{-# LANGUAGE ConstraintKinds #-} {-| Module : Database.LMDB.Simple Description : Simple Haskell API for LMDB Copyright : © 2017–2018 Robert Leslie License : BSD3 Maintainer : rob@mars.org Stability : experimental This module provides a simple Haskell API for the (LMDB). Example usage: @ module Main where import Database.LMDB.Simple import Control.Monad (forM_) main = do env <- openEnvironment "myenv" defaultLimits db <- readOnlyTransaction env $ getDatabase Nothing :: IO (Database String Int) transaction env $ forM_ [("one",1),("two",2),("three",3)] $ \\(k,v) -> put db k (Just v) print =<< readOnlyTransaction env (get db "two") -- Just 2 print =<< readOnlyTransaction env (get db "nine") -- Nothing @ These additional APIs are available: * "Database.LMDB.Simple.Extra" provides additional functions for querying and modifying LMDB databases from within the 'Transaction' monad * "Database.LMDB.Simple.View" provides a read-only snapshot of an LMDB database that can be accessed from pure code * "Database.LMDB.Simple.DBRef" provides a mutable variable (accessed from 'IO') that is tied to a particular key that persists in an LMDB database -} module Database.LMDB.Simple ( -- * Environments Environment , Limits (..) , defaultLimits , openEnvironment , openReadWriteEnvironment , openReadOnlyEnvironment , readOnlyEnvironment , clearStaleReaders -- * Transactions , Transaction , transaction , readWriteTransaction , readOnlyTransaction , nestTransaction , abort , AbortedTransaction -- * Databases , Database , getDatabase , get , put , clear -- * Access modes , ReadWrite , ReadOnly , Mode , SubMode ) where import Control.Concurrent ( runInBoundThread ) import Control.Exception ( Exception , throwIO , try , tryJust , bracketOnError ) import Control.Monad ( guard , void ) import Data.Coerce ( coerce ) import Database.LMDB.Raw ( LMDB_Error (LMDB_Error, e_code) , MDB_EnvFlag (MDB_NOSUBDIR, MDB_RDONLY) , MDB_DbFlag (MDB_CREATE) , mdb_env_create , mdb_env_open , mdb_env_set_mapsize , mdb_env_set_maxdbs , mdb_env_set_maxreaders , mdb_dbi_open' , mdb_txn_begin , mdb_txn_commit , mdb_txn_abort , mdb_txn_env , mdb_clear' , mdb_reader_check ) import Database.LMDB.Simple.Internal ( ReadWrite , ReadOnly , Mode , SubMode , Environment (Env) , Transaction (Txn) , Database (Db) , Serialise , isReadOnlyEnvironment , isReadOnlyTransaction , isReadWriteTransaction ) import qualified Database.LMDB.Simple.Internal as Internal import Foreign.C ( Errno (Errno) , eNOTDIR ) -- | LMDB environments have various limits on the size and number of databases -- and concurrent readers. data Limits = Limits { mapSize :: Int -- ^ memory map size, in bytes (also the maximum size -- of all databases) , maxDatabases :: Int -- ^ maximum number of named databases , maxReaders :: Int -- ^ maximum number of concurrent 'ReadOnly' -- transactions (also the number of slots in the lock -- table) } -- | The default limits are 1 MiB map size, 0 named databases, and 126 -- concurrent readers. These can be adjusted freely, and in particular the -- 'mapSize' may be set very large (limited only by available address -- space). However, LMDB is not optimized for a large number of named -- databases so 'maxDatabases' should be kept to a minimum. -- -- The default 'mapSize' is intentionally small, and should be changed to -- something appropriate for your application. It ought to be a multiple of -- the OS page size, and should be chosen as large as possible to accommodate -- future growth of the database(s). Once set for an environment, this limit -- cannot be reduced to a value smaller than the space already consumed by the -- environment, however it can later be increased. -- -- If you are going to use any named databases then you will need to change -- 'maxDatabases' to the number of named databases you plan to use. However, -- you do not need to change this field if you are only going to use the -- single main (unnamed) database. defaultLimits :: Limits defaultLimits = Limits { mapSize = 1024 * 1024 -- 1 MiB , maxDatabases = 0 , maxReaders = 126 } -- | Open an LMDB environment in either 'ReadWrite' or 'ReadOnly' mode. The -- 'FilePath' argument may be either a directory or a regular file, but it -- must already exist. If a regular file, an additional file with "-lock" -- appended to the name is used for the reader lock table. -- -- Note that an environment must have been opened in 'ReadWrite' mode at least -- once before it can be opened in 'ReadOnly' mode. -- -- An environment opened in 'ReadOnly' mode may still modify the reader lock -- table (except when the filesystem is read-only, in which case no locks are -- used). openEnvironment :: Mode mode => FilePath -> Limits -> IO (Environment mode) openEnvironment path limits = do env <- mdb_env_create mdb_env_set_mapsize env (mapSize limits) mdb_env_set_maxdbs env (maxDatabases limits) mdb_env_set_maxreaders env (maxReaders limits) let environ = Env env :: Mode mode => Environment mode flags = [MDB_RDONLY | isReadOnlyEnvironment environ] r <- tryJust (guard . isNotDirectoryError) $ mdb_env_open env path flags case r of Left _ -> mdb_env_open env path (MDB_NOSUBDIR : flags) Right _ -> return () return environ where isNotDirectoryError :: LMDB_Error -> Bool isNotDirectoryError LMDB_Error { e_code = Left code } | Errno (fromIntegral code) == eNOTDIR = True isNotDirectoryError _ = False -- | Convenience function for opening an LMDB environment in 'ReadWrite' -- mode; see 'openEnvironment' openReadWriteEnvironment :: FilePath -> Limits -> IO (Environment ReadWrite) openReadWriteEnvironment = openEnvironment -- | Convenience function for opening an LMDB environment in 'ReadOnly' -- mode; see 'openEnvironment' openReadOnlyEnvironment :: FilePath -> Limits -> IO (Environment ReadOnly) openReadOnlyEnvironment = openEnvironment -- | For an LMDB environment opened in 'ReadWrite' mode, return the same -- environment restricted to 'ReadOnly' mode. This can be used to prevent -- modification of the environment for a subset of an application while -- retaining the ability to perform modifications in another subset. -- -- @since 0.4.0.0 readOnlyEnvironment :: Environment ReadWrite -> Environment ReadOnly readOnlyEnvironment = coerce -- | Check for stale entries in the reader lock table, and return the number -- of entries cleared. clearStaleReaders :: Environment mode -> IO Int clearStaleReaders (Env env) = mdb_reader_check env -- | Perform a top-level transaction in either 'ReadWrite' or 'ReadOnly' -- mode. A transaction may only be 'ReadWrite' if the environment is also -- 'ReadWrite' (enforced by the 'SubMode' constraint). -- -- Once completed, the transaction will be committed and the result -- returned. An exception will cause the transaction to be implicitly aborted. -- -- Note that there may be several concurrent 'ReadOnly' transactions (limited -- only by the 'maxReaders' field of the 'Limits' argument given to -- 'openEnvironment'), but there is at most one active 'ReadWrite' -- transaction, which is forced to run in a bound thread, and is protected by -- an internal mutex. -- -- In general, long-lived transactions should be avoided. 'ReadOnly' -- transactions prevent reuse of database pages freed by newer 'ReadWrite' -- transactions, and thus the database can grow quickly. 'ReadWrite' -- transactions prevent other 'ReadWrite' transactions, since writes are -- serialized. transaction :: (Mode tmode, SubMode emode tmode) => Environment emode -> Transaction tmode a -> IO a transaction (Env env) tx@(Txn tf) | isReadOnlyTransaction tx = run True | otherwise = runInBoundThread (run False) where run readOnly = bracketOnError (mdb_txn_begin env Nothing readOnly) mdb_txn_abort $ \txn -> tf txn >>= \result -> mdb_txn_commit txn >> return result -- | The exception type thrown when a (top-level) transaction is explicitly -- aborted data AbortedTransaction = AbortedTransaction instance Show AbortedTransaction where showsPrec _ AbortedTransaction = showString "aborted transaction" instance Exception AbortedTransaction -- | Convenience function for performing a top-level 'ReadWrite' transaction; -- see 'transaction' readWriteTransaction :: Environment ReadWrite -> Transaction ReadWrite a -> IO a readWriteTransaction = transaction -- | Convenience function for performing a top-level 'ReadOnly' transaction; -- see 'transaction' readOnlyTransaction :: Environment mode -> Transaction ReadOnly a -> IO a readOnlyTransaction = transaction -- | Nest a transaction within the current 'ReadWrite' transaction. -- Transactions may be nested to any level. -- -- If the nested transaction is aborted, 'Nothing' is returned. Otherwise, the -- nested transaction is committed and the result is returned in a 'Just' -- value. (The overall effect of a nested transaction depends on whether the -- parent transaction is ultimately committed.) -- -- An exception will cause the nested transaction to be implicitly aborted. nestTransaction :: Transaction ReadWrite a -> Transaction ReadWrite (Maybe a) nestTransaction (Txn tf) = Txn $ \ptxn -> let env = mdb_txn_env ptxn in maybeAborted $ bracketOnError (mdb_txn_begin env (Just ptxn) False) mdb_txn_abort $ \ctxn -> tf ctxn >>= \result -> mdb_txn_commit ctxn >> return result where maybeAborted :: IO a -> IO (Maybe a) maybeAborted io = either (\e -> let _ = e :: AbortedTransaction in Nothing) Just <$> try io -- | Explicitly abort the current transaction, nullifying its effects on the -- LMDB environment. No further actions will be performed within the -- transaction. -- -- In a nested transaction, this causes the child transaction to return -- 'Nothing' to its parent. In a top-level transaction, this throws an -- 'AbortedTransaction' exception, which can be caught. abort :: Transaction mode a abort = Txn $ \_ -> throwIO AbortedTransaction -- | Retrieve a database handle from the LMDB environment. The database may be -- specified by name, or 'Nothing' can be used to specify the main (unnamed) -- database for the environment. If a named database is specified, it must -- already exist, or it will be created if the transaction is 'ReadWrite'. -- -- There are a limited number of named databases you may use in an -- environment, set by the 'maxDatabases' field of the 'Limits' argument given -- to 'openEnvironment'. By default ('defaultLimits') this number is zero, so -- you must specify another limit in order to use any named databases. -- -- You should not use both named and unnamed databases in the same -- environment, because the unnamed database is used internally to store -- entries for each named database. -- -- You can (and should) retain the database handle returned by this action for -- use in future transactions. getDatabase :: Mode mode => Maybe String -> Transaction mode (Database k v) getDatabase name = tx where tx = Txn $ \txn -> Db (mdb_txn_env txn) <$> mdb_dbi_open' txn name flags flags = [MDB_CREATE | isReadWriteTransaction tx] -- | Lookup a key in a database and return the corresponding value, or return -- 'Nothing' if the key does not exist in the database. get :: (Serialise k, Serialise v) => Database k v -> k -> Transaction mode (Maybe v) get = Internal.get -- | Insert the given key/value pair into a database, or delete the key from -- the database if 'Nothing' is given for the value. put :: (Serialise k, Serialise v) => Database k v -> k -> Maybe v -> Transaction ReadWrite () put db key = maybe (void $ Internal.delete db key) (Internal.put db key) -- | Delete all key/value pairs from a database, leaving the database empty. clear :: Database k v -> Transaction ReadWrite () clear (Db _ dbi) = Txn $ \txn -> mdb_clear' txn dbi