Re: master 52e527a02f 2/2: Make ?\LF generate 10, not -1 (bug#55738)

[Mattias Engdegård]

3 juni 2022 kl. 03.22 skrev Stefan Monnier <monnier@iro.umontreal.ca>:

>>    Make ?\LF generate 10, not -1 (bug#55738)
> 
> I think I'd prefer if it signaled an error.

I would, too -- done. There is no way such code could be anything other than a 
mistake.
And look what turns up during bootstrapping:

(defvar idlwave-comment-indent-char ?\

The idlwave.el history provides some entertainment:

f32b3b911bb Carsten Dominik    1999-12-20  ?\SPC   (initial commit)
775591f7677 Juanma Barranquero 2002-10-18  ?\SPC -> ?\LF
93af03f0a1c Markus Rost        2003-07-21  ?\LF  -> ?\s
52a244ebe2b J.D. Smith         2004-11-17  ?\s   -> ?\SPC
b6a97790db0 Juanma Barranquero 2005-06-16  ?\SPC -> ?\LF
3938cb82d0f J.D. Smith         2005-07-04  ?\LF  -> ?\SPC
31b58798cba Juanma Barranquero 2005-07-04  ?\SPC -> ?\LF
f66f03dec52 J.D. Smith         2006-04-06  ?\LF  -> ?\SPC
4b1aaa8b07c Paul Eggert        2006-09-12  ?\SPC -> ?\LF

Not only is this a strong argument supporting your preference for an error 
here, it's also an indication that automatically removing trailing spaces isn't 
necessarily a wonderful idea (I never do that myself).

(This particular variable hasn't been used for some time so the damage is 
limited in this case.)