Re: [PATCH 3/3] mktime: improve tm_isdst heuristic

[Paul Eggert]

On 2025-01-04 19:23, Bruno Haible wrote:
Does this revert the change from
https://git.savannah.gnu.org/gitweb/? 
p=gnulib.git;a=commitdiff;h=836c32a5915826c529b310981a149050bdd9db14
that was done for the sake of 'timegm'?
Or is that change now implemented in a different way?

It's now implemented in a different way: since gmtime has no timestamps 
with tm_isdst!=0, an incoming positive tm_isdst is now ignored.