jpayne@69
|
1 # Copyright (C) 2001-2010 Python Software Foundation
|
jpayne@69
|
2 # Author: Barry Warsaw
|
jpayne@69
|
3 # Contact: email-sig@python.org
|
jpayne@69
|
4
|
jpayne@69
|
5 """Classes to generate plain text from a message object tree."""
|
jpayne@69
|
6
|
jpayne@69
|
7 __all__ = ['Generator', 'DecodedGenerator', 'BytesGenerator']
|
jpayne@69
|
8
|
jpayne@69
|
9 import re
|
jpayne@69
|
10 import sys
|
jpayne@69
|
11 import time
|
jpayne@69
|
12 import random
|
jpayne@69
|
13
|
jpayne@69
|
14 from copy import deepcopy
|
jpayne@69
|
15 from io import StringIO, BytesIO
|
jpayne@69
|
16 from email.utils import _has_surrogates
|
jpayne@69
|
17
|
jpayne@69
|
18 UNDERSCORE = '_'
|
jpayne@69
|
19 NL = '\n' # XXX: no longer used by the code below.
|
jpayne@69
|
20
|
jpayne@69
|
21 NLCRE = re.compile(r'\r\n|\r|\n')
|
jpayne@69
|
22 fcre = re.compile(r'^From ', re.MULTILINE)
|
jpayne@69
|
23
|
jpayne@69
|
24
|
jpayne@69
|
25
|
jpayne@69
|
26 class Generator:
|
jpayne@69
|
27 """Generates output from a Message object tree.
|
jpayne@69
|
28
|
jpayne@69
|
29 This basic generator writes the message to the given file object as plain
|
jpayne@69
|
30 text.
|
jpayne@69
|
31 """
|
jpayne@69
|
32 #
|
jpayne@69
|
33 # Public interface
|
jpayne@69
|
34 #
|
jpayne@69
|
35
|
jpayne@69
|
36 def __init__(self, outfp, mangle_from_=None, maxheaderlen=None, *,
|
jpayne@69
|
37 policy=None):
|
jpayne@69
|
38 """Create the generator for message flattening.
|
jpayne@69
|
39
|
jpayne@69
|
40 outfp is the output file-like object for writing the message to. It
|
jpayne@69
|
41 must have a write() method.
|
jpayne@69
|
42
|
jpayne@69
|
43 Optional mangle_from_ is a flag that, when True (the default if policy
|
jpayne@69
|
44 is not set), escapes From_ lines in the body of the message by putting
|
jpayne@69
|
45 a `>' in front of them.
|
jpayne@69
|
46
|
jpayne@69
|
47 Optional maxheaderlen specifies the longest length for a non-continued
|
jpayne@69
|
48 header. When a header line is longer (in characters, with tabs
|
jpayne@69
|
49 expanded to 8 spaces) than maxheaderlen, the header will split as
|
jpayne@69
|
50 defined in the Header class. Set maxheaderlen to zero to disable
|
jpayne@69
|
51 header wrapping. The default is 78, as recommended (but not required)
|
jpayne@69
|
52 by RFC 2822.
|
jpayne@69
|
53
|
jpayne@69
|
54 The policy keyword specifies a policy object that controls a number of
|
jpayne@69
|
55 aspects of the generator's operation. If no policy is specified,
|
jpayne@69
|
56 the policy associated with the Message object passed to the
|
jpayne@69
|
57 flatten method is used.
|
jpayne@69
|
58
|
jpayne@69
|
59 """
|
jpayne@69
|
60
|
jpayne@69
|
61 if mangle_from_ is None:
|
jpayne@69
|
62 mangle_from_ = True if policy is None else policy.mangle_from_
|
jpayne@69
|
63 self._fp = outfp
|
jpayne@69
|
64 self._mangle_from_ = mangle_from_
|
jpayne@69
|
65 self.maxheaderlen = maxheaderlen
|
jpayne@69
|
66 self.policy = policy
|
jpayne@69
|
67
|
jpayne@69
|
68 def write(self, s):
|
jpayne@69
|
69 # Just delegate to the file object
|
jpayne@69
|
70 self._fp.write(s)
|
jpayne@69
|
71
|
jpayne@69
|
72 def flatten(self, msg, unixfrom=False, linesep=None):
|
jpayne@69
|
73 r"""Print the message object tree rooted at msg to the output file
|
jpayne@69
|
74 specified when the Generator instance was created.
|
jpayne@69
|
75
|
jpayne@69
|
76 unixfrom is a flag that forces the printing of a Unix From_ delimiter
|
jpayne@69
|
77 before the first object in the message tree. If the original message
|
jpayne@69
|
78 has no From_ delimiter, a `standard' one is crafted. By default, this
|
jpayne@69
|
79 is False to inhibit the printing of any From_ delimiter.
|
jpayne@69
|
80
|
jpayne@69
|
81 Note that for subobjects, no From_ line is printed.
|
jpayne@69
|
82
|
jpayne@69
|
83 linesep specifies the characters used to indicate a new line in
|
jpayne@69
|
84 the output. The default value is determined by the policy specified
|
jpayne@69
|
85 when the Generator instance was created or, if none was specified,
|
jpayne@69
|
86 from the policy associated with the msg.
|
jpayne@69
|
87
|
jpayne@69
|
88 """
|
jpayne@69
|
89 # We use the _XXX constants for operating on data that comes directly
|
jpayne@69
|
90 # from the msg, and _encoded_XXX constants for operating on data that
|
jpayne@69
|
91 # has already been converted (to bytes in the BytesGenerator) and
|
jpayne@69
|
92 # inserted into a temporary buffer.
|
jpayne@69
|
93 policy = msg.policy if self.policy is None else self.policy
|
jpayne@69
|
94 if linesep is not None:
|
jpayne@69
|
95 policy = policy.clone(linesep=linesep)
|
jpayne@69
|
96 if self.maxheaderlen is not None:
|
jpayne@69
|
97 policy = policy.clone(max_line_length=self.maxheaderlen)
|
jpayne@69
|
98 self._NL = policy.linesep
|
jpayne@69
|
99 self._encoded_NL = self._encode(self._NL)
|
jpayne@69
|
100 self._EMPTY = ''
|
jpayne@69
|
101 self._encoded_EMPTY = self._encode(self._EMPTY)
|
jpayne@69
|
102 # Because we use clone (below) when we recursively process message
|
jpayne@69
|
103 # subparts, and because clone uses the computed policy (not None),
|
jpayne@69
|
104 # submessages will automatically get set to the computed policy when
|
jpayne@69
|
105 # they are processed by this code.
|
jpayne@69
|
106 old_gen_policy = self.policy
|
jpayne@69
|
107 old_msg_policy = msg.policy
|
jpayne@69
|
108 try:
|
jpayne@69
|
109 self.policy = policy
|
jpayne@69
|
110 msg.policy = policy
|
jpayne@69
|
111 if unixfrom:
|
jpayne@69
|
112 ufrom = msg.get_unixfrom()
|
jpayne@69
|
113 if not ufrom:
|
jpayne@69
|
114 ufrom = 'From nobody ' + time.ctime(time.time())
|
jpayne@69
|
115 self.write(ufrom + self._NL)
|
jpayne@69
|
116 self._write(msg)
|
jpayne@69
|
117 finally:
|
jpayne@69
|
118 self.policy = old_gen_policy
|
jpayne@69
|
119 msg.policy = old_msg_policy
|
jpayne@69
|
120
|
jpayne@69
|
121 def clone(self, fp):
|
jpayne@69
|
122 """Clone this generator with the exact same options."""
|
jpayne@69
|
123 return self.__class__(fp,
|
jpayne@69
|
124 self._mangle_from_,
|
jpayne@69
|
125 None, # Use policy setting, which we've adjusted
|
jpayne@69
|
126 policy=self.policy)
|
jpayne@69
|
127
|
jpayne@69
|
128 #
|
jpayne@69
|
129 # Protected interface - undocumented ;/
|
jpayne@69
|
130 #
|
jpayne@69
|
131
|
jpayne@69
|
132 # Note that we use 'self.write' when what we are writing is coming from
|
jpayne@69
|
133 # the source, and self._fp.write when what we are writing is coming from a
|
jpayne@69
|
134 # buffer (because the Bytes subclass has already had a chance to transform
|
jpayne@69
|
135 # the data in its write method in that case). This is an entirely
|
jpayne@69
|
136 # pragmatic split determined by experiment; we could be more general by
|
jpayne@69
|
137 # always using write and having the Bytes subclass write method detect when
|
jpayne@69
|
138 # it has already transformed the input; but, since this whole thing is a
|
jpayne@69
|
139 # hack anyway this seems good enough.
|
jpayne@69
|
140
|
jpayne@69
|
141 def _new_buffer(self):
|
jpayne@69
|
142 # BytesGenerator overrides this to return BytesIO.
|
jpayne@69
|
143 return StringIO()
|
jpayne@69
|
144
|
jpayne@69
|
145 def _encode(self, s):
|
jpayne@69
|
146 # BytesGenerator overrides this to encode strings to bytes.
|
jpayne@69
|
147 return s
|
jpayne@69
|
148
|
jpayne@69
|
149 def _write_lines(self, lines):
|
jpayne@69
|
150 # We have to transform the line endings.
|
jpayne@69
|
151 if not lines:
|
jpayne@69
|
152 return
|
jpayne@69
|
153 lines = NLCRE.split(lines)
|
jpayne@69
|
154 for line in lines[:-1]:
|
jpayne@69
|
155 self.write(line)
|
jpayne@69
|
156 self.write(self._NL)
|
jpayne@69
|
157 if lines[-1]:
|
jpayne@69
|
158 self.write(lines[-1])
|
jpayne@69
|
159 # XXX logic tells me this else should be needed, but the tests fail
|
jpayne@69
|
160 # with it and pass without it. (NLCRE.split ends with a blank element
|
jpayne@69
|
161 # if and only if there was a trailing newline.)
|
jpayne@69
|
162 #else:
|
jpayne@69
|
163 # self.write(self._NL)
|
jpayne@69
|
164
|
jpayne@69
|
165 def _write(self, msg):
|
jpayne@69
|
166 # We can't write the headers yet because of the following scenario:
|
jpayne@69
|
167 # say a multipart message includes the boundary string somewhere in
|
jpayne@69
|
168 # its body. We'd have to calculate the new boundary /before/ we write
|
jpayne@69
|
169 # the headers so that we can write the correct Content-Type:
|
jpayne@69
|
170 # parameter.
|
jpayne@69
|
171 #
|
jpayne@69
|
172 # The way we do this, so as to make the _handle_*() methods simpler,
|
jpayne@69
|
173 # is to cache any subpart writes into a buffer. The we write the
|
jpayne@69
|
174 # headers and the buffer contents. That way, subpart handlers can
|
jpayne@69
|
175 # Do The Right Thing, and can still modify the Content-Type: header if
|
jpayne@69
|
176 # necessary.
|
jpayne@69
|
177 oldfp = self._fp
|
jpayne@69
|
178 try:
|
jpayne@69
|
179 self._munge_cte = None
|
jpayne@69
|
180 self._fp = sfp = self._new_buffer()
|
jpayne@69
|
181 self._dispatch(msg)
|
jpayne@69
|
182 finally:
|
jpayne@69
|
183 self._fp = oldfp
|
jpayne@69
|
184 munge_cte = self._munge_cte
|
jpayne@69
|
185 del self._munge_cte
|
jpayne@69
|
186 # If we munged the cte, copy the message again and re-fix the CTE.
|
jpayne@69
|
187 if munge_cte:
|
jpayne@69
|
188 msg = deepcopy(msg)
|
jpayne@69
|
189 msg.replace_header('content-transfer-encoding', munge_cte[0])
|
jpayne@69
|
190 msg.replace_header('content-type', munge_cte[1])
|
jpayne@69
|
191 # Write the headers. First we see if the message object wants to
|
jpayne@69
|
192 # handle that itself. If not, we'll do it generically.
|
jpayne@69
|
193 meth = getattr(msg, '_write_headers', None)
|
jpayne@69
|
194 if meth is None:
|
jpayne@69
|
195 self._write_headers(msg)
|
jpayne@69
|
196 else:
|
jpayne@69
|
197 meth(self)
|
jpayne@69
|
198 self._fp.write(sfp.getvalue())
|
jpayne@69
|
199
|
jpayne@69
|
200 def _dispatch(self, msg):
|
jpayne@69
|
201 # Get the Content-Type: for the message, then try to dispatch to
|
jpayne@69
|
202 # self._handle_<maintype>_<subtype>(). If there's no handler for the
|
jpayne@69
|
203 # full MIME type, then dispatch to self._handle_<maintype>(). If
|
jpayne@69
|
204 # that's missing too, then dispatch to self._writeBody().
|
jpayne@69
|
205 main = msg.get_content_maintype()
|
jpayne@69
|
206 sub = msg.get_content_subtype()
|
jpayne@69
|
207 specific = UNDERSCORE.join((main, sub)).replace('-', '_')
|
jpayne@69
|
208 meth = getattr(self, '_handle_' + specific, None)
|
jpayne@69
|
209 if meth is None:
|
jpayne@69
|
210 generic = main.replace('-', '_')
|
jpayne@69
|
211 meth = getattr(self, '_handle_' + generic, None)
|
jpayne@69
|
212 if meth is None:
|
jpayne@69
|
213 meth = self._writeBody
|
jpayne@69
|
214 meth(msg)
|
jpayne@69
|
215
|
jpayne@69
|
216 #
|
jpayne@69
|
217 # Default handlers
|
jpayne@69
|
218 #
|
jpayne@69
|
219
|
jpayne@69
|
220 def _write_headers(self, msg):
|
jpayne@69
|
221 for h, v in msg.raw_items():
|
jpayne@69
|
222 self.write(self.policy.fold(h, v))
|
jpayne@69
|
223 # A blank line always separates headers from body
|
jpayne@69
|
224 self.write(self._NL)
|
jpayne@69
|
225
|
jpayne@69
|
226 #
|
jpayne@69
|
227 # Handlers for writing types and subtypes
|
jpayne@69
|
228 #
|
jpayne@69
|
229
|
jpayne@69
|
230 def _handle_text(self, msg):
|
jpayne@69
|
231 payload = msg.get_payload()
|
jpayne@69
|
232 if payload is None:
|
jpayne@69
|
233 return
|
jpayne@69
|
234 if not isinstance(payload, str):
|
jpayne@69
|
235 raise TypeError('string payload expected: %s' % type(payload))
|
jpayne@69
|
236 if _has_surrogates(msg._payload):
|
jpayne@69
|
237 charset = msg.get_param('charset')
|
jpayne@69
|
238 if charset is not None:
|
jpayne@69
|
239 # XXX: This copy stuff is an ugly hack to avoid modifying the
|
jpayne@69
|
240 # existing message.
|
jpayne@69
|
241 msg = deepcopy(msg)
|
jpayne@69
|
242 del msg['content-transfer-encoding']
|
jpayne@69
|
243 msg.set_payload(payload, charset)
|
jpayne@69
|
244 payload = msg.get_payload()
|
jpayne@69
|
245 self._munge_cte = (msg['content-transfer-encoding'],
|
jpayne@69
|
246 msg['content-type'])
|
jpayne@69
|
247 if self._mangle_from_:
|
jpayne@69
|
248 payload = fcre.sub('>From ', payload)
|
jpayne@69
|
249 self._write_lines(payload)
|
jpayne@69
|
250
|
jpayne@69
|
251 # Default body handler
|
jpayne@69
|
252 _writeBody = _handle_text
|
jpayne@69
|
253
|
jpayne@69
|
254 def _handle_multipart(self, msg):
|
jpayne@69
|
255 # The trick here is to write out each part separately, merge them all
|
jpayne@69
|
256 # together, and then make sure that the boundary we've chosen isn't
|
jpayne@69
|
257 # present in the payload.
|
jpayne@69
|
258 msgtexts = []
|
jpayne@69
|
259 subparts = msg.get_payload()
|
jpayne@69
|
260 if subparts is None:
|
jpayne@69
|
261 subparts = []
|
jpayne@69
|
262 elif isinstance(subparts, str):
|
jpayne@69
|
263 # e.g. a non-strict parse of a message with no starting boundary.
|
jpayne@69
|
264 self.write(subparts)
|
jpayne@69
|
265 return
|
jpayne@69
|
266 elif not isinstance(subparts, list):
|
jpayne@69
|
267 # Scalar payload
|
jpayne@69
|
268 subparts = [subparts]
|
jpayne@69
|
269 for part in subparts:
|
jpayne@69
|
270 s = self._new_buffer()
|
jpayne@69
|
271 g = self.clone(s)
|
jpayne@69
|
272 g.flatten(part, unixfrom=False, linesep=self._NL)
|
jpayne@69
|
273 msgtexts.append(s.getvalue())
|
jpayne@69
|
274 # BAW: What about boundaries that are wrapped in double-quotes?
|
jpayne@69
|
275 boundary = msg.get_boundary()
|
jpayne@69
|
276 if not boundary:
|
jpayne@69
|
277 # Create a boundary that doesn't appear in any of the
|
jpayne@69
|
278 # message texts.
|
jpayne@69
|
279 alltext = self._encoded_NL.join(msgtexts)
|
jpayne@69
|
280 boundary = self._make_boundary(alltext)
|
jpayne@69
|
281 msg.set_boundary(boundary)
|
jpayne@69
|
282 # If there's a preamble, write it out, with a trailing CRLF
|
jpayne@69
|
283 if msg.preamble is not None:
|
jpayne@69
|
284 if self._mangle_from_:
|
jpayne@69
|
285 preamble = fcre.sub('>From ', msg.preamble)
|
jpayne@69
|
286 else:
|
jpayne@69
|
287 preamble = msg.preamble
|
jpayne@69
|
288 self._write_lines(preamble)
|
jpayne@69
|
289 self.write(self._NL)
|
jpayne@69
|
290 # dash-boundary transport-padding CRLF
|
jpayne@69
|
291 self.write('--' + boundary + self._NL)
|
jpayne@69
|
292 # body-part
|
jpayne@69
|
293 if msgtexts:
|
jpayne@69
|
294 self._fp.write(msgtexts.pop(0))
|
jpayne@69
|
295 # *encapsulation
|
jpayne@69
|
296 # --> delimiter transport-padding
|
jpayne@69
|
297 # --> CRLF body-part
|
jpayne@69
|
298 for body_part in msgtexts:
|
jpayne@69
|
299 # delimiter transport-padding CRLF
|
jpayne@69
|
300 self.write(self._NL + '--' + boundary + self._NL)
|
jpayne@69
|
301 # body-part
|
jpayne@69
|
302 self._fp.write(body_part)
|
jpayne@69
|
303 # close-delimiter transport-padding
|
jpayne@69
|
304 self.write(self._NL + '--' + boundary + '--' + self._NL)
|
jpayne@69
|
305 if msg.epilogue is not None:
|
jpayne@69
|
306 if self._mangle_from_:
|
jpayne@69
|
307 epilogue = fcre.sub('>From ', msg.epilogue)
|
jpayne@69
|
308 else:
|
jpayne@69
|
309 epilogue = msg.epilogue
|
jpayne@69
|
310 self._write_lines(epilogue)
|
jpayne@69
|
311
|
jpayne@69
|
312 def _handle_multipart_signed(self, msg):
|
jpayne@69
|
313 # The contents of signed parts has to stay unmodified in order to keep
|
jpayne@69
|
314 # the signature intact per RFC1847 2.1, so we disable header wrapping.
|
jpayne@69
|
315 # RDM: This isn't enough to completely preserve the part, but it helps.
|
jpayne@69
|
316 p = self.policy
|
jpayne@69
|
317 self.policy = p.clone(max_line_length=0)
|
jpayne@69
|
318 try:
|
jpayne@69
|
319 self._handle_multipart(msg)
|
jpayne@69
|
320 finally:
|
jpayne@69
|
321 self.policy = p
|
jpayne@69
|
322
|
jpayne@69
|
323 def _handle_message_delivery_status(self, msg):
|
jpayne@69
|
324 # We can't just write the headers directly to self's file object
|
jpayne@69
|
325 # because this will leave an extra newline between the last header
|
jpayne@69
|
326 # block and the boundary. Sigh.
|
jpayne@69
|
327 blocks = []
|
jpayne@69
|
328 for part in msg.get_payload():
|
jpayne@69
|
329 s = self._new_buffer()
|
jpayne@69
|
330 g = self.clone(s)
|
jpayne@69
|
331 g.flatten(part, unixfrom=False, linesep=self._NL)
|
jpayne@69
|
332 text = s.getvalue()
|
jpayne@69
|
333 lines = text.split(self._encoded_NL)
|
jpayne@69
|
334 # Strip off the unnecessary trailing empty line
|
jpayne@69
|
335 if lines and lines[-1] == self._encoded_EMPTY:
|
jpayne@69
|
336 blocks.append(self._encoded_NL.join(lines[:-1]))
|
jpayne@69
|
337 else:
|
jpayne@69
|
338 blocks.append(text)
|
jpayne@69
|
339 # Now join all the blocks with an empty line. This has the lovely
|
jpayne@69
|
340 # effect of separating each block with an empty line, but not adding
|
jpayne@69
|
341 # an extra one after the last one.
|
jpayne@69
|
342 self._fp.write(self._encoded_NL.join(blocks))
|
jpayne@69
|
343
|
jpayne@69
|
344 def _handle_message(self, msg):
|
jpayne@69
|
345 s = self._new_buffer()
|
jpayne@69
|
346 g = self.clone(s)
|
jpayne@69
|
347 # The payload of a message/rfc822 part should be a multipart sequence
|
jpayne@69
|
348 # of length 1. The zeroth element of the list should be the Message
|
jpayne@69
|
349 # object for the subpart. Extract that object, stringify it, and
|
jpayne@69
|
350 # write it out.
|
jpayne@69
|
351 # Except, it turns out, when it's a string instead, which happens when
|
jpayne@69
|
352 # and only when HeaderParser is used on a message of mime type
|
jpayne@69
|
353 # message/rfc822. Such messages are generated by, for example,
|
jpayne@69
|
354 # Groupwise when forwarding unadorned messages. (Issue 7970.) So
|
jpayne@69
|
355 # in that case we just emit the string body.
|
jpayne@69
|
356 payload = msg._payload
|
jpayne@69
|
357 if isinstance(payload, list):
|
jpayne@69
|
358 g.flatten(msg.get_payload(0), unixfrom=False, linesep=self._NL)
|
jpayne@69
|
359 payload = s.getvalue()
|
jpayne@69
|
360 else:
|
jpayne@69
|
361 payload = self._encode(payload)
|
jpayne@69
|
362 self._fp.write(payload)
|
jpayne@69
|
363
|
jpayne@69
|
364 # This used to be a module level function; we use a classmethod for this
|
jpayne@69
|
365 # and _compile_re so we can continue to provide the module level function
|
jpayne@69
|
366 # for backward compatibility by doing
|
jpayne@69
|
367 # _make_boundary = Generator._make_boundary
|
jpayne@69
|
368 # at the end of the module. It *is* internal, so we could drop that...
|
jpayne@69
|
369 @classmethod
|
jpayne@69
|
370 def _make_boundary(cls, text=None):
|
jpayne@69
|
371 # Craft a random boundary. If text is given, ensure that the chosen
|
jpayne@69
|
372 # boundary doesn't appear in the text.
|
jpayne@69
|
373 token = random.randrange(sys.maxsize)
|
jpayne@69
|
374 boundary = ('=' * 15) + (_fmt % token) + '=='
|
jpayne@69
|
375 if text is None:
|
jpayne@69
|
376 return boundary
|
jpayne@69
|
377 b = boundary
|
jpayne@69
|
378 counter = 0
|
jpayne@69
|
379 while True:
|
jpayne@69
|
380 cre = cls._compile_re('^--' + re.escape(b) + '(--)?$', re.MULTILINE)
|
jpayne@69
|
381 if not cre.search(text):
|
jpayne@69
|
382 break
|
jpayne@69
|
383 b = boundary + '.' + str(counter)
|
jpayne@69
|
384 counter += 1
|
jpayne@69
|
385 return b
|
jpayne@69
|
386
|
jpayne@69
|
387 @classmethod
|
jpayne@69
|
388 def _compile_re(cls, s, flags):
|
jpayne@69
|
389 return re.compile(s, flags)
|
jpayne@69
|
390
|
jpayne@69
|
391
|
jpayne@69
|
392 class BytesGenerator(Generator):
|
jpayne@69
|
393 """Generates a bytes version of a Message object tree.
|
jpayne@69
|
394
|
jpayne@69
|
395 Functionally identical to the base Generator except that the output is
|
jpayne@69
|
396 bytes and not string. When surrogates were used in the input to encode
|
jpayne@69
|
397 bytes, these are decoded back to bytes for output. If the policy has
|
jpayne@69
|
398 cte_type set to 7bit, then the message is transformed such that the
|
jpayne@69
|
399 non-ASCII bytes are properly content transfer encoded, using the charset
|
jpayne@69
|
400 unknown-8bit.
|
jpayne@69
|
401
|
jpayne@69
|
402 The outfp object must accept bytes in its write method.
|
jpayne@69
|
403 """
|
jpayne@69
|
404
|
jpayne@69
|
405 def write(self, s):
|
jpayne@69
|
406 self._fp.write(s.encode('ascii', 'surrogateescape'))
|
jpayne@69
|
407
|
jpayne@69
|
408 def _new_buffer(self):
|
jpayne@69
|
409 return BytesIO()
|
jpayne@69
|
410
|
jpayne@69
|
411 def _encode(self, s):
|
jpayne@69
|
412 return s.encode('ascii')
|
jpayne@69
|
413
|
jpayne@69
|
414 def _write_headers(self, msg):
|
jpayne@69
|
415 # This is almost the same as the string version, except for handling
|
jpayne@69
|
416 # strings with 8bit bytes.
|
jpayne@69
|
417 for h, v in msg.raw_items():
|
jpayne@69
|
418 self._fp.write(self.policy.fold_binary(h, v))
|
jpayne@69
|
419 # A blank line always separates headers from body
|
jpayne@69
|
420 self.write(self._NL)
|
jpayne@69
|
421
|
jpayne@69
|
422 def _handle_text(self, msg):
|
jpayne@69
|
423 # If the string has surrogates the original source was bytes, so
|
jpayne@69
|
424 # just write it back out.
|
jpayne@69
|
425 if msg._payload is None:
|
jpayne@69
|
426 return
|
jpayne@69
|
427 if _has_surrogates(msg._payload) and not self.policy.cte_type=='7bit':
|
jpayne@69
|
428 if self._mangle_from_:
|
jpayne@69
|
429 msg._payload = fcre.sub(">From ", msg._payload)
|
jpayne@69
|
430 self._write_lines(msg._payload)
|
jpayne@69
|
431 else:
|
jpayne@69
|
432 super(BytesGenerator,self)._handle_text(msg)
|
jpayne@69
|
433
|
jpayne@69
|
434 # Default body handler
|
jpayne@69
|
435 _writeBody = _handle_text
|
jpayne@69
|
436
|
jpayne@69
|
437 @classmethod
|
jpayne@69
|
438 def _compile_re(cls, s, flags):
|
jpayne@69
|
439 return re.compile(s.encode('ascii'), flags)
|
jpayne@69
|
440
|
jpayne@69
|
441
|
jpayne@69
|
442
|
jpayne@69
|
443 _FMT = '[Non-text (%(type)s) part of message omitted, filename %(filename)s]'
|
jpayne@69
|
444
|
jpayne@69
|
445 class DecodedGenerator(Generator):
|
jpayne@69
|
446 """Generates a text representation of a message.
|
jpayne@69
|
447
|
jpayne@69
|
448 Like the Generator base class, except that non-text parts are substituted
|
jpayne@69
|
449 with a format string representing the part.
|
jpayne@69
|
450 """
|
jpayne@69
|
451 def __init__(self, outfp, mangle_from_=None, maxheaderlen=None, fmt=None, *,
|
jpayne@69
|
452 policy=None):
|
jpayne@69
|
453 """Like Generator.__init__() except that an additional optional
|
jpayne@69
|
454 argument is allowed.
|
jpayne@69
|
455
|
jpayne@69
|
456 Walks through all subparts of a message. If the subpart is of main
|
jpayne@69
|
457 type `text', then it prints the decoded payload of the subpart.
|
jpayne@69
|
458
|
jpayne@69
|
459 Otherwise, fmt is a format string that is used instead of the message
|
jpayne@69
|
460 payload. fmt is expanded with the following keywords (in
|
jpayne@69
|
461 %(keyword)s format):
|
jpayne@69
|
462
|
jpayne@69
|
463 type : Full MIME type of the non-text part
|
jpayne@69
|
464 maintype : Main MIME type of the non-text part
|
jpayne@69
|
465 subtype : Sub-MIME type of the non-text part
|
jpayne@69
|
466 filename : Filename of the non-text part
|
jpayne@69
|
467 description: Description associated with the non-text part
|
jpayne@69
|
468 encoding : Content transfer encoding of the non-text part
|
jpayne@69
|
469
|
jpayne@69
|
470 The default value for fmt is None, meaning
|
jpayne@69
|
471
|
jpayne@69
|
472 [Non-text (%(type)s) part of message omitted, filename %(filename)s]
|
jpayne@69
|
473 """
|
jpayne@69
|
474 Generator.__init__(self, outfp, mangle_from_, maxheaderlen,
|
jpayne@69
|
475 policy=policy)
|
jpayne@69
|
476 if fmt is None:
|
jpayne@69
|
477 self._fmt = _FMT
|
jpayne@69
|
478 else:
|
jpayne@69
|
479 self._fmt = fmt
|
jpayne@69
|
480
|
jpayne@69
|
481 def _dispatch(self, msg):
|
jpayne@69
|
482 for part in msg.walk():
|
jpayne@69
|
483 maintype = part.get_content_maintype()
|
jpayne@69
|
484 if maintype == 'text':
|
jpayne@69
|
485 print(part.get_payload(decode=False), file=self)
|
jpayne@69
|
486 elif maintype == 'multipart':
|
jpayne@69
|
487 # Just skip this
|
jpayne@69
|
488 pass
|
jpayne@69
|
489 else:
|
jpayne@69
|
490 print(self._fmt % {
|
jpayne@69
|
491 'type' : part.get_content_type(),
|
jpayne@69
|
492 'maintype' : part.get_content_maintype(),
|
jpayne@69
|
493 'subtype' : part.get_content_subtype(),
|
jpayne@69
|
494 'filename' : part.get_filename('[no filename]'),
|
jpayne@69
|
495 'description': part.get('Content-Description',
|
jpayne@69
|
496 '[no description]'),
|
jpayne@69
|
497 'encoding' : part.get('Content-Transfer-Encoding',
|
jpayne@69
|
498 '[no encoding]'),
|
jpayne@69
|
499 }, file=self)
|
jpayne@69
|
500
|
jpayne@69
|
501
|
jpayne@69
|
502
|
jpayne@69
|
503 # Helper used by Generator._make_boundary
|
jpayne@69
|
504 _width = len(repr(sys.maxsize-1))
|
jpayne@69
|
505 _fmt = '%%0%dd' % _width
|
jpayne@69
|
506
|
jpayne@69
|
507 # Backward compatibility
|
jpayne@69
|
508 _make_boundary = Generator._make_boundary
|