1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
|
{-# LANGUAGE OverloadedStrings #-}
--------------------------------------------------------------------
-- |
-- Module : Codec.MIME.Pare
-- Copyright : (c) 2006-2009, Galois, Inc.
-- License : BSD3
--
-- Maintainer: Sigbjorn Finne <sigbjorn.finne@gmail.com>
-- Stability : provisional
-- Portability: portable
--
-- Parsing MIME content.
--
--------------------------------------------------------------------
module Codec.MIME.Parse
( parseMIMEBody -- :: [(T.Text,T.Text)] -> T.Text -> MIMEValue
, parseMIMEType -- :: T.Text -> Maybe Type
, parseMIMEMessage -- :: T.Text -> MIMEValue
, parseHeaders -- :: T.Text -> ([(T.Text,T.Text)], T.Text)
, parseMultipart -- :: Type -> T.Text -> (MIMEValue, T.Text)
, parseContentType -- :: T.Text -> Maybe Type
, splitMulti -- :: T.Text -> T.Text -> ([MIMEValue], T.Text)
, normalizeCRLF
) where
import Codec.MIME.Type
import Codec.MIME.Decode
import Control.Arrow(second)
import Data.Char
import Data.Maybe
import qualified Data.List as L
import Debug.Trace ( trace )
import qualified Data.Text as T
import Data.Monoid(Monoid(..), (<>))
enableTrace :: Bool
enableTrace = False
doTrace :: String -> b -> b
doTrace | enableTrace = trace
| otherwise = \_ x -> x
parseMIMEBody :: [MIMEParam] -> T.Text -> MIMEValue
parseMIMEBody headers_in body = result { mime_val_headers = headers }
where
result = case mimeType mty of
Multipart{} -> fst (parseMultipart mty body)
Message{} -> fst (parseMultipart mty body)
_ -> nullMIMEValue { mime_val_type = mty
, mime_val_disp = parseContentDisp headers
, mime_val_content = Single (processBody headers body)
}
headers = [ MIMEParam (T.toLower k) v | (MIMEParam k v) <- headers_in ]
mty = fromMaybe defaultType
(parseContentType =<< lookupField "content-type" (paramPairs headers))
defaultType :: Type
defaultType = Type { mimeType = Text "plain"
, mimeParams = [MIMEParam "charset" "us-ascii"]
}
parseContentDisp :: [MIMEParam] -> Maybe Disposition
parseContentDisp headers =
(processDisp . dropFoldingWSP) =<< lookupField "content-disposition" (paramPairs headers)
where
processDisp t | T.null t = Nothing
| T.null bs = Just $ Disposition { dispType = toDispType (T.toLower as)
, dispParams = []
}
| otherwise = Just $ Disposition { dispType = toDispType (T.toLower as)
, dispParams = processParams (parseParams bs)
}
where (as,bs) = T.break (\ch -> isSpace ch || ch == ';') t
processParams = map procP
where
procP (MIMEParam as val)
| "name" == asl = Name val
| "filename" == asl = Filename val
| "creation-date" == asl = CreationDate val
| "modification-date" == asl = ModDate val
| "read-date" == asl = ReadDate val
| "size" == asl = Size val
| otherwise = OtherParam asl val
where asl = T.toLower as
toDispType t = if t == "inline" then DispInline
else if t == "attachment" then DispAttachment
else if t == "form-data" then DispFormData
else DispOther t
paramPairs :: [MIMEParam] -> [(T.Text, T.Text)]
paramPairs = map paramPair
where
paramPair (MIMEParam a b) = (a,b)
processBody :: [MIMEParam] -> T.Text -> T.Text
processBody headers body =
case lookupField "content-transfer-encoding" $ paramPairs headers of
Nothing -> body
Just v -> T.pack $ decodeBody (T.unpack v) $ T.unpack body
normalizeCRLF :: T.Text -> T.Text
normalizeCRLF t
| T.null t = ""
| "\r\n" `T.isPrefixOf` t = "\r\n" <> normalizeCRLF (T.drop 2 t)
| any (`T.isPrefixOf` t) ["\r", "\n"] = "\r\n" <> normalizeCRLF (T.drop 1 t)
| otherwise = let (a,b) = T.break (`elem` "\r\n") t in a <> normalizeCRLF b
parseMIMEMessage :: T.Text -> MIMEValue
parseMIMEMessage entity =
case parseHeaders (normalizeCRLF entity) of
(as,bs) -> parseMIMEBody as bs
parseHeaders :: T.Text -> ([MIMEParam], T.Text)
parseHeaders str =
case findFieldName "" str of
Left (nm, rs) -> parseFieldValue nm (dropFoldingWSP rs)
Right body -> ([],body)
where
findFieldName acc t
| T.null t = Right ""
| "\r\n" `T.isPrefixOf` t = Right $ T.drop 2 t
| ":" `T.isPrefixOf` t = Left (T.reverse $ T.dropWhile isHSpace acc, T.drop 1 t)
| otherwise = findFieldName (T.take 1 t <> acc) $ T.drop 1 t
parseFieldValue nm xs
| T.null bs = ([MIMEParam nm as], "")
| otherwise = let (zs,ys) = parseHeaders bs in (MIMEParam nm as :zs, ys)
where
(as,bs) = takeUntilCRLF xs
parseMultipart :: Type -> T.Text -> (MIMEValue, T.Text)
parseMultipart mty body =
case lookupField "boundary" (paramPairs $ mimeParams mty) of
Nothing -> doTrace ("Multipart mime type, " ++ T.unpack (showType mty) ++
", has no required boundary parameter. Defaulting to text/plain") $
(nullMIMEValue{ mime_val_type = defaultType
, mime_val_disp = Nothing
, mime_val_content = Single body
}, "")
Just bnd -> (nullMIMEValue { mime_val_type = mty
, mime_val_disp = Nothing
, mime_val_content = Multi vals
}, rs)
where (vals,rs) = splitMulti bnd body
splitMulti :: T.Text -> T.Text -> ([MIMEValue], T.Text)
splitMulti bnd body_in =
-- Note: we insert a CRLF if it looks as if the boundary string starts
-- right off the bat. No harm done if this turns out to be incorrect.
let body | "--" `T.isPrefixOf` body_in = "\r\n" <> body_in
| otherwise = body_in
in case untilMatch dashBoundary body of
Nothing -> mempty
Just xs | "--" `T.isPrefixOf` xs -> ([], T.drop 2 xs)
| otherwise -> splitMulti1 (dropTrailer xs)
where
dashBoundary = ("\r\n--" <> bnd)
splitMulti1 xs
| T.null as && T.null bs = ([], "")
| T.null bs = ([parseMIMEMessage as],"")
| T.isPrefixOf "--" bs = ([parseMIMEMessage as], dropTrailer bs)
| otherwise = let (zs,ys) = splitMulti1 (dropTrailer bs)
in ((parseMIMEMessage as) : zs,ys)
where
(as,bs) = matchUntil dashBoundary xs
dropTrailer xs
| "\r\n" `T.isPrefixOf` xs1 = T.drop 2 xs1
| otherwise = xs1 -- hmm, flag an error?
where
xs1 = T.dropWhile isHSpace xs
parseMIMEType :: T.Text -> Maybe Type
parseMIMEType = parseContentType
parseContentType :: T.Text -> Maybe Type
parseContentType str
| T.null minor0 = doTrace ("unable to parse content-type: " ++ show str) $ Nothing
| otherwise = Just Type { mimeType = toType maj as
, mimeParams = parseParams (T.dropWhile isHSpace bs)
}
where
(maj, minor0) = T.break (=='/') (dropFoldingWSP str)
minor = T.drop 1 minor0
(as, bs) = T.break (\ ch -> isHSpace ch || isTSpecial ch) minor
toType a b = case lookupField (T.toLower a) mediaTypes of
Just ctor -> ctor b
_ -> Other a b
parseParams :: T.Text -> [MIMEParam]
parseParams t | T.null t = []
| ';' == T.head t = let (nm_raw, vs0) = T.break (=='=') (dropFoldingWSP $ T.tail t)
nm = T.toLower nm_raw in
if T.null vs0
then []
else let vs = T.tail vs0 in
if not (T.null vs) && T.head vs == '"'
then let vs1 = T.tail vs
(val, zs0) = T.break (=='"') vs1 in
if T.null zs0
then [MIMEParam nm val]
else MIMEParam nm val : parseParams (T.dropWhile isHSpace $ T.tail zs0)
else let (val, zs) = T.break (\ch -> isHSpace ch || isTSpecial ch) vs in
MIMEParam nm val : parseParams (T.dropWhile isHSpace zs)
| otherwise = doTrace ("Codec.MIME.Parse.parseParams: curious param value -- " ++ show t) []
mediaTypes :: [(T.Text, T.Text -> MIMEType)]
mediaTypes =
[ ("multipart", (Multipart . toMultipart))
, ("application", Application)
, ("audio", Audio)
, ("image", Image)
, ("message", Message)
, ("model", Model)
, ("text", Text)
, ("video", Video)
]
where toMultipart b = fromMaybe other (lookupField (T.toLower b) multipartTypes)
where other | T.isPrefixOf "x-" b = Extension b
| otherwise = OtherMulti b
multipartTypes :: [(T.Text, Multipart)]
multipartTypes =
[ ("alternative", Alternative)
, ("byteranges", Byteranges)
, ("digest", Digest)
, ("encrypted", Encrypted)
, ("form-data", FormData)
, ("mixed", Mixed)
, ("parallel", Parallel)
, ("related", Related)
, ("signed", Signed)
]
untilMatch :: T.Text -> T.Text -> Maybe T.Text
untilMatch a b | T.null a = Just b
| T.null b = Nothing
| a `T.isPrefixOf` b = Just $ T.drop (T.length a) b
| otherwise = untilMatch a $ T.tail b
matchUntil :: T.Text -> T.Text -> (T.Text, T.Text)
-- searching str; returning parts before str and after str
matchUntil str = second (T.drop $ T.length str) . T.breakOn str
{-
matchUntil' :: T.Text -> T.Text -> (T.Text, T.Text)
matchUntil' _ "" = ("", "")
matchUntil' str xs
| T.null xs = mempty
-- slow, but it'll do for now.
| str `T.isPrefixOf` xs = ("", T.drop (T.length str) xs)
| otherwise = let (as,bs) = matchUntil' str $ T.tail xs in (T.take 1 xs <> as, bs)
-}
isHSpace :: Char -> Bool
isHSpace c = c == ' ' || c == '\t'
isTSpecial :: Char -> Bool
isTSpecial x = x `elem` "()<>@,;:\\\"/[]?=" -- "
dropFoldingWSP :: T.Text -> T.Text
dropFoldingWSP t | T.null t = ""
| isHSpace (T.head t) = dropFoldingWSP $ T.tail t
| "\r\n" `T.isPrefixOf` t && not (T.null $ T.drop 2 t) && isHSpace (T.head $ T.drop 2 t)
= dropFoldingWSP $ T.drop 3 t
| otherwise = t
takeUntilCRLF :: T.Text -> (T.Text, T.Text)
takeUntilCRLF str = go "" str
where
go acc t | T.null t = (T.reverse (T.dropWhile isHSpace acc), "")
| "\r\n" `T.isPrefixOf` t && not (T.null $ T.drop 2 t) && isHSpace (T.head $ T.drop 2 t)
= go (" " <> acc) (T.drop 3 t)
| "\r\n" `T.isPrefixOf` t && not (T.null $ T.drop 2 t)
= (T.reverse (T.dropWhile isHSpace acc), T.drop 2 t)
| otherwise = go (T.take 1 t <> acc) $ T.tail t
-- case in-sensitive lookup of field names or attributes\/parameters.
lookupField :: T.Text -> [(T.Text,a)] -> Maybe a
lookupField n ns =
-- assume that inputs have been mostly normalized already
-- (i.e., lower-cased), but should the lookup fail fall back
-- to a second try where we do normalize before giving up.
case lookup n ns of
x@Just{} -> x
Nothing ->
let nl = T.toLower n in
fmap snd $ L.find ((nl==) . T.toLower . fst) ns
|