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. Also we claim the :IEEE-FLOATING-POINT feature,
136 and I think that means we should support infinities.-- WHN] Support
137 for them should be restored.
140 The ANSI syntax for non-STANDARD method combination types in CLOS is
141 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
142 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
143 If you mess this up, omitting the PROGN qualifier in in DEFMETHOD,
144 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
145 (DEFMETHOD FOO ((X BAR)) (PRINT 'NUMBER))
146 the error mesage is not easy to understand:
147 INVALID-METHOD-ERROR was called outside the dynamic scope
148 of a method combination function (inside the body of
149 DEFINE-METHOD-COMBINATION or a method on the generic
150 function COMPUTE-EFFECTIVE-METHOD).
151 It would be better if it were more informative, a la
152 The method combination type for this method (STANDARD) does
153 not match the method combination type for the generic function
155 Also, after you make the mistake of omitting the PROGN qualifier
156 on a DEFMETHOD, doing a new DEFMETHOD with the correct qualifier
158 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
160 INVALID-METHOD-ERROR was called outside the dynamic scope
161 of a method combination function (inside the body of
162 DEFINE-METHOD-COMBINATION or a method on the generic
163 function COMPUTE-EFFECTIVE-METHOD).
164 This is not very helpful..
167 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
168 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
169 (Also, when this is fixed, we can enable the code in PROCLAIM which
170 checks for incompatible FTYPE redeclarations.)
173 from DTC on the CMU CL mailing list 25 Feb 2000:
174 ;;; Compiler fails when this file is compiled.
176 ;;; Problem shows up in delete-block within ir1util.lisp. The assertion
177 ;;; (assert (member (functional-kind lambda) '(:let :mv-let :assignment)))
178 ;;; fails within bind node branch.
180 ;;; Note that if c::*check-consistency* is enabled then an un-reached
181 ;;; entry is also reported.
184 (declare (values nil))
201 (let ((ttt #'(lambda () (go cccc))))
202 (declare (special ttt))
203 (return-from bbbb nil))
206 (return-from bbbb nil))))))
209 (I *think* this is a bug. It certainly seems like strange behavior. But
210 the ANSI spec is scary, dark, and deep..)
211 (FORMAT NIL "~,1G" 1.4) => "1. "
212 (FORMAT NIL "~3,1G" 1.4) => "1. "
215 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
217 (setf (find-class 'ccc1) (find-class 'ccc))
218 (defmethod zut ((c ccc1)) 123)
219 DTC's recommended workaround from the mailing list 3 Mar 2000:
220 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
223 The ANSI spec, in section "22.3.5.2 Tilde Less-Than-Sign: Logical Block",
224 says that an error is signalled if ~W, ~_, ~<...~:>, ~I, or ~:T is used
225 inside "~<..~>" (without the colon modifier on the closing syntax).
226 However, SBCL doesn't do this:
227 * (FORMAT T "~<munge~wegnum~>" 12)
232 When too many files are opened, OPEN will fail with an
233 uninformative error message
234 error in function OPEN: error opening #P"/tmp/foo.lisp": NIL
235 instead of saying that too many files are open.
238 Right now, when COMPILE-FILE has a read error, it actually pops
239 you into the debugger before giving up on the file. It should
240 instead handle the error, perhaps issuing (and handling)
241 a secondary error "caught ERROR: unrecoverable error during compilation"
242 and then return with FAILURE-P true,
245 reported by Sam Steingold on the cmucl-imp mailing list 12 May 2000:
247 Also, there is another bug: `array-displacement' should return an array
248 or nil as first value (as per ANSI CL), while CMUCL declares it as
249 returning an array as first value always.
252 Sometimes (SB-EXT:QUIT) fails with
253 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
254 Process inferior-lisp exited abnormally with code 1
255 I haven't noticed a repeatable case of this yet.
258 some sort of bug in inlining and RETURN-FROM in sbcl-0.6.5: Compiling
261 (BLOCK USED-BY-SOME-Y?
264 (UNLESS (REJECTED? Y)
265 (RETURN-FROM USED-BY-SOME-Y? T)))))
266 (DECLARE (INLINE FROB))
271 error in function SB-KERNEL:ASSERT-ERROR:
272 The assertion (EQ (SB-C::CONTINUATION-KIND SB-C::CONT) :BLOCK-START) failed.
273 This is still present in sbcl-0.6.8.
276 In some cases the compiler believes type declarations on array
277 elements without checking them, e.g.
278 (DECLAIM (OPTIMIZE (SAFETY 3) (SPEED 1) (SPACE 1)))
281 (DECLARE (TYPE (SIMPLE-ARRAY CONS 1) X))
282 (WHEN (CONSP (AREF X 0))
284 (BAR (VECTOR (MAKE-FOO :A 11 :B 12)))
287 in SBCL 0.6.5 (and also in CMU CL 18b). This does not happen for
288 all cases, e.g. the type assumption *is* checked if the array
289 elements are declared to be of some structure type instead of CONS.
292 The printer doesn't report closures very well. This is true in
296 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
297 It would be nice to make closures have a settable name slot,
298 and make things like DEFSTRUCT and FLET, which create closures,
299 set helpful values into this slot.
302 And as long as we're wishing, it would be awfully nice if INSPECT could
303 also report on closures, telling about the values of the bound variables.
306 The compiler assumes that any time a function of declared FTYPE
307 doesn't signal an error, its arguments were of the declared type.
308 E.g. compiling and loading
309 (DECLAIM (OPTIMIZE (SAFETY 3)))
310 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
311 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE) FACTORIAL)))
313 (COND ((> (FACTORIAL X) 1.0E6)
314 (FORMAT T "too big~%"))
316 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
318 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
321 will cause the INTEGERP case to be selected, giving bogus output a la
323 This violates the "declarations are assertions" principle.
324 According to the ANSI spec, in the section "System Class FUNCTION",
325 this is a case of "lying to the compiler", but the lying is done
326 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
327 not by the unexpectedly general definition of FACTORIAL. In any case,
328 "declarations are assertions" means that lying to the compiler should
329 cause an error to be signalled, and should not cause a bogus
330 result to be returned. Thus, the compiler should not assume
331 that arbitrary functions check their argument types. (It might
332 make sense to add another flag (CHECKED?) to DEFKNOWN to
333 identify functions which *do* check their argument types.)
336 DEFMETHOD doesn't check the syntax of &REST argument lists properly,
337 accepting &REST even when it's not followed by an argument name:
338 (DEFMETHOD FOO ((X T) &REST) NIL)
341 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
342 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
344 (DECLARE (TYPE INDEXOID X))
345 (THE (VALUES INDEXOID)
347 where the implementation of the type check in function FOO
348 includes a full call to %TYPEP. There are also some fundamental problems
349 with the interpretation of VALUES types (inherited from CMU CL, and
350 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
351 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
354 The definitions of SIGCONTEXT-FLOAT-REGISTER and
355 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
356 supported on FreeBSD because the floating point state is not saved,
357 but at least as of FreeBSD 4.0, the floating point state *is* saved,
358 so they could be supported after all. Very likely
359 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
362 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
363 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
364 issue with 'union, 'and, 'or etc. So even though according to the
365 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
366 (and now SBCL) interpret them as legal types.
369 ANSI specifies DEFINE-SYMBOL-MACRO, but it's not defined in SBCL.
370 CMU CL added it ca. Aug 13, 2000, after some discussion on the mailing
371 list, and it is probably possible to use substantially the same
372 patches to add it to SBCL.
375 a slew of floating-point-related errors reported by Peter Van Eynde
377 a: (SQRT -9.0) fails, because SB-KERNEL::COMPLEX-SQRT is undefined.
378 Similarly, COMPLEX-ASIN, COMPLEX-ACOS, COMPLEX-ACOSH, and others
380 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
381 should probably be 1.4012985e-45. In SBCL,
382 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
383 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
384 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
385 and LEAST-NEGATIVE-LONG-FLOAT.
386 c: Many expressions generate floating infinity:
391 PVE's regression tests want them to raise errors. SBCL
392 generates the infinities instead, which may or may not be
393 conforming behavior, but then blow it by being unable to
394 output the infinities, since support for infinities is generally
395 broken, and in particular SB-IMPL::OUTPUT-FLOAT-INFINITY is
397 d: (in section12.erg) various forms a la
398 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
399 don't give the right behavior.
402 type safety errors reported by Peter Van Eynde July 25, 2000:
403 a: (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
405 In general lengths of array type specifications aren't
406 checked by COERCE, so it fails when the spec is
407 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
408 b: CONCATENATE has the same problem of not checking the length
409 of specified output array types. MAKE-SEQUENCE and MAP and
410 MERGE also have the same problem.
411 c: (COERCE 'AND 'FUNCTION) returns something related to
412 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
413 d: ELT signals SIMPLE-ERROR if its index argument
414 isn't a valid index for its sequence argument, but should
415 signal TYPE-ERROR instead.
416 e: FILE-LENGTH is supposed to signal a type error when its
417 argument is not a stream associated with a file, but doesn't.
418 f: (FLOAT-RADIX 2/3) should signal an error instead of
420 g: (LOAD "*.lsp") should signal FILE-ERROR.
421 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
422 should signal TYPE-ERROR.
423 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
424 be used for input and output as needed. It should fail with
425 TYPE-ERROR when handed e.g. the results of
426 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
427 the inappropriate positions, but doesn't.
428 j: (PARSE-NAMESTRING (COERCE (LIST #\f #\o #\o (CODE-CHAR 0) #\4 #\8)
430 should probably signal an error instead of making a pathname with
432 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
433 not a binary input stream, but instead cheerfully reads from
434 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
437 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
438 a: (DEFCLASS FOO () (A B A)) should signal a PROGRAM-ERROR, and
440 b: (DEFCLASS FOO () (A B A) (:DEFAULT-INITARGS X A X B)) should
441 signal a PROGRAM-ERROR, and doesn't.
442 c: (DEFCLASS FOO07 NIL ((A :ALLOCATION :CLASS :ALLOCATION :CLASS))),
443 and other DEFCLASS forms with duplicate specifications in their
444 slots, should signal a PROGRAM-ERROR, and doesn't.
445 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
446 causes a COMPILER-ERROR.
449 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
450 a: (SYMBOL-MACROLET ((T TRUE)) ..) should probably signal
451 PROGRAM-ERROR, but SBCL accepts it instead.
452 b: SYMBOL-MACROLET should refuse to bind something which is
453 declared as a global variable, signalling PROGRAM-ERROR.
454 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
455 it binds is declared SPECIAL inside.
458 LOOP bugs reported by Peter Van Eynde July 25, 2000:
459 a: (LOOP WITH (A B) DO (PRINT 1)) is a syntax error according to
460 the definition of WITH clauses given in the ANSI spec, but
461 compiles and runs happily in SBCL.
462 b: a messy one involving package iteration:
463 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<))))
464 Should be: (("blah" "blah2") ("blah2"))
465 SBCL: (("blah") ("blah2"))
466 * (LET ((X 1)) (LOOP FOR I BY (INCF X) FROM X TO 10 COLLECT I))
467 doesn't work -- SBCL's LOOP says BY isn't allowed in a FOR clause.
470 type system errors reported by Peter Van Eynde July 25, 2000:
471 a: (SUBTYPEP 'BIGNUM 'INTEGER) => NIL, NIL
472 but should be (VALUES T T) instead.
473 b: (SUBTYPEP 'EXTENDED-CHAR 'CHARACTER) => NIL, NIL
474 but should be (VALUES T T) instead.
475 c: (SUBTYPEP '(INTEGER (0) (0)) 'NIL) dies with nested errors.
476 d: In general, the system doesn't like '(INTEGER (0) (0)) -- it
477 blows up at the level of SPECIFIER-TYPE with
478 "Lower bound (0) is greater than upper bound (0)." Probably
479 SPECIFIER-TYPE should return NIL instead.
480 e: (TYPEP 0 '(COMPLEX (EQL 0)) fails with
481 "Component type for Complex is not numeric: (EQL 0)."
482 This might be easy to fix; the type system already knows
483 that (SUBTYPEP '(EQL 0) 'NUMBER) is true.
484 f: The type system doesn't know about the condition system,
485 so that e.g. (TYPEP 'SIMPLE-ERROR 'ERROR)=>NIL.
486 g: The type system isn't all that smart about relationships
487 between hairy types, as shown in the type.erg test results,
488 e.g. (SUBTYPEP 'CONS '(NOT ATOM)) => NIL, NIL.
491 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
493 (DEFGENERIC FOO02 (X))
494 (DEFMETHOD FOO02 ((X NUMBER)) T)
495 (LET ((M (FIND-METHOD (FUNCTION FOO02)
497 (LIST (FIND-CLASS (QUOTE NUMBER))))))
498 (REMOVE-METHOD (FUNCTION FOO02) M)
499 (DEFGENERIC FOO03 (X))
500 (ADD-METHOD (FUNCTION FOO03) M)))
501 should give an error, but SBCL allows it.
502 b: READ should probably return READER-ERROR, not the bare
503 arithmetic error, when input a la "1/0" or "1e1000" causes
507 It has been reported (e.g. by Peter Van Eynde) that there are
508 several metaobject protocol "errors". (In order to fix them, we might
509 need to document exactly what metaobject protocol specification
510 we're following -- the current code is just inherited from PCL.)
513 another error from Peter Van Eynde 5 September 2000:
514 (FORMAT NIL "~F" "FOO") should work, but instead reports an error.
515 PVE submitted a patch to deal with this bug, but it exposes other
516 comparably serious bugs, so I didn't apply it. It looks as though
517 the FORMAT code needs a fair amount of rewriting in order to comply
518 with the various details of the ANSI spec.
521 The implementation of #'+ returns its single argument without
522 type checking, e.g. (+ "illegal") => "illegal".
525 In sbcl-0.6.7, there is no doc string for CL:PUSH, probably
526 because it's defined with the DEFMACRO-MUNDANELY macro and something
527 is wrong with doc string setting in that macro.
530 Attempting to use COMPILE on something defined by DEFMACRO fails:
531 (DEFMACRO FOO (X) (CONS X X))
533 Error in function C::GET-LAMBDA-TO-COMPILE:
534 #<Closure Over Function "DEFUN (SETF MACRO-FUNCTION)" {480E21B1}> was defined in a non-null environment.
537 (SUBTYPEP '(AND ZILCH INTEGER) 'ZILCH)
541 CL:*DEFAULT-PATHNAME-DEFAULTS* doesn't behave as ANSI suggests (reflecting
542 current working directory). And there's no supported way to update
543 or query the current working directory (a la Unix "chdir" and "pwd"),
544 which is functionality that ILISP needs (and currently gets with low-level
548 The debugger LIST-LOCATIONS command doesn't work properly.
551 Compiling and loading
552 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
554 then requesting a BACKTRACE at the debugger prompt gives no information
555 about where in the user program the problem occurred.
558 The compiler is supposed to do type inference well enough that
561 ((SIMPLE-ARRAY SINGLE-FLOAT)
563 (DECLARE (TYPE (SIMPLE-ARRAY SINGLE-FLOAT) X))
566 is redundant. However, as reported by Juan Jose Garcia Ripoll for
567 CMU CL, it sometimes doesn't. Adding declarations is a pretty good
568 workaround for the problem for now, but can't be done by the TYPECASE
569 macros themselves, since it's too hard for the macro to detect
570 assignments to the variable within the clause.
571 Note: The compiler *is* smart enough to do the type inference in
572 many cases. This case, derived from a couple of MACROEXPAND-1
573 calls on Ripoll's original test case,
575 (DECLARE (OPTIMIZE SPEED (SAFETY 0)))
576 (COND ((TYPEP A '(SIMPLE-ARRAY SINGLE-FLOAT)) NIL
577 (LET ((LENGTH (ARRAY-TOTAL-SIZE A)))
578 (LET ((I 0) (G2554 LENGTH))
579 (DECLARE (TYPE REAL G2554) (TYPE REAL I))
582 (WHEN (>= I G2554) (GO SB-LOOP::END-LOOP))
583 (SETF (ROW-MAJOR-AREF A I) (- (ROW-MAJOR-AREF A I)))
584 (GO SB-LOOP::NEXT-LOOP)
585 SB-LOOP::END-LOOP))))))
586 demonstrates the problem; but the problem goes away if the TAGBODY
587 and GO forms are removed (leaving the SETF in ordinary, non-looping
588 code), or if the TAGBODY and GO forms are retained, but the
589 assigned value becomes 0.0 instead of (- (ROW-MAJOR-AREF A I)).
592 Paul Werkowski wrote on cmucl-imp@cons.org 2000-11-15
593 I am looking into this problem that showed up on the cmucl-help
594 list. It seems to me that the "implementation specific environment
595 hacking functions" found in pcl/walker.lisp are completely messed
596 up. The good thing is that they appear to be barely used within
597 PCL and the munged environment object is passed to cmucl only
598 in calls to macroexpand-1, which is probably why this case fails.
599 SBCL uses essentially the same code, so if the environment hacking
600 is screwed up, it affects us too.
603 Using the pretty-printer from the command prompt gives funny
604 results, apparently because the pretty-printer doesn't know
605 about user's command input, including the user's carriage return
606 that the user, and therefore the pretty-printer thinks that
607 the new output block should start indented 2 or more characters
608 rightward of the correct location.
611 (probably related to bug #70)
612 As reported by Carl Witty on submit@bugs.debian.org 1999-05-08,
614 (in-package "CL-USER")
615 (defun equal-terms (termx termy)
617 ((alpha-equal-bound-term-lists (listx listy)
618 (or (and (null listx) (null listy))
620 (let ((bindings-x (bindings-of-bound-term (car listx)))
621 (bindings-y (bindings-of-bound-term (car listy))))
622 (if (and (null bindings-x) (null bindings-y))
623 (alpha-equal-terms (term-of-bound-term (car listx))
624 (term-of-bound-term (car listy)))
625 (and (= (length bindings-x) (length bindings-y))
627 (enter-binding-pairs (bindings-of-bound-term (car listx))
628 (bindings-of-bound-term (car listy)))
629 (alpha-equal-terms (term-of-bound-term (car listx))
630 (term-of-bound-term (car listy)))
631 (exit-binding-pairs (bindings-of-bound-term (car listx))
632 (bindings-of-bound-term (car listy)))))))
633 (alpha-equal-bound-term-lists (cdr listx) (cdr listy)))))
635 (alpha-equal-terms (termx termy)
636 (if (and (variable-p termx)
638 (equal-bindings (id-of-variable-term termx)
639 (id-of-variable-term termy))
640 (and (equal-operators-p (operator-of-term termx) (operator-of-term termy))
641 (alpha-equal-bound-term-lists (bound-terms-of-term termx)
642 (bound-terms-of-term termy))))))
646 (with-variable-invocation (alpha-equal-terms termx termy))))))
647 causes an assertion failure
648 The assertion (EQ (C::LAMBDA-TAIL-SET C::CALLER)
649 (C::LAMBDA-TAIL-SET (C::LAMBDA-HOME C::CALLEE))) failed.
651 Bob Rogers reports (1999-07-28 on cmucl-imp@cons.org) a smaller test
652 case with the same problem:
653 (defun parse-fssp-alignment ()
654 ;; Given an FSSP alignment file named by the argument . . .
655 (labels ((get-fssp-char ()
659 ;; Stub body, enough to tickle the bug.
660 (list (read-fssp-char)
664 ANSI specifies that the RESULT-TYPE argument of CONCATENATE must be
665 a subtype of SEQUENCE, but CONCATENATE doesn't check this properly:
666 (CONCATENATE 'SIMPLE-ARRAY #(1 2) '(3)) => #(1 2 3)
667 This also leads to funny behavior when derived type specifiers
668 are used, as originally reported by Milan Zamazal for CMU CL (on the
669 Debian bugs mailing list (?) 2000-02-27), then reported by Martin
670 Atzmueller for SBCL (2000-10-01 on sbcl-devel@lists.sourceforge.net):
671 (DEFTYPE FOO () 'SIMPLE-ARRAY)
672 (CONCATENATE 'FOO #(1 2) '(3))
673 => #<ARRAY-TYPE SIMPLE-ARRAY> is a bad type specifier for
675 The derived type specifier FOO should act the same way as the
676 built-in type SIMPLE-ARRAY here, but it doesn't. That problem
677 doesn't seem to exist for sequence types:
678 (DEFTYPE BAR () 'SIMPLE-VECTOR)
679 (CONCATENATE 'BAR #(1 2) '(3)) => #(1 2 3)
682 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
683 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
684 crashes SBCL. In general tracing anything which is used in the
685 implementation of TRACE is likely to have the same problem.
688 As reported by Daniel Solaz on cmucl-help@cons.org 2000-11-23,
689 SXHASH returns the same value for all non-STRUCTURE-OBJECT instances,
690 notably including all PCL instances. There's a limit to how much
691 SXHASH can do to return unique values for instances, but at least
692 it should probably look at the class name, the way that it does
693 for STRUCTURE-OBJECTs.
696 As reported by Martin Atzmueller on the sbcl-devel list 2000-11-22,
697 > There remains one issue, that is a bug in SBCL:
698 > According to my interpretation of the spec, the ":" and "@" modifiers
699 > should appear _after_ the comma-seperated arguments.
700 > Well, SBCL (and CMUCL for that matter) accept
701 > (ASSERT (STRING= (FORMAT NIL "~:8D" 1) " 1"))
702 > where the correct way (IMHO) should be
703 > (ASSERT (STRING= (FORMAT NIL "~8:D" 1) " 1"))
704 Probably SBCL should stop accepting the "~:8D"-style format arguments,
705 or at least issue a warning.
708 (probably related to bug #65)
709 The compiler doesn't like &OPTIONAL arguments in LABELS and FLET
711 (DEFUN FIND-BEFORE (ITEM SEQUENCE &KEY (TEST #'EQL))
712 (LABELS ((FIND-ITEM (OBJ SEQ TEST &OPTIONAL (VAL NIL))
713 (LET ((ITEM (FIRST SEQ)))
716 ((FUNCALL TEST OBJ ITEM)
719 (FIND-ITEM OBJ (REST SEQ) TEST (NCONC VAL `(,ITEM))))))))
720 (FIND-ITEM ITEM SEQUENCE TEST)))
721 from David Young's bug report on cmucl-help@cons.org 30 Nov 2000
722 causes sbcl-0.6.9 to fail with
723 error in function SB-KERNEL:ASSERT-ERROR:
724 The assertion (EQ (SB-C::LAMBDA-TAIL-SET SB-C::CALLER)
725 (SB-C::LAMBDA-TAIL-SET
726 (SB-C::LAMBDA-HOME SB-C::CALLEE))) failed.
729 (DECLAIM (OPTIMIZE ..)) doesn't work. E.g. even after
730 (DECLAIM (OPTIMIZE (SPEED 3))), things are still optimized with
731 the previous SPEED policy. This bug will probably get fixed in
732 0.6.9.x in a general cleanup of optimization policy.
735 (DECLAIM (OPTIMIZE ..)) doesn't work properly inside LOCALLY forms.
738 As noted in the ANSI specification for COERCE, (COERCE 3 'COMPLEX)
739 gives a result which isn't COMPLEX. The result type optimizer
740 for COERCE doesn't know this, perhaps because it was written before
741 ANSI threw this curveball: the optimizer thinks that COERCE always
742 returns a result of the specified type. Thus while the interpreted
744 (DEFUN TRICKY (X) (TYPEP (COERCE X 'COMPLEX) 'COMPLEX))
745 returns the correct result,
747 the compiled function
753 As reported by Martin Atzmueller on sbcl-devel 26 Dec 2000,
754 ANSI says that WITH-OUTPUT-TO-STRING should have a keyword
755 :ELEMENT-TYPE, but in sbcl-0.6.9 this is not defined for
756 WITH-OUTPUT-TO-STRING.
759 ANSI says in one place that type declarations can be abbreviated even
760 when the type name is not a symbol, e.g.
761 (DECLAIM ((VECTOR T) *FOOVECTOR*))
762 SBCL doesn't support this. But ANSI says in another place that this
763 isn't allowed. So it's not clear this is a bug after all. (See the
764 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
768 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
769 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
770 an easily guessable temporary filename in a way which might open
771 applications using LOAD-FOREIGN to hijacking by malicious users
772 on the same machine. Incantations for doing this safely are
773 floating around the net in various "how to write secure programs
774 despite Unix" documents, and it would be good to (1) fix this in
775 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
776 files and make it share the same new safe logic.
779 (fixed early Feb 2001 by MNA)
782 As reported by wbuss@TELDA.NET (Wolfhard Buss) on cmucl-help
785 (loop with (a . b) of-type float = '(0.0 . 1.0)
786 and (c . d) of-type float = '(2.0 . 3.0)
787 return (list a b c d))
788 should evaluate to (0.0 1.0 2.0 3.0). cmucl-18c disagrees and
789 invokes the debugger: "B is not of type list".
790 SBCL does the same thing.
793 Functions are assigned names based on the context in which they're
794 defined. This is less than ideal for the functions which are
795 used to implement CLOS methods. E.g. the output of
796 (DESCRIBE 'PRINT-OBJECT) lists functions like
797 #<FUNCTION "DEF!STRUCT (TRACE-INFO (:MAKE-LOAD-FORM-FUN SB-KERNEL:JUST-DUMP-IT-NORMALLY) (:PRINT-OBJECT #))" {1020E49}>
799 #<FUNCTION "MACROLET ((FORCE-DELAYED-DEF!METHODS NIL #))" {1242871}>
800 It would be better if these functions' names always identified
801 them as methods, and identified their generic functions and
805 KNOWN BUGS RELATED TO THE IR1 INTERPRETER
807 (Note: At some point, the pure interpreter (actually a semi-pure
808 interpreter aka "the IR1 interpreter") will probably go away, replaced
810 (DEFUN EVAL (X) (FUNCALL (COMPILE NIL (LAMBDA ..)))))
811 and at that time these bugs should either go away automatically or
812 become more tractable to fix. Until then, they'll probably remain,
813 since some of them aren't considered urgent, and the rest are too hard
814 to fix as long as so many special cases remain. After the IR1
815 interpreter goes away is also the preferred time to start
816 systematically exterminating cases where debugging functionality
817 (backtrace, breakpoint, etc.) breaks down, since getting rid of the
818 IR1 interpreter will reduce the number of special cases we need to
822 The FUNCTION special operator doesn't check properly whether its
823 argument is a function name. E.g. (FUNCTION (X Y)) returns a value
824 instead of failing with an error. (Later attempting to funcall the
825 value does cause an error.)
828 COMPILED-FUNCTION-P bogusly reports T for interpreted functions:
829 * (DEFUN FOO (X) (- 12 X))
831 * (COMPILED-FUNCTION-P #'FOO)
836 (DEFVAR *SUPPRESS-P* T)
837 (EVAL '(UNLESS *SUPPRESS-P*
838 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
839 (FORMAT T "surprise!"))))
840 prints "surprise!". Probably the entire EVAL-WHEN mechanism ought to be
841 rewritten from scratch to conform to the ANSI definition, abandoning
842 the *ALREADY-EVALED-THIS* hack which is used in sbcl-0.6.8.9 (and
843 in the original CMU CL source, too). This should be easier to do --
844 though still nontrivial -- once the various IR1 interpreter special
848 EVAL-WHEN's idea of what's a toplevel form is even more screwed up
849 than the example in IR1-3 would suggest, since COMPILE-FILE and
850 COMPILE both print both "right now!" messages when compiling the
854 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
855 (PRINT "yes! right now!"))
858 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
859 (PRINT "no! right now!"))
861 and while EVAL doesn't print the "right now!" messages, the first
862 FUNCALL on the value returned by EVAL causes both of them to be printed.
865 The system accepts DECLAIM in most places where DECLARE would be
866 accepted, without even issuing a warning. ANSI allows this, but since
867 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
868 meaning is rather different, and it's unlikely that the user
869 has a good reason for doing DECLAIM not at top level, it would be
870 good to issue a STYLE-WARNING when this happens. A possible
871 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
872 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
873 [This is considered an IR1-interpreter-related bug because until
874 EVAL-WHEN is rewritten, which won't happen until after the IR1
875 interpreter is gone, the system's notion of what's a top-level form
876 and what's not will remain too confused to fix this problem.]