Re: new-flex-completion-style

[Stefan Monnier]

> Yep, and does the right thing default-wise, too.  If I'm not mistaken,
> I'd say it behaves exactly like ido-switch-to-buffer now, which doesn't
> use score-based sorting, when icomplete-mode is used with the
> new flex completion style.

;-)

> BTW, has the council of completion elders reached any sort of
> white-smoke decision vis-à-vis where to score-sort?

It's been too foggy around here lately to see much smoke, sorry.

>> I think a more precise description would be that cycle-sort-function now
>> also overrides the use of minibuffer-history-variable.
> That's in the commit message's subject line, but OK I've changed it.

Who reads the first line anyway?
[ After reading that first line: ] but that didn't make it clear that
the change is about the use of minibuffer-history-variable (the only
part which used to prevail over cycle-sort-function, AFAICT).


        Stefan