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 please please include enough information in a bug report
9 that someone reading 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.
25 KNOWN BUGS RELATED TO THE IR1 INTERPRETER
27 At some point, the pure interpreter (aka the "IR1 interpreter") will
28 probably go away (replaced by constructs like
29 (DEFUN EVAL (X) (FUNCALL (COMPILE NIL (LAMBDA ..)))))
30 and at that time these bugs should go away automatically. Until then,
31 they'll probably remain, since they're not considered urgent.
34 The FUNCTION special operator doesn't check properly whether its
35 argument is a function name. E.g. (FUNCTION (X Y)) returns a value
36 instead of failing with an error. (Later attempting to funcall the
37 value does cause an error.)
40 COMPILED-FUNCTION-P bogusly reports T for interpreted functions:
41 * (DEFUN FOO (X) (- 12 X))
43 * (COMPILED-FUNCTION-P #'FOO)
49 (There is also some information on bugs in the manual page and in the
50 TODO file. Eventually more such information may move here.)
53 Failure in initialization files is not handled gracefully -- it's
54 a throw to TOP-LEVEL-CATCHER, which is not caught until we enter
55 TOPLEVEL-REPL. Code should be added to catch such THROWs even when
56 we're not in TOPLEVEL-REPL and do *something* with them (probably
57 complaining about an error outside TOPLEVEL-REPL, perhaps printing
58 a BACKTRACE, then terminating execution of SBCL).
61 DEFSTRUCT should almost certainly overwrite the old LAYOUT information
62 instead of just punting when a contradictory structure definition
66 It should cause a STYLE-WARNING, not a full WARNING, when a structure
67 slot default value does not match the declared structure slot type.
68 (The current behavior is consistent with SBCL's behavior elsewhere,
69 and would not be a problem, except that the other behavior is
70 specifically required by the ANSI spec.)
73 It should cause a STYLE-WARNING, not a WARNING, when the system ignores
74 an FTYPE proclamation for a slot accessor.
77 Error reporting on various stream-requiring operations is not
78 very good when the stream argument has the wrong type, because
79 the operation tries to fall through to Gray stream code, and then
80 dies because it's undefined. E.g.
81 (PRINT-UNREADABLE-OBJECT (*STANDARD-OUTPUT* 1))
82 gives the error message
83 error in SB-KERNEL::UNDEFINED-SYMBOL-ERROR-HANDLER:
84 The function SB-IMPL::STREAM-WRITE-STRING is undefined.
85 It would be more useful and correct to signal a TYPE-ERROR:
87 (It wouldn't be terribly difficult to write stubs for all the
88 Gray stream functions that the old CMU CL code expects, with
89 each stub just raising the appropriate TYPE-ERROR.)
92 bogus warnings about undefined functions for magic functions like
93 SB!C::%%DEFUN and SB!C::%DEFCONSTANT when cross-compiling files
94 like src/code/float.lisp
97 The "byte compiling top-level form:" output ought to be condensed.
98 Perhaps any number of such consecutive lines ought to turn into a
99 single "byte compiling top-level forms:" line.
102 Compiling a file containing the erroneous program
106 (DEFSTRUCT (BAR (:INCLUDE FOO))
109 gives only the not-very-useful message
111 (during macroexpansion)
112 Condition PROGRAM-ERROR was signalled.
113 (The specific message which says that the problem was duplicate
114 slot names gets lost.)
117 The handling of IGNORE declarations on lambda list arguments of
118 DEFMETHOD is at least weird, and in fact seems broken and useless.
119 I should fix up another layer of binding, declared IGNORABLE, for
120 typed lambda list arguments.
123 The way that the compiler munges types with arguments together
124 with types with no arguments (in e.g. TYPE-EXPAND) leads to
125 weirdness visible to the user:
126 (DEFTYPE FOO () 'FIXNUM)
128 (TYPEP 11 '(FOO)) => T, which seems weird
129 (TYPEP 11 'FIXNUM) => T
130 (TYPEP 11 '(FIXNUM)) signals an error, as it should
131 The situation is complicated by the presence of Common Lisp types
132 like UNSIGNED-BYTE (which can either be used in list form or alone)
133 so I'm not 100% sure that the behavior above is actually illegal.
134 But I'm 90+% sure, and someday perhaps I'll be motivated to look it up..
137 It would be nice if the
139 (during macroexpansion)
140 said what macroexpansion was at fault, e.g.
142 (during macroexpansion of IN-PACKAGE,
143 during macroexpansion of DEFFOO)
146 The type system doesn't understand the KEYWORD type very well:
147 (SUBTYPEP 'KEYWORD 'SYMBOL) => NIL, NIL
148 It might be possible to fix this by changing the definition of
149 KEYWORD to (AND SYMBOL (SATISFIES KEYWORDP)), but the type system
150 would need to be a bit smarter about AND types, too:
151 (SUBTYPEP '(AND SYMBOL KEYWORD) 'SYMBOL) => NIL, NIL
152 (The type system does know something about AND types already,
153 (SUBTYPEP '(AND INTEGER FLOAT) 'NUMBER) => T, T
154 (SUBTYPEP '(AND INTEGER FIXNUM) 'NUMBER) =>T, T
155 so likely this is a small patch.)
158 Floating point infinities are screwed up. [When I was converting CMU CL
159 to SBCL, I was looking for complexity to delete, and I thought it was safe
160 to just delete support for floating point infinities. It wasn't: they're
161 generated by the floating point hardware even when we remove support
162 for them in software. -- WHN] Support for them should be restored.
165 The ANSI syntax for non-STANDARD method combination types in CLOS is
166 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
167 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
168 If you mess this up, omitting the PROGN qualifier in in DEFMETHOD,
169 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
170 (DEFMETHOD FOO ((X BAR)) (PRINT 'NUMBER))
171 the error mesage is not easy to understand:
172 INVALID-METHOD-ERROR was called outside the dynamic scope
173 of a method combination function (inside the body of
174 DEFINE-METHOD-COMBINATION or a method on the generic
175 function COMPUTE-EFFECTIVE-METHOD).
176 It would be better if it were more informative, a la
177 The method combination type for this method (STANDARD) does
178 not match the method combination type for the generic function
180 Also, after you make the mistake of omitting the PROGN qualifier
181 on a DEFMETHOD, doing a new DEFMETHOD with the correct qualifier
183 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
185 INVALID-METHOD-ERROR was called outside the dynamic scope
186 of a method combination function (inside the body of
187 DEFINE-METHOD-COMBINATION or a method on the generic
188 function COMPUTE-EFFECTIVE-METHOD).
189 This is not very helpful..
192 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
193 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
194 (Also, when this is fixed, we can enable the code in PROCLAIM which
195 checks for incompatible FTYPE redeclarations.)
198 The ANSI spec says that CONS can be a compound type spec, e.g.
199 (CONS FIXNUM REAL). SBCL doesn't support this.
202 from Paolo Amoroso on the CMU CL mailing list 27 Feb 2000:
203 I use CMU CL 18b under Linux. When COMPILE-FILE is supplied a physical
204 pathname, the type of the corresponding compiled file is X86F:
205 * (compile-file "/home/paolo/lisp/tools/foo")
206 Python version 1.0, VM version Intel x86 on 27 FEB 0 06:00:46 pm.
207 Compiling: /home/paolo/lisp/tools/foo.lisp 27 FEB 0 05:57:42 pm
209 Compiling DEFUN SQUARE:
210 Byte Compiling Top-Level Form:
211 /home/paolo/lisp/tools/foo.x86f written.
212 Compilation finished in 0:00:00.
213 #p"/home/paolo/lisp/tools/foo.x86f"
216 But when the function is called with a logical pathname, the file type
218 * (compile-file "tools:foo")
219 Python version 1.0, VM version Intel x86 on 27 FEB 0 06:01:04 pm.
220 Compiling: /home/paolo/lisp/tools/foo.lisp 27 FEB 0 05:57:42 pm
222 Compiling DEFUN SQUARE:
223 Byte Compiling Top-Level Form:
224 TOOLS:FOO.FASL written.
225 Compilation finished in 0:00:00.
226 #p"/home/paolo/lisp/tools/foo.fasl"
231 from DTC on the CMU CL mailing list 25 Feb 2000:
232 ;;; Compiler fails when this file is compiled.
234 ;;; Problem shows up in delete-block within ir1util.lisp. The assertion
235 ;;; (assert (member (functional-kind lambda) '(:let :mv-let :assignment)))
236 ;;; fails within bind node branch.
238 ;;; Note that if c::*check-consistency* is enabled then an un-reached
239 ;;; entry is also reported.
242 (declare (values nil))
259 (let ((ttt #'(lambda () (go cccc))))
260 (declare (special ttt))
261 (return-from bbbb nil))
264 (return-from bbbb nil))))))
267 (I *think* this is a bug. It certainly seems like strange behavior. But
268 the ANSI spec is scary, dark, and deep..)
269 (FORMAT NIL "~,1G" 1.4) => "1. "
270 (FORMAT NIL "~3,1G" 1.4) => "1. "
273 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
275 (setf (find-class 'ccc1) (find-class 'ccc))
276 (defmethod zut ((c ccc1)) 123)
277 DTC's recommended workaround from the mailing list 3 Mar 2000:
278 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
281 There's probably a bug in the compiler handling of special variables
282 in closures, inherited from the CMU CL code, as reported on the
283 CMU CL mailing list. There's a patch for this on the CMU CL
285 Message-ID: <38C8E188.A1E38B5E@jeack.com.au>
286 Date: Fri, 10 Mar 2000 22:50:32 +1100
287 From: "Douglas T. Crosher" <dtc@jeack.com.au>
290 The ANSI spec, in section "22.3.5.2 Tilde Less-Than-Sign: Logical Block",
291 says that an error is signalled if ~W, ~_, ~<...~:>, ~I, or ~:T is used
292 inside "~<..~>" (without the colon modifier on the closing syntax).
293 However, SBCL doesn't do this:
294 * (FORMAT T "~<munge~wegnum~>" 12)
299 When too many files are opened, OPEN will fail with an
300 uninformative error message
301 error in function OPEN: error opening #P"/tmp/foo.lisp": NIL
302 instead of saying that too many files are open.
305 Right now, when COMPILE-FILE has a read error, it actually pops
306 you into the debugger before giving up on the file. It should
307 instead handle the error, perhaps issuing (and handling)
308 a secondary error "caught ERROR: unrecoverable error during compilation"
309 and then return with FAILURE-P true,
312 from CMU CL mailing list 01 May 2000
314 I realize I can take care of this by doing (proclaim (ignore pcl::.slots1.))
315 but seeing as .slots0. is not-exported, shouldn't it be ignored within the
319 In: DEFMETHOD FOO-BAR-BAZ (RESOURCE-TYPE)
320 (DEFMETHOD FOO-BAR-BAZ
321 ((SELF RESOURCE-TYPE))
322 (SETF (SLOT-VALUE SELF 'NAME) 3))
323 --> BLOCK MACROLET PCL::FAST-LEXICAL-METHOD-FUNCTIONS
324 --> PCL::BIND-FAST-LEXICAL-METHOD-MACROS MACROLET
325 --> PCL::BIND-LEXICAL-METHOD-FUNCTIONS LET PCL::BIND-ARGS LET* PCL::PV-BINDING
326 --> PCL::PV-BINDING1 PCL::PV-ENV LET
328 (LET ((PCL::.SLOTS0. #))
333 Warning: Variable PCL::.SLOTS0. defined but never used.
335 Compilation unit finished.
338 #<Standard-Method FOO-BAR-BAZ (RESOURCE-TYPE) {480918FD}>
341 reported by Sam Steingold on the cmucl-imp mailing list 12 May 2000:
343 Also, there is another bug: `array-displacement' should return an array
344 or nil as first value (as per ANSI CL), while CMUCL declares it as
345 returning an array as first value always.
348 Sometimes (SB-EXT:QUIT) fails with
349 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
350 Process inferior-lisp exited abnormally with code 1
351 I haven't noticed a repeatable case of this yet.
354 The system accepts DECLAIM in most places where DECLARE would be
355 accepted, without even issuing a warning. ANSI allows this, but since
356 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
357 meaning is rather different, and it's unlikely that the user
358 has a good reason for doing DECLAIM not at top level, it would be
359 good to issue a STYLE-WARNING when this happens. A possible
360 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
361 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
364 some sort of bug in inlining and RETURN-FROM in sbcl-0.6.5: Compiling
367 (BLOCK USED-BY-SOME-Y?
370 (UNLESS (REJECTED? Y)
371 (RETURN-FROM USED-BY-SOME-Y? T)))))
372 (DECLARE (INLINE FROB))
377 error in function SB-KERNEL:ASSERT-ERROR:
378 The assertion (EQ (SB-C::CONTINUATION-KIND SB-C::CONT) :BLOCK-START) failed.
379 This is still present in sbcl-0.6.8.
382 The CMU CL reader code takes liberties in binding the standard read table
383 when reading the names of characters. Tim Moore posted a patch to the
384 CMU CL mailing list Mon, 22 May 2000 21:30:41 -0700.
387 In some cases the compiler believes type declarations on array
388 elements without checking them, e.g.
389 (DECLAIM (OPTIMIZE (SAFETY 3) (SPEED 1) (SPACE 1)))
392 (DECLARE (TYPE (SIMPLE-ARRAY CONS 1) X))
393 (WHEN (CONSP (AREF X 0))
395 (BAR (VECTOR (MAKE-FOO :A 11 :B 12)))
398 in SBCL 0.6.5 (and also in CMU CL 18b). This does not happen for
399 all cases, e.g. the type assumption *is* checked if the array
400 elements are declared to be of some structure type instead of CONS.
403 The printer doesn't report closures very well. This is true in
407 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
408 It would be nice to make closures have a settable name slot,
409 and make things like DEFSTRUCT and FLET, which create closures,
410 set helpful values into this slot.
413 And as long as we're wishing, it would be awfully nice if INSPECT could
414 also report on closures, telling about the values of the bound variables.
417 as reported by Robert Strandh on the CMU CL mailing list 12 Jun 2000:
419 (defconstant +a-constant+ (make-instance 'a-class))
420 (defconstant +another-constant+ (vector +a-constant+))
422 CMU Common Lisp release x86-linux 2.4.19 8 February 2000 build 456,
425 Send bug reports and questions to your local CMU CL maintainer,
426 or to pvaneynd@debian.org
427 or to cmucl-help@cons.org. (prefered)
428 type (help) for help, (quit) to exit, and (demo) to see the demos
430 Python 1.0, target Intel x86
431 CLOS based on PCL version: September 16 92 PCL (f)
432 * (defclass a-class () ())
433 #<STANDARD-CLASS A-CLASS {48027BD5}>
434 * (compile-file "xx.lisp")
435 Python version 1.0, VM version Intel x86 on 12 JUN 00 08:12:55 am.
437 /home/strandh/Research/Functional/Common-Lisp/CLIM/Development/McCLIM
438 /xx.lisp 12 JUN 00 07:47:14 am
439 Compiling Load Time Value of (PCL::GET-MAKE-INSTANCE-FUNCTION-SYMBOL
441 Byte Compiling Top-Level Form:
442 Error in function C::DUMP-STRUCTURE: Attempt to dump invalid
444 #<A-CLASS {4803A5B5}>
448 The compiler assumes that any time a function of declared FTYPE
449 doesn't signal an error, its arguments were of the declared type.
450 E.g. compiling and loading
451 (DECLAIM (OPTIMIZE (SAFETY 3)))
452 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
453 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE) FACTORIAL)))
455 (COND ((> (FACTORIAL X) 1.0E6)
456 (FORMAT T "too big~%"))
458 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
460 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
463 will cause the INTEGERP case to be selected, giving bogus output a la
465 This violates the "declarations are assertions" principle.
466 According to the ANSI spec, in the section "System Class FUNCTION",
467 this is a case of "lying to the compiler", but the lying is done
468 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
469 not by the unexpectedly general definition of FACTORIAL. In any case,
470 "declarations are assertions" means that lying to the compiler should
471 cause an error to be signalled, and should not cause a bogus
472 result to be returned. Thus, the compiler should not assume
473 that arbitrary functions check their argument types. (It might
474 make sense to add another flag (CHECKED?) to DEFKNOWN to
475 identify functions which *do* check their argument types.)
478 As pointed out by Martin Cracauer on the CMU CL mailing list
479 13 Jun 2000, the :FILE-LENGTH operation for
480 FD-STREAM-MISC-ROUTINE is broken for large files: it says
481 (THE INDEX SIZE) even though SIZE can be larger than INDEX.
484 In SBCL 0.6.5 (and CMU CL 18b) compiling and loading
485 (in-package :cl-user)
486 (declaim (optimize (safety 3)
488 (compilation-speed 2)
491 #+nil (sb-ext:inhibit-warnings 2)))
492 (declaim (ftype (function * (values)) emptyvalues))
493 (defun emptyvalues (&rest rest) (declare (ignore rest)) (values))
495 (defgeneric assertoid ((x t)))
496 (defmethod assertoid ((x t)) "just a placeholder")
498 (declare (type hash-table ht))
504 (assertoid (hash-table-count ht)))))))
505 (unless (typep res 'foo)
507 (common-lisp-user::bad-result-from-assertive-typed-fun
511 (bar (make-hash-table))
513 Error in KERNEL::UNDEFINED-SYMBOL-ERROR-HANDLER:
514 the function C::%INSTANCE-TYPEP is undefined.
515 %INSTANCE-TYPEP is always supposed to be IR1-transformed away, but for
516 some reason -- the (VALUES) return value declaration? -- the optimizer is
517 confused and compiles a full call to %INSTANCE-TYPEP (which doesn't exist
518 as a function) instead.
521 DEFMETHOD doesn't check the syntax of &REST argument lists properly,
522 accepting &REST even when it's not followed by an argument name:
523 (DEFMETHOD FOO ((X T) &REST) NIL)
526 On the CMU CL mailing list 26 June 2000, Douglas Crosher wrote
528 Hannu Rummukainen wrote:
530 > There's something weird going on with the compilation of the attached
531 > code. Compiling and loading the file in a fresh lisp, then invoking
533 Thanks for the bug report, nice to have this one fixed. It was a bug
534 in the x86 backend, the < VOP. A fix has been committed to the main
535 source, see the file compiler/x86/float.lisp.
537 Probably the same bug exists in SBCL.
540 TYPEP treats the result of UPGRADED-ARRAY-ELEMENT-TYPE as gospel,
541 so that (TYPEP (MAKE-ARRAY 3) '(VECTOR SOMETHING-NOT-DEFINED-YET))
542 returns (VALUES T T). Probably it should be an error instead,
543 complaining that the type SOMETHING-NOT-DEFINED-YET is not defined.
546 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
547 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
549 (DECLARE (TYPE INDEXOID X))
550 (THE (VALUES INDEXOID)
552 where the implementation of the type check in function FOO
553 includes a full call to %TYPEP. There are also some fundamental problems
554 with the interpretation of VALUES types (inherited from CMU CL, and
555 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
556 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
559 The definitions of SIGCONTEXT-FLOAT-REGISTER and
560 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
561 supported on FreeBSD because the floating point state is not saved,
562 but at least as of FreeBSD 4.0, the floating point state *is* saved,
563 so they could be supported after all. Very likely
564 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
567 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
568 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
569 issue with 'union, 'and, 'or etc. So even though according to the
570 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
571 (and now SBCL) interpret them as legal types.
574 ANSI specifies DEFINE-SYMBOL-MACRO, but it's not defined in SBCL.
575 CMU CL added it ca. Aug 13, 2000, after some discussion on the mailing
576 list, and it is probably possible to use substantially the same
577 patches to add it to SBCL.
580 a slew of floating-point-related errors reported by Peter Van Eynde
582 a: (SQRT -9.0) fails, because SB-KERNEL::COMPLEX-SQRT is undefined.
583 Similarly, COMPLEX-ASIN, COMPLEX-ACOS, COMPLEX-ACOSH, and others
585 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
586 should probably be 1.4012985e-45. In SBCL,
587 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
588 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
589 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
590 and LEAST-NEGATIVE-LONG-FLOAT.
591 c: Many expressions generate floating infinity:
596 PVE's regression tests want them to raise errors. SBCL
597 generates the infinities instead, which may or may not be
598 conforming behavior, but then blow it by being unable to
599 output the infinities, since support for infinities is generally
600 broken, and in particular SB-IMPL::OUTPUT-FLOAT-INFINITY is
602 d: (in section12.erg) various forms a la
603 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
604 don't give the right behavior.
607 type safety errors reported by Peter Van Eynde July 25, 2000:
608 a: (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
610 In general lengths of array type specifications aren't
611 checked by COERCE, so it fails when the spec is
612 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
613 b: CONCATENATE has the same problem of not checking the length
614 of specified output array types. MAKE-SEQUENCE and MAP and
615 MERGE also have the same problem.
616 c: (COERCE 'AND 'FUNCTION) returns something related to
617 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
618 d: ELT signals SIMPLE-ERROR if its index argument
619 isn't a valid index for its sequence argument, but should
620 signal TYPE-ERROR instead.
621 e: FILE-LENGTH is supposed to signal a type error when its
622 argument is not a stream associated with a file, but doesn't.
623 f: (FLOAT-RADIX 2/3) should signal an error instead of
625 g: (LOAD "*.lsp") should signal FILE-ERROR.
626 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
627 should signal TYPE-ERROR.
628 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
629 be used for input and output as needed. It should fail with
630 TYPE-ERROR when handed e.g. the results of
631 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
632 the inappropriate positions, but doesn't.
633 j: (PARSE-NAMESTRING (COERCE (LIST #\f #\o #\o (CODE-CHAR 0) #\4 #\8)
635 should probably signal an error instead of making a pathname with
637 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
638 not a binary input stream, but instead cheerfully reads from
639 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
642 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
643 a: (DEFCLASS FOO () (A B A)) should signal a PROGRAM-ERROR, and
645 b: (DEFCLASS FOO () (A B A) (:DEFAULT-INITARGS X A X B)) should
646 signal a PROGRAM-ERROR, and doesn't.
647 c: (DEFCLASS FOO07 NIL ((A :ALLOCATION :CLASS :ALLOCATION :CLASS))),
648 and other DEFCLASS forms with duplicate specifications in their
649 slots, should signal a PROGRAM-ERROR, and doesn't.
650 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
651 causes a COMPILER-ERROR.
654 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
655 a: (SYMBOL-MACROLET ((T TRUE)) ..) should probably signal
656 PROGRAM-ERROR, but SBCL accepts it instead.
657 b: SYMBOL-MACROLET should refuse to bind something which is
658 declared as a global variable, signalling PROGRAM-ERROR.
659 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
660 it binds is declared SPECIAL inside.
663 LOOP bugs reported by Peter Van Eynde July 25, 2000:
664 a: (LOOP WITH (A B) DO (PRINT 1)) is a syntax error according to
665 the definition of WITH clauses given in the ANSI spec, but
666 compiles and runs happily in SBCL.
667 b: a messy one involving package iteration:
668 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<))))
669 Should be: (("blah" "blah2") ("blah2"))
670 SBCL: (("blah") ("blah2"))
671 * (LET ((X 1)) (LOOP FOR I BY (INCF X) FROM X TO 10 COLLECT I))
672 doesn't work -- SBCL's LOOP says BY isn't allowed in a FOR clause.
675 type system errors reported by Peter Van Eynde July 25, 2000:
676 a: (SUBTYPEP 'BIGNUM 'INTEGER) => NIL, NIL
677 but should be (VALUES T T) instead.
678 b: (SUBTYPEP 'EXTENDED-CHAR 'CHARACTER) => NIL, NIL
679 but should be (VALUES T T) instead.
680 c: (SUBTYPEP '(INTEGER (0) (0)) 'NIL) dies with nested errors.
681 d: In general, the system doesn't like '(INTEGER (0) (0)) -- it
682 blows up at the level of SPECIFIER-TYPE with
683 "Lower bound (0) is greater than upper bound (0)." Probably
684 SPECIFIER-TYPE should return NIL instead.
685 e: (TYPEP 0 '(COMPLEX (EQL 0)) fails with
686 "Component type for Complex is not numeric: (EQL 0)."
687 This might be easy to fix; the type system already knows
688 that (SUBTYPEP '(EQL 0) 'NUMBER) is true.
689 f: The type system doesn't know about the condition system,
690 so that e.g. (TYPEP 'SIMPLE-ERROR 'ERROR)=>NIL.
691 g: The type system isn't all that smart about relationships
692 between hairy types, as shown in the type.erg test results,
693 e.g. (SUBTYPEP 'CONS '(NOT ATOM)) => NIL, NIL.
696 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
698 (DEFGENERIC FOO02 (X))
699 (DEFMETHOD FOO02 ((X NUMBER)) T)
700 (LET ((M (FIND-METHOD (FUNCTION FOO02)
702 (LIST (FIND-CLASS (QUOTE NUMBER))))))
703 (REMOVE-METHOD (FUNCTION FOO02) M)
704 (DEFGENERIC FOO03 (X))
705 (ADD-METHOD (FUNCTION FOO03) M)))
706 should give an error, but SBCL allows it.
707 b: READ should probably return READER-ERROR, not the bare
708 arithmetic error, when input a la "1/0" or "1e1000" causes
710 c: (BUTLAST NIL) should return NIL. (This appears to be a compiler
711 bug, since the definition of BUTLAST, when interpreted, does
712 give (BUTLAST NIL)=>NIL.)
715 It has been reported (e.g. by Peter Van Eynde) that there are
716 several metaobject protocol "errors". (In order to fix them, we might
717 need to document exactly what metaobject protocol specification
718 we're following -- the current code is just inherited from PCL.)
721 another error from Peter Van Eynde 5 September 2000:
722 (FORMAT NIL "~F" "FOO") should work, but instead reports an error.
723 PVE submitted a patch to deal with this bug, but it exposes other
724 comparably serious bugs, so I didn't apply it. It looks as though
725 the FORMAT code needs a fair amount of rewriting in order to comply
726 with the various details of the ANSI spec.
729 The implementation of #'+ returns its single argument without
730 type checking, e.g. (+ "illegal") => "illegal".
733 In sbcl-0.6.7, there is no doc string for CL:PUSH, probably
734 because it's defined with the DEFMACRO-MUNDANELY macro and something
735 is wrong with doc string setting in that macro.
738 Attempting to use COMPILE on something defined by DEFMACRO fails:
739 (DEFMACRO FOO (X) (CONS X X))
741 Error in function C::GET-LAMBDA-TO-COMPILE:
742 #<Closure Over Function "DEFUN (SETF MACRO-FUNCTION)" {480E21B1}> was defined in a non-null environment.
745 In sbcl-0.6.7, the compiler accepted a bogus declaration
746 (TYPE INDEX LENGTH) in the definition of BUTLAST, and then died
747 with infinite regress of errors when the BUTLAST function was
748 executed with a LIST=NIL which would cause LENGTH to be -1.
749 I fixed the bogus declaration, but I should come back and see
750 whether the system's inability to recover from the bogus declaration
751 (by signalling a TYPE-ERROR and dropping into the debugger) was
752 a compiler problem which remains to be fixed, or one of the
753 unrelated infinite-regress-errors problems, many related to
754 revised signal handling, which were fixed around the same time.
757 (SUBTYPEP '(AND ZILCH INTEGER) 'ZILCH)
761 CL:*DEFAULT-PATHNAME-DEFAULTS* doesn't behave as ANSI suggests (reflecting
762 current working directory). And there's no supported way to update
763 or query the current working directory (a la Unix "chdir" and "pwd"),
764 which is functionality that ILISP needs (and currently gets with low-level
768 The debugger LIST-LOCATIONS command doesn't work properly.