bug#58168: string-lessp glitches and inconsistencies

[Eli Zaretskii]

> From: Richard Stallman <rms@gnu.org>
> Cc: mattias.engdegard@gmail.com, 58168@debbugs.gnu.org
> Date: Tue, 04 Oct 2022 13:40:18 -0400
> 
>   > If the fact that string= says strings are not equal, but string-lessp
>   > says they are equal, is what bothers you
> 
> That result seems paradoxical to me.

Not if you accept that comparing unibyte non-ASCII text with multibyte
text yields inherently unspecified results.

> Perhaps documenting the difference between these two relationships
> could make the current behavior comprehensible rather than anomalous.

Yes, that's one of the alternatives that I think should be on the
table.