|
1 """RFC 2822 message manipulation. |
|
2 |
|
3 Note: This is only a very rough sketch of a full RFC-822 parser; in particular |
|
4 the tokenizing of addresses does not adhere to all the quoting rules. |
|
5 |
|
6 Note: RFC 2822 is a long awaited update to RFC 822. This module should |
|
7 conform to RFC 2822, and is thus mis-named (it's not worth renaming it). Some |
|
8 effort at RFC 2822 updates have been made, but a thorough audit has not been |
|
9 performed. Consider any RFC 2822 non-conformance to be a bug. |
|
10 |
|
11 RFC 2822: http://www.faqs.org/rfcs/rfc2822.html |
|
12 RFC 822 : http://www.faqs.org/rfcs/rfc822.html (obsolete) |
|
13 |
|
14 Directions for use: |
|
15 |
|
16 To create a Message object: first open a file, e.g.: |
|
17 |
|
18 fp = open(file, 'r') |
|
19 |
|
20 You can use any other legal way of getting an open file object, e.g. use |
|
21 sys.stdin or call os.popen(). Then pass the open file object to the Message() |
|
22 constructor: |
|
23 |
|
24 m = Message(fp) |
|
25 |
|
26 This class can work with any input object that supports a readline method. If |
|
27 the input object has seek and tell capability, the rewindbody method will |
|
28 work; also illegal lines will be pushed back onto the input stream. If the |
|
29 input object lacks seek but has an `unread' method that can push back a line |
|
30 of input, Message will use that to push back illegal lines. Thus this class |
|
31 can be used to parse messages coming from a buffered stream. |
|
32 |
|
33 The optional `seekable' argument is provided as a workaround for certain stdio |
|
34 libraries in which tell() discards buffered data before discovering that the |
|
35 lseek() system call doesn't work. For maximum portability, you should set the |
|
36 seekable argument to zero to prevent that initial \code{tell} when passing in |
|
37 an unseekable object such as a a file object created from a socket object. If |
|
38 it is 1 on entry -- which it is by default -- the tell() method of the open |
|
39 file object is called once; if this raises an exception, seekable is reset to |
|
40 0. For other nonzero values of seekable, this test is not made. |
|
41 |
|
42 To get the text of a particular header there are several methods: |
|
43 |
|
44 str = m.getheader(name) |
|
45 str = m.getrawheader(name) |
|
46 |
|
47 where name is the name of the header, e.g. 'Subject'. The difference is that |
|
48 getheader() strips the leading and trailing whitespace, while getrawheader() |
|
49 doesn't. Both functions retain embedded whitespace (including newlines) |
|
50 exactly as they are specified in the header, and leave the case of the text |
|
51 unchanged. |
|
52 |
|
53 For addresses and address lists there are functions |
|
54 |
|
55 realname, mailaddress = m.getaddr(name) |
|
56 list = m.getaddrlist(name) |
|
57 |
|
58 where the latter returns a list of (realname, mailaddr) tuples. |
|
59 |
|
60 There is also a method |
|
61 |
|
62 time = m.getdate(name) |
|
63 |
|
64 which parses a Date-like field and returns a time-compatible tuple, |
|
65 i.e. a tuple such as returned by time.localtime() or accepted by |
|
66 time.mktime(). |
|
67 |
|
68 See the class definition for lower level access methods. |
|
69 |
|
70 There are also some utility functions here. |
|
71 """ |
|
72 # Cleanup and extensions by Eric S. Raymond <esr@thyrsus.com> |
|
73 |
|
74 import time |
|
75 |
|
76 from warnings import warnpy3k |
|
77 warnpy3k("in 3.x, rfc822 has been removed in favor of the email package", |
|
78 stacklevel=2) |
|
79 |
|
80 __all__ = ["Message","AddressList","parsedate","parsedate_tz","mktime_tz"] |
|
81 |
|
82 _blanklines = ('\r\n', '\n') # Optimization for islast() |
|
83 |
|
84 |
|
85 class Message: |
|
86 """Represents a single RFC 2822-compliant message.""" |
|
87 |
|
88 def __init__(self, fp, seekable = 1): |
|
89 """Initialize the class instance and read the headers.""" |
|
90 if seekable == 1: |
|
91 # Exercise tell() to make sure it works |
|
92 # (and then assume seek() works, too) |
|
93 try: |
|
94 fp.tell() |
|
95 except (AttributeError, IOError): |
|
96 seekable = 0 |
|
97 self.fp = fp |
|
98 self.seekable = seekable |
|
99 self.startofheaders = None |
|
100 self.startofbody = None |
|
101 # |
|
102 if self.seekable: |
|
103 try: |
|
104 self.startofheaders = self.fp.tell() |
|
105 except IOError: |
|
106 self.seekable = 0 |
|
107 # |
|
108 self.readheaders() |
|
109 # |
|
110 if self.seekable: |
|
111 try: |
|
112 self.startofbody = self.fp.tell() |
|
113 except IOError: |
|
114 self.seekable = 0 |
|
115 |
|
116 def rewindbody(self): |
|
117 """Rewind the file to the start of the body (if seekable).""" |
|
118 if not self.seekable: |
|
119 raise IOError, "unseekable file" |
|
120 self.fp.seek(self.startofbody) |
|
121 |
|
122 def readheaders(self): |
|
123 """Read header lines. |
|
124 |
|
125 Read header lines up to the entirely blank line that terminates them. |
|
126 The (normally blank) line that ends the headers is skipped, but not |
|
127 included in the returned list. If a non-header line ends the headers, |
|
128 (which is an error), an attempt is made to backspace over it; it is |
|
129 never included in the returned list. |
|
130 |
|
131 The variable self.status is set to the empty string if all went well, |
|
132 otherwise it is an error message. The variable self.headers is a |
|
133 completely uninterpreted list of lines contained in the header (so |
|
134 printing them will reproduce the header exactly as it appears in the |
|
135 file). |
|
136 """ |
|
137 self.dict = {} |
|
138 self.unixfrom = '' |
|
139 self.headers = lst = [] |
|
140 self.status = '' |
|
141 headerseen = "" |
|
142 firstline = 1 |
|
143 startofline = unread = tell = None |
|
144 if hasattr(self.fp, 'unread'): |
|
145 unread = self.fp.unread |
|
146 elif self.seekable: |
|
147 tell = self.fp.tell |
|
148 while 1: |
|
149 if tell: |
|
150 try: |
|
151 startofline = tell() |
|
152 except IOError: |
|
153 startofline = tell = None |
|
154 self.seekable = 0 |
|
155 line = self.fp.readline() |
|
156 if not line: |
|
157 self.status = 'EOF in headers' |
|
158 break |
|
159 # Skip unix From name time lines |
|
160 if firstline and line.startswith('From '): |
|
161 self.unixfrom = self.unixfrom + line |
|
162 continue |
|
163 firstline = 0 |
|
164 if headerseen and line[0] in ' \t': |
|
165 # It's a continuation line. |
|
166 lst.append(line) |
|
167 x = (self.dict[headerseen] + "\n " + line.strip()) |
|
168 self.dict[headerseen] = x.strip() |
|
169 continue |
|
170 elif self.iscomment(line): |
|
171 # It's a comment. Ignore it. |
|
172 continue |
|
173 elif self.islast(line): |
|
174 # Note! No pushback here! The delimiter line gets eaten. |
|
175 break |
|
176 headerseen = self.isheader(line) |
|
177 if headerseen: |
|
178 # It's a legal header line, save it. |
|
179 lst.append(line) |
|
180 self.dict[headerseen] = line[len(headerseen)+1:].strip() |
|
181 continue |
|
182 else: |
|
183 # It's not a header line; throw it back and stop here. |
|
184 if not self.dict: |
|
185 self.status = 'No headers' |
|
186 else: |
|
187 self.status = 'Non-header line where header expected' |
|
188 # Try to undo the read. |
|
189 if unread: |
|
190 unread(line) |
|
191 elif tell: |
|
192 self.fp.seek(startofline) |
|
193 else: |
|
194 self.status = self.status + '; bad seek' |
|
195 break |
|
196 |
|
197 def isheader(self, line): |
|
198 """Determine whether a given line is a legal header. |
|
199 |
|
200 This method should return the header name, suitably canonicalized. |
|
201 You may override this method in order to use Message parsing on tagged |
|
202 data in RFC 2822-like formats with special header formats. |
|
203 """ |
|
204 i = line.find(':') |
|
205 if i > 0: |
|
206 return line[:i].lower() |
|
207 return None |
|
208 |
|
209 def islast(self, line): |
|
210 """Determine whether a line is a legal end of RFC 2822 headers. |
|
211 |
|
212 You may override this method if your application wants to bend the |
|
213 rules, e.g. to strip trailing whitespace, or to recognize MH template |
|
214 separators ('--------'). For convenience (e.g. for code reading from |
|
215 sockets) a line consisting of \r\n also matches. |
|
216 """ |
|
217 return line in _blanklines |
|
218 |
|
219 def iscomment(self, line): |
|
220 """Determine whether a line should be skipped entirely. |
|
221 |
|
222 You may override this method in order to use Message parsing on tagged |
|
223 data in RFC 2822-like formats that support embedded comments or |
|
224 free-text data. |
|
225 """ |
|
226 return False |
|
227 |
|
228 def getallmatchingheaders(self, name): |
|
229 """Find all header lines matching a given header name. |
|
230 |
|
231 Look through the list of headers and find all lines matching a given |
|
232 header name (and their continuation lines). A list of the lines is |
|
233 returned, without interpretation. If the header does not occur, an |
|
234 empty list is returned. If the header occurs multiple times, all |
|
235 occurrences are returned. Case is not important in the header name. |
|
236 """ |
|
237 name = name.lower() + ':' |
|
238 n = len(name) |
|
239 lst = [] |
|
240 hit = 0 |
|
241 for line in self.headers: |
|
242 if line[:n].lower() == name: |
|
243 hit = 1 |
|
244 elif not line[:1].isspace(): |
|
245 hit = 0 |
|
246 if hit: |
|
247 lst.append(line) |
|
248 return lst |
|
249 |
|
250 def getfirstmatchingheader(self, name): |
|
251 """Get the first header line matching name. |
|
252 |
|
253 This is similar to getallmatchingheaders, but it returns only the |
|
254 first matching header (and its continuation lines). |
|
255 """ |
|
256 name = name.lower() + ':' |
|
257 n = len(name) |
|
258 lst = [] |
|
259 hit = 0 |
|
260 for line in self.headers: |
|
261 if hit: |
|
262 if not line[:1].isspace(): |
|
263 break |
|
264 elif line[:n].lower() == name: |
|
265 hit = 1 |
|
266 if hit: |
|
267 lst.append(line) |
|
268 return lst |
|
269 |
|
270 def getrawheader(self, name): |
|
271 """A higher-level interface to getfirstmatchingheader(). |
|
272 |
|
273 Return a string containing the literal text of the header but with the |
|
274 keyword stripped. All leading, trailing and embedded whitespace is |
|
275 kept in the string, however. Return None if the header does not |
|
276 occur. |
|
277 """ |
|
278 |
|
279 lst = self.getfirstmatchingheader(name) |
|
280 if not lst: |
|
281 return None |
|
282 lst[0] = lst[0][len(name) + 1:] |
|
283 return ''.join(lst) |
|
284 |
|
285 def getheader(self, name, default=None): |
|
286 """Get the header value for a name. |
|
287 |
|
288 This is the normal interface: it returns a stripped version of the |
|
289 header value for a given header name, or None if it doesn't exist. |
|
290 This uses the dictionary version which finds the *last* such header. |
|
291 """ |
|
292 return self.dict.get(name.lower(), default) |
|
293 get = getheader |
|
294 |
|
295 def getheaders(self, name): |
|
296 """Get all values for a header. |
|
297 |
|
298 This returns a list of values for headers given more than once; each |
|
299 value in the result list is stripped in the same way as the result of |
|
300 getheader(). If the header is not given, return an empty list. |
|
301 """ |
|
302 result = [] |
|
303 current = '' |
|
304 have_header = 0 |
|
305 for s in self.getallmatchingheaders(name): |
|
306 if s[0].isspace(): |
|
307 if current: |
|
308 current = "%s\n %s" % (current, s.strip()) |
|
309 else: |
|
310 current = s.strip() |
|
311 else: |
|
312 if have_header: |
|
313 result.append(current) |
|
314 current = s[s.find(":") + 1:].strip() |
|
315 have_header = 1 |
|
316 if have_header: |
|
317 result.append(current) |
|
318 return result |
|
319 |
|
320 def getaddr(self, name): |
|
321 """Get a single address from a header, as a tuple. |
|
322 |
|
323 An example return value: |
|
324 ('Guido van Rossum', 'guido@cwi.nl') |
|
325 """ |
|
326 # New, by Ben Escoto |
|
327 alist = self.getaddrlist(name) |
|
328 if alist: |
|
329 return alist[0] |
|
330 else: |
|
331 return (None, None) |
|
332 |
|
333 def getaddrlist(self, name): |
|
334 """Get a list of addresses from a header. |
|
335 |
|
336 Retrieves a list of addresses from a header, where each address is a |
|
337 tuple as returned by getaddr(). Scans all named headers, so it works |
|
338 properly with multiple To: or Cc: headers for example. |
|
339 """ |
|
340 raw = [] |
|
341 for h in self.getallmatchingheaders(name): |
|
342 if h[0] in ' \t': |
|
343 raw.append(h) |
|
344 else: |
|
345 if raw: |
|
346 raw.append(', ') |
|
347 i = h.find(':') |
|
348 if i > 0: |
|
349 addr = h[i+1:] |
|
350 raw.append(addr) |
|
351 alladdrs = ''.join(raw) |
|
352 a = AddressList(alladdrs) |
|
353 return a.addresslist |
|
354 |
|
355 def getdate(self, name): |
|
356 """Retrieve a date field from a header. |
|
357 |
|
358 Retrieves a date field from the named header, returning a tuple |
|
359 compatible with time.mktime(). |
|
360 """ |
|
361 try: |
|
362 data = self[name] |
|
363 except KeyError: |
|
364 return None |
|
365 return parsedate(data) |
|
366 |
|
367 def getdate_tz(self, name): |
|
368 """Retrieve a date field from a header as a 10-tuple. |
|
369 |
|
370 The first 9 elements make up a tuple compatible with time.mktime(), |
|
371 and the 10th is the offset of the poster's time zone from GMT/UTC. |
|
372 """ |
|
373 try: |
|
374 data = self[name] |
|
375 except KeyError: |
|
376 return None |
|
377 return parsedate_tz(data) |
|
378 |
|
379 |
|
380 # Access as a dictionary (only finds *last* header of each type): |
|
381 |
|
382 def __len__(self): |
|
383 """Get the number of headers in a message.""" |
|
384 return len(self.dict) |
|
385 |
|
386 def __getitem__(self, name): |
|
387 """Get a specific header, as from a dictionary.""" |
|
388 return self.dict[name.lower()] |
|
389 |
|
390 def __setitem__(self, name, value): |
|
391 """Set the value of a header. |
|
392 |
|
393 Note: This is not a perfect inversion of __getitem__, because any |
|
394 changed headers get stuck at the end of the raw-headers list rather |
|
395 than where the altered header was. |
|
396 """ |
|
397 del self[name] # Won't fail if it doesn't exist |
|
398 self.dict[name.lower()] = value |
|
399 text = name + ": " + value |
|
400 for line in text.split("\n"): |
|
401 self.headers.append(line + "\n") |
|
402 |
|
403 def __delitem__(self, name): |
|
404 """Delete all occurrences of a specific header, if it is present.""" |
|
405 name = name.lower() |
|
406 if not name in self.dict: |
|
407 return |
|
408 del self.dict[name] |
|
409 name = name + ':' |
|
410 n = len(name) |
|
411 lst = [] |
|
412 hit = 0 |
|
413 for i in range(len(self.headers)): |
|
414 line = self.headers[i] |
|
415 if line[:n].lower() == name: |
|
416 hit = 1 |
|
417 elif not line[:1].isspace(): |
|
418 hit = 0 |
|
419 if hit: |
|
420 lst.append(i) |
|
421 for i in reversed(lst): |
|
422 del self.headers[i] |
|
423 |
|
424 def setdefault(self, name, default=""): |
|
425 lowername = name.lower() |
|
426 if lowername in self.dict: |
|
427 return self.dict[lowername] |
|
428 else: |
|
429 text = name + ": " + default |
|
430 for line in text.split("\n"): |
|
431 self.headers.append(line + "\n") |
|
432 self.dict[lowername] = default |
|
433 return default |
|
434 |
|
435 def has_key(self, name): |
|
436 """Determine whether a message contains the named header.""" |
|
437 return name.lower() in self.dict |
|
438 |
|
439 def __contains__(self, name): |
|
440 """Determine whether a message contains the named header.""" |
|
441 return name.lower() in self.dict |
|
442 |
|
443 def __iter__(self): |
|
444 return iter(self.dict) |
|
445 |
|
446 def keys(self): |
|
447 """Get all of a message's header field names.""" |
|
448 return self.dict.keys() |
|
449 |
|
450 def values(self): |
|
451 """Get all of a message's header field values.""" |
|
452 return self.dict.values() |
|
453 |
|
454 def items(self): |
|
455 """Get all of a message's headers. |
|
456 |
|
457 Returns a list of name, value tuples. |
|
458 """ |
|
459 return self.dict.items() |
|
460 |
|
461 def __str__(self): |
|
462 return ''.join(self.headers) |
|
463 |
|
464 |
|
465 # Utility functions |
|
466 # ----------------- |
|
467 |
|
468 # XXX Should fix unquote() and quote() to be really conformant. |
|
469 # XXX The inverses of the parse functions may also be useful. |
|
470 |
|
471 |
|
472 def unquote(s): |
|
473 """Remove quotes from a string.""" |
|
474 if len(s) > 1: |
|
475 if s.startswith('"') and s.endswith('"'): |
|
476 return s[1:-1].replace('\\\\', '\\').replace('\\"', '"') |
|
477 if s.startswith('<') and s.endswith('>'): |
|
478 return s[1:-1] |
|
479 return s |
|
480 |
|
481 |
|
482 def quote(s): |
|
483 """Add quotes around a string.""" |
|
484 return s.replace('\\', '\\\\').replace('"', '\\"') |
|
485 |
|
486 |
|
487 def parseaddr(address): |
|
488 """Parse an address into a (realname, mailaddr) tuple.""" |
|
489 a = AddressList(address) |
|
490 lst = a.addresslist |
|
491 if not lst: |
|
492 return (None, None) |
|
493 return lst[0] |
|
494 |
|
495 |
|
496 class AddrlistClass: |
|
497 """Address parser class by Ben Escoto. |
|
498 |
|
499 To understand what this class does, it helps to have a copy of |
|
500 RFC 2822 in front of you. |
|
501 |
|
502 http://www.faqs.org/rfcs/rfc2822.html |
|
503 |
|
504 Note: this class interface is deprecated and may be removed in the future. |
|
505 Use rfc822.AddressList instead. |
|
506 """ |
|
507 |
|
508 def __init__(self, field): |
|
509 """Initialize a new instance. |
|
510 |
|
511 `field' is an unparsed address header field, containing one or more |
|
512 addresses. |
|
513 """ |
|
514 self.specials = '()<>@,:;.\"[]' |
|
515 self.pos = 0 |
|
516 self.LWS = ' \t' |
|
517 self.CR = '\r\n' |
|
518 self.atomends = self.specials + self.LWS + self.CR |
|
519 # Note that RFC 2822 now specifies `.' as obs-phrase, meaning that it |
|
520 # is obsolete syntax. RFC 2822 requires that we recognize obsolete |
|
521 # syntax, so allow dots in phrases. |
|
522 self.phraseends = self.atomends.replace('.', '') |
|
523 self.field = field |
|
524 self.commentlist = [] |
|
525 |
|
526 def gotonext(self): |
|
527 """Parse up to the start of the next address.""" |
|
528 while self.pos < len(self.field): |
|
529 if self.field[self.pos] in self.LWS + '\n\r': |
|
530 self.pos = self.pos + 1 |
|
531 elif self.field[self.pos] == '(': |
|
532 self.commentlist.append(self.getcomment()) |
|
533 else: break |
|
534 |
|
535 def getaddrlist(self): |
|
536 """Parse all addresses. |
|
537 |
|
538 Returns a list containing all of the addresses. |
|
539 """ |
|
540 result = [] |
|
541 ad = self.getaddress() |
|
542 while ad: |
|
543 result += ad |
|
544 ad = self.getaddress() |
|
545 return result |
|
546 |
|
547 def getaddress(self): |
|
548 """Parse the next address.""" |
|
549 self.commentlist = [] |
|
550 self.gotonext() |
|
551 |
|
552 oldpos = self.pos |
|
553 oldcl = self.commentlist |
|
554 plist = self.getphraselist() |
|
555 |
|
556 self.gotonext() |
|
557 returnlist = [] |
|
558 |
|
559 if self.pos >= len(self.field): |
|
560 # Bad email address technically, no domain. |
|
561 if plist: |
|
562 returnlist = [(' '.join(self.commentlist), plist[0])] |
|
563 |
|
564 elif self.field[self.pos] in '.@': |
|
565 # email address is just an addrspec |
|
566 # this isn't very efficient since we start over |
|
567 self.pos = oldpos |
|
568 self.commentlist = oldcl |
|
569 addrspec = self.getaddrspec() |
|
570 returnlist = [(' '.join(self.commentlist), addrspec)] |
|
571 |
|
572 elif self.field[self.pos] == ':': |
|
573 # address is a group |
|
574 returnlist = [] |
|
575 |
|
576 fieldlen = len(self.field) |
|
577 self.pos += 1 |
|
578 while self.pos < len(self.field): |
|
579 self.gotonext() |
|
580 if self.pos < fieldlen and self.field[self.pos] == ';': |
|
581 self.pos += 1 |
|
582 break |
|
583 returnlist = returnlist + self.getaddress() |
|
584 |
|
585 elif self.field[self.pos] == '<': |
|
586 # Address is a phrase then a route addr |
|
587 routeaddr = self.getrouteaddr() |
|
588 |
|
589 if self.commentlist: |
|
590 returnlist = [(' '.join(plist) + ' (' + \ |
|
591 ' '.join(self.commentlist) + ')', routeaddr)] |
|
592 else: returnlist = [(' '.join(plist), routeaddr)] |
|
593 |
|
594 else: |
|
595 if plist: |
|
596 returnlist = [(' '.join(self.commentlist), plist[0])] |
|
597 elif self.field[self.pos] in self.specials: |
|
598 self.pos += 1 |
|
599 |
|
600 self.gotonext() |
|
601 if self.pos < len(self.field) and self.field[self.pos] == ',': |
|
602 self.pos += 1 |
|
603 return returnlist |
|
604 |
|
605 def getrouteaddr(self): |
|
606 """Parse a route address (Return-path value). |
|
607 |
|
608 This method just skips all the route stuff and returns the addrspec. |
|
609 """ |
|
610 if self.field[self.pos] != '<': |
|
611 return |
|
612 |
|
613 expectroute = 0 |
|
614 self.pos += 1 |
|
615 self.gotonext() |
|
616 adlist = "" |
|
617 while self.pos < len(self.field): |
|
618 if expectroute: |
|
619 self.getdomain() |
|
620 expectroute = 0 |
|
621 elif self.field[self.pos] == '>': |
|
622 self.pos += 1 |
|
623 break |
|
624 elif self.field[self.pos] == '@': |
|
625 self.pos += 1 |
|
626 expectroute = 1 |
|
627 elif self.field[self.pos] == ':': |
|
628 self.pos += 1 |
|
629 else: |
|
630 adlist = self.getaddrspec() |
|
631 self.pos += 1 |
|
632 break |
|
633 self.gotonext() |
|
634 |
|
635 return adlist |
|
636 |
|
637 def getaddrspec(self): |
|
638 """Parse an RFC 2822 addr-spec.""" |
|
639 aslist = [] |
|
640 |
|
641 self.gotonext() |
|
642 while self.pos < len(self.field): |
|
643 if self.field[self.pos] == '.': |
|
644 aslist.append('.') |
|
645 self.pos += 1 |
|
646 elif self.field[self.pos] == '"': |
|
647 aslist.append('"%s"' % self.getquote()) |
|
648 elif self.field[self.pos] in self.atomends: |
|
649 break |
|
650 else: aslist.append(self.getatom()) |
|
651 self.gotonext() |
|
652 |
|
653 if self.pos >= len(self.field) or self.field[self.pos] != '@': |
|
654 return ''.join(aslist) |
|
655 |
|
656 aslist.append('@') |
|
657 self.pos += 1 |
|
658 self.gotonext() |
|
659 return ''.join(aslist) + self.getdomain() |
|
660 |
|
661 def getdomain(self): |
|
662 """Get the complete domain name from an address.""" |
|
663 sdlist = [] |
|
664 while self.pos < len(self.field): |
|
665 if self.field[self.pos] in self.LWS: |
|
666 self.pos += 1 |
|
667 elif self.field[self.pos] == '(': |
|
668 self.commentlist.append(self.getcomment()) |
|
669 elif self.field[self.pos] == '[': |
|
670 sdlist.append(self.getdomainliteral()) |
|
671 elif self.field[self.pos] == '.': |
|
672 self.pos += 1 |
|
673 sdlist.append('.') |
|
674 elif self.field[self.pos] in self.atomends: |
|
675 break |
|
676 else: sdlist.append(self.getatom()) |
|
677 return ''.join(sdlist) |
|
678 |
|
679 def getdelimited(self, beginchar, endchars, allowcomments = 1): |
|
680 """Parse a header fragment delimited by special characters. |
|
681 |
|
682 `beginchar' is the start character for the fragment. If self is not |
|
683 looking at an instance of `beginchar' then getdelimited returns the |
|
684 empty string. |
|
685 |
|
686 `endchars' is a sequence of allowable end-delimiting characters. |
|
687 Parsing stops when one of these is encountered. |
|
688 |
|
689 If `allowcomments' is non-zero, embedded RFC 2822 comments are allowed |
|
690 within the parsed fragment. |
|
691 """ |
|
692 if self.field[self.pos] != beginchar: |
|
693 return '' |
|
694 |
|
695 slist = [''] |
|
696 quote = 0 |
|
697 self.pos += 1 |
|
698 while self.pos < len(self.field): |
|
699 if quote == 1: |
|
700 slist.append(self.field[self.pos]) |
|
701 quote = 0 |
|
702 elif self.field[self.pos] in endchars: |
|
703 self.pos += 1 |
|
704 break |
|
705 elif allowcomments and self.field[self.pos] == '(': |
|
706 slist.append(self.getcomment()) |
|
707 continue # have already advanced pos from getcomment |
|
708 elif self.field[self.pos] == '\\': |
|
709 quote = 1 |
|
710 else: |
|
711 slist.append(self.field[self.pos]) |
|
712 self.pos += 1 |
|
713 |
|
714 return ''.join(slist) |
|
715 |
|
716 def getquote(self): |
|
717 """Get a quote-delimited fragment from self's field.""" |
|
718 return self.getdelimited('"', '"\r', 0) |
|
719 |
|
720 def getcomment(self): |
|
721 """Get a parenthesis-delimited fragment from self's field.""" |
|
722 return self.getdelimited('(', ')\r', 1) |
|
723 |
|
724 def getdomainliteral(self): |
|
725 """Parse an RFC 2822 domain-literal.""" |
|
726 return '[%s]' % self.getdelimited('[', ']\r', 0) |
|
727 |
|
728 def getatom(self, atomends=None): |
|
729 """Parse an RFC 2822 atom. |
|
730 |
|
731 Optional atomends specifies a different set of end token delimiters |
|
732 (the default is to use self.atomends). This is used e.g. in |
|
733 getphraselist() since phrase endings must not include the `.' (which |
|
734 is legal in phrases).""" |
|
735 atomlist = [''] |
|
736 if atomends is None: |
|
737 atomends = self.atomends |
|
738 |
|
739 while self.pos < len(self.field): |
|
740 if self.field[self.pos] in atomends: |
|
741 break |
|
742 else: atomlist.append(self.field[self.pos]) |
|
743 self.pos += 1 |
|
744 |
|
745 return ''.join(atomlist) |
|
746 |
|
747 def getphraselist(self): |
|
748 """Parse a sequence of RFC 2822 phrases. |
|
749 |
|
750 A phrase is a sequence of words, which are in turn either RFC 2822 |
|
751 atoms or quoted-strings. Phrases are canonicalized by squeezing all |
|
752 runs of continuous whitespace into one space. |
|
753 """ |
|
754 plist = [] |
|
755 |
|
756 while self.pos < len(self.field): |
|
757 if self.field[self.pos] in self.LWS: |
|
758 self.pos += 1 |
|
759 elif self.field[self.pos] == '"': |
|
760 plist.append(self.getquote()) |
|
761 elif self.field[self.pos] == '(': |
|
762 self.commentlist.append(self.getcomment()) |
|
763 elif self.field[self.pos] in self.phraseends: |
|
764 break |
|
765 else: |
|
766 plist.append(self.getatom(self.phraseends)) |
|
767 |
|
768 return plist |
|
769 |
|
770 class AddressList(AddrlistClass): |
|
771 """An AddressList encapsulates a list of parsed RFC 2822 addresses.""" |
|
772 def __init__(self, field): |
|
773 AddrlistClass.__init__(self, field) |
|
774 if field: |
|
775 self.addresslist = self.getaddrlist() |
|
776 else: |
|
777 self.addresslist = [] |
|
778 |
|
779 def __len__(self): |
|
780 return len(self.addresslist) |
|
781 |
|
782 def __str__(self): |
|
783 return ", ".join(map(dump_address_pair, self.addresslist)) |
|
784 |
|
785 def __add__(self, other): |
|
786 # Set union |
|
787 newaddr = AddressList(None) |
|
788 newaddr.addresslist = self.addresslist[:] |
|
789 for x in other.addresslist: |
|
790 if not x in self.addresslist: |
|
791 newaddr.addresslist.append(x) |
|
792 return newaddr |
|
793 |
|
794 def __iadd__(self, other): |
|
795 # Set union, in-place |
|
796 for x in other.addresslist: |
|
797 if not x in self.addresslist: |
|
798 self.addresslist.append(x) |
|
799 return self |
|
800 |
|
801 def __sub__(self, other): |
|
802 # Set difference |
|
803 newaddr = AddressList(None) |
|
804 for x in self.addresslist: |
|
805 if not x in other.addresslist: |
|
806 newaddr.addresslist.append(x) |
|
807 return newaddr |
|
808 |
|
809 def __isub__(self, other): |
|
810 # Set difference, in-place |
|
811 for x in other.addresslist: |
|
812 if x in self.addresslist: |
|
813 self.addresslist.remove(x) |
|
814 return self |
|
815 |
|
816 def __getitem__(self, index): |
|
817 # Make indexing, slices, and 'in' work |
|
818 return self.addresslist[index] |
|
819 |
|
820 def dump_address_pair(pair): |
|
821 """Dump a (name, address) pair in a canonicalized form.""" |
|
822 if pair[0]: |
|
823 return '"' + pair[0] + '" <' + pair[1] + '>' |
|
824 else: |
|
825 return pair[1] |
|
826 |
|
827 # Parse a date field |
|
828 |
|
829 _monthnames = ['jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', |
|
830 'aug', 'sep', 'oct', 'nov', 'dec', |
|
831 'january', 'february', 'march', 'april', 'may', 'june', 'july', |
|
832 'august', 'september', 'october', 'november', 'december'] |
|
833 _daynames = ['mon', 'tue', 'wed', 'thu', 'fri', 'sat', 'sun'] |
|
834 |
|
835 # The timezone table does not include the military time zones defined |
|
836 # in RFC822, other than Z. According to RFC1123, the description in |
|
837 # RFC822 gets the signs wrong, so we can't rely on any such time |
|
838 # zones. RFC1123 recommends that numeric timezone indicators be used |
|
839 # instead of timezone names. |
|
840 |
|
841 _timezones = {'UT':0, 'UTC':0, 'GMT':0, 'Z':0, |
|
842 'AST': -400, 'ADT': -300, # Atlantic (used in Canada) |
|
843 'EST': -500, 'EDT': -400, # Eastern |
|
844 'CST': -600, 'CDT': -500, # Central |
|
845 'MST': -700, 'MDT': -600, # Mountain |
|
846 'PST': -800, 'PDT': -700 # Pacific |
|
847 } |
|
848 |
|
849 |
|
850 def parsedate_tz(data): |
|
851 """Convert a date string to a time tuple. |
|
852 |
|
853 Accounts for military timezones. |
|
854 """ |
|
855 if not data: |
|
856 return None |
|
857 data = data.split() |
|
858 if data[0][-1] in (',', '.') or data[0].lower() in _daynames: |
|
859 # There's a dayname here. Skip it |
|
860 del data[0] |
|
861 else: |
|
862 # no space after the "weekday,"? |
|
863 i = data[0].rfind(',') |
|
864 if i >= 0: |
|
865 data[0] = data[0][i+1:] |
|
866 if len(data) == 3: # RFC 850 date, deprecated |
|
867 stuff = data[0].split('-') |
|
868 if len(stuff) == 3: |
|
869 data = stuff + data[1:] |
|
870 if len(data) == 4: |
|
871 s = data[3] |
|
872 i = s.find('+') |
|
873 if i > 0: |
|
874 data[3:] = [s[:i], s[i+1:]] |
|
875 else: |
|
876 data.append('') # Dummy tz |
|
877 if len(data) < 5: |
|
878 return None |
|
879 data = data[:5] |
|
880 [dd, mm, yy, tm, tz] = data |
|
881 mm = mm.lower() |
|
882 if not mm in _monthnames: |
|
883 dd, mm = mm, dd.lower() |
|
884 if not mm in _monthnames: |
|
885 return None |
|
886 mm = _monthnames.index(mm)+1 |
|
887 if mm > 12: mm = mm - 12 |
|
888 if dd[-1] == ',': |
|
889 dd = dd[:-1] |
|
890 i = yy.find(':') |
|
891 if i > 0: |
|
892 yy, tm = tm, yy |
|
893 if yy[-1] == ',': |
|
894 yy = yy[:-1] |
|
895 if not yy[0].isdigit(): |
|
896 yy, tz = tz, yy |
|
897 if tm[-1] == ',': |
|
898 tm = tm[:-1] |
|
899 tm = tm.split(':') |
|
900 if len(tm) == 2: |
|
901 [thh, tmm] = tm |
|
902 tss = '0' |
|
903 elif len(tm) == 3: |
|
904 [thh, tmm, tss] = tm |
|
905 else: |
|
906 return None |
|
907 try: |
|
908 yy = int(yy) |
|
909 dd = int(dd) |
|
910 thh = int(thh) |
|
911 tmm = int(tmm) |
|
912 tss = int(tss) |
|
913 except ValueError: |
|
914 return None |
|
915 tzoffset = None |
|
916 tz = tz.upper() |
|
917 if tz in _timezones: |
|
918 tzoffset = _timezones[tz] |
|
919 else: |
|
920 try: |
|
921 tzoffset = int(tz) |
|
922 except ValueError: |
|
923 pass |
|
924 # Convert a timezone offset into seconds ; -0500 -> -18000 |
|
925 if tzoffset: |
|
926 if tzoffset < 0: |
|
927 tzsign = -1 |
|
928 tzoffset = -tzoffset |
|
929 else: |
|
930 tzsign = 1 |
|
931 tzoffset = tzsign * ( (tzoffset//100)*3600 + (tzoffset % 100)*60) |
|
932 return (yy, mm, dd, thh, tmm, tss, 0, 1, 0, tzoffset) |
|
933 |
|
934 |
|
935 def parsedate(data): |
|
936 """Convert a time string to a time tuple.""" |
|
937 t = parsedate_tz(data) |
|
938 if t is None: |
|
939 return t |
|
940 return t[:9] |
|
941 |
|
942 |
|
943 def mktime_tz(data): |
|
944 """Turn a 10-tuple as returned by parsedate_tz() into a UTC timestamp.""" |
|
945 if data[9] is None: |
|
946 # No zone info, so localtime is better assumption than GMT |
|
947 return time.mktime(data[:8] + (-1,)) |
|
948 else: |
|
949 t = time.mktime(data[:8] + (0,)) |
|
950 return t - data[9] - time.timezone |
|
951 |
|
952 def formatdate(timeval=None): |
|
953 """Returns time format preferred for Internet standards. |
|
954 |
|
955 Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123 |
|
956 |
|
957 According to RFC 1123, day and month names must always be in |
|
958 English. If not for that, this code could use strftime(). It |
|
959 can't because strftime() honors the locale and could generated |
|
960 non-English names. |
|
961 """ |
|
962 if timeval is None: |
|
963 timeval = time.time() |
|
964 timeval = time.gmtime(timeval) |
|
965 return "%s, %02d %s %04d %02d:%02d:%02d GMT" % ( |
|
966 ("Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun")[timeval[6]], |
|
967 timeval[2], |
|
968 ("Jan", "Feb", "Mar", "Apr", "May", "Jun", |
|
969 "Jul", "Aug", "Sep", "Oct", "Nov", "Dec")[timeval[1]-1], |
|
970 timeval[0], timeval[3], timeval[4], timeval[5]) |
|
971 |
|
972 |
|
973 # When used as script, run a small test program. |
|
974 # The first command line argument must be a filename containing one |
|
975 # message in RFC-822 format. |
|
976 |
|
977 if __name__ == '__main__': |
|
978 import sys, os |
|
979 file = os.path.join(os.environ['HOME'], 'Mail/inbox/1') |
|
980 if sys.argv[1:]: file = sys.argv[1] |
|
981 f = open(file, 'r') |
|
982 m = Message(f) |
|
983 print 'From:', m.getaddr('from') |
|
984 print 'To:', m.getaddrlist('to') |
|
985 print 'Subject:', m.getheader('subject') |
|
986 print 'Date:', m.getheader('date') |
|
987 date = m.getdate_tz('date') |
|
988 tz = date[-1] |
|
989 date = time.localtime(mktime_tz(date)) |
|
990 if date: |
|
991 print 'ParsedDate:', time.asctime(date), |
|
992 hhmmss = tz |
|
993 hhmm, ss = divmod(hhmmss, 60) |
|
994 hh, mm = divmod(hhmm, 60) |
|
995 print "%+03d%02d" % (hh, mm), |
|
996 if ss: print ".%02d" % ss, |
|
997 print |
|
998 else: |
|
999 print 'ParsedDate:', None |
|
1000 m.rewindbody() |
|
1001 n = 0 |
|
1002 while f.readline(): |
|
1003 n += 1 |
|
1004 print 'Lines:', n |
|
1005 print '-'*70 |
|
1006 print 'len =', len(m) |
|
1007 if 'Date' in m: print 'Date =', m['Date'] |
|
1008 if 'X-Nonsense' in m: pass |
|
1009 print 'keys =', m.keys() |
|
1010 print 'values =', m.values() |
|
1011 print 'items =', m.items() |