NIIBE Yutaka <gni...@fsij.org> wrote: > Based on your work of tmp-ametzler-v2.4 branch, I created my own fork. > > My hope is that the migration from 2.4 won't introduce (much) surprise > to Debian users. > > https://salsa.debian.org/gniibe/gnupg2/-/tree/tmp-gniibe-v2.4 > > This work of mine is: > > - Keep systemd support (despite its deprecation in upstream) > - Introduce keyboxd package (with systemd support, for integrity) > - Put backport of BEGIN_ENCRYPTION status output (from master of upstream)
Further, I recover the patches/gpg-agent-idling series. - Enable patches/gpg-agent-idling In the upstream, scd monitoring has improved by introducing a watching thread. I updated the patches to address this change of scd monitoring. Actually, we can improve the code more. On Windows, since parent_pid is always -1 (running command by gpg-agent is not supported), the need_tick function should return 0. And the return value of the need_tick function is determined at initialization, we can clean up the code. Besides, in my opinion, the agent_cache_housekeeping function makes less sense (it's totally OK to only check the expiration on its use). Having expired entries on memory is no problem at all, than running gpg-agent process periodically; memory is cheap but buttery power is not (for my use case). --