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. ANSI justifies this specification by saying
54 The restriction against issuing a warning for type mismatches
55 between a slot-initform and the corresponding slot's :TYPE
56 option is necessary because a slot-initform must be specified
57 in order to specify slot options; in some cases, no suitable
59 In SBCL, as in CMU CL (or, for that matter, any compiler which
60 really understands Common Lisp types) a suitable default does
61 exist, in all cases, because the compiler understands the concept
62 of functions which never return (i.e. has return type NIL, e.g.
63 ERROR). Thus, as a portable workaround, you can use a call to
64 some known-never-to-return function as the default. E.g.
66 (BAR (ERROR "missing :BAR argument")
67 :TYPE SOME-TYPE-TOO-HAIRY-TO-CONSTRUCT-AN-INSTANCE-OF))
69 (DECLAIM (FTYPE () NIL) MISSING-ARG)
70 (DEFUN REQUIRED-ARG () ; workaround for SBCL non-ANSI slot init typing
71 (ERROR "missing required argument"))
73 (BAR (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
74 (BLETCH (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
75 (N-REFS-SO-FAR 0 :TYPE (INTEGER 0)))
76 Such code will compile without complaint and work correctly either
77 on SBCL or on a completely compliant Common Lisp system.
80 bogus warnings about undefined functions for magic functions like
81 SB!C::%%DEFUN and SB!C::%DEFCONSTANT when cross-compiling files
82 like src/code/float.lisp. Fixing this will probably require
83 straightening out enough bootstrap consistency issues that
84 the cross-compiler can run with *TYPE-SYSTEM-INITIALIZED*.
85 Instead, the cross-compiler runs in a slightly flaky state
86 which is sane enough to compile SBCL itself, but which is
87 also unstable in several ways, including its inability
88 to really grok function declarations.
91 The "byte compiling top-level form:" output ought to be condensed.
92 Perhaps any number of such consecutive lines ought to turn into a
93 single "byte compiling top-level forms:" line.
96 The way that the compiler munges types with arguments together
97 with types with no arguments (in e.g. TYPE-EXPAND) leads to
98 weirdness visible to the user:
99 (DEFTYPE FOO () 'FIXNUM)
101 (TYPEP 11 '(FOO)) => T, which seems weird
102 (TYPEP 11 'FIXNUM) => T
103 (TYPEP 11 '(FIXNUM)) signals an error, as it should
104 The situation is complicated by the presence of Common Lisp types
105 like UNSIGNED-BYTE (which can either be used in list form or alone)
106 so I'm not 100% sure that the behavior above is actually illegal.
107 But I'm 90+% sure, and the following related behavior,
109 treating the bare symbol AND as equivalent to '(AND), is specifically
110 forbidden (by the ANSI specification of the AND type).
113 It would be nice if the
115 (during macroexpansion)
116 said what macroexpansion was at fault, e.g.
118 (during macroexpansion of IN-PACKAGE,
119 during macroexpansion of DEFFOO)
122 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
123 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
124 (Also, when this is fixed, we can enable the code in PROCLAIM which
125 checks for incompatible FTYPE redeclarations.)
128 from DTC on the CMU CL mailing list 25 Feb 2000:
129 ;;; Compiler fails when this file is compiled.
131 ;;; Problem shows up in delete-block within ir1util.lisp. The assertion
132 ;;; (assert (member (functional-kind lambda) '(:let :mv-let :assignment)))
133 ;;; fails within bind node branch.
135 ;;; Note that if c::*check-consistency* is enabled then an un-reached
136 ;;; entry is also reported.
139 (declare (values nil))
156 (let ((ttt #'(lambda () (go cccc))))
157 (declare (special ttt))
158 (return-from bbbb nil))
161 (return-from bbbb nil))))))
164 (I *think* this is a bug. It certainly seems like strange behavior. But
165 the ANSI spec is scary, dark, and deep..)
166 (FORMAT NIL "~,1G" 1.4) => "1. "
167 (FORMAT NIL "~3,1G" 1.4) => "1. "
170 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
172 (setf (find-class 'ccc1) (find-class 'ccc))
173 (defmethod zut ((c ccc1)) 123)
174 DTC's recommended workaround from the mailing list 3 Mar 2000:
175 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
178 The ANSI spec, in section "22.3.5.2 Tilde Less-Than-Sign: Logical Block",
179 says that an error is signalled if ~W, ~_, ~<...~:>, ~I, or ~:T is used
180 inside "~<..~>" (without the colon modifier on the closing syntax).
181 However, SBCL doesn't do this:
182 * (FORMAT T "~<munge~wegnum~>" 12)
187 When too many files are opened, OPEN will fail with an
188 uninformative error message
189 error in function OPEN: error opening #P"/tmp/foo.lisp": NIL
190 instead of saying that too many files are open.
193 Right now, when COMPILE-FILE has a read error, it actually pops
194 you into the debugger before giving up on the file. It should
195 instead handle the error, perhaps issuing (and handling)
196 a secondary error "caught ERROR: unrecoverable error during compilation"
197 and then return with FAILURE-P true,
200 reported by Sam Steingold on the cmucl-imp mailing list 12 May 2000:
201 Also, there is another bug: `array-displacement' should return an
202 array or nil as first value (as per ANSI CL), while CMUCL declares
203 it as returning an array as first value always.
204 (Actually, I think the old CMU CL version in SBCL never returns NIL,
205 i.e. it's not just a declaration problem, but the definition doesn't
209 Sometimes (SB-EXT:QUIT) fails with
210 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
211 Process inferior-lisp exited abnormally with code 1
212 I haven't noticed a repeatable case of this yet.
215 some sort of bug in inlining and RETURN-FROM in sbcl-0.6.5: Compiling
218 (BLOCK USED-BY-SOME-Y?
221 (UNLESS (REJECTED? Y)
222 (RETURN-FROM USED-BY-SOME-Y? T)))))
223 (DECLARE (INLINE FROB))
228 error in function SB-KERNEL:ASSERT-ERROR:
229 The assertion (EQ (SB-C::CONTINUATION-KIND SB-C::CONT) :BLOCK-START) failed.
230 This is still present in sbcl-0.6.8.
233 In some cases the compiler believes type declarations on array
234 elements without checking them, e.g.
235 (DECLAIM (OPTIMIZE (SAFETY 3) (SPEED 1) (SPACE 1)))
238 (DECLARE (TYPE (SIMPLE-ARRAY CONS 1) X))
239 (WHEN (CONSP (AREF X 0))
241 (BAR (VECTOR (MAKE-FOO :A 11 :B 12)))
244 in SBCL 0.6.5 (and also in CMU CL 18b). This does not happen for
245 all cases, e.g. the type assumption *is* checked if the array
246 elements are declared to be of some structure type instead of CONS.
249 The printer doesn't report closures very well. This is true in
253 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
254 It would be nice to make closures have a settable name slot,
255 and make things like DEFSTRUCT and FLET, which create closures,
256 set helpful values into this slot.
259 And as long as we're wishing, it would be awfully nice if INSPECT could
260 also report on closures, telling about the values of the bound variables.
263 The compiler assumes that any time a function of declared FTYPE
264 doesn't signal an error, its arguments were of the declared type.
265 E.g. compiling and loading
266 (DECLAIM (OPTIMIZE (SAFETY 3)))
267 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
268 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE) FACTORIAL)))
270 (COND ((> (FACTORIAL X) 1.0E6)
271 (FORMAT T "too big~%"))
273 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
275 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
278 will cause the INTEGERP case to be selected, giving bogus output a la
280 This violates the "declarations are assertions" principle.
281 According to the ANSI spec, in the section "System Class FUNCTION",
282 this is a case of "lying to the compiler", but the lying is done
283 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
284 not by the unexpectedly general definition of FACTORIAL. In any case,
285 "declarations are assertions" means that lying to the compiler should
286 cause an error to be signalled, and should not cause a bogus
287 result to be returned. Thus, the compiler should not assume
288 that arbitrary functions check their argument types. (It might
289 make sense to add another flag (CHECKED?) to DEFKNOWN to
290 identify functions which *do* check their argument types.)
293 DEFMETHOD doesn't check the syntax of &REST argument lists properly,
294 accepting &REST even when it's not followed by an argument name:
295 (DEFMETHOD FOO ((X T) &REST) NIL)
298 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
299 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
301 (DECLARE (TYPE INDEXOID X))
302 (THE (VALUES INDEXOID)
304 where the implementation of the type check in function FOO
305 includes a full call to %TYPEP. There are also some fundamental problems
306 with the interpretation of VALUES types (inherited from CMU CL, and
307 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
308 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
311 The definitions of SIGCONTEXT-FLOAT-REGISTER and
312 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
313 supported on FreeBSD because the floating point state is not saved,
314 but at least as of FreeBSD 4.0, the floating point state *is* saved,
315 so they could be supported after all. Very likely
316 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
319 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
320 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
321 issue with 'union, 'and, 'or etc. So even though according to the
322 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
323 (and now SBCL) interpret them as legal types.
326 ANSI specifies DEFINE-SYMBOL-MACRO, but it's not defined in SBCL.
327 CMU CL added it ca. Aug 13, 2000, after some discussion on the mailing
328 list, and it is probably possible to use substantially the same
329 patches to add it to SBCL.
332 a slew of floating-point-related errors reported by Peter Van Eynde
334 a: (fixed in sbcl-0.6.11.25)
335 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
336 should probably be 1.4012985e-45. In SBCL,
337 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
338 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
339 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
340 and LEAST-NEGATIVE-LONG-FLOAT.
341 c: Many expressions generate floating infinity:
346 PVE's regression tests want them to raise errors. SBCL
347 generates the infinities instead, which may or may not be
349 d: (in section12.erg) various forms a la
350 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
351 don't give the right behavior.
354 type safety errors reported by Peter Van Eynde July 25, 2000:
355 a: (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
357 In general lengths of array type specifications aren't
358 checked by COERCE, so it fails when the spec is
359 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
360 b: CONCATENATE has the same problem of not checking the length
361 of specified output array types. MAKE-SEQUENCE and MAP and
362 MERGE also have the same problem.
363 c: (COERCE 'AND 'FUNCTION) returns something related to
364 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
365 d: ELT signals SIMPLE-ERROR if its index argument
366 isn't a valid index for its sequence argument, but should
367 signal TYPE-ERROR instead.
368 e: FILE-LENGTH is supposed to signal a type error when its
369 argument is not a stream associated with a file, but doesn't.
370 f: (FLOAT-RADIX 2/3) should signal an error instead of
372 g: (LOAD "*.lsp") should signal FILE-ERROR.
373 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
374 should signal TYPE-ERROR.
375 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
376 be used for input and output as needed. It should fail with
377 TYPE-ERROR when handed e.g. the results of
378 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
379 the inappropriate positions, but doesn't.
380 j: (PARSE-NAMESTRING (COERCE (LIST #\f #\o #\o (CODE-CHAR 0) #\4 #\8)
382 should probably signal an error instead of making a pathname with
384 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
385 not a binary input stream, but instead cheerfully reads from
386 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
389 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
390 a: (DEFCLASS FOO () (A B A)) should signal a PROGRAM-ERROR, and
392 b: (DEFCLASS FOO () (A B A) (:DEFAULT-INITARGS X A X B)) should
393 signal a PROGRAM-ERROR, and doesn't.
394 c: (DEFCLASS FOO07 NIL ((A :ALLOCATION :CLASS :ALLOCATION :CLASS))),
395 and other DEFCLASS forms with duplicate specifications in their
396 slots, should signal a PROGRAM-ERROR, and doesn't.
397 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
398 causes a COMPILER-ERROR.
401 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
402 a: (SYMBOL-MACROLET ((T TRUE)) ..) should probably signal
403 PROGRAM-ERROR, but SBCL accepts it instead.
404 b: SYMBOL-MACROLET should refuse to bind something which is
405 declared as a global variable, signalling PROGRAM-ERROR.
406 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
407 it binds is declared SPECIAL inside.
410 LOOP bugs reported by Peter Van Eynde July 25, 2000:
411 a: (LOOP WITH (A B) DO (PRINT 1)) is a syntax error according to
412 the definition of WITH clauses given in the ANSI spec, but
413 compiles and runs happily in SBCL.
414 b: a messy one involving package iteration:
415 interpreted Form: (LET ((PACKAGE (MAKE-PACKAGE "LOOP-TEST"))) (INTERN "blah" PACKAGE) (LET ((BLAH2 (INTERN "blah2" PACKAGE))) (EXPORT BLAH2 PACKAGE)) (LIST (SORT (LOOP FOR SYM BEING EACH PRESENT-SYMBOL OF PACKAGE FOR SYM-NAME = (SYMBOL-NAME SYM) COLLECT SYM-NAME) (FUNCTION STRING<)) (SORT (LOOP FOR SYM BEING EACH EXTERNAL-SYMBOL OF PACKAGE FOR SYM-NAME = (SYMBOL-NAME SYM) COLLECT SYM-NAME) (FUNCTION STRING<))))
416 Should be: (("blah" "blah2") ("blah2"))
417 SBCL: (("blah") ("blah2"))
418 * (LET ((X 1)) (LOOP FOR I BY (INCF X) FROM X TO 10 COLLECT I))
419 doesn't work -- SBCL's LOOP says BY isn't allowed in a FOR clause.
422 type system errors reported by Peter Van Eynde July 25, 2000:
423 a: (SUBTYPEP 'BIGNUM 'INTEGER) => NIL, NIL
424 but should be (VALUES T T) instead.
425 b: (SUBTYPEP 'EXTENDED-CHAR 'CHARACTER) => NIL, NIL
426 but should be (VALUES T T) instead.
427 c: (SUBTYPEP '(INTEGER (0) (0)) 'NIL) dies with nested errors.
428 d: In general, the system doesn't like '(INTEGER (0) (0)) -- it
429 blows up at the level of SPECIFIER-TYPE with
430 "Lower bound (0) is greater than upper bound (0)." Probably
431 SPECIFIER-TYPE should return NIL instead.
432 e: (TYPEP 0 '(COMPLEX (EQL 0)) fails with
433 "Component type for Complex is not numeric: (EQL 0)."
434 This might be easy to fix; the type system already knows
435 that (SUBTYPEP '(EQL 0) 'NUMBER) is true.
436 f: The type system doesn't know about the condition system,
437 so that e.g. (TYPEP 'SIMPLE-ERROR 'ERROR)=>NIL.
438 g: The type system isn't all that smart about relationships
439 between hairy types, as shown in the type.erg test results,
440 e.g. (SUBTYPEP 'CONS '(NOT ATOM)) => NIL, NIL.
443 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
445 (DEFGENERIC FOO02 (X))
446 (DEFMETHOD FOO02 ((X NUMBER)) T)
447 (LET ((M (FIND-METHOD (FUNCTION FOO02)
449 (LIST (FIND-CLASS (QUOTE NUMBER))))))
450 (REMOVE-METHOD (FUNCTION FOO02) M)
451 (DEFGENERIC FOO03 (X))
452 (ADD-METHOD (FUNCTION FOO03) M)))
453 should give an error, but SBCL allows it.
454 b: READ should probably return READER-ERROR, not the bare
455 arithmetic error, when input a la "1/0" or "1e1000" causes
459 It has been reported (e.g. by Peter Van Eynde) that there are
460 several metaobject protocol "errors". (In order to fix them, we might
461 need to document exactly what metaobject protocol specification
462 we're following -- the current code is just inherited from PCL.)
465 another error from Peter Van Eynde 5 September 2000:
466 (FORMAT NIL "~F" "FOO") should work, but instead reports an error.
467 PVE submitted a patch to deal with this bug, but it exposes other
468 comparably serious bugs, so I didn't apply it. It looks as though
469 the FORMAT code needs a fair amount of rewriting in order to comply
470 with the various details of the ANSI spec.
473 The implementation of #'+ returns its single argument without
474 type checking, e.g. (+ "illegal") => "illegal".
477 In sbcl-0.6.7, there is no doc string for CL:PUSH, probably
478 because it's defined with the DEFMACRO-MUNDANELY macro and something
479 is wrong with doc string setting in that macro.
482 Attempting to use COMPILE on something defined by DEFMACRO fails:
483 (DEFMACRO FOO (X) (CONS X X))
485 Error in function C::GET-LAMBDA-TO-COMPILE:
486 #<Closure Over Function "DEFUN (SETF MACRO-FUNCTION)" {480E21B1}> was defined in a non-null environment.
489 (SUBTYPEP '(AND ZILCH INTEGER) 'ZILCH) => NIL, NIL
490 Note: I looked into fixing this in 0.6.11.15, but gave up. The
491 problem seems to be that there are two relevant type methods for
492 the subtypep operation, HAIRY :COMPLEX-SUBTYPEP-ARG2 and
493 INTERSECTION :COMPLEX-SUBTYPEP-ARG1, and only the first is
494 called. This could be fixed, but type dispatch is messy and
495 confusing enough already, I don't want to complicate it further.
496 Perhaps someday we can make CLOS cross-compiled (instead of compiled
497 after bootstrapping) so that we don't need to have the type system
498 available before CLOS, and then we can rewrite the type methods to
499 CLOS methods, and then expressing the solutions to stuff like this
500 should become much more straightforward. -- WHN 2001-03-14
503 CL:*DEFAULT-PATHNAME-DEFAULTS* doesn't behave as ANSI suggests (reflecting
504 current working directory). And there's no supported way to update
505 or query the current working directory (a la Unix "chdir" and "pwd"),
506 which is functionality that ILISP needs (and currently gets with low-level
508 When this is fixed, probably the more-or-less-parallel Unix-level
511 %SET-DEFAULT-DIRECTORY
513 should go away. Also we need to figure out what's the proper way to
514 deal with the interaction of users assigning new values to
515 *DEFAULT-PATHNAME-DEFAULTS* and cores being saved and restored.
516 (Perhaps just make restoring from a save always overwrite the old
517 value with the new Unix-level default directory?)
520 The debugger LIST-LOCATIONS command doesn't work properly.
523 Compiling and loading
524 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
526 then requesting a BACKTRACE at the debugger prompt gives no information
527 about where in the user program the problem occurred.
530 The compiler is supposed to do type inference well enough that
533 ((SIMPLE-ARRAY SINGLE-FLOAT)
535 (DECLARE (TYPE (SIMPLE-ARRAY SINGLE-FLOAT) X))
538 is redundant. However, as reported by Juan Jose Garcia Ripoll for
539 CMU CL, it sometimes doesn't. Adding declarations is a pretty good
540 workaround for the problem for now, but can't be done by the TYPECASE
541 macros themselves, since it's too hard for the macro to detect
542 assignments to the variable within the clause.
543 Note: The compiler *is* smart enough to do the type inference in
544 many cases. This case, derived from a couple of MACROEXPAND-1
545 calls on Ripoll's original test case,
547 (DECLARE (OPTIMIZE SPEED (SAFETY 0)))
548 (COND ((TYPEP A '(SIMPLE-ARRAY SINGLE-FLOAT)) NIL
549 (LET ((LENGTH (ARRAY-TOTAL-SIZE A)))
550 (LET ((I 0) (G2554 LENGTH))
551 (DECLARE (TYPE REAL G2554) (TYPE REAL I))
554 (WHEN (>= I G2554) (GO SB-LOOP::END-LOOP))
555 (SETF (ROW-MAJOR-AREF A I) (- (ROW-MAJOR-AREF A I)))
556 (GO SB-LOOP::NEXT-LOOP)
557 SB-LOOP::END-LOOP))))))
558 demonstrates the problem; but the problem goes away if the TAGBODY
559 and GO forms are removed (leaving the SETF in ordinary, non-looping
560 code), or if the TAGBODY and GO forms are retained, but the
561 assigned value becomes 0.0 instead of (- (ROW-MAJOR-AREF A I)).
564 Paul Werkowski wrote on cmucl-imp@cons.org 2000-11-15
565 I am looking into this problem that showed up on the cmucl-help
566 list. It seems to me that the "implementation specific environment
567 hacking functions" found in pcl/walker.lisp are completely messed
568 up. The good thing is that they appear to be barely used within
569 PCL and the munged environment object is passed to cmucl only
570 in calls to macroexpand-1, which is probably why this case fails.
571 SBCL uses essentially the same code, so if the environment hacking
572 is screwed up, it affects us too.
575 Using the pretty-printer from the command prompt gives funny
576 results, apparently because the pretty-printer doesn't know
577 about user's command input, including the user's carriage return
578 that the user, and therefore the pretty-printer thinks that
579 the new output block should start indented 2 or more characters
580 rightward of the correct location.
583 (probably related to bug #70)
584 As reported by Carl Witty on submit@bugs.debian.org 1999-05-08,
586 (in-package "CL-USER")
587 (defun equal-terms (termx termy)
589 ((alpha-equal-bound-term-lists (listx listy)
590 (or (and (null listx) (null listy))
592 (let ((bindings-x (bindings-of-bound-term (car listx)))
593 (bindings-y (bindings-of-bound-term (car listy))))
594 (if (and (null bindings-x) (null bindings-y))
595 (alpha-equal-terms (term-of-bound-term (car listx))
596 (term-of-bound-term (car listy)))
597 (and (= (length bindings-x) (length bindings-y))
599 (enter-binding-pairs (bindings-of-bound-term (car listx))
600 (bindings-of-bound-term (car listy)))
601 (alpha-equal-terms (term-of-bound-term (car listx))
602 (term-of-bound-term (car listy)))
603 (exit-binding-pairs (bindings-of-bound-term (car listx))
604 (bindings-of-bound-term (car listy)))))))
605 (alpha-equal-bound-term-lists (cdr listx) (cdr listy)))))
607 (alpha-equal-terms (termx termy)
608 (if (and (variable-p termx)
610 (equal-bindings (id-of-variable-term termx)
611 (id-of-variable-term termy))
612 (and (equal-operators-p (operator-of-term termx) (operator-of-term termy))
613 (alpha-equal-bound-term-lists (bound-terms-of-term termx)
614 (bound-terms-of-term termy))))))
618 (with-variable-invocation (alpha-equal-terms termx termy))))))
619 causes an assertion failure
620 The assertion (EQ (C::LAMBDA-TAIL-SET C::CALLER)
621 (C::LAMBDA-TAIL-SET (C::LAMBDA-HOME C::CALLEE))) failed.
623 Bob Rogers reports (1999-07-28 on cmucl-imp@cons.org) a smaller test
624 case with the same problem:
625 (defun parse-fssp-alignment ()
626 ;; Given an FSSP alignment file named by the argument . . .
627 (labels ((get-fssp-char ()
631 ;; Stub body, enough to tickle the bug.
632 (list (read-fssp-char)
636 ANSI specifies that the RESULT-TYPE argument of CONCATENATE must be
637 a subtype of SEQUENCE, but CONCATENATE doesn't check this properly:
638 (CONCATENATE 'SIMPLE-ARRAY #(1 2) '(3)) => #(1 2 3)
639 This also leads to funny behavior when derived type specifiers
640 are used, as originally reported by Milan Zamazal for CMU CL (on the
641 Debian bugs mailing list (?) 2000-02-27), then reported by Martin
642 Atzmueller for SBCL (2000-10-01 on sbcl-devel@lists.sourceforge.net):
643 (DEFTYPE FOO () 'SIMPLE-ARRAY)
644 (CONCATENATE 'FOO #(1 2) '(3))
645 => #<ARRAY-TYPE SIMPLE-ARRAY> is a bad type specifier for
647 The derived type specifier FOO should act the same way as the
648 built-in type SIMPLE-ARRAY here, but it doesn't. That problem
649 doesn't seem to exist for sequence types:
650 (DEFTYPE BAR () 'SIMPLE-VECTOR)
651 (CONCATENATE 'BAR #(1 2) '(3)) => #(1 2 3)
654 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
655 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
656 crashes SBCL. In general tracing anything which is used in the
657 implementation of TRACE is likely to have the same problem.
660 As reported by Daniel Solaz on cmucl-help@cons.org 2000-11-23,
661 SXHASH returns the same value for all non-STRUCTURE-OBJECT instances,
662 notably including all PCL instances. There's a limit to how much
663 SXHASH can do to return unique values for instances, but at least
664 it should probably look at the class name, the way that it does
665 for STRUCTURE-OBJECTs.
668 As reported by Martin Atzmueller on the sbcl-devel list 2000-11-22,
669 > There remains one issue, that is a bug in SBCL:
670 > According to my interpretation of the spec, the ":" and "@" modifiers
671 > should appear _after_ the comma-seperated arguments.
672 > Well, SBCL (and CMUCL for that matter) accept
673 > (ASSERT (STRING= (FORMAT NIL "~:8D" 1) " 1"))
674 > where the correct way (IMHO) should be
675 > (ASSERT (STRING= (FORMAT NIL "~8:D" 1) " 1"))
676 Probably SBCL should stop accepting the "~:8D"-style format arguments,
677 or at least issue a warning.
680 (probably related to bug #65)
681 The compiler doesn't like &OPTIONAL arguments in LABELS and FLET
683 (DEFUN FIND-BEFORE (ITEM SEQUENCE &KEY (TEST #'EQL))
684 (LABELS ((FIND-ITEM (OBJ SEQ TEST &OPTIONAL (VAL NIL))
685 (LET ((ITEM (FIRST SEQ)))
688 ((FUNCALL TEST OBJ ITEM)
691 (FIND-ITEM OBJ (REST SEQ) TEST (NCONC VAL `(,ITEM))))))))
692 (FIND-ITEM ITEM SEQUENCE TEST)))
693 from David Young's bug report on cmucl-help@cons.org 30 Nov 2000
694 causes sbcl-0.6.9 to fail with
695 error in function SB-KERNEL:ASSERT-ERROR:
696 The assertion (EQ (SB-C::LAMBDA-TAIL-SET SB-C::CALLER)
697 (SB-C::LAMBDA-TAIL-SET
698 (SB-C::LAMBDA-HOME SB-C::CALLEE))) failed.
701 (DECLAIM (OPTIMIZE ..)) doesn't work. E.g. even after
702 (DECLAIM (OPTIMIZE (SPEED 3))), things are still optimized with
703 the previous SPEED policy. This bug will probably get fixed in
704 0.6.9.x in a general cleanup of optimization policy.
707 (DECLAIM (OPTIMIZE ..)) doesn't work properly inside LOCALLY forms.
710 As noted in the ANSI specification for COERCE, (COERCE 3 'COMPLEX)
711 gives a result which isn't COMPLEX. The result type optimizer
712 for COERCE doesn't know this, perhaps because it was written before
713 ANSI threw this curveball: the optimizer thinks that COERCE always
714 returns a result of the specified type. Thus while the interpreted
716 (DEFUN TRICKY (X) (TYPEP (COERCE X 'COMPLEX) 'COMPLEX))
717 returns the correct result,
719 the compiled function
725 As reported by Martin Atzmueller on sbcl-devel 26 Dec 2000,
726 ANSI says that WITH-OUTPUT-TO-STRING should have a keyword
727 :ELEMENT-TYPE, but in sbcl-0.6.9 this is not defined for
728 WITH-OUTPUT-TO-STRING.
731 ANSI says in one place that type declarations can be abbreviated even
732 when the type name is not a symbol, e.g.
733 (DECLAIM ((VECTOR T) *FOOVECTOR*))
734 SBCL doesn't support this. But ANSI says in another place that this
735 isn't allowed. So it's not clear this is a bug after all. (See the
736 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
740 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
741 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
742 an easily guessable temporary filename in a way which might open
743 applications using LOAD-FOREIGN to hijacking by malicious users
744 on the same machine. Incantations for doing this safely are
745 floating around the net in various "how to write secure programs
746 despite Unix" documents, and it would be good to (1) fix this in
747 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
748 files and make it share the same new safe logic.
751 (fixed early Feb 2001 by MNA)
754 As reported by wbuss@TELDA.NET (Wolfhard Buss) on cmucl-help
757 (loop with (a . b) of-type float = '(0.0 . 1.0)
758 and (c . d) of-type float = '(2.0 . 3.0)
759 return (list a b c d))
760 should evaluate to (0.0 1.0 2.0 3.0). cmucl-18c disagrees and
761 invokes the debugger: "B is not of type list".
762 SBCL does the same thing.
765 Functions are assigned names based on the context in which they're
766 defined. This is less than ideal for the functions which are
767 used to implement CLOS methods. E.g. the output of
768 (DESCRIBE 'PRINT-OBJECT) lists functions like
769 #<FUNCTION "DEF!STRUCT (TRACE-INFO (:MAKE-LOAD-FORM-FUN SB-KERNEL:JUST-DUMP-IT-NORMALLY) (:PRINT-OBJECT #))" {1020E49}>
771 #<FUNCTION "MACROLET ((FORCE-DELAYED-DEF!METHODS NIL #))" {1242871}>
772 It would be better if these functions' names always identified
773 them as methods, and identified their generic functions and
777 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
778 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
779 the mean of the distribution can be systematically O(0.1%) wrong.
780 Just increasing R-I-E-B is probably not a good solution, since
781 it would decrease efficiency more than is probably necessary. Perhaps
782 using some sort of accept/reject method would be better.
785 (SUBTYPEP '(SATISFIES SOME-UNDEFINED-FUN) NIL)=>NIL,T (should be NIL,NIL)
788 Internally the compiler sometimes evaluates
789 (sb-kernel:type/= (specifier-type '*) (specifier-type t))
790 (I stumbled across this when I added an
791 (assert (not (eq type1 *wild-type*)))
792 in the NAMED :SIMPLE-= type method.) '* isn't really a type, and
793 in a type context should probably be translated to T, and so it's
794 probably to ask whether it's equal to the T type and then (using the
795 EQ type comparison in the NAMED :SIMPLE-= type method) return NIL.
796 (I haven't tried to investigate this bug enough to guess whether
797 there might be any user-level symptoms.)
800 a latent cross-compilation/bootstrapping bug: The cross-compilation
801 host's CL:CHAR-CODE-LIMIT is used in target code in readtable.lisp
802 and possibly elsewhere. Instead, we should use the target system's
803 CHAR-CODE-LIMIT. This will probably cause problems if we try to
804 bootstrap on a system which uses a different value of CHAR-CODE-LIMIT
808 (subtypep '(or (integer -1 1)
812 (integer -1 1))) => NIL,T
813 An analogous problem with SINGLE-FLOAT and REAL types was fixed in
814 sbcl-0.6.11.22, but some peculiarites of the RATIO type make it
815 awkward to generalize the fix to INTEGER and RATIONAL. It's not
816 clear what's the best fix. (See the "bug in type handling" discussion
817 on cmucl-imp ca. 2001-03-22 and ca. 2001-02-12.)
820 In sbcl-0.6.11.26, (COMPILE 'IN-HOST-COMPILATION-MODE) in
821 src/cold/shared.lisp doesn't correctly translate the
823 (defun in-host-compilation-mode (fn)
824 (let ((*features* (cons :sb-xc-host *features*))
825 ;; the CROSS-FLOAT-INFINITY-KLUDGE, as documented in
826 ;; base-target-features.lisp-expr:
827 (*shebang-features* (set-difference *shebang-features*
828 '(:sb-propagate-float-type
829 :sb-propagate-fun-type))))
830 (with-additional-nickname ("SB-XC" "SB!XC")
832 No error is reported by the compiler, but when the function is executed,
834 TYPE-ERROR in SB-KERNEL::OBJECT-NOT-TYPE-ERROR-HANDLER:
835 (:LINUX :X86 :IEEE-FLOATING-POINT :SB-CONSTRAIN-FLOAT-TYPE :SB-TEST
836 :SB-INTERPRETER :SB-DOC :UNIX ...) is not of type SYMBOL.
839 Inconsistencies between derived and declared VALUES return types for
840 DEFUN aren't checked very well. E.g. the logic which successfully
841 catches problems like
842 (declaim (ftype (function (fixnum) float) foo))
844 (declare (type integer x))
845 (values x)) ; wrong return type, detected, gives warning, good!
847 (declaim (ftype (function (t) (values t t)) bar))
849 (values x)) ; wrong number of return values, no warning, bad!
850 The cause of this is seems to be that (1) the internal function
851 VALUES-TYPES-EQUAL-OR-INTERSECT used to make the check handles its
852 arguments symmetrically, and (2) when the type checking code was
853 written back when when SBCL's code was still CMU CL, the intent
855 (declaim (ftype (function (t) t) bar))
857 (values x x)) ; wrong number of return values; should give warning?
858 not be warned for, because a two-valued return value is considered
859 to be compatible with callers who expects a single value to be
860 returned. That intent is probably not appropriate for modern ANSI
861 Common Lisp, but fixing this might be complicated because of other
862 divergences between auld-style and new-style handling of
863 multiple-VALUES types. (Some issues related to this were discussed
864 on cmucl-imp at some length sometime in 2000.)
867 The facility for dumping a running Lisp image to disk gets confused
868 when run without the PURIFY option, and creates an unnecessarily large
869 core file (apparently representing memory usage up to the previous
870 high-water mark). Moreover, when the file is loaded, it confuses the
871 GC, so that thereafter memory usage can never be reduced below that
875 The TRACE facility can't be used on some kinds of functions.
876 (Basically, the breakpoint facility was incompletely implemented
877 in the X86 port of CMU CL, and hasn't been fixed in SBCL.)
880 In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
881 CL), out-of-line structure slot setters are horribly inefficient
882 whenever the type of the slot is declared, because out-of-line
883 structure slot setters are implemented as closures to save space,
884 so the compiler doesn't compile the type test into code, but
885 instead just saves the type in a lexical closure and interprets it
887 A proper solution involves deciding whether it's really worth
888 saving space by implementing structure slot accessors as closures.
889 (If it's not worth it, the problem vanishes automatically. If it
890 is worth it, there are hacks we could use to force type tests to
891 be compiled anyway, and even shared. E.g. we could implement
892 an EQUAL hash table mapping from types to compiled type tests,
893 and save the appropriate compiled type test as part of each lexical
894 closure; or we could make the lexical closures be placeholders
895 which overwrite their old definition as a lexical closure with
896 a new compiled definition the first time that they're called.)
897 As a workaround for the problem, #'(SETF FOO) expressions can
898 be replaced with (EFFICIENT-SETF-FUNCTION FOO), where
899 (defmacro efficient-setf-function (place-function-name)
900 (or #+sbcl (and (sb-impl::info :function :accessor-for place-function-name)
901 ;; a workaround for the problem, encouraging the
902 ;; inline expansion of the structure accessor, so
903 ;; that the compiler can optimize its type test
904 (let ((new-value (gensym "NEW-VALUE-"))
905 (structure-value (gensym "STRUCTURE-VALUE-")))
906 `(lambda (,new-value ,structure-value)
907 (setf (,place-function-name ,structure-value)
909 ;; no problem, can just use the ordinary expansion
910 `(function (setf ,place-function-name))))
913 DESCRIBE interacts poorly with *PRINT-CIRCLE*, e.g. the output from
914 (let ((*print-circle* t)) (describe (make-hash-table)))
916 #<HASH-TABLE :TEST EQL :COUNT 0 {90BBFC5}> is an . (EQL)
918 Its REHASH-SIZE is 1.5. Its REHASH-THRESHOLD is . (1.0)
919 It holds 0 key/value pairs.
920 where the ". (EQL)" and ". (1.0)" substrings are screwups.
921 (This is likely a pretty-printer problem which happens to
922 be exercised by DESCRIBE, not actually a DESCRIBE problem.)
925 There's apparently a bug in CEILING optimization which caused
926 Douglas Crosher to patch the CMU CL version. Martin Atzmueller
927 applied the patches to SBCL and they didn't seem to cause problems
928 (as reported sbcl-devel 2001-05-04). However, since the patches
929 modify nontrivial code which was apparently written incorrectly
930 the first time around, until regression tests are written I'm not
931 comfortable merging the patches in the CVS version of SBCL.
934 The error message for calls to structure accessors with the
935 wrong number of arguments is confusing and of the wrong
936 condition class (TYPE-ERROR instead of PROGRAM-ERROR):
937 * (defstruct foo x y)
939 debugger invoked on condition of type SIMPLE-TYPE-ERROR:
940 Structure for accessor FOO-X is not a FOO:
944 As reported by Arthur Lemmens sbcl-devel 2001-05-05, ANSI
945 requires that SYMBOL-MACROLET refuse to rebind special variables,
946 but SBCL doesn't do this. (Also as reported by AL in the same
947 message, SBCL depended on this nonconforming behavior to build
948 itself, because of the way that **CURRENT-SEGMENT** was implemented.
949 As of sbcl-0.6.12.x, this dependence on the nonconforming behavior
950 has been fixed, but the nonconforming behavior remains.)
953 As reported by Arthur Lemmens sbcl-devel 2001-05-05, ANSI's
954 definition of (LOOP .. DO ..) requires that the terms following
955 DO all be compound forms. SBCL's implementation of LOOP allows
956 non-compound forms (like the bare symbol COUNT, in his example)
960 (DESCRIBE 'SB-ALIEN:DEF-ALIEN-TYPE) reports the macro argument list
964 in #<PACKAGE "SB-ALIEN">.
965 Macro-function: #<FUNCTION "DEF!MACRO DEF-ALIEN-TYPE" {19F4A39}>
966 Macro arguments: (#:whole-470 #:environment-471)
967 On Sat, May 26, 2001 09:45:57 AM CDT it was compiled from:
968 /usr/stuff/sbcl/src/code/host-alieneval.lisp
969 Created: Monday, March 12, 2001 07:47:43 AM CST
972 (DESCRIBE 'STREAM-READ-BYTE)
975 (reported by Eric Marsden on cmucl-imp 2001-06-15)
977 (TYPEP 0 '(COMPLEX (EQL 0)))
978 signals an error in sbcl-0.6.12.34,
979 The component type for COMPLEX is not numeric: (EQL 0)
980 This is funny since sbcl-0.6.12.34 knows
981 (SUBTYPEP '(EQL 0) 'NUMBER) => T
984 (reported as a CMU CL bug by Erik Naggum on comp.lang.lisp
986 * (write #*101 :radix t :base 36)
992 KNOWN BUGS RELATED TO THE IR1 INTERPRETER
994 (Note: At some point, the pure interpreter (actually a semi-pure
995 interpreter aka "the IR1 interpreter") will probably go away, replaced
997 (DEFUN EVAL (X) (FUNCALL (COMPILE NIL (LAMBDA ..)))))
998 and at that time these bugs should either go away automatically or
999 become more tractable to fix. Until then, they'll probably remain,
1000 since some of them aren't considered urgent, and the rest are too hard
1001 to fix as long as so many special cases remain. After the IR1
1002 interpreter goes away is also the preferred time to start
1003 systematically exterminating cases where debugging functionality
1004 (backtrace, breakpoint, etc.) breaks down, since getting rid of the
1005 IR1 interpreter will reduce the number of special cases we need to
1009 The FUNCTION special operator doesn't check properly whether its
1010 argument is a function name. E.g. (FUNCTION (X Y)) returns a value
1011 instead of failing with an error. (Later attempting to funcall the
1012 value does cause an error.)
1015 COMPILED-FUNCTION-P bogusly reports T for interpreted functions:
1016 * (DEFUN FOO (X) (- 12 X))
1018 * (COMPILED-FUNCTION-P #'FOO)
1023 (DEFVAR *SUPPRESS-P* T)
1024 (EVAL '(UNLESS *SUPPRESS-P*
1025 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
1026 (FORMAT T "surprise!"))))
1027 prints "surprise!". Probably the entire EVAL-WHEN mechanism ought to be
1028 rewritten from scratch to conform to the ANSI definition, abandoning
1029 the *ALREADY-EVALED-THIS* hack which is used in sbcl-0.6.8.9 (and
1030 in the original CMU CL source, too). This should be easier to do --
1031 though still nontrivial -- once the various IR1 interpreter special
1035 EVAL-WHEN's idea of what's a toplevel form is even more screwed up
1036 than the example in IR1-3 would suggest, since COMPILE-FILE and
1037 COMPILE both print both "right now!" messages when compiling the
1041 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
1042 (PRINT "yes! right now!"))
1045 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
1046 (PRINT "no! right now!"))
1048 and while EVAL doesn't print the "right now!" messages, the first
1049 FUNCALL on the value returned by EVAL causes both of them to be printed.
1052 The system accepts DECLAIM in most places where DECLARE would be
1053 accepted, without even issuing a warning. ANSI allows this, but since
1054 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
1055 meaning is rather different, and it's unlikely that the user
1056 has a good reason for doing DECLAIM not at top level, it would be
1057 good to issue a STYLE-WARNING when this happens. A possible
1058 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
1059 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
1060 [This is considered an IR1-interpreter-related bug because until
1061 EVAL-WHEN is rewritten, which won't happen until after the IR1
1062 interpreter is gone, the system's notion of what's a top-level form
1063 and what's not will remain too confused to fix this problem.]