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.
92 As of sbcl-0.7.5, sbcl's cross-compiler does run with
93 *TYPE-SYSTEM-INITIALIZED*; however, this bug remains.
96 The "compiling top-level form:" output ought to be condensed.
97 Perhaps any number of such consecutive lines ought to turn into a
98 single "compiling top-level forms:" line.
101 The way that the compiler munges types with arguments together
102 with types with no arguments (in e.g. TYPE-EXPAND) leads to
103 weirdness visible to the user:
104 (DEFTYPE FOO () 'FIXNUM)
106 (TYPEP 11 '(FOO)) => T, which seems weird
107 (TYPEP 11 'FIXNUM) => T
108 (TYPEP 11 '(FIXNUM)) signals an error, as it should
109 The situation is complicated by the presence of Common Lisp types
110 like UNSIGNED-BYTE (which can either be used in list form or alone)
111 so I'm not 100% sure that the behavior above is actually illegal.
112 But I'm 90+% sure, and the following related behavior,
114 treating the bare symbol AND as equivalent to '(AND), is specifically
115 forbidden (by the ANSI specification of the AND type).
118 It would be nice if the
120 (during macroexpansion)
121 said what macroexpansion was at fault, e.g.
123 (during macroexpansion of IN-PACKAGE,
124 during macroexpansion of DEFFOO)
127 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
128 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
129 (Also, when this is fixed, we can enable the code in PROCLAIM which
130 checks for incompatible FTYPE redeclarations.)
133 (I *think* this is a bug. It certainly seems like strange behavior. But
134 the ANSI spec is scary, dark, and deep.. -- WHN)
135 (FORMAT NIL "~,1G" 1.4) => "1. "
136 (FORMAT NIL "~3,1G" 1.4) => "1. "
139 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
141 (setf (find-class 'ccc1) (find-class 'ccc))
142 (defmethod zut ((c ccc1)) 123)
143 In sbcl-0.7.1.13, this gives an error,
144 There is no class named CCC1.
145 DTC's recommended workaround from the mailing list 3 Mar 2000:
146 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
149 Sometimes (SB-EXT:QUIT) fails with
150 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
151 Process inferior-lisp exited abnormally with code 1
152 I haven't noticed a repeatable case of this yet.
155 The printer doesn't report closures very well. This is true in
159 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
160 It would be nice to make closures have a settable name slot,
161 and make things like DEFSTRUCT and FLET, which create closures,
162 set helpful values into this slot.
165 And as long as we're wishing, it would be awfully nice if INSPECT could
166 also report on closures, telling about the values of the bound variables.
169 The compiler assumes that any time a function of declared FTYPE
170 doesn't signal an error, its arguments were of the declared type.
171 E.g. compiling and loading
172 (DECLAIM (OPTIMIZE (SAFETY 3)))
173 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
175 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE)) FACTORIAL))
177 (COND ((> (FACTORIAL X) 1.0E6)
178 (FORMAT T "too big~%"))
180 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
182 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
185 will cause the INTEGERP case to be selected, giving bogus output a la
187 (or (FOO 1000.5), "exactly 1001.5")
188 This violates the "declarations are assertions" principle.
189 According to the ANSI spec, in the section "System Class FUNCTION",
190 this is a case of "lying to the compiler", but the lying is done
191 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
192 not by the unexpectedly general definition of FACTORIAL. In any case,
193 "declarations are assertions" means that lying to the compiler should
194 cause an error to be signalled, and should not cause a bogus
195 result to be returned. Thus, the compiler should not assume
196 that arbitrary functions check their argument types. (It might
197 make sense to add another flag (CHECKED?) to DEFKNOWN to
198 identify functions which *do* check their argument types.)
199 (Also, verify that the compiler handles declared function
200 return types as assertions.)
203 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
204 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
206 (DECLARE (TYPE INDEXOID X))
207 (THE (VALUES INDEXOID)
209 where the implementation of the type check in function FOO
210 includes a full call to %TYPEP. There are also some fundamental problems
211 with the interpretation of VALUES types (inherited from CMU CL, and
212 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
213 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
216 The definitions of SIGCONTEXT-FLOAT-REGISTER and
217 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
218 supported on FreeBSD because the floating point state is not saved,
219 but at least as of FreeBSD 4.0, the floating point state *is* saved,
220 so they could be supported after all. Very likely
221 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
224 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
225 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
226 issue with 'union, 'and, 'or etc. So even though according to the
227 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
228 (and now SBCL) interpret them as legal types.
231 a slew of floating-point-related errors reported by Peter Van Eynde
233 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
234 should probably be 1.4012985e-45. In SBCL,
235 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
236 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
237 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
238 and LEAST-NEGATIVE-LONG-FLOAT.
239 c: Many expressions generate floating infinity on x86/Linux:
244 PVE's regression tests want them to raise errors. sbcl-0.7.0.5
245 on x86/Linux generates the infinities instead. That might or
246 might not be conforming behavior, but it's also inconsistent,
247 which is almost certainly wrong. (Inconsistency: (/ 1 0.0)
248 should give the same result as (/ 1.0 0.0), but instead (/ 1 0.0)
249 generates SINGLE-FLOAT-POSITIVE-INFINITY and (/ 1.0 0.0)
251 d: (in section12.erg) various forms a la
252 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
253 don't give the right behavior.
256 type safety errors reported by Peter Van Eynde July 25, 2000:
257 a: (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
259 In general lengths of array type specifications aren't
260 checked by COERCE, so it fails when the spec is
261 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
262 b: CONCATENATE has the same problem of not checking the length
263 of specified output array types. MAKE-SEQUENCE and MAP and
264 MERGE also have the same problem.
265 c: (COERCE 'AND 'FUNCTION) returns something related to
266 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
267 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
268 should signal TYPE-ERROR.
269 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
270 be used for input and output as needed. It should fail with
271 TYPE-ERROR when handed e.g. the results of
272 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
273 the inappropriate positions, but doesn't.
274 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
275 not a binary input stream, but instead cheerfully reads from
276 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
279 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
280 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
281 causes a COMPILER-ERROR.
284 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
285 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
286 it binds is declared SPECIAL inside.
289 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
291 (DEFGENERIC FOO02 (X))
292 (DEFMETHOD FOO02 ((X NUMBER)) T)
293 (LET ((M (FIND-METHOD (FUNCTION FOO02)
295 (LIST (FIND-CLASS (QUOTE NUMBER))))))
296 (REMOVE-METHOD (FUNCTION FOO02) M)
297 (DEFGENERIC FOO03 (X))
298 (ADD-METHOD (FUNCTION FOO03) M)))
299 should give an error, but SBCL allows it.
302 It has been reported (e.g. by Peter Van Eynde) that there are
303 several metaobject protocol "errors". (In order to fix them, we might
304 need to document exactly what metaobject protocol specification
305 we're following -- the current code is just inherited from PCL.)
308 The implementation of #'+ returns its single argument without
309 type checking, e.g. (+ "illegal") => "illegal".
312 The debugger LIST-LOCATIONS command doesn't work properly.
315 Compiling and loading
316 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
318 then requesting a BACKTRACE at the debugger prompt gives no information
319 about where in the user program the problem occurred.
322 The compiler is supposed to do type inference well enough that
325 ((SIMPLE-ARRAY SINGLE-FLOAT)
327 (DECLARE (TYPE (SIMPLE-ARRAY SINGLE-FLOAT) X))
330 is redundant. However, as reported by Juan Jose Garcia Ripoll for
331 CMU CL, it sometimes doesn't. Adding declarations is a pretty good
332 workaround for the problem for now, but can't be done by the TYPECASE
333 macros themselves, since it's too hard for the macro to detect
334 assignments to the variable within the clause.
335 Note: The compiler *is* smart enough to do the type inference in
336 many cases. This case, derived from a couple of MACROEXPAND-1
337 calls on Ripoll's original test case,
339 (DECLARE (OPTIMIZE SPEED (SAFETY 0)))
340 (COND ((TYPEP A '(SIMPLE-ARRAY SINGLE-FLOAT)) NIL
341 (LET ((LENGTH (ARRAY-TOTAL-SIZE A)))
342 (LET ((I 0) (G2554 LENGTH))
343 (DECLARE (TYPE REAL G2554) (TYPE REAL I))
346 (WHEN (>= I G2554) (GO SB-LOOP::END-LOOP))
347 (SETF (ROW-MAJOR-AREF A I) (- (ROW-MAJOR-AREF A I)))
348 (GO SB-LOOP::NEXT-LOOP)
349 SB-LOOP::END-LOOP))))))
350 demonstrates the problem; but the problem goes away if the TAGBODY
351 and GO forms are removed (leaving the SETF in ordinary, non-looping
352 code), or if the TAGBODY and GO forms are retained, but the
353 assigned value becomes 0.0 instead of (- (ROW-MAJOR-AREF A I)).
356 Paul Werkowski wrote on cmucl-imp@cons.org 2000-11-15
357 I am looking into this problem that showed up on the cmucl-help
358 list. It seems to me that the "implementation specific environment
359 hacking functions" found in pcl/walker.lisp are completely messed
360 up. The good thing is that they appear to be barely used within
361 PCL and the munged environment object is passed to cmucl only
362 in calls to macroexpand-1, which is probably why this case fails.
363 SBCL uses essentially the same code, so if the environment hacking
364 is screwed up, it affects us too.
367 Using the pretty-printer from the command prompt gives funny
368 results, apparently because the pretty-printer doesn't know
369 about user's command input, including the user's carriage return
370 that the user, and therefore the pretty-printer thinks that
371 the new output block should start indented 2 or more characters
372 rightward of the correct location.
375 ANSI specifies that the RESULT-TYPE argument of CONCATENATE must be
376 a subtype of SEQUENCE, but CONCATENATE doesn't check this properly:
377 (CONCATENATE 'SIMPLE-ARRAY #(1 2) '(3)) => #(1 2 3)
378 This also leads to funny behavior when derived type specifiers
379 are used, as originally reported by Milan Zamazal for CMU CL (on the
380 Debian bugs mailing list (?) 2000-02-27), then reported by Martin
381 Atzmueller for SBCL (2000-10-01 on sbcl-devel@lists.sourceforge.net):
382 (DEFTYPE FOO () 'SIMPLE-ARRAY)
383 (CONCATENATE 'FOO #(1 2) '(3))
384 => #<ARRAY-TYPE SIMPLE-ARRAY> is a bad type specifier for
386 The derived type specifier FOO should act the same way as the
387 built-in type SIMPLE-ARRAY here, but it doesn't. That problem
388 doesn't seem to exist for sequence types:
389 (DEFTYPE BAR () 'SIMPLE-VECTOR)
390 (CONCATENATE 'BAR #(1 2) '(3)) => #(1 2 3)
391 See also bug #46a./b., and discussion and patch sbcl-devel and
395 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
396 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
397 crashes SBCL. In general tracing anything which is used in the
398 implementation of TRACE is likely to have the same problem.
401 (DECLAIM (OPTIMIZE ..)) doesn't work properly inside LOCALLY forms.
404 As reported by Martin Atzmueller on sbcl-devel 26 Dec 2000,
405 ANSI says that WITH-OUTPUT-TO-STRING should have a keyword
406 :ELEMENT-TYPE, but in sbcl-0.6.9 this is not defined for
407 WITH-OUTPUT-TO-STRING.
410 ANSI says in one place that type declarations can be abbreviated even
411 when the type name is not a symbol, e.g.
412 (DECLAIM ((VECTOR T) *FOOVECTOR*))
413 SBCL doesn't support this. But ANSI says in another place that this
414 isn't allowed. So it's not clear this is a bug after all. (See the
415 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
419 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
420 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
421 an easily guessable temporary filename in a way which might open
422 applications using LOAD-FOREIGN to hijacking by malicious users
423 on the same machine. Incantations for doing this safely are
424 floating around the net in various "how to write secure programs
425 despite Unix" documents, and it would be good to (1) fix this in
426 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
427 files and make it share the same new safe logic.
430 Functions are assigned names based on the context in which they're
431 defined. This is less than ideal for the functions which are
432 used to implement CLOS methods. E.g. the output of
433 (DESCRIBE 'PRINT-OBJECT) lists functions like
434 #<FUNCTION "DEF!STRUCT (TRACE-INFO (:MAKE-LOAD-FORM-FUN SB-KERNEL:JUST-DUMP-IT-NORMALLY) (:PRINT-OBJECT #))" {1020E49}>
436 #<FUNCTION "MACROLET ((FORCE-DELAYED-DEF!METHODS NIL #))" {1242871}>
437 It would be better if these functions' names always identified
438 them as methods, and identified their generic functions and
442 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
443 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
444 the mean of the distribution can be systematically O(0.1%) wrong.
445 Just increasing R-I-E-B is probably not a good solution, since
446 it would decrease efficiency more than is probably necessary. Perhaps
447 using some sort of accept/reject method would be better.
450 Internally the compiler sometimes evaluates
451 (sb-kernel:type/= (specifier-type '*) (specifier-type t))
452 (I stumbled across this when I added an
453 (assert (not (eq type1 *wild-type*)))
454 in the NAMED :SIMPLE-= type method.) '* isn't really a type, and
455 in a type context should probably be translated to T, and so it's
456 probably wrong to ask whether it's equal to the T type and then (using
457 the EQ type comparison in the NAMED :SIMPLE-= type method) return NIL.
458 (I haven't tried to investigate this bug enough to guess whether
459 there might be any user-level symptoms.)
462 Inconsistencies between derived and declared VALUES return types for
463 DEFUN aren't checked very well. E.g. the logic which successfully
464 catches problems like
465 (declaim (ftype (function (fixnum) float) foo))
467 (declare (type integer x))
468 (values x)) ; wrong return type, detected, gives warning, good!
470 (declaim (ftype (function (t) (values t t)) bar))
472 (values x)) ; wrong number of return values, no warning, bad!
473 The cause of this is seems to be that (1) the internal function
474 VALUES-TYPES-EQUAL-OR-INTERSECT used to make the check handles its
475 arguments symmetrically, and (2) when the type checking code was
476 written back when when SBCL's code was still CMU CL, the intent
478 (declaim (ftype (function (t) t) bar))
480 (values x x)) ; wrong number of return values; should give warning?
481 not be warned for, because a two-valued return value is considered
482 to be compatible with callers who expects a single value to be
483 returned. That intent is probably not appropriate for modern ANSI
484 Common Lisp, but fixing this might be complicated because of other
485 divergences between auld-style and new-style handling of
486 multiple-VALUES types. (Some issues related to this were discussed
487 on cmucl-imp at some length sometime in 2000.)
490 The facility for dumping a running Lisp image to disk gets confused
491 when run without the PURIFY option, and creates an unnecessarily large
492 core file (apparently representing memory usage up to the previous
493 high-water mark). Moreover, when the file is loaded, it confuses the
494 GC, so that thereafter memory usage can never be reduced below that
498 In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
499 CL), out-of-line structure slot setters are horribly inefficient
500 whenever the type of the slot is declared, because out-of-line
501 structure slot setters are implemented as closures to save space,
502 so the compiler doesn't compile the type test into code, but
503 instead just saves the type in a lexical closure and interprets it
505 A proper solution involves deciding whether it's really worth
506 saving space by implementing structure slot accessors as closures.
507 (If it's not worth it, the problem vanishes automatically. If it
508 is worth it, there are hacks we could use to force type tests to
509 be compiled anyway, and even shared. E.g. we could implement
510 an EQUAL hash table mapping from types to compiled type tests,
511 and save the appropriate compiled type test as part of each lexical
512 closure; or we could make the lexical closures be placeholders
513 which overwrite their old definition as a lexical closure with
514 a new compiled definition the first time that they're called.)
515 As a workaround for the problem, #'(SETF FOO) expressions can
516 be replaced with (EFFICIENT-SETF-FUNCTION FOO), where
517 (defmacro efficient-setf-function (place-function-name)
518 (or #+sbcl (and (sb-impl::info :function :accessor-for place-function-name)
519 ;; a workaround for the problem, encouraging the
520 ;; inline expansion of the structure accessor, so
521 ;; that the compiler can optimize its type test
522 (let ((new-value (gensym "NEW-VALUE-"))
523 (structure-value (gensym "STRUCTURE-VALUE-")))
524 `(lambda (,new-value ,structure-value)
525 (setf (,place-function-name ,structure-value)
527 ;; no problem, can just use the ordinary expansion
528 `(function (setf ,place-function-name))))
531 There's apparently a bug in CEILING optimization which caused
532 Douglas Crosher to patch the CMU CL version. Martin Atzmueller
533 applied the patches to SBCL and they didn't seem to cause problems
534 (as reported sbcl-devel 2001-05-04). However, since the patches
535 modify nontrivial code which was apparently written incorrectly
536 the first time around, until regression tests are written I'm not
537 comfortable merging the patches in the CVS version of SBCL.
540 (DESCRIBE 'SB-ALIEN:DEF-ALIEN-TYPE) reports the macro argument list
544 in #<PACKAGE "SB-ALIEN">.
545 Macro-function: #<FUNCTION "DEF!MACRO DEF-ALIEN-TYPE" {19F4A39}>
546 Macro arguments: (#:whole-470 #:environment-471)
547 On Sat, May 26, 2001 09:45:57 AM CDT it was compiled from:
548 /usr/stuff/sbcl/src/code/host-alieneval.lisp
549 Created: Monday, March 12, 2001 07:47:43 AM CST
552 (TIME (ROOM T)) reports more than 200 Mbytes consed even for
553 a clean, just-started SBCL system. And it seems to be right:
554 (ROOM T) can bring a small computer to its knees for a *long*
555 time trying to GC afterwards. Surely there's some more economical
556 way to implement (ROOM T).
559 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
561 ;;; The compiler is flushing the argument type test, and the default
562 ;;; case in the cond, so that calling with say a fixnum 0 causes a
564 (declaim (optimize (safety 2) (speed 3)))
566 (declare (type (or string stream) x))
567 (cond ((typep x 'string) 'string)
568 ((typep x 'stream) 'stream)
571 The symptom in sbcl-0.6.12.42 on OpenBSD is actually (TST 0)=>STREAM
572 (not the SIGBUS reported in the comment) but that's broken too;
573 type declarations are supposed to be treated as assertions unless
574 SAFETY 0, so we should be getting a TYPE-ERROR.
577 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
579 (in-package :cl-user)
580 ;;; The following invokes a compiler error.
581 (declaim (optimize (speed 2) (debug 3)))
584 (unwind-protect nil)))
588 The error message in sbcl-0.6.12.42 is
589 internal error, failed AVER:
590 "(COMMON-LISP:EQ (SB!C::TN-ENVIRONMENT SB!C:TN) SB!C::TN-ENV)"
593 When the compiler inline expands functions, it may be that different
594 kinds of return values are generated from different code branches.
595 E.g. an inline expansion of POSITION generates integer results
596 from one branch, and NIL results from another. When that inline
597 expansion is used in a context where only one of those results
600 (aref *a1* (position x *a2*)))
601 and the compiler can't prove that the unacceptable branch is
602 never taken, then bogus type mismatch warnings can be generated.
603 If you need to suppress the type mismatch warnings, you can
604 suppress the inline expansion,
606 #+sbcl (declare (notinline position)) ; to suppress bug 117 bogowarnings
607 (aref *a1* (position x *a2*)))
608 or, sometimes, suppress them by declaring the result to be of an
611 (aref *a1* (the integer (position x *a2*))))
613 This is not a new compiler problem in 0.7.0, but the new compiler
614 transforms for FIND, POSITION, FIND-IF, and POSITION-IF make it
615 more conspicuous. If you don't need performance from these functions,
616 and the bogus warnings are a nuisance for you, you can return to
617 your pre-0.7.0 state of grace with
618 #+sbcl (declaim (notinline find position find-if position-if)) ; bug 117..
621 as reported by Eric Marsden on cmucl-imp@cons.org 2001-08-14:
622 (= (FLOAT 1 DOUBLE-FLOAT-EPSILON)
623 (+ (FLOAT 1 DOUBLE-FLOAT-EPSILON) DOUBLE-FLOAT-EPSILON)) => T
624 when of course it should be NIL. (He says it only fails for X86,
625 not SPARC; dunno about Alpha.)
627 Also, "the same problem exists for LONG-FLOAT-EPSILON,
628 DOUBLE-FLOAT-NEGATIVE-EPSILON, LONG-FLOAT-NEGATIVE-EPSILON (though
629 for the -negative- the + is replaced by a - in the test)."
631 Raymond Toy comments that this is tricky on the X86 since its FPU
632 uses 80-bit precision internally.
635 Even in sbcl-0.pre7.x, which is supposed to be free of the old
636 non-ANSI behavior of treating the function return type inferred
637 from the current function definition as a declaration of the
638 return type from any function of that name, the return type of NIL
639 is attached to FOO in 120a above, and used to optimize code which
643 There was some sort of screwup in handling of
644 (IF (NOT (IGNORE-ERRORS ..))). E.g.
646 (if (not (ignore-errors
647 (make-pathname :host "foo"
651 (error "notunlessnot")))
652 The (NOT (IGNORE-ERRORS ..)) form evaluates to T, so this should be
653 printing "ok", but instead it's going to the ERROR. This problem
654 seems to've been introduced by MNA's HANDLER-CASE patch (sbcl-devel
655 2001-07-17) and as a workaround (put in sbcl-0.pre7.14.flaky4.12)
656 I reverted back to the old weird HANDLER-CASE code. However, I
657 think the problem looks like a compiler bug in handling RETURN-FROM,
658 so I left the MNA-patched code in HANDLER-CASE (suppressed with
659 #+NIL) and I'd like to go back to see whether this really is
660 a compiler bug before I delete this BUGS entry.
663 As of version 0.pre7.14, SBCL's implementation of MACROLET makes
664 the entire lexical environment at the point of MACROLET available
665 in the bodies of the macroexpander functions. In particular, it
666 allows the function bodies (which run at compile time) to try to
667 access lexical variables (which are only defined at runtime).
668 It doesn't even issue a warning, which is bad.
670 The SBCL behavior arguably conforms to the ANSI spec (since the
671 spec says that the behavior is undefined, ergo anything conforms).
672 However, it would be better to issue a compile-time error.
673 Unfortunately I (WHN) don't see any simple way to detect this
674 condition in order to issue such an error, so for the meantime
675 SBCL just does this weird broken "conforming" thing.
677 The ANSI standard says, in the definition of the special operator
679 The macro-expansion functions defined by MACROLET are defined
680 in the lexical environment in which the MACROLET form appears.
681 Declarations and MACROLET and SYMBOL-MACROLET definitions affect
682 the local macro definitions in a MACROLET, but the consequences
683 are undefined if the local macro definitions reference any
684 local variable or function bindings that are visible in that
686 Then it seems to contradict itself by giving the example
688 (macrolet ((fudge (z)
689 ;The parameters x and flag are not accessible
690 ; at this point; a reference to flag would be to
691 ; the global variable of that name.
692 ` (if flag (* ,z ,z) ,z)))
693 ;The parameters x and flag are accessible here.
697 The comment "a reference to flag would be to the global variable
698 of the same name" sounds like good behavior for the system to have.
699 but actual specification quoted above says that the actual behavior
703 (as reported by Gabe Garza on cmucl-help 2001-09-21)
705 (defun test-pred (x y)
709 (func (lambda () x)))
710 (print (eq func func))
711 (print (test-pred func func))
712 (delete func (list func))))
713 Now calling (TEST-CASE) gives output
716 (#<FUNCTION {500A9EF9}>)
717 Evidently Python thinks of the lambda as a code transformation so
718 much that it forgets that it's also an object.
721 The DEFSTRUCT section of the ANSI spec, in the :CONC-NAME section,
722 specifies a precedence rule for name collisions between slot accessors of
723 structure classes related by inheritance. As of 0.7.0, SBCL still
727 insufficient syntax checking in MACROLET:
729 (macrolet ((defmacro bar (z) `(+ z z)))
731 shouldn't compile without error (because of the extra DEFMACRO symbol).
734 Ideally, uninterning a symbol would allow it, and its associated
735 FDEFINITION and PROCLAIM data, to be reclaimed by the GC. However,
736 at least as of sbcl-0.7.0, this isn't the case. Information about
737 FDEFINITIONs and PROCLAIMed properties is stored in globaldb.lisp
738 essentially in ordinary (non-weak) hash tables keyed by symbols.
739 Thus, once a system has an entry in this system, it tends to live
740 forever, even when it is uninterned and all other references to it
744 (reported by Arnaud Rouanet on cmucl-imp 2001-12-18)
745 (defmethod foo ((x integer))
747 (defmethod foo :around ((x integer))
750 Now (FOO 3) should return 3, but instead it returns 4.
753 (reported by Alexey Dejneka sbcl-devel 2002-01-03)
755 SUBTYPEP does not work well with redefined classes:
764 * (defclass b (a) ())
777 This bug was fixed in sbcl-0.7.4.1 by invalidating the PCL wrapper
778 class upon redefinition. Unfortunately, doing so causes bug #176 to
779 appear. Pending further investication, one or other of these bugs
780 might be present at any given time.
783 Pretty-printing nested backquotes doesn't work right, as
784 reported by Alexey Dejneka sbcl-devel 2002-01-13:
786 ``(FOO SB-IMPL::BACKQ-COMMA-AT S)
787 * (lisp-implementation-version)
791 (as reported by Lynn Quam on cmucl-imp ca. 2002-01-16)
792 %NATURALIZE-C-STRING conses a lot, like 16 bytes per byte
793 of the naturalized string. We could probably port the patches
794 from the cmucl-imp mailing list.
797 (reported by Jesse Bouwman 2001-10-24 through the unfortunately
798 prominent SourceForge web/db bug tracking system, which is
799 unfortunately not a reliable way to get a timely response from
800 the SBCL maintainers)
801 In the course of trying to build a test case for an
802 application error, I encountered this behavior:
803 If you start up sbcl, and then lay on CTRL-C for a
804 minute or two, the lisp process will eventually say:
805 %PRIMITIVE HALT called; the party is over.
806 and throw you into the monitor. If I start up lisp,
807 attach to the process with strace, and then do the same
808 (abusive) thing, I get instead:
809 access failure in heap page not marked as write-protected
810 and the monitor again. I don't know enough to have the
811 faintest idea of what is going on here.
812 This is with sbcl 6.12, uname -a reports:
813 Linux prep 2.2.19 #4 SMP Tue Apr 24 13:59:52 CDT 2001 i686 unknown
814 I (WHN) have verified that the same thing occurs on sbcl-0.pre7.141
815 under OpenBSD 2.9 on my X86 laptop. Do be patient when you try it:
816 it took more than two minutes (but less than five) for me.
819 (This was once known as IR1-4, but it lived on even after the
820 IR1 interpreter went to the big bit bucket in the sky.)
821 The system accepts DECLAIM in most places where DECLARE would be
822 accepted, without even issuing a warning. ANSI allows this, but since
823 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
824 meaning is rather different, and it's unlikely that the user
825 has a good reason for doing DECLAIM not at top level, it would be
826 good to issue a STYLE-WARNING when this happens. A possible
827 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
828 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
829 [This is considered an IR1-interpreter-related bug because until
830 EVAL-WHEN is rewritten, which won't happen until after the IR1
831 interpreter is gone, the system's notion of what's a top-level form
832 and what's not will remain too confused to fix this problem.]
835 ANSI allows types `(COMPLEX ,FOO) to use very hairy values for
836 FOO, e.g. (COMPLEX (AND REAL (SATISFIES ODDP))). The old CMU CL
837 COMPLEX implementation didn't deal with this, and hasn't been
838 upgraded to do so. (This doesn't seem to be a high priority
839 conformance problem, since seems hard to construct useful code
843 Floating point errors are reported poorly. E.g. on x86 OpenBSD
846 debugger invoked on condition of type SB-KERNEL:FLOATING-POINT-EXCEPTION:
847 An arithmetic error SB-KERNEL:FLOATING-POINT-EXCEPTION was signalled.
848 No traps are enabled? How can this be?
849 It should be possible to be much more specific (overflow, division
850 by zero, etc.) and of course the "How can this be?" should be fixable.
852 See also bugs #45.c and #183
855 In sbcl-0.7.1.3 on x86, COMPILE-FILE on the file
856 (in-package :cl-user)
859 (defstruct foo bar bletch)
861 (labels ((kidify1 (kid)
869 (declare (inline kid-frob))
872 (the simple-vector (foo-bar perd)))))
874 debugger invoked on condition of type TYPE-ERROR:
875 The value NIL is not of type SB-C::NODE.
876 The location of this failure has moved around as various related
877 issues were cleaned up. As of sbcl-0.7.1.9, it occurs in
878 NODE-BLOCK called by LAMBDA-COMPONENT called by IR2-CONVERT-CLOSURE.
881 (essentially the same problem as a CMU CL bug reported by Martin
882 Cracauer on cmucl-imp 2002-02-19)
883 There is a hole in structure slot type checking. Compiling and LOADing
884 (declaim (optimize safety))
886 (bla 0 :type fixnum))
888 (let ((foo (make-foo)))
889 (setf (foo-bla foo) '(1 . 1))
890 (format t "Is ~a of type ~a a cons? => ~a~%"
892 (type-of (foo-bla foo))
893 (consp (foo-bla foo)))))
895 should signal an error, but in sbcl-0.7.1.21 instead gives the output
896 Is (1 . 1) of type CONS a cons? => NIL
897 without signalling an error.
900 Functions SUBTYPEP, TYPEP, UPGRADED-ARRAY-ELEMENT-TYPE, and
901 UPGRADED-COMPLEX-PART-TYPE should have an optional environment argument.
902 (reported by Alexey Dejneka sbcl-devel 2002-04-12)
905 (reported by Robert E. Brown 2002-04-16)
906 When a function is called with too few arguments, causing the
907 debugger to be entered, the uninitialized slots in the bad call frame
908 seem to cause GCish problems, being interpreted as tagged data even
909 though they're not. In particular, executing ROOM in the
910 debugger at that point causes AVER failures:
913 * (lisp-implementation-version)
919 failed AVER: "(SAP= CURRENT END)"
920 (Christophe Rhodes reports that this doesn't occur on the SPARC, which
921 isn't too surprising since there are many differences in stack
922 implementation and GC conservatism between the X86 and other ports.)
926 (in-package :cl-user)
928 (defmethod permanentize ((uustk uustk))
929 (flet ((frob (hash-table test-for-deletion)
931 (obj-entry.stale? (oe)
932 (destructuring-bind (key . datum) oe
933 (declare (type simple-vector key))
934 (deny0 (void? datum))
935 (some #'stale? key))))
936 (declare (inline frob obj-entry.stale?))
937 (frob (uustk.args-hash->obj-alist uustk)
939 (frob (uustk.hash->memoized-objs-list uustk)
942 in sbcl-0.7.3.11 causes an assertion failure,
945 (AND (NULL (BLOCK-SUCC B))
946 (NOT (BLOCK-DELETE-P B))
947 (NOT (EQ B (COMPONENT-HEAD #)))))"
950 In sbcl-0.7.3.11, compiling the (illegal) code
951 (in-package :cl-user)
952 (defmethod prove ((uustk uustk))
955 gives the (not terribly clear) error message
957 ; (during macroexpansion of (DEFMETHOD PROVE ...))
958 ; can't get template for (FROB NIL NIL)
959 The problem seems to be that the code walker used by the DEFMETHOD
960 macro is unhappy with the illegal syntax in the method body, and
961 is giving an unclear error message.
964 sbcl's treatment of at least macro lambda lists is too permissive;
965 e.g., in sbcl-0.7.3.7:
966 (defmacro foo (&rest rest bar) `(,bar ,rest))
967 (macroexpand '(foo quux zot)) -> (QUUX (QUUX ZOT))
968 whereas section 3.4.4 of the CLHS doesn't allow required parameters
969 to come after the rest argument.
972 The compiler sometimes tries to constant-fold expressions before
973 it checks to see whether they can be reached. This can lead to
974 bogus warnings about errors in the constant folding, e.g. in code
977 (WRITE-STRING (> X 0) "+" "0"))
978 compiled in a context where the compiler can prove that X is NIL,
979 and the compiler complains that (> X 0) causes a type error because
980 NIL isn't a valid argument to #'>. Until sbcl-0.7.4.10 or so this
981 caused a full WARNING, which made the bug really annoying because then
982 COMPILE and COMPILE-FILE returned FAILURE-P=T for perfectly legal
983 code. Since then the warning has been downgraded to STYLE-WARNING,
984 so it's still a bug but at least it's a little less annoying.
987 reported by Alexey Dejneka 08 Jun 2002 in sbcl-devel:
988 Playing with McCLIM, I've received an error "Unbound variable WRAPPER
989 in SB-PCL::CHECK-WRAPPER-VALIDITY".
990 (defun check-wrapper-validity (instance)
991 (let* ((owrapper (wrapper-of instance)))
992 (if (not (invalid-wrapper-p owrapper))
994 (let* ((state (wrapper-state wrapper)) ; !!!
996 I've tried to replace it with OWRAPPER, but now OBSOLETE-INSTANCE-TRAP
997 breaks with "NIL is not of type SB-KERNEL:LAYOUT".
999 partial fix: The undefined variable WRAPPER resulted from an error
1000 in recent refactoring, as can be seen by comparing to the code in e.g.
1001 sbcl-0.7.2. Replacing WRAPPER with OWRAPPER (done by WHN in sbcl-0.7.4.22)
1002 should bring the code back to its behavior as of sbcl-0.7.2, but
1003 that still leaves the OBSOLETE-INSTANCE-TRAP bug. An example of
1004 input which triggers that bug is
1006 (let ((lastname (intern (format nil "C~D" (1- i))))
1007 (name (intern (format nil "C~D" i))))
1008 (eval `(defclass ,name
1009 (,@(if (= i 0) nil (list lastname)))
1011 (eval `(defmethod initialize-instance :after ((x ,name) &rest any)
1012 (declare (ignore any))))))
1014 (defclass c0 (b) ())
1015 (make-instance 'c19)
1019 178: "AVER failure compiling confused THEs in FUNCALL"
1020 In sbcl-0.7.4.24, compiling
1022 (funcall (the function (the standard-object x))))
1025 "(AND (EQ (IR2-CONTINUATION-PRIMITIVE-TYPE 2CONT) FUNCTION-PTYPE) (EQ CHECK T))"
1026 This variant compiles OK, though:
1027 (defun bug178alternative (x)
1028 (funcall (the nil x)))
1030 183: "IEEE floating point issues"
1031 Even where floating point handling is being dealt with relatively
1032 well (as of sbcl-0.7.5, on sparc/sunos and alpha; see bug #146), the
1033 accrued-exceptions and current-exceptions part of the fp control
1034 word don't seem to bear much relation to reality. E.g. on
1038 debugger invoked on condition of type DIVISION-BY-ZERO:
1039 arithmetic error DIVISION-BY-ZERO signalled
1040 0] (sb-vm::get-floating-point-modes)
1042 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
1043 :ROUNDING-MODE :NEAREST
1044 :CURRENT-EXCEPTIONS NIL
1045 :ACCRUED-EXCEPTIONS (:INEXACT)
1048 * (sb-vm::get-floating-point-modes)
1049 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
1050 :ROUNDING-MODE :NEAREST
1051 :CURRENT-EXCEPTIONS (:INEXACT)
1052 :ACCRUED-EXCEPTIONS (:INEXACT)
1055 185: "top-level forms at the REPL"
1056 * (locally (defstruct foo (a 0 :type fixnum)))
1059 ; (in macroexpansion of (SB-KERNEL::%DELAYED-GET-COMPILER-LAYOUT BAR))
1060 however, compiling and loading the same expression in a file works
1063 187: "type inference confusion around DEFTRANSFORM time"
1064 (reported even more verbosely on sbcl-devel 2002-06-28 as "strange
1065 bug in DEFTRANSFORM")
1066 After the file below is compiled and loaded in sbcl-0.7.5, executing
1067 (TCX (MAKE-ARRAY 4 :FILL-POINTER 2) 0)
1068 at the REPL returns an adjustable vector, which is wrong. Presumably
1069 somehow the DERIVE-TYPE information for the output values of %WAD is
1070 being mispropagated as a type constraint on the input values of %WAD,
1071 and so causing the type test to be optimized away. It's unclear how
1072 hand-expanding the DEFTRANSFORM would change this, but it suggests
1073 the DEFTRANSFORM machinery (or at least the way DEFTRANSFORMs are
1074 invoked at a particular phase) is involved.
1075 (cl:in-package :sb-c)
1076 (eval-when (:compile-toplevel)
1077 ;;; standin for %DATA-VECTOR-AND-INDEX
1078 (defknown %dvai (array index)
1080 (foldable flushable))
1081 (deftransform %dvai ((array index)
1085 (let* ((atype (continuation-type array))
1086 (eltype (array-type-specialized-element-type atype)))
1087 (when (eq eltype *wild-type*)
1088 (give-up-ir1-transform
1089 "specialized array element type not known at compile-time"))
1090 (when (not (array-type-complexp atype))
1091 (give-up-ir1-transform "SIMPLE array!"))
1092 `(if (array-header-p array)
1093 (%wad array index nil)
1094 (values array index))))
1095 ;;; standin for %WITH-ARRAY-DATA
1096 (defknown %wad (array index (or index null))
1097 (values (simple-array * (*)) index index index)
1098 (foldable flushable))
1099 ;;; (Commenting out this optimizer causes the bug to go away.)
1100 (defoptimizer (%wad derive-type) ((array start end))
1101 (let ((atype (continuation-type array)))
1102 (when (array-type-p atype)
1103 (values-specifier-type
1104 `(values (simple-array ,(type-specifier
1105 (array-type-specialized-element-type atype))
1107 index index index)))))
1109 (defun %wad (array start end)
1110 (format t "~&in %WAD~%")
1111 (%with-array-data array start end))
1112 (cl:in-package :cl-user)
1114 (declare (type (vector t) v))
1115 (declare (notinline sb-kernel::%with-array-data))
1116 ;; (Hand-expending DEFTRANSFORM %DVAI here also causes the bug to
1120 188: "compiler performance fiasco involving type inference and UNION-TYPE"
1121 (In sbcl-0.7.6.10, DEFTRANSFORM CONCATENATE was commented out until this
1122 bug could be fixed properly, so you won't see the bug unless you restore
1123 the DEFTRANSFORM by hand.) In sbcl-0.7.5.11 on a 700 MHz Pentium III,
1127 (declare (optimize (safety 3)))
1128 (declare (optimize (compilation-speed 2)))
1129 (declare (optimize (speed 1) (debug 1) (space 1)))
1130 (let ((fn "if-this-file-exists-the-universe-is-strange"))
1131 (load fn :if-does-not-exist nil)
1132 (load (concatenate 'string fn ".lisp") :if-does-not-exist nil)
1133 (load (concatenate 'string fn ".fasl") :if-does-not-exist nil)
1134 (load (concatenate 'string fn ".misc-garbage")
1135 :if-does-not-exist nil)))))
1137 134.552 seconds of real time
1138 133.35156 seconds of user run time
1139 0.03125 seconds of system run time
1140 [Run times include 2.787 seconds GC run time.]
1142 246883368 bytes consed.
1143 BACKTRACE from Ctrl-C in the compilation shows that the compiler is
1144 thinking about type relationships involving types like
1146 (OR (INTEGER 576 576)
1157 190: "PPC/Linux pipe? buffer? bug"
1158 In sbcl-0.7.6, the run-program.test.sh test script sometimes hangs
1159 on the PPC/Linux platform, waiting for a zombie env process. This
1160 is a classic symptom of buffer filling and deadlock, but it seems
1161 only sporadically reproducible.
1163 191: "Miscellaneous PCL deficiencies"
1164 (reported by Alexey Dejneka sbcl-devel 2002-08-04)
1165 a. DEFCLASS does not inform the compiler about generated
1166 functions. Compiling a file with
1167 (DEFCLASS A-CLASS ()
1169 (DEFUN A-CLASS-X (A)
1170 (WITH-SLOTS (A-CLASS-X) A
1172 results in a STYLE-WARNING:
1174 SB-SLOT-ACCESSOR-NAME::|COMMON-LISP-USER A-CLASS-X slot READER|
1176 APD's fix for this was checked in to sbcl-0.7.6.20, but Pierre
1177 Mai points out that the declamation of functions is in fact
1178 incorrect in some cases (most notably for structure
1179 classes). This means that at present erroneous attempts to use
1180 WITH-SLOTS and the like on classes with metaclass STRUCTURE-CLASS
1181 won't get the corresponding STYLE-WARNING.
1182 c. the examples in CLHS 7.6.5.1 (regarding generic function lambda
1183 lists and &KEY arguments) do not signal errors when they should.
1185 192: "Python treats free type declarations as promises."
1186 b. What seemed like the same fundamental problem as bug 192a, but
1187 was not fixed by the same (APD "more strict type checking
1188 sbcl-devel 2002-08-97) patch:
1189 (DOTIMES (I ...) (DOTIMES (J ...) (DECLARE ...) ...)):
1190 (declaim (optimize (speed 1) (safety 3)))
1191 (defun trust-assertion (i)
1193 (declare (type (mod 4) i)) ; when commented out, behavior changes!
1196 (trust-assertion 6) ; prints nothing unless DECLARE is commented out
1198 193: "unhelpful CLOS error reporting when the primary method is missing"
1200 (defmethod foo :before ((x t)) (print x))
1201 is the only method defined on FOO, the error reporting when e.g.
1203 is relatively unhelpful:
1204 There is no primary method for the generic function
1205 #<STANDARD-GENERIC-FUNCTION FOO (1)>.
1206 with the offending argument nowhere visible in the backtrace. This
1207 continues even if there *are* primary methods, just not for the
1208 specified arg type, e.g.
1209 (defmethod foo ((x character)) (print x))
1210 (defmethod foo ((x string)) (print x))
1211 (defmethod foo ((x pathname)) ...)
1212 In that case it could be very helpful to know what argument value is
1213 falling through the cracks of the defined primary methods, but the
1214 error message stays the same (even BACKTRACE doesn't tell you what the
1215 bad argument value is).
1217 194: "no error from (THE REAL '(1 2 3)) in some cases"
1220 (multiple-value-prog1 (progn (the real '(1 2 3))))
1221 returns (1 2 3) instead of signalling an error. This was fixed by
1222 APD's "more strict type checking patch", but although the fixed
1223 code (in sbcl-0.7.7.19) works (signals TYPE-ERROR) interactively,
1224 it's difficult to write a regression test for it, because
1225 (IGNORE-ERRORS (MULTIPLE-VALUE-PROG1 (PROGN (THE REAL '(1 2 3)))))
1226 still returns (1 2 3).
1228 b. (IGNORE-ERRORS (MULTIPLE-VALUE-PROG1 (PROGN (THE REAL '(1 2 3)))))
1229 returns (1 2 3). (As above, this shows up when writing regression
1230 tests for fixed-ness of part a.)
1231 c. Also in sbcl-0.7.7.9, (IGNORE-ERRORS (THE REAL '(1 2 3))) => (1 2 3).
1233 (null (ignore-errors
1235 (arg2 (identity (the real #(1 2 3)))))
1236 (if (< arg1 arg2) arg1 arg2))))
1238 but putting the same expression inside (DEFUN FOO () ...),
1241 * Actually this entry is probably multiple bugs, as
1242 Alexey Dejneka commented on sbcl-devel 2002-09-03:)
1243 I don't think that placing these two bugs in one entry is
1244 a good idea: they have different explanations. The second
1245 (min 1 nil) is caused by flushing of unused code--IDENTITY
1246 can do nothing with it. So it is really bug 122. The first
1247 (min nil) is due to M-V-PROG1: substituting a continuation
1248 for the result, it forgets about type assertion. The purpose
1249 of IDENTITY is to save the restricted continuation from
1250 inaccurate transformations.
1251 * Alexey Dejneka pointed out that
1252 (IGNORE-ERRORS (IDENTITY (THE REAL '(1 2 3))))
1253 works as it should. Also
1254 (IGNORE-ERRORS (VALUES (THE REAL '(1 2 3))))
1255 works as it should. Perhaps this is another case of VALUES type
1256 intersections behaving in non-useful ways?
1258 199: "hairy FUNCTION types confuse the compiler"
1259 (reported by APD sbcl-devel 2002-09-15)
1261 (EQ NIL (FUNCALL F)))
1264 (DECLARE (TYPE (AND FUNCTION (SATISFIES MUR)) F))
1267 fails to compile, printing
1269 "(AND (EQ (IR2-CONTINUATION-PRIMITIVE-TYPE 2CONT) FUNCTION-PTYPE) (EQ CHECK T))"
1271 APD further reports that this bug is not present in CMUCL.
1273 201: "Incautious type inference from compound CONS types"
1274 (reported by APD sbcl-devel 2002-09-17)
1276 (LET ((Y (CAR (THE (CONS INTEGER *) X))))
1278 (FORMAT NIL "~S IS ~S, Y = ~S"
1285 (FOO ' (1 . 2)) => "NIL IS INTEGER, Y = 1"
1288 In 0.6.12.43 compilation of a function definition, contradicting its
1289 FTYPE proclamation, causes an error, e.g. COMPILE-FILE on
1291 (declaim (ftype (function () null) foo))
1296 debugger invoked on condition of type UNBOUND-VARIABLE:
1297 The variable SB-C::*ERROR-FUNCTION* is unbound.
1302 "~@<The previously declared FTYPE~2I ~_~S~I ~_~
1303 conflicts with the definition type ~2I~_~S~:>"
1307 (In 0.7.0 the variable was renamed to SB-C::*LOSSAGE-FUN*.)
1309 DEFUNCT CATEGORIES OF BUGS
1311 These labels were used for bugs related to the old IR1 interpreter.
1312 The # values reached 6 before the category was closed down.