as outlined in bug #77069, a stream filter may loose (aka. skip) the
final data block. That bug report has some duplicates: #48725, #79984
Several weaks ago I made PR #6001 which would fix this issue, but I'm
not sure whether that change would violate the assumption that the
PSFS_FLAG_FLUSH_CLOSE flag will only ever be passed if no further data
is to be processed.
Could someone more proficient with the streams layer than I am please
clarify whether that is a valid assumption in the first place?