bug with equalp hash-table?

[Nicolas Neuß]

According to the Spec, equalp should not distinguish between arrays depending on element-type and I guess the same should be true for equalp hash tables. Thus the following should probably return T,T:

(let ((table (make-hash-table :test 'equalp)))
  (let ((x (vector 1.0d0 1.0d0))
        (y (make-array 2 :element-type 'double-float :initial-contents '(1.0d0 1.0d0))))
    (setf (gethash x table) t)
    (gethash y table)))

It does so in SBCL 2.0.1. However, in recent SBCLs (I compiled SBCL 2.1.8.7-fb7df1c4d from source) it returns NIL,NIL.

uname -a
Linux mini 5.4.0-81-generic #91-Ubuntu SMP Thu Jul 15 19:09:17 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

*features*
(:SWANK :QUICKLISP :ASDF3.3 :ASDF3.2 :ASDF3.1 :ASDF3 :ASDF2 :ASDF :OS-UNIX
 :NON-BASE-CHARS-EXIST-P :ASDF-UNICODE :X86-64 :GENCGC :64-BIT :ANSI-CL
 :COMMON-LISP :ELF :IEEE-FLOATING-POINT :LINUX :LITTLE-ENDIAN
 :PACKAGE-LOCAL-NICKNAMES :SB-LDB :SB-PACKAGE-LOCKS :SB-THREAD :SB-UNICODE
 :SBCL :UNIX)

[Douglas Katzman (dougk)]

looks like e63feff826 should have used a psxhash where it used sxhash.
The fix isn't just that though, because to do it so simply would essentially nullify the effect of the change as far as avoiding consing