mirror of
https://github.com/scylladb/scylladb.git
synced 2026-04-25 19:10:42 +00:00
That's intended to fix a bad layer violation as table was given the responsibility of disabling compaction for a given table T, but that logic clearly belongs to compaction_manager instead. Additionally, gate will be used instead of counter, as former provides manager with a way to synchronize with functions running under run_with_compaction_disabled. so remove() can wait for their termination. Signed-off-by: Raphael S. Carvalho <raphaelsc@scylladb.com>