a real problem.)
991 289: "type checking and source-transforms"
993 (block nil (let () (funcall #'+ (eval 'nil) (eval '1) (return :good))))
996 Our policy is to check argument types at the moment of a call. It
997 disagrees with ANSI, which says that type assertions are put
998 immediately onto argument expressions, but is easier to implement in
999 IR1 and is more compatible to type inference, inline expansion,
1000 etc. IR1-transforms automatically keep this policy, but source
1001 transforms for associative functions (such as +), being applied
1002 during IR1-convertion, do not. It may be tolerable for direct calls
1003 (+ x y z), but for (FUNCALL #'+ x y z) it is non-conformant.
1005 b. Another aspect of this problem is efficiency. [x y + z +]
1006 requires less registers than [x y z + +]. This transformation is
1007 currently performed with source transforms, but it would be good to
1008 also perform it in IR1 optimization phase.
1010 290: Alpha floating point and denormalized traps
1011 In SBCL 0.8.3.6x on the alpha, we work around what appears to be a
1012 hardware or kernel deficiency: the status of the enable/disable
1013 denormalized-float traps bit seems to be ambiguous; by the time we
1014 get to os_restore_fp_control after a trap, denormalized traps seem
1015 to be enabled. Since we don't want a trap every time someone uses a
1016 denormalized float, in general, we mask out that bit when we restore
1017 the control word; however, this clobbers any change the user might
1021 (reported by Adam Warner, sbcl-devel 2003-09-23)
1023 The --load toplevel argument does not perform any sanitization of its
1024 argument. As a result, files with Lisp pathname pattern characters
1025 (#\* or #\?, for instance) or quotation marks can cause the system
1026 to perform arbitrary behaviour.
1029 LOOP with non-constant arithmetic step clauses suffers from overzealous
1030 type constraint: code of the form
1031 (loop for d of-type double-float from 0d0 to 10d0 by x collect d)
1032 compiles to a type restriction on X of (AND DOUBLE-FLOAT (REAL
1033 (0))). However, an integral value of X should be legal, because
1034 successive adds of integers to double-floats produces double-floats,
1035 so none of the type restrictions in the code is violated.
1037 300: (reported by Peter Graves) Function PEEK-CHAR checks PEEK-TYPE
1038 argument type only after having read a character. This is caused
1039 with EXPLICIT-CHECK attribute in DEFKNOWN. The similar problem
1040 exists with =, /=, <, >, <=, >=. They were fixed, but it is probably
1041 less error prone to have EXPLICIT-CHECK be a local declaration,
1042 being put into the definition, instead of an attribute being kept in
1043 a separate file; maybe also put it into SB-EXT?
1045 301: ARRAY-SIMPLE-=-TYPE-METHOD breaks on corner cases which can arise
1046 in NOTE-ASSUMED-TYPES
1047 In sbcl-0.8.7.32, compiling the file
1049 (declare (type integer x))
1050 (declare (type (vector (or hash-table bit)) y))
1053 (declare (type integer x))
1054 (declare (type (simple-array base (2)) y))
1057 failed AVER: "(NOT (AND (NOT EQUALP) CERTAINP))"
1059 303: "nonlinear LVARs" (aka MISC.293)
1061 (multiple-value-call #'list
1063 (multiple-value-prog1
1064 (eval '(values :a :b :c))
1070 (throw 'bar (values 3 4)))))))))))
1072 (BUU 1) returns garbage.
1074 The problem is that both EVALs sequentially write to the same LVAR.
1077 (Reported by Dave Roberts.)
1078 Local INLINE/NOTINLINE declaration removes local FTYPE declaration:
1081 (declare (ftype (function () (integer 0 10)) fee)
1085 uses generic arithmetic with INLINE and fixnum without.
1087 306: "Imprecise unions of array types"
1089 (declare (optimize speed)
1090 (type (or (array cons) (array vector)) x))
1092 (foo #((0))) => TYPE-ERROR
1099 ,@(loop for x across sb-vm:*specialized-array-element-type-properties*
1100 collect `(array ,(sb-vm:saetp-specifier x)))))
1101 => NIL, T (when it should be T, T)
1103 309: "Dubious values for implementation limits"
1104 (reported by Bruno Haible sbcl-devel "Incorrect value of
1105 multiple-values-limit" 2004-04-19)
1106 (values-list (make-list 1000000)), on x86/linux, signals a stack
1107 exhaustion condition, despite MULTIPLE-VALUES-LIMIT being
1108 significantly larger than 1000000. There are probably similar
1109 dubious values for CALL-ARGUMENTS-LIMIT (see cmucl-help/cmucl-imp
1110 around the same time regarding a call to LIST on sparc with 1000
1111 arguments) and other implementation limit constants.
1113 311: "Tokeniser not thread-safe"
1114 (see also Robert Marlow sbcl-help "Multi threaded read chucking a
1116 The tokenizer's use of *read-buffer* and *read-buffer-length* causes
1117 spurious errors should two threads attempt to tokenise at the same
1120 314: "LOOP :INITIALLY clauses and scope of initializers"
1121 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1122 test suite, originally by Thomas F. Burdick.
1123 ;; <http://www.lisp.org/HyperSpec/Body/sec_6-1-7-2.html>
1124 ;; According to the HyperSpec 6.1.2.1.4, in for-as-equals-then, var is
1125 ;; initialized to the result of evaluating form1. 6.1.7.2 says that
1126 ;; initially clauses are evaluated in the loop prologue, which precedes all
1127 ;; loop code except for the initial settings provided by with, for, or as.
1128 (loop :for x = 0 :then (1+ x)
1129 :for y = (1+ x) :then (ash y 1)
1130 :for z :across #(1 3 9 27 81 243)
1132 :initially (assert (zerop x)) :initially (assert (= 2 w))
1133 :until (>= w 100) :collect w)
1134 Expected: (2 6 15 38)
1137 318: "stack overflow in compiler warning with redefined class"
1138 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1141 (setf (find-class 'foo) nil)
1142 (defstruct foo slot-1)
1143 This used to give a stack overflow from within the printer, which has
1144 been fixed as of 0.8.16.11. Current result:
1146 ; can't compile TYPEP of anonymous or undefined class:
1147 ; #<SB-KERNEL:STRUCTURE-CLASSOID FOO>
1149 debugger invoked on a TYPE-ERROR in thread 19973:
1150 The value NIL is not of type FUNCTION.
1152 CSR notes: it's not really clear what it should give: is (SETF FIND-CLASS)
1153 meant to be enough to delete structure classes from the system?
1155 319: "backquote with comma inside array"
1156 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1158 (read-from-string "`#1A(1 2 ,(+ 2 2) 4)")
1160 #(1 2 ((SB-IMPL::|,|) + 2 2) 4)
1161 which probably isn't intentional.
1163 323: "REPLACE, BIT-BASH and large strings"
1164 The transform for REPLACE on simple-base-strings uses BIT-BASH, which
1165 at present has an upper limit in size. Consequently, in sbcl-0.8.10
1167 (declare (optimize speed (safety 1)))
1168 (let ((x (make-string 140000000))
1169 (y (make-string 140000000)))
1170 (length (replace x y))))
1173 debugger invoked on a TYPE-ERROR in thread 2412:
1174 The value 1120000000 is not of type (MOD 536870911).
1175 (see also "more and better sequence transforms" sbcl-devel 2004-05-10)
1177 324: "STREAMs and :ELEMENT-TYPE with large bytesize"
1178 In theory, (open foo :element-type '(unsigned-byte <x>)) should work
1179 for all positive integral <x>. At present, it only works for <x> up
1180 to about 1024 (and similarly for signed-byte), so
1181 (open "/dev/zero" :element-type '(unsigned-byte 1025))
1182 gives an error in sbcl-0.8.10.
1184 325: "CLOSE :ABORT T on supeseding streams"
1185 Closing a stream opened with :IF-EXISTS :SUPERSEDE with :ABORT T leaves no
1186 file on disk, even if one existed before opening.
1188 The illegality of this is not crystal clear, as the ANSI dictionary
1189 entry for CLOSE says that when :ABORT is T superseded files are not
1190 superseded (ie. the original should be restored), whereas the OPEN
1191 entry says about :IF-EXISTS :SUPERSEDE "If possible, the
1192 implementation should not destroy the old file until the new stream
1193 is closed." -- implying that even though undesirable, early deletion
1194 is legal. Restoring the original would none the less be the polite
1197 326: "*PRINT-CIRCLE* crosstalk between streams"
1198 In sbcl-0.8.10.48 it's possible for *PRINT-CIRCLE* references to be
1199 mixed between streams when output operations are intermingled closely
1200 enough (as by doing output on S2 from within (PRINT-OBJECT X S1) in the
1201 test case below), so that e.g. the references #2# appears on a stream
1202 with no preceding #2= on that stream to define it (because the #2= was
1203 sent to another stream).
1204 (cl:in-package :cl-user)
1205 (defstruct foo index)
1206 (defparameter *foo* (make-foo :index 4))
1208 (defparameter *bar* (make-bar))
1209 (defparameter *tangle* (list *foo* *bar* *foo*))
1210 (defmethod print-object ((foo foo) stream)
1211 (let ((index (foo-index foo)))
1212 (format *trace-output*
1213 "~&-$- emitting FOO ~D, ambient *BAR*=~S~%"
1215 (format stream "[FOO ~D]" index))
1217 (let ((tsos (make-string-output-stream))
1218 (ssos (make-string-output-stream)))
1219 (let ((*print-circle* t)
1220 (*trace-output* tsos)
1221 (*standard-output* ssos))
1222 (prin1 *tangle* *standard-output*))
1223 (let ((string (get-output-stream-string ssos)))
1224 (unless (string= string "(#1=[FOO 4] #S(BAR) #1#)")
1225 ;; In sbcl-0.8.10.48 STRING was "(#1=[FOO 4] #2# #1#)".:-(
1226 (error "oops: ~S" string)))))
1227 It might be straightforward to fix this by turning the
1228 *CIRCULARITY-HASH-TABLE* and *CIRCULARITY-COUNTER* variables into
1229 per-stream slots, but (1) it would probably be sort of messy faking
1230 up the special variable binding semantics using UNWIND-PROTECT and
1231 (2) it might be sort of a pain to test that no other bugs had been
1234 328: "Profiling generic functions", transplanted from #241
1235 (from tonyms on #lisp IRC 2003-02-25)
1236 In sbcl-0.7.12.55, typing
1237 (defclass foo () ((bar :accessor foo-bar)))
1240 (defclass foo () ((bar :accessor foo-bar)))
1241 gives the error message
1242 "#:FOO-BAR already names an ordinary function or a macro."
1244 Problem: when a generic function is profiled, it appears as an ordinary
1245 function to PCL. (Remembering the uninterned accessor is OK, as the
1246 redefinition must be able to remove old accessors from their generic
1249 329: "Sequential class redefinition"
1250 reported by Bruno Haible:
1251 (defclass reactor () ((max-temp :initform 10000000)))
1252 (defvar *r1* (make-instance 'reactor))
1253 (defvar *r2* (make-instance 'reactor))
1254 (slot-value *r1* 'max-temp)
1255 (slot-value *r2* 'max-temp)
1256 (defclass reactor () ((uptime :initform 0)))
1257 (slot-value *r1* 'uptime)
1258 (defclass reactor () ((uptime :initform 0) (max-temp :initform 10000)))
1259 (slot-value *r1* 'max-temp) ; => 10000
1260 (slot-value *r2* 'max-temp) ; => 10000000 oops...
1263 The method effective when the wrapper is obsoleted can be saved
1264 in the wrapper, and then to update the instance just run through
1265 all the old wrappers in order from oldest to newest.
1267 332: "fasl stack inconsistency in structure redefinition"
1268 (reported by Tim Daly Jr sbcl-devel 2004-05-06)
1269 Even though structure redefinition is undefined by the standard, the
1270 following behaviour is suboptimal: running
1271 (defun stimulate-sbcl ()
1272 (let ((filename (format nil "/tmp/~A.lisp" (gensym))))
1273 ;;create a file which redefines a structure incompatibly
1274 (with-open-file (f filename :direction :output :if-exists :supersede)
1275 (print '(defstruct astruct foo) f)
1276 (print '(defstruct astruct foo bar) f))
1277 ;;compile and load the file, then invoke the continue restart on
1278 ;;the structure redefinition error
1279 (handler-bind ((error (lambda (c) (continue c))))
1280 (load (compile-file filename)))))
1282 and choosing the CONTINUE restart yields the message
1283 debugger invoked on a SB-INT:BUG in thread 27726:
1284 fasl stack not empty when it should be
1286 336: "slot-definitions must retain the generic functions of accessors"
1287 reported by Tony Martinez:
1288 (defclass foo () ((bar :reader foo-bar)))
1289 (defun foo-bar (x) x)
1290 (defclass foo () ((bar :reader get-bar))) ; => error, should work
1292 Note: just punting the accessor removal if the fdefinition
1293 is not a generic function is not enough:
1295 (defclass foo () ((bar :reader foo-bar)))
1296 (defvar *reader* #'foo-bar)
1297 (defun foo-bar (x) x)
1298 (defclass foo () ((bar :initform 'ok :reader get-bar)))
1299 (funcall *reader* (make-instance 'foo)) ; should be an error, since
1300 ; the method must be removed
1301 ; by the class redefinition
1303 Fixing this should also fix a subset of #328 -- update the
1304 description with a new test-case then.
1306 337: MAKE-METHOD and user-defined method classes
1307 (reported by Bruno Haible sbcl-devel 2004-06-11)
1311 (defclass user-method (standard-method) (myslot))
1312 (defmacro def-user-method (name &rest rest)
1313 (let* ((lambdalist-position (position-if #'listp rest))
1314 (qualifiers (subseq rest 0 lambdalist-position))
1315 (lambdalist (elt rest lambdalist-position))
1316 (body (subseq rest (+ lambdalist-position 1)))
1318 (subseq lambdalist 0 (or
1320 (lambda (x) (member x lambda-list-keywords))
1322 (length lambdalist))))
1323 (specializers (mapcar #'find-class
1324 (mapcar (lambda (x) (if (consp x) (second x) t))
1326 (unspecialized-required-part
1327 (mapcar (lambda (x) (if (consp x) (first x) x)) required-part))
1328 (unspecialized-lambdalist
1329 (append unspecialized-required-part
1330 (subseq lambdalist (length required-part)))))
1333 (MAKE-INSTANCE 'USER-METHOD
1334 :QUALIFIERS ',qualifiers
1335 :LAMBDA-LIST ',unspecialized-lambdalist
1336 :SPECIALIZERS ',specializers
1338 (LAMBDA (ARGUMENTS NEXT-METHODS-LIST)
1339 (FLET ((NEXT-METHOD-P () NEXT-METHODS-LIST)
1340 (CALL-NEXT-METHOD (&REST NEW-ARGUMENTS)
1341 (UNLESS NEW-ARGUMENTS (SETQ NEW-ARGUMENTS ARGUMENTS))
1342 (IF (NULL NEXT-METHODS-LIST)
1343 (ERROR "no next method for arguments ~:S" ARGUMENTS)
1344 (FUNCALL (SB-PCL:METHOD-FUNCTION
1345 (FIRST NEXT-METHODS-LIST))
1346 NEW-ARGUMENTS (REST NEXT-METHODS-LIST)))))
1347 (APPLY #'(LAMBDA ,unspecialized-lambdalist ,@body) ARGUMENTS)))))
1351 (defgeneric test-um03 (x))
1352 (defmethod test-um03 ((x integer))
1353 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1354 (def-user-method test-um03 ((x rational))
1355 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1356 (defmethod test-um03 ((x real))
1357 (list 'real x (not (null (next-method-p)))))
1362 (defgeneric test-um10 (x))
1363 (defmethod test-um10 ((x integer))
1364 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1365 (defmethod test-um10 ((x rational))
1366 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1367 (defmethod test-um10 ((x real))
1368 (list 'real x (not (null (next-method-p)))))
1369 (defmethod test-um10 :after ((x real)))
1370 (def-user-method test-um10 :around ((x integer))
1371 (list* 'around-integer x
1372 (not (null (next-method-p))) (call-next-method)))
1373 (defmethod test-um10 :around ((x rational))
1374 (list* 'around-rational x
1375 (not (null (next-method-p))) (call-next-method)))
1376 (defmethod test-um10 :around ((x real))
1377 (list* 'around-real x (not (null (next-method-p))) (call-next-method)))
1379 fails with a type error, and
1382 (defgeneric test-um12 (x))
1383 (defmethod test-um12 ((x integer))
1384 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1385 (defmethod test-um12 ((x rational))
1386 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1387 (defmethod test-um12 ((x real))
1388 (list 'real x (not (null (next-method-p)))))
1389 (defmethod test-um12 :after ((x real)))
1390 (defmethod test-um12 :around ((x integer))
1391 (list* 'around-integer x
1392 (not (null (next-method-p))) (call-next-method)))
1393 (defmethod test-um12 :around ((x rational))
1394 (list* 'around-rational x
1395 (not (null (next-method-p))) (call-next-method)))
1396 (def-user-method test-um12 :around ((x real))
1397 (list* 'around-real x (not (null (next-method-p))) (call-next-method)))
1399 fails with NO-APPLICABLE-METHOD.
1401 339: "DEFINE-METHOD-COMBINATION bugs"
1402 (reported by Bruno Haible via the clisp test suite)
1404 a. Syntax checking laxity (should produce errors):
1405 i. (define-method-combination foo :documentation :operator)
1406 ii. (define-method-combination foo :documentation nil)
1407 iii. (define-method-combination foo nil)
1408 iv. (define-method-combination foo nil nil
1409 (:arguments order &aux &key))
1410 v. (define-method-combination foo nil nil (:arguments &whole))
1411 vi. (define-method-combination foo nil nil (:generic-function))
1412 vii. (define-method-combination foo nil nil (:generic-function bar baz))
1413 viii. (define-method-combination foo nil nil (:generic-function (bar)))
1414 ix. (define-method-combination foo nil ((3)))
1415 x. (define-method-combination foo nil ((a)))
1417 b. define-method-combination arguments lambda list badness
1418 i. &aux args are currently unsupported;
1419 ii. default values of &optional and &key arguments are ignored;
1420 iii. supplied-p variables for &optional and &key arguments are not
1423 c. qualifier matching incorrect
1425 (define-method-combination mc27 ()
1427 (ignored (:ignore :unused)))
1429 ,@(mapcar #'(lambda (method) `(call-method ,method)) normal)))
1430 (defgeneric test-mc27 (x)
1431 (:method-combination mc27)
1432 (:method :ignore ((x number)) (/ 0)))
1435 should signal an invalid-method-error, as the :IGNORE (NUMBER)
1436 method is applicable, and yet matches neither of the method group
1439 341: PPRINT-LOGICAL-BLOCK / PPRINT-FILL / PPRINT-LINEAR sharing detection.
1440 (from Paul Dietz' test suite)
1442 CLHS on PPRINT-LINEAR and PPRINT-FILL (and PPRINT-TABULAR, though
1443 that's slightly different) states that these functions perform
1444 circular and shared structure detection on their object. Therefore,
1446 a.(let ((*print-circle* t))
1447 (pprint-linear *standard-output* (let ((x '(a))) (list x x))))
1448 should print "(#1=(A) #1#)"
1450 b.(let ((*print-circle* t))
1451 (pprint-linear *standard-output*
1452 (let ((x (cons nil nil))) (setf (cdr x) x) x)))
1453 should print "#1=(NIL . #1#)"
1455 (it is likely that the fault lies in PPRINT-LOGICAL-BLOCK, as
1456 suggested by the suggested implementation of PPRINT-TABULAR)
1458 343: MOP:COMPUTE-DISCRIMINATING-FUNCTION overriding causes error
1459 Even the simplest possible overriding of
1460 COMPUTE-DISCRIMINATING-FUNCTION, suggested in the PCL implementation
1461 as "canonical", does not work:
1462 (defclass my-generic-function (standard-generic-function) ()
1463 (:metaclass funcallable-standard-class))
1464 (defmethod compute-discriminating-function ((gf my-generic-function))
1465 (let ((dfun (call-next-method)))
1466 (lambda (&rest args)
1467 (apply dfun args))))
1469 (:generic-function-class my-generic-function))
1470 (defmethod foo (x) (+ x x))
1472 signals an error. This error is the same even if the LAMBDA is
1473 replaced by (FUNCTION (SB-KERNEL:INSTANCE-LAMBDA ...)). Maybe the
1474 SET-FUNCALLABLE-INSTANCE-FUN scary stuff in
1475 src/code/target-defstruct.lisp is broken? This seems to be broken
1476 in CMUCL 18e, so it's not caused by a recent change.
1478 344: more (?) ROOM T problems (possibly part of bug 108)
1479 In sbcl-0.8.12.51, and off and on leading up to it, the
1480 SB!VM:MEMORY-USAGE operations in ROOM T caused
1481 unhandled condition (of type SB-INT:BUG):
1482 failed AVER: "(SAP= CURRENT END)"
1483 Several clever people have taken a shot at this without fixing
1484 it; this time around (before sbcl-0.8.13 release) I (WHN) just
1485 commented out the SB!VM:MEMORY-USAGE calls until someone figures
1486 out how to make them work reliably with the rest of the GC.
1488 (Note: there's at least one dubious thing in room.lisp: see the
1489 comment in VALID-OBJ)
1491 346: alpha backtrace
1492 In sbcl-0.8.13, all backtraces from errors caused by internal errors
1493 on the alpha seem to have a "bogus stack frame".
1495 349: PPRINT-INDENT rounding implementation decisions
1496 At present, pprint-indent (and indeed the whole pretty printer)
1497 more-or-less assumes that it's using a monospace font. That's
1498 probably not too silly an assumption, but one piece of information
1499 the current implementation loses is from requests to indent by a
1500 non-integral amount. As of sbcl-0.8.15.9, the system silently
1501 truncates the indentation to an integer at the point of request, but
1502 maybe the non-integral value should be propagated through the
1503 pprinter and only truncated at output? (So that indenting by 1/2
1504 then 3/2 would indent by two spaces, not one?)
1506 352: forward-referenced-class trouble
1507 reported by Bruno Haible on sbcl-devel
1509 (setf (class-name (find-class 'a)) 'b)
1513 Expected: an instance of c, with a slot named x
1514 Got: debugger invoked on a SIMPLE-ERROR in thread 78906:
1515 While computing the class precedence list of the class named C.
1516 The class named B is a forward referenced class.
1517 The class named B is a direct superclass of the class named C.
1519 353: debugger suboptimalities on x86
1520 On x86 backtraces for undefined functions start with a bogus stack
1521 frame, and backtraces for throws to unknown catch tags with a "no
1522 debug information" frame. These are both due to CODE-COMPONENT-FROM-BITS
1523 (used on non-x86 platforms) being a more complete solution then what
1526 On x86/linux large portions of tests/debug.impure.lisp have been commented
1527 out as failures. The probable culprit for these problems is in x86-call-context
1528 (things work fine on x86/freebsd).
1530 More generally, the debugger internals suffer from excessive x86/non-x86
1531 conditionalization and OAOOMization: refactoring the common parts would
1534 354: XEPs in backtraces
1535 Under default compilation policy
1539 Has the XEP for TEST in the backtrace, not the TEST frame itself.
1540 (sparc and x86 at least)
1542 Since SBCL 0.8.20.1 this is hidden unless *SHOW-ENTRY-POINT-DETAILS*
1543 is true (instead there appear two TEST frames at least on ppc). The
1544 underlying cause seems to be that SB-C::TAIL-ANNOTATE will not merge
1545 the tail-call for the XEP, since Python has by that time proved that
1546 the function can never return; same happens if the function holds an
1547 unconditional call to ERROR.
1549 355: change-class of generic-function
1550 (reported by Bruno Haible)
1551 The MOP doesn't support change-class on a generic-function. However, SBCL
1552 apparently supports it, since it doesn't give an error or warning when doing
1553 so so. Then, however, it produces wrong results for calls to this generic
1555 ;;; The effective-methods cache:
1557 (defgeneric testgf35 (x))
1558 (defmethod testgf35 ((x integer))
1559 (cons 'integer (if (next-method-p) (call-next-method))))
1560 (defmethod testgf35 ((x real))
1561 (cons 'real (if (next-method-p) (call-next-method))))
1562 (defclass customized5-generic-function (standard-generic-function)
1564 (:metaclass sb-pcl:funcallable-standard-class))
1565 (defmethod sb-pcl:compute-effective-method ((gf customized5-generic-function) method-combination methods)
1566 `(REVERSE ,(call-next-method)))
1570 (change-class #'testgf35 'customized5-generic-function)
1572 Expected: ((INTEGER REAL) (REAL INTEGER))
1573 Got: ((INTEGER REAL) (INTEGER REAL))
1574 ;;; The discriminating-function cache:
1576 (defgeneric testgf36 (x))
1577 (defmethod testgf36 ((x integer))
1578 (cons 'integer (if (next-method-p) (call-next-method))))
1579 (defmethod testgf36 ((x real))
1580 (cons 'real (if (next-method-p) (call-next-method))))
1581 (defclass customized6-generic-function (standard-generic-function)
1583 (:metaclass sb-pcl:funcallable-standard-class))
1584 (defmethod sb-pcl:compute-discriminating-function ((gf customized6-generic-function))
1585 (let ((orig-df (call-next-method)))
1586 #'(lambda (&rest arguments)
1587 (reverse (apply orig-df arguments)))))
1591 (change-class #'testgf36 'customized6-generic-function)
1593 Expected: ((INTEGER REAL) (REAL INTEGER))
1594 Got: ((INTEGER REAL) (INTEGER REAL))
1597 (reported by Bruno Haible)
1598 After the "layout depth conflict" error, the CLOS is left in a state where
1599 it's not possible to define new standard-class subclasses any more.
1601 (defclass prioritized-dispatcher ()
1602 ((dependents :type list :initform nil)))
1603 (defmethod sb-pcl:validate-superclass ((c1 sb-pcl:funcallable-standard-class)
1604 (c2 (eql (find-class 'prioritized-dispatcher))))
1606 (defclass prioritized-generic-function (prioritized-dispatcher standard-generic-function)
1608 (:metaclass sb-pcl:funcallable-standard-class))
1609 ;; ERROR, Quit the debugger with ABORT
1610 (defclass typechecking-reader-class (standard-class)
1612 Expected: #<STANDARD-CLASS TYPECHECKING-READER-CLASS>
1613 Got: ERROR "The assertion SB-PCL::WRAPPERS failed."
1615 357: defstruct inheritance of initforms
1616 (reported by Bruno Haible)
1617 When defstruct and defclass (with :metaclass structure-class) are mixed,
1618 1. some slot initforms are ignored by the DEFSTRUCT generated constructor
1620 2. all slot initforms are ignored by MAKE-INSTANCE. (This can be arguably
1621 OK for initforms that were given in a DEFSTRUCT form, but for those
1622 given in a DEFCLASS form, I think it qualifies as a bug.)
1624 (defstruct structure02a
1628 (defclass structure02b (structure02a)
1629 ((slot4 :initform -44)
1632 (slot7 :initform (floor (* pi pi)))
1633 (slot8 :initform 88))
1634 (:metaclass structure-class))
1635 (defstruct (structure02c (:include structure02b (slot8 -88)))
1638 (slot11 (floor (exp 3))))
1640 (let ((a (make-structure02c)))
1641 (list (structure02c-slot4 a)
1642 (structure02c-slot5 a)
1643 (structure02c-slot6 a)
1644 (structure02c-slot7 a)))
1645 Expected: (-44 nil t 9)
1646 Got: (SB-PCL::..SLOT-UNBOUND.. SB-PCL::..SLOT-UNBOUND..
1647 SB-PCL::..SLOT-UNBOUND.. SB-PCL::..SLOT-UNBOUND..)
1649 (let ((b (make-instance 'structure02c)))
1650 (list (structure02c-slot2 b)
1651 (structure02c-slot3 b)
1652 (structure02c-slot4 b)
1653 (structure02c-slot6 b)
1654 (structure02c-slot7 b)
1655 (structure02c-slot8 b)
1656 (structure02c-slot10 b)
1657 (structure02c-slot11 b)))
1658 Expected: (t 3 -44 t 9 -88 t 20)
1659 Got: (0 0 0 0 0 0 0 0)
1661 358: :DECLARE argument to ENSURE-GENERIC-FUNCTION
1662 (reported by Bruno Haible)
1663 According to ANSI CL, ensure-generic-function must accept a :DECLARE
1664 keyword argument. In SBCL 0.8.16 it does not.
1667 (ensure-generic-function 'foo113 :declare '((optimize (speed 3))))
1668 (sb-pcl:generic-function-declarations #'foo113))
1669 Expected: ((OPTIMIZE (SPEED 3)))
1671 Invalid initialization argument:
1673 in call for class #<SB-MOP:FUNCALLABLE-STANDARD-CLASS STANDARD-GENERIC-FUNCTION>.
1675 The ANSI Standard, Section 7.1.2
1677 Bruno notes: The MOP specifies that ensure-generic-function accepts :DECLARATIONS.
1678 The easiest way to be compliant to both specs is to accept both (exclusively
1681 359: wrong default value for ensure-generic-function's :generic-function-class argument
1682 (reported by Bruno Haible)
1683 ANSI CL is silent on this, but the MOP's specification of ENSURE-GENERIC-FUNCTION says:
1684 "The remaining arguments are the complete set of keyword arguments
1685 received by ENSURE-GENERIC-FUNCTION."
1686 and the spec of ENSURE-GENERIC-FUNCTION-USING-CLASS:
1687 ":GENERIC-FUNCTION-CLASS - a class metaobject or a class name. If it is not
1688 supplied, it defaults to the class named STANDARD-GENERIC-FUNCTION."
1689 This is not the case in SBCL. Test case:
1690 (defclass my-generic-function (standard-generic-function)
1692 (:metaclass sb-pcl:funcallable-standard-class))
1693 (setf (fdefinition 'foo1)
1694 (make-instance 'my-generic-function :name 'foo1))
1695 (ensure-generic-function 'foo1
1696 :generic-function-class (find-class 'standard-generic-function))
1698 ; => #<SB-MOP:FUNCALLABLE-STANDARD-CLASS STANDARD-GENERIC-FUNCTION>
1699 (setf (fdefinition 'foo2)
1700 (make-instance 'my-generic-function :name 'foo2))
1701 (ensure-generic-function 'foo2)
1703 Expected: #<SB-MOP:FUNCALLABLE-STANDARD-CLASS STANDARD-GENERIC-FUNCTION>
1704 Got: #<SB-MOP:FUNCALLABLE-STANDARD-CLASS MY-GENERIC-FUNCTION>
1706 360: CALL-METHOD not recognized in method-combination body
1707 (reported by Bruno Haible)
1708 This method combination, which adds 'redo' and 'return' restarts for each
1709 method invocation to standard method combination, gives an error in SBCL.
1710 (defun prompt-for-new-values ()
1711 (format *debug-io* "~&New values: ")
1712 (list (read *debug-io*)))
1713 (defun add-method-restarts (form method)
1714 (let ((block (gensym))
1722 :REPORT (LAMBDA (STREAM) (FORMAT STREAM "Try calling ~S again." ,method))
1725 :REPORT (LAMBDA (STREAM) (FORMAT STREAM "Specify return values for ~S call." ,method))
1726 :INTERACTIVE (LAMBDA () (PROMPT-FOR-NEW-VALUES))
1727 (RETURN-FROM ,block (VALUES-LIST L)))))))))
1728 (defun convert-effective-method (efm)
1730 (if (eq (car efm) 'CALL-METHOD)
1731 (let ((method-list (third efm)))
1732 (if (or (typep (first method-list) 'method) (rest method-list))
1733 ; Reduce the case of multiple methods to a single one.
1734 ; Make the call to the next-method explicit.
1735 (convert-effective-method
1736 `(CALL-METHOD ,(second efm)
1738 (CALL-METHOD ,(first method-list) ,(rest method-list))))))
1739 ; Now the case of at most one method.
1740 (if (typep (second efm) 'method)
1741 ; Wrap the method call in a RESTART-CASE.
1742 (add-method-restarts
1743 (cons (convert-effective-method (car efm))
1744 (convert-effective-method (cdr efm)))
1746 ; Normal recursive processing.
1747 (cons (convert-effective-method (car efm))
1748 (convert-effective-method (cdr efm))))))
1749 (cons (convert-effective-method (car efm))
1750 (convert-effective-method (cdr efm))))
1752 (define-method-combination standard-with-restarts ()
1755 (primary () :required t)
1757 (flet ((call-methods-sequentially (methods)
1758 (mapcar #'(lambda (method)
1759 `(CALL-METHOD ,method))
1761 (let ((form (if (or before after (rest primary))
1762 `(MULTIPLE-VALUE-PROG1
1764 ,@(call-methods-sequentially before)
1765 (CALL-METHOD ,(first primary) ,(rest primary)))
1766 ,@(call-methods-sequentially (reverse after)))
1767 `(CALL-METHOD ,(first primary)))))
1770 `(CALL-METHOD ,(first around)
1771 (,@(rest around) (MAKE-METHOD ,form)))))
1772 (convert-effective-method form))))
1773 (defgeneric testgf16 (x) (:method-combination standard-with-restarts))
1774 (defclass testclass16a () ())
1775 (defclass testclass16b (testclass16a) ())
1776 (defclass testclass16c (testclass16a) ())
1777 (defclass testclass16d (testclass16b testclass16c) ())
1778 (defmethod testgf16 ((x testclass16a))
1780 (not (null (find-restart 'method-redo)))
1781 (not (null (find-restart 'method-return)))))
1782 (defmethod testgf16 ((x testclass16b))
1783 (cons 'b (call-next-method)))
1784 (defmethod testgf16 ((x testclass16c))
1785 (cons 'c (call-next-method)))
1786 (defmethod testgf16 ((x testclass16d))
1787 (cons 'd (call-next-method)))
1788 (testgf16 (make-instance 'testclass16d))
1790 Expected: (D B C A T T)
1791 Got: ERROR CALL-METHOD outside of a effective method form
1793 This is a bug because ANSI CL HyperSpec/Body/locmac_call-m__make-method
1795 "The macro call-method invokes the specified method, supplying it with
1796 arguments and with definitions for call-next-method and for next-method-p.
1797 If the invocation of call-method is lexically inside of a make-method,
1798 the arguments are those that were supplied to that method. Otherwise
1799 the arguments are those that were supplied to the generic function."
1800 and the example uses nothing more than these two cases (as you can see by
1801 doing (trace convert-effective-method)).
1803 361: initialize-instance of standard-reader-method ignores :function argument
1804 (reported by Bruno Haible)
1805 Pass a custom :function argument to initialize-instance of a
1806 standard-reader-method instance, but it has no effect.
1807 ;; Check that it's possible to define reader methods that do typechecking.
1809 (defclass typechecking-reader-method (sb-pcl:standard-reader-method)
1811 (defmethod initialize-instance ((method typechecking-reader-method) &rest initargs
1812 &key slot-definition)
1813 (let ((name (sb-pcl:slot-definition-name slot-definition))
1814 (type (sb-pcl:slot-definition-type slot-definition)))
1815 (apply #'call-next-method method
1816 :function #'(lambda (args next-methods)
1817 (declare (ignore next-methods))
1818 (apply #'(lambda (instance)
1819 (let ((value (slot-value instance name)))
1820 (unless (typep value type)
1821 (error "Slot ~S of ~S is not of type ~S: ~S"
1822 name instance type value))
1826 (defclass typechecking-reader-class (standard-class)
1828 (defmethod sb-pcl:validate-superclass ((c1 typechecking-reader-class) (c2 standard-class))
1830 (defmethod reader-method-class ((class typechecking-reader-class) direct-slot &rest args)
1831 (find-class 'typechecking-reader-method))
1832 (defclass testclass25 ()
1833 ((pair :type (cons symbol (cons symbol null)) :initarg :pair :accessor testclass25-pair))
1834 (:metaclass typechecking-reader-class))
1835 (macrolet ((succeeds (form)
1836 `(not (nth-value 1 (ignore-errors ,form)))))
1837 (let ((p (list 'abc 'def))
1838 (x (make-instance 'testclass25)))
1839 (list (succeeds (make-instance 'testclass25 :pair '(seventeen 17)))
1840 (succeeds (setf (testclass25-pair x) p))
1841 (succeeds (setf (second p) 456))
1842 (succeeds (testclass25-pair x))
1843 (succeeds (slot-value x 'pair))))))
1844 Expected: (t t t nil t)
1847 (inspect (first (sb-pcl:generic-function-methods #'testclass25-pair)))
1848 shows that the method was created with a FAST-FUNCTION slot but with a
1849 FUNCTION slot of NIL.
1851 362: missing error when a slot-definition is created without a name
1852 (reported by Bruno Haible)
1853 The MOP says about slot-definition initialization:
1854 "The :NAME argument is a slot name. An ERROR is SIGNALled if this argument
1855 is not a symbol which can be used as a variable name. An ERROR is SIGNALled
1856 if this argument is not supplied."
1858 (make-instance (find-class 'sb-pcl:standard-direct-slot-definition))
1860 Got: #<SB-MOP:STANDARD-DIRECT-SLOT-DEFINITION NIL>
1862 363: missing error when a slot-definition is created with a wrong documentation object
1863 (reported by Bruno Haible)
1864 The MOP says about slot-definition initialization:
1865 "The :DOCUMENTATION argument is a STRING or NIL. An ERROR is SIGNALled
1866 if it is not. This argument default to NIL during initialization."
1868 (make-instance (find-class 'sb-pcl:standard-direct-slot-definition)
1870 :documentation 'not-a-string)
1872 Got: #<SB-MOP:STANDARD-DIRECT-SLOT-DEFINITION FOO>
1874 364: does not support class objects as specializer names
1875 (reported by Bruno Haible)
1876 According to ANSI CL 7.6.2, class objects are valid specializer names,
1877 and "Parameter specializer names are used in macros intended as the
1878 user-level interface (defmethod)". DEFMETHOD's syntax section doesn't
1879 mention this possibility in the BNF for parameter-specializer-name;
1880 however, this appears to be an editorial omission, since the CLHS
1881 mentions issue CLASS-OBJECT-SPECIALIZER:AFFIRM as being approved
1882 by X3J13. SBCL doesn't support it:
1883 (defclass foo () ())
1884 (defmethod goo ((x #.(find-class 'foo))) x)
1885 Expected: #<STANDARD-METHOD GOO (#<STANDARD-CLASS FOO>)>
1886 Got: ERROR "#<STANDARD-CLASS FOO> is not a legal class name."
1888 365: mixin on generic-function subclass
1889 (reported by Bruno Haible)
1891 (defclass prioritized-dispatcher ()
1892 ((dependents :type list :initform nil)))
1893 on a generic-function subclass:
1894 (defclass prioritized-generic-function (prioritized-dispatcher standard-generic-function)
1896 (:metaclass sb-pcl:funcallable-standard-class))
1897 SBCL gives an error on this, telling to define a method on SB-MOP:VALIDATE-SUPERCLASS. If done,
1898 (defmethod sb-pcl:validate-superclass ((c1 sb-pcl:funcallable-standard-class)
1899 (c2 (eql (find-class 'prioritized-dispatcher))))
1902 (defclass prioritized-generic-function (prioritized-dispatcher standard-generic-function)
1904 (:metaclass sb-pcl:funcallable-standard-class))
1905 => debugger invoked on a SIMPLE-ERROR in thread 6687:
1906 layout depth conflict: #(#<SB-KERNEL:LAYOUT for T {500E1E9}> ...)
1908 Further discussion on this: http://thread.gmane.org/gmane.lisp.steel-bank.general/491
1910 366: cannot define two generic functions with user-defined class
1911 (reported by Bruno Haible)
1912 it is possible to define one generic function class and an instance
1913 of it. But attempting to do the same thing again, in the same session,
1914 leads to a "Control stack exhausted" error. Test case:
1915 (defclass my-generic-function-1 (standard-generic-function)
1917 (:metaclass sb-pcl:funcallable-standard-class))
1918 (defgeneric testgf-1 (x) (:generic-function-class my-generic-function-1)
1919 (:method ((x integer)) (cons 'integer nil)))
1920 (defclass my-generic-function-2 (standard-generic-function)
1922 (:metaclass sb-pcl:funcallable-standard-class))
1923 (defgeneric testgf-2 (x) (:generic-function-class my-generic-function-2)
1924 (:method ((x integer)) (cons 'integer nil)))
1925 => SB-KERNEL::CONTROL-STACK-EXHAUSTED
1927 367: TYPE-ERROR at compile time, undetected TYPE-ERROR at runtime
1929 (declaim (optimize (safety 3) (debug 2) (speed 2) (space 1)))
1933 (i367s (make-array 0 :fill-pointer t) :type (or (vector i367) null)))
1935 (g367 (error "missing :G367") :type g367 :read-only t))
1936 ;;; In sbcl-0.8.18, commenting out this (DECLAIM (FTYPE ... R367))
1937 ;;; gives an internal error at compile time:
1938 ;;; The value #<SB-KERNEL:NAMED-TYPE NIL> is not of
1939 ;;; type SB-KERNEL:VALUES-TYPE.
1940 (declaim (ftype (function ((vector i367) e367) (or s367 null)) r367))
1941 (declaim (ftype (function ((vector e367)) (values)) h367))
1943 (let ((x (g367-i367s (make-g367))))
1944 (let* ((y (or (r367 x w)
1947 (format t "~&Y=~S Z=~S~%" y z)
1949 (defun r367 (x y) (declare (ignore x y)) nil)
1950 (defun h367 (x) (declare (ignore x)) (values))
1951 ;;; In sbcl-0.8.18, executing this form causes an low-level error
1952 ;;; segmentation violation at #X9B0E1F4
1953 ;;; (instead of the TYPE-ERROR that one might like).
1954 (frob 0 (make-e367))
1955 can be made to cause two different problems, as noted in the comments:
1956 bug 367a: Compile and load the file. No TYPE-ERROR is signalled at
1957 run time (in the (S367-G367 Y) form of FROB, when Y is NIL
1958 instead of an instance of S367). Instead (on x86/Linux at least)
1959 we end up with a segfault.
1960 bug 367b: Comment out the (DECLAIM (FTYPE ... R367)), and compile
1961 the file. The compiler fails with TYPE-ERROR at compile time.
1963 368: miscompiled OR (perhaps related to bug 367)
1964 Trying to relax type declarations to find a workaround for bug 367,
1965 it turns out that even when the return type isn't declared (or
1966 declared to be T, anyway) the system remains confused about type
1967 inference in code similar to that for bug 367:
1968 (in-package :cl-user)
1969 (declaim (optimize (safety 3) (debug 2) (speed 2) (space 1)))
1973 (i368s (make-array 0 :fill-pointer t) :type (or (vector i368) null)))
1975 (g368 (error "missing :G368") :type g368 :read-only t))
1976 (declaim (ftype (function (fixnum (vector i368) e368) t) r368))
1977 (declaim (ftype (function (fixnum (vector e368)) t) h368))
1978 (defparameter *h368-was-called-p* nil)
1979 (defun nsu (vertices e368)
1980 (let ((i368s (g368-i368s (make-g368))))
1981 (let ((fuis (r368 0 i368s e368)))
1982 (format t "~&FUIS=~S~%" fuis)
1983 (or fuis (h368 0 i368s)))))
1985 (declare (ignore w x y))
1988 (declare (ignore w x))
1989 (setf *h368-was-called-p* t)
1990 (make-s368 :g368 (make-g368)))
1992 (format t "~&calling NSU~%")
1993 (let ((nsu (nsu #() (make-e368))))
1994 (format t "~&NSU returned ~S~%" nsu)
1995 (format t "~&*H368-WAS-CALLED-P*=~S~%" *h368-was-called-p*)
1996 (assert (s368-p nsu))
1997 (assert *h368-was-called-p*))
1998 In sbcl-0.8.18, both ASSERTs fail, and (DISASSEMBLE 'NSU) shows
1999 that no call to H368 is compiled.
2001 369: unlike-an-intersection behavior of VALUES-TYPE-INTERSECTION
2002 In sbcl-0.8.18.2, the identity $(x \cap y \cap y)=(x \cap y)$
2003 does not hold for VALUES-TYPE-INTERSECTION, even for types which
2004 can be intersected exactly, so that ASSERTs fail in this test case:
2005 (in-package :cl-user)
2006 (let ((types (mapcar #'sb-c::values-specifier-type
2007 '((values (vector package) &optional)
2008 (values (vector package) &rest t)
2009 (values (vector hash-table) &rest t)
2010 (values (vector hash-table) &optional)
2011 (values t &optional)
2013 (values nil &optional)
2014 (values nil &rest t)
2015 (values sequence &optional)
2016 (values sequence &rest t)
2017 (values list &optional)
2018 (values list &rest t)))))
2021 (let ((i (sb-c::values-type-intersection x y)))
2022 (assert (sb-c::type= i (sb-c::values-type-intersection i x)))
2023 (assert (sb-c::type= i (sb-c::values-type-intersection i y)))))))
2025 370: reader misbehaviour on large-exponent floats
2026 (read-from-string "1.0s1000000000000000000000000000000000000000")
2027 causes the reader to attempt to create a very large bignum (which it
2028 will then attempt to coerce to a rational). While this isn't
2029 completely wrong, it is probably not ideal -- checking the floating
2030 point control word state and then returning the relevant float
2031 (most-positive-short-float or short-float-infinity) or signalling an
2032 error immediately would seem to make more sense.
2034 372: floating-point overflow not signalled on ppc/darwin
2035 The following assertions in float.pure.lisp fail on ppc/darwin
2036 (Mac OS X version 10.3.7):
2037 (assert (raises-error? (scale-float 1.0 most-positive-fixnum)
2038 floating-point-overflow))
2039 (assert (raises-error? (scale-float 1.0d0 (1+ most-positive-fixnum))
2040 floating-point-overflow)))
2041 as the SCALE-FLOAT just returns
2042 #.SB-EXT:SINGLE/DOUBLE-FLOAT-POSITIVE-INFINITY. These tests have been
2043 disabled on Darwin for now.
2045 373: profiling issues on ppc/darwin
2046 The following bit from smoke.impure.lisp fails on ppc/darwin:
2048 (defun profiled-fun ()
2050 (profile profiled-fun)
2051 (loop repeat 100000 do (profiled-fun))
2053 dropping into the debugger with a TYPE-ERROR:
2054 The value -1073741382 is not of type UNSIGNED-BYTE.
2055 The test has been disabled on Darwin till the bug is fixed.
2057 374: BIT-AND problem on ppc/darwin:
2058 The BIT-AND test in bit-vector.impure-cload.lisp results in
2059 fatal error encountered in SBCL pid 8356:
2060 GC invariant lost, file "gc-common.c", line 605
2061 on ppc/darwin. Test disabled for the duration.
2064 (compile nil '(lambda (p1)
2065 (declare (optimize (speed 1) (safety 2) (debug 2) (space 0))
2069 fails on hairy type check in IR2.
2071 1. KEYWORDP is MAYBE-INLINE expanded (before TYPEP-like
2072 transformation could eliminate it).
2074 2. From the only call of KEYWORDP the type of its argument is
2075 derived to be KEYWORD.
2077 2. Type check for P1 is generated; it uses KEYWORDP to perform the
2078 check, and so references the local function; from the KEYWORDP
2079 argument type new CAST to KEYWORD is generated. The compiler
2083 Type deriver for CONJUGATE thinks that it returns an object of the
2084 same type as its argument, which is wrong for such types as (EQL
2087 377: Memory fault error reporting
2088 On those architectures where :C-STACK-IS-CONTROL-STACK is in
2089 *FEATURES*, we handle SIG_MEMORY_FAULT (SEGV or BUS) on an altstack,
2090 so we cannot handle the signal directly (as in interrupt_handle_now())
2091 in the case when the signal comes from some external agent (the user
2092 using kill(1), or a fault in some foreign code, for instance). As
2093 of sbcl-0.8.20.20, this is fixed by calling
2094 arrange_return_to_lisp_function() to a new error-signalling
2095 function, but as a result the error reporting is poor: we cannot
2096 even tell the user at which address the fault occurred. We should
2097 arrange such that arguments can be passed to the function called from
2098 arrange_return_to_lisp_function(), but this looked hard to do in
2099 general without suffering from memory leaks.