mode
- Skip automatic compaction when `data.mdb` is 10GB or larger, and
surface an in-app notification instead of surprising users with a long
startup wait
- Keep manual compaction available for large databases by tracking
whether a next-launch compaction request was scheduled automatically
or explicitly by the user
- Add a large-database explanation to the compaction loading screen to
clarify that long runtimes are expected for oversized databases and
are usually a one-time catch-up cost
- Add a developer-only `Every minute` auto-compact schedule to make
rollout and reminder behavior easier to test locally
- Add and update unit tests covering large-database auto-compaction
skips, request-source persistence, reminder scheduling, and the
developer testing interval
Motivation:
TestFlight users with long-lived databases were hitting the brand-new
compaction flow for the first time, which could turn startup into a
multi-minute wait. These changes make automatic compaction less
disruptive for very large databases, preserve an explicit manual path
for users who want to optimize immediately, and add a fast
developer-only schedule to make the new behavior easier to validate
during testing.
This is an enhancement for an unreleased feature, so therefore no
changelog is needed.
Closes: https://github.com/damus-io/damus/issues/3730
Changelog-None
Signed-off-by: Daniel D’Aquino <daniel@daquino.me>
Daniel D’Aquino
2af65320ef