bug#58168: string-lessp glitches and inconsistencies

[Eli Zaretskii]

> From: Mattias Engdegård <mattias.engdegard@gmail.com>
> Date: Fri, 14 Oct 2022 16:39:55 +0200
> Cc: 58168@debbugs.gnu.org
> 
> As performance is more acceptable now I'm not going to take any further 
> action with respect of string<, but let me just answer your questions:
> 
> 8 okt. 2022 kl. 09.35 skrev Eli Zaretskii <eliz@gnu.org>:
> 
> > I suggested to use memchr to find whether a string has any
> > C0 or C1 bytes, _before_ doing the actual comparison, to find out
> > whether a multibyte string includes any raw bytes, which would then
> > require slower comparisons.
> 
> That isn't practical; we would traverse each argument in full, twice, even if 
> there is a difference early on. While memchr is fast for what it does, it 
> will still need to look at every bit of its input.

I fail to see how the number of times we'd traverse the strings is of
concern, as long as it's fast enough.  And memchr _is_ very fast.