bug#52593: 28.0.90; (thing-at-point thing) has so much overhead since commit 7db376e560448e61485ba054def8c82b21f33d6a

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> So basically, the "thing" should have the field property on its
> characters?  And testing for that property at point, and refraining
> from narrowing if there's no such property, still leaves the code too
> slow?

No, if we do that, it should be plenty fast.  But it's slightly
different from what `(goto-char field-beginning)' (for instance) does --
if we're not in a field, it'll take us to the end of the previous field.
(I.e., it considers a section of text that has no `field' parameter to
be a field in itself.)

But I think we can just document that small quirk in `thing-at-point'
and still get the intended `field' action in most things...  perhaps.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no