NIIBE Yutaka <gni...@fsij.org> wrote: > I was wrong. The ticket for agent_cache_housekeeping is: > > https://dev.gnupg.org/T3829 > > It was introduced because of some risk keeping passphrase. > > I'd like to consider to improve the implementation of cache and > expiration, not using handle_tick.
In the upstream, I created a ticket for this task. https://dev.gnupg.org/T6681 Then, I did some work. I believe that I somehow sorted out and deal with problems for T3829 and gpg-agent-idling. I'm going to backport the improvement to my branch of tmp-gniibe-v2.4 for Debian. * * * During my implementing the code for T6681, I realized that the possible reason why Christoph disabled use of debian/patches/gpg-agent-idling when he packaged 2.3.1-1; The patches were written in Oct/Nov of 2016. In 2018, to solve an issue of T3829, cache expiration code was added at the handle_tick function. --