GNU bug report logs - #40972
27.0.91; assoc TESTFN's args reversed?

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 40972 in the body.
You can then email your comments to 40972 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Shigeru Fukaya <shigeru.fukaya <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: 27.0.91; assoc TESTFN's args reversed?
Date: Thu, 30 Apr 2020 06:15:09 +0900

It seems to me, for third argument of `assoc', arguments are passed in a
reverse order.

(assoc "assoc" '(("cl-assoc" A) ("cl-member" B))
       (lambda (x y) (equal (concat "cl-" x) y)))
==> nil

(assoc "assoc" '(("cl-assoc" A) ("cl-member" B))
       (lambda (y x) (equal (concat "cl-" x) y)))
==> ("cl-assoc" A)

Or, is this intended behavior??

Thanks,
Shigeru

Message #8 received at 40972 <at> debbugs.gnu.org (full text, mbox):

From: Michael Heerdegen <michael_heerdegen <at> web.de>
To: Shigeru Fukaya <shigeru.fukaya <at> gmail.com>
Cc: 40972 <at> debbugs.gnu.org
Subject: Re: bug#40972: 27.0.91; assoc TESTFN's args reversed?
Date: Thu, 30 Apr 2020 01:30:13 +0200

Shigeru Fukaya <shigeru.fukaya <at> gmail.com> writes:

> It seems to me, for third argument of `assoc', arguments are passed in a
> reverse order.
>
> (assoc "assoc" '(("cl-assoc" A) ("cl-member" B))
>        (lambda (x y) (equal (concat "cl-" x) y)))
> ==> nil
>
> (assoc "assoc" '(("cl-assoc" A) ("cl-member" B))
>        (lambda (y x) (equal (concat "cl-" x) y)))
> ==> ("cl-assoc" A)
>
> Or, is this intended behavior??

Dunno what others expect, but note that your predicate is not an
equality predicate (it's not symmetric), and the docstring doesn't
specify a behavior for this case.

Michael.

Message #13 received at 40972 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefan <at> marxist.se>
To: Michael Heerdegen <michael_heerdegen <at> web.de>
Cc: 40972 <at> debbugs.gnu.org, Shigeru Fukaya <shigeru.fukaya <at> gmail.com>
Subject: Re: bug#40972: 27.0.91; assoc TESTFN's args reversed?
Date: Wed, 9 Sep 2020 06:22:18 -0700

tags 40972 + wontfix notabug
close 40972
thanks

Michael Heerdegen <michael_heerdegen <at> web.de> writes:

> Shigeru Fukaya <shigeru.fukaya <at> gmail.com> writes:
>
>> It seems to me, for third argument of `assoc', arguments are passed in a
>> reverse order.
>>
>> (assoc "assoc" '(("cl-assoc" A) ("cl-member" B))
>>        (lambda (x y) (equal (concat "cl-" x) y)))
>> ==> nil
>>
>> (assoc "assoc" '(("cl-assoc" A) ("cl-member" B))
>>        (lambda (y x) (equal (concat "cl-" x) y)))
>> ==> ("cl-assoc" A)
>>
>> Or, is this intended behavior??
>
> Dunno what others expect, but note that your predicate is not an
> equality predicate (it's not symmetric), and the docstring doesn't
> specify a behavior for this case.

As Michael points out, the equality operation is symmetric and therefore
by definition the order does not matter.  We therefore cannot say if X=Y
or Y=X is "more correct": they are equivalent.

In any case, I think it will cause more problems to switch the argument
order at this point, since there might be code out there that relies on
this (undocumented) behavior.  So I'm closing this bug.

Message #20 received at 40972 <at> debbugs.gnu.org (full text, mbox):

From: Mattias Engdegård <mattiase <at> acm.org>
To: Stefan Kangas <stefan <at> marxist.se>
Cc: Michael Heerdegen <michael_heerdegen <at> web.de>, 40972 <at> debbugs.gnu.org,
 Shigeru Fukaya <shigeru.fukaya <at> gmail.com>
Subject: Re: bug#40972: 27.0.91; assoc TESTFN's args reversed? 
Date: Wed, 9 Sep 2020 16:46:43 +0200

> As Michael points out, the equality operation is symmetric and therefore by definition the order does not matter. We therefore cannot say if X=Y or Y=X is "more correct": they are equivalent.
> 
> In any case, I think it will cause more problems to switch the argument order at this point, since there might be code out there that relies on this (undocumented) behavior. So I'm closing this bug. 

It's quite possible that the current argument order was intended but never really documented. It's occasionally exploited (see image-converter--filter-formats). The asymmetry is stated more explicitly in the doc string and manual entry for assoc-default. Closing the bug is probably wisest.

Message #23 received at 40972 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefan <at> marxist.se>
To: Mattias Engdegård <mattiase <at> acm.org>
Cc: Michael Heerdegen <michael_heerdegen <at> web.de>, 40972 <at> debbugs.gnu.org,
 Shigeru Fukaya <shigeru.fukaya <at> gmail.com>
Subject: Re: bug#40972: 27.0.91; assoc TESTFN's args reversed?
Date: Wed, 9 Sep 2020 08:18:30 -0700

Mattias Engdegård <mattiase <at> acm.org> writes:

> It's quite possible that the current argument order was intended but
> never really documented. It's occasionally exploited (see
> image-converter--filter-formats). The asymmetry is stated more
> explicitly in the doc string and manual entry for
> assoc-default. Closing the bug is probably wisest.

Should we fix the doc string for assoc to match assoc-default?

Message #26 received at 40972 <at> debbugs.gnu.org (full text, mbox):

From: Mattias Engdegård <mattiase <at> acm.org>
To: Stefan Kangas <stefan <at> marxist.se>, Nicolas Petton <nicolas <at> petton.fr>
Cc: Michael Heerdegen <michael_heerdegen <at> web.de>, 40972 <at> debbugs.gnu.org,
 Shigeru Fukaya <shigeru.fukaya <at> gmail.com>
Subject: Re: bug#40972: 27.0.91; assoc TESTFN's args reversed?
Date: Wed, 9 Sep 2020 17:51:46 +0200

9 sep. 2020 kl. 17.18 skrev Stefan Kangas <stefan <at> marxist.se>:

>> It's quite possible that the current argument order was intended but
>> never really documented. It's occasionally exploited (see
>> image-converter--filter-formats). The asymmetry is stated more
>> explicitly in the doc string and manual entry for
>> assoc-default. Closing the bug is probably wisest.
> 
> Should we fix the doc string for assoc to match assoc-default?

Good question. The TESTFN parameter was added recently (in Emacs time) to assoc (0bece6c6815). Nicolas, did you intend it to be symmetric, or defined in an asymmetric way like in assoc-default?

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.