GNU bug report logs - #42904
[PATCH] Non-Unicode frame title crashes Emacs on macOS

Previous Next

Full log

Message #46 received at 42904 <at> debbugs.gnu.org (full text, mbox):

From: Mattias Engdegård <mattiase <at> acm.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 42904 <at> debbugs.gnu.org, Alan Third <alan <at> idiocy.org>
Subject: Re: bug#42904: [PATCH] Non-Unicode frame title crashes Emacs on macOS
Date: Tue, 18 Aug 2020 16:11:02 +0200

18 aug. 2020 kl. 14.24 skrev Eli Zaretskii <eliz <at> gnu.org>:

> It isn't wrong (and there's no need to worry about alignment in this
> case, AFAIK).

Do you mean that SDATA is guaranteed to be aligned, or that no NS platforms that Emacs runs on (or is likely to run on in the near future, such as macOS on arm64) trap on unaligned?

> Thanks, but I don't think we should modify the :default-char attribute
> of the UTF-* encodings as part of this change.  It's a separate issue,
> and is a backward-incompatible change of sorts.  For instance, we
> should consider what this will do to display on TTY frames that don't
> support Unicode.  So I think we should discuss this issue separately
> before we make such a change.

Yes, we can certainly make it a separate change. All bug fixes are backward-incompatible in some respect; it is not reasonable to depend on non-Unicode characters being translated to spaces when converted to UTF-16 since that is neither documented nor reasonably expected behaviour.

> Why is it a problem to display a space instead of invalid bytes in
> this case?

A problem is not necessary for a change to be desirable. The Unicode replacement character clearly indicates that something could not be encoded correctly, and the exact position for it; it's universally recognised and valuable for users and developers alike. Space is the default value for :default-char, and that it isn't U+FFFD for UTF-16 (or other Unicode encodings) is a clear bug, since that is the correct character to use for that purpose.

My guess is that space was chosen as default because it's a character that occurs in all coding systems, but it is clearly wrong for UTF-16. 'us-ascii' uses '?' for :default-char, which is a better character in that repertoire.

Previous Next

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