|Portability||portable to Hugs and GHC.|
|Maintainer||Adam Langley <firstname.lastname@example.org>|
This is a version of the Get monad for incremental parsing. The parser is written as if a single, huge, strict ByteString was to be parsed. It produces results as it parses by calling yield.
However, if the parser runs out of data, rather than failing the caller sees a Partial result, which includes the list of yielded values so far and a continuation. By calling the continuation with more data, the parser continues, none the wiser.
Take the following example
testParse = do a <- getWord16be b <- getWord16be return $ a + b test = runGet testParse $ B.pack [1,0,0]
testParse needs to read 4 bytes in order to complete, so test is
a Partial, which includes the continuation function, so which you can pass
more data until it completes
The lookahead functions have been removed from this parser because of their incompatibility with the incremental monad at the moment.
- data Get r a
- data Result a
- runGet :: Get r r -> ByteString -> Result r
- skip :: Int -> Get r ()
- bytesRead :: Get r Int
- remaining :: Get r Int
- isEmpty :: Get r Bool
- plus :: Get r a -> Get r a -> Get r a
- zero :: Get r a
- spanOf :: (Word8 -> Bool) -> Get r ByteString
- suspend :: Get r ()
- getWord8 :: Get r Word8
- getByteString :: Int -> Get r ByteString
- getWord16be :: Get r Word16
- getWord32be :: Get r Word32
- getWord64be :: Get r Word64
- getWord16le :: Get r Word16
- getWord32le :: Get r Word32
- getWord64le :: Get r Word64
- getWordhost :: Get r Word
- getWord16host :: Get r Word16
- getWord32host :: Get r Word32
- getWord64host :: Get r Word64
The Get type
The result of a partial parse
the parse failed with the given error message
|Finished ByteString a|
the parse finished and produced the given list of results doing so. Any unparsed data is returned.
|Partial (ByteString -> Result a)|
the parse ran out of data before finishing, but produced the given list of results before doing so. To continue the parse pass more data to the given continuation
Get the number of remaining unparsed bytes. Useful for checking whether all input has been consumed.
Test whether all input has been consumed, i.e. there are no remaining unparsed bytes.
This is the choice operator. If the first option fails, the second is tried. The failure of the first option must happen within this function otherwise rollback is not attempted.
Parsing particular types
get method for strict ByteStrings. Fails if fewer
n bytes are left in the input.