Re: Lisp reader and #1=... abbreviations

>>>>> "Raymond" == Raymond Toy <[email protected]> writes:

    Raymond> Jared C. Davis wrote:
    >> Hi,
    >> I'm having trouble reading in files which use a lot of
    >> structure-sharing abbreviations.  For example, here is a typical file
    >> (490 KB) which has 18,000+ abbreviations.  Unfortunately, reading this
    >> (with "read") takes about 13 minutes in CMUCL 19d on my test machine.
    >> http://www.cs.utexas.edu/users/jared/Milawa/Sources/ACL2/bootstrap/utilities/proofs/thm-equal-of-booleans-rewrite.proof

    Raymond> Jared has sent me a patch (for sbcl) which basically works on
    Raymond> cmucl. For the test case above, it is now read in about 2 sec on my
    Raymond> test machine.

    Raymond> Don't know if it's right or not, though.  And I haven't tested the
    Raymond> speed of the reader using hash-tables instead of alists.

I've made a few changes to the patch and reading "normal" stuff is as
fast as before.  Only when reading #1= stuff is there a difference.
For the test file above, cmucl takes some 40 minutes to read it.  With
the patch, it takes 2.8 sec.  That's a nice savings!

However, with a different test:

(defun make-circ (n)
  (let ((list (make-list n)))
    (do ((head list (cdr head)))
	((null head))
      (setf (car head) head))
    list))

cmucl with the patch takes about 5-10% less time for lists of length
5000.  For lists of length 2, it's about twice as slow and conses
about 3 times more.

Ray