bug#72753: Different output between SBCL `loop` and Emacs `cl-loop`

bug#72753: Different output between SBCL `loop` and Emacs `cl-loop`

[Okamsn via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Hello,

The following in SBCL 2.1.11 produces 3, which I expected.

     (loop for i in '(1 2 3)
           for j from 1
           finally (return j))

The following in Emacs 29 produces 4, which I did not expect.

     (cl-loop for i in '(1 2 3)
              for j from 1
              finally (cl-return j))

This is a simplified example that only shows the difference.

I don't know what the standard behavior is, but I personally believe 
that the SBCL output makes more sense.

Thank you.

bug#72753: Different output between SBCL `loop` and Emacs `cl-loop`

[Stephen Berman via Bug reports for GNU Emacs, the Swiss army knife of text editors]

On Thu, 22 Aug 2024 01:55:07 +0000 Okamsn via "Bug reports for GNU Emacs, the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org> wrote:

> Hello,
>
> The following in SBCL 2.1.11 produces 3, which I expected.
>
>      (loop for i in '(1 2 3)
>            for j from 1
>            finally (return j))

I get the same result with SBCL 2.4.5.

> The following in Emacs 29 produces 4, which I did not expect.
>
>      (cl-loop for i in '(1 2 3)
>               for j from 1
>               finally (cl-return j))

I also get 4 with Emacs built from master.

> This is a simplified example that only shows the difference.
>
> I don't know what the standard behavior is, but I personally believe
> that the SBCL output makes more sense.

FWIW, GNU CLISP 2.49.93+ (built from the git repo) returns 4, not 3, for
the above Common Lisp loop sexp.

Steve Berman

bug#72753: Different output between SBCL `loop` and Emacs `cl-loop`

[Robert Pluim]

>>>>> On Thu, 22 Aug 2024 09:57:08 +0200, Stephen Berman via "Bug reports for GNU Emacs, the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org> said:

    >> 
    >> I don't know what the standard behavior is, but I personally believe
    >> that the SBCL output makes more sense.

    Stephen> FWIW, GNU CLISP 2.49.93+ (built from the git repo) returns 4, not 3, for
    Stephen> the above Common Lisp loop sexp.

The CLHS section 6.1.7.2 says

"The finally construct causes the supplied compound-forms to be
evaluated in the loop epilogue *after* normal iteration terminates."
(emphasis mine).

To me that means all bets are off regarding the values of variables
created in the context of the loop iteration. But we should ask a
Common Lisp expert to be sure.

Robert
--