On 09/12/2023 16:35, Eli Zaretskii wrote:
>> Date: Sat, 9 Dec 2023 16:13:02 +0200
>> From: Dmitry Gutov<dmitry@gutov.dev>
>>
>> Eli, should we put this in Emacs 29.2? It's not a regression (a fairly
>> old problem), but the fix looks simple enough.
> Yes, please.

Very good, pushed to emacs-29.

Here's a second patch on top which gets rid of the whitelist altogether 
(the new heuristic seems to cover the cases better anyway), though I'm 
on the fence whether this should go here or into master.

It seems to work well with some testing, but it is more adventurous than 
the previous one.