/[gcl]/gcl/BUGS
ViewVC logotype

Diff of /gcl/BUGS

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.9 by pfdietz, Wed Aug 17 03:47:18 2005 UTC revision 1.10 by pfdietz, Wed Aug 17 12:50:10 2005 UTC
# Line 589  number in place so it isn't reused. Line 589  number in place so it isn't reused.
589  138. (lcm) should yield 1, not an error.  138. (lcm) should yield 1, not an error.
590       ansi-tests: lcm.1       ansi-tests: lcm.1
591    
592  139. (typep 1 '(real (0))) ==> error  139. (typep 1 '(real (0))) ==> error  (and similarly for
593            (typep 1 '(integer (0))).)
594       ansi-tests: random.error.3       ansi-tests: random.error.3
595    
596  140. (phase 0) ==> error (logarithmic singularity)  140. (phase 0) ==> error (logarithmic singularity)
# Line 603  number in place so it isn't reused. Line 604  number in place so it isn't reused.
604  142. rational and rationalize should signal type errors on arguments  142. rational and rationalize should signal type errors on arguments
605       that are not of type real in safe code.       that are not of type real in safe code.
606       ansi-tests: rational.error.4 rationalize.error.4       ansi-tests: rational.error.4 rationalize.error.4
607    
608    143. Apparent compiler problem in this function:
609    
610         (defun digit-char.1.body.old ()
611           (loop for r from 2 to 36 always
612                (loop for i from 0 to 36
613                   always (let* ((c (digit-char i r))
614                                 (result
615                                  (if (>= i r) (null c)
616                                      (eqlt c (char +extended-digit-chars+ i)))))
617                            (unless result
618                              (format t "~A ~A ~A~%" r i c))
619                            result))))
620    
621          When called, it prints  2 2 p and returns nil.  However,
622          (digit-char 2 2) ==> nil when called at the REPL.
623    
624          (this is from digit-char.1, but that test has been changed.)
625    
626    144. digit-char should take at most two arguments, but it is failing
627         to signal a program error in safe code when given more.
628         ansi-tests: digit-char.error.2
629    
630    145. standard-char-p should signal a type error in safe code when
631         called on a non-character.  ansi-tests: standard-char-p.3
632    
633    146. code-char should signal an error in safe code when called with
634         more than one argument.  ansi-tests: code-char.error.2
635    
636    147. adjust-array does not have the correct semantics when performed
637         on the middle of a chain of three displaced bit vectors.  The chain
638         appears to have been improperly short-circuited.
639         ansi-tests: adjust-array.bit-vector.adjustable.13
640    
641    148. fill-pointer should signal a type error when called with vectors
642         that lack fill pointers.  It does, but the expected-type field
643         is wrong.  ansi-tests: fill-pointer.error.[36]
644    
645    149. The value (upgraded-array-element-type 'base-char) should be
646         type-equivalent to base-char.  However, it is string-char,
647         and (subtypep 'string-char 'base-char) ==> nil t
648         ansi-tests: upgraded-array-element-type.2
649    
650    150. upgraded-array-element-type violates the 'preserves subtype'
651         properties for numerous types.  For example:
652    
653          (upgraded-array-element-type '(eql 2)) ==> lisp:signed-char
654          but
655          (upgraded-array-element-type '(unsigned-byte 16)) ==>
656                 lisp:unsigned-short
657    
658         ansi-tests: upgraded-array-element-type.8
659    
660    151. (upgraded-array-element-type nil) ==> t, but it should be nil.
661    
662    152. vector-pop should signal a type-error when passed a vector
663         that lacks a fill pointer.  It does, but the expected-type
664         slot is wrong.  ansi-tests: vector-pop.error.[15]
665    
666    153. (make-hash-table :size 0) signals an error, but 0 is a legal
667         value for the size parameter.  ansi-tests: make-hash-table.2
668    
669    154. equalp hash tables are not supported.
670         ansi-tests: make-hash-table.1[01] hash-table-count.[59]
671         hash-table-rehash-threshold.[23] hash-table-test.[234]
672         gethash.zero.4 remhash.4 clrhash.4 maphash.[36]
673         with-hash-table-iterator.7
674    
675    155. (make-hash-table :rehash-threshold x) errors:
676    
677        a. x == 0 (or 0.0s0, or 0.0)  signals an error, but zero
678         is a legal value for this parameter.  ansi-tests:
679         make-hash-table.1[6789] make-hash-table.20
680        
681        b. x == 1/2 signals an error, but this is a legal value for
682         this parameter.  ansi-tests: make-hash-table.21
683    
684    156. The third argument to GETHASH in the SETF form should be
685        allowed, but instead is causing an error.
686        ansi-tests: gethash.5 gethash.order.4
687    
688    157. The last invocation of the local macro in WITH-HASH-TABLE-ITERATOR
689        should return just a single value, NIL, but it is returning
690        three values.  ansi-tests: with-hash-table-iterator.4
691    
692    158. PATHNAME, when called on a stream, should return the pathname
693        that was used to create the stream.  However, for logical pathnames,
694        this is not happening -- the physical pathname corresponding to the
695        logical pathname is returned instead.
696        ansi-tests: logical-pathname.3
697    
698    159. ensure-directories-exist should signal a file-error if part of
699        the directory is wild, but it does not.
700        ansi-tests: ensure-directories-exist.error.1

Legend:
Removed from v.1.9  
changed lines
  Added in v.1.10

savannah-hackers-public@gnu.org
ViewVC Help
Powered by ViewVC 1.1.26