mime.lua

   1 #!/usr/bin/lua
   2
   3 require 'lpeg'
   4
   5 -- MIME implementation in Lua
   6
   7 -- XXX: at section 3.4
   8
   9
  10 local string = string
  11 local type = type
  12 local tostring = tostring
  13 local select = select
  14 local setfenv = setfenv
  15 local CRLF = '\r\n'
  16 local lpeg = lpeg
  17 local print = print
  18
  19 -- this collects both the table of named captures and the
  20 -- substitution captureof the given pattern
  21 lpeg.Cst = function (patt)
  22         return (#lpeg.Cs(patt) * lpeg.Ct(patt)) / function (s, t)
  23                 local ret = {}
  24                 ret [1] = s
  25                 for k, v in pairs(t) do
  26                         if type(k)~='number' then
  27                                 ret[k] = v
  28                         end
  29                 end
  30                 return ret
  31         end
  32 end
  33
  34 -- rfc2822 Section 2.1:
  35 --
  36 -- A message consists of header fields (collectively called "the header
  37 -- of the message") followed, optionally, by a body.  The header is a
  38 -- sequence of lines of characters with special syntax as defined in
  39 -- this standard. The body is simply a sequence of characters that
  40 -- follows the header and is separated from the header by an empty line
  41 -- (i.e., a line with nothing preceding the CRLF).
  42 --
  43 local split_content = function (c)
  44         -- NOTE: there is no guarantee that the CRLF separating header
  45         -- and body is the first one. However no headers admit a blank
  46         -- line so far
  47         local h, b = string.match(c, '^(.-)'..CRLF..CRLF..'(.*)$')
  48         assert(type(h)=='string' and type(b)=='string')
  49         return h, b
  50 end
  51
  52 -- rfc2822 Section 2.2.3
  53 --
  54 -- The process of moving from this folded multiple-line representation
  55 -- of a header field to its single line representation is called
  56 -- "unfolding". Unfolding is accomplished by simply removing any CRLF
  57 -- that is immediately followed by WSP.  Each header field should be
  58 -- treated in its unfolded form for further syntactic and semantic
  59 -- evaluation.
  60 --
  61 local unfold_header = function (h)
  62         return string.gsub(h, CRLF..'[ \t]', ' ')
  63 end
  64
  65 -- rfc2822 Section 2.1
  66 --
  67 -- At the most basic level, a message is a series of characters.  A
  68 -- message that is conformant with this standard is comprised of
  69 -- characters with values in the range 1 through 127 and interpreted as
  70 -- US-ASCII characters [ASCII].  For brevity, this document sometimes
  71 -- refers to this range of characters as simply "US-ASCII characters".
  72 --
  73
  74 -- rfc2822 Section 2.1.1
  75 --
  76 -- There are two limits that this standard places on the number of
  77 -- characters in a line. Each line of characters MUST be no more than
  78 -- 998 characters, and SHOULD be no more than 78 characters, excluding
  79 -- the CRLF.
  80 --
  81 local check_content = function (c)
  82         local err, warn
  83         -- TODO: implement in a portable way
  84         return err, warn
  85 end
  86
  87 -- rfc2822 Section 2.3
  88 --
  89 -- The body of a message is simply lines of US-ASCII characters.  The
  90 -- only two limitations on the body are as follows:
  91 --
  92 -- - CR and LF MUST only occur together as CRLF; they MUST NOT appear
  93 -- independently in the body.
  94 --
  95 -- - Lines of characters in the body MUST be limited to 998 characters,
  96 --
  97 -- and SHOULD be limited to 78 characters, excluding the CRLF.
  98 --
  99 local check_body = function (b)
 100         local err, warn
 101         -- TODO: implement
 102         return err, warn
 103 end
 104
 105 -- some core values (not implemented literally)
 106 -- see RFC 2234 Section 6.1
 107 local core_values = {
 108         ["ALPHA"] = lpeg.R("AZ", "az"),
 109         ["BIT"] = lpeg.P("0") + lpeg.P("1"),
 110         ["CHAR"] = lpeg.R("\01\127"),
 111         ["CR"] = lpeg.P("\13"),
 112         ["CRLF"] = lpeg.P("\13\10"),
 113         ["DIGIT"] = lpeg.R("09"),
 114         ["LF"] = lpeg.P("\10"),
 115         ["WSP"] = lpeg.S("\32\09"),
 116 }
 117
 118 -- lexical tokens used in the specification
 119 -- TODO: could be optimized
 120 -- check for non-obfuscated optimizations
 121 -- TODO: write in a less obfuscated way
 122 -- check if parenthesis can be removed
 123 local lex_tokens = function ()
 124         setfenv(1, lpeg)
 125         return {
 126         -- control characters without whitespaces
 127         ["NO-WS-CTL"] = R("\1\8") +
 128         P("\11") +
 129         P("\12") +
 130         R("\14\13") +
 131         P("\127"), -- RFC 2822 Section 3.2.1
 132         -- a character in a text
 133         ["text"] =      R("\1\9") +
 134         P("\11") +
 135         P("\12") +
 136         R("\14\127"), -- RFC 2822 Section 3.2.1
 137         -- a special character
 138         ["specials"] =  P("(") + P(")") +
 139         P("<") + P(">") +
 140         P("[") + P("]") +
 141         P(":") + P(";") +
 142         P("@") + P("\\") +
 143         P(",") + P(".") +
 144         P("\""), -- RFC 2822 Section 3.2.1
 145         -- a quoted pair should return only the second character
 146         ["quoted-pair"] = (P("\\") * C(V("text"))) + V("obs-qp"), -- RFC 2822 Section 3.2.2
 147         -- a folding white space (a whitespace that can include a CRLF)
 148         -- should be substituted by a single whitespace
 149         ["FWS"] = (((V("WSP")^0 * V("CRLF"))^-1 * V("WSP")^1) + V("obs-FWS")) / " ", -- RFC 2822 Section 3.2.3
 150         -- a text character allowed inside a comment
 151         ["ctext"] =     V("NO-WS-CTL") +
 152         R("\33\39") +
 153         R("\42\91") +
 154         R("\93\126"), -- RFC 2822 Section 3.2.3
 155         -- the content of a comment (comments can nest)
 156         ["ccontent"] =  V("ctext") + V("quoted-pair") + V("comment"), -- RFC 2822 Section 3.2.3
 157         -- an actual comment
 158         -- should be substituted by a single whitespace
 159         ["comment"] =   (P("(") * (V("FWS")^-1 * V("ccontent"))^0 * V("FWS")^-1 * P(")")) / " ", -- RFC 2822 Section 3.2.3
 160         -- a comment or a folding white space
 161         -- should be substituted by a single whitespace
 162         --
 163         -- Folding white spaces should not be placed in a way that
 164         -- creates lines containing only whitespaces.
 165         -- This requirement Is not necessarily enforced by this grammar
 166         ["CFWS"] =      ( (V("FWS") + (V("comment")*V("FWS")^-1)) * (V("comment")*V("FWS")^-1)^0 ) / " ", -- RFC 2822 Section 3.2.3
 167         -- character that can appear in an atom
 168         ["atext"] =     V("ALPHA") + V("DIGIT") +
 169         P("!") + P("#") +
 170         P("$") + P("%") +
 171         P("&") + P("'") +
 172         P("*") + P("+") +
 173         P("-") + P("/") +
 174         P("=") + P("?") +
 175         P("^") + P("_") +
 176         P("`") + P("{") +
 177         P("|") + P("}") +
 178         P("~"), -- RFC 2822 Section 3.2.4
 179         -- an atom is equal to the content only discarding comments and whitespace
 180         ["atom"] =      V("CFWS")^-1 * C(V("atext")^1) * V("CFWS")^-1, -- RFC 2822 Section 3.2.4
 181         -- an atom with dots is only the content, discarding CFWSs
 182         ["dot-atom"] =  Cs((V("CFWS")^-1/'') * C(V("dot-atom-text")) * (V("CFWS")^-1/'')), -- RFC 2822 Section 3.2.4
 183         -- the content of an atom text with dots
 184         ["dot-atom-text"] =     V("atext")^1 * (P(".") * V("atext")^1)^0, -- RFC 2822 Section 3.2.4
 185         -- character that can appear in a quoted string
 186         ["qtext"] =     V("NO-WS-CTL") +
 187         P("\33") +
 188         R("\35\91") +
 189         R("\93\126"), -- RFC 2822 Section 3.2.5
 190         -- character or quoted pair (both can appear in a quoted string)
 191         -- it is equivalent to the character itself or to the result of the
 192         -- quoted pair
 193         ["qcontent"] =  C(V("qtext")) + V("quoted-pair"),  -- RFC 2822 Section 3.2.5
 194         -- a quoted string is equal to its content
 195         ["quoted-string"] =     V("CFWS")^-1 *
 196         P("\"") * Cs((V("FWS")^-1 * V("qcontent"))^0 * V("FWS")^-1) * P("\"") * V("CFWS")^-1, -- RFC 2822 Section 3.2.5
 197         -- unstructured patterns for unspecified headers
 198         -- what should these be equal to?
 199         --
 200         -- an generic word
 201         ["word"] =      V("atom") + V("quoted-string"), -- RFC 2822 Section 3.2.6
 202         -- an generic phrase
 203         ["phrase"] =    V("word")^1 + V("obs-phrase"), -- RFC 2822 Section 3.2.6
 204         -- a character for unstructured text
 205         ["utext"] =     V("NO-WS-CTL") + R("\33\126") + V("obs-utext"), -- RFC 2822 Section 3.2.6
 206         -- an unstructured text
 207         ["unstructured"] =      (V("FWS")^-1 * V("utext"))^0 * V("FWS")^-1, -- RFC 2822 Section 3.2.6
 208 }
 209 end
 210
 211 local show = my.show
 212 local deb = function(...) for i=1,select('#',...) do local v=select(i,...)print(type(v), v)end end
 213
 214 local date_time = function ()
 215         setfenv(1, lpeg)
 216         return {
 217         -- date and time specification
 218         -- dates and times should be valid
 219         -- this grammar does not enforce this yet
 220         ["date-time"] = Cst(( V"day-of-week" * P"," )^-1 * V"date" * V"FWS" * V"time" * V"CFWS"^-1), -- RFC 2822 Section 3.3
 221         ["day-of-week"] = ( V"FWS"^-1 * V"day-name" ) + V"obs-day-of-week", -- RFC 2822 Section 3.3
 222         ["day-name"] = Cg(P"Mon" + P"Tue" + P"Wed" + P"Thu" + P"Fri" + P"Sat" + P"Sun", "weekday"), -- RFC 2822 Section 3.3
 223         ["date"] = V"day" * V"month" * V"year", -- RFC 2822 Section 3.3
 224         ["year"] = Cg(C(V"DIGIT"^4), "year") + V"obs-year", -- RFC 2822 Section 3.3
 225         ["month"] = (V"FWS" * V"month-name" * V"FWS") + V"obs-month", -- RFC 2822 Section 3.3
 226         ["month-name"] = Cg(C"Jan" + C"Feb" + C"Mar" + C"Apr" + C"May" + C"Jun" + C"Jul" + C"Aug" + C"Sep" + C"Oct" + C"Nov" + C"Dec", "month"), -- RFC 2822 Section 3.3
 227         ["day"] = (V"FWS"^-1 * Cg(C(V"DIGIT" * V"DIGIT"^-1), "day")) + V"obs-day", -- RFC 2822 Section 3.3
 228         ["time"] = V"time-of-day" * V"FWS" * V"zone", -- RFC 2822 Section 3.3
 229         ["time-of-day"] = V"hour" * P":" * V"minute" * (P":" * V"second")^-1, -- RFC 2822 Section 3.3
 230         ["hour"] = Cg(C(V"DIGIT" * V"DIGIT"), "hour") + V"obs-hour", -- RFC 2822 Section 3.3
 231         ["minute"] = Cg((V"DIGIT" * V"DIGIT"), "minute") + V"obs-minute", -- RFC 2822 Section 3.3
 232         ["second"] = Cg((V"DIGIT" * V"DIGIT"), "second") + V"obs-second", -- RFC 2822 Section 3.3
 233         ["zone"] = Cg( (P"+" + P"-") * V"DIGIT" * V"DIGIT" * V"DIGIT" * V"DIGIT", "zone" ) + V"obs-zone", -- RFC 2822 Section 3.3
 234 }
 235 end
 236
 237 local address = function ()
 238         setfenv(1, lpeg)
 239         return {
 240         -- address specification
 241         -- dates and times should be valid
 242         -- this grammar does not enforce this yet
 243         ["address"] = V"mailbox" + V"group", -- RFC 2822 Section 3.4
 244         ["mailbox"] = V"name-addr" + V"addr-spec", -- RFC 2822 Section 3.4
 245         ["name-addr"] = V"display-name"^-1 * V"angle-addr", -- RFC 2822 Section 3.4
 246         ["angle-addr"] = (V"CFWS"^-1 * P"<" * V"addr-spec" * P">" * V"CFWS"^-1) + V"obs-angle-addr", -- RFC 2822 Section 3.4
 247         ["group"] = V"display-name" * P":" * (V"mailbox-list" + V"CFWS") * P";" * V"CFWS"^-1, -- RFC 2822 Section 3.4
 248         ["display-name"] = V"phrase", -- RFC 2822 Section 3.4
 249         ["mailbox-list"] = (V"mailbox" * (P"," * V"mailbox")^0) + V"obs-mailbox-list", -- RFC 2822 Section 3.4
 250         ["address-list"] = (V"address" * (P"," * V"address")^0) + V"obs-address-list", -- RFC 2822 Section 3.4
 251         -- address specification (name@host.domain)
 252         ["addr-spec"] = V"local-part" * P"@" * V"domain", -- RFC 2822 Section 3.4.1
 253         ["local-part"] = V"dot-atom" + V"quoted-string" + V"obs-local-part", -- RFC 2822 Section 3.4.1
 254         ["domain"] = V"dot-atom" + V"domain-literal" + V"obs-domain", -- RFC 2822 Section 3.4.1
 255         ["domain-literal"] = V"CFWS"^-1 * P"[" * (V"FWS"^-1 * V"dcontent")^0 * V"FWS"^-1 * P"]" * V"CFWS"^-1, -- RFC 2822 Section 3.4.1
 256         ["dcontent"] = V"dtext" + V"quoted-pair", -- RFC 2822 Section 3.4.1
 257         ["dtext"] = V"NO-WS-CTL" + R"\33\90" + R"\94\126", -- RFC 2822 Section 3.4.1
 258 }
 259 end
 260
 261 local overall_message = function ()
 262         setfenv(1, lpeg)
 263         return {
 264         -- overall message specification
 265         --
 266         -- RRC 2822 Section 3.5:
 267         --
 268         -- A message consists of header fields, optionally followed by a message
 269         -- body.  Lines in a message MUST be a maximum of 998 characters
 270         -- excluding the CRLF, but it is RECOMMENDED that lines be limited to 78
 271         -- characters excluding the CRLF.
 272         --
 273         -- this grammar does not enforce the 78 chars limit yet
 274         ["message"] = (V"fields" + V"obs-fields") * (V"CRLF" * V"body")^-1, -- RFC 2822 Section 3.5
 275         ["body"] = (V"text"^-998 * V"CRLF")^0 * V"text"^-998, -- RFC 2822 Section 3.5
 276 }
 277 end
 278
 279 local fields = function ()
 280         setfenv(1, lpeg)
 281         return {
 282         -- spec for all fields in the header
 283         --
 284         -- RRC 2822 Section 3.6
 285         --
 286         --the "fields" rule is incomplete
 287         ["fields"] = -- trace fields missing
 288                         (V"orig-date" + V"from" + V"sender" + V"reply-to" + V"to" + V"cc" + V"bcc"
 289                         + V"message-id" + V"in-reply-to" + V"to" + V"cc" + V"bcc")^0, -- RFC 2822 Section 3.6
 290         -- origination date field: the time at which the author decided
 291         -- the message was ready to be sent (e.g. hit the button)
 292         ["orig-date"] = P"Date:" * V"date-time" * V"CRLF", -- RFC 2822 Section 3.6.1
 293         -- the author of the message
 294         ["from"] = P"From:" * V"mailbox-list" * V"CRLF", -- RFC 2822 Section 3.6.2
 295         -- the actual agent that sent the message
 296         ["sender"] = P"Sender:" * V"mailbox" * V"CRLF", -- RFC 2822 Section 3.6.2
 297         -- address which should be replied to
 298         -- see Section 3.6.3 for forming replies
 299         ["reply-to"] = P"Reply-To:" * V"address-list" * V"CRLF", -- RFC 2822 Section 3.6.2
 300         -- The "To:" field contains the address(es) of the primary recipient(s)
 301         -- of the message.
 302         ["to"] = P"To:" * V"address-list" * V"CRLF", -- RFC 2822 Section 3.6.3
 303         -- The "Cc:" field contains the addresses of others who are to receive
 304         -- the message, though the content of the message may not be directed
 305         -- at them.
 306         ["cc"] = P"Cc:" * V"address-list" * V"CRLF", -- RFC 2822 Section 3.6.3
 307         -- The "Bcc:" field contains addresses of recipients of the message
 308         -- whose addresses are not to be revealed to other recipients of the
 309         -- message.
 310         ["bcc"] = P"Bcc:" * (V"address-list" + V"CFWS") * V"CRLF", -- RFC 2822 Section 3.6.3
 311         ["message-id"] = P"Message-Id:" * V"msg-id" * V"CRLF", -- RFC 2822 Section 3.6.4
 312         ["in-reply-to"] = P"In-Reply-To:" * V"msg-id"^1 * V"CRLF", -- RFC 2822 Section 3.6.4
 313         ["references"] = P"References:" * V"msg-id"^1 * V"CRLF", -- RFC 2822 Section 3.6.4
 314         ["msg-id"] = V"CFWS"^-1 * P"<" * V"id-left" * P"@" * V"id-right" * P">" * V"CFWS"^-1, -- RFC 2822 Section 3.6.4
 315         ["id-left"] = V"dot-atom-text" + V"no-fold-quote" + V"obs-id-left", -- RFC 2822 Section 3.6.4
 316         ["id-right"] = V"dot-atom-text" + V"no-fold-literal" + V"obs-id-right", -- RFC 2822 Section 3.6.4
 317         ["no-fold-quote"] = P"\"" * (V"qtext" + V"quoted-pair")^0 * P"\"", -- RFC 2822 Section 3.6.4
 318         ["no-fold-literal"] = P"[" * (V"dtext" + V"quoted-pair")^0 * P"]", -- RFC 2822 Section 3.6.4
 319 }
 320 end
 321
 322 local obs_strict = {
 323         ["obs-FWS"] = lpeg.P(false),
 324         ["obs-qp"] = lpeg.P(false),
 325         ["obs-phrase"] = lpeg.P(false),
 326         ["obs-utext"] = lpeg.P(false),
 327         ["obs-day-of-week"] = lpeg.P(false),
 328         ["obs-year"] = lpeg.P(false),
 329         ["obs-month"] = lpeg.P(false),
 330         ["obs-day"] = lpeg.P(false),
 331         ["obs-hour"] = lpeg.P(false),
 332         ["obs-minute"] = lpeg.P(false),
 333         ["obs-second"] = lpeg.P(false),
 334         ["obs-zone"] = lpeg.P(false),
 335         ["obs-angle-addr"] = lpeg.P(false),
 336         ["obs-mailbox-list"] = lpeg.P(false),
 337         ["obs-address-list"] = lpeg.P(false),
 338         ["obs-local-part"] = lpeg.P(false),
 339         ["obs-domain"] = lpeg.P(false),
 340         ["obs-fields"] = lpeg.P(false),
 341         ["obs-id-left"] = lpeg.P(false),
 342         ["obs-id-right"] = lpeg.P(false),
 343 }
 344
 345 local os = setmetatable({}, {__index = function() return lpeg.P(false) end})
 346
 347 local join_set = function (...)
 348         local n = select('#', ...)
 349         local ret = {}
 350         for i = 1, n do
 351                 local t = select(i, ...)
 352                 if type(t)=='table' then
 353                         for k, v in pairs(t) do
 354                                 ret[k] = v
 355                         end
 356                 elseif type(t)=='string' then
 357                         ret[1] = t
 358                 elseif type(t)=='boolean' then
 359                         -- TODO: check no overwrite
 360                 else
 361                         error('join_set: bad argument number '..i..' of type '..type(t))
 362                 end
 363         end
 364         return ret
 365 end
 366
 367 local gr = join_set(core_values, lex_tokens(), date_time(), address(), overall_message(), fields(), obs_strict)
 368
 369 return gr
 370