Re: hash tables: print, read, print-readably
Mike Kupfer writes:
So it's punching a hole in the hash table abstraction and
exposing
an implementation detail.
So what? It prints an object ID, too, which is also internal. The
usage ratio is a useful number to know in some cases (it has a bearing
on the performance of hash tables) and easy enough to print.
To make matters worse, the printed representation uses the same
"size"
keyword. So "size" means one thing in the default printed
representation, but it means something else in the readable
representation and in the API.
It actually means different things in the readable representation and
in `make-hash-table' it seems to me. In the readable representation,
it's the current number of items, while in the API it's the expected
long-run number. Thus
(let ((print-readably t)
(foo (make-hash-table :size 100)))
(puthash 1 1 foo)
(prin1 foo))
#s(hash-table:size 1 :data (1 1))
#<hash-table:size 1/197 :data (1 1) 0x20a1d>
But if you find a better word for the size in the unreadable
representation, we can use it there. I don't much like either of
"usage" or "ratio", and "usage ratio" is too long. It
should be
short, too, no more than about 5-6 characters.
The docstring for print and friends should mention print-readably.
A
specific list of objects for which print-readably is relevant (e.g.,
hash table) would be nice, but I can see arguments for not including
such a list.
The docstring for print-readably already has such a list, but it's
only two types long, specifically missing hash tables. I think that
is where it belongs.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-beta