bug#69709: `sort` interface improvement and universal ordering predicate

[Dmitry Gutov]

On 10/03/2024 17:56, Mattias Engdegård wrote:
> 10 mars 2024 kl. 16.48 skrev Dmitry Gutov<dmitry@gutov.dev>:
>
> > Here's a concern: a lot of existing code is written with either mutability in 
> > mind (the source list is a throwaway one, owned by the caller), or coupled with 
> > copy-sequence already.
> >
> > If the new 'sort' default to non-destructive, wouldn't that make those existing 
> > callsites slower? Or at least some of them.
> No, with the old calling convention they would get destructive (in-place) 
> sorting. There should be no incompatibilities at all.

I see. That sounds more confusing, though (you drop the 'pred' argument, 
and the function changes behavior, possibly becoming slower too).