[ moving to libraries@haskell.org from glasgow-haskell-users@haskell.org ] Bulat Ziganshin wrote:
Wednesday, April 19, 2006, 4:45:19 PM, you wrote:
Believe me I've looked in detail at your streams library. Performance-wise it is great but the design needs to be reworked IMO.
The main problem is that it doesn't have enough type structure. There are many combinations of stream transformers that don't make sense, and should therefore be ruled out by the type system. There are operations that don't work on some streams. There should at the least be a type distinction between directly accessible memory streams, byte streams, and text streams. Additionally I would add separate classes for seekable and buffered streams. I believe these changes would improve performance by reducing the size of dictionaries.
you have written this in February, but this discussion was not finished due to my laziness. now i tried to split Stream interface to several parts. so
I've attached a sketched design. It doesn't compile, but it illustrates the structure I have in mind. The main improvement since the new-io library is the addition of memory streams. This is an idea from your library and I like it a lot, although I changed the type of the methods: -- | An input stream accessed directly via a memory buffer. -- Ordinary 'InputStream's may be converted to 'MemInputStream's by -- adding buffering; see 'bufferInputStream'. class MemInputStream s where -- | Consume some bytes from the memory stream. The second argument -- is an IO action that is passed a buffer and its size (the size must -- be non-zero), and it should return the number of bytes consumed. withStreamInputBuffer :: s -> (Ptr Word8 -> Int -> IO Int) -> IO () -- | An output stream accessed directly via a memory buffer. -- Ordinary 'OutputStream's may be converted to 'MemOutputStream's by -- adding buffering; see 'bufferOutputStream'. class MemOutputStream s where -- | Write some bytes to a memory stream. The second argument -- is an IO action that is passed a buffer and its size (the size must -- be non-zero), and it should return the number of bytes written. withStreamOutputBuffer :: s -> (Ptr Word8 -> Int -> IO Int) -> IO ()
1) that you think - Stream should be base for all other stream classes or each Stream class should be independent? i.e.
Superclasses aren't necessary, but they might help to reduce the size of contexts in practice. We should probably experiment with both.
2) separation of Stream classes make some automatic definitions impossible. for example, released version contains vGetBuf implementation that is defined via vGetChar and works ok for streams that provide only vGetChar as base function.
No class should implement both reading bytes and reading chars. Encoding/decoding should be a stream transformer that turns a byte stream into a text stream. So there's no duplication of these methods. I believe splitting up the classes should lead to less duplication, not more, partly because you don't have to implement a lot of methods that don't do anything or are errors (eg. writing to an input stream). I admit I haven't actually written all the code, though.
3) the problems are substantially growed now - when i tried to separate input and output streams (the same will apply to detaching of seekable streams into the separate class). the problem is what i need either to provide 2 or 3 separate implementations for buffering of read-only, write-only and read-write streams or have some universal definition that should work even when base Stream don't provide part of operations. the last seems to be impossible - may be i don't understand enough Haskell's class system?
let's see:
data BufferedStream h = Buf h ....
vClose (Buf h ...) = vPutBuf ... - flush buffer's contents
how i can implement this if `h` may not support vPutBuf operation? especially to allow read/write streams to work???
You can only buffer a byte stream. See my sketch design.
4) what you mean by "There are many combinations of stream transformers that don't make sense" ? splitting Stream class to the BlockStream/TextStream/ByteStream or something else?
Yes - adding decoding to a TextStream doesn't make sense. Directly accessing the memory of a byte stream doesn't make sense: you need to buffer it first, or use a memory-mapped stream. It is still possible to implement read/write files using this structure. There's nothing stopping you having an type that is an instance of both InputStream and OutputStream (eg. a read/write file), and layering buffering on top of this would yield a buffered input/output stream in which the buffer contains only input or output data. Cheers, Simon {-# OPTIONS -fglasgow-exts -cpp #-} ----------------------------------------------------------------------------- -- | -- Module : System.IO.Stream -- Copyright : (c) various -- License : see libraries/base/LICENSE -- -- Maintainer : simonmar@microsoft.com -- Stability : experimental -- Portability : non-portable (existentials, ghc extensions) -- -- InputStreams and OutputStreams are classes of objects which support -- input and output respectively. Streams can be layered on top of various -- underlying I/O objects (such as files or sockets). Stream transformers -- can be applied to turn streams of one type into streams of another type. -- ----------------------------------------------------------------------------- module System.IO.Stream ( -- * Streams {-class-} Stream(..), {-class-} InputStream(..), {-class-} OutputStream(..), streamGet, streamReadBuffer, streamPut, streamWriteBuffer, -- * Stream connections PipeInputStream, PipeOutputStream, streamPipe, streamConnect, -- * Memory streams {-class-} MemInputStream, {-class-} MemOutputStream, -- ** Converting memory streams to I/O streams MemToInputStream, memToInputStream, MemToOutputStream, memToOutputStream, -- * Buffering BufferMode(..), BufferedInputStream, bufferIntputStream, BufferedOutputStream, bufferOutputStream, ) where import System.IO.Buffer import Foreign import Data.Word ( Word8 ) import System.IO ( BufferMode(..) ) import System.IO.Error ( mkIOError, eofErrorType ) import Control.Exception ( assert ) import Control.Monad ( when, liftM ) import Control.Concurrent import Data.IORef import GHC.Exts import GHC.Ptr ( Ptr(..) ) import GHC.IOBase ( IO(..), ioException ) import GHC.Handle ( ioe_EOF ) #define UPK {-# UNPACK #-} ! -- ----------------------------------------------------------------------------- -- Streams class Stream s where -- | closes a stream closeStream :: s -> IO () -- | returns 'True' if the stream is open streamIsOpen :: s -> IO Bool -- | ToDo: objections have been raised about this method, and -- are still to be resolved. It doesn't make as much sense -- for output streams as it does for input streams. streamIsEOS :: s -> IO Bool -- | Returns 'True' if there is data available to read from this -- stream. Returns 'False' if either there is no data available, or -- the end of the stream has been reached. streamReady :: s -> IO Bool -- | Returns the number of bytes that can be transfered to/from -- this stream, if known. streamRemaining :: s -> IO (Maybe Integer) -- | An 'InputStream' is a basic I/O object which supports reading a -- stream of 'Word8's. It is expected that 'InputStream's are unbuffered: -- buffering is layered on top of one of these. class Stream s => InputStream s where -- | Grabs data without blocking, but only if there is data available. -- If there is none, then waits for some. This function may only -- return zero if either the requested length is zero or the end of stream -- has been reached. streamReadBufferNonBlocking :: s -> Integer -> Ptr Word8 -> IO Integer -- | Reads a single 'Word8' from a stream. streamGet :: InputStream s => s -> IO Word8 streamGet s = alloca $ \p -> do r <- streamReadBufferNonBlocking s 1 p if r == 0 then ioe_EOF else peek p -- | Reads data from the stream into a 'Buffer'. Returns the -- number of elements that were read, which may only be less than the -- requested length if the end of the stream was reached. streamReadBuffer :: InputStream s => s -> Integer -> Ptr Word8 -> IO Integer streamReadBuffer s 0 buf = return 0 streamReadBuffer s len ptr = streamReadBufferLoop s ptr 0 (fromIntegral len) streamReadBufferLoop :: InputStream s => s -> Ptr Word8 -> Integer -> Integer -> IO Integer streamReadBufferLoop s ptr off len = do r <- streamReadBufferNonBlocking s len ptr if r == 0 then return (fromIntegral off) else if (r < len) then streamReadBufferLoop s (ptr `plusPtr` fromIntegral r) (off+r) (len-r) else return (off+r) -- ----------------------------------------------------------------------------- -- Output streams class Stream s => OutputStream s where -- | Writes data to an output stream. It will write at least one -- byte, but will only write further bytes if it can do so without -- blocking. -- -- The result may never be 0 if the requested write size was > 0. -- If no bytes can be written to the stream, then the -- 'streamWriteBufferNonBlocking' should raise an exception -- indicating the cause of the problem (eg. the stream is closed). streamWriteBufferNonBlocking :: s -> Integer -> Ptr Word8 -> IO Integer -- | Writes a single byte to an output stream. streamPut :: OutputStream s => s -> Word8 -> IO () streamPut s word = with word $ \p -> streamWriteBuffer s 1 p -- | Writes data to a stream, only returns when all the data has been -- written. streamWriteBuffer :: OutputStream s => s -> Integer -> Ptr Word8 -> IO () streamWriteBuffer s 0 ptr = return () streamWriteBuffer s len ptr = streamWriteBufferLoop s ptr 0 len streamWriteBufferLoop :: OutputStream s => s -> Ptr Word8 -> Integer -> Integer -> IO () streamWriteBufferLoop s ptr off len = seq off $ -- strictness hack if len == 0 then return () else do r <- streamWriteBufferNonBlocking s len ptr assert (r /= 0) $ do if (r < len) then streamWriteBufferLoop s (ptr `plusPtr` fromIntegral r) (off+r) (len-r) else return () -- ---------------------------------------------------------------------------- -- Connecting streams data PipeInputStream data PipeOutputStream instance Stream PipeInputStream -- ToDo instance InputStream PipeInputStream -- ToDo instance Stream PipeOutputStream -- ToDo instance OutputStream PipeOutputStream -- ToDo streamPipe :: IO (PipeInputStream, PipeOutputStream) streamPipe = error "unimplemented: streamOutputToInput" -- | Takes an output stream and an input stream, and pipes all the -- data from the former into the latter. streamConnect :: (OutputStream o, InputStream i) => o -> i -> IO () streamConnect = error "unimplemented: streamInputToOutput" -- ---------------------------------------------------------------------------- -- Memory streams -- | An input stream accessed directly via a memory buffer. -- Ordinary 'InputStream's may be converted to 'MemInputStream's by -- adding buffering; see 'bufferInputStream'. class MemInputStream s where -- | Consume some bytes from the memory stream. The second argument -- is an IO action that is passed a buffer and its size (the size must -- be non-zero), and it should return the number of bytes consumed. withStreamInputBuffer :: s -> (Ptr Word8 -> Int -> IO Int) -> IO () -- | An output stream accessed directly via a memory buffer. -- Ordinary 'OutputStream's may be converted to 'MemOutputStream's by -- adding buffering; see 'bufferOutputStream'. class MemOutputStream s where -- | Write some bytes to a memory stream. The second argument -- is an IO action that is passed a buffer and its size (the size must -- be non-zero), and it should return the number of bytes written. withStreamOutputBuffer :: s -> (Ptr Word8 -> Int -> IO Int) -> IO () -- ----------------------------------------------------------------------------- -- A memory stream can be converted to an ordinary byte stream newtype MemToInputStream s = MemToInputStream s deriving (Stream, MemInputStream) newtype MemToOutputStream s = MemToOutputStream s deriving (Stream, MemOutputStream) -- Rationale: what we really want is -- instance MemInputStream s => InputStream s -- but that overlaps. So instead we provide a way to convert -- every MemInputStream into something that is an instance of -- InputStream. instance InputStream (MemToInputStream s) where instance OutputStream (MemToOutputStream s) where memToInputStream :: MemInputStream s => MemToInputStream s memToInputStream = MemToInputStream memToOutputStream :: MemOutputStream s => MemToOutputStream s memToOutputStream = MemToOutputStream -- ----------------------------------------------------------------------------- -- Buffering -- | Operations on a stream with a buffer class Buffered s where -- | Sets the buffering mode on the stream. Returns 'True' if -- the buffereing mode was set, or 'False' if it wasn't. If the -- stream does not support buffering, it may return 'False'. setBufferMode :: s -> BufferMode -> IO Bool -- | Returns the current buffering mode for a stream. On a -- stream that does not support buffering, the result will always -- be 'NoBuffering'. getBufferMode :: s -> IO BufferMode -- | Returns the number of bytes of data in the buffer countBufferedBytes :: s -> IO Int -- | Operations on an output stream with a buffer class OutputBuffered s where -- | Flushes the buffer to the operating system flush :: s -> IO () -- | Flushes the buffered data as far as possible, even to the -- physical media if it can. It returns 'True' if the data -- has definitely been flushed as far as it can go: to the -- disk for a disk file, to the screen for a terminal, and so on. sync :: s -> IO Bool -- | Operations on an input stream with a buffer class InputBuffered s where -- | Discards the input buffer discard :: s -> IO () -- | Pushes back the buffered data, if possible. Returns 'True' if -- the buffer could be pushed back, 'False' otherwise. pushback :: s -> IO Bool -- TODO: Seekable superclass allows pushback? -- | An 'InputStream' with buffering added data BufferedInputStream s = BufferedInputStream s BufferMode !(IORef (Buffer Word8)) -- | An 'OutputStream' with buffering added data BufferedOutputStream s = BufferedOutputStream s BufferMode !(IORef (Buffer Word8)) bufferSize :: BufferMode -> Int bufferSize (BlockBuffering (Just size)) = size bufferSize _ = dEFAULT_BUFFER_SIZE -- | Add buffering to an 'InputStream' bufferInputStream :: InputStream s => s -> BufferMode -> BufferedInputStream s bufferInputStream stream bmode = do buffer <- allocateBuffer (bufferSize bmode) r <- newIORef buffer return (BufferedInputStream stream bmode r) -- | Add buffering to an 'OutputStream' bufferOutputStream :: OutputStream s => s -> BufferedOutputStream s bufferOutputStream stream bmode = do buffer <- allocateBuffer (bufferSize bmode) r <- newIORef buffer return (BufferedOutputStream stream bmode r) instance Stream s => MemInputStream (BufferedInputStream s) where withStreamInputBuffer b@(BufferedInputStream s bmode ref) action = do buffer <- readIORef ref if emptyBuffer buffer then do buffer' <- fillReadBuffer s buffer writeIORef ref buffer' withStreamInputBuffer b action else do let used = bufferUsed buffer count <- withBuffer buffer $ \ptr -> action (ptr `plusPtr` bufRPtr buffer) used let buffer' = bufferRemove (min used count) checkBufferInvariants buffer' writeIORef ref $! buffer' instance Stream s => MemOutputStream (BufferedOutputStream s) where withStreamOutputBuffer b@(BufferedOutputStream s bmode ref) action = do buffer <- readIORef ref let avail = bufferAvailable buffer count <- withBuffer buffer $ \ptr -> action (ptr `plusPtr` bufWPtr buffer) avail let buffer' = bufferAdd (max count avail) if fullBuffer buffer' then do writeBuffer s buffer; writeIORef ref (emptyBuffer buffer') else writeIORef ref buffer' instance Stream s => Stream (BufferedInputStream s) instance Stream s => Stream (BufferedOutputStream s) instance Buffered (BufferedOutputStream s) instance OutputBuffered (BufferedOutputStream s) instance Buffered (BufferedInputStream s) instance InputBuffered (BufferedInputStream s)
Simon Marlow
You can only buffer a byte stream. See my sketch design.
IMHO there should be buffered character streams too; of course in this case the buffer holds characters. This allows to amortize the cost of character recoding. Buffering should usually be the topmost layer. -- __("< Marcin Kowalczyk \__/ qrczak@knm.org.pl ^^ http://qrnik.knm.org.pl/~qrczak/
Hello Simon, Thursday, April 20, 2006, 12:57:10 PM, you wrote:
Believe me I've looked in detail at your streams library.
The main problem is that it doesn't have enough type structure. There are many combinations of stream transformers that don't make sense, and should therefore be ruled out by the type system. There are operations that don't work on some streams. There should at the least be a type distinction between directly accessible memory streams, byte streams, and text streams. Additionally I would add separate classes for seekable and buffered streams. I believe these changes would improve performance by reducing the size of dictionaries.
well, your last answer shows that you don't understand my problems. i'm entirely want to have precise classes, but when i run into IMPLEMENTATION, Haskell restrictions bite me again and again. for beginning, i should describe class structure what was modified according to your critics (i yet not released version with these modifications, but it available as http://freearc.narod.ru/Binary_20060329184510.rar) currently, Stream lib includes class Streams, what includes "unclassified", minor operations, such as vIsEOF or vShow, and classes BlockStream, MemoryStream, ByteStream, TextStream class BlockStream supports reading and writing of memory blocks: class (Stream m h) => BlockStream m h | h->m where vGetBuf :: h -> Ptr a -> Int -> m Int vPutBuf :: h -> Ptr a -> Int -> m () it's a natural class for low-level, "raw" streams such as FD or network sockets: instance BlockStream IO FD where vGetBuf = fdGetBuf vPutBuf = fdPutBuf vGetBuf, as you suggested, should implement non-blocking reads, while the vPutBuf should write the whole buffer. MemoryStream class supports low-level, "raw" streams whose data are ALREADY RESIDE IN MEMORY. examples are MemBuf, MemoryMappedFile, circular buffer for inter-thread communication what i plan to implement. the main difference against BlockStream is what data are not read in user-supplied buffer, but already somewhere in memory, and MemoryStream functions just give to user address and size of next part of data (for reading) or next place to fill up (for writing). so: class (Stream m h) => MemoryStream m h | h->m where -- | Receive next buffer which contains data / should be filled with data vReceiveBuf :: h -> ReadWrite -> m (Ptr a, Int) -- | Release buffer after reading `len` bytes / Send buffer filled with `len` bytes vSendBuf :: h -> Ptr a -> Int -> Int -> m () data ReadWrite = READING | WRITING | UNKNOWN deriving (Eq, Show) this scheme will allow to create MemBuf not as one huge buffer, as in current implementation and all other Binary/... libraries, but as list of buffers of some fixed size. the same applies to MemoryMappedFile - this scheme allows to map just fixed-size buffer each time instead of mapping the whole file next level is the ByteStream class, what is just the way to quickly read/write one byte at time: class (Stream m h) => ByteStream m h | h->m where vGetByte :: h -> m Word8 vPutByte :: h -> Word8 -> m () each buffering transformer implements ByteStream via BlockStream or MemoryStream: instance (BlockStream IO h) => ByteStream IO (BufferedBlockStream h) instance (MemoryStream IO h) => ByteStream IO (BufferedMemoryStream h) instance (MemoryStream IO h) => ByteStream IO (UncheckedBufferedMemoryStream h) These data type constructors (BufferedBlockStream...) joins raw stream with buffer, r/w pointer and other data required to implement buffering: data BufferedBlockStream h = BBuf h -- raw stream (IOURef BytePtr) -- buffer (IOURef BytePtr) -- buffer end (IOURef BytePtr) -- r/w pointer .... type BytePtr = Ptr Word8 Next level is class TextStream that implements text I/O operations: class (Stream m h) => TextStream m h | h->m where vGetChar :: h -> m Char vGetLine :: h -> m String vGetContents :: h -> m String vPutChar :: h -> Char -> m () vPutStr :: h -> String -> m () Encoding transformer attaches encoding to the ByteStream that allows to implement text I/O: instance (ByteStream m h) => TextStream m (WithEncoding m h) data WithEncoding m h = WithEncoding h (CharEncoding m) where type "CharEncoding m" provides vGetByte->vGetChar and vPutByte->vPutChar transformers operating in monad m all these can be named a canonical Streams hierarchy and it already works. i will be glad to add BufferedStream and SeekableStream classes and split BlockStream..TextStream to the reading and writing ones, but this is, as i said, limited by implementation issues -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com
Hi Bulat, Bulat Ziganshin wrote:
well, your last answer shows that you don't understand my problems. i'm entirely want to have precise classes, but when i run into IMPLEMENTATION, Haskell restrictions bite me again and again.
Ok, I think you need to describe these problems in more detail. The message you just wrote describes the structure of the library which I think is mostly fine, and corresponds fairly well with what I had in mind.
Could you provide a .zip or .tar.gz instead?
currently, Stream lib includes class Streams, what includes "unclassified", minor operations, such as vIsEOF or vShow, and classes BlockStream, MemoryStream, ByteStream, TextStream
class BlockStream supports reading and writing of memory blocks:
class (Stream m h) => BlockStream m h | h->m where vGetBuf :: h -> Ptr a -> Int -> m Int vPutBuf :: h -> Ptr a -> Int -> m ()
fine, that's closely equivalent to my InputStream/OutputStream.
it's a natural class for low-level, "raw" streams such as FD or network sockets:
instance BlockStream IO FD where vGetBuf = fdGetBuf vPutBuf = fdPutBuf
vGetBuf, as you suggested, should implement non-blocking reads, while the vPutBuf should write the whole buffer.
vPutBuf should be non-blocking too.
MemoryStream class supports low-level, "raw" streams whose data are ALREADY RESIDE IN MEMORY. examples are MemBuf, MemoryMappedFile, circular buffer for inter-thread communication what i plan to implement. the main difference against BlockStream is what data are not read in user-supplied buffer, but already somewhere in memory, and MemoryStream functions just give to user address and size of next part of data (for reading) or next place to fill up (for writing). so:
class (Stream m h) => MemoryStream m h | h->m where -- | Receive next buffer which contains data / should be filled with data vReceiveBuf :: h -> ReadWrite -> m (Ptr a, Int) -- | Release buffer after reading `len` bytes / Send buffer filled with `len` bytes vSendBuf :: h -> Ptr a -> Int -> Int -> m ()
which is equivalent to my MemInputStream/MemOutputStream. I believe the with-style interface that I use is better though.
data ReadWrite = READING | WRITING | UNKNOWN deriving (Eq, Show)
this scheme will allow to create MemBuf not as one huge buffer, as in current implementation and all other Binary/... libraries, but as list of buffers of some fixed size. the same applies to MemoryMappedFile - this scheme allows to map just fixed-size buffer each time instead of mapping the whole file
next level is the ByteStream class, what is just the way to quickly read/write one byte at time:
class (Stream m h) => ByteStream m h | h->m where vGetByte :: h -> m Word8 vPutByte :: h -> Word8 -> m ()
each buffering transformer implements ByteStream via BlockStream or MemoryStream:
I didn't have an equivalent to this class in my design. Why is it necessary? vGetByte/vPutByte can be implemented for an arbitrary BlockStream or indeed a MemoryStream.
instance (BlockStream IO h) => ByteStream IO (BufferedBlockStream h)
BufferedBlockStream == my BufferedInputStream/BufferedOutputStream
instance (MemoryStream IO h) => ByteStream IO (BufferedMemoryStream h)
what's a BufferedMemoryStream for? Isn't a memory stream already buffered by definition?
instance (MemoryStream IO h) => ByteStream IO (UncheckedBufferedMemoryStream h)
These data type constructors (BufferedBlockStream...) joins raw stream with buffer, r/w pointer and other data required to implement buffering:
data BufferedBlockStream h = BBuf h -- raw stream (IOURef BytePtr) -- buffer (IOURef BytePtr) -- buffer end (IOURef BytePtr) -- r/w pointer .... type BytePtr = Ptr Word8
Next level is class TextStream that implements text I/O operations:
class (Stream m h) => TextStream m h | h->m where vGetChar :: h -> m Char vGetLine :: h -> m String vGetContents :: h -> m String vPutChar :: h -> Char -> m () vPutStr :: h -> String -> m ()
Ok, I wouldn't put all those method in the class, I think. Also it is necessary to have buffering at the TextStream level, as Marcin pointed out. I haven't thought through the design carefully here.
Encoding transformer attaches encoding to the ByteStream that allows to implement text I/O:
instance (ByteStream m h) => TextStream m (WithEncoding m h)
data WithEncoding m h = WithEncoding h (CharEncoding m)
yes
where type "CharEncoding m" provides vGetByte->vGetChar and vPutByte->vPutChar transformers operating in monad m
all these can be named a canonical Streams hierarchy and it already works.
So what are the problems you were referring to? eg. this from your previous message:
2) separation of Stream classes make some automatic definitions impossible. for example, released version contains vGetBuf implementation that is defined via vGetChar and works ok for streams that provide only vGetChar as base function.
and
i will be glad to add BufferedStream and SeekableStream classes and split BlockStream..TextStream to the reading and writing ones, but this is, as i said, limited by implementation issues
which implementation issues? Cheers, Simon
Hello Simon, Friday, April 21, 2006, 2:20:26 PM, you wrote:
well, your last answer shows that you don't understand my problems. i'm entirely want to have precise classes, but when i run into IMPLEMENTATION, Haskell restrictions bite me again and again.
Ok, I think you need to describe these problems in more detail. The message you just wrote describes the structure of the library which I think is mostly fine, and corresponds fairly well with what I had in mind.
Could you provide a .zip or .tar.gz instead?
yes, the last version is http://freearc.narod.ru/StreamsBeta.zip
vPutBuf should be non-blocking too.
so it should return number of bytes written? and it should be named vPutBufNonBlocking? you are already said in February that vGetBuf should be used in buffering transformer gently - i.e. i should use just the returned number of bytes and don't enforce filling of whole buffer. how should be the policy of using vPutBufNonBlocking? what should be vGEtBuf/vPutBuf (i.e. functions provided to users) - are they should be blocking or non-blocking?
which is equivalent to my MemInputStream/MemOutputStream. I believe the with-style interface that I use is better though.
main benefit of my structure is that MemoryStream and BlockStream are very close so i hope to implement common buffering layer. your variant makes this much harder. on the other side, i don't see any benefits in using your scheme
each buffering transformer implements ByteStream via BlockStream or MemoryStream:
I didn't have an equivalent to this class in my design. Why is it necessary? vGetByte/vPutByte can be implemented for an arbitrary BlockStream or indeed a MemoryStream.
:) yes, you never mind about implementation issues! how the naked FD, for example, can implement vGetByte? i'm not Copperfield to extract buffer from nowhere. Buffering transformer is not just a bunch of functions, it's a whole DATA STRUCTURE what joins stream handle and buffer pointers together! see my previous letter what described this the main idea of the whole Streams library is what low-level stream types don't carry any data that will be used only in high-level routines. FD don't carry buffer or CharEncoding information. it's just can read or write block of data - it's all! all other data items are added by corresponding transformers. so FD CAN'T implement vGetByte with buffering and its implementation without buffering will be imho just source of errors - users will constantly asking why it is so slow. moreover i can't add vGetByte/vPutByte to the BlockStream or MemoryStream class (because both can implement it), so it anyway should be in separate class
instance (MemoryStream IO h) => ByteStream IO (BufferedMemoryStream h)
what's a BufferedMemoryStream for? Isn't a memory stream already buffered by definition?
we just use word "buffering" in different meanings. for me buffering means that underlying stream give us data in large enough chunks and this can be implemented inefficiently. buffering transformer efficiently implements byte-oriented or char-oriented operations by working with buffer (either provided by MemoryStream or allocated explicitly for BlockStream) and make rather infrequent calls to the block i/o operations of underlying stream so, for me MemoryStream is not buffered, it just support block i/o operations in slightly different way than BlockStream. going your way, as i understand, we should call withStreamInputBuffer on each i/o operation what mean that withStreamInputBuffer should be implemented very efficiently in each MemoryStream. well, it's also possible but i prefer to use the same algorithms (and i hope to join implementations) for byte&char i/o over both MemoryStreams and BlockStreams on the other side, now i think about making BlockStream->MemoryStream transformer (it should just alloc it's own buffer) and then your's withStreamInputBuffer may be used for any I/O through buffer (i.e. both for BlockStreams and MemoryStreams). it will be great for implementing functions requiring lookahead such as vGetLine
class (Stream m h) => TextStream m h | h->m where vGetChar :: h -> m Char vGetLine :: h -> m String vGetContents :: h -> m String vPutChar :: h -> Char -> m () vPutStr :: h -> String -> m ()
Ok, I wouldn't put all those method in the class, I think.
it is because you never mind about fast implementation :) i already moved vPutStrLn/vPrint out of this class.
Also it is necessary to have buffering at the TextStream level, as Marcin pointed out. I haven't thought through the design carefully here.
implementation is rather obvious and therefore boring :) we should convert data to the UCS-4 and then work with buffer containing 4-byte chars. are you not done this in 6.6 compiler when you read UTF-8 sources? we will lose vTell operation (on the other side, we anyway lose it on text streams in windows :) )
all these can be named a canonical Streams hierarchy and it already works.
So what are the problems you were referring to?
i said only about working things. as you see, i don't mentioned, for example, InTextStream or SeekableStream classes, because i had problems with implementing this
2) separation of Stream classes make some automatic definitions impossible. for example, released version contains vGetBuf implementation that is defined via vGetChar and works ok for streams that provide only vGetChar as base function.
it's a minor problem, so you can skip it. just for completeness: my first lib contained the following definitions: class Stream m h where vGetChar :: ... vGetPuf :: ... -- default implementations vGetPuf = ... some code using vGetChar instance Stream IO StringBuffer vGetChar = ... instance Stream IO StringReader vGetChar = ... and vGetBuf for StringBuffer and StringReader was defined automatically. try to make this effect for current design. it should define: instance (TextStream m h) => BlockStream m h vGetPuf = ... some code using vGetChar but such code seen by Haskell as potential overlapping instance definition
i will be glad to add BufferedStream and SeekableStream classes and split BlockStream..TextStream to the reading and writing ones, but this is, as i said, limited by implementation issues
which implementation issues?
well, let's we have 3 different FDs - ReadFD, WriteFD and ReadWriteFD. it's essentially the same Ints, just supporting different sets of I/O operations: instance InBlockStream ReadFD where ... -- implements vGetBuf instance InBlockStream ReadWriteFD where ... instance OutBlockStream WriteFD where ... -- implements vPutBuf instance OutBlockStream ReadWriteFD where ... now i'm going to implement InByteStream interface (vGetByte) and OutByteStream (vPutByte) for buffered FDs. i want to use the same data type constructor to buffer all 3 types of FD streams. well, we can write something like: instance (InBlockStream h) => InByteStream (Buffered h) where vGetByte = ... something involving vGetBuf at last instance (OutBlockStream h) => OutByteStream (Buffered h) where vPutByte = ... something involving vPutBuf at last at the first look, it seems fine. but we also need to implement some common functionality: instance (Stream h) => Stream (Buffered h) where vSeek = ... let's examine vSeek implementation. it should flush buffer on output streams, i.e. it should call vPutBuf. but that's impossible - vPutBuf is not in Streams dictionary! well, let's replicate implementation of Stream class: instance (InBlockStream h) => Stream (Buffered h) where ... instance (OutBlockStream h) => Stream (Buffered h) where ... but this, besides repeating almost the same code two times, again creates overlapped instances! what instance should use compiler for ReadWriteFD? yes, i can create 3 different data types for buffering input, output and input-output streams but this means even more code duplication. and multiplying these 3 i/o modes to the 3 buffering transformers i already have, i got limitless code growth! the same problem strikes me on each transformation stage: BlockStream->ByteStream, ByteStream->TextStream, ByteStream->BitsStream. it also bites me when i try to separate SeekableStream operations - vSeek operation required to discard input buffer, what is required in vSetBuf, for example. implementation of different Stream operations contains numerous dependencies from each other. may be it is great that i discover these dependencies at the early development stage instead of bothering with numerous errors when library will be really used in various environments btw, that you think about specifying buffering mode at the Stream creation? so vSetBuf will just gone away. to be exact, it will be still possible to use different buffering schemes, by creating new buffered Streams over the one raw Stream: h <- openRawFD "text" h1 <- withBuffering (BlockBuffering 512) h ... working with h1 vFlush h1 -- discards buffer and returns `h` pointer to exact the -- position of first byte that is not consumed by `h1` h2 <- withBuffering LineBuffering h ... working with h2 -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com
So I think I've extracted the important points here: 1. seek needs to flush the buffer, so it needs to know whether it is on a buffered handle or not. My plan was to do it like this: class Seekable s where seek :: s -> SeekLocation -> IO () instance Seekable s => Seekable (BufferedInputStream s) where seek (BufferedInputStream s buf) loc = do .. flush the buffer .. seek s loc 2. performance of putByte, getByte. You have the ByteStream interface, as far as I can tell, just to improve the performance of these operations. I didn't have such a class because I believe that shovelling large numbers of bytes is more important from a performance perspective than shovelling one byte at a time. For instance, I gave the byte reader that works on any InputStream earlier: -- | Reads a single 'Word8' from a stream. streamGet :: InputStream s => s -> IO Word8 streamGet s = alloca $ \p -> do r <- streamReadBufferNonBlocking s 1 p if r == 0 then ioe_EOF else peek p Still, you can have full speed single byte operations by just adding methods to the MemInputStream/MemOutputStream class: class MemInputStream s where withStreamInputBuffer :: s -> (Ptr Word8 -> Int -> IO (a,Int)) -> IO a memStreamGet :: s -> IO Word8 memStreamGet s = withStreamInputBuffer s $ \p i -> do x <- peek p return (x, i+1) (the default obviously isn't very fast, but individual instances would provide faster implementations). 3. code explosion. Firstly, removing ByteStream will reduce the code explosion. In my design I had separate types for BufferedInputStream and BufferedOutputStream, and there is hardly any duplication because buffering input is just different to buffering output. Read/write buffering is just a combination of these two. I still don't see why you need another layer of buffering over a memory stream. It supports direct operations on the memory already. As I said above, if you want fast(er) single byte read on these, then add it to the MemInputStream class. There is some boilerplate: the need to translate Stream, Seekable, Buffered, InputBuffered etc. through upper layers to lower layers means some boilerplate wrapper instances. But I don't think that's a big deal, and splitting up these classes hasn't made it any worse. Admittedly though, separating input and output has doubled the amount of boilerplate wrappers. I don't see a way around that, other than using some preprocessing or TH to generate the instances. -------------- So here's how I see the layering for a file reading stream using iconv: file or socket (fd-based) :: FileInputStream instance of: InputStream, Stream, Seekable byte buffer :: BufferedInputStream instance of: MemInputStream, Stream, Seekable, Buffered, InputBuffered iconv :: BufferedTextInputStream instance of: MemTextInputStream, TextInputStream, Stream, Seekable, Buffered, InputBuffered lock :: LockedTextInputStream instance of: MemTextInputStream, TextInputStream, Stream, Seekable, Buffered, InputBuffered Two layers of buffering aren't strictly necessary, because the byte buffer will usually be emptied immediately by the decoder, except that it might leave a few bytes when there isn't a complete character to decode. It is simpler to do it this way though, because the decoding layer then works with any MemInputStream (eg. a memory mapped file or ByteString, as well as a buffered file), and it doesn't lose any efficiency. A memory-mapped file becomes a MemInputStream directly, so it only gets 3 layers instead of 4, similarly for a ByteString or UArray Word8. I haven't said what the MemTextInputStream or TextInputStream classes look like. I imagine something like this: class MemTextInputStream s where withCharInputBuffer :: s -> (Ptr Char -> Int -> IO (a,Int)) -> IO a class TextInputStream s where getChar :: s -> IO Char getContents ... getLine ... A StringReader can be made an instance of TextInputStream directly. So do you see any problem with this design? Performance should be fine: getContents reads directly from the buffer filled by iconv, and all other operations are done at buffer sizes. Fast byte and multibyte operations are supported on MemInputStreams so fast Binary I/O and serialisation/deserialisation are possible. Cheers, Simon Bulat Ziganshin wrote:
Hello Simon,
Friday, April 21, 2006, 2:20:26 PM, you wrote:
well, your last answer shows that you don't understand my problems. i'm entirely want to have precise classes, but when i run into IMPLEMENTATION, Haskell restrictions bite me again and again.
Ok, I think you need to describe these problems in more detail. The message you just wrote describes the structure of the library which I think is mostly fine, and corresponds fairly well with what I had in mind.
Could you provide a .zip or .tar.gz instead?
yes, the last version is http://freearc.narod.ru/StreamsBeta.zip
vPutBuf should be non-blocking too.
so it should return number of bytes written? and it should be named vPutBufNonBlocking? you are already said in February that vGetBuf should be used in buffering transformer gently - i.e. i should use just the returned number of bytes and don't enforce filling of whole buffer. how should be the policy of using vPutBufNonBlocking? what should be vGEtBuf/vPutBuf (i.e. functions provided to users) - are they should be blocking or non-blocking?
which is equivalent to my MemInputStream/MemOutputStream. I believe the with-style interface that I use is better though.
main benefit of my structure is that MemoryStream and BlockStream are very close so i hope to implement common buffering layer. your variant makes this much harder. on the other side, i don't see any benefits in using your scheme
each buffering transformer implements ByteStream via BlockStream or MemoryStream:
I didn't have an equivalent to this class in my design. Why is it necessary? vGetByte/vPutByte can be implemented for an arbitrary BlockStream or indeed a MemoryStream.
:) yes, you never mind about implementation issues! how the naked FD, for example, can implement vGetByte? i'm not Copperfield to extract buffer from nowhere. Buffering transformer is not just a bunch of functions, it's a whole DATA STRUCTURE what joins stream handle and buffer pointers together! see my previous letter what described this
the main idea of the whole Streams library is what low-level stream types don't carry any data that will be used only in high-level routines. FD don't carry buffer or CharEncoding information. it's just can read or write block of data - it's all! all other data items are added by corresponding transformers. so FD CAN'T implement vGetByte with buffering and its implementation without buffering will be imho just source of errors - users will constantly asking why it is so slow. moreover i can't add vGetByte/vPutByte to the BlockStream or MemoryStream class (because both can implement it), so it anyway should be in separate class
instance (MemoryStream IO h) => ByteStream IO (BufferedMemoryStream h)
what's a BufferedMemoryStream for? Isn't a memory stream already buffered by definition?
we just use word "buffering" in different meanings. for me buffering means that underlying stream give us data in large enough chunks and this can be implemented inefficiently. buffering transformer efficiently implements byte-oriented or char-oriented operations by working with buffer (either provided by MemoryStream or allocated explicitly for BlockStream) and make rather infrequent calls to the block i/o operations of underlying stream
so, for me MemoryStream is not buffered, it just support block i/o operations in slightly different way than BlockStream. going your way, as i understand, we should call withStreamInputBuffer on each i/o operation what mean that withStreamInputBuffer should be implemented very efficiently in each MemoryStream. well, it's also possible but i prefer to use the same algorithms (and i hope to join implementations) for byte&char i/o over both MemoryStreams and BlockStreams
on the other side, now i think about making BlockStream->MemoryStream transformer (it should just alloc it's own buffer) and then your's withStreamInputBuffer may be used for any I/O through buffer (i.e. both for BlockStreams and MemoryStreams). it will be great for implementing functions requiring lookahead such as vGetLine
class (Stream m h) => TextStream m h | h->m where vGetChar :: h -> m Char vGetLine :: h -> m String vGetContents :: h -> m String vPutChar :: h -> Char -> m () vPutStr :: h -> String -> m ()
Ok, I wouldn't put all those method in the class, I think.
it is because you never mind about fast implementation :) i already moved vPutStrLn/vPrint out of this class.
Also it is necessary to have buffering at the TextStream level, as Marcin pointed out. I haven't thought through the design carefully here.
implementation is rather obvious and therefore boring :) we should convert data to the UCS-4 and then work with buffer containing 4-byte chars. are you not done this in 6.6 compiler when you read UTF-8 sources? we will lose vTell operation (on the other side, we anyway lose it on text streams in windows :) )
all these can be named a canonical Streams hierarchy and it already works.
So what are the problems you were referring to?
i said only about working things. as you see, i don't mentioned, for example, InTextStream or SeekableStream classes, because i had problems with implementing this
2) separation of Stream classes make some automatic definitions impossible. for example, released version contains vGetBuf implementation that is defined via vGetChar and works ok for streams that provide only vGetChar as base function.
it's a minor problem, so you can skip it. just for completeness: my first lib contained the following definitions:
class Stream m h where vGetChar :: ... vGetPuf :: ...
-- default implementations vGetPuf = ... some code using vGetChar
instance Stream IO StringBuffer vGetChar = ... instance Stream IO StringReader vGetChar = ...
and vGetBuf for StringBuffer and StringReader was defined automatically. try to make this effect for current design. it should define:
instance (TextStream m h) => BlockStream m h vGetPuf = ... some code using vGetChar
but such code seen by Haskell as potential overlapping instance definition
i will be glad to add BufferedStream and SeekableStream classes and split BlockStream..TextStream to the reading and writing ones, but this is, as i said, limited by implementation issues
which implementation issues?
well, let's we have 3 different FDs - ReadFD, WriteFD and ReadWriteFD. it's essentially the same Ints, just supporting different sets of I/O operations:
instance InBlockStream ReadFD where ... -- implements vGetBuf instance InBlockStream ReadWriteFD where ... instance OutBlockStream WriteFD where ... -- implements vPutBuf instance OutBlockStream ReadWriteFD where ...
now i'm going to implement InByteStream interface (vGetByte) and OutByteStream (vPutByte) for buffered FDs. i want to use the same data type constructor to buffer all 3 types of FD streams. well, we can write something like:
instance (InBlockStream h) => InByteStream (Buffered h) where vGetByte = ... something involving vGetBuf at last
instance (OutBlockStream h) => OutByteStream (Buffered h) where vPutByte = ... something involving vPutBuf at last
at the first look, it seems fine. but we also need to implement some common functionality:
instance (Stream h) => Stream (Buffered h) where vSeek = ...
let's examine vSeek implementation. it should flush buffer on output streams, i.e. it should call vPutBuf. but that's impossible - vPutBuf is not in Streams dictionary! well, let's replicate implementation of Stream class:
instance (InBlockStream h) => Stream (Buffered h) where ... instance (OutBlockStream h) => Stream (Buffered h) where ...
but this, besides repeating almost the same code two times, again creates overlapped instances! what instance should use compiler for ReadWriteFD?
yes, i can create 3 different data types for buffering input, output and input-output streams but this means even more code duplication. and multiplying these 3 i/o modes to the 3 buffering transformers i already have, i got limitless code growth!
the same problem strikes me on each transformation stage: BlockStream->ByteStream, ByteStream->TextStream, ByteStream->BitsStream. it also bites me when i try to separate SeekableStream operations - vSeek operation required to discard input buffer, what is required in vSetBuf, for example. implementation of different Stream operations contains numerous dependencies from each other. may be it is great that i discover these dependencies at the early development stage instead of bothering with numerous errors when library will be really used in various environments
btw, that you think about specifying buffering mode at the Stream creation? so vSetBuf will just gone away. to be exact, it will be still possible to use different buffering schemes, by creating new buffered Streams over the one raw Stream:
h <- openRawFD "text" h1 <- withBuffering (BlockBuffering 512) h ... working with h1 vFlush h1 -- discards buffer and returns `h` pointer to exact the -- position of first byte that is not consumed by `h1` h2 <- withBuffering LineBuffering h ... working with h2
Hello Bulat, Friday, April 21, 2006, 1:10:29 PM, you wrote:
according to your critics (i yet not released version with these modifications, but it available as http://freearc.narod.ru/Binary_20060329184510.rar)
sorry, it is http://freearc.narod.ru/Binary_20060331174635.rar -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com
participants (3)
-
Bulat Ziganshin -
Marcin 'Qrczak' Kowalczyk -
Simon Marlow