bug#55156: [PATCH] eval.c: New functions `defvar-f` and `defconst-f`

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Date: Thu, 28 Apr 2022 00:33:47 +0200
> Cc: 55156@debbugs.gnu.org
> 
> Oh, if we call a function containing the defvar...  Yes, that's probably
> rare enough that nobody's noticed.

Famous last words.

> I think I'd prefer keeping the behaviour we currently promise, but I
> don't have a strong opinion.

I sincerely question the wisdom of messing with this, for the reasons
that were described, which seem to be some inelegant code somewhere in
the bowels of the byte compiler.  Do we really care enough about such
inelegance to make potentially breaking changes in code that works for
decades and causes zero trouble to Lisp programmers?

And I'm quite sure that the replacement code might look no more
elegant to people other than the author.

I suggest that we all take a step back and re-evaluate the need for
this.  It is IME exactly the kind of change that prevents Emacs from
becoming steadily more and more stable as time goes by.