' + re.escape(separator) +
r')(?P--)?(?P[ \t]*)(?P\r\n|\r|\n)?$')
capturing_preamble = True
preamble = []
linesep = False
while True:
line = self._input.readline()
if line is NeedMoreData:
yield NeedMoreData
continue
if line == '':
break
mo = boundaryre.match(line)
if mo:
# If we're looking at the end boundary, we're done with
# this multipart. If there was a newline at the end of
# the closing boundary, then we need to initialize the
# epilogue with the empty string (see below).
if mo.group('end'):
linesep = mo.group('linesep')
break
# We saw an inter-part boundary. Were we in the preamble?
if capturing_preamble:
if preamble:
# According to RFC 2046, the last newline belongs
# to the boundary.
lastline = preamble[-1]
eolmo = NLCRE_eol.search(lastline)
if eolmo:
preamble[-1] = lastline[:-len(eolmo.group(0))]
self._cur.preamble = EMPTYSTRING.join(preamble)
capturing_preamble = False
self._input.unreadline(line)
continue
# We saw a boundary separating two parts. Consume any
# multiple boundary lines that may be following. Our
# interpretation of RFC 2046 BNF grammar does not produce
# body parts within such double boundaries.
while True:
line = self._input.readline()
if line is NeedMoreData:
yield NeedMoreData
continue
mo = boundaryre.match(line)
if not mo:
self._input.unreadline(line)
break
# Recurse to parse this subpart; the input stream points
# at the subpart's first line.
self._input.push_eof_matcher(boundaryre.match)
for retval in self._parsegen():
if retval is NeedMoreData:
yield NeedMoreData
continue
break
# Because of RFC 2046, the newline preceding the boundary
# separator actually belongs to the boundary, not the
# previous subpart's payload (or epilogue if the previous
# part is a multipart).
if self._last.get_content_maintype() == 'multipart':
epilogue = self._last.epilogue
if epilogue == '':
self._last.epilogue = None
elif epilogue is not None:
mo = NLCRE_eol.search(epilogue)
if mo:
end = len(mo.group(0))
self._last.epilogue = epilogue[:-end]
else:
payload = self._last.get_payload()
if isinstance(payload, basestring):
mo = NLCRE_eol.search(payload)
if mo:
payload = payload[:-len(mo.group(0))]
self._last.set_payload(payload)
self._input.pop_eof_matcher()
self._pop_message()
# Set the multipart up for newline cleansing, which will
# happen if we're in a nested multipart.
self._last = self._cur
else:
# I think we must be in the preamble
assert capturing_preamble
preamble.append(line)
# We've seen either the EOF or the end boundary. If we're still
# capturing the preamble, we never saw the start boundary. Note
# that as a defect and store the captured text as the payload.
# Everything from here to the EOF is epilogue.
if capturing_preamble:
self._cur.defects.append(errors.StartBoundaryNotFoundDefect())
self._cur.set_payload(EMPTYSTRING.join(preamble))
epilogue = []
for line in self._input:
if line is NeedMoreData:
yield NeedMoreData
continue
self._cur.epilogue = EMPTYSTRING.join(epilogue)
return
# If the end boundary ended in a newline, we'll need to make sure
# the epilogue isn't None
if linesep:
epilogue = ['']
else:
epilogue = []
for line in self._input:
if line is NeedMoreData:
yield NeedMoreData
continue
epilogue.append(line)
# Any CRLF at the front of the epilogue is not technically part of
# the epilogue. Also, watch out for an empty string epilogue,
# which means a single newline.
if epilogue:
firstline = epilogue[0]
bolmo = NLCRE_bol.match(firstline)
if bolmo:
epilogue[0] = firstline[len(bolmo.group(0)):]
self._cur.epilogue = EMPTYSTRING.join(epilogue)
return
# Otherwise, it's some non-multipart type, so the entire rest of the
# file contents becomes the payload.
lines = []
for line in self._input:
if line is NeedMoreData:
yield NeedMoreData
continue
lines.append(line)
self._cur.set_payload(EMPTYSTRING.join(lines))
def _parse_headers(self, lines):
# Passed a list of lines that make up the headers for the current msg
lastheader = ''
lastvalue = []
for lineno, line in enumerate(lines):
# Check for continuation
if line[0] in ' \t':
if not lastheader:
# The first line of the headers was a continuation. This
# is illegal, so let's note the defect, store the illegal
# line, and ignore it for purposes of headers.
defect = errors.FirstHeaderLineIsContinuationDefect(line)
self._cur.defects.append(defect)
continue
lastvalue.append(line)
continue
if lastheader:
# XXX reconsider the joining of folded lines
lhdr = EMPTYSTRING.join(lastvalue)[:-1].rstrip('\r\n')
self._cur[lastheader] = lhdr
lastheader, lastvalue = '', []
# Check for envelope header, i.e. unix-from
if line.startswith('From '):
if lineno == 0:
# Strip off the trailing newline
mo = NLCRE_eol.search(line)
if mo:
line = line[:-len(mo.group(0))]
self._cur.set_unixfrom(line)
continue
elif lineno == len(lines) - 1:
# Something looking like a unix-from at the end - it's
# probably the first line of the body, so push back the
# line and stop.
self._input.unreadline(line)
return
else:
# Weirdly placed unix-from line. Note this as a defect
# and ignore it.
defect = errors.MisplacedEnvelopeHeaderDefect(line)
self._cur.defects.append(defect)
continue
# Split the line on the colon separating field name from value.
i = line.find(':')
if i < 0:
defect = errors.MalformedHeaderDefect(line)
self._cur.defects.append(defect)
continue
lastheader = line[:i]
lastvalue = [line[i+1:].lstrip()]
# Done with all the lines, so handle the last header.
if lastheader:
# XXX reconsider the joining of folded lines
self._cur[lastheader] = EMPTYSTRING.join(lastvalue).rstrip('\r\n')