Operator <> was removed in Python 3, but still appears in Grammar/Grammar (and hence in Doc/reference/grammar.rst)

Reported by Alexander Ivanyuta on the docs mailing list

The relevant code in Parser/parsetok.c is:

#ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
        if (type == NOTEQUAL) {
            if (!(ps->p_flags & CO_FUTURE_BARRY_AS_BDFL) &&
                            strcmp(str, "!=")) {
                err_ret->error = E_SYNTAX;
                break;
            }
            else if ((ps->p_flags & CO_FUTURE_BARRY_AS_BDFL) &&
                            strcmp(str, "<>")) {
                err_ret->text = "with Barry as BDFL, use '<>' "
                                "instead of '!='";
                err_ret->error = E_SYNTAX;
                break;
            }
        }
#endif


Hmm... I'm too new to remember this joke, but here it presents a practical problem - since Grammar/Grammar is being directly reflected into the documentation (.. literalinclude:: ../../Grammar/Grammar) and thus can confuse.

Is it safe to just remove the whole thing from Grammar/Grammar and correspondingly Parser/parsetok.c ?

This is PEP 401.

"[Because] the != inequality operator ... was a horrible, finger pain inducing mistake, the FLUFL reinstates the <> diamond operator as the sole spelling. This change is important enough to be implemented for, and released in Python 3.1. To help transition to this feature, a new future statement, from __future__ import barry_as_FLUFL has been added."

Probably need a comment in the Grammar file so people know why an
unused operator is in there.

Or perhaps we don't need joke backward compatibility? (That's nearly 3 years old.)

On Fri, Oct 21, 2011 at 11:35, Benjamin Peterson <report@bugs.python.org> wrote:
>
> Benjamin Peterson <benjamin@python.org> added the comment:
>
> Or perhaps we don't need joke backward compatibility? (That's nearly 3 years old.)

Then you tell the FLUFL that you want to take his precious operator away. =)

On Oct 21, 2011, at 06:35 PM, Benjamin Peterson wrote:

>Benjamin Peterson <benjamin@python.org> added the comment:
>
>Or perhaps we don't need joke backward compatibility? (That's nearly 3 years
>old.)

OTOH, __future__ imports (even jokes) should never be removed.

> OTOH, __future__ imports (even jokes) should never be removed.

But their meaning can be altered?
(as part of another joke if you want :)

On Oct 21, 2011, at 07:33 PM, Antoine Pitrou wrote:

>Antoine Pitrou <pitrou@free.fr> added the comment:
>
>> OTOH, __future__ imports (even jokes) should never be removed.
>
>But their meaning can be altered?
>(as part of another joke if you want :)

Well, you have 6 months to work that out then. :)

Attaching a patch with a clarifying comment in Grammar/Grammar. Should be enough for now?

I think the clarification should be enough.

+1 for a comment too.  I’d even make it shorter:

# don't look at <>, it's not a real operator (see PEP 401)

Éric, do you feel strongly about the wording, or can I just go ahead and commit my version if I like it more :) ?

New changeset a259511351d9 by Eli Bendersky in branch '3.2':
Clarify the existence of the <> operator in Grammar/Grammar with a comment, for issue 13239
http://hg.python.org/cpython/rev/a259511351d9

New changeset 410115400838 by Eli Bendersky in branch 'default':
Clarify the existence of the <> operator in Grammar/Grammar with a comment. Closes issue 13239
http://hg.python.org/cpython/rev/410115400838

+# <> isn't actually a valid comparison operator in Python. It's here for the
+# sake of a __future__ import described in PEP 401

If we wanted to be exact, the operator isn’t here for a __future__ import but for a feature enabled by a __future__ import.  But I don’t feel strongly about this at all :)