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