On Thu, Jun 13, 2019 at 11:28:42AM +0300, Alexander Lakhin wrote:
> Yes, you're right. I've completed the patch for a possible elimination
> of the field.
For now I have discarded this one, and committed the rest as the
inconsistencies stand out. Good catches by the way.
Your patch was actually incorrect in checkpointer.c. 3eb77eb has
refactored the fsync queue and has removed FORGET_DATABASE_FSYNC, but
it has been replaced by SYNC_FILTER_REQUEST as equivalent in the
shared queue to forget database-level stuff.
--
Michael