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).
35 KNOWN BUGS OF NO SPECIAL CLASS:
38 DEFSTRUCT almost certainly should overwrite the old LAYOUT information
39 instead of just punting when a contradictory structure definition
40 is loaded. As it is, if you redefine DEFSTRUCTs in a way which
41 changes their layout, you probably have to rebuild your entire
42 program, even if you know or guess enough about the internals of
43 SBCL to wager that this (undefined in ANSI) operation would be safe.
46 ANSI specifies that a type mismatch in a structure slot
47 initialization value should not cause a warning.
49 This one might not be fixed for a while because while we're big
50 believers in ANSI compatibility and all, (1) there's no obvious
51 simple way to do it (short of disabling all warnings for type
52 mismatches everywhere), and (2) there's a good portable
53 workaround, and (3) by their own reasoning, it looks as though
54 ANSI may have gotten it wrong. ANSI justifies this specification
56 The restriction against issuing a warning for type mismatches
57 between a slot-initform and the corresponding slot's :TYPE
58 option is necessary because a slot-initform must be specified
59 in order to specify slot options; in some cases, no suitable
61 However, in SBCL (as in CMU CL or, for that matter, any compiler
62 which really understands Common Lisp types) a suitable default
63 does exist, in all cases, because the compiler understands the
64 concept of functions which never return (i.e. has return type NIL).
65 Thus, as a portable workaround, you can use a call to some
66 known-never-to-return function as the default. E.g.
68 (BAR (ERROR "missing :BAR argument")
69 :TYPE SOME-TYPE-TOO-HAIRY-TO-CONSTRUCT-AN-INSTANCE-OF))
71 (DECLAIM (FTYPE (FUNCTION () NIL) MISSING-ARG))
72 (DEFUN REQUIRED-ARG () ; workaround for SBCL non-ANSI slot init typing
73 (ERROR "missing required argument"))
75 (BAR (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
76 (BLETCH (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
77 (N-REFS-SO-FAR 0 :TYPE (INTEGER 0)))
78 Such code should compile without complaint and work correctly either
79 on SBCL or on any other completely compliant Common Lisp system.
82 bogus warnings about undefined functions for magic functions like
83 SB!C::%%DEFUN and SB!C::%DEFCONSTANT when cross-compiling files
84 like src/code/float.lisp. Fixing this will probably require
85 straightening out enough bootstrap consistency issues that
86 the cross-compiler can run with *TYPE-SYSTEM-INITIALIZED*.
87 Instead, the cross-compiler runs in a slightly flaky state
88 which is sane enough to compile SBCL itself, but which is
89 also unstable in several ways, including its inability
90 to really grok function declarations.
93 The "compiling top-level form:" output ought to be condensed.
94 Perhaps any number of such consecutive lines ought to turn into a
95 single "compiling top-level forms:" line.
98 The way that the compiler munges types with arguments together
99 with types with no arguments (in e.g. TYPE-EXPAND) leads to
100 weirdness visible to the user:
101 (DEFTYPE FOO () 'FIXNUM)
103 (TYPEP 11 '(FOO)) => T, which seems weird
104 (TYPEP 11 'FIXNUM) => T
105 (TYPEP 11 '(FIXNUM)) signals an error, as it should
106 The situation is complicated by the presence of Common Lisp types
107 like UNSIGNED-BYTE (which can either be used in list form or alone)
108 so I'm not 100% sure that the behavior above is actually illegal.
109 But I'm 90+% sure, and the following related behavior,
111 treating the bare symbol AND as equivalent to '(AND), is specifically
112 forbidden (by the ANSI specification of the AND type).
115 It would be nice if the
117 (during macroexpansion)
118 said what macroexpansion was at fault, e.g.
120 (during macroexpansion of IN-PACKAGE,
121 during macroexpansion of DEFFOO)
124 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
125 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
126 (Also, when this is fixed, we can enable the code in PROCLAIM which
127 checks for incompatible FTYPE redeclarations.)
130 (I *think* this is a bug. It certainly seems like strange behavior. But
131 the ANSI spec is scary, dark, and deep.. -- WHN)
132 (FORMAT NIL "~,1G" 1.4) => "1. "
133 (FORMAT NIL "~3,1G" 1.4) => "1. "
136 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
138 (setf (find-class 'ccc1) (find-class 'ccc))
139 (defmethod zut ((c ccc1)) 123)
140 In sbcl-0.7.1.13, this gives an error,
141 There is no class named CCC1.
142 DTC's recommended workaround from the mailing list 3 Mar 2000:
143 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
146 The ANSI spec, in section "22.3.5.2 Tilde Less-Than-Sign: Logical Block",
147 says that an error is signalled if ~W, ~_, ~<...~:>, ~I, or ~:T is used
148 inside "~<..~>" (without the colon modifier on the closing syntax).
149 However, SBCL doesn't do this:
150 * (FORMAT T "~<munge~wegnum~>" 12)
155 Sometimes (SB-EXT:QUIT) fails with
156 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
157 Process inferior-lisp exited abnormally with code 1
158 I haven't noticed a repeatable case of this yet.
161 In some cases the compiler believes type declarations on array
162 elements without checking them, e.g.
163 (DECLAIM (OPTIMIZE (SAFETY 3) (SPEED 1) (SPACE 1)))
166 (DECLARE (TYPE (SIMPLE-ARRAY CONS 1) X))
167 (WHEN (CONSP (AREF X 0))
169 (BAR (VECTOR (MAKE-FOO :A 11 :B 12)))
172 in SBCL 0.6.5 (and also in CMU CL 18b). This does not happen for
173 all cases, e.g. the type assumption *is* checked if the array
174 elements are declared to be of some structure type instead of CONS.
177 The printer doesn't report closures very well. This is true in
181 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
182 It would be nice to make closures have a settable name slot,
183 and make things like DEFSTRUCT and FLET, which create closures,
184 set helpful values into this slot.
187 And as long as we're wishing, it would be awfully nice if INSPECT could
188 also report on closures, telling about the values of the bound variables.
191 The compiler assumes that any time a function of declared FTYPE
192 doesn't signal an error, its arguments were of the declared type.
193 E.g. compiling and loading
194 (DECLAIM (OPTIMIZE (SAFETY 3)))
195 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
197 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE)) FACTORIAL))
199 (COND ((> (FACTORIAL X) 1.0E6)
200 (FORMAT T "too big~%"))
202 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
204 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
207 will cause the INTEGERP case to be selected, giving bogus output a la
209 This violates the "declarations are assertions" principle.
210 According to the ANSI spec, in the section "System Class FUNCTION",
211 this is a case of "lying to the compiler", but the lying is done
212 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
213 not by the unexpectedly general definition of FACTORIAL. In any case,
214 "declarations are assertions" means that lying to the compiler should
215 cause an error to be signalled, and should not cause a bogus
216 result to be returned. Thus, the compiler should not assume
217 that arbitrary functions check their argument types. (It might
218 make sense to add another flag (CHECKED?) to DEFKNOWN to
219 identify functions which *do* check their argument types.)
220 (Also, verify that the compiler handles declared function
221 return types as assertions.)
224 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
225 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
227 (DECLARE (TYPE INDEXOID X))
228 (THE (VALUES INDEXOID)
230 where the implementation of the type check in function FOO
231 includes a full call to %TYPEP. There are also some fundamental problems
232 with the interpretation of VALUES types (inherited from CMU CL, and
233 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
234 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
237 The definitions of SIGCONTEXT-FLOAT-REGISTER and
238 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
239 supported on FreeBSD because the floating point state is not saved,
240 but at least as of FreeBSD 4.0, the floating point state *is* saved,
241 so they could be supported after all. Very likely
242 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
245 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
246 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
247 issue with 'union, 'and, 'or etc. So even though according to the
248 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
249 (and now SBCL) interpret them as legal types.
252 a slew of floating-point-related errors reported by Peter Van Eynde
254 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
255 should probably be 1.4012985e-45. In SBCL,
256 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
257 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
258 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
259 and LEAST-NEGATIVE-LONG-FLOAT.
260 c: Many expressions generate floating infinity on x86/Linux:
265 PVE's regression tests want them to raise errors. sbcl-0.7.0.5
266 on x86/Linux generates the infinities instead. That might or
267 might not be conforming behavior, but it's also inconsistent,
268 which is almost certainly wrong. (Inconsistency: (/ 1 0.0)
269 should give the same result as (/ 1.0 0.0), but instead (/ 1 0.0)
270 generates SINGLE-FLOAT-POSITIVE-INFINITY and (/ 1.0 0.0)
272 d: (in section12.erg) various forms a la
273 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
274 don't give the right behavior.
277 type safety errors reported by Peter Van Eynde July 25, 2000:
278 a: (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
280 In general lengths of array type specifications aren't
281 checked by COERCE, so it fails when the spec is
282 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
283 b: CONCATENATE has the same problem of not checking the length
284 of specified output array types. MAKE-SEQUENCE and MAP and
285 MERGE also have the same problem.
286 c: (COERCE 'AND 'FUNCTION) returns something related to
287 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
288 f: (FLOAT-RADIX 2/3) should signal an error instead of
290 g: (LOAD "*.lsp") should signal FILE-ERROR.
291 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
292 should signal TYPE-ERROR.
293 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
294 be used for input and output as needed. It should fail with
295 TYPE-ERROR when handed e.g. the results of
296 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
297 the inappropriate positions, but doesn't.
298 j: (PARSE-NAMESTRING (COERCE (LIST #\f #\o #\o (CODE-CHAR 0) #\4 #\8)
300 should probably signal an error instead of making a pathname with
302 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
303 not a binary input stream, but instead cheerfully reads from
304 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
307 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
308 a: (DEFCLASS FOO () (A B A)) should signal a PROGRAM-ERROR, and
310 b: (DEFCLASS FOO () (A B A) (:DEFAULT-INITARGS X A X B)) should
311 signal a PROGRAM-ERROR, and doesn't.
312 c: (DEFCLASS FOO07 NIL ((A :ALLOCATION :CLASS :ALLOCATION :CLASS))),
313 and other DEFCLASS forms with duplicate specifications in their
314 slots, should signal a PROGRAM-ERROR, and doesn't.
315 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
316 causes a COMPILER-ERROR.
319 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
320 a: (SYMBOL-MACROLET ((T TRUE)) ..) should probably signal
321 PROGRAM-ERROR, but SBCL accepts it instead.
322 b: SYMBOL-MACROLET should refuse to bind something which is
323 declared as a global variable, signalling PROGRAM-ERROR.
324 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
325 it binds is declared SPECIAL inside.
328 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
330 (DEFGENERIC FOO02 (X))
331 (DEFMETHOD FOO02 ((X NUMBER)) T)
332 (LET ((M (FIND-METHOD (FUNCTION FOO02)
334 (LIST (FIND-CLASS (QUOTE NUMBER))))))
335 (REMOVE-METHOD (FUNCTION FOO02) M)
336 (DEFGENERIC FOO03 (X))
337 (ADD-METHOD (FUNCTION FOO03) M)))
338 should give an error, but SBCL allows it.
339 b: READ should probably return READER-ERROR, not the bare
340 arithmetic error, when input a la "1/0" or "1e1000" causes
344 It has been reported (e.g. by Peter Van Eynde) that there are
345 several metaobject protocol "errors". (In order to fix them, we might
346 need to document exactly what metaobject protocol specification
347 we're following -- the current code is just inherited from PCL.)
350 The implementation of #'+ returns its single argument without
351 type checking, e.g. (+ "illegal") => "illegal".
354 The debugger LIST-LOCATIONS command doesn't work properly.
357 Compiling and loading
358 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
360 then requesting a BACKTRACE at the debugger prompt gives no information
361 about where in the user program the problem occurred.
364 The compiler is supposed to do type inference well enough that
367 ((SIMPLE-ARRAY SINGLE-FLOAT)
369 (DECLARE (TYPE (SIMPLE-ARRAY SINGLE-FLOAT) X))
372 is redundant. However, as reported by Juan Jose Garcia Ripoll for
373 CMU CL, it sometimes doesn't. Adding declarations is a pretty good
374 workaround for the problem for now, but can't be done by the TYPECASE
375 macros themselves, since it's too hard for the macro to detect
376 assignments to the variable within the clause.
377 Note: The compiler *is* smart enough to do the type inference in
378 many cases. This case, derived from a couple of MACROEXPAND-1
379 calls on Ripoll's original test case,
381 (DECLARE (OPTIMIZE SPEED (SAFETY 0)))
382 (COND ((TYPEP A '(SIMPLE-ARRAY SINGLE-FLOAT)) NIL
383 (LET ((LENGTH (ARRAY-TOTAL-SIZE A)))
384 (LET ((I 0) (G2554 LENGTH))
385 (DECLARE (TYPE REAL G2554) (TYPE REAL I))
388 (WHEN (>= I G2554) (GO SB-LOOP::END-LOOP))
389 (SETF (ROW-MAJOR-AREF A I) (- (ROW-MAJOR-AREF A I)))
390 (GO SB-LOOP::NEXT-LOOP)
391 SB-LOOP::END-LOOP))))))
392 demonstrates the problem; but the problem goes away if the TAGBODY
393 and GO forms are removed (leaving the SETF in ordinary, non-looping
394 code), or if the TAGBODY and GO forms are retained, but the
395 assigned value becomes 0.0 instead of (- (ROW-MAJOR-AREF A I)).
398 Paul Werkowski wrote on cmucl-imp@cons.org 2000-11-15
399 I am looking into this problem that showed up on the cmucl-help
400 list. It seems to me that the "implementation specific environment
401 hacking functions" found in pcl/walker.lisp are completely messed
402 up. The good thing is that they appear to be barely used within
403 PCL and the munged environment object is passed to cmucl only
404 in calls to macroexpand-1, which is probably why this case fails.
405 SBCL uses essentially the same code, so if the environment hacking
406 is screwed up, it affects us too.
409 Using the pretty-printer from the command prompt gives funny
410 results, apparently because the pretty-printer doesn't know
411 about user's command input, including the user's carriage return
412 that the user, and therefore the pretty-printer thinks that
413 the new output block should start indented 2 or more characters
414 rightward of the correct location.
417 (probably related to bug #70; maybe related to bug #109)
418 As reported by Carl Witty on submit@bugs.debian.org 1999-05-08,
420 (in-package "CL-USER")
421 (defun equal-terms (termx termy)
423 ((alpha-equal-bound-term-lists (listx listy)
424 (or (and (null listx) (null listy))
426 (let ((bindings-x (bindings-of-bound-term (car listx)))
427 (bindings-y (bindings-of-bound-term (car listy))))
428 (if (and (null bindings-x) (null bindings-y))
429 (alpha-equal-terms (term-of-bound-term (car listx))
430 (term-of-bound-term (car listy)))
431 (and (= (length bindings-x) (length bindings-y))
433 (enter-binding-pairs (bindings-of-bound-term (car listx))
434 (bindings-of-bound-term (car listy)))
435 (alpha-equal-terms (term-of-bound-term (car listx))
436 (term-of-bound-term (car listy)))
437 (exit-binding-pairs (bindings-of-bound-term (car listx))
438 (bindings-of-bound-term (car listy)))))))
439 (alpha-equal-bound-term-lists (cdr listx) (cdr listy)))))
441 (alpha-equal-terms (termx termy)
442 (if (and (variable-p termx)
444 (equal-bindings (id-of-variable-term termx)
445 (id-of-variable-term termy))
446 (and (equal-operators-p (operator-of-term termx) (operator-of-term termy))
447 (alpha-equal-bound-term-lists (bound-terms-of-term termx)
448 (bound-terms-of-term termy))))))
452 (with-variable-invocation (alpha-equal-terms termx termy))))))
453 causes an assertion failure
454 The assertion (EQ (C::LAMBDA-TAIL-SET C::CALLER)
455 (C::LAMBDA-TAIL-SET (C::LAMBDA-HOME C::CALLEE))) failed.
457 Bob Rogers reports (1999-07-28 on cmucl-imp@cons.org) a smaller test
458 case with the same problem:
459 (defun parse-fssp-alignment ()
460 ;; Given an FSSP alignment file named by the argument . . .
461 (labels ((get-fssp-char ()
465 ;; Stub body, enough to tickle the bug.
466 (list (read-fssp-char)
470 ANSI specifies that the RESULT-TYPE argument of CONCATENATE must be
471 a subtype of SEQUENCE, but CONCATENATE doesn't check this properly:
472 (CONCATENATE 'SIMPLE-ARRAY #(1 2) '(3)) => #(1 2 3)
473 This also leads to funny behavior when derived type specifiers
474 are used, as originally reported by Milan Zamazal for CMU CL (on the
475 Debian bugs mailing list (?) 2000-02-27), then reported by Martin
476 Atzmueller for SBCL (2000-10-01 on sbcl-devel@lists.sourceforge.net):
477 (DEFTYPE FOO () 'SIMPLE-ARRAY)
478 (CONCATENATE 'FOO #(1 2) '(3))
479 => #<ARRAY-TYPE SIMPLE-ARRAY> is a bad type specifier for
481 The derived type specifier FOO should act the same way as the
482 built-in type SIMPLE-ARRAY here, but it doesn't. That problem
483 doesn't seem to exist for sequence types:
484 (DEFTYPE BAR () 'SIMPLE-VECTOR)
485 (CONCATENATE 'BAR #(1 2) '(3)) => #(1 2 3)
488 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
489 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
490 crashes SBCL. In general tracing anything which is used in the
491 implementation of TRACE is likely to have the same problem.
494 As reported by Daniel Solaz on cmucl-help@cons.org 2000-11-23,
495 SXHASH returns the same value for all non-STRUCTURE-OBJECT instances,
496 notably including all PCL instances. There's a limit to how much
497 SXHASH can do to return unique values for instances, but at least
498 it should probably look at the class name, the way that it does
499 for STRUCTURE-OBJECTs.
502 (probably related to bug #65; maybe related to bug #109)
503 The compiler doesn't like &OPTIONAL arguments in LABELS and FLET
505 (DEFUN FIND-BEFORE (ITEM SEQUENCE &KEY (TEST #'EQL))
506 (LABELS ((FIND-ITEM (OBJ SEQ TEST &OPTIONAL (VAL NIL))
507 (LET ((ITEM (FIRST SEQ)))
510 ((FUNCALL TEST OBJ ITEM)
513 (FIND-ITEM OBJ (REST SEQ) TEST (NCONC VAL `(,ITEM))))))))
514 (FIND-ITEM ITEM SEQUENCE TEST)))
515 from David Young's bug report on cmucl-help@cons.org 30 Nov 2000
516 causes sbcl-0.6.9 to fail with
517 error in function SB-KERNEL:ASSERT-ERROR:
518 The assertion (EQ (SB-C::LAMBDA-TAIL-SET SB-C::CALLER)
519 (SB-C::LAMBDA-TAIL-SET
520 (SB-C::LAMBDA-HOME SB-C::CALLEE))) failed.
523 (DECLAIM (OPTIMIZE ..)) doesn't work. E.g. even after
524 (DECLAIM (OPTIMIZE (SPEED 3))), things are still optimized with
525 the previous SPEED policy. This bug will probably get fixed in
526 0.6.9.x in a general cleanup of optimization policy.
529 (DECLAIM (OPTIMIZE ..)) doesn't work properly inside LOCALLY forms.
532 As reported by Martin Atzmueller on sbcl-devel 26 Dec 2000,
533 ANSI says that WITH-OUTPUT-TO-STRING should have a keyword
534 :ELEMENT-TYPE, but in sbcl-0.6.9 this is not defined for
535 WITH-OUTPUT-TO-STRING.
538 ANSI says in one place that type declarations can be abbreviated even
539 when the type name is not a symbol, e.g.
540 (DECLAIM ((VECTOR T) *FOOVECTOR*))
541 SBCL doesn't support this. But ANSI says in another place that this
542 isn't allowed. So it's not clear this is a bug after all. (See the
543 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
547 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
548 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
549 an easily guessable temporary filename in a way which might open
550 applications using LOAD-FOREIGN to hijacking by malicious users
551 on the same machine. Incantations for doing this safely are
552 floating around the net in various "how to write secure programs
553 despite Unix" documents, and it would be good to (1) fix this in
554 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
555 files and make it share the same new safe logic.
558 Functions are assigned names based on the context in which they're
559 defined. This is less than ideal for the functions which are
560 used to implement CLOS methods. E.g. the output of
561 (DESCRIBE 'PRINT-OBJECT) lists functions like
562 #<FUNCTION "DEF!STRUCT (TRACE-INFO (:MAKE-LOAD-FORM-FUN SB-KERNEL:JUST-DUMP-IT-NORMALLY) (:PRINT-OBJECT #))" {1020E49}>
564 #<FUNCTION "MACROLET ((FORCE-DELAYED-DEF!METHODS NIL #))" {1242871}>
565 It would be better if these functions' names always identified
566 them as methods, and identified their generic functions and
570 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
571 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
572 the mean of the distribution can be systematically O(0.1%) wrong.
573 Just increasing R-I-E-B is probably not a good solution, since
574 it would decrease efficiency more than is probably necessary. Perhaps
575 using some sort of accept/reject method would be better.
578 Internally the compiler sometimes evaluates
579 (sb-kernel:type/= (specifier-type '*) (specifier-type t))
580 (I stumbled across this when I added an
581 (assert (not (eq type1 *wild-type*)))
582 in the NAMED :SIMPLE-= type method.) '* isn't really a type, and
583 in a type context should probably be translated to T, and so it's
584 probably wrong to ask whether it's equal to the T type and then (using
585 the EQ type comparison in the NAMED :SIMPLE-= type method) return NIL.
586 (I haven't tried to investigate this bug enough to guess whether
587 there might be any user-level symptoms.)
590 a latent cross-compilation/bootstrapping bug: The cross-compilation
591 host's CL:CHAR-CODE-LIMIT is used in target code in readtable.lisp
592 and possibly elsewhere. Instead, we should use the target system's
593 CHAR-CODE-LIMIT. This will probably cause problems if we try to
594 bootstrap on a system which uses a different value of CHAR-CODE-LIMIT
598 (subtypep '(or (integer -1 1)
602 (integer -1 1))) => NIL,T
603 An analogous problem with SINGLE-FLOAT and REAL types was fixed in
604 sbcl-0.6.11.22, but some peculiarites of the RATIO type make it
605 awkward to generalize the fix to INTEGER and RATIONAL. It's not
606 clear what's the best fix. (See the "bug in type handling" discussion
607 on cmucl-imp ca. 2001-03-22 and ca. 2001-02-12.)
610 Inconsistencies between derived and declared VALUES return types for
611 DEFUN aren't checked very well. E.g. the logic which successfully
612 catches problems like
613 (declaim (ftype (function (fixnum) float) foo))
615 (declare (type integer x))
616 (values x)) ; wrong return type, detected, gives warning, good!
618 (declaim (ftype (function (t) (values t t)) bar))
620 (values x)) ; wrong number of return values, no warning, bad!
621 The cause of this is seems to be that (1) the internal function
622 VALUES-TYPES-EQUAL-OR-INTERSECT used to make the check handles its
623 arguments symmetrically, and (2) when the type checking code was
624 written back when when SBCL's code was still CMU CL, the intent
626 (declaim (ftype (function (t) t) bar))
628 (values x x)) ; wrong number of return values; should give warning?
629 not be warned for, because a two-valued return value is considered
630 to be compatible with callers who expects a single value to be
631 returned. That intent is probably not appropriate for modern ANSI
632 Common Lisp, but fixing this might be complicated because of other
633 divergences between auld-style and new-style handling of
634 multiple-VALUES types. (Some issues related to this were discussed
635 on cmucl-imp at some length sometime in 2000.)
638 The facility for dumping a running Lisp image to disk gets confused
639 when run without the PURIFY option, and creates an unnecessarily large
640 core file (apparently representing memory usage up to the previous
641 high-water mark). Moreover, when the file is loaded, it confuses the
642 GC, so that thereafter memory usage can never be reduced below that
646 The TRACE facility can't be used on some kinds of functions.
647 (Basically, the breakpoint facility was incompletely implemented
648 in the X86 port of CMU CL, and hasn't been fixed in SBCL.)
651 In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
652 CL), out-of-line structure slot setters are horribly inefficient
653 whenever the type of the slot is declared, because out-of-line
654 structure slot setters are implemented as closures to save space,
655 so the compiler doesn't compile the type test into code, but
656 instead just saves the type in a lexical closure and interprets it
658 A proper solution involves deciding whether it's really worth
659 saving space by implementing structure slot accessors as closures.
660 (If it's not worth it, the problem vanishes automatically. If it
661 is worth it, there are hacks we could use to force type tests to
662 be compiled anyway, and even shared. E.g. we could implement
663 an EQUAL hash table mapping from types to compiled type tests,
664 and save the appropriate compiled type test as part of each lexical
665 closure; or we could make the lexical closures be placeholders
666 which overwrite their old definition as a lexical closure with
667 a new compiled definition the first time that they're called.)
668 As a workaround for the problem, #'(SETF FOO) expressions can
669 be replaced with (EFFICIENT-SETF-FUNCTION FOO), where
670 (defmacro efficient-setf-function (place-function-name)
671 (or #+sbcl (and (sb-impl::info :function :accessor-for place-function-name)
672 ;; a workaround for the problem, encouraging the
673 ;; inline expansion of the structure accessor, so
674 ;; that the compiler can optimize its type test
675 (let ((new-value (gensym "NEW-VALUE-"))
676 (structure-value (gensym "STRUCTURE-VALUE-")))
677 `(lambda (,new-value ,structure-value)
678 (setf (,place-function-name ,structure-value)
680 ;; no problem, can just use the ordinary expansion
681 `(function (setf ,place-function-name))))
684 There's apparently a bug in CEILING optimization which caused
685 Douglas Crosher to patch the CMU CL version. Martin Atzmueller
686 applied the patches to SBCL and they didn't seem to cause problems
687 (as reported sbcl-devel 2001-05-04). However, since the patches
688 modify nontrivial code which was apparently written incorrectly
689 the first time around, until regression tests are written I'm not
690 comfortable merging the patches in the CVS version of SBCL.
693 As reported by Arthur Lemmens sbcl-devel 2001-05-05, ANSI
694 requires that SYMBOL-MACROLET refuse to rebind special variables,
695 but SBCL doesn't do this. (Also as reported by AL in the same
696 message, SBCL depended on this nonconforming behavior to build
697 itself, because of the way that **CURRENT-SEGMENT** was implemented.
698 As of sbcl-0.6.12.x, this dependence on the nonconforming behavior
699 has been fixed, but the nonconforming behavior remains.)
702 (DESCRIBE 'SB-ALIEN:DEF-ALIEN-TYPE) reports the macro argument list
706 in #<PACKAGE "SB-ALIEN">.
707 Macro-function: #<FUNCTION "DEF!MACRO DEF-ALIEN-TYPE" {19F4A39}>
708 Macro arguments: (#:whole-470 #:environment-471)
709 On Sat, May 26, 2001 09:45:57 AM CDT it was compiled from:
710 /usr/stuff/sbcl/src/code/host-alieneval.lisp
711 Created: Monday, March 12, 2001 07:47:43 AM CST
714 (TIME (ROOM T)) reports more than 200 Mbytes consed even for
715 a clean, just-started SBCL system. And it seems to be right:
716 (ROOM T) can bring a small computer to its knees for a *long*
717 time trying to GC afterwards. Surely there's some more economical
718 way to implement (ROOM T).
721 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
723 ;;; This file fails to compile.
724 ;;; Maybe this bug is related to bugs #65, #70 in the BUGS file.
725 (in-package :cl-user)
731 ;; Uncomment and it works
734 In SBCL 0.6.12.42, the problem is
735 internal error, failed AVER:
736 "(COMMON-LISP:EQ (SB!C::LAMBDA-TAIL-SET SB!C::CALLER)
737 (SB!C::LAMBDA-TAIL-SET (SB!C::LAMBDA-HOME SB!C::CALLEE)))"
740 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
742 ;;; The compiler is flushing the argument type test, and the default
743 ;;; case in the cond, so that calling with say a fixnum 0 causes a
745 (declaim (optimize (safety 2) (speed 3)))
747 (declare (type (or string stream) x))
748 (cond ((typep x 'string) 'string)
749 ((typep x 'stream) 'stream)
752 The symptom in sbcl-0.6.12.42 on OpenBSD is actually (TST 0)=>STREAM
753 (not the SIGBUS reported in the comment) but that's broken too;
754 type declarations are supposed to be treated as assertions unless
755 SAFETY 0, so we should be getting a TYPE-ERROR.
758 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
760 (in-package :cl-user)
761 ;;; From: David Gadbois <gadbois@cyc.com>
763 ;;; Logical pathnames aren't externalizable.
765 (let ((tempfile "/tmp/test.lisp"))
766 (setf (logical-pathname-translations "XXX")
767 '(("XXX:**;*.*" "/tmp/**/*.*")))
768 (with-open-file (out tempfile :direction :output)
769 (write-string "(defvar *path* #P\"XXX:XXX;FOO.LISP\")" out))
770 (compile-file tempfile))
771 The error message in sbcl-0.6.12.42 is
773 ; (while making load form for #<SB-IMPL::LOGICAL-HOST "XXX">)
774 ; A logical host can't be dumped as a constant: #<SB-IMPL::LOGICAL-HOST "XXX">
777 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
779 (in-package :cl-user)
780 ;;; The following invokes a compiler error.
781 (declaim (optimize (speed 2) (debug 3)))
784 (unwind-protect nil)))
788 The error message in sbcl-0.6.12.42 is
789 internal error, failed AVER:
790 "(COMMON-LISP:EQ (SB!C::TN-ENVIRONMENT SB!C:TN) SB!C::TN-ENV)"
793 When the compiler inline expands functions, it may be that different
794 kinds of return values are generated from different code branches.
795 E.g. an inline expansion of POSITION generates integer results
796 from one branch, and NIL results from another. When that inline
797 expansion is used in a context where only one of those results
800 (aref *a1* (position x *a2*)))
801 and the compiler can't prove that the unacceptable branch is
802 never taken, then bogus type mismatch warnings can be generated.
803 If you need to suppress the type mismatch warnings, you can
804 suppress the inline expansion,
806 #+sbcl (declare (notinline position)) ; to suppress bug 117 bogowarnings
807 (aref *a1* (position x *a2*)))
808 or, sometimes, suppress them by declaring the result to be of an
811 (aref *a1* (the integer (position x *a2*))))
813 This is not a new compiler problem in 0.7.0, but the new compiler
814 transforms for FIND, POSITION, FIND-IF, and POSITION-IF make it
815 more conspicuous. If you don't need performance from these functions,
816 and the bogus warnings are a nuisance for you, you can return to
817 your pre-0.7.0 state of grace with
818 #+sbcl (declaim (notinline find position find-if position-if)) ; bug 117..
821 as reported by Eric Marsden on cmucl-imp@cons.org 2001-08-14:
822 (= (FLOAT 1 DOUBLE-FLOAT-EPSILON)
823 (+ (FLOAT 1 DOUBLE-FLOAT-EPSILON) DOUBLE-FLOAT-EPSILON)) => T
824 when of course it should be NIL. (He says it only fails for X86,
825 not SPARC; dunno about Alpha.)
827 Also, "the same problem exists for LONG-FLOAT-EPSILON,
828 DOUBLE-FLOAT-NEGATIVE-EPSILON, LONG-FLOAT-NEGATIVE-EPSILON (though
829 for the -negative- the + is replaced by a - in the test)."
831 Raymond Toy comments that this is tricky on the X86 since its FPU
832 uses 80-bit precision internally.
835 The compiler incorrectly figures the return type of
836 (DEFUN FOO (FRAME UP-FRAME)
843 This problem exists in CMU CL 18c too. When I reported it on
844 cmucl-imp@cons.org, Raymond Toy replied 23 Aug 2001 with
845 a partial explanation, but no fix has been found yet.
848 Even in sbcl-0.pre7.x, which is supposed to be free of the old
849 non-ANSI behavior of treating the function return type inferred
850 from the current function definition as a declaration of the
851 return type from any function of that name, the return type of NIL
852 is attached to FOO in 120a above, and used to optimize code which
856 There was some sort of screwup in handling of
857 (IF (NOT (IGNORE-ERRORS ..))). E.g.
859 (if (not (ignore-errors
860 (make-pathname :host "foo" :directory "!bla" :name "bar")))
862 (error "notunlessnot")))
863 The (NOT (IGNORE-ERRORS ..)) form evaluates to T, so this should be
864 printing "ok", but instead it's going to the ERROR. This problem
865 seems to've been introduced by MNA's HANDLER-CASE patch (sbcl-devel
866 2001-07-17) and as a workaround (put in sbcl-0.pre7.14.flaky4.12)
867 I reverted back to the old weird HANDLER-CASE code. However, I
868 think the problem looks like a compiler bug in handling RETURN-FROM,
869 so I left the MNA-patched code in HANDLER-CASE (suppressed with
870 #+NIL) and I'd like to go back to see whether this really is
871 a compiler bug before I delete this BUGS entry.
874 The *USE-IMPLEMENTATION-TYPES* hack causes bugs, particularly
875 (IN-PACKAGE :SB-KERNEL)
876 (TYPE= (SPECIFIER-TYPE '(VECTOR T))
877 (SPECIFIER-TYPE '(VECTOR UNDEFTYPE)))
878 Then because of this, the compiler bogusly optimizes
879 (TYPEP #(11) '(SIMPLE-ARRAY UNDEF-TYPE 1))
880 to T. Unfortunately, just setting *USE-IMPLEMENTATION-TYPES* to
881 NIL around sbcl-0.pre7.14.flaky4.12 didn't work: the compiler complained
882 about type mismatches (probably harmlessly, another instance of bug 117);
883 and then cold init died with a segmentation fault.
886 As of version 0.pre7.14, SBCL's implementation of MACROLET makes
887 the entire lexical environment at the point of MACROLET available
888 in the bodies of the macroexpander functions. In particular, it
889 allows the function bodies (which run at compile time) to try to
890 access lexical variables (which are only defined at runtime).
891 It doesn't even issue a warning, which is bad.
893 The SBCL behavior arguably conforms to the ANSI spec (since the
894 spec says that the behavior is undefined, ergo anything conforms).
895 However, it would be better to issue a compile-time error.
896 Unfortunately I (WHN) don't see any simple way to detect this
897 condition in order to issue such an error, so for the meantime
898 SBCL just does this weird broken "conforming" thing.
900 The ANSI standard says, in the definition of the special operator
902 The macro-expansion functions defined by MACROLET are defined
903 in the lexical environment in which the MACROLET form appears.
904 Declarations and MACROLET and SYMBOL-MACROLET definitions affect
905 the local macro definitions in a MACROLET, but the consequences
906 are undefined if the local macro definitions reference any
907 local variable or function bindings that are visible in that
909 Then it seems to contradict itself by giving the example
911 (macrolet ((fudge (z)
912 ;The parameters x and flag are not accessible
913 ; at this point; a reference to flag would be to
914 ; the global variable of that name.
915 ` (if flag (* ,z ,z) ,z)))
916 ;The parameters x and flag are accessible here.
920 The comment "a reference to flag would be to the global variable
921 of the same name" sounds like good behavior for the system to have.
922 but actual specification quoted above says that the actual behavior
926 (as reported by Gabe Garza on cmucl-help 2001-09-21)
928 (defun test-pred (x y)
932 (func (lambda () x)))
933 (print (eq func func))
934 (print (test-pred func func))
935 (delete func (list func))))
936 Now calling (TEST-CASE) gives output
939 (#<FUNCTION {500A9EF9}>)
940 Evidently Python thinks of the lambda as a code transformation so
941 much that it forgets that it's also an object.
947 The DEFSTRUCT section of the ANSI spec, in the :CONC-NAME section,
948 specifies a precedence rule for name collisions between slot accessors of
949 structure classes related by inheritance. As of 0.7.0, SBCL still
953 insufficient syntax checking in MACROLET:
955 (macrolet ((defmacro bar (z) `(+ z z)))
957 shouldn't compile without error (because of the extra DEFMACRO symbol).
960 As of sbcl-0.pre7.86.flaky7.3, the cross-compiler, and probably
961 the CL:COMPILE function (which is based on the same %COMPILE
962 mechanism) get confused by
964 (labels ((sxhash-number (x)
966 (fixnum (sxhash x)) ; through DEFTRANSFORM
967 (integer (sb!bignum:sxhash-bignum x))
968 (single-float (sxhash x)) ; through DEFTRANSFORM
969 (double-float (sxhash x)) ; through DEFTRANSFORM
970 #!+long-float (long-float (error "stub: no LONG-FLOAT"))
971 (ratio (let ((result 127810327))
972 (declare (type fixnum result))
973 (mixf result (sxhash-number (numerator x)))
974 (mixf result (sxhash-number (denominator x)))
976 (complex (let ((result 535698211))
977 (declare (type fixnum result))
978 (mixf result (sxhash-number (realpart x)))
979 (mixf result (sxhash-number (imagpart x)))
981 (sxhash-recurse (x &optional (depthoid +max-hash-depthoid+))
982 (declare (type index depthoid))
986 (mix (sxhash-recurse (car x) (1- depthoid))
987 (sxhash-recurse (cdr x) (1- depthoid)))
990 (if (typep x 'structure-object)
992 (sxhash ; through DEFTRANSFORM
993 (class-name (layout-class (%instance-layout x)))))
995 (symbol (sxhash x)) ; through DEFTRANSFORM
996 (number (sxhash-number x))
999 (simple-string (sxhash x)) ; through DEFTRANSFORM
1000 (string (%sxhash-substring x))
1001 (bit-vector (let ((result 410823708))
1002 (declare (type fixnum result))
1003 (dotimes (i (min depthoid (length x)))
1004 (mixf result (aref x i)))
1006 (t (logxor 191020317 (sxhash (array-rank x))))))
1009 (sxhash (char-code x)))) ; through DEFTRANSFORM
1011 (sxhash-recurse x)))
1012 complaining "function called with two arguments, but wants exactly
1013 one" about SXHASH-RECURSE. (This might not be strictly a new bug,
1014 since IIRC post-fork CMU CL has also had problems with &OPTIONAL
1015 arguments in FLET/LABELS: it might be an old Python bug which is
1016 only exercised by the new arrangement of the SBCL compiler.)
1019 Ideally, uninterning a symbol would allow it, and its associated
1020 FDEFINITION and PROCLAIM data, to be reclaimed by the GC. However,
1021 at least as of sbcl-0.7.0, this isn't the case. Information about
1022 FDEFINITIONs and PROCLAIMed properties is stored in globaldb.lisp
1023 essentially in ordinary (non-weak) hash tables keyed by symbols.
1024 Thus, once a system has an entry in this system, it tends to live
1025 forever, even when it is uninterned and all other references to it
1029 (reported by Arnaud Rouanet on cmucl-imp 2001-12-18)
1030 (defmethod foo ((x integer))
1032 (defmethod foo :around ((x integer))
1034 (call-next-method)))
1035 Now (FOO 3) should return 3, but instead it returns 4.
1038 (SB-DEBUG:BACKTRACE) output should start with something
1039 including the name BACKTRACE, not (as in 0.pre7.88)
1040 just "0: (\"hairy arg processor\" ...)". Until about
1041 sbcl-0.pre7.109, the names in BACKTRACE were all screwed
1042 up compared to the nice useful names in sbcl-0.6.13.
1043 Around sbcl-0.pre7.109, they were mostly fixed by using
1044 NAMED-LAMBDA to implement DEFUN. However, there are still
1045 some screwups left, e.g. as of sbcl-0.pre7.109, there are
1046 still some functions named "hairy arg processor" and
1047 "SB-INT:&MORE processor".
1050 (reported by Alexey Dejneka sbcl-devel 2002-01-03)
1052 SUBTYPEP does not work well with redefined classes:
1054 * (defclass a () ())
1056 * (defclass b () ())
1061 * (defclass b (a) ())
1066 * (defclass b () ())
1074 This is probably due to underzealous clearing of the type caches; a
1075 brute-force solution in that case would be to make a defclass expand
1076 into something that included a call to SB-KERNEL::CLEAR-TYPE-CACHES,
1077 but there may be a better solution.
1080 Pretty-printing nested backquotes doesn't work right, as
1081 reported by Alexey Dejneka sbcl-devel 2002-01-13:
1083 ``(FOO SB-IMPL::BACKQ-COMMA-AT S)
1084 * (lisp-implementation-version)
1088 (as reported by Lynn Quam on cmucl-imp ca. 2002-01-16)
1089 %NATURALIZE-C-STRING conses a lot, like 16 bytes per byte
1090 of the naturalized string. We could probably port the patches
1091 from the cmucl-imp mailing list.
1094 (reported by Jesse Bouwman 2001-10-24 through the unfortunately
1095 prominent SourceForge web/db bug tracking system, which is
1096 unfortunately not a reliable way to get a timely response from
1097 the SBCL maintainers)
1098 In the course of trying to build a test case for an
1099 application error, I encountered this behavior:
1100 If you start up sbcl, and then lay on CTRL-C for a
1101 minute or two, the lisp process will eventually say:
1102 %PRIMITIVE HALT called; the party is over.
1103 and throw you into the monitor. If I start up lisp,
1104 attach to the process with strace, and then do the same
1105 (abusive) thing, I get instead:
1106 access failure in heap page not marked as write-protected
1107 and the monitor again. I don't know enough to have the
1108 faintest idea of what is going on here.
1109 This is with sbcl 6.12, uname -a reports:
1110 Linux prep 2.2.19 #4 SMP Tue Apr 24 13:59:52 CDT 2001 i686 unknown
1111 I (WHN) have verified that the same thing occurs on sbcl-0.pre7.141
1112 under OpenBSD 2.9 on my X86 laptop. Do be patient when you try it:
1113 it took more than two minutes (but less than five) for me.
1116 (This was once known as IR1-4, but it lived on even after the
1117 IR1 interpreter went to the big bit bucket in the sky.)
1118 The system accepts DECLAIM in most places where DECLARE would be
1119 accepted, without even issuing a warning. ANSI allows this, but since
1120 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
1121 meaning is rather different, and it's unlikely that the user
1122 has a good reason for doing DECLAIM not at top level, it would be
1123 good to issue a STYLE-WARNING when this happens. A possible
1124 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
1125 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
1126 [This is considered an IR1-interpreter-related bug because until
1127 EVAL-WHEN is rewritten, which won't happen until after the IR1
1128 interpreter is gone, the system's notion of what's a top-level form
1129 and what's not will remain too confused to fix this problem.]
1132 ANSI allows types `(COMPLEX ,FOO) to use very hairy values for
1133 FOO, e.g. (COMPLEX (AND REAL (SATISFIES ODDP))). The old CMU CL
1134 COMPLEX implementation didn't deal with this, and hasn't been
1135 upgraded to do so. (This doesn't seem to be a high priority
1136 conformance problem, since seems hard to construct useful code
1140 Floating point errors are reported poorly. E.g. on x86 OpenBSD
1143 debugger invoked on condition of type SB-KERNEL:FLOATING-POINT-EXCEPTION:
1144 An arithmetic error SB-KERNEL:FLOATING-POINT-EXCEPTION was signalled.
1145 No traps are enabled? How can this be?
1146 It should be possible to be much more specific (overflow, division
1147 by zero, etc.) and of course the "How can this be?" should be fixable.
1150 (reported by Alexey Dejneka sbcl-devel 2002-01-28)
1151 Compiling a file containing
1152 (deftype digit () '(member #\1))
1153 (defun parse-num (string ind)
1156 (if (and (< ind ind)
1157 (typep (char string ind) 'digit))
1159 in sbcl-0.7.1 causes the compiler to fail with
1160 internal error, failed AVER: "(= (LENGTH (BLOCK-SUCC CALL-BLOCK)) 1)"
1161 This problem seems to have been introduced by the sbcl-0.pre7.* compiler
1162 changes, since 0.pre7.73 and 0.6.13 don't suffer from it. A related
1164 (defun parse-num (index)
1171 (when (and (digs) (digs)) x))))
1172 In sbcl-0.7.1, this second test case failed with the same
1173 internal error, failed AVER: "(= (LENGTH (BLOCK-SUCC CALL-BLOCK)) 1)"
1174 After the APD patches in sbcl-0.7.1.2 (new consistency check in
1175 TARGET-IF-DESIRABLE, plus a fix in meta-vmdef.lisp to keep the
1176 new consistency check from failing routinely) this second test case
1177 failed in FIND-IN-PHYSENV instead. Fixes in sbcl-0.7.1.3 (not
1178 closing over unreferenced variables) made this second test case
1179 compile without error, but the original test case still fails.
1181 Another way to get rid of the DEFTYPE without changing the symptom
1184 (defun parse-num (string ind)
1187 (if (and (< ind ind)
1188 (sb-int:memq *ch* '(#\1)))
1190 In sbcl-0.7.1.3, this fails with
1191 internal error, failed AVER: "(= (LENGTH (BLOCK-SUCC CALL-BLOCK)) 1)"
1192 The problem occurs while the inline expansion of MEMQ,
1193 #<LAMBDA :%DEBUG-NAME "varargs entry point for SB-C::.ANONYMOUS.">
1194 is being LET-converted after having its second REF deleted, leaving
1195 it with only one entry in LEAF-REFS.
1198 In sbcl-0.7.1.3 on x86, COMPILE-FILE on the file
1199 (in-package :cl-user)
1202 (defstruct foo bar bletch)
1204 (labels ((kidify1 (kid)
1209 (m+ (frobnicate kid)
1212 (declare (inline kid-frob))
1215 (the simple-vector (foo-bar perd)))))
1217 debugger invoked on condition of type TYPE-ERROR:
1218 The value NIL is not of type SB-C::NODE.
1219 The location of this failure has moved around as various related
1220 issues were cleaned up. As of sbcl-0.7.1.9, it occurs in
1221 NODE-BLOCK called by LAMBDA-COMPONENT called by IR2-CONVERT-CLOSURE.
1224 (essentially the same problem as a CMU CL bug reported by Martin
1225 Cracauer on cmucl-imp 2002-02-19)
1226 There is a hole in structure slot type checking. Compiling and LOADing
1227 (declaim (optimize safety))
1229 (bla 0 :type fixnum))
1231 (let ((foo (make-foo)))
1232 (setf (foo-bla foo) '(1 . 1))
1233 (format t "Is ~a of type ~a a cons? => ~a~%"
1235 (type-of (foo-bla foo))
1236 (consp (foo-bla foo)))))
1238 should signal an error, but in sbcl-0.7.1.21 instead gives the output
1239 Is (1 . 1) of type CONS a cons? => NIL
1240 without signalling an error.
1243 There's some sort of problem with aborting back out of the debugger
1244 after a %DETECT-STACK-EXHAUSTION error in sbcl-0.7.1.38. In some cases
1245 telling the debugger to ABORT doesn't get you back to the main REPL,
1246 but instead just gives you another stack exhaustion error. The problem
1247 doesn't occur in the trivial case
1248 * (defun frob () (frob) (frob))
1251 but it has happened in more complicated cases (which I haven't
1252 figured out how to reproduce).
1256 (defclass standard-gadget (basic-gadget) ())
1257 (defclass basic-gadget () ())
1259 The slot SB-PCL::DIRECT-SUPERCLASSES is unbound in the
1260 object #<SB-PCL::STANDARD-CLASS "unbound">.
1261 (reported by Brian Spilsbury sbcl-devel 2002-04-09)
1264 FUNCTION-LAMBDA-EXPRESSION doesn't work right in 0.7.0 or 0.7.2.9:
1265 * (function-lambda-expression #'(lambda (x) x))
1266 debugger invoked on condition of type TYPE-ERROR:
1267 The value NIL is not of type SB-C::DEBUG-SOURCE
1268 (reported by Alexey Dejneka sbcl-devel 2002-04-12)
1271 Functions SUBTYPEP, TYPEP, UPGRADED-ARRAY-ELEMENT-TYPE, and
1272 UPGRADED-COMPLEX-PART-TYPE should have an optional environment argument.
1273 (reported by Alexey Dejneka sbcl-devel 2002-04-12)
1276 Compiling the following code causes SBCL 0.7.2 to bug. This only
1277 happens with optimization enabled, and only when the loop variable is
1278 being incremented by more than 1.
1280 (declare (optimize (safety 0) (space 0) (debug 0) (speed 3)))
1281 (loop for i from 0 to 10 by 2
1282 do (foo (svref array i))) (svref array (1+ i)))
1283 (reported by Eric Marsden sbcl-devel 2002-04-15)
1286 * (lisp-implementation-version)
1289 (:constructor make-foo (&key (bar nil bar-p)
1290 &aux (baz (if bar-p bar 2)))))
1294 debugger invoked on condition of type SB-KERNEL::ARG-COUNT-ERROR:
1295 error while parsing arguments to DESTRUCTURING-BIND:
1296 invalid number of elements in
1298 to satisfy lambda list
1299 (SB-KERNEL::WOT &OPTIONAL (SB-KERNEL::DEF NIL SB-KERNEL::DEF-P)):
1300 between 1 and 2 expected, but 3 found
1301 (reported by Christophe Rhodes and Martin Atzmueller sbcl-devel
1305 USER-HOMEDIR-PATHNAME returns a pathname that SBCL can't do anything
1306 with. Probably we should return an absolute physical pathname
1307 instead. (Reported by Peter van Eynde sbcl-devel 2002-03-29)
1310 Typep on certain SATISFIES types doesn't take account of the fact
1311 that the function could cause an error; e.g. (TYPEP #\! '(SATISFIES
1312 FBOUNDP)) raises an error when it should return NIL.
1315 (reported by Robert E. Brown 2002-04-16)
1316 When a function is called with too few arguments, causing the
1317 debugger to be entered, the uninitialized slots in the bad call frame
1318 seem to cause GCish problems, being interpreted as tagged data even
1319 though they're not. In particular, executing ROOM in the
1320 debugger at that point causes AVER failures:
1323 * (lisp-implementation-version)
1329 failed AVER: "(SAP= CURRENT END)"
1330 (Christophe Rhodes reports that this doesn't occur on the SPARC, which
1331 isn't too surprising since there are many differences in stack
1332 implementation and GC conservatism between the X86 and other ports.)
1334 DEFUNCT CATEGORIES OF BUGS
1336 These labels were used for bugs related to the old IR1 interpreter.
1337 The # values reached 6 before the category was closed down.