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 Under sbcl-1.2.3, when I compile and load the file
15 (DEFSTRUCT (FOO (:PRINT-OBJECT (LAMBDA (X Y)
16 (LET ((*PRINT-LENGTH* 4))
19 then at the command line type
21 the program loops endlessly instead of printing the object.
24 KNOWN PORT-SPECIFIC BUGS
26 The breakpoint-based TRACE facility doesn't work properly in the
27 OpenBSD port of sbcl-0.6.7.
31 (There is also some information on bugs in the manual page and in the
32 TODO file. Eventually more such information may move here.)
34 * (DESCRIBE NIL) causes an endless loop.
36 * The FUNCTION special operator doesn't check properly whether its
37 argument is a function name. E.g. (FUNCTION (X Y)) returns a value
38 instead of failing with an error.
40 * (DESCRIBE 'GF) fails where GF is the name of a generic function:
41 The function SB-IMPL::DESCRIBE-INSTANCE is undefined.
43 * Failure in initialization files is not handled gracefully -- it's
44 a throw to TOP-LEVEL-CATCHER, which is not caught until we enter
45 TOPLEVEL-REPL. Code should be added to catch such THROWs even when
46 we're not in TOPLEVEL-REPL and do *something* with them (probably
47 complaining about an error outside TOPLEVEL-REPL, perhaps printing
48 a BACKTRACE, then terminating execution of SBCL).
50 * COMPILED-FUNCTION-P bogusly reports T for interpreted functions:
51 * (DEFUN FOO (X) (- 12 X))
53 * (COMPILED-FUNCTION-P #'FOO)
56 * The CL:STEP macro is undefined.
58 * DEFSTRUCT should almost certainly overwrite the old LAYOUT information
59 instead of just punting when a contradictory structure definition
62 * It should cause a STYLE-WARNING, not a full WARNING, when a structure
63 slot default value does not match the declared structure slot type.
64 (The current behavior is consistent with SBCL's behavior elsewhere,
65 and would not be a problem, except that the other behavior is
66 specifically required by the ANSI spec.)
68 * It should cause a STYLE-WARNING, not a WARNING, when the system ignores
69 an FTYPE proclamation for a slot accessor.
71 * Missing ordinary arguments in a macro call aren't reported when the
72 macro lambda list contains &KEY:
73 (DEFMACRO FOO (BAR &KEY) BAR) => FOO
75 Also in DESTRUCTURING-BIND:
76 (DESTRUCTURING-BIND (X Y &REST REST) '(1) (VECTOR X Y REST))
78 Also with &REST lists:
79 (DEFMACRO FOO (BAR &REST REST) BAR) => FOO
82 * Error reporting on various stream-requiring operations is not
83 very good when the stream argument has the wrong type, because
84 the operation tries to fall through to Gray stream code, and then
85 dies because it's undefined. E.g.
86 (PRINT-UNREADABLE-OBJECT (*STANDARD-OUTPUT* 1))
87 gives the error message
88 error in SB-KERNEL::UNDEFINED-SYMBOL-ERROR-HANDLER:
89 The function SB-IMPL::STREAM-WRITE-STRING is undefined.
90 It would be more useful and correct to signal a TYPE-ERROR:
92 (It wouldn't be terribly difficult to write stubs for all the
93 Gray stream functions that the old CMU CL code expects, with
94 each stub just raising the appropriate TYPE-ERROR.)
96 * bogus warnings about undefined functions for magic functions like
97 SB!C::%%DEFUN and SB!C::%DEFCONSTANT when cross-compiling files
98 like src/code/float.lisp
100 * The "byte compiling top-level form:" output ought to be condensed.
101 Perhaps any number of such consecutive lines ought to turn into a
102 single "byte compiling top-level forms:" line.
104 * The handling of IGNORE declarations on lambda list arguments of DEFMETHOD
105 is at least weird, and in fact seems broken and useless. I should
106 fix up another layer of binding, declared IGNORABLE, for typed
107 lambda list arguments.
109 * Compiling a file containing the erroneous program
113 (DEFSTRUCT (BAR (:INCLUDE FOO))
116 gives only the not-very-useful message
118 (during macroexpansion)
119 Condition PROGRAM-ERROR was signalled.
120 (The specific message which says that the problem was duplicate
121 slot names gets lost.)
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..
136 * It would be nice if the
138 (during macroexpansion)
139 said what macroexpansion was at fault, e.g.
141 (during macroexpansion of IN-PACKAGE,
142 during macroexpansion of DEFFOO)
144 * The type system doesn't understand the KEYWORD type very well:
145 (SUBTYPEP 'KEYWORD 'SYMBOL) => NIL, NIL
146 It might be possible to fix this by changing the definition of
147 KEYWORD to (AND SYMBOL (SATISFIES KEYWORDP)), but the type system
148 would need to be a bit smarter about AND types, too:
149 (SUBTYPEP '(AND SYMBOL KEYWORD) 'SYMBOL) => NIL, NIL
150 (The type system does know something about AND types already,
151 (SUBTYPEP '(AND INTEGER FLOAT) 'NUMBER) => T, T
152 (SUBTYPEP '(AND INTEGER FIXNUM) 'NUMBER) =>T, T
153 so likely this is a small patch.)
155 * Floating point infinities are screwed up. [When I was converting CMU CL
156 to SBCL, I was looking for complexity to delete, and I thought it was safe
157 to just delete support for floating point infinities. It wasn't: they're
158 generated by the floating point hardware even when we remove support
159 for them in software. -- WHN] Support for them should be restored.
161 * The ANSI syntax for non-STANDARD method combination types in CLOS is
162 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
163 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
164 If you mess this up, omitting the PROGN qualifier in in DEFMETHOD,
165 (DEFGENERIC FOO (X) (:METHOD-COMBINATION PROGN))
166 (DEFMETHOD FOO ((X BAR)) (PRINT 'NUMBER))
167 the error mesage is not easy to understand:
168 INVALID-METHOD-ERROR was called outside the dynamic scope
169 of a method combination function (inside the body of
170 DEFINE-METHOD-COMBINATION or a method on the generic
171 function COMPUTE-EFFECTIVE-METHOD).
172 It would be better if it were more informative, a la
173 The method combination type for this method (STANDARD) does
174 not match the method combination type for the generic function
176 Also, after you make the mistake of omitting the PROGN qualifier
177 on a DEFMETHOD, doing a new DEFMETHOD with the correct qualifier
179 (DEFMETHOD FOO PROGN ((X BAR)) (PRINT 'NUMBER))
181 INVALID-METHOD-ERROR was called outside the dynamic scope
182 of a method combination function (inside the body of
183 DEFINE-METHOD-COMBINATION or a method on the generic
184 function COMPUTE-EFFECTIVE-METHOD).
185 This is not very helpful..
187 * The message "The top of the stack was encountered." from the debugger
188 is not helpful when I type "FRAME 0" -- I know I'm going to the top
191 * (SUBTYPEP '(FUNCTION (T BOOLEAN) NIL)
192 '(FUNCTION (FIXNUM FIXNUM) NIL)) => T, T
193 (Also, when this is fixed, we can enable the code in PROCLAIM which
194 checks for incompatible FTYPE redeclarations.)
196 * The ANSI spec says that CONS can be a compound type spec, e.g.
197 (CONS FIXNUM REAL). SBCL doesn't support this.
199 * from Paolo Amoroso on the CMU CL mailing list 27 Feb 2000:
200 I use CMU CL 18b under Linux. When COMPILE-FILE is supplied a physical
201 pathname, the type of the corresponding compiled file is X86F:
202 * (compile-file "/home/paolo/lisp/tools/foo")
203 Python version 1.0, VM version Intel x86 on 27 FEB 0 06:00:46 pm.
204 Compiling: /home/paolo/lisp/tools/foo.lisp 27 FEB 0 05:57:42 pm
206 Compiling DEFUN SQUARE:
207 Byte Compiling Top-Level Form:
208 /home/paolo/lisp/tools/foo.x86f written.
209 Compilation finished in 0:00:00.
210 #p"/home/paolo/lisp/tools/foo.x86f"
213 But when the function is called with a logical pathname, the file type
215 * (compile-file "tools:foo")
216 Python version 1.0, VM version Intel x86 on 27 FEB 0 06:01:04 pm.
217 Compiling: /home/paolo/lisp/tools/foo.lisp 27 FEB 0 05:57:42 pm
219 Compiling DEFUN SQUARE:
220 Byte Compiling Top-Level Form:
221 TOOLS:FOO.FASL written.
222 Compilation finished in 0:00:00.
223 #p"/home/paolo/lisp/tools/foo.fasl"
227 * from DTC on the CMU CL mailing list 25 Feb 2000:
228 ;;; Compiler fails when this file is compiled.
230 ;;; Problem shows up in delete-block within ir1util.lisp. The assertion
231 ;;; (assert (member (functional-kind lambda) '(:let :mv-let :assignment)))
232 ;;; fails within bind node branch.
234 ;;; Note that if c::*check-consistency* is enabled then an un-reached
235 ;;; entry is also reported.
238 (declare (values nil))
255 (let ((ttt #'(lambda () (go cccc))))
256 (declare (special ttt))
257 (return-from bbbb nil))
260 (return-from bbbb nil))))))
262 * (I *think* this is a bug. It certainly seems like strange behavior. But
263 the ANSI spec is scary, dark, and deep..)
264 (FORMAT NIL "~,1G" 1.4) => "1. "
265 (FORMAT NIL "~3,1G" 1.4) => "1. "
267 * from Marco Antoniotti on cmucl-imp mailing list 1 Mar 2000:
269 (setf (find-class 'ccc1) (find-class 'ccc))
270 (defmethod zut ((c ccc1)) 123)
271 DTC's recommended workaround from the mailing list 3 Mar 2000:
272 (setf (pcl::find-class 'ccc1) (pcl::find-class 'ccc))
274 * There's probably a bug in the compiler handling of special variables
275 in closures, inherited from the CMU CL code, as reported on the
276 CMU CL mailing list. There's a patch for this on the CMU CL
278 Message-ID: <38C8E188.A1E38B5E@jeack.com.au>
279 Date: Fri, 10 Mar 2000 22:50:32 +1100
280 From: "Douglas T. Crosher" <dtc@jeack.com.au>
282 * The ANSI spec, in section "22.3.5.2 Tilde Less-Than-Sign: Logical Block",
283 says that an error is signalled if ~W, ~_, ~<...~:>, ~I, or ~:T is used
284 inside "~<..~>" (without the colon modifier on the closing syntax).
285 However, SBCL doesn't do this:
286 * (FORMAT T "~<munge~wegnum~>" 12)
290 * When too many files are opened, OPEN will fail with an
291 uninformative error message
292 error in function OPEN: error opening #P"/tmp/foo.lisp": NIL
293 instead of saying that too many files are open.
295 * Right now, when COMPILE-FILE has a read error, it actually pops
296 you into the debugger before giving up on the file. It should
297 instead handle the error, perhaps issuing (and handling)
298 a secondary error "caught ERROR: unrecoverable error during compilation"
299 and then return with FAILURE-P true,
301 * The print system doesn't conform to ANSI
302 "22.1.3.3.1 Package Prefixes for Symbols" for keywords printed when
303 *PACKAGE* is the KEYWORD package.
305 from a message by Ray Toy on CMU CL mailing list Fri, 28 Apr 2000:
307 In a discussion on comp.lang.lisp, the following code was given (by
310 (let ((*package* (find-package :keyword)))
311 (write-to-string object :readably t))
313 If OBJECT is a keyword, CMUCL prints out the keyword, but without a
314 colon. Hence, it's not readable, as requested.
316 I think the following patch will make this work as expected. The
317 patch just basically checks for the keyword package first before
318 checking the current package.
322 --- ../cmucl-18c/src/code/print.lisp Wed Dec 8 14:33:47 1999
323 +++ ../cmucl-18c/new/code/print.lisp Fri Apr 28 09:21:29 2000
324 @@ -605,12 +605,12 @@
325 (let ((package (symbol-package object))
326 (name (symbol-name object)))
328 - ;; If the symbol's home package is the current one, then a
329 - ;; prefix is never necessary.
330 - ((eq package *package*))
331 ;; If the symbol is in the keyword package, output a colon.
332 ((eq package *keyword-package*)
333 (write-char #\: stream))
334 + ;; If the symbol's home package is the current one, then a
335 + ;; prefix is never necessary.
336 + ((eq package *package*))
337 ;; Uninterned symbols print with a leading #:.
339 (when (or *print-gensym* *print-readably*)
341 * from CMU CL mailing list 01 May 2000
343 I realize I can take care of this by doing (proclaim (ignore pcl::.slots1.))
344 but seeing as .slots0. is not-exported, shouldn't it be ignored within the
348 In: DEFMETHOD FOO-BAR-BAZ (RESOURCE-TYPE)
349 (DEFMETHOD FOO-BAR-BAZ
350 ((SELF RESOURCE-TYPE))
351 (SETF (SLOT-VALUE SELF 'NAME) 3))
352 --> BLOCK MACROLET PCL::FAST-LEXICAL-METHOD-FUNCTIONS
353 --> PCL::BIND-FAST-LEXICAL-METHOD-MACROS MACROLET
354 --> PCL::BIND-LEXICAL-METHOD-FUNCTIONS LET PCL::BIND-ARGS LET* PCL::PV-BINDING
355 --> PCL::PV-BINDING1 PCL::PV-ENV LET
357 (LET ((PCL::.SLOTS0. #))
362 Warning: Variable PCL::.SLOTS0. defined but never used.
364 Compilation unit finished.
367 #<Standard-Method FOO-BAR-BAZ (RESOURCE-TYPE) {480918FD}>
369 * reported by Sam Steingold on the cmucl-imp mailing list 12 May 2000:
371 Also, there is another bug: `array-displacement' should return an array
372 or nil as first value (as per ANSI CL), while CMUCL declares it as
373 returning an array as first value always.
375 * Sometimes (SB-EXT:QUIT) fails with
376 Argh! maximum interrupt nesting depth (4096) exceeded, exiting
377 Process inferior-lisp exited abnormally with code 1
378 I haven't noticed a repeatable case of this yet.
380 * The system accepts DECLAIM in most places where DECLARE would be
381 accepted, without even issuing a warning. ANSI allows this, but since
382 it's fairly easy to mistype DECLAIM instead of DECLARE, and the
383 meaning is rather different, and it's unlikely that the user
384 has a good reason for doing DECLAIM not at top level, it would be
385 good to issue a STYLE-WARNING when this happens. A possible
386 fix would be to issue STYLE-WARNINGs for DECLAIMs not at top level,
387 or perhaps to issue STYLE-WARNINGs for any EVAL-WHEN not at top level.
389 * There seems to be some sort of bug in the interaction of the
390 normal compiler, the byte compiler, and type predicates.
391 Compiling and loading this file
392 (IN-PACKAGE :CL-USER)
395 (DECLAIM (FTYPE (FUNCTION (FOO) FOO) FOO-BAR))
396 (DECLAIM (INLINE FOO-BAR))
398 (DECLARE (TYPE FOO FOO))
399 (LET ((RESULT2605 (BLOCK FOO-BAR (PROGN (THE FOO (FOO-A FOO))))))
400 (UNLESS (TYPEP RESULT2605 'FOO)
401 (LOCALLY (ERROR "OOPS")))
402 (THE FOO RESULT2605)))
404 (DEFPARAMETER *FOO* (MAKE-FOO :A (MAKE-FOO)))
405 (UNLESS (EQ *PRINT-LEVEL* 133)
408 (WHEN (TYPEP *X* 'FOO)
411 (PRINT (FOO-BAR *FOO*))
413 in sbcl-0.6.5 (or also in CMU CL 18b for FreeBSD) gives a call
414 to the undefined function SB-C::%INSTANCE-TYPEP. %INSTANCE-TYPEP
415 is not defined as a function because it's supposed to
416 be transformed away. My guess is what's happening is that
417 the mixture of toplevel and non-toplevel stuff and inlining
418 is confusing the system into compiling an %INSTANCE-TYPEP
419 form into byte code, where the DEFTRANSFORM which is supposed
420 to get rid of such forms is not effective.
422 * some sort of bug in inlining and RETURN-FROM in sbcl-0.6.5: Compiling
425 (BLOCK USED-BY-SOME-Y?
428 (UNLESS (REJECTED? Y)
429 (RETURN-FROM USED-BY-SOME-Y? T)))))
430 (DECLARE (INLINE FROB))
435 error in function SB-KERNEL:ASSERT-ERROR:
436 The assertion (EQ (SB-C::CONTINUATION-KIND SB-C::CONT) :BLOCK-START) failed.
438 * The CMU CL reader code takes liberties in binding the standard read table
439 when reading the names of characters. Tim Moore posted a patch to the
440 CMU CL mailing list Mon, 22 May 2000 21:30:41 -0700.
442 * In some cases the compiler believes type declarations on array
443 elements without checking them, e.g.
444 (DECLAIM (OPTIMIZE (SAFETY 3) (SPEED 1) (SPACE 1)))
447 (DECLARE (TYPE (SIMPLE-ARRAY CONS 1) X))
448 (WHEN (CONSP (AREF X 0))
450 (BAR (VECTOR (MAKE-FOO :A 11 :B 12)))
453 in SBCL 0.6.5 (and also in CMU CL 18b). This does not happen for
454 all cases, e.g. the type assumption *is* checked if the array
455 elements are declared to be of some structure type instead of CONS.
457 * The printer doesn't report closures very well. This is true in
461 #<Closure Over Function "DEFUN STRUCTURE-SLOT-ACCESSOR" {134D1A1}>
462 It would be nice to make closures have a settable name slot,
463 and make things like DEFSTRUCT and FLET, which create closures,
464 set helpful values into this slot.
466 * And as long as we're wishing, it would be awfully nice if INSPECT could
467 also report on closures, telling about the values of the bound variables.
469 * as reported by Robert Strandh on the CMU CL mailing list 12 Jun 2000:
471 (defconstant +a-constant+ (make-instance 'a-class))
472 (defconstant +another-constant+ (vector +a-constant+))
474 CMU Common Lisp release x86-linux 2.4.19 8 February 2000 build 456,
477 Send bug reports and questions to your local CMU CL maintainer,
478 or to pvaneynd@debian.org
479 or to cmucl-help@cons.org. (prefered)
480 type (help) for help, (quit) to exit, and (demo) to see the demos
482 Python 1.0, target Intel x86
483 CLOS based on PCL version: September 16 92 PCL (f)
484 * (defclass a-class () ())
485 #<STANDARD-CLASS A-CLASS {48027BD5}>
486 * (compile-file "xx.lisp")
487 Python version 1.0, VM version Intel x86 on 12 JUN 00 08:12:55 am.
489 /home/strandh/Research/Functional/Common-Lisp/CLIM/Development/McCLIM
490 /xx.lisp 12 JUN 00 07:47:14 am
491 Compiling Load Time Value of (PCL::GET-MAKE-INSTANCE-FUNCTION-SYMBOL
493 Byte Compiling Top-Level Form:
494 Error in function C::DUMP-STRUCTURE: Attempt to dump invalid
496 #<A-CLASS {4803A5B5}>
499 * The compiler assumes that any time a function of declared FTYPE
500 doesn't signal an error, its arguments were of the declared type.
501 E.g. compiling and loading
502 (DECLAIM (OPTIMIZE (SAFETY 3)))
503 (DEFUN FACTORIAL (X) (GAMMA (1+ X)))
504 (DECLAIM (FTYPE (FUNCTION (UNSIGNED-BYTE) FACTORIAL)))
506 (COND ((> (FACTORIAL X) 1.0E6)
507 (FORMAT T "too big~%"))
509 (FORMAT T "exactly ~S~%" (FACTORIAL X)))
511 (FORMAT T "approximately ~S~%" (FACTORIAL X)))))
514 will cause the INTEGERP case to be selected, giving bogus output a la
516 This violates the "declarations are assertions" principle.
517 According to the ANSI spec, in the section "System Class FUNCTION",
518 this is a case of "lying to the compiler", but the lying is done
519 by the code which calls FACTORIAL with non-UNSIGNED-BYTE arguments,
520 not by the unexpectedly general definition of FACTORIAL. In any case,
521 "declarations are assertions" means that lying to the compiler should
522 cause an error to be signalled, and should not cause a bogus
523 result to be returned. Thus, the compiler should not assume
524 that arbitrary functions check their argument types. (It might
525 make sense to add another flag (CHECKED?) to DEFKNOWN to
526 identify functions which *do* check their argument types.)
528 * As pointed out by Martin Cracauer on the CMU CL mailing list
529 13 Jun 2000, the :FILE-LENGTH operation for
530 FD-STREAM-MISC-ROUTINE is broken for large files: it says
531 (THE INDEX SIZE) even though SIZE can be larger than INDEX.
533 * In SBCL 0.6.5 (and CMU CL 18b) compiling and loading
534 (in-package :cl-user)
535 (declaim (optimize (safety 3)
537 (compilation-speed 2)
540 #+nil (sb-ext:inhibit-warnings 2)))
541 (declaim (ftype (function * (values)) emptyvalues))
542 (defun emptyvalues (&rest rest) (declare (ignore rest)) (values))
544 (defgeneric assertoid ((x t)))
545 (defmethod assertoid ((x t)) "just a placeholder")
547 (declare (type hash-table ht))
553 (assertoid (hash-table-count ht)))))))
554 (unless (typep res 'foo)
556 (common-lisp-user::bad-result-from-assertive-typed-fun
560 (bar (make-hash-table))
562 Error in KERNEL::UNDEFINED-SYMBOL-ERROR-HANDLER:
563 the function C::%INSTANCE-TYPEP is undefined.
564 %INSTANCE-TYPEP is always supposed to be IR1-transformed away, but for
565 some reason -- the (VALUES) return value declaration? -- the optimizer is
566 confused and compiles a full call to %INSTANCE-TYPEP (which doesn't exist
567 as a function) instead.
569 * DEFMETHOD doesn't check the syntax of &REST argument lists properly,
570 accepting &REST even when it's not followed by an argument name:
571 (DEFMETHOD FOO ((X T) &REST) NIL)
573 * On the CMU CL mailing list 26 June 2000, Douglas Crosher wrote
575 Hannu Rummukainen wrote:
577 > There's something weird going on with the compilation of the attached
578 > code. Compiling and loading the file in a fresh lisp, then invoking
580 Thanks for the bug report, nice to have this one fixed. It was a bug
581 in the x86 backend, the < VOP. A fix has been committed to the main
582 source, see the file compiler/x86/float.lisp.
584 Probably the same bug exists in SBCL.
586 * TYPEP treats the result of UPGRADED-ARRAY-ELEMENT-TYPE as gospel,
587 so that (TYPEP (MAKE-ARRAY 3) '(VECTOR SOMETHING-NOT-DEFINED-YET))
588 returns (VALUES T T). Probably it should be an error instead,
589 complaining that the type SOMETHING-NOT-DEFINED-YET is not defined.
591 * TYPEP of VALUES types is sometimes implemented very inefficiently, e.g. in
592 (DEFTYPE INDEXOID () '(INTEGER 0 1000))
594 (DECLARE (TYPE INDEXOID X))
595 (THE (VALUES INDEXOID)
597 where the implementation of the type check in function FOO
598 includes a full call to %TYPEP. There are also some fundamental problems
599 with the interpretation of VALUES types (inherited from CMU CL, and
600 from the ANSI CL standard) as discussed on the cmucl-imp@cons.org
601 mailing list, e.g. in Robert Maclachlan's post of 21 Jun 2000.
603 * The definitions of SIGCONTEXT-FLOAT-REGISTER and
604 %SET-SIGCONTEXT-FLOAT-REGISTER in x86-vm.lisp say they're not
605 supported on FreeBSD because the floating point state is not saved,
606 but at least as of FreeBSD 4.0, the floating point state *is* saved,
607 so they could be supported after all. Very likely
608 SIGCONTEXT-FLOATING-POINT-MODES could now be supported, too.
610 * (as discussed by Douglas Crosher on the cmucl-imp mailing list ca.
611 Aug. 10, 2000): CMUCL currently interprets 'member as '(member); same issue
612 with 'union, 'and, 'or etc. So even though according to the ANSI spec,
613 bare 'MEMBER, 'AND, and 'OR are not legal types, CMUCL (and now
614 SBCL) interpret them as legal types.
616 * ANSI specifies DEFINE-SYMBOL-MACRO, but it's not defined in SBCL.
617 CMU CL added it ca. Aug 13, 2000, after some discussion on the mailing
618 list, and it is probably possible to use substantially the same
619 patches to add it to SBCL.
621 * a slew of floating-point-related errors reported by Peter Van Eynde
623 * (SQRT -9.0) fails, because SB-KERNEL::COMPLEX-SQRT is undefined.
624 Similarly, COMPLEX-ASIN, COMPLEX-ACOS, COMPLEX-ACOSH, and others
626 * SBCL's value for LEAST-POSITIVE-SHORT-FLOAT is bogus, and
627 should probably be 1.4012985e-45. In SBCL,
628 (/ LEAST-POSITIVE-SHORT-FLOAT 2) returns a number smaller
629 than LEAST-POSITIVE-SHORT-FLOAT. Similar problems
630 exist for LEAST-NEGATIVE-SHORT-FLOAT, LEAST-POSITIVE-LONG-FLOAT,
631 and LEAST-NEGATIVE-LONG-FLOAT.
632 * Many expressions generate floating infinity:
637 PVE's regression tests want them to raise errors. SBCL
638 generates the infinities instead, which may or may not be
639 conforming behavior, but then blow it by being unable to
640 output the infinities, since support for infinities is generally
641 broken, and in particular SB-IMPL::OUTPUT-FLOAT-INFINITY is
643 * (in section12.erg) various forms a la
644 (FLOAT 1 DOUBLE-FLOAT-EPSILON) don't give the right behavior.
646 * type safety errors reported by Peter Van Eynde July 25, 2000:
647 * (COERCE (QUOTE (A B C)) (QUOTE (VECTOR * 4)))
649 In general lengths of array type specifications aren't
650 checked by COERCE, so it fails when the spec is
651 (VECTOR 4), (STRING 2), (SIMPLE-BIT-VECTOR 3), or whatever.
652 * CONCATENATE has the same problem of not checking the length
653 of specified output array types. MAKE-SEQUENCE and MAP and
654 MERGE also have the same problem.
655 * (COERCE 'AND 'FUNCTION) returns something related to
656 (MACRO-FUNCTION 'AND), but ANSI says it should raise an error.
657 * ELT signals SIMPLE-ERROR if its index argument
658 isn't a valid index for its sequence argument, but should
659 signal TYPE-ERROR instead.
660 * FILE-LENGTH is supposed to signal a type error when its
661 argument is not a stream associated with a file, but doesn't.
662 * (FLOAT-RADIX 2/3) should signal an error instead of
664 * (LOAD "*.lsp") should signal FILE-ERROR.
665 * (MAKE-CONCATENATED-STREAM (MAKE-STRING-OUTPUT-STREAM))
666 should signal TYPE-ERROR.
667 * MAKE-TWO-WAY-STREAM doesn't check that its arguments can
668 be used for input and output as needed. It should fail with
669 TYPE-ERROR when handed e.g. the results of MAKE-STRING-INPUT-STREAM
670 or MAKE-STRING-OUTPUT-STREAM in the inappropriate positions,
672 * (PARSE-NAMESTRING (COERCE (LIST #\f #\o #\o (CODE-CHAR 0) #\4 #\8)
674 should probably signal an error instead of making a pathname with
676 * READ-BYTE is supposed to signal TYPE-ERROR when its argument is
677 not a binary input stream, but instead cheerfully reads from
678 character streams, e.g. (MAKE-STRING-INPUT-STREAM "abc").
680 * DEFCLASS bugs reported by Peter Van Eynde July 25, 2000:
681 * (DEFCLASS FOO () (A B A)) should signal a PROGRAM-ERROR, and doesn't.
682 * (DEFCLASS FOO () (A B A) (:DEFAULT-INITARGS X A X B)) should
683 signal a PROGRAM-ERROR, and doesn't.
684 * (DEFCLASS FOO07 NIL ((A :ALLOCATION :CLASS :ALLOCATION :CLASS))),
685 and other DEFCLASS forms with duplicate specifications in their
686 slots, should signal a PROGRAM-ERROR, and doesn't.
687 * (DEFGENERIC IF (X)) should signal a PROGRAM-ERROR, but instead
688 causes a COMPILER-ERROR.
690 * SYMBOL-MACROLET bugs reported by Peter Van Eynde July 25, 2000:
691 * (SYMBOL-MACROLET ((T TRUE)) ..) should probably signal
692 PROGRAM-ERROR, but SBCL accepts it instead.
693 * SYMBOL-MACROLET should refuse to bind something which is
694 declared as a global variable, signalling PROGRAM-ERROR.
695 * SYMBOL-MACROLET should signal PROGRAM-ERROR if something
696 it binds is declared SPECIAL inside.
698 * LOOP bugs reported by Peter Van Eynde July 25, 2000:
699 * (LOOP WITH (A B) DO (PRINT 1)) is a syntax error according to
700 the definition of WITH clauses given in the ANSI spec, but
701 compiles and runs happily in SBCL.
702 * a messy one involving package iteration:
703 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<))))
704 Should be: (("blah" "blah2") ("blah2"))
705 SBCL: (("blah") ("blah2"))
706 * (LET ((X 1)) (LOOP FOR I BY (INCF X) FROM X TO 10 COLLECT I))
707 doesn't work -- SBCL's LOOP says BY isn't allowed in a FOR clause.
709 * type system errors reported by Peter Van Eynde July 25, 2000:
710 * (SUBTYPEP 'BIGNUM 'INTEGER) => NIL, NIL
711 but should be (VALUES T T) instead.
712 * (SUBTYPEP 'EXTENDED-CHAR 'CHARACTER) => NIL, NIL
713 but should be (VALUES T T) instead.
714 * (SUBTYPEP '(INTEGER (0) (0)) 'NIL) dies with nested errors.
715 * In general, the system doesn't like '(INTEGER (0) (0)) -- it
716 blows up at the level of SPECIFIER-TYPE with
717 "Lower bound (0) is greater than upper bound (0)." Probably
718 SPECIFIER-TYPE should return NIL instead.
719 * (TYPEP 0 '(COMPLEX (EQL 0)) fails with
720 "Component type for Complex is not numeric: (EQL 0)."
721 This might be easy to fix; the type system already knows
722 that (SUBTYPEP '(EQL 0) 'NUMBER) is true.
723 * The type system doesn't know about the condition system,
724 so that e.g. (TYPEP 'SIMPLE-ERROR 'ERROR)=>NIL.
725 * The type system isn't all that smart about relationships
726 between hairy types, as shown in the type.erg test results,
727 e.g. (SUBTYPEP 'CONS '(NOT ATOM)) => NIL, NIL.
729 * miscellaneous errors reported by Peter Van Eynde July 25, 2000:
731 (DEFGENERIC FOO02 (X))
732 (DEFMETHOD FOO02 ((X NUMBER)) T)
733 (LET ((M (FIND-METHOD (FUNCTION FOO02)
735 (LIST (FIND-CLASS (QUOTE NUMBER))))))
736 (REMOVE-METHOD (FUNCTION FOO02) M)
737 (DEFGENERIC FOO03 (X))
738 (ADD-METHOD (FUNCTION FOO03) M)))
739 should give an error, but SBCL allows it.
740 * READ should probably return READER-ERROR, not the bare
741 arithmetic error, when input a la "1/0" or "1e1000" causes
743 * There are several metaobject protocol "errors". (In order to fix
744 them, we might need to document exactly what metaobject
745 protocol specification we're following -- the current code is
746 just inherited from PCL.)
747 * (BUTLAST NIL) should return NIL. (This appears to be a compiler
748 bug, since the definition of BUTLAST, when interpreted, does
749 give (BUTLAST NIL)=>NIL.)
751 * another error from Peter Van Eynde 5 September 2000:
752 (FORMAT NIL "~F" "FOO") should work, but instead reports an error.
753 PVE submitted a patch to deal with this bug, but it exposes other
754 comparably serious bugs, so I didn't apply it. It looks as though
755 the FORMAT code needs a fair amount of rewriting in order to comply
756 with the various details of the ANSI spec.
758 * The bug discussed on the cmucl-imp@cons.org mailing list ca. 5 September,
759 simplified by Douglas Crosher down to
776 causes the same problem on SBCL: compiling it fails with
777 :LET fell through ECASE expression.
778 Very likely the patch discussed there is appropriate for SBCL
779 as well, but I don't understand it, so I didn't apply it.