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.
97 Currently INSPECT and DESCRIBE do show the values, but showing the
98 names of the bindings would be even nicer.
101 The compiler assumes that any time a function of declared FTYPE
102 doesn't signal an error, its arguments were of the declared type.
103 E.g. compiling and loading
104 (DECLAIM (OPTIMIZE (SAFETY 3)))
105 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
107 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE)) FACTORIAL))
109 (COND ((> (FACTORIAL X) 1.0E6)
110 (FORMAT T "too big~%"))
112 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
114 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
117 will cause the INTEGERP case to be selected, giving bogus output a la
119 This violates the "declarations are assertions" principle.
120 According to the ANSI spec, in the section "System Class FUNCTION",
121 this is a case of "lying to the compiler", but the lying is done
122 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
123 not by the unexpectedly general definition of FACTORIAL. In any case,
124 "declarations are assertions" means that lying to the compiler should
125 cause an error to be signalled, and should not cause a bogus
126 result to be returned. Thus, the compiler should not assume
127 that arbitrary functions check their argument types. (It might
128 make sense to add another flag (CHECKED?) to DEFKNOWN to
129 identify functions which *do* check their argument types.)
130 (Also, verify that the compiler handles declared function
131 return types as assertions.)
134 The definitions of SIGCONTEXT-FLOAT-REGISTER and
135 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
136 supported on FreeBSD because the floating point state is not saved,
137 but at least as of FreeBSD 4.0, the floating point state *is* saved,
138 so they could be supported after all. Very likely
139 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
142 Compiling and loading
143 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
145 then requesting a BACKTRACE at the debugger prompt gives no information
146 about where in the user program the problem occurred.
148 (this is apparently mostly fixed on the SPARC, PPC, and x86 architectures:
149 while giving the backtrace the non-x86 systems complains about "unknown
150 source location: using block start", but apart from that the
151 backtrace seems reasonable. On x86 this is masked by bug 353. See
152 tests/debug.impure.lisp for a test case)
155 Using the pretty-printer from the command prompt gives funny
156 results, apparently because the pretty-printer doesn't know
157 about user's command input, including the user's carriage return
158 that the user, and therefore the pretty-printer thinks that
159 the new output block should start indented 2 or more characters
160 rightward of the correct location.
163 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
164 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
165 crashes SBCL. In general tracing anything which is used in the
166 implementation of TRACE is likely to have the same problem.
169 ANSI says in one place that type declarations can be abbreviated even
170 when the type name is not a symbol, e.g.
171 (DECLAIM ((VECTOR T) *FOOVECTOR*))
172 SBCL doesn't support this. But ANSI says in another place that this
173 isn't allowed. So it's not clear this is a bug after all. (See the
174 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
178 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
179 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
180 the mean of the distribution can be systematically O(0.1%) wrong.
181 Just increasing R-I-E-B is probably not a good solution, since
182 it would decrease efficiency more than is probably necessary. Perhaps
183 using some sort of accept/reject method would be better.
186 Internally the compiler sometimes evaluates
187 (sb-kernel:type/= (specifier-type '*) (specifier-type t))
188 (I stumbled across this when I added an
189 (assert (not (eq type1 *wild-type*)))
190 in the NAMED :SIMPLE-= type method.) '* isn't really a type, and
191 in a type context should probably be translated to T, and so it's
192 probably wrong to ask whether it's equal to the T type and then (using
193 the EQ type comparison in the NAMED :SIMPLE-= type method) return NIL.
194 (I haven't tried to investigate this bug enough to guess whether
195 there might be any user-level symptoms.)
197 In fact, the type system is likely to depend on this inequality not
198 holding... * is not equivalent to T in many cases, such as
199 (VECTOR *) /= (VECTOR T).
202 In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
203 CL), out-of-line structure slot setters are horribly inefficient
204 whenever the type of the slot is declared, because out-of-line
205 structure slot setters are implemented as closures to save space,
206 so the compiler doesn't compile the type test into code, but
207 instead just saves the type in a lexical closure and interprets it
209 To exercise the problem, compile and load
210 (cl:in-package :cl-user)
212 (bar (error "missing") :type bar))
215 (loop (setf (foo-bar *foo*) x)))
217 (defvar *bar* (make-bar))
218 (defvar *foo* (make-foo :bar *bar*))
219 (defvar *setf-foo-bar* #'(setf foo-bar))
221 (loop (funcall *setf-foo-bar* x *foo*)))
222 then run (WASTREL1 *BAR*) or (WASTREL2 *BAR*), hit Ctrl-C, and
223 use BACKTRACE, to see it's spending all essentially all its time
224 in %TYPEP and VALUES-SPECIFIER-TYPE and so forth.
225 One possible solution would be simply to give up on
226 representing structure slot accessors as functions, and represent
227 them as macroexpansions instead. This can be inconvenient for users,
228 but it's not clear that it's worse than trying to help by expanding
229 into a horribly inefficient implementation.
230 As a workaround for the problem, #'(SETF FOO) expressions
231 can be replaced with (EFFICIENT-SETF-FUNCTION FOO), where
232 (defmacro efficient-setf-function (place-function-name)
233 (or #+sbcl (and (sb-int:info :function :accessor-for place-function-name)
234 ;; a workaround for the problem, encouraging the
235 ;; inline expansion of the structure accessor, so
236 ;; that the compiler can optimize its type test
237 (let ((new-value (gensym "NEW-VALUE-"))
238 (structure-value (gensym "STRUCTURE-VALUE-")))
239 `(lambda (,new-value ,structure-value)
240 (setf (,place-function-name ,structure-value)
242 ;; no problem, can just use the ordinary expansion
243 `(function (setf ,place-function-name))))
246 There's apparently a bug in CEILING optimization which caused
247 Douglas Crosher to patch the CMU CL version. Martin Atzmueller
248 applied the patches to SBCL and they didn't seem to cause problems
249 (as reported sbcl-devel 2001-05-04). However, since the patches
250 modify nontrivial code which was apparently written incorrectly
251 the first time around, until regression tests are written I'm not
252 comfortable merging the patches in the CVS version of SBCL.
255 (TIME (ROOM T)) reports more than 200 Mbytes consed even for
256 a clean, just-started SBCL system. And it seems to be right:
257 (ROOM T) can bring a small computer to its knees for a *long*
258 time trying to GC afterwards. Surely there's some more economical
259 way to implement (ROOM T).
261 Daniel Barlow doesn't know what fixed this, but observes that it
262 doesn't seem to be the case in 0.8.7.3 any more. Instead, (ROOM T)
263 in a fresh SBCL causes
265 debugger invoked on a SB-INT:BUG in thread 5911:
266 failed AVER: "(SAP= CURRENT END)"
268 unless a GC has happened beforehand.
271 When the compiler inline expands functions, it may be that different
272 kinds of return values are generated from different code branches.
273 E.g. an inline expansion of POSITION generates integer results
274 from one branch, and NIL results from another. When that inline
275 expansion is used in a context where only one of those results
278 (aref *a1* (position x *a2*)))
279 and the compiler can't prove that the unacceptable branch is
280 never taken, then bogus type mismatch warnings can be generated.
281 If you need to suppress the type mismatch warnings, you can
282 suppress the inline expansion,
284 #+sbcl (declare (notinline position)) ; to suppress bug 117 bogowarnings
285 (aref *a1* (position x *a2*)))
286 or, sometimes, suppress them by declaring the result to be of an
289 (aref *a1* (the integer (position x *a2*))))
291 This is not a new compiler problem in 0.7.0, but the new compiler
292 transforms for FIND, POSITION, FIND-IF, and POSITION-IF make it
293 more conspicuous. If you don't need performance from these functions,
294 and the bogus warnings are a nuisance for you, you can return to
295 your pre-0.7.0 state of grace with
296 #+sbcl (declaim (notinline find position find-if position-if)) ; bug 117..
301 As of version 0.pre7.14, SBCL's implementation of MACROLET makes
302 the entire lexical environment at the point of MACROLET available
303 in the bodies of the macroexpander functions. In particular, it
304 allows the function bodies (which run at compile time) to try to
305 access lexical variables (which are only defined at runtime).
306 It doesn't even issue a warning, which is bad.
308 The SBCL behavior arguably conforms to the ANSI spec (since the
309 spec says that the behavior is undefined, ergo anything conforms).
310 However, it would be better to issue a compile-time error.
311 Unfortunately I (WHN) don't see any simple way to detect this
312 condition in order to issue such an error, so for the meantime
313 SBCL just does this weird broken "conforming" thing.
315 The ANSI standard says, in the definition of the special operator
317 The macro-expansion functions defined by MACROLET are defined
318 in the lexical environment in which the MACROLET form appears.
319 Declarations and MACROLET and SYMBOL-MACROLET definitions affect
320 the local macro definitions in a MACROLET, but the consequences
321 are undefined if the local macro definitions reference any
322 local variable or function bindings that are visible in that
324 Then it seems to contradict itself by giving the example
326 (macrolet ((fudge (z)
327 ;The parameters x and flag are not accessible
328 ; at this point; a reference to flag would be to
329 ; the global variable of that name.
330 ` (if flag (* ,z ,z) ,z)))
331 ;The parameters x and flag are accessible here.
335 The comment "a reference to flag would be to the global variable
336 of the same name" sounds like good behavior for the system to have.
337 but actual specification quoted above says that the actual behavior
340 (Since 0.7.8.23 macroexpanders are defined in a restricted version
341 of the lexical environment, containing no lexical variables and
342 functions, which seems to conform to ANSI and CLtL2, but signalling
343 a STYLE-WARNING for references to variables similar to locals might
347 Ideally, uninterning a symbol would allow it, and its associated
348 FDEFINITION and PROCLAIM data, to be reclaimed by the GC. However,
349 at least as of sbcl-0.7.0, this isn't the case. Information about
350 FDEFINITIONs and PROCLAIMed properties is stored in globaldb.lisp
351 essentially in ordinary (non-weak) hash tables keyed by symbols.
352 Thus, once a system has an entry in this system, it tends to live
353 forever, even when it is uninterned and all other references to it
357 (reported by Jesse Bouwman 2001-10-24 through the unfortunately
358 prominent SourceForge web/db bug tracking system, which is
359 unfortunately not a reliable way to get a timely response from
360 the SBCL maintainers)
361 In the course of trying to build a test case for an
362 application error, I encountered this behavior:
363 If you start up sbcl, and then lay on CTRL-C for a
364 minute or two, the lisp process will eventually say:
365 %PRIMITIVE HALT called; the party is over.
366 and throw you into the monitor. If I start up lisp,
367 attach to the process with strace, and then do the same
368 (abusive) thing, I get instead:
369 access failure in heap page not marked as write-protected
370 and the monitor again. I don't know enough to have the
371 faintest idea of what is going on here.
372 This is with sbcl 6.12, uname -a reports:
373 Linux prep 2.2.19 #4 SMP Tue Apr 24 13:59:52 CDT 2001 i686 unknown
374 I (WHN) have verified that the same thing occurs on sbcl-0.pre7.141
375 under OpenBSD 2.9 on my X86 laptop. Do be patient when you try it:
376 it took more than two minutes (but less than five) for me.
380 ANSI allows types `(COMPLEX ,FOO) to use very hairy values for
381 FOO, e.g. (COMPLEX (AND REAL (SATISFIES ODDP))). The old CMU CL
382 COMPLEX implementation didn't deal with this, and hasn't been
383 upgraded to do so. (This doesn't seem to be a high priority
384 conformance problem, since seems hard to construct useful code
387 [ partially fixed by CSR in 0.8.17.17 because of a PFD ansi-tests
388 report that (COMPLEX RATIO) was failing; still failing on types of
389 the form (AND NUMBER (SATISFIES REALP) (SATISFIES ZEROP)). ]
391 b. (fixed in 0.8.3.43)
394 Floating point errors are reported poorly. E.g. on x86 OpenBSD
397 debugger invoked on condition of type SB-KERNEL:FLOATING-POINT-EXCEPTION:
398 An arithmetic error SB-KERNEL:FLOATING-POINT-EXCEPTION was signalled.
399 No traps are enabled? How can this be?
400 It should be possible to be much more specific (overflow, division
401 by zero, etc.) and of course the "How can this be?" should be fixable.
403 See also bugs #45.c and #183
406 (reported by Robert E. Brown 2002-04-16)
407 When a function is called with too few arguments, causing the
408 debugger to be entered, the uninitialized slots in the bad call frame
409 seem to cause GCish problems, being interpreted as tagged data even
410 though they're not. In particular, executing ROOM in the
411 debugger at that point causes AVER failures:
414 * (lisp-implementation-version)
420 failed AVER: "(SAP= CURRENT END)"
421 (Christophe Rhodes reports that this doesn't occur on the SPARC, which
422 isn't too surprising since there are many differences in stack
423 implementation and GC conservatism between the X86 and other ports.)
425 This is probably the same bug as 216
428 The compiler sometimes tries to constant-fold expressions before
429 it checks to see whether they can be reached. This can lead to
430 bogus warnings about errors in the constant folding, e.g. in code
433 (WRITE-STRING (> X 0) "+" "0"))
434 compiled in a context where the compiler can prove that X is NIL,
435 and the compiler complains that (> X 0) causes a type error because
436 NIL isn't a valid argument to #'>. Until sbcl-0.7.4.10 or so this
437 caused a full WARNING, which made the bug really annoying because then
438 COMPILE and COMPILE-FILE returned FAILURE-P=T for perfectly legal
439 code. Since then the warning has been downgraded to STYLE-WARNING,
440 so it's still a bug but at least it's a little less annoying.
442 183: "IEEE floating point issues"
443 Even where floating point handling is being dealt with relatively
444 well (as of sbcl-0.7.5, on sparc/sunos and alpha; see bug #146), the
445 accrued-exceptions and current-exceptions part of the fp control
446 word don't seem to bear much relation to reality. E.g. on
450 debugger invoked on condition of type DIVISION-BY-ZERO:
451 arithmetic error DIVISION-BY-ZERO signalled
452 0] (sb-vm::get-floating-point-modes)
454 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
455 :ROUNDING-MODE :NEAREST
456 :CURRENT-EXCEPTIONS NIL
457 :ACCRUED-EXCEPTIONS (:INEXACT)
460 * (sb-vm::get-floating-point-modes)
461 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
462 :ROUNDING-MODE :NEAREST
463 :CURRENT-EXCEPTIONS (:INEXACT)
464 :ACCRUED-EXCEPTIONS (:INEXACT)
467 188: "compiler performance fiasco involving type inference and UNION-TYPE"
471 (declare (optimize (safety 3)))
472 (declare (optimize (compilation-speed 2)))
473 (declare (optimize (speed 1) (debug 1) (space 1)))
475 (declare (type (integer 0) start))
476 (print (incf start 22))
477 (print (incf start 26))
478 (print (incf start 28)))
480 (declare (type (integer 0) start))
481 (print (incf start 22))
482 (print (incf start 26)))
484 (declare (type (integer 0) start))
485 (print (incf start 22))
486 (print (incf start 26))))))
488 This example could be solved with clever enough constraint
489 propagation or with SSA, but consider
494 The careful type of X is {2k} :-(. Is it really important to be
495 able to work with unions of many intervals?
497 191: "Miscellaneous PCL deficiencies"
498 (reported by Alexey Dejneka sbcl-devel 2002-08-04)
499 a. DEFCLASS does not inform the compiler about generated
500 functions. Compiling a file with
504 (WITH-SLOTS (A-CLASS-X) A
506 results in a STYLE-WARNING:
508 SB-SLOT-ACCESSOR-NAME::|COMMON-LISP-USER A-CLASS-X slot READER|
510 APD's fix for this was checked in to sbcl-0.7.6.20, but Pierre
511 Mai points out that the declamation of functions is in fact
512 incorrect in some cases (most notably for structure
513 classes). This means that at present erroneous attempts to use
514 WITH-SLOTS and the like on classes with metaclass STRUCTURE-CLASS
515 won't get the corresponding STYLE-WARNING.
516 c. (fixed in 0.8.4.23)
518 201: "Incautious type inference from compound types"
519 a. (reported by APD sbcl-devel 2002-09-17)
521 (LET ((Y (CAR (THE (CONS INTEGER *) X))))
523 (FORMAT NIL "~S IS ~S, Y = ~S"
530 (FOO ' (1 . 2)) => "NIL IS INTEGER, Y = 1"
534 (declare (type (array * (4 4)) x))
536 (setq x (make-array '(4 4)))
537 (adjust-array y '(3 5))
538 (= (array-dimension y 0) (eval `(array-dimension ,y 0)))))
540 * (foo (make-array '(4 4) :adjustable t))
543 205: "environment issues in cross compiler"
544 (These bugs have no impact on user code, but should be fixed or
546 a. Macroexpanders introduced with MACROLET are defined in the null
548 b. The body of (EVAL-WHEN (:COMPILE-TOPLEVEL) ...) is evaluated in
549 the null lexical environment.
550 c. The cross-compiler cannot inline functions defined in a non-null
553 206: ":SB-FLUID feature broken"
554 (reported by Antonio Martinez-Shotton sbcl-devel 2002-10-07)
555 Enabling :SB-FLUID in the target-features list in sbcl-0.7.8 breaks
558 207: "poorly distributed SXHASH results for compound data"
559 SBCL's SXHASH could probably try a little harder. ANSI: "the
560 intent is that an implementation should make a good-faith
561 effort to produce hash-codes that are well distributed
562 within the range of non-negative fixnums". But
563 (let ((hits (make-hash-table)))
566 (let* ((ij (cons i j))
567 (newlist (push ij (gethash (sxhash ij) hits))))
569 (format t "~&collision: ~S~%" newlist))))))
570 reports lots of collisions in sbcl-0.7.8. A stronger MIX function
571 would be an obvious way of fix. Maybe it would be acceptably efficient
572 to redo MIX using a lookup into a 256-entry s-box containing
573 29-bit pseudorandom numbers?
575 211: "keywords processing"
576 a. :ALLOW-OTHER-KEYS T should allow a function to receive an odd
577 number of keyword arguments.
579 212: "Sequence functions and circular arguments"
580 COERCE, MERGE and CONCATENATE go into an infinite loop when given
581 circular arguments; it would be good for the user if they could be
582 given an error instead (ANSI 17.1.1 allows this behaviour on the part
583 of the implementation, as conforming code cannot give non-proper
584 sequences to these functions. MAP also has this problem (and
585 solution), though arguably the convenience of being able to do
586 (MAP 'LIST '+ FOO '#1=(1 . #1#))
587 might be classed as more important (though signalling an error when
588 all of the arguments are circular is probably desireable).
590 213: "Sequence functions and type checking"
591 b. MAP, when given a type argument that is SUBTYPEP LIST, does not
592 check that it will return a sequence of the given type. Fixing
593 it along the same lines as the others (cf. work done around
594 sbcl-0.7.8.45) is possible, but doing so efficiently didn't look
595 entirely straightforward.
596 c. All of these functions will silently accept a type of the form
598 whether or not the return value is of this type. This is
599 probably permitted by ANSI (see "Exceptional Situations" under
600 ANSI MAKE-SEQUENCE), but the DERIVE-TYPE mechanism does not
601 know about this escape clause, so code of the form
602 (INTEGERP (CAR (MAKE-SEQUENCE '(CONS INTEGER *) 2)))
603 can erroneously return T.
605 215: ":TEST-NOT handling by functions"
606 a. FIND and POSITION currently signal errors when given non-NIL for
607 both their :TEST and (deprecated) :TEST-NOT arguments, but by
608 ANSI 17.2 "the consequences are unspecified", which by ANSI 1.4.2
609 means that the effect is "unpredictable but harmless". It's not
610 clear what that actually means; it may preclude conforming
611 implementations from signalling errors.
612 b. COUNT, REMOVE and the like give priority to a :TEST-NOT argument
613 when conflict occurs. As a quality of implementation issue, it
614 might be preferable to treat :TEST and :TEST-NOT as being in some
615 sense the same &KEY, and effectively take the first test function in
617 c. Again, a quality of implementation issue: it would be good to issue a
618 STYLE-WARNING at compile-time for calls with :TEST-NOT, and a
619 WARNING for calls with both :TEST and :TEST-NOT; possibly this
620 latter should be WARNed about at execute-time too.
622 216: "debugger confused by frames with invalid number of arguments"
623 In sbcl-0.7.8.51, executing e.g. (VECTOR-PUSH-EXTEND T), BACKTRACE, Q
624 leaves the system confused, enough so that (QUIT) no longer works.
625 It's as though the process of working with the uninitialized slot in
626 the bad VECTOR-PUSH-EXTEND frame causes GC problems, though that may
627 not be the actual problem. (CMU CL 18c doesn't have problems with this.)
629 This is probably the same bug as 162
631 217: "Bad type operations with FUNCTION types"
634 * (values-type-union (specifier-type '(function (base-char)))
635 (specifier-type '(function (integer))))
637 #<FUN-TYPE (FUNCTION (BASE-CHAR) *)>
639 It causes insertion of wrong type assertions into generated
643 (let ((f (etypecase x
644 (character #'write-char)
645 (integer #'write-byte))))
648 (character (write-char x s))
649 (integer (write-byte x s)))))
651 Then (FOO #\1 *STANDARD-OUTPUT*) signals type error.
653 (In 0.7.9.1 the result type is (FUNCTION * *), so Python does not
654 produce invalid code, but type checking is not accurate.)
656 233: bugs in constraint propagation
658 (declaim (optimize (speed 2) (safety 3)))
660 (if (typep (prog1 x (setq x y)) 'double-float)
663 (foo 1d0 5) => segmentation violation
665 235: "type system and inline expansion"
667 (declaim (ftype (function (cons) number) acc))
668 (declaim (inline acc))
670 (the number (car c)))
673 (values (locally (declare (optimize (safety 0)))
675 (locally (declare (optimize (safety 3)))
678 (foo '(nil) '(t)) => NIL, T.
680 237: "Environment arguments to type functions"
681 a. Functions SUBTYPEP, TYPEP, UPGRADED-ARRAY-ELEMENT-TYPE, and
682 UPGRADED-COMPLEX-PART-TYPE now have an optional environment
683 argument, but they ignore it completely. This is almost
684 certainly not correct.
685 b. Also, the compiler's optimizers for TYPEP have not been informed
686 about the new argument; consequently, they will not transform
687 calls of the form (TYPEP 1 'INTEGER NIL), even though this is
688 just as optimizeable as (TYPEP 1 'INTEGER).
690 238: "REPL compiler overenthusiasm for CLOS code"
692 * (defclass foo () ())
693 * (defmethod bar ((x foo) (foo foo)) (call-next-method))
694 causes approximately 100 lines of code deletion notes. Some
695 discussion on this issue happened under the title 'Three "interesting"
696 bugs in PCL', resulting in a fix for this oververbosity from the
697 compiler proper; however, the problem persists in the interactor
698 because the notion of original source is not preserved: for the
699 compiler, the original source of the above expression is (DEFMETHOD
700 BAR ((X FOO) (FOO FOO)) (CALL-NEXT-METHOD)), while by the time the
701 compiler gets its hands on the code needing compilation from the REPL,
702 it has been macroexpanded several times.
704 A symptom of the same underlying problem, reported by Tony Martinez:
706 (with-input-from-string (*query-io* " no")
708 (simple-type-error () 'error))
710 ; (SB-KERNEL:FLOAT-WAIT)
712 ; note: deleting unreachable code
713 ; compilation unit finished
716 242: "WRITE-SEQUENCE suboptimality"
717 (observed from clx performance)
718 In sbcl-0.7.13, WRITE-SEQUENCE of a sequence of type
719 (SIMPLE-ARRAY (UNSIGNED-BYTE 8) (*)) on a stream with element-type
720 (UNSIGNED-BYTE 8) will write to the stream one byte at a time,
721 rather than writing the sequence in one go, leading to severe
722 performance degradation.
723 As of sbcl-0.9.0.36, this is solved for fd-streams, so is less of a
724 problem in practice. (Fully fixing this would require adding a
725 ansi-stream-n-bout slot and associated methods to write a byte
726 sequence to ansi-stream, similar to the existing ansi-stream-sout
729 243: "STYLE-WARNING overenthusiasm for unused variables"
730 (observed from clx compilation)
731 In sbcl-0.7.14, in the presence of the macros
732 (DEFMACRO FOO (X) `(BAR ,X))
733 (DEFMACRO BAR (X) (DECLARE (IGNORABLE X)) 'NIL)
734 somewhat surprising style warnings are emitted for
735 (COMPILE NIL '(LAMBDA (Y) (FOO Y))):
737 ; (LAMBDA (Y) (FOO Y))
739 ; caught STYLE-WARNING:
740 ; The variable Y is defined but never used.
742 245: bugs in disassembler
743 b. On X86 operand size prefix is not recognized.
746 (defun foo (&key (a :x))
750 does not cause a warning. (BTW: old SBCL issued a warning, but for a
751 function, which was never called!)
754 Compiler does not emit warnings for
756 a. (lambda () (svref (make-array 8 :adjustable t) 1))
759 (list (let ((y (the real x)))
760 (unless (floatp y) (error ""))
765 (declare (optimize (debug 0)))
766 (declare (type vector x))
767 (list (fill-pointer x)
771 Complex array type does not have corresponding type specifier.
773 This is a problem because the compiler emits optimization notes when
774 you use a non-simple array, and without a type specifier for hairy
775 array types, there's no good way to tell it you're doing it
776 intentionally so that it should shut up and just compile the code.
778 Another problem is confusing error message "asserted type ARRAY
779 conflicts with derived type (VALUES SIMPLE-VECTOR &OPTIONAL)" during
780 compiling (LAMBDA (V) (VALUES (SVREF V 0) (VECTOR-POP V))).
782 The last problem is that when type assertions are converted to type
783 checks, types are represented with type specifiers, so we could lose
784 complex attribute. (Now this is probably not important, because
785 currently checks for complex arrays seem to be performed by
789 (compile nil '(lambda () (aref (make-array 0) 0))) compiles without
790 warning. Analogous cases with the index and length being equal and
791 greater than 0 are warned for; the problem here seems to be that the
792 type required for an array reference of this type is (INTEGER 0 (0))
793 which is canonicalized to NIL.
798 (t1 (specifier-type s)))
799 (eval `(defstruct ,s))
800 (type= t1 (specifier-type s)))
805 b. The same for CSUBTYPEP.
807 262: "yet another bug in inline expansion of local functions"
808 During inline expansion of a local function Python can try to
809 reference optimized away objects (functions, variables, CTRANs from
810 tags and blocks), which later may lead to problems. Some of the
811 cases are worked around by forbidding expansion in such cases, but
812 the better way would be to reimplement inline expansion by copying
816 David Lichteblau provided (sbcl-devel 2003-06-01) a patch to fix
817 behaviour of streams with element-type (SIGNED-BYTE 8). The patch
818 looks reasonable, if not obviously correct; however, it caused the
819 PPC/Linux port to segfault during warm-init while loading
820 src/pcl/std-class.fasl. A workaround patch was made, but it would
821 be nice to understand why the first patch caused problems, and to
822 fix the cause if possible.
824 268: "wrong free declaration scope"
825 The following code must signal type error:
827 (locally (declare (optimize (safety 3)))
828 (flet ((foo (x &optional (y (car x)))
829 (declare (optimize (safety 0)))
831 (funcall (eval #'foo) 1)))
834 In the following function constraint propagator optimizes nothing:
837 (declare (integer x))
838 (declare (optimize speed))
846 Compilation of the following two forms causes "X is unbound" error:
848 (symbol-macrolet ((x pi))
849 (macrolet ((foo (y) (+ x y)))
850 (declaim (inline bar))
856 (See (COERCE (CDR X) 'FUNCTION) in IR1-CONVERT-INLINE-LAMBDA.)
859 CLHS says that type declaration of a symbol macro should not affect
860 its expansion, but in SBCL it does. (If you like magic and want to
861 fix it, don't forget to change all uses of MACROEXPAND to
865 The following code (taken from CLOCC) takes a lot of time to compile:
868 (declare (type (integer 0 #.large-constant) n))
871 (fixed in 0.8.2.51, but a test case would be good)
874 b. The same as in a., but using MULTIPLE-VALUE-SETQ instead of SETQ.
876 (defmethod faa ((*faa* double-float))
877 (set '*faa* (when (< *faa* 0) (- *faa*)))
879 (faa 1d0) => type error
884 (declare (optimize speed))
885 (loop for i of-type (integer 0) from 0 by 2 below 10
888 uses generic arithmetic.
890 b. (fixed in 0.8.3.6)
892 279: type propagation error -- correctly inferred type goes astray?
893 In sbcl-0.8.3 and sbcl-0.8.1.47, the warning
894 The binding of ABS-FOO is a (VALUES (INTEGER 0 0)
895 &OPTIONAL), not a (INTEGER 1 536870911)
896 is emitted when compiling this file:
897 (declaim (ftype (function ((integer 0 #.most-positive-fixnum))
898 (integer #.most-negative-fixnum 0))
903 (let* (;; Uncomment this for a type mismatch warning indicating
904 ;; that the type of (FOO X) is correctly understood.
905 #+nil (fs-foo (float-sign (foo x)))
906 ;; Uncomment this for a type mismatch warning
907 ;; indicating that the type of (ABS (FOO X)) is
908 ;; correctly understood.
909 #+nil (fs-abs-foo (float-sign (abs (foo x))))
910 ;; something wrong with this one though
911 (abs-foo (abs (foo x))))
912 (declare (type (integer 1 100) abs-foo))
917 283: Thread safety: libc functions
918 There are places that we call unsafe-for-threading libc functions
919 that we should find alternatives for, or put locks around. Known or
920 strongly suspected problems, as of 0.8.3.10: please update this
921 bug instead of creating new ones
923 localtime() - called for timezone calculations in code/time.lisp
925 284: Thread safety: special variables
926 There are lots of special variables in SBCL, and I feel sure that at
927 least some of them are indicative of potentially thread-unsafe
928 parts of the system. See doc/internals/notes/threading-specials
930 286: "recursive known functions"
931 Self-call recognition conflicts with known function
932 recognition. Currently cross compiler and target COMPILE do not
933 recognize recursion, and in target compiler it can be disabled. We
934 can always disable it for known functions with RECURSIVE attribute,
935 but there remains a possibility of a function with a
936 (tail)-recursive simplification pass and transforms/VOPs for base
939 287: PPC/Linux miscompilation or corruption in first GC
940 When the runtime is compiled with -O3 on certain PPC/Linux machines, a
941 segmentation fault is reported at the point of first triggered GC,
942 during the compilation of DEFSTRUCT WRAPPER. As a temporary workaround,
943 the runtime is no longer compiled with -O3 on PPC/Linux, but it is likely
944 that this merely obscures, not solves, the underlying problem; as and when
945 underlying problems are fixed, it would be worth trying again to provoke
948 288: fundamental cross-compilation issues (from old UGLINESS file)
949 Using host floating point numbers to represent target floating point
950 numbers, or host characters to represent target characters, is
951 theoretically shaky. (The characters are OK as long as the characters
952 are in the ANSI-guaranteed character set, though, so they aren't a
953 real problem as long as the sources don't need anything but that;
954 the floats are a real problem.)
956 289: "type checking and source-transforms"
958 (block nil (let () (funcall #'+ (eval 'nil) (eval '1) (return :good))))
961 Our policy is to check argument types at the moment of a call. It
962 disagrees with ANSI, which says that type assertions are put
963 immediately onto argument expressions, but is easier to implement in
964 IR1 and is more compatible to type inference, inline expansion,
965 etc. IR1-transforms automatically keep this policy, but source
966 transforms for associative functions (such as +), being applied
967 during IR1-convertion, do not. It may be tolerable for direct calls
968 (+ x y z), but for (FUNCALL #'+ x y z) it is non-conformant.
970 b. Another aspect of this problem is efficiency. [x y + z +]
971 requires less registers than [x y z + +]. This transformation is
972 currently performed with source transforms, but it would be good to
973 also perform it in IR1 optimization phase.
975 290: Alpha floating point and denormalized traps
976 In SBCL 0.8.3.6x on the alpha, we work around what appears to be a
977 hardware or kernel deficiency: the status of the enable/disable
978 denormalized-float traps bit seems to be ambiguous; by the time we
979 get to os_restore_fp_control after a trap, denormalized traps seem
980 to be enabled. Since we don't want a trap every time someone uses a
981 denormalized float, in general, we mask out that bit when we restore
982 the control word; however, this clobbers any change the user might
986 (reported by Adam Warner, sbcl-devel 2003-09-23)
988 The --load toplevel argument does not perform any sanitization of its
989 argument. As a result, files with Lisp pathname pattern characters
990 (#\* or #\?, for instance) or quotation marks can cause the system
991 to perform arbitrary behaviour.
994 LOOP with non-constant arithmetic step clauses suffers from overzealous
995 type constraint: code of the form
996 (loop for d of-type double-float from 0d0 to 10d0 by x collect d)
997 compiles to a type restriction on X of (AND DOUBLE-FLOAT (REAL
998 (0))). However, an integral value of X should be legal, because
999 successive adds of integers to double-floats produces double-floats,
1000 so none of the type restrictions in the code is violated.
1002 300: (reported by Peter Graves) Function PEEK-CHAR checks PEEK-TYPE
1003 argument type only after having read a character. This is caused
1004 with EXPLICIT-CHECK attribute in DEFKNOWN. The similar problem
1005 exists with =, /=, <, >, <=, >=. They were fixed, but it is probably
1006 less error prone to have EXPLICIT-CHECK be a local declaration,
1007 being put into the definition, instead of an attribute being kept in
1008 a separate file; maybe also put it into SB-EXT?
1010 301: ARRAY-SIMPLE-=-TYPE-METHOD breaks on corner cases which can arise
1011 in NOTE-ASSUMED-TYPES
1012 In sbcl-0.8.7.32, compiling the file
1014 (declare (type integer x))
1015 (declare (type (vector (or hash-table bit)) y))
1018 (declare (type integer x))
1019 (declare (type (simple-array base (2)) y))
1022 failed AVER: "(NOT (AND (NOT EQUALP) CERTAINP))"
1024 303: "nonlinear LVARs" (aka MISC.293)
1026 (multiple-value-call #'list
1028 (multiple-value-prog1
1029 (eval '(values :a :b :c))
1035 (throw 'bar (values 3 4)))))))))))
1037 (BUU 1) returns garbage.
1039 The problem is that both EVALs sequentially write to the same LVAR.
1041 306: "Imprecise unions of array types"
1043 (declare (optimize speed)
1044 (type (or (array cons) (array vector)) x))
1046 (foo #((0))) => TYPE-ERROR
1053 ,@(loop for x across sb-vm:*specialized-array-element-type-properties*
1054 collect `(array ,(sb-vm:saetp-specifier x)))))
1055 => NIL, T (when it should be T, T)
1057 309: "Dubious values for implementation limits"
1058 (reported by Bruno Haible sbcl-devel "Incorrect value of
1059 multiple-values-limit" 2004-04-19)
1060 (values-list (make-list 1000000)), on x86/linux, signals a stack
1061 exhaustion condition, despite MULTIPLE-VALUES-LIMIT being
1062 significantly larger than 1000000. There are probably similar
1063 dubious values for CALL-ARGUMENTS-LIMIT (see cmucl-help/cmucl-imp
1064 around the same time regarding a call to LIST on sparc with 1000
1065 arguments) and other implementation limit constants.
1067 314: "LOOP :INITIALLY clauses and scope of initializers"
1068 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1069 test suite, originally by Thomas F. Burdick.
1070 ;; <http://www.lisp.org/HyperSpec/Body/sec_6-1-7-2.html>
1071 ;; According to the HyperSpec 6.1.2.1.4, in for-as-equals-then, var is
1072 ;; initialized to the result of evaluating form1. 6.1.7.2 says that
1073 ;; initially clauses are evaluated in the loop prologue, which precedes all
1074 ;; loop code except for the initial settings provided by with, for, or as.
1075 (loop :for x = 0 :then (1+ x)
1076 :for y = (1+ x) :then (ash y 1)
1077 :for z :across #(1 3 9 27 81 243)
1079 :initially (assert (zerop x)) :initially (assert (= 2 w))
1080 :until (>= w 100) :collect w)
1081 Expected: (2 6 15 38)
1084 318: "stack overflow in compiler warning with redefined class"
1085 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1088 (setf (find-class 'foo) nil)
1089 (defstruct foo slot-1)
1090 This used to give a stack overflow from within the printer, which has
1091 been fixed as of 0.8.16.11. Current result:
1093 ; can't compile TYPEP of anonymous or undefined class:
1094 ; #<SB-KERNEL:STRUCTURE-CLASSOID FOO>
1096 debugger invoked on a TYPE-ERROR in thread 19973:
1097 The value NIL is not of type FUNCTION.
1099 CSR notes: it's not really clear what it should give: is (SETF FIND-CLASS)
1100 meant to be enough to delete structure classes from the system?
1102 319: "backquote with comma inside array"
1103 reported by Bruno Haible sbcl-devel "various SBCL bugs" from CLISP
1105 (read-from-string "`#1A(1 2 ,(+ 2 2) 4)")
1107 #(1 2 ((SB-IMPL::|,|) + 2 2) 4)
1108 which probably isn't intentional.
1110 324: "STREAMs and :ELEMENT-TYPE with large bytesize"
1111 In theory, (open foo :element-type '(unsigned-byte <x>)) should work
1112 for all positive integral <x>. At present, it only works for <x> up
1113 to about 1024 (and similarly for signed-byte), so
1114 (open "/dev/zero" :element-type '(unsigned-byte 1025))
1115 gives an error in sbcl-0.8.10.
1117 325: "CLOSE :ABORT T on supeseding streams"
1118 Closing a stream opened with :IF-EXISTS :SUPERSEDE with :ABORT T leaves no
1119 file on disk, even if one existed before opening.
1121 The illegality of this is not crystal clear, as the ANSI dictionary
1122 entry for CLOSE says that when :ABORT is T superseded files are not
1123 superseded (ie. the original should be restored), whereas the OPEN
1124 entry says about :IF-EXISTS :SUPERSEDE "If possible, the
1125 implementation should not destroy the old file until the new stream
1126 is closed." -- implying that even though undesirable, early deletion
1127 is legal. Restoring the original would none the less be the polite
1130 326: "*PRINT-CIRCLE* crosstalk between streams"
1131 In sbcl-0.8.10.48 it's possible for *PRINT-CIRCLE* references to be
1132 mixed between streams when output operations are intermingled closely
1133 enough (as by doing output on S2 from within (PRINT-OBJECT X S1) in the
1134 test case below), so that e.g. the references #2# appears on a stream
1135 with no preceding #2= on that stream to define it (because the #2= was
1136 sent to another stream).
1137 (cl:in-package :cl-user)
1138 (defstruct foo index)
1139 (defparameter *foo* (make-foo :index 4))
1141 (defparameter *bar* (make-bar))
1142 (defparameter *tangle* (list *foo* *bar* *foo*))
1143 (defmethod print-object ((foo foo) stream)
1144 (let ((index (foo-index foo)))
1145 (format *trace-output*
1146 "~&-$- emitting FOO ~D, ambient *BAR*=~S~%"
1148 (format stream "[FOO ~D]" index))
1150 (let ((tsos (make-string-output-stream))
1151 (ssos (make-string-output-stream)))
1152 (let ((*print-circle* t)
1153 (*trace-output* tsos)
1154 (*standard-output* ssos))
1155 (prin1 *tangle* *standard-output*))
1156 (let ((string (get-output-stream-string ssos)))
1157 (unless (string= string "(#1=[FOO 4] #S(BAR) #1#)")
1158 ;; In sbcl-0.8.10.48 STRING was "(#1=[FOO 4] #2# #1#)".:-(
1159 (error "oops: ~S" string)))))
1160 It might be straightforward to fix this by turning the
1161 *CIRCULARITY-HASH-TABLE* and *CIRCULARITY-COUNTER* variables into
1162 per-stream slots, but (1) it would probably be sort of messy faking
1163 up the special variable binding semantics using UNWIND-PROTECT and
1164 (2) it might be sort of a pain to test that no other bugs had been
1167 328: "Profiling generic functions", transplanted from #241
1168 (from tonyms on #lisp IRC 2003-02-25)
1169 In sbcl-0.7.12.55, typing
1170 (defclass foo () ((bar :accessor foo-bar)))
1173 (defclass foo () ((bar :accessor foo-bar)))
1174 gives the error message
1175 "#:FOO-BAR already names an ordinary function or a macro."
1177 Problem: when a generic function is profiled, it appears as an ordinary
1178 function to PCL. (Remembering the uninterned accessor is OK, as the
1179 redefinition must be able to remove old accessors from their generic
1182 329: "Sequential class redefinition"
1183 reported by Bruno Haible:
1184 (defclass reactor () ((max-temp :initform 10000000)))
1185 (defvar *r1* (make-instance 'reactor))
1186 (defvar *r2* (make-instance 'reactor))
1187 (slot-value *r1* 'max-temp)
1188 (slot-value *r2* 'max-temp)
1189 (defclass reactor () ((uptime :initform 0)))
1190 (slot-value *r1* 'uptime)
1191 (defclass reactor () ((uptime :initform 0) (max-temp :initform 10000)))
1192 (slot-value *r1* 'max-temp) ; => 10000
1193 (slot-value *r2* 'max-temp) ; => 10000000 oops...
1196 The method effective when the wrapper is obsoleted can be saved
1197 in the wrapper, and then to update the instance just run through
1198 all the old wrappers in order from oldest to newest.
1200 332: "fasl stack inconsistency in structure redefinition"
1201 (reported by Tim Daly Jr sbcl-devel 2004-05-06)
1202 Even though structure redefinition is undefined by the standard, the
1203 following behaviour is suboptimal: running
1204 (defun stimulate-sbcl ()
1205 (let ((filename (format nil "/tmp/~A.lisp" (gensym))))
1206 ;;create a file which redefines a structure incompatibly
1207 (with-open-file (f filename :direction :output :if-exists :supersede)
1208 (print '(defstruct astruct foo) f)
1209 (print '(defstruct astruct foo bar) f))
1210 ;;compile and load the file, then invoke the continue restart on
1211 ;;the structure redefinition error
1212 (handler-bind ((error (lambda (c) (continue c))))
1213 (load (compile-file filename)))))
1215 and choosing the CONTINUE restart yields the message
1216 debugger invoked on a SB-INT:BUG in thread 27726:
1217 fasl stack not empty when it should be
1219 336: "slot-definitions must retain the generic functions of accessors"
1220 reported by Tony Martinez:
1221 (defclass foo () ((bar :reader foo-bar)))
1222 (defun foo-bar (x) x)
1223 (defclass foo () ((bar :reader get-bar))) ; => error, should work
1225 Note: just punting the accessor removal if the fdefinition
1226 is not a generic function is not enough:
1228 (defclass foo () ((bar :reader foo-bar)))
1229 (defvar *reader* #'foo-bar)
1230 (defun foo-bar (x) x)
1231 (defclass foo () ((bar :initform 'ok :reader get-bar)))
1232 (funcall *reader* (make-instance 'foo)) ; should be an error, since
1233 ; the method must be removed
1234 ; by the class redefinition
1236 Fixing this should also fix a subset of #328 -- update the
1237 description with a new test-case then.
1239 337: MAKE-METHOD and user-defined method classes
1240 (reported by Bruno Haible sbcl-devel 2004-06-11)
1244 (defclass user-method (standard-method) (myslot))
1245 (defmacro def-user-method (name &rest rest)
1246 (let* ((lambdalist-position (position-if #'listp rest))
1247 (qualifiers (subseq rest 0 lambdalist-position))
1248 (lambdalist (elt rest lambdalist-position))
1249 (body (subseq rest (+ lambdalist-position 1)))
1251 (subseq lambdalist 0 (or
1253 (lambda (x) (member x lambda-list-keywords))
1255 (length lambdalist))))
1256 (specializers (mapcar #'find-class
1257 (mapcar (lambda (x) (if (consp x) (second x) t))
1259 (unspecialized-required-part
1260 (mapcar (lambda (x) (if (consp x) (first x) x)) required-part))
1261 (unspecialized-lambdalist
1262 (append unspecialized-required-part
1263 (subseq lambdalist (length required-part)))))
1266 (MAKE-INSTANCE 'USER-METHOD
1267 :QUALIFIERS ',qualifiers
1268 :LAMBDA-LIST ',unspecialized-lambdalist
1269 :SPECIALIZERS ',specializers
1271 (LAMBDA (ARGUMENTS NEXT-METHODS-LIST)
1272 (FLET ((NEXT-METHOD-P () NEXT-METHODS-LIST)
1273 (CALL-NEXT-METHOD (&REST NEW-ARGUMENTS)
1274 (UNLESS NEW-ARGUMENTS (SETQ NEW-ARGUMENTS ARGUMENTS))
1275 (IF (NULL NEXT-METHODS-LIST)
1276 (ERROR "no next method for arguments ~:S" ARGUMENTS)
1277 (FUNCALL (SB-PCL:METHOD-FUNCTION
1278 (FIRST NEXT-METHODS-LIST))
1279 NEW-ARGUMENTS (REST NEXT-METHODS-LIST)))))
1280 (APPLY #'(LAMBDA ,unspecialized-lambdalist ,@body) ARGUMENTS)))))
1284 (defgeneric test-um03 (x))
1285 (defmethod test-um03 ((x integer))
1286 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1287 (def-user-method test-um03 ((x rational))
1288 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1289 (defmethod test-um03 ((x real))
1290 (list 'real x (not (null (next-method-p)))))
1295 (defgeneric test-um10 (x))
1296 (defmethod test-um10 ((x integer))
1297 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1298 (defmethod test-um10 ((x rational))
1299 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1300 (defmethod test-um10 ((x real))
1301 (list 'real x (not (null (next-method-p)))))
1302 (defmethod test-um10 :after ((x real)))
1303 (def-user-method test-um10 :around ((x integer))
1304 (list* 'around-integer x
1305 (not (null (next-method-p))) (call-next-method)))
1306 (defmethod test-um10 :around ((x rational))
1307 (list* 'around-rational x
1308 (not (null (next-method-p))) (call-next-method)))
1309 (defmethod test-um10 :around ((x real))
1310 (list* 'around-real x (not (null (next-method-p))) (call-next-method)))
1312 fails with a type error, and
1315 (defgeneric test-um12 (x))
1316 (defmethod test-um12 ((x integer))
1317 (list* 'integer x (not (null (next-method-p))) (call-next-method)))
1318 (defmethod test-um12 ((x rational))
1319 (list* 'rational x (not (null (next-method-p))) (call-next-method)))
1320 (defmethod test-um12 ((x real))
1321 (list 'real x (not (null (next-method-p)))))
1322 (defmethod test-um12 :after ((x real)))
1323 (defmethod test-um12 :around ((x integer))
1324 (list* 'around-integer x
1325 (not (null (next-method-p))) (call-next-method)))
1326 (defmethod test-um12 :around ((x rational))
1327 (list* 'around-rational x
1328 (not (null (next-method-p))) (call-next-method)))
1329 (def-user-method test-um12 :around ((x real))
1330 (list* 'around-real x (not (null (next-method-p))) (call-next-method)))
1332 fails with NO-APPLICABLE-METHOD.
1334 339: "DEFINE-METHOD-COMBINATION bugs"
1335 (reported by Bruno Haible via the clisp test suite)
1337 a. Syntax checking laxity (should produce errors):
1338 i. (define-method-combination foo :documentation :operator)
1339 ii. (define-method-combination foo :documentation nil)
1340 iii. (define-method-combination foo nil)
1341 iv. (define-method-combination foo nil nil
1342 (:arguments order &aux &key))
1343 v. (define-method-combination foo nil nil (:arguments &whole))
1344 vi. (define-method-combination foo nil nil (:generic-function))
1345 vii. (define-method-combination foo nil nil (:generic-function bar baz))
1346 viii. (define-method-combination foo nil nil (:generic-function (bar)))
1347 ix. (define-method-combination foo nil ((3)))
1348 x. (define-method-combination foo nil ((a)))
1350 b. define-method-combination arguments lambda list badness
1351 i. &aux args are currently unsupported;
1352 ii. default values of &optional and &key arguments are ignored;
1353 iii. supplied-p variables for &optional and &key arguments are not
1356 c. qualifier matching incorrect
1358 (define-method-combination mc27 ()
1360 (ignored (:ignore :unused)))
1362 ,@(mapcar #'(lambda (method) `(call-method ,method)) normal)))
1363 (defgeneric test-mc27 (x)
1364 (:method-combination mc27)
1365 (:method :ignore ((x number)) (/ 0)))
1368 should signal an invalid-method-error, as the :IGNORE (NUMBER)
1369 method is applicable, and yet matches neither of the method group
1372 343: MOP:COMPUTE-DISCRIMINATING-FUNCTION overriding causes error
1373 Even the simplest possible overriding of
1374 COMPUTE-DISCRIMINATING-FUNCTION, suggested in the PCL implementation
1375 as "canonical", does not work:
1376 (defclass my-generic-function (standard-generic-function) ()
1377 (:metaclass funcallable-standard-class))
1378 (defmethod compute-discriminating-function ((gf my-generic-function))
1379 (let ((dfun (call-next-method)))
1380 (lambda (&rest args)
1381 (apply dfun args))))
1383 (:generic-function-class my-generic-function))
1384 (defmethod foo (x) (+ x x))
1386 signals an error. This error is the same even if the LAMBDA is
1387 replaced by (FUNCTION (SB-KERNEL:INSTANCE-LAMBDA ...)). Maybe the
1388 SET-FUNCALLABLE-INSTANCE-FUN scary stuff in
1389 src/code/target-defstruct.lisp is broken? This seems to be broken
1390 in CMUCL 18e, so it's not caused by a recent change.
1392 344: more (?) ROOM T problems (possibly part of bug 108)
1393 In sbcl-0.8.12.51, and off and on leading up to it, the
1394 SB!VM:MEMORY-USAGE operations in ROOM T caused
1395 unhandled condition (of type SB-INT:BUG):
1396 failed AVER: "(SAP= CURRENT END)"
1397 Several clever people have taken a shot at this without fixing
1398 it; this time around (before sbcl-0.8.13 release) I (WHN) just
1399 commented out the SB!VM:MEMORY-USAGE calls until someone figures
1400 out how to make them work reliably with the rest of the GC.
1402 (Note: there's at least one dubious thing in room.lisp: see the
1403 comment in VALID-OBJ)
1405 346: alpha backtrace
1406 In sbcl-0.8.13, all backtraces from errors caused by internal errors
1407 on the alpha seem to have a "bogus stack frame".
1409 349: PPRINT-INDENT rounding implementation decisions
1410 At present, pprint-indent (and indeed the whole pretty printer)
1411 more-or-less assumes that it's using a monospace font. That's
1412 probably not too silly an assumption, but one piece of information
1413 the current implementation loses is from requests to indent by a
1414 non-integral amount. As of sbcl-0.8.15.9, the system silently
1415 truncates the indentation to an integer at the point of request, but
1416 maybe the non-integral value should be propagated through the
1417 pprinter and only truncated at output? (So that indenting by 1/2
1418 then 3/2 would indent by two spaces, not one?)
1420 352: forward-referenced-class trouble
1421 reported by Bruno Haible on sbcl-devel
1423 (setf (class-name (find-class 'a)) 'b)
1427 Expected: an instance of c, with a slot named x
1428 Got: debugger invoked on a SIMPLE-ERROR in thread 78906:
1429 While computing the class precedence list of the class named C.
1430 The class named B is a forward referenced class.
1431 The class named B is a direct superclass of the class named C.
1433 353: debugger suboptimalities on x86
1434 On x86 backtraces for undefined functions start with a bogus stack
1435 frame, and backtraces for throws to unknown catch tags with a "no
1436 debug information" frame. These are both due to CODE-COMPONENT-FROM-BITS
1437 (used on non-x86 platforms) being a more complete solution then what
1440 On x86/linux large portions of tests/debug.impure.lisp have been commented
1441 out as failures. The probable culprit for these problems is in x86-call-context
1442 (things work fine on x86/freebsd).
1444 More generally, the debugger internals suffer from excessive x86/non-x86
1445 conditionalization and OAOOMization: refactoring the common parts would
1448 354: XEPs in backtraces
1449 Under default compilation policy
1453 Has the XEP for TEST in the backtrace, not the TEST frame itself.
1454 (sparc and x86 at least)
1456 Since SBCL 0.8.20.1 this is hidden unless *SHOW-ENTRY-POINT-DETAILS*
1457 is true (instead there appear two TEST frames at least on ppc). The
1458 underlying cause seems to be that SB-C::TAIL-ANNOTATE will not merge
1459 the tail-call for the XEP, since Python has by that time proved that
1460 the function can never return; same happens if the function holds an
1461 unconditional call to ERROR.
1463 355: change-class of generic-function
1464 (reported by Bruno Haible)
1465 The MOP doesn't support change-class on a generic-function. However, SBCL
1466 apparently supports it, since it doesn't give an error or warning when doing
1467 so so. Then, however, it produces wrong results for calls to this generic
1469 ;;; The effective-methods cache:
1471 (defgeneric testgf35 (x))
1472 (defmethod testgf35 ((x integer))
1473 (cons 'integer (if (next-method-p) (call-next-method))))
1474 (defmethod testgf35 ((x real))
1475 (cons 'real (if (next-method-p) (call-next-method))))
1476 (defclass customized5-generic-function (standard-generic-function)
1478 (:metaclass sb-pcl:funcallable-standard-class))
1479 (defmethod sb-pcl:compute-effective-method ((gf customized5-generic-function) method-combination methods)
1480 `(REVERSE ,(call-next-method)))
1484 (change-class #'testgf35 'customized5-generic-function)
1486 Expected: ((INTEGER REAL) (REAL INTEGER))
1487 Got: ((INTEGER REAL) (INTEGER REAL))
1488 ;;; The discriminating-function cache:
1490 (defgeneric testgf36 (x))
1491 (defmethod testgf36 ((x integer))
1492 (cons 'integer (if (next-method-p) (call-next-method))))
1493 (defmethod testgf36 ((x real))
1494 (cons 'real (if (next-method-p) (call-next-method))))
1495 (defclass customized6-generic-function (standard-generic-function)
1497 (:metaclass sb-pcl:funcallable-standard-class))
1498 (defmethod sb-pcl:compute-discriminating-function ((gf customized6-generic-function))
1499 (let ((orig-df (call-next-method)))
1500 #'(lambda (&rest arguments)
1501 (reverse (apply orig-df arguments)))))
1505 (change-class #'testgf36 'customized6-generic-function)
1507 Expected: ((INTEGER REAL) (REAL INTEGER))
1508 Got: ((INTEGER REAL) (INTEGER REAL))
1511 (reported by Bruno Haible)
1512 After the "layout depth conflict" error, the CLOS is left in a state where
1513 it's not possible to define new standard-class subclasses any more.
1515 (defclass prioritized-dispatcher ()
1516 ((dependents :type list :initform nil)))
1517 (defmethod sb-pcl:validate-superclass ((c1 sb-pcl:funcallable-standard-class)
1518 (c2 (eql (find-class 'prioritized-dispatcher))))
1520 (defclass prioritized-generic-function (prioritized-dispatcher standard-generic-function)
1522 (:metaclass sb-pcl:funcallable-standard-class))
1523 ;; ERROR, Quit the debugger with ABORT
1524 (defclass typechecking-reader-class (standard-class)
1526 Expected: #<STANDARD-CLASS TYPECHECKING-READER-CLASS>
1527 Got: ERROR "The assertion SB-PCL::WRAPPERS failed."
1529 357: defstruct inheritance of initforms
1530 (reported by Bruno Haible)
1531 When defstruct and defclass (with :metaclass structure-class) are mixed,
1532 1. some slot initforms are ignored by the DEFSTRUCT generated constructor
1534 2. all slot initforms are ignored by MAKE-INSTANCE. (This can be arguably
1535 OK for initforms that were given in a DEFSTRUCT form, but for those
1536 given in a DEFCLASS form, I think it qualifies as a bug.)
1538 (defstruct structure02a
1542 (defclass structure02b (structure02a)
1543 ((slot4 :initform -44)
1546 (slot7 :initform (floor (* pi pi)))
1547 (slot8 :initform 88))
1548 (:metaclass structure-class))
1549 (defstruct (structure02c (:include structure02b (slot8 -88)))
1552 (slot11 (floor (exp 3))))
1554 (let ((a (make-structure02c)))
1555 (list (structure02c-slot4 a)
1556 (structure02c-slot5 a)
1557 (structure02c-slot6 a)
1558 (structure02c-slot7 a)))
1559 Expected: (-44 nil t 9)
1560 Got: (SB-PCL::..SLOT-UNBOUND.. SB-PCL::..SLOT-UNBOUND..
1561 SB-PCL::..SLOT-UNBOUND.. SB-PCL::..SLOT-UNBOUND..)
1563 (let ((b (make-instance 'structure02c)))
1564 (list (structure02c-slot2 b)
1565 (structure02c-slot3 b)
1566 (structure02c-slot4 b)
1567 (structure02c-slot6 b)
1568 (structure02c-slot7 b)
1569 (structure02c-slot8 b)
1570 (structure02c-slot10 b)
1571 (structure02c-slot11 b)))
1572 Expected: (t 3 -44 t 9 -88 t 20)
1573 Got: (0 0 0 0 0 0 0 0)
1575 358: :DECLARE argument to ENSURE-GENERIC-FUNCTION
1576 (reported by Bruno Haible)
1577 According to ANSI CL, ensure-generic-function must accept a :DECLARE
1578 keyword argument. In SBCL 0.8.16 it does not.
1581 (ensure-generic-function 'foo113 :declare '((optimize (speed 3))))
1582 (sb-pcl:generic-function-declarations #'foo113))
1583 Expected: ((OPTIMIZE (SPEED 3)))
1585 Invalid initialization argument:
1587 in call for class #<SB-MOP:FUNCALLABLE-STANDARD-CLASS STANDARD-GENERIC-FUNCTION>.
1589 The ANSI Standard, Section 7.1.2
1591 Bruno notes: The MOP specifies that ensure-generic-function accepts :DECLARATIONS.
1592 The easiest way to be compliant to both specs is to accept both (exclusively
1595 359: wrong default value for ensure-generic-function's :generic-function-class argument
1596 (reported by Bruno Haible)
1597 ANSI CL is silent on this, but the MOP's specification of ENSURE-GENERIC-FUNCTION says:
1598 "The remaining arguments are the complete set of keyword arguments
1599 received by ENSURE-GENERIC-FUNCTION."
1600 and the spec of ENSURE-GENERIC-FUNCTION-USING-CLASS:
1601 ":GENERIC-FUNCTION-CLASS - a class metaobject or a class name. If it is not
1602 supplied, it defaults to the class named STANDARD-GENERIC-FUNCTION."
1603 This is not the case in SBCL. Test case:
1604 (defclass my-generic-function (standard-generic-function)
1606 (:metaclass sb-pcl:funcallable-standard-class))
1607 (setf (fdefinition 'foo1)
1608 (make-instance 'my-generic-function :name 'foo1))
1609 (ensure-generic-function 'foo1
1610 :generic-function-class (find-class 'standard-generic-function))
1612 ; => #<SB-MOP:FUNCALLABLE-STANDARD-CLASS STANDARD-GENERIC-FUNCTION>
1613 (setf (fdefinition 'foo2)
1614 (make-instance 'my-generic-function :name 'foo2))
1615 (ensure-generic-function 'foo2)
1617 Expected: #<SB-MOP:FUNCALLABLE-STANDARD-CLASS STANDARD-GENERIC-FUNCTION>
1618 Got: #<SB-MOP:FUNCALLABLE-STANDARD-CLASS MY-GENERIC-FUNCTION>
1620 360: CALL-METHOD not recognized in method-combination body
1621 (reported by Bruno Haible)
1622 This method combination, which adds 'redo' and 'return' restarts for each
1623 method invocation to standard method combination, gives an error in SBCL.
1624 (defun prompt-for-new-values ()
1625 (format *debug-io* "~&New values: ")
1626 (list (read *debug-io*)))
1627 (defun add-method-restarts (form method)
1628 (let ((block (gensym))
1636 :REPORT (LAMBDA (STREAM) (FORMAT STREAM "Try calling ~S again." ,method))
1639 :REPORT (LAMBDA (STREAM) (FORMAT STREAM "Specify return values for ~S call." ,method))
1640 :INTERACTIVE (LAMBDA () (PROMPT-FOR-NEW-VALUES))
1641 (RETURN-FROM ,block (VALUES-LIST L)))))))))
1642 (defun convert-effective-method (efm)
1644 (if (eq (car efm) 'CALL-METHOD)
1645 (let ((method-list (third efm)))
1646 (if (or (typep (first method-list) 'method) (rest method-list))
1647 ; Reduce the case of multiple methods to a single one.
1648 ; Make the call to the next-method explicit.
1649 (convert-effective-method
1650 `(CALL-METHOD ,(second efm)
1652 (CALL-METHOD ,(first method-list) ,(rest method-list))))))
1653 ; Now the case of at most one method.
1654 (if (typep (second efm) 'method)
1655 ; Wrap the method call in a RESTART-CASE.
1656 (add-method-restarts
1657 (cons (convert-effective-method (car efm))
1658 (convert-effective-method (cdr efm)))
1660 ; Normal recursive processing.
1661 (cons (convert-effective-method (car efm))
1662 (convert-effective-method (cdr efm))))))
1663 (cons (convert-effective-method (car efm))
1664 (convert-effective-method (cdr efm))))
1666 (define-method-combination standard-with-restarts ()
1669 (primary () :required t)
1671 (flet ((call-methods-sequentially (methods)
1672 (mapcar #'(lambda (method)
1673 `(CALL-METHOD ,method))
1675 (let ((form (if (or before after (rest primary))
1676 `(MULTIPLE-VALUE-PROG1
1678 ,@(call-methods-sequentially before)
1679 (CALL-METHOD ,(first primary) ,(rest primary)))
1680 ,@(call-methods-sequentially (reverse after)))
1681 `(CALL-METHOD ,(first primary)))))
1684 `(CALL-METHOD ,(first around)
1685 (,@(rest around) (MAKE-METHOD ,form)))))
1686 (convert-effective-method form))))
1687 (defgeneric testgf16 (x) (:method-combination standard-with-restarts))
1688 (defclass testclass16a () ())
1689 (defclass testclass16b (testclass16a) ())
1690 (defclass testclass16c (testclass16a) ())
1691 (defclass testclass16d (testclass16b testclass16c) ())
1692 (defmethod testgf16 ((x testclass16a))
1694 (not (null (find-restart 'method-redo)))
1695 (not (null (find-restart 'method-return)))))
1696 (defmethod testgf16 ((x testclass16b))
1697 (cons 'b (call-next-method)))
1698 (defmethod testgf16 ((x testclass16c))
1699 (cons 'c (call-next-method)))
1700 (defmethod testgf16 ((x testclass16d))
1701 (cons 'd (call-next-method)))
1702 (testgf16 (make-instance 'testclass16d))
1704 Expected: (D B C A T T)
1705 Got: ERROR CALL-METHOD outside of a effective method form
1707 This is a bug because ANSI CL HyperSpec/Body/locmac_call-m__make-method
1709 "The macro call-method invokes the specified method, supplying it with
1710 arguments and with definitions for call-next-method and for next-method-p.
1711 If the invocation of call-method is lexically inside of a make-method,
1712 the arguments are those that were supplied to that method. Otherwise
1713 the arguments are those that were supplied to the generic function."
1714 and the example uses nothing more than these two cases (as you can see by
1715 doing (trace convert-effective-method)).
1717 361: initialize-instance of standard-reader-method ignores :function argument
1718 (reported by Bruno Haible)
1719 Pass a custom :function argument to initialize-instance of a
1720 standard-reader-method instance, but it has no effect.
1721 ;; Check that it's possible to define reader methods that do typechecking.
1723 (defclass typechecking-reader-method (sb-pcl:standard-reader-method)
1725 (defmethod initialize-instance ((method typechecking-reader-method) &rest initargs
1726 &key slot-definition)
1727 (let ((name (sb-pcl:slot-definition-name slot-definition))
1728 (type (sb-pcl:slot-definition-type slot-definition)))
1729 (apply #'call-next-method method
1730 :function #'(lambda (args next-methods)
1731 (declare (ignore next-methods))
1732 (apply #'(lambda (instance)
1733 (let ((value (slot-value instance name)))
1734 (unless (typep value type)
1735 (error "Slot ~S of ~S is not of type ~S: ~S"
1736 name instance type value))
1740 (defclass typechecking-reader-class (standard-class)
1742 (defmethod sb-pcl:validate-superclass ((c1 typechecking-reader-class) (c2 standard-class))
1744 (defmethod reader-method-class ((class typechecking-reader-class) direct-slot &rest args)
1745 (find-class 'typechecking-reader-method))
1746 (defclass testclass25 ()
1747 ((pair :type (cons symbol (cons symbol null)) :initarg :pair :accessor testclass25-pair))
1748 (:metaclass typechecking-reader-class))
1749 (macrolet ((succeeds (form)
1750 `(not (nth-value 1 (ignore-errors ,form)))))
1751 (let ((p (list 'abc 'def))
1752 (x (make-instance 'testclass25)))
1753 (list (succeeds (make-instance 'testclass25 :pair '(seventeen 17)))
1754 (succeeds (setf (testclass25-pair x) p))
1755 (succeeds (setf (second p) 456))
1756 (succeeds (testclass25-pair x))
1757 (succeeds (slot-value x 'pair))))))
1758 Expected: (t t t nil t)
1761 (inspect (first (sb-pcl:generic-function-methods #'testclass25-pair)))
1762 shows that the method was created with a FAST-FUNCTION slot but with a
1763 FUNCTION slot of NIL.
1765 362: missing error when a slot-definition is created without a name
1766 (reported by Bruno Haible)
1767 The MOP says about slot-definition initialization:
1768 "The :NAME argument is a slot name. An ERROR is SIGNALled if this argument
1769 is not a symbol which can be used as a variable name. An ERROR is SIGNALled
1770 if this argument is not supplied."
1772 (make-instance (find-class 'sb-pcl:standard-direct-slot-definition))
1774 Got: #<SB-MOP:STANDARD-DIRECT-SLOT-DEFINITION NIL>
1776 363: missing error when a slot-definition is created with a wrong documentation object
1777 (reported by Bruno Haible)
1778 The MOP says about slot-definition initialization:
1779 "The :DOCUMENTATION argument is a STRING or NIL. An ERROR is SIGNALled
1780 if it is not. This argument default to NIL during initialization."
1782 (make-instance (find-class 'sb-pcl:standard-direct-slot-definition)
1784 :documentation 'not-a-string)
1786 Got: #<SB-MOP:STANDARD-DIRECT-SLOT-DEFINITION FOO>
1788 364: does not support class objects as specializer names
1789 (reported by Bruno Haible)
1790 According to ANSI CL 7.6.2, class objects are valid specializer names,
1791 and "Parameter specializer names are used in macros intended as the
1792 user-level interface (defmethod)". DEFMETHOD's syntax section doesn't
1793 mention this possibility in the BNF for parameter-specializer-name;
1794 however, this appears to be an editorial omission, since the CLHS
1795 mentions issue CLASS-OBJECT-SPECIALIZER:AFFIRM as being approved
1796 by X3J13. SBCL doesn't support it:
1797 (defclass foo () ())
1798 (defmethod goo ((x #.(find-class 'foo))) x)
1799 Expected: #<STANDARD-METHOD GOO (#<STANDARD-CLASS FOO>)>
1800 Got: ERROR "#<STANDARD-CLASS FOO> is not a legal class name."
1802 365: mixin on generic-function subclass
1803 (reported by Bruno Haible)
1805 (defclass prioritized-dispatcher ()
1806 ((dependents :type list :initform nil)))
1807 on a generic-function subclass:
1808 (defclass prioritized-generic-function (prioritized-dispatcher standard-generic-function)
1810 (:metaclass sb-pcl:funcallable-standard-class))
1811 SBCL gives an error on this, telling to define a method on SB-MOP:VALIDATE-SUPERCLASS. If done,
1812 (defmethod sb-pcl:validate-superclass ((c1 sb-pcl:funcallable-standard-class)
1813 (c2 (eql (find-class 'prioritized-dispatcher))))
1816 (defclass prioritized-generic-function (prioritized-dispatcher standard-generic-function)
1818 (:metaclass sb-pcl:funcallable-standard-class))
1819 => debugger invoked on a SIMPLE-ERROR in thread 6687:
1820 layout depth conflict: #(#<SB-KERNEL:LAYOUT for T {500E1E9}> ...)
1822 Further discussion on this: http://thread.gmane.org/gmane.lisp.steel-bank.general/491
1824 366: cannot define two generic functions with user-defined class
1825 (reported by Bruno Haible)
1826 it is possible to define one generic function class and an instance
1827 of it. But attempting to do the same thing again, in the same session,
1828 leads to a "Control stack exhausted" error. Test case:
1829 (defclass my-generic-function-1 (standard-generic-function)
1831 (:metaclass sb-pcl:funcallable-standard-class))
1832 (defgeneric testgf-1 (x) (:generic-function-class my-generic-function-1)
1833 (:method ((x integer)) (cons 'integer nil)))
1834 (defclass my-generic-function-2 (standard-generic-function)
1836 (:metaclass sb-pcl:funcallable-standard-class))
1837 (defgeneric testgf-2 (x) (:generic-function-class my-generic-function-2)
1838 (:method ((x integer)) (cons 'integer nil)))
1839 => SB-KERNEL::CONTROL-STACK-EXHAUSTED
1841 367: TYPE-ERROR at compile time, undetected TYPE-ERROR at runtime
1843 (declaim (optimize (safety 3) (debug 2) (speed 2) (space 1)))
1847 (i367s (make-array 0 :fill-pointer t) :type (or (vector i367) null)))
1849 (g367 (error "missing :G367") :type g367 :read-only t))
1850 ;;; In sbcl-0.8.18, commenting out this (DECLAIM (FTYPE ... R367))
1851 ;;; gives an internal error at compile time:
1852 ;;; The value #<SB-KERNEL:NAMED-TYPE NIL> is not of
1853 ;;; type SB-KERNEL:VALUES-TYPE.
1854 (declaim (ftype (function ((vector i367) e367) (or s367 null)) r367))
1855 (declaim (ftype (function ((vector e367)) (values)) h367))
1857 (let ((x (g367-i367s (make-g367))))
1858 (let* ((y (or (r367 x w)
1861 (format t "~&Y=~S Z=~S~%" y z)
1863 (defun r367 (x y) (declare (ignore x y)) nil)
1864 (defun h367 (x) (declare (ignore x)) (values))
1865 ;;; In sbcl-0.8.18, executing this form causes an low-level error
1866 ;;; segmentation violation at #X9B0E1F4
1867 ;;; (instead of the TYPE-ERROR that one might like).
1868 (frob 0 (make-e367))
1869 can be made to cause two different problems, as noted in the comments:
1870 bug 367a: Compile and load the file. No TYPE-ERROR is signalled at
1871 run time (in the (S367-G367 Y) form of FROB, when Y is NIL
1872 instead of an instance of S367). Instead (on x86/Linux at least)
1873 we end up with a segfault.
1874 bug 367b: Comment out the (DECLAIM (FTYPE ... R367)), and compile
1875 the file. The compiler fails with TYPE-ERROR at compile time.
1877 368: miscompiled OR (perhaps related to bug 367)
1878 Trying to relax type declarations to find a workaround for bug 367,
1879 it turns out that even when the return type isn't declared (or
1880 declared to be T, anyway) the system remains confused about type
1881 inference in code similar to that for bug 367:
1882 (in-package :cl-user)
1883 (declaim (optimize (safety 3) (debug 2) (speed 2) (space 1)))
1887 (i368s (make-array 0 :fill-pointer t) :type (or (vector i368) null)))
1889 (g368 (error "missing :G368") :type g368 :read-only t))
1890 (declaim (ftype (function (fixnum (vector i368) e368) t) r368))
1891 (declaim (ftype (function (fixnum (vector e368)) t) h368))
1892 (defparameter *h368-was-called-p* nil)
1893 (defun nsu (vertices e368)
1894 (let ((i368s (g368-i368s (make-g368))))
1895 (let ((fuis (r368 0 i368s e368)))
1896 (format t "~&FUIS=~S~%" fuis)
1897 (or fuis (h368 0 i368s)))))
1899 (declare (ignore w x y))
1902 (declare (ignore w x))
1903 (setf *h368-was-called-p* t)
1904 (make-s368 :g368 (make-g368)))
1906 (format t "~&calling NSU~%")
1907 (let ((nsu (nsu #() (make-e368))))
1908 (format t "~&NSU returned ~S~%" nsu)
1909 (format t "~&*H368-WAS-CALLED-P*=~S~%" *h368-was-called-p*)
1910 (assert (s368-p nsu))
1911 (assert *h368-was-called-p*))
1912 In sbcl-0.8.18, both ASSERTs fail, and (DISASSEMBLE 'NSU) shows
1913 that no call to H368 is compiled.
1915 369: unlike-an-intersection behavior of VALUES-TYPE-INTERSECTION
1916 In sbcl-0.8.18.2, the identity $(x \cap y \cap y)=(x \cap y)$
1917 does not hold for VALUES-TYPE-INTERSECTION, even for types which
1918 can be intersected exactly, so that ASSERTs fail in this test case:
1919 (in-package :cl-user)
1920 (let ((types (mapcar #'sb-c::values-specifier-type
1921 '((values (vector package) &optional)
1922 (values (vector package) &rest t)
1923 (values (vector hash-table) &rest t)
1924 (values (vector hash-table) &optional)
1925 (values t &optional)
1927 (values nil &optional)
1928 (values nil &rest t)
1929 (values sequence &optional)
1930 (values sequence &rest t)
1931 (values list &optional)
1932 (values list &rest t)))))
1935 (let ((i (sb-c::values-type-intersection x y)))
1936 (assert (sb-c::type= i (sb-c::values-type-intersection i x)))
1937 (assert (sb-c::type= i (sb-c::values-type-intersection i y)))))))
1939 370: reader misbehaviour on large-exponent floats
1940 (read-from-string "1.0s1000000000000000000000000000000000000000")
1941 causes the reader to attempt to create a very large bignum (which it
1942 will then attempt to coerce to a rational). While this isn't
1943 completely wrong, it is probably not ideal -- checking the floating
1944 point control word state and then returning the relevant float
1945 (most-positive-short-float or short-float-infinity) or signalling an
1946 error immediately would seem to make more sense.
1948 372: floating-point overflow not signalled on ppc/darwin
1949 The following assertions in float.pure.lisp fail on ppc/darwin
1950 (Mac OS X version 10.3.7):
1951 (assert (raises-error? (scale-float 1.0 most-positive-fixnum)
1952 floating-point-overflow))
1953 (assert (raises-error? (scale-float 1.0d0 (1+ most-positive-fixnum))
1954 floating-point-overflow)))
1955 as the SCALE-FLOAT just returns
1956 #.SB-EXT:SINGLE/DOUBLE-FLOAT-POSITIVE-INFINITY. These tests have been
1957 disabled on Darwin for now.
1960 (compile nil '(lambda (p1)
1961 (declare (optimize (speed 1) (safety 2) (debug 2) (space 0))
1965 fails on hairy type check in IR2.
1967 1. KEYWORDP is MAYBE-INLINE expanded (before TYPEP-like
1968 transformation could eliminate it).
1970 2. From the only call of KEYWORDP the type of its argument is
1971 derived to be KEYWORD.
1973 2. Type check for P1 is generated; it uses KEYWORDP to perform the
1974 check, and so references the local function; from the KEYWORDP
1975 argument type new CAST to KEYWORD is generated. The compiler
1978 377: Memory fault error reporting
1979 On those architectures where :C-STACK-IS-CONTROL-STACK is in
1980 *FEATURES*, we handle SIG_MEMORY_FAULT (SEGV or BUS) on an altstack,
1981 so we cannot handle the signal directly (as in interrupt_handle_now())
1982 in the case when the signal comes from some external agent (the user
1983 using kill(1), or a fault in some foreign code, for instance). As
1984 of sbcl-0.8.20.20, this is fixed by calling
1985 arrange_return_to_lisp_function() to a new error-signalling
1986 function, but as a result the error reporting is poor: we cannot
1987 even tell the user at which address the fault occurred. We should
1988 arrange such that arguments can be passed to the function called from
1989 arrange_return_to_lisp_function(), but this looked hard to do in
1990 general without suffering from memory leaks.
1992 379: TRACE :ENCAPSULATE NIL broken on ppc/darwin
1993 See commented-out test-case in debug.impure.lisp.
1995 380: Accessor redefinition fails because of old accessor name
1996 When redefining an accessor, SB-PCL::FIX-SLOT-ACCESSORS may try to
1997 find the generic function named by the old accessor name using
1998 ENSURE-GENERIC-FUNCTION and then remove the old accessor's method in
1999 the GF. If the old name does not name a function, or if the old name
2000 does not name a generic function, no attempt to find the GF or remove
2001 any methods is made.
2003 However, if an unrelated GF with an incompatible lambda list exists,
2004 the class redefinition will fail when SB-PCL::REMOVE-READER-METHOD
2005 tries to find and remove a method with an incompatible lambda list
2006 from the unrelated generic function.
2008 381: incautious calls to EQUAL in fasl dumping
2010 (frob #(#1=(a #1#)))
2011 (frob #(#1=(b #1#)))
2012 (frob #(#1=(a #1#)))
2013 in sbcl-0.9.0 causes CONTROL-STACK-EXHAUSTED. My (WHN) impression
2014 is that this follows from the use of (MAKE-HASH-TABLE :TEST 'EQUAL)
2015 to detect sharing, in which case fixing it might require either
2016 getting less ambitious about detecting shared list structure, or
2017 implementing the moral equivalent of EQUAL hash tables in a
2020 382: externalization unexpectedly changes array simplicity
2021 COMPILE-FILE and LOAD
2023 (let ((x #.(make-array 4 :fill-pointer 0)))
2024 (values (eval `(typep ',x 'simple-array))
2025 (typep x 'simple-array))))
2026 then (FOO) => T, NIL.
2028 Similar problems exist with SIMPLE-ARRAY-P, ARRAY-HEADER accessors
2029 and all array dimension functions.
2031 383: ASH'ing non-constant zeros
2034 (declare (type (integer -2 14) b))
2035 (declare (ignorable b))
2036 (ash (imagpart b) 57))
2037 on PPC (and other platforms, presumably) gives an error during the
2038 emission of FASH-ASH-LEFT/FIXNUM=>FIXNUM as the assembler attempts to
2039 stuff a too-large constant into the immediate field of a PPC
2040 instruction. Either the VOP should be fixed or the compiler should be
2041 taught how to transform this case away, paying particular attention
2042 to side-effects that might occur in the arguments to ASH.
2044 384: Compiler runaway on very large character types
2046 (compile nil '(lambda (x)
2047 (declare (type (member #\a 1) x))
2048 (the (member 1 nil) x)))
2050 The types apparently normalize into a very large type, and the compiler
2051 gets lost in REMOVE-DUPLICATES. Perhaps the latter should use
2052 a better algorithm (one based on hash tables, say) on very long lists
2053 when :TEST has its default value?
2057 (compile nil '(lambda (x) (the (not (eql #\a)) x)))
2059 (partially fixed in 0.9.3.1, but a better representation for these
2063 (format nil "~4,1F" 0.001) => "0.00" (should be " 0.0");
2064 (format nil "~4,1@F" 0.001) => "+.00" (should be "+0.0").
2066 386: SunOS/x86 stack exhaustion handling broken
2067 According to <http://alfa.s145.xrea.com/sbcl/solaris-x86.html>, the
2068 stack exhaustion checking (implemented with a write-protected guard
2069 page) does not work on SunOS/x86.
2072 12:10 < jsnell> the package-lock test is basically due to a change in the test
2073 behaviour when you install a handler for error around it. I
2074 thought I'd disabled the test for now, but apparently that was
2076 12:19 < Xophe> jsnell: ah, I see the problem in the package-locks stuff
2077 12:19 < Xophe> it's the same problem as we had with compiler-error conditions
2078 12:19 < Xophe> the thing that's signalled up and down the stack is a subtype of
2079 ERROR, where it probably shouldn't be
2082 (found by Dmitry Bogomolov)
2084 (defclass foo () ((x :type (unsigned-byte 8))))
2085 (defclass bar () ((x :type symbol)))
2086 (defclass baz (foo bar) ())
2090 SB-PCL::SPECIALIZER-APPLICABLE-USING-TYPE-P cannot handle the second argument
2094 (reported several times on sbcl-devel, by Rick Taube, Brian Rowe and
2097 ROUND-NUMERIC-BOUND assumes that float types always have a FORMAT
2098 specifying whether they're SINGLE or DOUBLE. This is true for types
2099 computed by the type system itself, but the compiler type derivation
2100 short-circuits this and constructs non-canonical types. A temporary
2101 fix was made to ROUND-NUMERIC-BOUND for the sbcl-0.9.6 release, but
2102 the right fix is to remove the abstraction violation in the
2103 compiler's type deriver.
2106 (fixed in sbcl-0.9.7.1)