Re: Upcoming loss of usability of Emacs source files and Emacs.

[Tassilo Horn]

Paul Eggert <address@hidden> writes:

>> It is good to _permit_ putting non-ASCII characters in strings and
>> comments, as we do; but Lisp code (including its normal conventions)
>> shouldn't require use of such characters.
>
> This shouldn't be a problem.  The approach in Emacs master does not
> require the use of non-ASCII characters in source-code docstrings or
> diagnostics, and further changes along these lines should still
> support ASCII-only strings and comments containing quoted symbols that
> continue to work as before.

But with respect to the coding conventions Oleh cited, it is true that
the use of ‘foo-bar’ instead of `foo-bar' is at least encouraged for
lisp docstrings (and comments) although both will be displayed like the
former with describe-*, right?

To me, having two alternative ways to name lisp symbols in docstrings is
even worse than switching from `...' to ‘...’ in the first place [1].
I'd rather declare the former obsolete and make checkdoc bark at it to
get to some uniformity as soon as possible (which might mean in a bit
less than a decade).

[1] But I agree with Drew that the `...' in `foo-bar' isn't quoting but
    a markup which unambiguously says "foo-bar is a symbol."  In
    contrast, ‘foo-bar’ is ambiguous.  It could be a quote or a symbol.
    Well, of course there aren't too many quotes in lisp files so this
    problem probably isn't too relevant in practice.

Bye,
Tassilo