avoid uncessary keys db writes; doubled speed!

When running eg git-annex get, for each file it has to read from and
write to the keys database. But it's reading exclusively from one table,
and writing to a different table. So, it is not necessary to flush the
write to the database before reading. This avoids writing the database
once per file, instead it will buffer 1000 changes before writing.

Benchmarking getting 1000 small files from a local origin,
git-annex get now takes 13.62s, down from 22.41s!
git-annex drop now takes 9.07s, down from 18.63s!
Wowowowowowowow!

(It would perhaps have been better if there were separate databases for
the two tables. At least it would have avoided this complexity. Ah well,
this is better than splitting the table in a annex.version upgrade.)

Sponsored-by: Dartmouth College's Datalad project

Database/Keys/Handle.hs

@@ -15,6 +15,7 @@ module Database.Keys.Handle (
 ) where
 
 import qualified Database.Queue as H
+import Database.Keys.Tables
 import Utility.Exception
 import Utility.DebugLocks
 
@@ -29,7 +30,7 @@ newtype DbHandle = DbHandle (MVar DbState)
 
 -- The database can be closed or open, but it also may have been
 -- tried to open (for read) and didn't exist yet or is not readable.
-data DbState = DbClosed | DbOpen H.DbQueue | DbUnavailable
+data DbState = DbClosed | DbOpen (H.DbQueue, DbTablesChanged) | DbUnavailable
 
 newDbHandle :: IO DbHandle
 newDbHandle = DbHandle <$> newMVar DbClosed
@@ -52,15 +53,17 @@ withDbState (DbHandle mvar) a = do
 		return v
 
 flushDbQueue :: DbHandle -> IO ()
-flushDbQueue (DbHandle mvar) = go =<< debugLocks (readMVar mvar)
+flushDbQueue h = withDbState h go
   where
-	go (DbOpen qh) = H.flushDbQueue qh
-	go _ = return ()
+	go (DbOpen (qh, _)) = do
+		H.flushDbQueue qh
+		return ((), DbOpen (qh, mempty))
+	go st = return ((), st)
 
 closeDbHandle :: DbHandle -> IO ()
 closeDbHandle h = withDbState h go
   where
-	go (DbOpen qh) = do
+	go (DbOpen (qh, _)) = do
 		H.closeDbQueue qh
 		return ((), DbClosed)
 	go st = return ((), st)