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

[Eli Zaretskii <eliz@gnu.org>]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: 52593@debbugs.gnu.org,  Kang Niu <isgniuk@gmail.com>
> Date: Sun, 19 Dec 2021 18:33:37 +0100
> 
> Lars Ingebrigtsen <larsi@gnus.org> writes:
> 
> > To answer my question: Yes.  For instance, `list' may span several lines.
> > So...  Eli's suggestion of adding a no-field parameter here may be
> > solution, unless we somehow rethink the thing-at-point algorithm.
> 
> Another idea: We could change thing-at-point to only respond to
> "positive" fields.  That is, only consider a thing to be a field if
> there's a `field' property under point.
> 
> (insert (propertize "foo" 'field 1) "bar" (propertize "zot" 'field 2))
> 
> This would mean that (symbol-at-point) in the "foo" bit would return
> `foo', but in the "bar" part, it would return `foobarzot'.  (Before
> adding field support, it would return the latter in the first case,
> too.)

I admit that I lack the context here: why was thing-at-point changed
to honor fields? what was that supposed to achieve?