bug#81606: Mojibake with UTF-8 file names

> From: Arash Esbati <[email protected]>
> Cc: [email protected],  [email protected]
> Date: Sun, 16 Aug 2026 13:26:08 +0200
> 
> Eli Zaretskii <[email protected]> writes:
> 
> > So this can only work if the encoding of the main file's name is
> > easily determined from its byte sequence.  The patch below tries to
> > do the best we can.  It is still not 100% reliable: if the default
> > locale's encoding is not UTF-8, there are chances it will still not
> > work.  But I don't think there's a solution in that case, except by
> > providing the \inputencoding directive or the coding: cookie.
> 
> I think the above is a safe assumption.  This is an excerpt from LaTeX
> news for release 2018-04-01:
> 
>   UTF-8: the new default input encoding
>   [...]
>   With this release, the default encoding for LaTeX files has been
>   changed from the ``fall through raw'' encoding to UTF-8 if used with
>   classic TeX or pdfTeX. The implementation is essentially the same as
>   the existing UTF-8 support from \usepackage[utf8]{inputenc}.  The
>   LuaTeX and XeTeX engines always supported the UTF-8 encoding as their
>   native (and only) input encoding, so with these engines inputenc was
>   always a no-op.
> 
>   This means that with new documents one can assume UTF-8 input and it
>   is no longer required to always specify
>   \usepackage[utf8]{inputenc}. But if this line is present it will not
>   hurt either.  [...]
> 
>   Only documents that have been stored in a legacy encoding and used
>   accented letters from the keyboard without loading inputenc (relying
>   on the similarities between the input used and the T1 font encoding)
>   are affected.
> 
>   These documents will now generate an error that they contain invalid
>   UTF-8 sequences.  [...]
> 
>   Possible alternatives are reencoding the file to UTF-8 using a tool
>   (such as recode or iconv or an editor) or adding the line
>   \usepackage[<encoding>]{inputenc}
>   to the preamble specifying the <encoding> that fits the file encoding.
> 
> IIUC, LaTeX users with old documents are now forced to touch their
> files, and with your patch below, it should cover most cases, I think.

Once again: if the default encoding is not UTF-8, what I suggest is
not reliable enough.  For example, on MS-Windows it doesn't always
work, unless I invoke prefer-coding-system first to prefer UTF-8.

I don't know how widely the above convention is accepted nowadays, but
if it's wide enough, perhaps replace 'undecided' with 'utf-8' in the
decode-coding-string call.

> > Here's the patch I can suggest:
> >
> > diff --git a/lisp/international/latexenc.el b/lisp/international/latexenc.el
> > index 6e0cde6..1e0f8e7 100644
> > --- a/lisp/international/latexenc.el
> > +++ b/lisp/international/latexenc.el
> > @@ -160,6 +160,8 @@ latexenc-find-file-coding-system
> >                                       (stringp TeX-master)
> >                                       TeX-master)
> >                                  (bound-and-true-p tex-main-file)))))
> > +                ;; Decode the main file name as best as we can...
> > +                (setq file (decode-coding-string file 'undecided))
> >                  (dolist (ext `("" ,(if (boundp 'TeX-default-extension)
> >                                         (concat "." TeX-default-extension)
> >                                       "")
> 
> Thanks, I tested it and it works.

So should I install the above, or should I install the utf-8 variant?
I'm not a user of AUCTeX and don't edit LaTeX files, so I cannot
reason about this myself.