mongodb/Database/MongoDB/Internal/Protocol.hs
2011-07-22 10:38:50 -04:00

336 lines
12 KiB
Haskell
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

{-| Low-level messaging between this client and the MongoDB server, see Mongo Wire Protocol (<http://www.mongodb.org/display/DOCS/Mongo+Wire+Protocol>).
This module is not intended for direct use. Use the high-level interface at "Database.MongoDB.Query" and "Database.MongoDB.Connection" instead. -}
{-# LANGUAGE RecordWildCards, StandaloneDeriving, OverloadedStrings, FlexibleContexts, TupleSections, TypeSynonymInstances, MultiParamTypeClasses, FlexibleInstances, UndecidableInstances #-}
module Database.MongoDB.Internal.Protocol (
FullCollection,
-- * Pipe
Pipe, newPipe, send, call,
-- ** Notice
Notice(..), InsertOption(..), UpdateOption(..), DeleteOption(..), CursorId,
-- ** Request
Request(..), QueryOption(..),
-- ** Reply
Reply(..), ResponseFlag(..),
-- * Authentication
Username, Password, Nonce, pwHash, pwKey
) where
import Prelude as X
import Control.Applicative ((<$>))
import Control.Arrow ((***))
import Data.ByteString.Lazy as B (length, hPut)
import System.IO.Pipeline (IOE, Pipeline, newPipeline, IOStream(..))
import qualified System.IO.Pipeline as P (send, call)
import System.IO (Handle, hClose)
import Data.Bson (Document, UString)
import Data.Bson.Binary
import Data.Binary.Put
import Data.Binary.Get
import Data.Int
import Data.Bits
import Data.IORef
import System.IO.Unsafe (unsafePerformIO)
import qualified Crypto.Hash.MD5 as MD5 (hash)
import Data.UString as U (pack, append, toByteString)
import System.IO.Error as E (try)
import Control.Monad.Error
import System.IO (hFlush)
import Database.MongoDB.Internal.Util (whenJust, hGetN, bitOr, byteStringHex)
-- * Pipe
type Pipe = Pipeline Response Message
-- ^ Thread-safe TCP connection with pipelined requests
newPipe :: Handle -> IO Pipe
-- ^ Create pipe over handle
newPipe handle = newPipeline $ IOStream (writeMessage handle) (readMessage handle) (hClose handle)
send :: Pipe -> [Notice] -> IOE ()
-- ^ Send notices as a contiguous batch to server with no reply. Throw IOError if connection fails.
send pipe notices = P.send pipe (notices, Nothing)
call :: Pipe -> [Notice] -> Request -> IOE (IOE Reply)
-- ^ Send notices and request as a contiguous batch to server and return reply promise, which will block when invoked until reply arrives. This call and resulting promise will throw IOError if connection fails.
call pipe notices request = do
requestId <- genRequestId
promise <- P.call pipe (notices, Just (request, requestId))
return $ check requestId <$> promise
where
check requestId (responseTo, reply) = if requestId == responseTo then reply else
error $ "expected response id (" ++ show responseTo ++ ") to match request id (" ++ show requestId ++ ")"
-- * Message
type Message = ([Notice], Maybe (Request, RequestId))
-- ^ A write notice(s) with getLastError request, or just query request.
-- Note, that requestId will be out of order because request ids will be generated for notices after the request id supplied was generated. This is ok because the mongo server does not care about order just uniqueness.
writeMessage :: Handle -> Message -> IOE ()
-- ^ Write message to socket
writeMessage handle (notices, mRequest) = ErrorT . E.try $ do
forM_ notices $ \n -> writeReq . (Left n,) =<< genRequestId
whenJust mRequest $ writeReq . (Right *** id)
hFlush handle
where
writeReq (e, requestId) = do
hPut handle lenBytes
hPut handle bytes
where
bytes = runPut $ (either putNotice putRequest e) requestId
lenBytes = encodeSize . toEnum . fromEnum $ B.length bytes
encodeSize = runPut . putInt32 . (+ 4)
type Response = (ResponseTo, Reply)
-- ^ Message received from a Mongo server in response to a Request
readMessage :: Handle -> IOE Response
-- ^ read response from socket
readMessage handle = ErrorT $ E.try readResp where
readResp = do
len <- fromEnum . decodeSize <$> hGetN handle 4
runGet getReply <$> hGetN handle len
decodeSize = subtract 4 . runGet getInt32
type FullCollection = UString
-- ^ Database name and collection name with period (.) in between. Eg. \"myDb.myCollection\"
-- ** Header
type Opcode = Int32
type RequestId = Int32
-- ^ A fresh request id is generated for every message
type ResponseTo = RequestId
genRequestId :: (MonadIO m) => m RequestId
-- ^ Generate fresh request id
genRequestId = liftIO $ atomicModifyIORef counter $ \n -> (n + 1, n) where
counter :: IORef RequestId
counter = unsafePerformIO (newIORef 0)
{-# NOINLINE counter #-}
-- *** Binary format
putHeader :: Opcode -> RequestId -> Put
-- ^ Note, does not write message length (first int32), assumes caller will write it
putHeader opcode requestId = do
putInt32 requestId
putInt32 0
putInt32 opcode
getHeader :: Get (Opcode, ResponseTo)
-- ^ Note, does not read message length (first int32), assumes it was already read
getHeader = do
_requestId <- getInt32
responseTo <- getInt32
opcode <- getInt32
return (opcode, responseTo)
-- ** Notice
-- | A notice is a message that is sent with no reply
data Notice =
Insert {
iFullCollection :: FullCollection,
iOptions :: [InsertOption],
iDocuments :: [Document]}
| Update {
uFullCollection :: FullCollection,
uOptions :: [UpdateOption],
uSelector :: Document,
uUpdater :: Document}
| Delete {
dFullCollection :: FullCollection,
dOptions :: [DeleteOption],
dSelector :: Document}
| KillCursors {
kCursorIds :: [CursorId]}
deriving (Show, Eq)
data InsertOption = KeepGoing -- ^ If set, the database will not stop processing a bulk insert if one fails (eg due to duplicate IDs). This makes bulk insert behave similarly to a series of single inserts, except lastError will be set if any insert fails, not just the last one. (new in 1.9.1)
deriving (Show, Eq)
data UpdateOption =
Upsert -- ^ If set, the database will insert the supplied object into the collection if no matching document is found
| MultiUpdate -- ^ If set, the database will update all matching objects in the collection. Otherwise only updates first matching doc
deriving (Show, Eq)
data DeleteOption = SingleRemove -- ^ If set, the database will remove only the first matching document in the collection. Otherwise all matching documents will be removed
deriving (Show, Eq)
type CursorId = Int64
-- *** Binary format
nOpcode :: Notice -> Opcode
nOpcode Update{} = 2001
nOpcode Insert{} = 2002
nOpcode Delete{} = 2006
nOpcode KillCursors{} = 2007
putNotice :: Notice -> RequestId -> Put
putNotice notice requestId = do
putHeader (nOpcode notice) requestId
case notice of
Insert{..} -> do
putInt32 (iBits iOptions)
putCString iFullCollection
mapM_ putDocument iDocuments
Update{..} -> do
putInt32 0
putCString uFullCollection
putInt32 (uBits uOptions)
putDocument uSelector
putDocument uUpdater
Delete{..} -> do
putInt32 0
putCString dFullCollection
putInt32 (dBits dOptions)
putDocument dSelector
KillCursors{..} -> do
putInt32 0
putInt32 $ toEnum (X.length kCursorIds)
mapM_ putInt64 kCursorIds
iBit :: InsertOption -> Int32
iBit KeepGoing = bit 0
iBits :: [InsertOption] -> Int32
iBits = bitOr . map iBit
uBit :: UpdateOption -> Int32
uBit Upsert = bit 0
uBit MultiUpdate = bit 1
uBits :: [UpdateOption] -> Int32
uBits = bitOr . map uBit
dBit :: DeleteOption -> Int32
dBit SingleRemove = bit 0
dBits :: [DeleteOption] -> Int32
dBits = bitOr . map dBit
-- ** Request
-- | A request is a message that is sent with a 'Reply' expected in return
data Request =
Query {
qOptions :: [QueryOption],
qFullCollection :: FullCollection,
qSkip :: Int32, -- ^ Number of initial matching documents to skip
qBatchSize :: Int32, -- ^ The number of document to return in each batch response from the server. 0 means use Mongo default. Negative means close cursor after first batch and use absolute value as batch size.
qSelector :: Document, -- ^ \[\] = return all documents in collection
qProjector :: Document -- ^ \[\] = return whole document
} | GetMore {
gFullCollection :: FullCollection,
gBatchSize :: Int32,
gCursorId :: CursorId}
deriving (Show, Eq)
data QueryOption =
TailableCursor -- ^ Tailable means cursor is not closed when the last data is retrieved. Rather, the cursor marks the final object's position. You can resume using the cursor later, from where it was located, if more data were received. Like any "latent cursor", the cursor may become invalid at some point for example if the final object it references were deleted. Thus, you should be prepared to requery on CursorNotFound exception.
| SlaveOK -- ^ Allow query of replica slave. Normally these return an error except for namespace "local".
| NoCursorTimeout -- ^ The server normally times out idle cursors after 10 minutes to prevent a memory leak in case a client forgets to close a cursor. Set this option to allow a cursor to live forever until it is closed.
| AwaitData -- ^ Use with TailableCursor. If we are at the end of the data, block for a while rather than returning no data. After a timeout period, we do return as normal.
-- | Exhaust -- ^ Stream the data down full blast in multiple "more" packages, on the assumption that the client will fully read all data queried. Faster when you are pulling a lot of data and know you want to pull it all down. Note: the client is not allowed to not read all the data unless it closes the connection.
-- Exhaust commented out because not compatible with current `Pipeline` implementation
| Partial -- ^ Get partial results from a _mongos_ if some shards are down, instead of throwing an error.
deriving (Show, Eq)
-- *** Binary format
qOpcode :: Request -> Opcode
qOpcode Query{} = 2004
qOpcode GetMore{} = 2005
putRequest :: Request -> RequestId -> Put
putRequest request requestId = do
putHeader (qOpcode request) requestId
case request of
Query{..} -> do
putInt32 (qBits qOptions)
putCString qFullCollection
putInt32 qSkip
putInt32 qBatchSize
putDocument qSelector
unless (null qProjector) (putDocument qProjector)
GetMore{..} -> do
putInt32 0
putCString gFullCollection
putInt32 gBatchSize
putInt64 gCursorId
qBit :: QueryOption -> Int32
qBit TailableCursor = bit 1
qBit SlaveOK = bit 2
qBit NoCursorTimeout = bit 4
qBit AwaitData = bit 5
--qBit Exhaust = bit 6
qBit Partial = bit 7
qBits :: [QueryOption] -> Int32
qBits = bitOr . map qBit
-- ** Reply
-- | A reply is a message received in response to a 'Request'
data Reply = Reply {
rResponseFlags :: [ResponseFlag],
rCursorId :: CursorId, -- ^ 0 = cursor finished
rStartingFrom :: Int32,
rDocuments :: [Document]
} deriving (Show, Eq)
data ResponseFlag =
CursorNotFound -- ^ Set when getMore is called but the cursor id is not valid at the server. Returned with zero results.
| QueryError -- ^ Query error. Returned with one document containing an "$err" field holding the error message.
| AwaitCapable -- ^ For backward compatability: Set when the server supports the AwaitData query option. if it doesn't, a replica slave client should sleep a little between getMore's
deriving (Show, Eq, Enum)
-- * Binary format
replyOpcode :: Opcode
replyOpcode = 1
getReply :: Get (ResponseTo, Reply)
getReply = do
(opcode, responseTo) <- getHeader
unless (opcode == replyOpcode) $ fail $ "expected reply opcode (1) but got " ++ show opcode
rResponseFlags <- rFlags <$> getInt32
rCursorId <- getInt64
rStartingFrom <- getInt32
numDocs <- fromIntegral <$> getInt32
rDocuments <- replicateM numDocs getDocument
return (responseTo, Reply{..})
rFlags :: Int32 -> [ResponseFlag]
rFlags bits = filter (testBit bits . rBit) [CursorNotFound ..]
rBit :: ResponseFlag -> Int
rBit CursorNotFound = 0
rBit QueryError = 1
rBit AwaitCapable = 3
-- * Authentication
type Username = UString
type Password = UString
type Nonce = UString
pwHash :: Username -> Password -> UString
pwHash u p = pack . byteStringHex . MD5.hash . toByteString $ u `U.append` ":mongo:" `U.append` p
pwKey :: Nonce -> Username -> Password -> UString
pwKey n u p = pack . byteStringHex . MD5.hash . toByteString . U.append n . U.append u $ pwHash u p
{- Authors: Tony Hannan <tony@10gen.com>
Copyright 2011 10gen Inc.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at: http://www.apache.org/licenses/LICENSE-2.0. Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. -}