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 The type system doesn't understand the KEYWORD type very well:
123 (SUBTYPEP 'KEYWORD 'SYMBOL) => NIL, NIL
124 It might be possible to fix this by changing the definition of
125 KEYWORD to (AND SYMBOL (SATISFIES KEYWORDP)), but the type system
126 would need to be a bit smarter about AND types, too:
127 (SUBTYPEP '(AND SYMBOL KEYWORD) 'SYMBOL) => NIL, NIL
128 (The type system does know something about AND types already,
129 (SUBTYPEP '(AND INTEGER FLOAT) 'NUMBER) => T, T
130 (SUBTYPEP '(AND INTEGER FIXNUM) 'NUMBER) =>T, T
131 so likely this is a small patch.)
134 Floating point infinities are screwed up. [When I was converting CMU CL
135 to SBCL, I was looking for complexity to delete, and I thought it was safe
136 to just delete support for floating point infinities. It wasn't: they're
137 generated by the floating point hardware even when we remove support
138 for them in software. Also we claim the :IEEE-FLOATING-POINT feature,
139 and I think that means we should support infinities.-- WHN] Support
140 for them should be restored.
143 The ANSI syntax for non-STANDARD method combination types in CLOS is
144 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
145 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
146 If you mess this up, omitting the PROGN qualifier in in DEFMETHOD,
147 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
148 (DEFMETHOD FOO ((X BAR)) (PRINT 'NUMBER))
149 the error mesage is not easy to understand:
150 INVALID-METHOD-ERROR was called outside the dynamic scope
151 of a method combination function (inside the body of
152 DEFINE-METHOD-COMBINATION or a method on the generic
153 function COMPUTE-EFFECTIVE-METHOD).
154 It would be better if it were more informative, a la
155 The method combination type for this method (STANDARD) does
156 not match the method combination type for the generic function
158 Also, after you make the mistake of omitting the PROGN qualifier
159 on a DEFMETHOD, doing a new DEFMETHOD with the correct qualifier
161 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
163 INVALID-METHOD-ERROR was called outside the dynamic scope
164 of a method combination function (inside the body of
165 DEFINE-METHOD-COMBINATION or a method on the generic
166 function COMPUTE-EFFECTIVE-METHOD).
167 This is not very helpful..
170 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
171 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
172 (Also, when this is fixed, we can enable the code in PROCLAIM which
173 checks for incompatible FTYPE redeclarations.)
176 from DTC on the CMU CL mailing list 25 Feb 2000:
177 ;;; Compiler fails when this file is compiled.
179 ;;; Problem shows up in delete-block within ir1util.lisp. The assertion
180 ;;; (assert (member (functional-kind lambda) '(:let :mv-let :assignment)))
181 ;;; fails within bind node branch.
183 ;;; Note that if c::*check-consistency* is enabled then an un-reached
184 ;;; entry is also reported.
187 (declare (values nil))
204 (let ((ttt #'(lambda () (go cccc))))
205 (declare (special ttt))
206 (return-from bbbb nil))
209 (return-from bbbb nil))))))
212 (I *think* this is a bug. It certainly seems like strange behavior. But
213 the ANSI spec is scary, dark, and deep..)
214 (FORMAT NIL "~,1G" 1.4) => "1. "
215 (FORMAT NIL "~3,1G" 1.4) => "1. "
218 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
220 (setf (find-class 'ccc1) (find-class 'ccc))
221 (defmethod zut ((c ccc1)) 123)
222 DTC's recommended workaround from the mailing list 3 Mar 2000:
223 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
226 The ANSI spec, in section "22.3.5.2 Tilde Less-Than-Sign: Logical Block",
227 says that an error is signalled if ~W, ~_, ~<...~:>, ~I, or ~:T is used
228 inside "~<..~>" (without the colon modifier on the closing syntax).
229 However, SBCL doesn't do this:
230 * (FORMAT T "~<munge~wegnum~>" 12)
235 When too many files are opened, OPEN will fail with an
236 uninformative error message
237 error in function OPEN: error opening #P"/tmp/foo.lisp": NIL
238 instead of saying that too many files are open.
241 Right now, when COMPILE-FILE has a read error, it actually pops
242 you into the debugger before giving up on the file. It should
243 instead handle the error, perhaps issuing (and handling)
244 a secondary error "caught ERROR: unrecoverable error during compilation"
245 and then return with FAILURE-P true,
248 reported by Sam Steingold on the cmucl-imp mailing list 12 May 2000:
250 Also, there is another bug: `array-displacement' should return an array
251 or nil as first value (as per ANSI CL), while CMUCL declares it as
252 returning an array as first value always.
255 Sometimes (SB-EXT:QUIT) fails with
256 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
257 Process inferior-lisp exited abnormally with code 1
258 I haven't noticed a repeatable case of this yet.
261 some sort of bug in inlining and RETURN-FROM in sbcl-0.6.5: Compiling
264 (BLOCK USED-BY-SOME-Y?
267 (UNLESS (REJECTED? Y)
268 (RETURN-FROM USED-BY-SOME-Y? T)))))
269 (DECLARE (INLINE FROB))
274 error in function SB-KERNEL:ASSERT-ERROR:
275 The assertion (EQ (SB-C::CONTINUATION-KIND SB-C::CONT) :BLOCK-START) failed.
276 This is still present in sbcl-0.6.8.
279 In some cases the compiler believes type declarations on array
280 elements without checking them, e.g.
281 (DECLAIM (OPTIMIZE (SAFETY 3) (SPEED 1) (SPACE 1)))
284 (DECLARE (TYPE (SIMPLE-ARRAY CONS 1) X))
285 (WHEN (CONSP (AREF X 0))
287 (BAR (VECTOR (MAKE-FOO :A 11 :B 12)))
290 in SBCL 0.6.5 (and also in CMU CL 18b). This does not happen for
291 all cases, e.g. the type assumption *is* checked if the array
292 elements are declared to be of some structure type instead of CONS.
295 The printer doesn't report closures very well. This is true in
299 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
300 It would be nice to make closures have a settable name slot,
301 and make things like DEFSTRUCT and FLET, which create closures,
302 set helpful values into this slot.
305 And as long as we're wishing, it would be awfully nice if INSPECT could
306 also report on closures, telling about the values of the bound variables.
309 The compiler assumes that any time a function of declared FTYPE
310 doesn't signal an error, its arguments were of the declared type.
311 E.g. compiling and loading
312 (DECLAIM (OPTIMIZE (SAFETY 3)))
313 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
314 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE) FACTORIAL)))
316 (COND ((> (FACTORIAL X) 1.0E6)
317 (FORMAT T "too big~%"))
319 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
321 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
324 will cause the INTEGERP case to be selected, giving bogus output a la
326 This violates the "declarations are assertions" principle.
327 According to the ANSI spec, in the section "System Class FUNCTION",
328 this is a case of "lying to the compiler", but the lying is done
329 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
330 not by the unexpectedly general definition of FACTORIAL. In any case,
331 "declarations are assertions" means that lying to the compiler should
332 cause an error to be signalled, and should not cause a bogus
333 result to be returned. Thus, the compiler should not assume
334 that arbitrary functions check their argument types. (It might
335 make sense to add another flag (CHECKED?) to DEFKNOWN to
336 identify functions which *do* check their argument types.)
339 DEFMETHOD doesn't check the syntax of &REST argument lists properly,
340 accepting &REST even when it's not followed by an argument name:
341 (DEFMETHOD FOO ((X T) &REST) NIL)
344 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
345 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
347 (DECLARE (TYPE INDEXOID X))
348 (THE (VALUES INDEXOID)
350 where the implementation of the type check in function FOO
351 includes a full call to %TYPEP. There are also some fundamental problems
352 with the interpretation of VALUES types (inherited from CMU CL, and
353 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
354 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
357 The definitions of SIGCONTEXT-FLOAT-REGISTER and
358 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
359 supported on FreeBSD because the floating point state is not saved,
360 but at least as of FreeBSD 4.0, the floating point state *is* saved,
361 so they could be supported after all. Very likely
362 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
365 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
366 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
367 issue with 'union, 'and, 'or etc. So even though according to the
368 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
369 (and now SBCL) interpret them as legal types.
372 ANSI specifies DEFINE-SYMBOL-MACRO, but it's not defined in SBCL.
373 CMU CL added it ca. Aug 13, 2000, after some discussion on the mailing
374 list, and it is probably possible to use substantially the same
375 patches to add it to SBCL.
378 a slew of floating-point-related errors reported by Peter Van Eynde
380 a: (SQRT -9.0) fails, because SB-KERNEL::COMPLEX-SQRT is undefined.
381 Similarly, COMPLEX-ASIN, COMPLEX-ACOS, COMPLEX-ACOSH, and others
383 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
384 should probably be 1.4012985e-45. In SBCL,
385 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
386 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
387 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
388 and LEAST-NEGATIVE-LONG-FLOAT.
389 c: Many expressions generate floating infinity:
394 PVE's regression tests want them to raise errors. SBCL
395 generates the infinities instead, which may or may not be
396 conforming behavior, but then blow it by being unable to
397 output the infinities, since support for infinities is generally
398 broken, and in particular SB-IMPL::OUTPUT-FLOAT-INFINITY is
400 d: (in section12.erg) various forms a la
401 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
402 don't give the right behavior.
405 type safety errors reported by Peter Van Eynde July 25, 2000:
406 a: (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
408 In general lengths of array type specifications aren't
409 checked by COERCE, so it fails when the spec is
410 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
411 b: CONCATENATE has the same problem of not checking the length
412 of specified output array types. MAKE-SEQUENCE and MAP and
413 MERGE also have the same problem.
414 c: (COERCE 'AND 'FUNCTION) returns something related to
415 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
416 d: ELT signals SIMPLE-ERROR if its index argument
417 isn't a valid index for its sequence argument, but should
418 signal TYPE-ERROR instead.
419 e: FILE-LENGTH is supposed to signal a type error when its
420 argument is not a stream associated with a file, but doesn't.
421 f: (FLOAT-RADIX 2/3) should signal an error instead of
423 g: (LOAD "*.lsp") should signal FILE-ERROR.
424 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
425 should signal TYPE-ERROR.
426 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
427 be used for input and output as needed. It should fail with
428 TYPE-ERROR when handed e.g. the results of
429 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
430 the inappropriate positions, but doesn't.
431 j: (PARSE-NAMESTRING (COERCE (LIST #\f #\o #\o (CODE-CHAR 0) #\4 #\8)
433 should probably signal an error instead of making a pathname with
435 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
436 not a binary input stream, but instead cheerfully reads from
437 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
440 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
441 a: (DEFCLASS FOO () (A B A)) should signal a PROGRAM-ERROR, and
443 b: (DEFCLASS FOO () (A B A) (:DEFAULT-INITARGS X A X B)) should
444 signal a PROGRAM-ERROR, and doesn't.
445 c: (DEFCLASS FOO07 NIL ((A :ALLOCATION :CLASS :ALLOCATION :CLASS))),
446 and other DEFCLASS forms with duplicate specifications in their
447 slots, should signal a PROGRAM-ERROR, and doesn't.
448 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
449 causes a COMPILER-ERROR.
452 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
453 a: (SYMBOL-MACROLET ((T TRUE)) ..) should probably signal
454 PROGRAM-ERROR, but SBCL accepts it instead.
455 b: SYMBOL-MACROLET should refuse to bind something which is
456 declared as a global variable, signalling PROGRAM-ERROR.
457 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
458 it binds is declared SPECIAL inside.
461 LOOP bugs reported by Peter Van Eynde July 25, 2000:
462 a: (LOOP WITH (A B) DO (PRINT 1)) is a syntax error according to
463 the definition of WITH clauses given in the ANSI spec, but
464 compiles and runs happily in SBCL.
465 b: a messy one involving package iteration:
466 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<))))
467 Should be: (("blah" "blah2") ("blah2"))
468 SBCL: (("blah") ("blah2"))
469 * (LET ((X 1)) (LOOP FOR I BY (INCF X) FROM X TO 10 COLLECT I))
470 doesn't work -- SBCL's LOOP says BY isn't allowed in a FOR clause.
473 type system errors reported by Peter Van Eynde July 25, 2000:
474 a: (SUBTYPEP 'BIGNUM 'INTEGER) => NIL, NIL
475 but should be (VALUES T T) instead.
476 b: (SUBTYPEP 'EXTENDED-CHAR 'CHARACTER) => NIL, NIL
477 but should be (VALUES T T) instead.
478 c: (SUBTYPEP '(INTEGER (0) (0)) 'NIL) dies with nested errors.
479 d: In general, the system doesn't like '(INTEGER (0) (0)) -- it
480 blows up at the level of SPECIFIER-TYPE with
481 "Lower bound (0) is greater than upper bound (0)." Probably
482 SPECIFIER-TYPE should return NIL instead.
483 e: (TYPEP 0 '(COMPLEX (EQL 0)) fails with
484 "Component type for Complex is not numeric: (EQL 0)."
485 This might be easy to fix; the type system already knows
486 that (SUBTYPEP '(EQL 0) 'NUMBER) is true.
487 f: The type system doesn't know about the condition system,
488 so that e.g. (TYPEP 'SIMPLE-ERROR 'ERROR)=>NIL.
489 g: The type system isn't all that smart about relationships
490 between hairy types, as shown in the type.erg test results,
491 e.g. (SUBTYPEP 'CONS '(NOT ATOM)) => NIL, NIL.
494 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
496 (DEFGENERIC FOO02 (X))
497 (DEFMETHOD FOO02 ((X NUMBER)) T)
498 (LET ((M (FIND-METHOD (FUNCTION FOO02)
500 (LIST (FIND-CLASS (QUOTE NUMBER))))))
501 (REMOVE-METHOD (FUNCTION FOO02) M)
502 (DEFGENERIC FOO03 (X))
503 (ADD-METHOD (FUNCTION FOO03) M)))
504 should give an error, but SBCL allows it.
505 b: READ should probably return READER-ERROR, not the bare
506 arithmetic error, when input a la "1/0" or "1e1000" causes
510 It has been reported (e.g. by Peter Van Eynde) that there are
511 several metaobject protocol "errors". (In order to fix them, we might
512 need to document exactly what metaobject protocol specification
513 we're following -- the current code is just inherited from PCL.)
516 another error from Peter Van Eynde 5 September 2000:
517 (FORMAT NIL "~F" "FOO") should work, but instead reports an error.
518 PVE submitted a patch to deal with this bug, but it exposes other
519 comparably serious bugs, so I didn't apply it. It looks as though
520 the FORMAT code needs a fair amount of rewriting in order to comply
521 with the various details of the ANSI spec.
524 The implementation of #'+ returns its single argument without
525 type checking, e.g. (+ "illegal") => "illegal".
528 In sbcl-0.6.7, there is no doc string for CL:PUSH, probably
529 because it's defined with the DEFMACRO-MUNDANELY macro and something
530 is wrong with doc string setting in that macro.
533 Attempting to use COMPILE on something defined by DEFMACRO fails:
534 (DEFMACRO FOO (X) (CONS X X))
536 Error in function C::GET-LAMBDA-TO-COMPILE:
537 #<Closure Over Function "DEFUN (SETF MACRO-FUNCTION)" {480E21B1}> was defined in a non-null environment.
540 (SUBTYPEP '(AND ZILCH INTEGER) 'ZILCH)
544 CL:*DEFAULT-PATHNAME-DEFAULTS* doesn't behave as ANSI suggests (reflecting
545 current working directory). And there's no supported way to update
546 or query the current working directory (a la Unix "chdir" and "pwd"),
547 which is functionality that ILISP needs (and currently gets with low-level
551 The debugger LIST-LOCATIONS command doesn't work properly.
554 Compiling and loading
555 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
557 then requesting a BACKTRACE at the debugger prompt gives no information
558 about where in the user program the problem occurred.
561 The compiler is supposed to do type inference well enough that
564 ((SIMPLE-ARRAY SINGLE-FLOAT)
566 (DECLARE (TYPE (SIMPLE-ARRAY SINGLE-FLOAT) X))
569 is redundant. However, as reported by Juan Jose Garcia Ripoll for
570 CMU CL, it sometimes doesn't. Adding declarations is a pretty good
571 workaround for the problem for now, but can't be done by the TYPECASE
572 macros themselves, since it's too hard for the macro to detect
573 assignments to the variable within the clause.
574 Note: The compiler *is* smart enough to do the type inference in
575 many cases. This case, derived from a couple of MACROEXPAND-1
576 calls on Ripoll's original test case,
578 (DECLARE (OPTIMIZE SPEED (SAFETY 0)))
579 (COND ((TYPEP A '(SIMPLE-ARRAY SINGLE-FLOAT)) NIL
580 (LET ((LENGTH (ARRAY-TOTAL-SIZE A)))
581 (LET ((I 0) (G2554 LENGTH))
582 (DECLARE (TYPE REAL G2554) (TYPE REAL I))
585 (WHEN (>= I G2554) (GO SB-LOOP::END-LOOP))
586 (SETF (ROW-MAJOR-AREF A I) (- (ROW-MAJOR-AREF A I)))
587 (GO SB-LOOP::NEXT-LOOP)
588 SB-LOOP::END-LOOP))))))
589 demonstrates the problem; but the problem goes away if the TAGBODY
590 and GO forms are removed (leaving the SETF in ordinary, non-looping
591 code), or if the TAGBODY and GO forms are retained, but the
592 assigned value becomes 0.0 instead of (- (ROW-MAJOR-AREF A I)).
595 Paul Werkowski wrote on cmucl-imp@cons.org 2000-11-15
596 I am looking into this problem that showed up on the cmucl-help
597 list. It seems to me that the "implementation specific environment
598 hacking functions" found in pcl/walker.lisp are completely messed
599 up. The good thing is that they appear to be barely used within
600 PCL and the munged environment object is passed to cmucl only
601 in calls to macroexpand-1, which is probably why this case fails.
602 SBCL uses essentially the same code, so if the environment hacking
603 is screwed up, it affects us too.
606 Using the pretty-printer from the command prompt gives funny
607 results, apparently because the pretty-printer doesn't know
608 about user's command input, including the user's carriage return
609 that the user, and therefore the pretty-printer thinks that
610 the new output block should start indented 2 or more characters
611 rightward of the correct location.
614 (probably related to bug #70)
615 As reported by Carl Witty on submit@bugs.debian.org 1999-05-08,
617 (in-package "CL-USER")
618 (defun equal-terms (termx termy)
620 ((alpha-equal-bound-term-lists (listx listy)
621 (or (and (null listx) (null listy))
623 (let ((bindings-x (bindings-of-bound-term (car listx)))
624 (bindings-y (bindings-of-bound-term (car listy))))
625 (if (and (null bindings-x) (null bindings-y))
626 (alpha-equal-terms (term-of-bound-term (car listx))
627 (term-of-bound-term (car listy)))
628 (and (= (length bindings-x) (length bindings-y))
630 (enter-binding-pairs (bindings-of-bound-term (car listx))
631 (bindings-of-bound-term (car listy)))
632 (alpha-equal-terms (term-of-bound-term (car listx))
633 (term-of-bound-term (car listy)))
634 (exit-binding-pairs (bindings-of-bound-term (car listx))
635 (bindings-of-bound-term (car listy)))))))
636 (alpha-equal-bound-term-lists (cdr listx) (cdr listy)))))
638 (alpha-equal-terms (termx termy)
639 (if (and (variable-p termx)
641 (equal-bindings (id-of-variable-term termx)
642 (id-of-variable-term termy))
643 (and (equal-operators-p (operator-of-term termx) (operator-of-term termy))
644 (alpha-equal-bound-term-lists (bound-terms-of-term termx)
645 (bound-terms-of-term termy))))))
649 (with-variable-invocation (alpha-equal-terms termx termy))))))
650 causes an assertion failure
651 The assertion (EQ (C::LAMBDA-TAIL-SET C::CALLER)
652 (C::LAMBDA-TAIL-SET (C::LAMBDA-HOME C::CALLEE))) failed.
654 Bob Rogers reports (1999-07-28 on cmucl-imp@cons.org) a smaller test
655 case with the same problem:
656 (defun parse-fssp-alignment ()
657 ;; Given an FSSP alignment file named by the argument . . .
658 (labels ((get-fssp-char ()
662 ;; Stub body, enough to tickle the bug.
663 (list (read-fssp-char)
667 ANSI specifies that the RESULT-TYPE argument of CONCATENATE must be
668 a subtype of SEQUENCE, but CONCATENATE doesn't check this properly:
669 (CONCATENATE 'SIMPLE-ARRAY #(1 2) '(3)) => #(1 2 3)
670 This also leads to funny behavior when derived type specifiers
671 are used, as originally reported by Milan Zamazal for CMU CL (on the
672 Debian bugs mailing list (?) 2000-02-27), then reported by Martin
673 Atzmueller for SBCL (2000-10-01 on sbcl-devel@lists.sourceforge.net):
674 (DEFTYPE FOO () 'SIMPLE-ARRAY)
675 (CONCATENATE 'FOO #(1 2) '(3))
676 => #<ARRAY-TYPE SIMPLE-ARRAY> is a bad type specifier for
678 The derived type specifier FOO should act the same way as the
679 built-in type SIMPLE-ARRAY here, but it doesn't. That problem
680 doesn't seem to exist for sequence types:
681 (DEFTYPE BAR () 'SIMPLE-VECTOR)
682 (CONCATENATE 'BAR #(1 2) '(3)) => #(1 2 3)
685 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
686 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
687 crashes SBCL. In general tracing anything which is used in the
688 implementation of TRACE is likely to have the same problem.
691 As reported by Daniel Solaz on cmucl-help@cons.org 2000-11-23,
692 SXHASH returns the same value for all non-STRUCTURE-OBJECT instances,
693 notably including all PCL instances. There's a limit to how much
694 SXHASH can do to return unique values for instances, but at least
695 it should probably look at the class name, the way that it does
696 for STRUCTURE-OBJECTs.
699 As reported by Martin Atzmueller on the sbcl-devel list 2000-11-22,
700 > There remains one issue, that is a bug in SBCL:
701 > According to my interpretation of the spec, the ":" and "@" modifiers
702 > should appear _after_ the comma-seperated arguments.
703 > Well, SBCL (and CMUCL for that matter) accept
704 > (ASSERT (STRING= (FORMAT NIL "~:8D" 1) " 1"))
705 > where the correct way (IMHO) should be
706 > (ASSERT (STRING= (FORMAT NIL "~8:D" 1) " 1"))
707 Probably SBCL should stop accepting the "~:8D"-style format arguments,
708 or at least issue a warning.
711 (probably related to bug #65)
712 The compiler doesn't like &OPTIONAL arguments in LABELS and FLET
714 (DEFUN FIND-BEFORE (ITEM SEQUENCE &KEY (TEST #'EQL))
715 (LABELS ((FIND-ITEM (OBJ SEQ TEST &OPTIONAL (VAL NIL))
716 (LET ((ITEM (FIRST SEQ)))
719 ((FUNCALL TEST OBJ ITEM)
722 (FIND-ITEM OBJ (REST SEQ) TEST (NCONC VAL `(,ITEM))))))))
723 (FIND-ITEM ITEM SEQUENCE TEST)))
724 from David Young's bug report on cmucl-help@cons.org 30 Nov 2000
725 causes sbcl-0.6.9 to fail with
726 error in function SB-KERNEL:ASSERT-ERROR:
727 The assertion (EQ (SB-C::LAMBDA-TAIL-SET SB-C::CALLER)
728 (SB-C::LAMBDA-TAIL-SET
729 (SB-C::LAMBDA-HOME SB-C::CALLEE))) failed.
732 (DECLAIM (OPTIMIZE ..)) doesn't work. E.g. even after
733 (DECLAIM (OPTIMIZE (SPEED 3))), things are still optimized with
734 the previous SPEED policy. This bug will probably get fixed in
735 0.6.9.x in a general cleanup of optimization policy.
738 (DECLAIM (OPTIMIZE ..)) doesn't work properly inside LOCALLY forms.
741 As noted in the ANSI specification for COERCE, (COERCE 3 'COMPLEX)
742 gives a result which isn't COMPLEX. The result type optimizer
743 for COERCE doesn't know this, perhaps because it was written before
744 ANSI threw this curveball: the optimizer thinks that COERCE always
745 returns a result of the specified type. Thus while the interpreted
747 (DEFUN TRICKY (X) (TYPEP (COERCE X 'COMPLEX) 'COMPLEX))
748 returns the correct result,
750 the compiled function
756 As reported by Martin Atzmueller on sbcl-devel 26 Dec 2000,
757 ANSI says that WITH-OUTPUT-TO-STRING should have a keyword
758 :ELEMENT-TYPE, but in sbcl-0.6.9 this is not defined for
759 WITH-OUTPUT-TO-STRING.
762 ANSI says in one place that type declarations can be abbreviated even
763 when the type name is not a symbol, e.g.
764 (DECLAIM ((VECTOR T) *FOOVECTOR*))
765 SBCL doesn't support this. But ANSI says in another place that this
766 isn't allowed. So it's not clear this is a bug after all. (See the
767 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
771 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
772 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
773 an easily guessable temporary filename in a way which might open
774 applications using LOAD-FOREIGN to hijacking by malicious users
775 on the same machine. Incantations for doing this safely are
776 floating around the net in various "how to write secure programs
777 despite Unix" documents, and it would be good to (1) fix this in
778 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
779 files and make it share the same new safe logic.
782 (fixed early Feb 2001 by MNA)
785 As reported by wbuss@TELDA.NET (Wolfhard Buss) on cmucl-help
788 (loop with (a . b) of-type float = '(0.0 . 1.0)
789 and (c . d) of-type float = '(2.0 . 3.0)
790 return (list a b c d))
791 should evaluate to (0.0 1.0 2.0 3.0). cmucl-18c disagrees and
792 invokes the debugger: "B is not of type list".
793 SBCL does the same thing.
796 Functions are assigned names based on the context in which they're
797 defined. This is less than ideal for the functions which are
798 used to implement CLOS methods. E.g. the output of
799 (DESCRIBE 'PRINT-OBJECT) lists functions like
800 #<FUNCTION "DEF!STRUCT (TRACE-INFO (:MAKE-LOAD-FORM-FUN SB-KERNEL:JUST-DUMP-IT-NORMALLY) (:PRINT-OBJECT #))" {1020E49}>
802 #<FUNCTION "MACROLET ((FORCE-DELAYED-DEF!METHODS NIL #))" {1242871}>
803 It would be better if these functions' names always identified
804 them as methods, and identified their generic functions and
808 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
809 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
810 the mean of the distribution can be systematically O(0.1%) wrong.
811 Just increasing R-I-E-B is probably not a good solution, since
812 it would decrease efficiency more than is probably necessary. Perhaps
813 using some sort of accept/reject method would be better.
815 KNOWN BUGS RELATED TO THE IR1 INTERPRETER
817 (Note: At some point, the pure interpreter (actually a semi-pure
818 interpreter aka "the IR1 interpreter") will probably go away, replaced
820 (DEFUN EVAL (X) (FUNCALL (COMPILE NIL (LAMBDA ..)))))
821 and at that time these bugs should either go away automatically or
822 become more tractable to fix. Until then, they'll probably remain,
823 since some of them aren't considered urgent, and the rest are too hard
824 to fix as long as so many special cases remain. After the IR1
825 interpreter goes away is also the preferred time to start
826 systematically exterminating cases where debugging functionality
827 (backtrace, breakpoint, etc.) breaks down, since getting rid of the
828 IR1 interpreter will reduce the number of special cases we need to
832 The FUNCTION special operator doesn't check properly whether its
833 argument is a function name. E.g. (FUNCTION (X Y)) returns a value
834 instead of failing with an error. (Later attempting to funcall the
835 value does cause an error.)
838 COMPILED-FUNCTION-P bogusly reports T for interpreted functions:
839 * (DEFUN FOO (X) (- 12 X))
841 * (COMPILED-FUNCTION-P #'FOO)
846 (DEFVAR *SUPPRESS-P* T)
847 (EVAL '(UNLESS *SUPPRESS-P*
848 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
849 (FORMAT T "surprise!"))))
850 prints "surprise!". Probably the entire EVAL-WHEN mechanism ought to be
851 rewritten from scratch to conform to the ANSI definition, abandoning
852 the *ALREADY-EVALED-THIS* hack which is used in sbcl-0.6.8.9 (and
853 in the original CMU CL source, too). This should be easier to do --
854 though still nontrivial -- once the various IR1 interpreter special
858 EVAL-WHEN's idea of what's a toplevel form is even more screwed up
859 than the example in IR1-3 would suggest, since COMPILE-FILE and
860 COMPILE both print both "right now!" messages when compiling the
864 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
865 (PRINT "yes! right now!"))
868 (EVAL-WHEN (:COMPILE-TOPLEVEL :LOAD-TOPLEVEL :EXECUTE)
869 (PRINT "no! right now!"))
871 and while EVAL doesn't print the "right now!" messages, the first
872 FUNCALL on the value returned by EVAL causes both of them to be printed.
875 The system accepts DECLAIM in most places where DECLARE would be
876 accepted, without even issuing a warning. ANSI allows this, but since
877 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
878 meaning is rather different, and it's unlikely that the user
879 has a good reason for doing DECLAIM not at top level, it would be
880 good to issue a STYLE-WARNING when this happens. A possible
881 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
882 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
883 [This is considered an IR1-interpreter-related bug because until
884 EVAL-WHEN is rewritten, which won't happen until after the IR1
885 interpreter is gone, the system's notion of what's a top-level form
886 and what's not will remain too confused to fix this problem.]