Renamed Data.SExpression to non-conflicting (but terribly punny) name Data.SCargot
Getty Ritter
9 years ago
1 | -- | Contains the type of atoms that Common Lisp understands, as | |
2 | -- well as the built-in reader macros that Common Lisp provides. | |
3 | -- Given a Common Lisp source file that contains no extra reader | |
4 | -- macro definitions, this module should successfully parse and | |
5 | -- desugar even quoted lists and vector literals. | |
6 | ||
7 | module Data.SCargot.CommonLisp where | |
8 | ||
9 | data Atom | |
10 | = Symbol Text | |
11 | | String Text | |
12 | | Integer Int | |
13 | | True | |
14 | deriving (Eq, Show, Read) | |
15 | ||
16 | parseSexpr :: Text -> Either SExprError |
1 | module Data.SCargot.General where | |
2 | ||
3 | import Control.Applicative | |
4 | import Data.Attoparsec.Text | |
5 | import Data.Map.String (Map) | |
6 | import qualified Data.Map.String as M | |
7 | ||
8 | import Data.SCargot.Repr | |
9 | ||
10 | type ReaderMacroMap atom = Map Char (Reader atom) | |
11 | type Reader atom = (Parser (SExpr atom) -> Parser (SExpr atom)) | |
12 | type Serializer atom = atom -> Text | |
13 | ||
14 | -- | A 'SExprSpec' describes a parser and emitter for a particular | |
15 | -- variant of S-Expressions. The @atom@ type corresponds to a | |
16 | -- Haskell type used to represent the atoms, and the @carrier@ | |
17 | -- type corresponds to the parsed S-Expression structure. This | |
18 | -- is deliberately opaque so that it must be constructed and | |
19 | -- modified with other helper functions. | |
20 | data SExprSpec atom carrier = SExprSpec | |
21 | { sesPAtom :: Parser atom | |
22 | , sesSAtom :: Serializer atom | |
23 | , rmMap :: ReaderMacroMap atom | |
24 | , postparse :: SExpr atom -> Either String carrier | |
25 | , preserial :: carrier -> SExpr atom | |
26 | } | |
27 | ||
28 | -- | This creates a basic 'SExprSpec' when given a parser and serializer | |
29 | -- for an atom type. | |
30 | mkSpec :: Parser atom -> Serializer atom -> SExprSpec atom (SExpr atom) | |
31 | mkSpec p s = SExprSpec | |
32 | { sesPAtom = p | |
33 | , sesSAtom = s | |
34 | , rmMap = M.empty | |
35 | , postparse = return | |
36 | , preserial = id | |
37 | } | |
38 | ||
39 | -- | This is used to modify the carrier type for a 'SExprSpec'. This is | |
40 | -- used internally to convert between various 'SExpr' representations, | |
41 | -- but could also be used externally to add an extra conversion layer | |
42 | -- onto a 'SExprSpec', e.g. for a custom Lisp-like language: | |
43 | -- | |
44 | -- > mySpec :: SExprSpec MyAtomType MyAST | |
45 | -- > mySpec = convertSpec sexprToMyAST myASTToSexpr spec | |
46 | -- > where spec = mkSpec myParser mySerializer | |
47 | convertSpec :: (b -> Either String c) -> (c -> b) -> SExprSpec a b -> SExprSpec a c | |
48 | convertSpec f g spec = spec | |
49 | { postparse = postparse spec >=> f | |
50 | , preserial = g . preserial spec | |
51 | } | |
52 | ||
53 | addReader :: Char -> Reader a -> SExprSpec a c -> SExprSpec a c | |
54 | addReader c reader spec = spec { rmMap = insert c reader (rmMap spec) } | |
55 | ||
56 | quote :: atom -> Reader atom | |
57 | quote q parse = go <$> parse | |
58 | where go v = SCons q (SCons v SNil) | |
59 | ||
60 | toRich :: SExprSpec a (SExpr b) -> SExprSpec a (RichSExpr b) | |
61 | toRich = convertSpec (return . toRich) fromRich | |
62 | ||
63 | toWellFormed :: SExprSpec a (SExpr b) -> SExprSpec a (WellFormedSExpr b) | |
64 | toWellFormed = convertSpec toWellFormed fromWellFormed | |
65 | ||
66 | parseGenericSExpr :: Parser atom -> ReaderMacroMap atom -> Parser (SExpr atom) | |
67 | ||
68 | -- | | |
69 | parseSExpr :: SExprSpec atom carrier -> Text -> Either String carrier | |
70 | parseSExpr spec = undefined | |
71 | ||
72 | -- | blah | |
73 | serializeSExpr :: SExprSpec atom carrier -> carrier -> Text | |
74 | serializeSExpr spec = serializeGenericSExpr ses . preserial |
1 | {-# LANGUAGE PatternSynonyms #-} | |
2 | ||
3 | module Data.SCargot.Repr.Rich | |
4 | ( pattern List | |
5 | , pattern DotList | |
6 | , pattern Atom | |
7 | ) where | |
8 | ||
9 | import Data.SCargot.Repr as R | |
10 | ||
11 | pattern List xs = R.RSList xs | |
12 | pattern DotList xs = R.RSDotted xs | |
13 | pattern Atom a = R.RSAtom a |
1 | {-# LANGUAGE PatternSynonyms #-} | |
2 | ||
3 | module Data.SCargot.Repr.Rich | |
4 | ( pattern List | |
5 | , pattern Atom | |
6 | ) where | |
7 | ||
8 | import Data.SCargot.Repr as R | |
9 | ||
10 | pattern List xs = R.WFSList xs | |
11 | pattern Atom a = R.WFSAtom a |
1 | module Data.SCargot.Repr | |
2 | ( SExpr(..) | |
3 | , RichSExpr(..) | |
4 | , toRich | |
5 | , fromRich | |
6 | , WellFormedSExpr(..) | |
7 | , toWellFormed | |
8 | , fromWellFormed | |
9 | ) where | |
10 | ||
11 | -- | All S-Expressions can be understood as a sequence | |
12 | -- of @cons@ cells (represented here by 'SCons'), the | |
13 | -- empty list @nil@ (represented by 'SNil') or an | |
14 | -- @atom@. | |
15 | data SExpr atom | |
16 | = SCons (SExpr atom) (SExpr atom) | |
17 | | SAtom atom | |
18 | | SNil | |
19 | deriving (Eq, Show, Read) | |
20 | ||
21 | -- | Sometimes, the cons-based interface is too low | |
22 | -- level, and we'd rather have the lists themselves | |
23 | -- exposed. In this case, we have 'RSList' to | |
24 | -- represent a well-formed cons list, and 'RSDotted' | |
25 | -- to represent an improper list of the form | |
26 | -- @(a b c . d)@. | |
27 | data RichSExpr atom | |
28 | = RSList [RichSExpr atom] | |
29 | | RSDotted [RichSExpr atom] atom | |
30 | | RSAtom atom | |
31 | deriving (Eq, Show, Read) | |
32 | ||
33 | -- | A Rich S-Expression might be a nicer interface | |
34 | -- for certain libraries. It should always be true | |
35 | -- that | |
36 | -- | |
37 | -- > fromRich . toRich == id | |
38 | -- | |
39 | -- and that | |
40 | -- | |
41 | -- > toRich . fromRich == id | |
42 | toRich :: SExpr atom -> RichSExpr atom | |
43 | toRich (SAtom a) = RSAtom a | |
44 | toRich (SCons x xs) = go xs [toRich x] | |
45 | where go (SAtom a) rs = RSDotted rs a | |
46 | go SNil rs = RSList rs | |
47 | go (SCons x xs) rs = go xs (toRich x:rs) | |
48 | ||
49 | -- | This follows the same laws as 'toRich'. | |
50 | fromRich :: RichSExpr atom -> SExpr atom | |
51 | fromRich (RSAtom a) = SAtom a | |
52 | fromRich (RSList xs) = foldr SCons SNil (map fromRich xs) | |
53 | fromRich (RSDotted xs x) = foldr SCons (SAtom x) (map fromRich xs) | |
54 | ||
55 | -- | A well-formed s-expression is one which does not | |
56 | -- contain any dotted lists. This means that not | |
57 | -- every value of @SExpr a@ can be converted to a | |
58 | -- @WellFormedSExpr a@, although the opposite is | |
59 | -- fine. | |
60 | data WellFormedSExpr atom | |
61 | = WFSList [WellFormedSExpr atom] | |
62 | | WFSAtom atom | |
63 | deriving (Eq, Show, Read) | |
64 | ||
65 | -- | This will be @Nothing@ is the argument contains an | |
66 | -- improper list. It should hold that | |
67 | -- | |
68 | -- > toWellFormed . fromWellFormed == Right | |
69 | toWellFormed :: SExpr atom -> Either String (WellFormedSExpr atom) | |
70 | toWellFormed (SAtom a) = return (WFSAtom a) | |
71 | toWellFormed (SCons x xs) = do | |
72 | x' <- toWellFormed x | |
73 | go xs [x'] | |
74 | where go (SAtom a) rs = Left "Found atom in cdr position" | |
75 | go SNil rs = return (WFSList rs) | |
76 | go (SCons x xs) rs = do | |
77 | x' <- toWellFormed x | |
78 | go xs (x':rs) | |
79 | ||
80 | -- | Convert a WellFormedSExpr back into a SExpr. | |
81 | fromWellFormed :: WellFormedSExpr atom -> SExpr atom | |
82 | fromWellFormed (WFSAtom a) = SAtom a | |
83 | fromWellFormed (WFSList xs) = | |
84 | foldr SCons SNil (map fromWellFormed xs) |
1 | module Data.SCargot.Rivest where | |
2 | ||
3 | import Data.ByteString (ByteString) | |
4 | import qualified Data.ByteString as BS | |
5 | import qualified Data.ByteString.Base64 as B64 | |
6 | import Data.Text (Text) | |
7 | import qualified Data.Text as T | |
8 | ||
9 | newtype Atom = Atom { fromAtom :: ByteString } deriving (Eq, Show, Read) | |
10 | ||
11 | pToken :: Parser ByteString | |
12 | pToken = undefined | |
13 | ||
14 | pQuoted :: Maybe Int -> Parser ByteString | |
15 | pQuoted = do | |
16 | char '"' | |
17 | ss <- many1 quoteChar | |
18 | char '"' | |
19 | return ss | |
20 | ||
21 | pHex :: Parser ByteString | |
22 | pHex = undefined | |
23 | ||
24 | pVerbatim :: Int -> Parser ByteString | |
25 | pVerbatim = do | |
26 | char ':' | |
27 | take n | |
28 | ||
29 | pBase64Verbatim :: Parser ByteString | |
30 | pBase64 :: Parser ByteString |
1 | {-| The "s-cargot" library attempts to be as general as possible, and | |
2 | to support a wide range of use-cases for s-expressions. It is built | |
3 | around a core of primitives which are then exposed in various | |
4 | ways, and can be easily and flexibly extended. This tutorial | |
5 | describes particular use-cases, and then shows how to adapt this | |
6 | library to that use-case. | |
7 | -} | |
8 | ||
9 | module Data.SCargot.Tutorial | |
10 | ( -- * Basic Usage and Organization | |
11 | -- $usage | |
12 | -- * Analyzing Scheme code | |
13 | -- $scheme | |
14 | -- * Building a Custom Config Format | |
15 | -- $config | |
16 | -- * Building a Custom Lisp | |
17 | -- $lisp | |
18 | ) where | |
19 | ||
20 | {- $usage | |
21 | ||
22 | -} | |
23 | ||
24 | {- $scheme | |
25 | ||
26 | -} | |
27 | ||
28 | ||
29 | {- $config | |
30 | ||
31 | -} | |
32 | ||
33 | {- $lisp | |
34 | ||
35 | -} |
1 | -- | Contains the type of atoms that Common Lisp understands, as | |
2 | -- well as the built-in reader macros that Common Lisp provides. | |
3 | -- Given a Common Lisp source file that contains no extra reader | |
4 | -- macro definitions, this module should successfully parse and | |
5 | -- desugar even quoted lists and vector literals. | |
6 | ||
7 | module Data.SExpression.CommonLisp where | |
8 | ||
9 | data Atom | |
10 | = Symbol Text | |
11 | | String Text | |
12 | | Integer Int | |
13 | | True | |
14 | deriving (Eq, Show, Read) | |
15 | ||
16 | parseSexpr :: Text -> Either SExprError |
1 | module Data.SExpression.General where | |
2 | ||
3 | import Control.Applicative | |
4 | import Data.Attoparsec.Text | |
5 | import Data.Map.String (Map) | |
6 | import qualified Data.Map.String as M | |
7 | ||
8 | type ReaderMacroMap atom = Map Char (Reader atom) | |
9 | type Reader atom = (Parser (SExpr atom) -> Parser (SExpr atom)) | |
10 | type Serializer atom = atom -> Text | |
11 | ||
12 | -- | A 'SExprSpec' describes a parser and emitter for a particular | |
13 | -- variant of S-Expressions. The @atom@ type corresponds to a | |
14 | -- Haskell type used to represent the atoms, and the @carrier@ | |
15 | -- type corresponds to the parsed S-Expression structure. This | |
16 | -- is deliberately opaque so that it must be constructed and | |
17 | -- modified with other helper functions. | |
18 | data SExprSpec atom carrier = SExprSpec | |
19 | { sesPAtom :: Parser atom | |
20 | , sesSAtom :: Serializer atom | |
21 | , rmMap :: ReaderMacroMap atom | |
22 | , postparse :: SExpr atom -> Either String carrier | |
23 | , preserial :: carrier -> SExpr atom | |
24 | } | |
25 | ||
26 | -- | This creates a basic 'SExprSpec' when given a parser and serializer | |
27 | -- for an atom type. | |
28 | mkSpec :: Parser atom -> Serializer atom -> SExprSpec atom (SExpr atom) | |
29 | mkSpec p s = SExprSpec | |
30 | { sesPAtom = p | |
31 | , sesSAtom = s | |
32 | , rmMap = M.empty | |
33 | , postparse = return | |
34 | , preserial = id | |
35 | } | |
36 | ||
37 | -- | This is used to modify the carrier type for a 'SExprSpec'. This is | |
38 | -- used internally to convert between various 'SExpr' representations, | |
39 | -- but could also be used externally to add an extra conversion layer | |
40 | -- onto a 'SExprSpec', e.g. for a custom Lisp-like language: | |
41 | -- | |
42 | -- > mySpec :: SExprSpec MyAtomType MyAST | |
43 | -- > mySpec = convertSpec sexprToMyAST myASTToSexpr spec | |
44 | -- > where spec = mkSpec myParser mySerializer | |
45 | convertSpec :: (b -> Either String c) -> (c -> b) -> SExprSpec a b -> SExprSpec a c | |
46 | convertSpec f g spec = spec | |
47 | { postparse = postparse spec >=> f | |
48 | , preserial = g . preserial spec | |
49 | } | |
50 | ||
51 | addReader :: Char -> Reader a -> SExprSpec a c -> SExprSpec a c | |
52 | addReader c reader spec = spec { rmMap = insert c reader (rmMap spec) } | |
53 | ||
54 | quote :: atom -> Reader atom | |
55 | quote q parse = go <$> parse | |
56 | where go v = SCons q (SCons v SNil) | |
57 | ||
58 | toRich :: SExprSpec a (SExpr b) -> SExprSpec a (RichSExpr b) | |
59 | toRich = convertSpec (return . toRich) fromRich | |
60 | ||
61 | toWellFormed :: SExprSpec a (SExpr b) -> SExprSpec a (WellFormedSExpr b) | |
62 | toWellFormed = convertSpec toWellFormed fromWellFormed | |
63 | ||
64 | parseGenericSExpr :: Parser atom -> ReaderMacroMap atom -> Parser (SExpr atom) | |
65 | ||
66 | -- | | |
67 | parseSExpr :: SExprSpec atom carrier -> Text -> Either String carrier | |
68 | parseSExpr spec = undefined | |
69 | ||
70 | -- | blah | |
71 | serializeSExpr :: SExprSpec atom carrier -> carrier -> Text | |
72 | serializeSExpr spec = serializeGenericSExpr ses . preserial |
1 | {-# LANGUAGE PatternSynonyms #-} | |
2 | ||
3 | module Data.SExpression.Repr.Rich | |
4 | ( pattern List | |
5 | , pattern DotList | |
6 | , pattern Atom | |
7 | ) where | |
8 | ||
9 | import Data.SExpression.Repr as R | |
10 | ||
11 | pattern List xs = R.RSList xs | |
12 | pattern DotList xs = R.RSDotted xs | |
13 | pattern Atom a = R.RSAtom a |
1 | {-# LANGUAGE PatternSynonyms #-} | |
2 | ||
3 | module Data.SExpression.Repr.Rich | |
4 | ( pattern List | |
5 | , pattern Atom | |
6 | ) where | |
7 | ||
8 | import Data.SExpression.Repr as R | |
9 | ||
10 | pattern List xs = R.WFSList xs | |
11 | pattern Atom a = R.WFSAtom a |
1 | module Data.SExpression.Repr | |
2 | ( SExpr(..) | |
3 | , RichSExpr(..) | |
4 | , toRich | |
5 | , fromRich | |
6 | , WellFormedSExpr(..) | |
7 | , toWellFormed | |
8 | , fromWellFormed | |
9 | ) where | |
10 | ||
11 | -- | All S-Expressions can be understood as a sequence | |
12 | -- of @cons@ cells (represented here by 'SCons'), the | |
13 | -- empty list @nil@ (represented by 'SNil') or an | |
14 | -- @atom@. | |
15 | data SExpr atom | |
16 | = SCons (SExpr atom) (SExpr atom) | |
17 | | SAtom atom | |
18 | | SNil | |
19 | deriving (Eq, Show, Read) | |
20 | ||
21 | -- | Sometimes, the cons-based interface is too low | |
22 | -- level, and we'd rather have the lists themselves | |
23 | -- exposed. In this case, we have 'RSList' to | |
24 | -- represent a well-formed cons list, and 'RSDotted' | |
25 | -- to represent an improper list of the form | |
26 | -- @(a b c . d)@. | |
27 | data RichSExpr atom | |
28 | = RSList [RichSExpr atom] | |
29 | | RSDotted [RichSExpr atom] atom | |
30 | | RSAtom atom | |
31 | deriving (Eq, Show, Read) | |
32 | ||
33 | -- | A Rich S-Expression might be a nicer interface | |
34 | -- for certain libraries. It should always be true | |
35 | -- that | |
36 | -- | |
37 | -- > fromRich . toRich == id | |
38 | -- | |
39 | -- and that | |
40 | -- | |
41 | -- > toRich . fromRich == id | |
42 | toRich :: SExpr atom -> RichSExpr atom | |
43 | toRich (SAtom a) = RSAtom a | |
44 | toRich (SCons x xs) = go xs [toRich x] | |
45 | where go (SAtom a) rs = RSDotted rs a | |
46 | go SNil rs = RSList rs | |
47 | go (SCons x xs) rs = go xs (toRich x:rs) | |
48 | ||
49 | -- | This follows the same laws as 'toRich'. | |
50 | fromRich :: RichSExpr atom -> SExpr atom | |
51 | fromRich (RSAtom a) = SAtom a | |
52 | fromRich (RSList xs) = foldr SCons SNil (map fromRich xs) | |
53 | fromRich (RSDotted xs x) = foldr SCons (SAtom x) (map fromRich xs) | |
54 | ||
55 | -- | A well-formed s-expression is one which does not | |
56 | -- contain any dotted lists. This means that not | |
57 | -- every value of @SExpr a@ can be converted to a | |
58 | -- @WellFormedSExpr a@, although the opposite is | |
59 | -- fine. | |
60 | data WellFormedSExpr atom | |
61 | = WFSList [WellFormedSExpr atom] | |
62 | | WFSAtom atom | |
63 | deriving (Eq, Show, Read) | |
64 | ||
65 | -- | This will be @Nothing@ is the argument contains an | |
66 | -- improper list. It should hold that | |
67 | -- | |
68 | -- > toWellFormed . fromWellFormed == Right | |
69 | toWellFormed :: SExpr atom -> Either String (WellFormedSExpr atom) | |
70 | toWellFormed (SAtom a) = return (WFSAtom a) | |
71 | toWellFormed (SCons x xs) = do | |
72 | x' <- toWellFormed x | |
73 | go xs [x'] | |
74 | where go (SAtom a) rs = Left "Found atom in cdr position" | |
75 | go SNil rs = return (WFSList rs) | |
76 | go (SCons x xs) rs = do | |
77 | x' <- toWellFormed x | |
78 | go xs (x':rs) | |
79 | ||
80 | -- | Convert a WellFormedSExpr back into a SExpr. | |
81 | fromWellFormed :: WellFormedSExpr atom -> SExpr atom | |
82 | fromWellFormed (WFSAtom a) = SAtom a | |
83 | fromWellFormed (WFSList xs) = | |
84 | foldr SCons SNil (map fromWellFormed xs) |
1 | module Data.SExpression.Rivest where | |
2 | ||
3 | import Data.ByteString (ByteString) | |
4 | import qualified Data.ByteString as BS | |
5 | import qualified Data.ByteString.Base64 as B64 | |
6 | import Data.Text (Text) | |
7 | import qualified Data.Text as T | |
8 | ||
9 | newtype Atom = Atom { fromAtom :: ByteString } deriving (Eq, Show, Read) | |
10 | ||
11 | pToken :: Parser ByteString | |
12 | pToken = undefined | |
13 | ||
14 | pQuoted :: Maybe Int -> Parser ByteString | |
15 | pQuoted = do | |
16 | char '"' | |
17 | ss <- many1 quoteChar | |
18 | char '"' | |
19 | return ss | |
20 | ||
21 | pHex :: Parser ByteString | |
22 | pHex = undefined | |
23 | ||
24 | pVerbatim :: Int -> Parser ByteString | |
25 | pVerbatim = do | |
26 | char ':' | |
27 | take n | |
28 | ||
29 | pBase64Verbatim :: Parser ByteString | |
30 | pBase64 :: Parser ByteString |
1 | {-| The @blah@ library attempts to be as general as possible, and to | |
2 | support a wide range of use-cases for s-expressions. It is built | |
3 | around a core of primitives which are then exposed in various | |
4 | ways, and can be easily and flexibly extended. This tutorial | |
5 | describes particular use-cases, and then shows how to adapt this | |
6 | library to that use-case. | |
7 | -} | |
8 | ||
9 | module Data.SExpression.Tutorial | |
10 | ( -- * Basic Usage and Organization | |
11 | -- $usage | |
12 | -- * Analyzing Scheme code | |
13 | -- $scheme | |
14 | -- * Building a Custom Config Format | |
15 | -- $config | |
16 | -- * Building a Custom Lisp | |
17 | -- $lisp | |
18 | ) where | |
19 | ||
20 | {- $usage | |
21 | ||
22 | -} | |
23 | ||
24 | {- $scheme | |
25 | ||
26 | -} | |
27 | ||
28 | ||
29 | {- $config | |
30 | ||
31 | -} | |
32 | ||
33 | {- $lisp | |
34 | ||
35 | -} |