Revert "validation: Only run BlockTx validation when DB fully synced"

If the validation runs grindingly slow, then it is likely that
the generation of table indices is missing.

This reverts commit 1a8dbc9.

cardano-db/app/Cardano/Db/App/Validate/BlockTxs.hs

@@ -24,11 +24,10 @@ import qualified System.Random as Random
 
 validateEpochBlockTxs :: IO ()
 validateEpochBlockTxs = do
-  fullySynced <- runDbNoLogging queryIsFullySynced
   mLatestEpoch <- runDbNoLogging queryLatestCachedEpochNo
   case mLatestEpoch of
     Nothing -> putStrLn "Epoch table is empty"
-    Just latest -> validateLatestBlockTxs fullySynced latest
+    Just latest -> validateLatestBlockTxs latest
 
 -- -----------------------------------------------------------------------------
 
@@ -38,15 +37,10 @@ data ValidateError = ValidateError
   , veTxCountExpected :: !Word64
   }
 
-validateLatestBlockTxs :: Bool -> Word64 -> IO ()
-validateLatestBlockTxs fullySynced latestEpoch = do
-  if not fullySynced
-    then putStrLn "Not fully synced so not running BlockTx validation"
-    else do
-      -- This validation seems to be quite DB intensive, so only run it
-      -- when the DB is fully synced.
-      validateBlockTxs latestEpoch
-      validateBlockTxs =<< Random.randomRIO (0, latestEpoch - 1)
+validateLatestBlockTxs :: Word64 -> IO ()
+validateLatestBlockTxs latestEpoch = do
+  validateBlockTxs latestEpoch
+  validateBlockTxs =<< Random.randomRIO (0, latestEpoch - 1)
 
 validateBlockTxs :: Word64 -> IO ()
 validateBlockTxs epoch = do