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, and (3) by their own reasoning, it looks as though
54 ANSI may have gotten it wrong. ANSI justifies this specification
56 The restriction against issuing a warning for type mismatches
57 between a slot-initform and the corresponding slot's :TYPE
58 option is necessary because a slot-initform must be specified
59 in order to specify slot options; in some cases, no suitable
61 However, in SBCL (as in CMU CL or, for that matter, any compiler
62 which really understands Common Lisp types) a suitable default
63 does exist, in all cases, because the compiler understands the
64 concept of functions which never return (i.e. has return type NIL).
65 Thus, as a portable workaround, you can use a call to some
66 known-never-to-return function as the default. E.g.
68 (BAR (ERROR "missing :BAR argument")
69 :TYPE SOME-TYPE-TOO-HAIRY-TO-CONSTRUCT-AN-INSTANCE-OF))
71 (DECLAIM (FTYPE (FUNCTION () NIL) MISSING-ARG))
72 (DEFUN REQUIRED-ARG () ; workaround for SBCL non-ANSI slot init typing
73 (ERROR "missing required argument"))
75 (BAR (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
76 (BLETCH (REQUIRED-ARG) :TYPE TRICKY-TYPE-OF-SOME-SORT)
77 (N-REFS-SO-FAR 0 :TYPE (INTEGER 0)))
78 Such code should compile without complaint and work correctly either
79 on SBCL or on any other completely compliant Common Lisp system.
82 bogus warnings about undefined functions for magic functions like
83 SB!C::%%DEFUN and SB!C::%DEFCONSTANT when cross-compiling files
84 like src/code/float.lisp. Fixing this will probably require
85 straightening out enough bootstrap consistency issues that
86 the cross-compiler can run with *TYPE-SYSTEM-INITIALIZED*.
87 Instead, the cross-compiler runs in a slightly flaky state
88 which is sane enough to compile SBCL itself, but which is
89 also unstable in several ways, including its inability
90 to really grok function declarations.
92 As of sbcl-0.7.5, sbcl's cross-compiler does run with
93 *TYPE-SYSTEM-INITIALIZED*; however, this bug remains.
96 The "compiling top-level form:" output ought to be condensed.
97 Perhaps any number of such consecutive lines ought to turn into a
98 single "compiling top-level forms:" line.
101 The way that the compiler munges types with arguments together
102 with types with no arguments (in e.g. TYPE-EXPAND) leads to
103 weirdness visible to the user:
104 (DEFTYPE FOO () 'FIXNUM)
106 (TYPEP 11 '(FOO)) => T, which seems weird
107 (TYPEP 11 'FIXNUM) => T
108 (TYPEP 11 '(FIXNUM)) signals an error, as it should
109 The situation is complicated by the presence of Common Lisp types
110 like UNSIGNED-BYTE (which can either be used in list form or alone)
111 so I'm not 100% sure that the behavior above is actually illegal.
112 But I'm 90+% sure, and the following related behavior,
114 treating the bare symbol AND as equivalent to '(AND), is specifically
115 forbidden (by the ANSI specification of the AND type).
118 It would be nice if the
120 (during macroexpansion)
121 said what macroexpansion was at fault, e.g.
123 (during macroexpansion of IN-PACKAGE,
124 during macroexpansion of DEFFOO)
127 (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
128 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
129 (Also, when this is fixed, we can enable the code in PROCLAIM which
130 checks for incompatible FTYPE redeclarations.)
133 (I *think* this is a bug. It certainly seems like strange behavior. But
134 the ANSI spec is scary, dark, and deep.. -- WHN)
135 (FORMAT NIL "~,1G" 1.4) => "1. "
136 (FORMAT NIL "~3,1G" 1.4) => "1. "
139 from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
141 (setf (find-class 'ccc1) (find-class 'ccc))
142 (defmethod zut ((c ccc1)) 123)
143 In sbcl-0.7.1.13, this gives an error,
144 There is no class named CCC1.
145 DTC's recommended workaround from the mailing list 3 Mar 2000:
146 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
149 Sometimes (SB-EXT:QUIT) fails with
150 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
151 Process inferior-lisp exited abnormally with code 1
152 I haven't noticed a repeatable case of this yet.
155 The printer doesn't report closures very well. This is true in
159 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
160 It would be nice to make closures have a settable name slot,
161 and make things like DEFSTRUCT and FLET, which create closures,
162 set helpful values into this slot.
165 And as long as we're wishing, it would be awfully nice if INSPECT could
166 also report on closures, telling about the values of the bound variables.
169 The compiler assumes that any time a function of declared FTYPE
170 doesn't signal an error, its arguments were of the declared type.
171 E.g. compiling and loading
172 (DECLAIM (OPTIMIZE (SAFETY 3)))
173 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
175 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE)) FACTORIAL))
177 (COND ((> (FACTORIAL X) 1.0E6)
178 (FORMAT T "too big~%"))
180 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
182 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
185 will cause the INTEGERP case to be selected, giving bogus output a la
187 (or (FOO 1000.5), "exactly 1001.5")
188 This violates the "declarations are assertions" principle.
189 According to the ANSI spec, in the section "System Class FUNCTION",
190 this is a case of "lying to the compiler", but the lying is done
191 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
192 not by the unexpectedly general definition of FACTORIAL. In any case,
193 "declarations are assertions" means that lying to the compiler should
194 cause an error to be signalled, and should not cause a bogus
195 result to be returned. Thus, the compiler should not assume
196 that arbitrary functions check their argument types. (It might
197 make sense to add another flag (CHECKED?) to DEFKNOWN to
198 identify functions which *do* check their argument types.)
199 (Also, verify that the compiler handles declared function
200 return types as assertions.)
203 TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
204 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
206 (DECLARE (TYPE INDEXOID X))
207 (THE (VALUES INDEXOID)
209 where the implementation of the type check in function FOO
210 includes a full call to %TYPEP. There are also some fundamental problems
211 with the interpretation of VALUES types (inherited from CMU CL, and
212 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
213 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
216 The definitions of SIGCONTEXT-FLOAT-REGISTER and
217 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
218 supported on FreeBSD because the floating point state is not saved,
219 but at least as of FreeBSD 4.0, the floating point state *is* saved,
220 so they could be supported after all. Very likely
221 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
224 (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
225 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same
226 issue with 'union, 'and, 'or etc. So even though according to the
227 ANSI spec, bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL
228 (and now SBCL) interpret them as legal types.
231 a slew of floating-point-related errors reported by Peter Van Eynde
233 b: SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
234 should probably be 1.4012985e-45. In SBCL,
235 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
236 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
237 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
238 and LEAST-NEGATIVE-LONG-FLOAT.
239 c: Many expressions generate floating infinity on x86/Linux:
244 PVE's regression tests want them to raise errors. sbcl-0.7.0.5
245 on x86/Linux generates the infinities instead. That might or
246 might not be conforming behavior, but it's also inconsistent,
247 which is almost certainly wrong. (Inconsistency: (/ 1 0.0)
248 should give the same result as (/ 1.0 0.0), but instead (/ 1 0.0)
249 generates SINGLE-FLOAT-POSITIVE-INFINITY and (/ 1.0 0.0)
251 d: (in section12.erg) various forms a la
252 (FLOAT 1 DOUBLE-FLOAT-EPSILON)
253 don't give the right behavior.
256 type safety errors reported by Peter Van Eynde July 25, 2000:
257 c: (COERCE 'AND 'FUNCTION) returns something related to
258 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
259 h: (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
260 should signal TYPE-ERROR.
261 i: MAKE-TWO-WAY-STREAM doesn't check that its arguments can
262 be used for input and output as needed. It should fail with
263 TYPE-ERROR when handed e.g. the results of
264 MAKE-STRING-INPUT-STREAM or MAKE-STRING-OUTPUT-STREAM in
265 the inappropriate positions, but doesn't.
266 k: READ-BYTE is supposed to signal TYPE-ERROR when its argument is
267 not a binary input stream, but instead cheerfully reads from
268 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
271 DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
272 d: (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
273 causes a COMPILER-ERROR.
276 SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
277 c: SYMBOL-MACROLET should signal PROGRAM-ERROR if something
278 it binds is declared SPECIAL inside.
281 miscellaneous errors reported by Peter Van Eynde July 25, 2000:
283 (DEFGENERIC FOO02 (X))
284 (DEFMETHOD FOO02 ((X NUMBER)) T)
285 (LET ((M (FIND-METHOD (FUNCTION FOO02)
287 (LIST (FIND-CLASS (QUOTE NUMBER))))))
288 (REMOVE-METHOD (FUNCTION FOO02) M)
289 (DEFGENERIC FOO03 (X))
290 (ADD-METHOD (FUNCTION FOO03) M)))
291 should give an error, but SBCL allows it.
294 It has been reported (e.g. by Peter Van Eynde) that there are
295 several metaobject protocol "errors". (In order to fix them, we might
296 need to document exactly what metaobject protocol specification
297 we're following -- the current code is just inherited from PCL.)
300 The implementation of #'+ returns its single argument without
301 type checking, e.g. (+ "illegal") => "illegal".
304 The debugger LIST-LOCATIONS command doesn't work properly.
307 Compiling and loading
308 (DEFUN FAIL (X) (THROW 'FAIL-TAG X))
310 then requesting a BACKTRACE at the debugger prompt gives no information
311 about where in the user program the problem occurred.
314 The compiler is supposed to do type inference well enough that
317 ((SIMPLE-ARRAY SINGLE-FLOAT)
319 (DECLARE (TYPE (SIMPLE-ARRAY SINGLE-FLOAT) X))
322 is redundant. However, as reported by Juan Jose Garcia Ripoll for
323 CMU CL, it sometimes doesn't. Adding declarations is a pretty good
324 workaround for the problem for now, but can't be done by the TYPECASE
325 macros themselves, since it's too hard for the macro to detect
326 assignments to the variable within the clause.
327 Note: The compiler *is* smart enough to do the type inference in
328 many cases. This case, derived from a couple of MACROEXPAND-1
329 calls on Ripoll's original test case,
331 (DECLARE (OPTIMIZE SPEED (SAFETY 0)))
332 (COND ((TYPEP A '(SIMPLE-ARRAY SINGLE-FLOAT)) NIL
333 (LET ((LENGTH (ARRAY-TOTAL-SIZE A)))
334 (LET ((I 0) (G2554 LENGTH))
335 (DECLARE (TYPE REAL G2554) (TYPE REAL I))
338 (WHEN (>= I G2554) (GO SB-LOOP::END-LOOP))
339 (SETF (ROW-MAJOR-AREF A I) (- (ROW-MAJOR-AREF A I)))
340 (GO SB-LOOP::NEXT-LOOP)
341 SB-LOOP::END-LOOP))))))
342 demonstrates the problem; but the problem goes away if the TAGBODY
343 and GO forms are removed (leaving the SETF in ordinary, non-looping
344 code), or if the TAGBODY and GO forms are retained, but the
345 assigned value becomes 0.0 instead of (- (ROW-MAJOR-AREF A I)).
348 Paul Werkowski wrote on cmucl-imp@cons.org 2000-11-15
349 I am looking into this problem that showed up on the cmucl-help
350 list. It seems to me that the "implementation specific environment
351 hacking functions" found in pcl/walker.lisp are completely messed
352 up. The good thing is that they appear to be barely used within
353 PCL and the munged environment object is passed to cmucl only
354 in calls to macroexpand-1, which is probably why this case fails.
355 SBCL uses essentially the same code, so if the environment hacking
356 is screwed up, it affects us too.
359 Using the pretty-printer from the command prompt gives funny
360 results, apparently because the pretty-printer doesn't know
361 about user's command input, including the user's carriage return
362 that the user, and therefore the pretty-printer thinks that
363 the new output block should start indented 2 or more characters
364 rightward of the correct location.
367 As reported by Winton Davies on a CMU CL mailing list 2000-01-10,
368 and reported for SBCL by Martin Atzmueller 2000-10-20: (TRACE GETHASH)
369 crashes SBCL. In general tracing anything which is used in the
370 implementation of TRACE is likely to have the same problem.
373 (DECLAIM (OPTIMIZE ..)) doesn't work properly inside LOCALLY forms.
376 As reported by Martin Atzmueller on sbcl-devel 26 Dec 2000,
377 ANSI says that WITH-OUTPUT-TO-STRING should have a keyword
378 :ELEMENT-TYPE, but in sbcl-0.6.9 this is not defined for
379 WITH-OUTPUT-TO-STRING.
382 ANSI says in one place that type declarations can be abbreviated even
383 when the type name is not a symbol, e.g.
384 (DECLAIM ((VECTOR T) *FOOVECTOR*))
385 SBCL doesn't support this. But ANSI says in another place that this
386 isn't allowed. So it's not clear this is a bug after all. (See the
387 e-mail on cmucl-help@cons.org on 2001-01-16 and 2001-01-17 from WHN
391 as pointed out by Dan Barlow on sbcl-devel 2000-07-02:
392 The PICK-TEMPORARY-FILE-NAME utility used by LOAD-FOREIGN uses
393 an easily guessable temporary filename in a way which might open
394 applications using LOAD-FOREIGN to hijacking by malicious users
395 on the same machine. Incantations for doing this safely are
396 floating around the net in various "how to write secure programs
397 despite Unix" documents, and it would be good to (1) fix this in
398 LOAD-FOREIGN, and (2) hunt for any other code which uses temporary
399 files and make it share the same new safe logic.
402 Functions are assigned names based on the context in which they're
403 defined. This is less than ideal for the functions which are
404 used to implement CLOS methods. E.g. the output of
405 (DESCRIBE 'PRINT-OBJECT) lists functions like
406 #<FUNCTION "DEF!STRUCT (TRACE-INFO (:MAKE-LOAD-FORM-FUN SB-KERNEL:JUST-DUMP-IT-NORMALLY) (:PRINT-OBJECT #))" {1020E49}>
408 #<FUNCTION "MACROLET ((FORCE-DELAYED-DEF!METHODS NIL #))" {1242871}>
409 It would be better if these functions' names always identified
410 them as methods, and identified their generic functions and
414 RANDOM-INTEGER-EXTRA-BITS=10 may not be large enough for the RANDOM
415 RNG to be high quality near RANDOM-FIXNUM-MAX; it looks as though
416 the mean of the distribution can be systematically O(0.1%) wrong.
417 Just increasing R-I-E-B is probably not a good solution, since
418 it would decrease efficiency more than is probably necessary. Perhaps
419 using some sort of accept/reject method would be better.
422 Internally the compiler sometimes evaluates
423 (sb-kernel:type/= (specifier-type '*) (specifier-type t))
424 (I stumbled across this when I added an
425 (assert (not (eq type1 *wild-type*)))
426 in the NAMED :SIMPLE-= type method.) '* isn't really a type, and
427 in a type context should probably be translated to T, and so it's
428 probably wrong to ask whether it's equal to the T type and then (using
429 the EQ type comparison in the NAMED :SIMPLE-= type method) return NIL.
430 (I haven't tried to investigate this bug enough to guess whether
431 there might be any user-level symptoms.)
434 Inconsistencies between derived and declared VALUES return types for
435 DEFUN aren't checked very well. E.g. the logic which successfully
436 catches problems like
437 (declaim (ftype (function (fixnum) float) foo))
439 (declare (type integer x))
440 (values x)) ; wrong return type, detected, gives warning, good!
442 (declaim (ftype (function (t) (values t t)) bar))
444 (values x)) ; wrong number of return values, no warning, bad!
445 The cause of this is seems to be that (1) the internal function
446 VALUES-TYPES-EQUAL-OR-INTERSECT used to make the check handles its
447 arguments symmetrically, and (2) when the type checking code was
448 written back when when SBCL's code was still CMU CL, the intent
450 (declaim (ftype (function (t) t) bar))
452 (values x x)) ; wrong number of return values; should give warning?
453 not be warned for, because a two-valued return value is considered
454 to be compatible with callers who expects a single value to be
455 returned. That intent is probably not appropriate for modern ANSI
456 Common Lisp, but fixing this might be complicated because of other
457 divergences between auld-style and new-style handling of
458 multiple-VALUES types. (Some issues related to this were discussed
459 on cmucl-imp at some length sometime in 2000.)
462 The facility for dumping a running Lisp image to disk gets confused
463 when run without the PURIFY option, and creates an unnecessarily large
464 core file (apparently representing memory usage up to the previous
465 high-water mark). Moreover, when the file is loaded, it confuses the
466 GC, so that thereafter memory usage can never be reduced below that
470 In sbcl-0.6.11.41 (and in all earlier SBCL, and in CMU
471 CL), out-of-line structure slot setters are horribly inefficient
472 whenever the type of the slot is declared, because out-of-line
473 structure slot setters are implemented as closures to save space,
474 so the compiler doesn't compile the type test into code, but
475 instead just saves the type in a lexical closure and interprets it
477 A proper solution involves deciding whether it's really worth
478 saving space by implementing structure slot accessors as closures.
479 (If it's not worth it, the problem vanishes automatically. If it
480 is worth it, there are hacks we could use to force type tests to
481 be compiled anyway, and even shared. E.g. we could implement
482 an EQUAL hash table mapping from types to compiled type tests,
483 and save the appropriate compiled type test as part of each lexical
484 closure; or we could make the lexical closures be placeholders
485 which overwrite their old definition as a lexical closure with
486 a new compiled definition the first time that they're called.)
487 As a workaround for the problem, #'(SETF FOO) expressions can
488 be replaced with (EFFICIENT-SETF-FUNCTION FOO), where
489 (defmacro efficient-setf-function (place-function-name)
490 (or #+sbcl (and (sb-impl::info :function :accessor-for place-function-name)
491 ;; a workaround for the problem, encouraging the
492 ;; inline expansion of the structure accessor, so
493 ;; that the compiler can optimize its type test
494 (let ((new-value (gensym "NEW-VALUE-"))
495 (structure-value (gensym "STRUCTURE-VALUE-")))
496 `(lambda (,new-value ,structure-value)
497 (setf (,place-function-name ,structure-value)
499 ;; no problem, can just use the ordinary expansion
500 `(function (setf ,place-function-name))))
503 There's apparently a bug in CEILING optimization which caused
504 Douglas Crosher to patch the CMU CL version. Martin Atzmueller
505 applied the patches to SBCL and they didn't seem to cause problems
506 (as reported sbcl-devel 2001-05-04). However, since the patches
507 modify nontrivial code which was apparently written incorrectly
508 the first time around, until regression tests are written I'm not
509 comfortable merging the patches in the CVS version of SBCL.
512 (DESCRIBE 'SB-ALIEN:DEF-ALIEN-TYPE) reports the macro argument list
516 in #<PACKAGE "SB-ALIEN">.
517 Macro-function: #<FUNCTION "DEF!MACRO DEF-ALIEN-TYPE" {19F4A39}>
518 Macro arguments: (#:whole-470 #:environment-471)
519 On Sat, May 26, 2001 09:45:57 AM CDT it was compiled from:
520 /usr/stuff/sbcl/src/code/host-alieneval.lisp
521 Created: Monday, March 12, 2001 07:47:43 AM CST
524 (TIME (ROOM T)) reports more than 200 Mbytes consed even for
525 a clean, just-started SBCL system. And it seems to be right:
526 (ROOM T) can bring a small computer to its knees for a *long*
527 time trying to GC afterwards. Surely there's some more economical
528 way to implement (ROOM T).
531 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
533 ;;; The compiler is flushing the argument type test, and the default
534 ;;; case in the cond, so that calling with say a fixnum 0 causes a
536 (declaim (optimize (safety 2) (speed 3)))
538 (declare (type (or string stream) x))
539 (cond ((typep x 'string) 'string)
540 ((typep x 'stream) 'stream)
543 The symptom in sbcl-0.6.12.42 on OpenBSD is actually (TST 0)=>STREAM
544 (not the SIGBUS reported in the comment) but that's broken too;
545 type declarations are supposed to be treated as assertions unless
546 SAFETY 0, so we should be getting a TYPE-ERROR.
549 reported by Martin Atzmueller 2001-06-25; originally from CMU CL bugs
551 (in-package :cl-user)
552 ;;; The following invokes a compiler error.
553 (declaim (optimize (speed 2) (debug 3)))
556 (unwind-protect nil)))
560 The error message in sbcl-0.6.12.42 is
561 internal error, failed AVER:
562 "(COMMON-LISP:EQ (SB!C::TN-ENVIRONMENT SB!C:TN) SB!C::TN-ENV)"
565 When the compiler inline expands functions, it may be that different
566 kinds of return values are generated from different code branches.
567 E.g. an inline expansion of POSITION generates integer results
568 from one branch, and NIL results from another. When that inline
569 expansion is used in a context where only one of those results
572 (aref *a1* (position x *a2*)))
573 and the compiler can't prove that the unacceptable branch is
574 never taken, then bogus type mismatch warnings can be generated.
575 If you need to suppress the type mismatch warnings, you can
576 suppress the inline expansion,
578 #+sbcl (declare (notinline position)) ; to suppress bug 117 bogowarnings
579 (aref *a1* (position x *a2*)))
580 or, sometimes, suppress them by declaring the result to be of an
583 (aref *a1* (the integer (position x *a2*))))
585 This is not a new compiler problem in 0.7.0, but the new compiler
586 transforms for FIND, POSITION, FIND-IF, and POSITION-IF make it
587 more conspicuous. If you don't need performance from these functions,
588 and the bogus warnings are a nuisance for you, you can return to
589 your pre-0.7.0 state of grace with
590 #+sbcl (declaim (notinline find position find-if position-if)) ; bug 117..
593 as reported by Eric Marsden on cmucl-imp@cons.org 2001-08-14:
594 (= (FLOAT 1 DOUBLE-FLOAT-EPSILON)
595 (+ (FLOAT 1 DOUBLE-FLOAT-EPSILON) DOUBLE-FLOAT-EPSILON)) => T
596 when of course it should be NIL. (He says it only fails for X86,
597 not SPARC; dunno about Alpha.)
599 Also, "the same problem exists for LONG-FLOAT-EPSILON,
600 DOUBLE-FLOAT-NEGATIVE-EPSILON, LONG-FLOAT-NEGATIVE-EPSILON (though
601 for the -negative- the + is replaced by a - in the test)."
603 Raymond Toy comments that this is tricky on the X86 since its FPU
604 uses 80-bit precision internally.
607 Even in sbcl-0.pre7.x, which is supposed to be free of the old
608 non-ANSI behavior of treating the function return type inferred
609 from the current function definition as a declaration of the
610 return type from any function of that name, the return type of NIL
611 is attached to FOO in 120a above, and used to optimize code which
615 There was some sort of screwup in handling of
616 (IF (NOT (IGNORE-ERRORS ..))). E.g.
618 (if (not (ignore-errors
619 (make-pathname :host "foo"
623 (error "notunlessnot")))
624 The (NOT (IGNORE-ERRORS ..)) form evaluates to T, so this should be
625 printing "ok", but instead it's going to the ERROR. This problem
626 seems to've been introduced by MNA's HANDLER-CASE patch (sbcl-devel
627 2001-07-17) and as a workaround (put in sbcl-0.pre7.14.flaky4.12)
628 I reverted back to the old weird HANDLER-CASE code. However, I
629 think the problem looks like a compiler bug in handling RETURN-FROM,
630 so I left the MNA-patched code in HANDLER-CASE (suppressed with
631 #+NIL) and I'd like to go back to see whether this really is
632 a compiler bug before I delete this BUGS entry.
635 As of version 0.pre7.14, SBCL's implementation of MACROLET makes
636 the entire lexical environment at the point of MACROLET available
637 in the bodies of the macroexpander functions. In particular, it
638 allows the function bodies (which run at compile time) to try to
639 access lexical variables (which are only defined at runtime).
640 It doesn't even issue a warning, which is bad.
642 The SBCL behavior arguably conforms to the ANSI spec (since the
643 spec says that the behavior is undefined, ergo anything conforms).
644 However, it would be better to issue a compile-time error.
645 Unfortunately I (WHN) don't see any simple way to detect this
646 condition in order to issue such an error, so for the meantime
647 SBCL just does this weird broken "conforming" thing.
649 The ANSI standard says, in the definition of the special operator
651 The macro-expansion functions defined by MACROLET are defined
652 in the lexical environment in which the MACROLET form appears.
653 Declarations and MACROLET and SYMBOL-MACROLET definitions affect
654 the local macro definitions in a MACROLET, but the consequences
655 are undefined if the local macro definitions reference any
656 local variable or function bindings that are visible in that
658 Then it seems to contradict itself by giving the example
660 (macrolet ((fudge (z)
661 ;The parameters x and flag are not accessible
662 ; at this point; a reference to flag would be to
663 ; the global variable of that name.
664 ` (if flag (* ,z ,z) ,z)))
665 ;The parameters x and flag are accessible here.
669 The comment "a reference to flag would be to the global variable
670 of the same name" sounds like good behavior for the system to have.
671 but actual specification quoted above says that the actual behavior
675 (as reported by Gabe Garza on cmucl-help 2001-09-21)
677 (defun test-pred (x y)
681 (func (lambda () x)))
682 (print (eq func func))
683 (print (test-pred func func))
684 (delete func (list func))))
685 Now calling (TEST-CASE) gives output
688 (#<FUNCTION {500A9EF9}>)
689 Evidently Python thinks of the lambda as a code transformation so
690 much that it forgets that it's also an object.
693 The DEFSTRUCT section of the ANSI spec, in the :CONC-NAME section,
694 specifies a precedence rule for name collisions between slot accessors of
695 structure classes related by inheritance. As of 0.7.0, SBCL still
699 insufficient syntax checking in MACROLET:
701 (macrolet ((defmacro bar (z) `(+ z z)))
703 shouldn't compile without error (because of the extra DEFMACRO symbol).
706 Ideally, uninterning a symbol would allow it, and its associated
707 FDEFINITION and PROCLAIM data, to be reclaimed by the GC. However,
708 at least as of sbcl-0.7.0, this isn't the case. Information about
709 FDEFINITIONs and PROCLAIMed properties is stored in globaldb.lisp
710 essentially in ordinary (non-weak) hash tables keyed by symbols.
711 Thus, once a system has an entry in this system, it tends to live
712 forever, even when it is uninterned and all other references to it
716 (reported by Arnaud Rouanet on cmucl-imp 2001-12-18)
717 (defmethod foo ((x integer))
719 (defmethod foo :around ((x integer))
722 Now (FOO 3) should return 3, but instead it returns 4.
725 (reported by Alexey Dejneka sbcl-devel 2002-01-03)
727 SUBTYPEP does not work well with redefined classes:
736 * (defclass b (a) ())
749 This bug was fixed in sbcl-0.7.4.1 by invalidating the PCL wrapper
750 class upon redefinition. Unfortunately, doing so causes bug #176 to
751 appear. Pending further investication, one or other of these bugs
752 might be present at any given time.
755 Pretty-printing nested backquotes doesn't work right, as
756 reported by Alexey Dejneka sbcl-devel 2002-01-13:
758 ``(FOO SB-IMPL::BACKQ-COMMA-AT S)
759 * (lisp-implementation-version)
763 (as reported by Lynn Quam on cmucl-imp ca. 2002-01-16)
764 %NATURALIZE-C-STRING conses a lot, like 16 bytes per byte
765 of the naturalized string. We could probably port the patches
766 from the cmucl-imp mailing list.
769 (reported by Jesse Bouwman 2001-10-24 through the unfortunately
770 prominent SourceForge web/db bug tracking system, which is
771 unfortunately not a reliable way to get a timely response from
772 the SBCL maintainers)
773 In the course of trying to build a test case for an
774 application error, I encountered this behavior:
775 If you start up sbcl, and then lay on CTRL-C for a
776 minute or two, the lisp process will eventually say:
777 %PRIMITIVE HALT called; the party is over.
778 and throw you into the monitor. If I start up lisp,
779 attach to the process with strace, and then do the same
780 (abusive) thing, I get instead:
781 access failure in heap page not marked as write-protected
782 and the monitor again. I don't know enough to have the
783 faintest idea of what is going on here.
784 This is with sbcl 6.12, uname -a reports:
785 Linux prep 2.2.19 #4 SMP Tue Apr 24 13:59:52 CDT 2001 i686 unknown
786 I (WHN) have verified that the same thing occurs on sbcl-0.pre7.141
787 under OpenBSD 2.9 on my X86 laptop. Do be patient when you try it:
788 it took more than two minutes (but less than five) for me.
791 (This was once known as IR1-4, but it lived on even after the
792 IR1 interpreter went to the big bit bucket in the sky.)
793 The system accepts DECLAIM in most places where DECLARE would be
794 accepted, without even issuing a warning. ANSI allows this, but since
795 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
796 meaning is rather different, and it's unlikely that the user
797 has a good reason for doing DECLAIM not at top level, it would be
798 good to issue a STYLE-WARNING when this happens. A possible
799 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
800 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
801 [This is considered an IR1-interpreter-related bug because until
802 EVAL-WHEN is rewritten, which won't happen until after the IR1
803 interpreter is gone, the system's notion of what's a top-level form
804 and what's not will remain too confused to fix this problem.]
807 ANSI allows types `(COMPLEX ,FOO) to use very hairy values for
808 FOO, e.g. (COMPLEX (AND REAL (SATISFIES ODDP))). The old CMU CL
809 COMPLEX implementation didn't deal with this, and hasn't been
810 upgraded to do so. (This doesn't seem to be a high priority
811 conformance problem, since seems hard to construct useful code
815 Floating point errors are reported poorly. E.g. on x86 OpenBSD
818 debugger invoked on condition of type SB-KERNEL:FLOATING-POINT-EXCEPTION:
819 An arithmetic error SB-KERNEL:FLOATING-POINT-EXCEPTION was signalled.
820 No traps are enabled? How can this be?
821 It should be possible to be much more specific (overflow, division
822 by zero, etc.) and of course the "How can this be?" should be fixable.
824 See also bugs #45.c and #183
827 In sbcl-0.7.1.3 on x86, COMPILE-FILE on the file
828 (in-package :cl-user)
831 (defstruct foo bar bletch)
833 (labels ((kidify1 (kid)
841 (declare (inline kid-frob))
844 (the simple-vector (foo-bar perd)))))
846 debugger invoked on condition of type TYPE-ERROR:
847 The value NIL is not of type SB-C::NODE.
848 The location of this failure has moved around as various related
849 issues were cleaned up. As of sbcl-0.7.1.9, it occurs in
850 NODE-BLOCK called by LAMBDA-COMPONENT called by IR2-CONVERT-CLOSURE.
853 (essentially the same problem as a CMU CL bug reported by Martin
854 Cracauer on cmucl-imp 2002-02-19)
855 There is a hole in structure slot type checking. Compiling and LOADing
856 (declaim (optimize safety))
858 (bla 0 :type fixnum))
860 (let ((foo (make-foo)))
861 (setf (foo-bla foo) '(1 . 1))
862 (format t "Is ~a of type ~a a cons? => ~a~%"
864 (type-of (foo-bla foo))
865 (consp (foo-bla foo)))))
867 should signal an error, but in sbcl-0.7.1.21 instead gives the output
868 Is (1 . 1) of type CONS a cons? => NIL
869 without signalling an error.
872 Functions SUBTYPEP, TYPEP, UPGRADED-ARRAY-ELEMENT-TYPE, and
873 UPGRADED-COMPLEX-PART-TYPE should have an optional environment argument.
874 (reported by Alexey Dejneka sbcl-devel 2002-04-12)
877 (reported by Robert E. Brown 2002-04-16)
878 When a function is called with too few arguments, causing the
879 debugger to be entered, the uninitialized slots in the bad call frame
880 seem to cause GCish problems, being interpreted as tagged data even
881 though they're not. In particular, executing ROOM in the
882 debugger at that point causes AVER failures:
885 * (lisp-implementation-version)
891 failed AVER: "(SAP= CURRENT END)"
892 (Christophe Rhodes reports that this doesn't occur on the SPARC, which
893 isn't too surprising since there are many differences in stack
894 implementation and GC conservatism between the X86 and other ports.)
898 (in-package :cl-user)
900 (defmethod permanentize ((uustk uustk))
901 (flet ((frob (hash-table test-for-deletion)
903 (obj-entry.stale? (oe)
904 (destructuring-bind (key . datum) oe
905 (declare (type simple-vector key))
906 (deny0 (void? datum))
907 (some #'stale? key))))
908 (declare (inline frob obj-entry.stale?))
909 (frob (uustk.args-hash->obj-alist uustk)
911 (frob (uustk.hash->memoized-objs-list uustk)
914 in sbcl-0.7.3.11 causes an assertion failure,
917 (AND (NULL (BLOCK-SUCC B))
918 (NOT (BLOCK-DELETE-P B))
919 (NOT (EQ B (COMPONENT-HEAD #)))))"
922 In sbcl-0.7.3.11, compiling the (illegal) code
923 (in-package :cl-user)
924 (defmethod prove ((uustk uustk))
927 gives the (not terribly clear) error message
929 ; (during macroexpansion of (DEFMETHOD PROVE ...))
930 ; can't get template for (FROB NIL NIL)
931 The problem seems to be that the code walker used by the DEFMETHOD
932 macro is unhappy with the illegal syntax in the method body, and
933 is giving an unclear error message.
936 sbcl's treatment of at least macro lambda lists is too permissive;
937 e.g., in sbcl-0.7.3.7:
938 (defmacro foo (&rest rest bar) `(,bar ,rest))
939 (macroexpand '(foo quux zot)) -> (QUUX (QUUX ZOT))
940 whereas section 3.4.4 of the CLHS doesn't allow required parameters
941 to come after the rest argument.
944 The compiler sometimes tries to constant-fold expressions before
945 it checks to see whether they can be reached. This can lead to
946 bogus warnings about errors in the constant folding, e.g. in code
949 (WRITE-STRING (> X 0) "+" "0"))
950 compiled in a context where the compiler can prove that X is NIL,
951 and the compiler complains that (> X 0) causes a type error because
952 NIL isn't a valid argument to #'>. Until sbcl-0.7.4.10 or so this
953 caused a full WARNING, which made the bug really annoying because then
954 COMPILE and COMPILE-FILE returned FAILURE-P=T for perfectly legal
955 code. Since then the warning has been downgraded to STYLE-WARNING,
956 so it's still a bug but at least it's a little less annoying.
959 reported by Alexey Dejneka 08 Jun 2002 in sbcl-devel:
960 Playing with McCLIM, I've received an error "Unbound variable WRAPPER
961 in SB-PCL::CHECK-WRAPPER-VALIDITY".
962 (defun check-wrapper-validity (instance)
963 (let* ((owrapper (wrapper-of instance)))
964 (if (not (invalid-wrapper-p owrapper))
966 (let* ((state (wrapper-state wrapper)) ; !!!
968 I've tried to replace it with OWRAPPER, but now OBSOLETE-INSTANCE-TRAP
969 breaks with "NIL is not of type SB-KERNEL:LAYOUT".
971 partial fix: The undefined variable WRAPPER resulted from an error
972 in recent refactoring, as can be seen by comparing to the code in e.g.
973 sbcl-0.7.2. Replacing WRAPPER with OWRAPPER (done by WHN in sbcl-0.7.4.22)
974 should bring the code back to its behavior as of sbcl-0.7.2, but
975 that still leaves the OBSOLETE-INSTANCE-TRAP bug. An example of
976 input which triggers that bug is
978 (let ((lastname (intern (format nil "C~D" (1- i))))
979 (name (intern (format nil "C~D" i))))
980 (eval `(defclass ,name
981 (,@(if (= i 0) nil (list lastname)))
983 (eval `(defmethod initialize-instance :after ((x ,name) &rest any)
984 (declare (ignore any))))))
991 178: "AVER failure compiling confused THEs in FUNCALL"
992 In sbcl-0.7.4.24, compiling
994 (funcall (the function (the standard-object x))))
997 "(AND (EQ (IR2-CONTINUATION-PRIMITIVE-TYPE 2CONT) FUNCTION-PTYPE) (EQ CHECK T))"
998 This variant compiles OK, though:
999 (defun bug178alternative (x)
1000 (funcall (the nil x)))
1002 183: "IEEE floating point issues"
1003 Even where floating point handling is being dealt with relatively
1004 well (as of sbcl-0.7.5, on sparc/sunos and alpha; see bug #146), the
1005 accrued-exceptions and current-exceptions part of the fp control
1006 word don't seem to bear much relation to reality. E.g. on
1010 debugger invoked on condition of type DIVISION-BY-ZERO:
1011 arithmetic error DIVISION-BY-ZERO signalled
1012 0] (sb-vm::get-floating-point-modes)
1014 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
1015 :ROUNDING-MODE :NEAREST
1016 :CURRENT-EXCEPTIONS NIL
1017 :ACCRUED-EXCEPTIONS (:INEXACT)
1020 * (sb-vm::get-floating-point-modes)
1021 (:TRAPS (:OVERFLOW :INVALID :DIVIDE-BY-ZERO)
1022 :ROUNDING-MODE :NEAREST
1023 :CURRENT-EXCEPTIONS (:INEXACT)
1024 :ACCRUED-EXCEPTIONS (:INEXACT)
1027 185: "top-level forms at the REPL"
1028 * (locally (defstruct foo (a 0 :type fixnum)))
1031 ; (in macroexpansion of (SB-KERNEL::%DELAYED-GET-COMPILER-LAYOUT BAR))
1032 however, compiling and loading the same expression in a file works
1035 187: "type inference confusion around DEFTRANSFORM time"
1036 (reported even more verbosely on sbcl-devel 2002-06-28 as "strange
1037 bug in DEFTRANSFORM")
1038 After the file below is compiled and loaded in sbcl-0.7.5, executing
1039 (TCX (MAKE-ARRAY 4 :FILL-POINTER 2) 0)
1040 at the REPL returns an adjustable vector, which is wrong. Presumably
1041 somehow the DERIVE-TYPE information for the output values of %WAD is
1042 being mispropagated as a type constraint on the input values of %WAD,
1043 and so causing the type test to be optimized away. It's unclear how
1044 hand-expanding the DEFTRANSFORM would change this, but it suggests
1045 the DEFTRANSFORM machinery (or at least the way DEFTRANSFORMs are
1046 invoked at a particular phase) is involved.
1047 (cl:in-package :sb-c)
1048 (eval-when (:compile-toplevel)
1049 ;;; standin for %DATA-VECTOR-AND-INDEX
1050 (defknown %dvai (array index)
1052 (foldable flushable))
1053 (deftransform %dvai ((array index)
1057 (let* ((atype (continuation-type array))
1058 (eltype (array-type-specialized-element-type atype)))
1059 (when (eq eltype *wild-type*)
1060 (give-up-ir1-transform
1061 "specialized array element type not known at compile-time"))
1062 (when (not (array-type-complexp atype))
1063 (give-up-ir1-transform "SIMPLE array!"))
1064 `(if (array-header-p array)
1065 (%wad array index nil)
1066 (values array index))))
1067 ;;; standin for %WITH-ARRAY-DATA
1068 (defknown %wad (array index (or index null))
1069 (values (simple-array * (*)) index index index)
1070 (foldable flushable))
1071 ;;; (Commenting out this optimizer causes the bug to go away.)
1072 (defoptimizer (%wad derive-type) ((array start end))
1073 (let ((atype (continuation-type array)))
1074 (when (array-type-p atype)
1075 (values-specifier-type
1076 `(values (simple-array ,(type-specifier
1077 (array-type-specialized-element-type atype))
1079 index index index)))))
1081 (defun %wad (array start end)
1082 (format t "~&in %WAD~%")
1083 (%with-array-data array start end))
1084 (cl:in-package :cl-user)
1086 (declare (type (vector t) v))
1087 (declare (notinline sb-kernel::%with-array-data))
1088 ;; (Hand-expending DEFTRANSFORM %DVAI here also causes the bug to
1092 188: "compiler performance fiasco involving type inference and UNION-TYPE"
1093 (In sbcl-0.7.6.10, DEFTRANSFORM CONCATENATE was commented out until this
1094 bug could be fixed properly, so you won't see the bug unless you restore
1095 the DEFTRANSFORM by hand.) In sbcl-0.7.5.11 on a 700 MHz Pentium III,
1099 (declare (optimize (safety 3)))
1100 (declare (optimize (compilation-speed 2)))
1101 (declare (optimize (speed 1) (debug 1) (space 1)))
1102 (let ((fn "if-this-file-exists-the-universe-is-strange"))
1103 (load fn :if-does-not-exist nil)
1104 (load (concatenate 'string fn ".lisp") :if-does-not-exist nil)
1105 (load (concatenate 'string fn ".fasl") :if-does-not-exist nil)
1106 (load (concatenate 'string fn ".misc-garbage")
1107 :if-does-not-exist nil)))))
1109 134.552 seconds of real time
1110 133.35156 seconds of user run time
1111 0.03125 seconds of system run time
1112 [Run times include 2.787 seconds GC run time.]
1114 246883368 bytes consed.
1115 BACKTRACE from Ctrl-C in the compilation shows that the compiler is
1116 thinking about type relationships involving types like
1118 (OR (INTEGER 576 576)
1129 190: "PPC/Linux pipe? buffer? bug"
1130 In sbcl-0.7.6, the run-program.test.sh test script sometimes hangs
1131 on the PPC/Linux platform, waiting for a zombie env process. This
1132 is a classic symptom of buffer filling and deadlock, but it seems
1133 only sporadically reproducible.
1135 191: "Miscellaneous PCL deficiencies"
1136 (reported by Alexey Dejneka sbcl-devel 2002-08-04)
1137 a. DEFCLASS does not inform the compiler about generated
1138 functions. Compiling a file with
1139 (DEFCLASS A-CLASS ()
1141 (DEFUN A-CLASS-X (A)
1142 (WITH-SLOTS (A-CLASS-X) A
1144 results in a STYLE-WARNING:
1146 SB-SLOT-ACCESSOR-NAME::|COMMON-LISP-USER A-CLASS-X slot READER|
1148 APD's fix for this was checked in to sbcl-0.7.6.20, but Pierre
1149 Mai points out that the declamation of functions is in fact
1150 incorrect in some cases (most notably for structure
1151 classes). This means that at present erroneous attempts to use
1152 WITH-SLOTS and the like on classes with metaclass STRUCTURE-CLASS
1153 won't get the corresponding STYLE-WARNING.
1154 c. the examples in CLHS 7.6.5.1 (regarding generic function lambda
1155 lists and &KEY arguments) do not signal errors when they should.
1157 192: "Python treats free type declarations as promises."
1158 b. What seemed like the same fundamental problem as bug 192a, but
1159 was not fixed by the same (APD "more strict type checking
1160 sbcl-devel 2002-08-97) patch:
1161 (DOTIMES (I ...) (DOTIMES (J ...) (DECLARE ...) ...)):
1162 (declaim (optimize (speed 1) (safety 3)))
1163 (defun trust-assertion (i)
1165 (declare (type (mod 4) i)) ; when commented out, behavior changes!
1168 (trust-assertion 6) ; prints nothing unless DECLARE is commented out
1170 193: "unhelpful CLOS error reporting when the primary method is missing"
1172 (defmethod foo :before ((x t)) (print x))
1173 is the only method defined on FOO, the error reporting when e.g.
1175 is relatively unhelpful:
1176 There is no primary method for the generic function
1177 #<STANDARD-GENERIC-FUNCTION FOO (1)>.
1178 with the offending argument nowhere visible in the backtrace. This
1179 continues even if there *are* primary methods, just not for the
1180 specified arg type, e.g.
1181 (defmethod foo ((x character)) (print x))
1182 (defmethod foo ((x string)) (print x))
1183 (defmethod foo ((x pathname)) ...)
1184 In that case it could be very helpful to know what argument value is
1185 falling through the cracks of the defined primary methods, but the
1186 error message stays the same (even BACKTRACE doesn't tell you what the
1187 bad argument value is).
1189 194: "no error from (THE REAL '(1 2 3)) in some cases"
1192 (multiple-value-prog1 (progn (the real '(1 2 3))))
1193 returns (1 2 3) instead of signalling an error. This was fixed by
1194 APD's "more strict type checking patch", but although the fixed
1195 code (in sbcl-0.7.7.19) works (signals TYPE-ERROR) interactively,
1196 it's difficult to write a regression test for it, because
1197 (IGNORE-ERRORS (MULTIPLE-VALUE-PROG1 (PROGN (THE REAL '(1 2 3)))))
1198 still returns (1 2 3).
1200 b. (IGNORE-ERRORS (MULTIPLE-VALUE-PROG1 (PROGN (THE REAL '(1 2 3)))))
1201 returns (1 2 3). (As above, this shows up when writing regression
1202 tests for fixed-ness of part a.)
1203 c. Also in sbcl-0.7.7.9, (IGNORE-ERRORS (THE REAL '(1 2 3))) => (1 2 3).
1205 (null (ignore-errors
1207 (arg2 (identity (the real #(1 2 3)))))
1208 (if (< arg1 arg2) arg1 arg2))))
1210 but putting the same expression inside (DEFUN FOO () ...),
1213 * Actually this entry is probably multiple bugs, as
1214 Alexey Dejneka commented on sbcl-devel 2002-09-03:)
1215 I don't think that placing these two bugs in one entry is
1216 a good idea: they have different explanations. The second
1217 (min 1 nil) is caused by flushing of unused code--IDENTITY
1218 can do nothing with it. So it is really bug 122. The first
1219 (min nil) is due to M-V-PROG1: substituting a continuation
1220 for the result, it forgets about type assertion. The purpose
1221 of IDENTITY is to save the restricted continuation from
1222 inaccurate transformations.
1223 * Alexey Dejneka pointed out that
1224 (IGNORE-ERRORS (IDENTITY (THE REAL '(1 2 3))))
1225 works as it should. Also
1226 (IGNORE-ERRORS (VALUES (THE REAL '(1 2 3))))
1227 works as it should. Perhaps this is another case of VALUES type
1228 intersections behaving in non-useful ways?
1230 199: "hairy FUNCTION types confuse the compiler"
1231 (reported by APD sbcl-devel 2002-09-15)
1233 (EQ NIL (FUNCALL F)))
1236 (DECLARE (TYPE (AND FUNCTION (SATISFIES MUR)) F))
1239 fails to compile, printing
1241 "(AND (EQ (IR2-CONTINUATION-PRIMITIVE-TYPE 2CONT) FUNCTION-PTYPE) (EQ CHECK T))"
1243 APD further reports that this bug is not present in CMUCL.
1245 201: "Incautious type inference from compound CONS types"
1246 (reported by APD sbcl-devel 2002-09-17)
1248 (LET ((Y (CAR (THE (CONS INTEGER *) X))))
1250 (FORMAT NIL "~S IS ~S, Y = ~S"
1257 (FOO ' (1 . 2)) => "NIL IS INTEGER, Y = 1"
1260 In 0.6.12.43 compilation of a function definition, contradicting its
1261 FTYPE proclamation, causes an error, e.g. COMPILE-FILE on
1263 (declaim (ftype (function () null) foo))
1268 debugger invoked on condition of type UNBOUND-VARIABLE:
1269 The variable SB-C::*ERROR-FUNCTION* is unbound.
1274 "~@<The previously declared FTYPE~2I ~_~S~I ~_~
1275 conflicts with the definition type ~2I~_~S~:>"
1279 (In 0.7.0 the variable was renamed to SB-C::*LOSSAGE-FUN*.)
1281 DEFUNCT CATEGORIES OF BUGS
1283 These labels were used for bugs related to the old IR1 interpreter.
1284 The # values reached 6 before the category was closed down.