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 someday perhaps I'll be motivated to look it up..
110 It would be nice if the
112 (during macroexpansion)
113 said what macroexpansion was at fault, e.g.
115 (during macroexpansion of IN-PACKAGE,
116 during macroexpansion of DEFFOO)
119 The type system doesn't understand the KEYWORD type very well:
120 (SUBTYPEP 'KEYWORD 'SYMBOL) => NIL, NIL
121 It might be possible to fix this by changing the definition of
122 KEYWORD to (AND SYMBOL (SATISFIES KEYWORDP)), but the type system
123 would need to be a bit smarter about AND types, too:
124 (SUBTYPEP '(AND SYMBOL KEYWORD) 'SYMBOL) => NIL, NIL
125 (The type system does know something about AND types already,
126 (SUBTYPEP '(AND INTEGER FLOAT) 'NUMBER) => T, T
127 (SUBTYPEP '(AND INTEGER FIXNUM) 'NUMBER) =>T, T
128 so likely this is a small patch.)
131 Floating point infinities are screwed up. [When I was converting CMU CL
132 to SBCL, I was looking for complexity to delete, and I thought it was safe
133 to just delete support for floating point infinities. It wasn't: they're
134 generated by the floating point hardware even when we remove support
135 for them in software. -- WHN] Support for them should be restored.
138 The ANSI syntax for non-STANDARD method combination types in CLOS is
139 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
140 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
141 If you mess this up, omitting the PROGN qualifier in in DEFMETHOD,
142 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
143 (DEFMETHOD FOO ((X BAR)) (PRINT 'NUMBER))
144 the error mesage is not easy to understand:
145 INVALID-METHOD-ERROR was called outside the dynamic scope
146 of a method combination function (inside the body of
147 DEFINE-METHOD-COMBINATION or a method on the generic
148 function COMPUTE-EFFECTIVE-METHOD).
149 It would be better if it were more informative, a la
150 The method combination type for this method (STANDARD) does
151 not match the method combination type for the generic function
153 Also, after you make the mistake of omitting the PROGN qualifier
154 on a DEFMETHOD, doing a new DEFMETHOD with the correct qualifier
156 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
158 INVALID-METHOD-ERROR was called outside the dynamic scope
159 of a method combination function (inside the body of
160 DEFINE-METHOD-COMBINATION or a method on the generic
161 function COMPUTE-EFFECTIVE-METHOD).
162 This is not very helpful..
165 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
166 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
167 (Also, when this is fixed, we can enable the code in PROCLAIM which
168 checks for incompatible FTYPE redeclarations.)
171 from DTC on the CMU CL mailing list 25 Feb 2000:
172 ;;; Compiler fails when this file is compiled.
174 ;;; Problem shows up in delete-block within ir1util.lisp. The assertion
175 ;;; (assert (member (functional-kind lambda) '(:let :mv-let :assignment)))
176 ;;; fails within bind node branch.
178 ;;; Note that if c::*check-consistency* is enabled then an un-reached
179 ;;; entry is also reported.
182 (declare (values nil))
199 (let ((ttt #'(lambda () (go cccc))))
200 (declare (special ttt))
201 (return-from bbbb nil))
204 (return-from bbbb nil))))))
207 (I *think* this is a bug. It certainly seems like strange behavior. But
208 the ANSI spec is scary, dark, and deep..)
209 (FORMAT NIL "~,1G" 1.4) => "1. "
210 (FORMAT NIL "~3,1G" 1.4) => "1. "
213 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
215 (setf (find-class 'ccc1) (find-class 'ccc))
216 (defmethod zut ((c ccc1)) 123)
217 DTC's recommended workaround from the mailing list 3 Mar 2000:
218 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
221 The ANSI spec, in section "22.3.5.2 Tilde Less-Than-Sign: Logical Block",
222 says that an error is signalled if ~W, ~_, ~<...~:>, ~I, or ~:T is used
223 inside "~<..~>" (without the colon modifier on the closing syntax).
224 However, SBCL doesn't do this:
225 * (FORMAT T "~<munge~wegnum~>" 12)
230 When too many files are opened, OPEN will fail with an
231 uninformative error message
232 error in function OPEN: error opening #P"/tmp/foo.lisp": NIL
233 instead of saying that too many files are open.
236 Right now, when COMPILE-FILE has a read error, it actually pops
237 you into the debugger before giving up on the file. It should
238 instead handle the error, perhaps issuing (and handling)
239 a secondary error "caught ERROR: unrecoverable error during compilation"
240 and then return with FAILURE-P true,
243 reported by Sam Steingold on the cmucl-imp mailing list 12 May 2000:
245 Also, there is another bug: `array-displacement' should return an array
246 or nil as first value (as per ANSI CL), while CMUCL declares it as
247 returning an array as first value always.
250 Sometimes (SB-EXT:QUIT) fails with
251 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
252 Process inferior-lisp exited abnormally with code 1
253 I haven't noticed a repeatable case of this yet.
256 some sort of bug in inlining and RETURN-FROM in sbcl-0.6.5: Compiling
259 (BLOCK USED-BY-SOME-Y?
262 (UNLESS (REJECTED? Y)
263 (RETURN-FROM USED-BY-SOME-Y? T)))))
264 (DECLARE (INLINE FROB))
269 error in function SB-KERNEL:ASSERT-ERROR:
270 The assertion (EQ (SB-C::CONTINUATION-KIND SB-C::CONT) :BLOCK-START) failed.
271 This is still present in sbcl-0.6.8.
274 In some cases the compiler believes type declarations on array
275 elements without checking them, e.g.
276 (DECLAIM (OPTIMIZE (SAFETY 3) (SPEED 1) (SPACE 1)))
279 (DECLARE (TYPE (SIMPLE-ARRAY CONS 1) X))
280 (WHEN (CONSP (AREF X 0))
282 (BAR (VECTOR (MAKE-FOO :A 11 :B 12)))
285 in SBCL 0.6.5 (and also in CMU CL 18b). This does not happen for
286 all cases, e.g. the type assumption *is* checked if the array
287 elements are declared to be of some structure type instead of CONS.
290 The printer doesn't report closures very well. This is true in
294 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
295 It would be nice to make closures have a settable name slot,
296 and make things like DEFSTRUCT and FLET, which create closures,
297 set helpful values into this slot.
300 And as long as we're wishing, it would be awfully nice if INSPECT could
301 also report on closures, telling about the values of the bound variables.
304 The compiler assumes that any time a function of declared FTYPE
305 doesn't signal an error, its arguments were of the declared type.
306 E.g. compiling and loading
307 (DECLAIM (OPTIMIZE (SAFETY 3)))
308 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
309 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE) FACTORIAL)))
311 (COND ((> (FACTORIAL X) 1.0E6)
312 (FORMAT T "too big~%"))
314 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
316 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
319 will cause the INTEGERP case to be selected, giving bogus output a la
321 This violates the "declarations are assertions" principle.
322 According to the ANSI spec, in the section "System Class FUNCTION",
323 this is a case of "lying to the compiler", but the lying is done
324 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
325 not by the unexpectedly general definition of FACTORIAL. In any case,
326 "declarations are assertions" means that lying to the compiler should
327 cause an error to be signalled, and should not cause a bogus
328 result to be returned. Thus, the compiler should not assume
329 that arbitrary functions check their argument types. (It might
330 make sense to add another flag (CHECKED?) to DEFKNOWN to
331 identify functions which *do* check their argument types.)
334 DEFMETHOD doesn't check the syntax of &REST argument lists properly,
335 accepting &REST even when it's not followed by an argument name:
336 (DEFMETHOD FOO ((X T) &REST) NIL)
339 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
340 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
342 (DECLARE (TYPE INDEXOID X))
343 (THE (VALUES INDEXOID)
345 where the implementation of the type check in function FOO
346 includes a full call to %TYPEP. There are also some fundamental problems
347 with the interpretation of VALUES types (inherited from CMU CL, and
348 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
349 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
352 The definitions of SIGCONTEXT-FLOAT-REGISTER and
353 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
354 supported on FreeBSD because the floating point state is not saved,
355 but at least as of FreeBSD 4.0, the floating point state *is* saved,
356 so they could be supported after all. Very likely
357 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
360 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
361 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
362 issue with 'union, 'and, 'or etc. So even though according to the
363 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
364 (and now SBCL) interpret them as legal types.
367 ANSI specifies DEFINE-SYMBOL-MACRO, but it's not defined in SBCL.
368 CMU CL added it ca. Aug 13, 2000, after some discussion on the mailing
369 list, and it is probably possible to use substantially the same
370 patches to add it to SBCL.
373 a slew of floating-point-related errors reported by Peter Van Eynde
375 a: (SQRT -9.0) fails, because SB-KERNEL::COMPLEX-SQRT is undefined.
376 Similarly, COMPLEX-ASIN, COMPLEX-ACOS, COMPLEX-ACOSH, and others
378 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
379 should probably be 1.4012985e-45. In SBCL,
380 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
381 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
382 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
383 and LEAST-NEGATIVE-LONG-FLOAT.
384 c: Many expressions generate floating infinity:
389 PVE's regression tests want them to raise errors. SBCL
390 generates the infinities instead, which may or may not be
391 conforming behavior, but then blow it by being unable to
392 output the infinities, since support for infinities is generally
393 broken, and in particular SB-IMPL::OUTPUT-FLOAT-INFINITY is
395 d: (in section12.erg) various forms a la
396 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
397 don't give the right behavior.
400 type safety errors reported by Peter Van Eynde July 25, 2000:
401 a: (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
403 In general lengths of array type specifications aren't
404 checked by COERCE, so it fails when the spec is
405 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
406 b: CONCATENATE has the same problem of not checking the length
407 of specified output array types. MAKE-SEQUENCE and MAP and
408 MERGE also have the same problem.
409 c: (COERCE 'AND 'FUNCTION) returns something related to
410 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
411 d: ELT signals SIMPLE-ERROR if its index argument
412 isn't a valid index for its sequence argument, but should
413 signal TYPE-ERROR instead.
414 e: FILE-LENGTH is supposed to signal a type error when its
415 argument is not a stream associated with a file, but doesn't.
416 f: (FLOAT-RADIX 2/3) should signal an error instead of
418 g: (LOAD "*.lsp") should signal FILE-ERROR.
419 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
420 should signal TYPE-ERROR.
421 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
422 be used for input and output as needed. It should fail with
423 TYPE-ERROR when handed e.g. the results of
424 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
425 the inappropriate positions, but doesn't.
426 j: (PARSE-NAMESTRING (COERCE (LIST #\f #\o #\o (CODE-CHAR 0) #\4 #\8)
428 should probably signal an error instead of making a pathname with
430 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
431 not a binary input stream, but instead cheerfully reads from
432 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
435 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
436 a: (DEFCLASS FOO () (A B A)) should signal a PROGRAM-ERROR, and
438 b: (DEFCLASS FOO () (A B A) (:DEFAULT-INITARGS X A X B)) should
439 signal a PROGRAM-ERROR, and doesn't.
440 c: (DEFCLASS FOO07 NIL ((A :ALLOCATION :CLASS :ALLOCATION :CLASS))),
441 and other DEFCLASS forms with duplicate specifications in their
442 slots, should signal a PROGRAM-ERROR, and doesn't.
443 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
444 causes a COMPILER-ERROR.
447 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
448 a: (SYMBOL-MACROLET ((T TRUE)) ..) should probably signal
449 PROGRAM-ERROR, but SBCL accepts it instead.
450 b: SYMBOL-MACROLET should refuse to bind something which is
451 declared as a global variable, signalling PROGRAM-ERROR.
452 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
453 it binds is declared SPECIAL inside.
456 LOOP bugs reported by Peter Van Eynde July 25, 2000:
457 a: (LOOP WITH (A B) DO (PRINT 1)) is a syntax error according to
458 the definition of WITH clauses given in the ANSI spec, but
459 compiles and runs happily in SBCL.
460 b: a messy one involving package iteration:
461 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<))))
462 Should be: (("blah" "blah2") ("blah2"))
463 SBCL: (("blah") ("blah2"))
464 * (LET ((X 1)) (LOOP FOR I BY (INCF X) FROM X TO 10 COLLECT I))
465 doesn't work -- SBCL's LOOP says BY isn't allowed in a FOR clause.
468 type system errors reported by Peter Van Eynde July 25, 2000:
469 a: (SUBTYPEP 'BIGNUM 'INTEGER) => NIL, NIL
470 but should be (VALUES T T) instead.
471 b: (SUBTYPEP 'EXTENDED-CHAR 'CHARACTER) => NIL, NIL
472 but should be (VALUES T T) instead.
473 c: (SUBTYPEP '(INTEGER (0) (0)) 'NIL) dies with nested errors.
474 d: In general, the system doesn't like '(INTEGER (0) (0)) -- it
475 blows up at the level of SPECIFIER-TYPE with
476 "Lower bound (0) is greater than upper bound (0)." Probably
477 SPECIFIER-TYPE should return NIL instead.
478 e: (TYPEP 0 '(COMPLEX (EQL 0)) fails with
479 "Component type for Complex is not numeric: (EQL 0)."
480 This might be easy to fix; the type system already knows
481 that (SUBTYPEP '(EQL 0) 'NUMBER) is true.
482 f: The type system doesn't know about the condition system,
483 so that e.g. (TYPEP 'SIMPLE-ERROR 'ERROR)=>NIL.
484 g: The type system isn't all that smart about relationships
485 between hairy types, as shown in the type.erg test results,
486 e.g. (SUBTYPEP 'CONS '(NOT ATOM)) => NIL, NIL.
489 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
491 (DEFGENERIC FOO02 (X))
492 (DEFMETHOD FOO02 ((X NUMBER)) T)
493 (LET ((M (FIND-METHOD (FUNCTION FOO02)
495 (LIST (FIND-CLASS (QUOTE NUMBER))))))
496 (REMOVE-METHOD (FUNCTION FOO02) M)
497 (DEFGENERIC FOO03 (X))
498 (ADD-METHOD (FUNCTION FOO03) M)))
499 should give an error, but SBCL allows it.
500 b: READ should probably return READER-ERROR, not the bare
501 arithmetic error, when input a la "1/0" or "1e1000" causes
505 It has been reported (e.g. by Peter Van Eynde) that there are
506 several metaobject protocol "errors". (In order to fix them, we might
507 need to document exactly what metaobject protocol specification
508 we're following -- the current code is just inherited from PCL.)
511 another error from Peter Van Eynde 5 September 2000:
512 (FORMAT NIL "~F" "FOO") should work, but instead reports an error.
513 PVE submitted a patch to deal with this bug, but it exposes other
514 comparably serious bugs, so I didn't apply it. It looks as though
515 the FORMAT code needs a fair amount of rewriting in order to comply
516 with the various details of the ANSI spec.
519 The implementation of #'+ returns its single argument without
520 type checking, e.g. (+ "illegal") => "illegal".
523 In sbcl-0.6.7, there is no doc string for CL:PUSH, probably
524 because it's defined with the DEFMACRO-MUNDANELY macro and something
525 is wrong with doc string setting in that macro.
528 Attempting to use COMPILE on something defined by DEFMACRO fails:
529 (DEFMACRO FOO (X) (CONS X X))
531 Error in function C::GET-LAMBDA-TO-COMPILE:
532 #<Closure Over Function "DEFUN (SETF MACRO-FUNCTION)" {480E21B1}> was defined in a non-null environment.
535 (SUBTYPEP '(AND ZILCH INTEGER) 'ZILCH)
539 CL:*DEFAULT-PATHNAME-DEFAULTS* doesn't behave as ANSI suggests (reflecting
540 current working directory). And there's no supported way to update
541 or query the current working directory (a la Unix "chdir" and "pwd"),
542 which is functionality that ILISP needs (and currently gets with low-level
546 The debugger LIST-LOCATIONS command doesn't work properly.
549 Compiling and loading
550 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
552 then requesting a BACKTRACE at the debugger prompt gives no information
553 about where in the user program the problem occurred.
556 The compiler is supposed to do type inference well enough that
559 ((SIMPLE-ARRAY SINGLE-FLOAT)
561 (DECLARE (TYPE (SIMPLE-ARRAY SINGLE-FLOAT) X))
564 is redundant. However, as reported by Juan Jose Garcia Ripoll for
565 CMU CL, it sometimes doesn't. Adding declarations is a pretty good
566 workaround for the problem for now, but can't be done by the TYPECASE
567 macros themselves, since it's too hard for the macro to detect
568 assignments to the variable within the clause.
569 Note: The compiler *is* smart enough to do the type inference in
570 many cases. This case, derived from a couple of MACROEXPAND-1
571 calls on Ripoll's original test case,
573 (DECLARE (OPTIMIZE SPEED (SAFETY 0)))
574 (COND ((TYPEP A '(SIMPLE-ARRAY SINGLE-FLOAT)) NIL
575 (LET ((LENGTH (ARRAY-TOTAL-SIZE A)))
576 (LET ((I 0) (G2554 LENGTH))
577 (DECLARE (TYPE REAL G2554) (TYPE REAL I))
580 (WHEN (>= I G2554) (GO SB-LOOP::END-LOOP))
581 (SETF (ROW-MAJOR-AREF A I) (- (ROW-MAJOR-AREF A I)))
582 (GO SB-LOOP::NEXT-LOOP)
583 SB-LOOP::END-LOOP))))))
584 demonstrates the problem; but the problem goes away if the TAGBODY
585 and GO forms are removed (leaving the SETF in ordinary, non-looping
586 code), or if the TAGBODY and GO forms are retained, but the
587 assigned value becomes 0.0 instead of (- (ROW-MAJOR-AREF A I)).
590 Paul Werkowski wrote on cmucl-imp@cons.org 2000-11-15
591 I am looking into this problem that showed up on the cmucl-help
592 list. It seems to me that the "implementation specific environment
593 hacking functions" found in pcl/walker.lisp are completely messed
594 up. The good thing is that they appear to be barely used within
595 PCL and the munged environment object is passed to cmucl only
596 in calls to macroexpand-1, which is probably why this case fails.
597 SBCL uses essentially the same code, so if the environment hacking
598 is screwed up, it affects us too.
601 Using the pretty-printer from the command prompt gives funny
602 results, apparently because the pretty-printer doesn't know
603 about user's command input, including the user's carriage return
604 that the user, and therefore the pretty-printer thinks that
605 the new output block should start indented 2 or more characters
606 rightward of the correct location.
609 (probably related to bug #70)
610 As reported by Carl Witty on submit@bugs.debian.org 1999-05-08,
612 (in-package "CL-USER")
613 (defun equal-terms (termx termy)
615 ((alpha-equal-bound-term-lists (listx listy)
616 (or (and (null listx) (null listy))
618 (let ((bindings-x (bindings-of-bound-term (car listx)))
619 (bindings-y (bindings-of-bound-term (car listy))))
620 (if (and (null bindings-x) (null bindings-y))
621 (alpha-equal-terms (term-of-bound-term (car listx))
622 (term-of-bound-term (car listy)))
623 (and (= (length bindings-x) (length bindings-y))
625 (enter-binding-pairs (bindings-of-bound-term (car listx))
626 (bindings-of-bound-term (car listy)))
627 (alpha-equal-terms (term-of-bound-term (car listx))
628 (term-of-bound-term (car listy)))
629 (exit-binding-pairs (bindings-of-bound-term (car listx))
630 (bindings-of-bound-term (car listy)))))))
631 (alpha-equal-bound-term-lists (cdr listx) (cdr listy)))))
633 (alpha-equal-terms (termx termy)
634 (if (and (variable-p termx)
636 (equal-bindings (id-of-variable-term termx)
637 (id-of-variable-term termy))
638 (and (equal-operators-p (operator-of-term termx) (operator-of-term termy))
639 (alpha-equal-bound-term-lists (bound-terms-of-term termx)
640 (bound-terms-of-term termy))))))
644 (with-variable-invocation (alpha-equal-terms termx termy))))))
645 causes an assertion failure
646 The assertion (EQ (C::LAMBDA-TAIL-SET C::CALLER)
647 (C::LAMBDA-TAIL-SET (C::LAMBDA-HOME C::CALLEE))) failed.
649 Bob Rogers reports (1999-07-28 on cmucl-imp@cons.org) a smaller test
650 case with the same problem:
651 (defun parse-fssp-alignment ()
652 ;; Given an FSSP alignment file named by the argument . . .
653 (labels ((get-fssp-char ()
657 ;; Stub body, enough to tickle the bug.
658 (list (read-fssp-char)
662 ANSI specifies that the RESULT-TYPE argument of CONCATENATE must be
663 a subtype of SEQUENCE, but CONCATENATE doesn't check this properly:
664 (CONCATENATE 'SIMPLE-ARRAY #(1 2) '(3)) => #(1 2 3)
665 This also leads to funny behavior when derived type specifiers
666 are used, as originally reported by Milan Zamazal for CMU CL (on the
667 Debian bugs mailing list (?) 2000-02-27), then reported by Martin
668 Atzmueller for SBCL (2000-10-01 on sbcl-devel@lists.sourceforge.net):
669 (DEFTYPE FOO () 'SIMPLE-ARRAY)
670 (CONCATENATE 'FOO #(1 2) '(3))
671 => #<ARRAY-TYPE SIMPLE-ARRAY> is a bad type specifier for
673 The derived type specifier FOO should act the same way as the
674 built-in type SIMPLE-ARRAY here, but it doesn't. That problem
675 doesn't seem to exist for sequence types:
676 (DEFTYPE BAR () 'SIMPLE-VECTOR)
677 (CONCATENATE 'BAR #(1 2) '(3)) => #(1 2 3)
680 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
681 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
682 crashes SBCL. In general tracing anything which is used in the
683 implementation of TRACE is likely to have the same problem.
686 As reported by Daniel Solaz on cmucl-help@cons.org 2000-11-23,
687 SXHASH returns the same value for all non-STRUCTURE-OBJECT instances,
688 notably including all PCL instances. There's a limit to how much
689 SXHASH can do to return unique values for instances, but at least
690 it should probably look at the class name, the way that it does
691 for STRUCTURE-OBJECTs.
694 As reported by Martin Atzmueller on the sbcl-devel list 2000-11-22,
695 > There remains one issue, that is a bug in SBCL:
696 > According to my interpretation of the spec, the ":" and "@" modifiers
697 > should appear _after_ the comma-seperated arguments.
698 > Well, SBCL (and CMUCL for that matter) accept
699 > (ASSERT (STRING= (FORMAT NIL "~:8D" 1) " 1"))
700 > where the correct way (IMHO) should be
701 > (ASSERT (STRING= (FORMAT NIL "~8:D" 1) " 1"))
702 Probably SBCL should stop accepting the "~:8D"-style format arguments,
703 or at least issue a warning.
706 (probably related to bug #65)
707 The compiler doesn't like &OPTIONAL arguments in LABELS and FLET
709 (DEFUN FIND-BEFORE (ITEM SEQUENCE &KEY (TEST #'EQL))
710 (LABELS ((FIND-ITEM (OBJ SEQ TEST &OPTIONAL (VAL NIL))
711 (LET ((ITEM (FIRST SEQ)))
714 ((FUNCALL TEST OBJ ITEM)
717 (FIND-ITEM OBJ (REST SEQ) TEST (NCONC VAL `(,ITEM))))))))
718 (FIND-ITEM ITEM SEQUENCE TEST)))
719 from David Young's bug report on cmucl-help@cons.org 30 Nov 2000
720 causes sbcl-0.6.9 to fail with
721 error in function SB-KERNEL:ASSERT-ERROR:
722 The assertion (EQ (SB-C::LAMBDA-TAIL-SET SB-C::CALLER)
723 (SB-C::LAMBDA-TAIL-SET
724 (SB-C::LAMBDA-HOME SB-C::CALLEE))) failed.
727 (DECLAIM (OPTIMIZE ..)) doesn't work. E.g. even after
728 (DECLAIM (OPTIMIZE (SPEED 3))), things are still optimized with
729 the previous SPEED policy. This bug will probably get fixed in
730 0.6.9.x in a general cleanup of optimization policy.
733 (DECLAIM (OPTIMIZE ..)) doesn't work properly inside LOCALLY forms.
736 As noted in the ANSI specification for COERCE, (COERCE 3 'COMPLEX)
737 gives a result which isn't COMPLEX. The result type optimizer
738 for COERCE doesn't know this, perhaps because it was written before
739 ANSI threw this curveball: the optimizer thinks that COERCE always
740 returns a result of the specified type. Thus while the interpreted
742 (DEFUN TRICKY (X) (TYPEP (COERCE X 'COMPLEX) 'COMPLEX))
743 returns the correct result,
745 the compiled function
751 As reported by Martin Atzmueller on sbcl-devel 26 Dec 2000,
752 ANSI says that WITH-OUTPUT-TO-STRING should have a keyword
753 :ELEMENT-TYPE, but in sbcl-0.6.9 this is not defined for
754 WITH-OUTPUT-TO-STRING.
757 ANSI says in one place that type declarations can be abbreviated even
758 when the type name is not a symbol, e.g.
759 (DECLAIM ((VECTOR T) *FOOVECTOR*))
760 SBCL doesn't support this. But ANSI says in another place that this
761 isn't allowed. So it's not clear this is a bug after all. (See the
762 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
766 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
767 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
768 an easily guessable temporary filename in a way which might open
769 applications using LOAD-FOREIGN to hijacking by malicious users
770 on the same machine. Incantations for doing this safely are
771 floating around the net in various "how to write secure programs
772 despite Unix" documents, and it would be good to (1) fix this in
773 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
774 files and make it share the same new safe logic.
777 (fixed early Feb 2001 by MNA)
780 As reported by wbuss@TELDA.NET (Wolfhard Buss) on cmucl-help
783 (loop with (a . b) of-type float = '(0.0 . 1.0)
784 and (c . d) of-type float = '(2.0 . 3.0)
785 return (list a b c d))
786 should evaluate to (0.0 1.0 2.0 3.0). cmucl-18c disagrees and
787 invokes the debugger: "B is not of type list".
788 SBCL does the same thing.
791 KNOWN BUGS RELATED TO THE IR1 INTERPRETER
793 (Note: At some point, the pure interpreter (actually a semi-pure
794 interpreter aka "the IR1 interpreter") will probably go away, replaced
796 (DEFUN EVAL (X) (FUNCALL (COMPILE NIL (LAMBDA ..)))))
797 and at that time these bugs should either go away automatically or
798 become more tractable to fix. Until then, they'll probably remain,
799 since some of them aren't considered urgent, and the rest are too hard
800 to fix as long as so many special cases remain. After the IR1
801 interpreter goes away is also the preferred time to start
802 systematically exterminating cases where debugging functionality
803 (backtrace, breakpoint, etc.) breaks down, since getting rid of the
804 IR1 interpreter will reduce the number of special cases we need to
808 The FUNCTION special operator doesn't check properly whether its
809 argument is a function name. E.g. (FUNCTION (X Y)) returns a value
810 instead of failing with an error. (Later attempting to funcall the
811 value does cause an error.)
814 COMPILED-FUNCTION-P bogusly reports T for interpreted functions:
815 * (DEFUN FOO (X) (- 12 X))
817 * (COMPILED-FUNCTION-P #'FOO)
822 (DEFVAR *SUPPRESS-P* T)
823 (EVAL '(UNLESS *SUPPRESS-P*
824 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
825 (FORMAT T "surprise!"))))
826 prints "surprise!". Probably the entire EVAL-WHEN mechanism ought to be
827 rewritten from scratch to conform to the ANSI definition, abandoning
828 the *ALREADY-EVALED-THIS* hack which is used in sbcl-0.6.8.9 (and
829 in the original CMU CL source, too). This should be easier to do --
830 though still nontrivial -- once the various IR1 interpreter special
834 EVAL-WHEN's idea of what's a toplevel form is even more screwed up
835 than the example in IR1-3 would suggest, since COMPILE-FILE and
836 COMPILE both print both "right now!" messages when compiling the
840 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
841 (PRINT "yes! right now!"))
844 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
845 (PRINT "no! right now!"))
847 and while EVAL doesn't print the "right now!" messages, the first
848 FUNCALL on the value returned by EVAL causes both of them to be printed.
851 The system accepts DECLAIM in most places where DECLARE would be
852 accepted, without even issuing a warning. ANSI allows this, but since
853 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
854 meaning is rather different, and it's unlikely that the user
855 has a good reason for doing DECLAIM not at top level, it would be
856 good to issue a STYLE-WARNING when this happens. A possible
857 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
858 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
859 [This is considered an IR1-interpreter-related bug because until
860 EVAL-WHEN is rewritten, which won't happen until after the IR1
861 interpreter is gone, the system's notion of what's a top-level form
862 and what's not will remain too confused to fix this problem.]