summaryrefslogtreecommitdiffstats
path: root/Lib/email/message.py
diff options
context:
space:
mode:
authorGuido van Rossum <guido@python.org>2007-08-25 13:43:02 (GMT)
committerGuido van Rossum <guido@python.org>2007-08-25 13:43:02 (GMT)
commit6398b7a351d04408fa8ce0204d78559e8d6ac98f (patch)
tree5f274983fa91009d1a7b4bfa26de59a83942581f /Lib/email/message.py
parentf616b224506ef17ac9ab4378c92b99ae816b88f7 (diff)
downloadcpython-6398b7a351d04408fa8ce0204d78559e8d6ac98f.zip
cpython-6398b7a351d04408fa8ce0204d78559e8d6ac98f.tar.gz
cpython-6398b7a351d04408fa8ce0204d78559e8d6ac98f.tar.bz2
Remove the email package for now.
Once Barry and the email-sig have a working new version we'll add it back. If it doesn't make the 3.0a deadline (release August 31), too bad.
Diffstat (limited to 'Lib/email/message.py')
-rw-r--r--Lib/email/message.py786
1 files changed, 0 insertions, 786 deletions
diff --git a/Lib/email/message.py b/Lib/email/message.py
deleted file mode 100644
index 78e7d82..0000000
--- a/Lib/email/message.py
+++ /dev/null
@@ -1,786 +0,0 @@
-# Copyright (C) 2001-2006 Python Software Foundation
-# Author: Barry Warsaw
-# Contact: email-sig@python.org
-
-"""Basic message object for the email package object model."""
-
-__all__ = ['Message']
-
-import re
-import uu
-import binascii
-import warnings
-from io import StringIO
-
-# Intrapackage imports
-import email.charset
-from email import utils
-from email import errors
-
-SEMISPACE = '; '
-
-# Regular expression used to split header parameters. BAW: this may be too
-# simple. It isn't strictly RFC 2045 (section 5.1) compliant, but it catches
-# most headers found in the wild. We may eventually need a full fledged
-# parser eventually.
-paramre = re.compile(r'\s*;\s*')
-# Regular expression that matches `special' characters in parameters, the
-# existance of which force quoting of the parameter value.
-tspecials = re.compile(r'[ \(\)<>@,;:\\"/\[\]\?=]')
-
-
-
-# Helper functions
-def _formatparam(param, value=None, quote=True):
- """Convenience function to format and return a key=value pair.
-
- This will quote the value if needed or if quote is true.
- """
- if value is not None and len(value) > 0:
- # A tuple is used for RFC 2231 encoded parameter values where items
- # are (charset, language, value). charset is a string, not a Charset
- # instance.
- if isinstance(value, tuple):
- # Encode as per RFC 2231
- param += '*'
- value = utils.encode_rfc2231(value[2], value[0], value[1])
- # BAW: Please check this. I think that if quote is set it should
- # force quoting even if not necessary.
- if quote or tspecials.search(value):
- return '%s="%s"' % (param, utils.quote(value))
- else:
- return '%s=%s' % (param, value)
- else:
- return param
-
-def _parseparam(s):
- plist = []
- while s[:1] == ';':
- s = s[1:]
- end = s.find(';')
- while end > 0 and s.count('"', 0, end) % 2:
- end = s.find(';', end + 1)
- if end < 0:
- end = len(s)
- f = s[:end]
- if '=' in f:
- i = f.index('=')
- f = f[:i].strip().lower() + '=' + f[i+1:].strip()
- plist.append(f.strip())
- s = s[end:]
- return plist
-
-
-def _unquotevalue(value):
- # This is different than utils.collapse_rfc2231_value() because it doesn't
- # try to convert the value to a unicode. Message.get_param() and
- # Message.get_params() are both currently defined to return the tuple in
- # the face of RFC 2231 parameters.
- if isinstance(value, tuple):
- return value[0], value[1], utils.unquote(value[2])
- else:
- return utils.unquote(value)
-
-
-
-class Message:
- """Basic message object.
-
- A message object is defined as something that has a bunch of RFC 2822
- headers and a payload. It may optionally have an envelope header
- (a.k.a. Unix-From or From_ header). If the message is a container (i.e. a
- multipart or a message/rfc822), then the payload is a list of Message
- objects, otherwise it is a string.
-
- Message objects implement part of the `mapping' interface, which assumes
- there is exactly one occurrance of the header per message. Some headers
- do in fact appear multiple times (e.g. Received) and for those headers,
- you must use the explicit API to set or get all the headers. Not all of
- the mapping methods are implemented.
- """
- def __init__(self):
- self._headers = []
- self._unixfrom = None
- self._payload = None
- self._charset = None
- # Defaults for multipart messages
- self.preamble = self.epilogue = None
- self.defects = []
- # Default content type
- self._default_type = 'text/plain'
-
- def __str__(self):
- """Return the entire formatted message as a string.
- This includes the headers, body, and envelope header.
- """
- return self.as_string(unixfrom=True)
-
- def as_string(self, unixfrom=False):
- """Return the entire formatted message as a string.
- Optional `unixfrom' when True, means include the Unix From_ envelope
- header.
-
- This is a convenience method and may not generate the message exactly
- as you intend because by default it mangles lines that begin with
- "From ". For more flexibility, use the flatten() method of a
- Generator instance.
- """
- from email.Generator import Generator
- fp = StringIO()
- g = Generator(fp)
- g.flatten(self, unixfrom=unixfrom)
- return fp.getvalue()
-
- def is_multipart(self):
- """Return True if the message consists of multiple parts."""
- return isinstance(self._payload, list)
-
- #
- # Unix From_ line
- #
- def set_unixfrom(self, unixfrom):
- self._unixfrom = unixfrom
-
- def get_unixfrom(self):
- return self._unixfrom
-
- #
- # Payload manipulation.
- #
- def attach(self, payload):
- """Add the given payload to the current payload.
-
- The current payload will always be a list of objects after this method
- is called. If you want to set the payload to a scalar object, use
- set_payload() instead.
- """
- if self._payload is None:
- self._payload = [payload]
- else:
- self._payload.append(payload)
-
- def get_payload(self, i=None, decode=False):
- """Return a reference to the payload.
-
- The payload will either be a list object or a string. If you mutate
- the list object, you modify the message's payload in place. Optional
- i returns that index into the payload.
-
- Optional decode is a flag indicating whether the payload should be
- decoded or not, according to the Content-Transfer-Encoding header
- (default is False).
-
- When True and the message is not a multipart, the payload will be
- decoded if this header's value is `quoted-printable' or `base64'. If
- some other encoding is used, or the header is missing, or if the
- payload has bogus data (i.e. bogus base64 or uuencoded data), the
- payload is returned as-is.
-
- If the message is a multipart and the decode flag is True, then None
- is returned.
- """
- if i is None:
- payload = self._payload
- elif not isinstance(self._payload, list):
- raise TypeError('Expected list, got %s' % type(self._payload))
- else:
- payload = self._payload[i]
- if decode:
- if self.is_multipart():
- return None
- cte = self.get('content-transfer-encoding', '').lower()
- if cte == 'quoted-printable':
- return utils._qdecode(payload)
- elif cte == 'base64':
- try:
- return utils._bdecode(payload)
- except binascii.Error:
- # Incorrect padding
- return payload
- elif cte in ('x-uuencode', 'uuencode', 'uue', 'x-uue'):
- sfp = StringIO()
- try:
- uu.decode(StringIO(payload+'\n'), sfp, quiet=True)
- payload = sfp.getvalue()
- except uu.Error:
- # Some decoding problem
- return payload
- # Everything else, including encodings with 8bit or 7bit are returned
- # unchanged.
- return payload
-
- def set_payload(self, payload, charset=None):
- """Set the payload to the given value.
-
- Optional charset sets the message's default character set. See
- set_charset() for details.
- """
- self._payload = payload
- if charset is not None:
- self.set_charset(charset)
-
- def set_charset(self, charset):
- """Set the charset of the payload to a given character set.
-
- charset can be a Charset instance, a string naming a character set, or
- None. If it is a string it will be converted to a Charset instance.
- If charset is None, the charset parameter will be removed from the
- Content-Type field. Anything else will generate a TypeError.
-
- The message will be assumed to be of type text/* encoded with
- charset.input_charset. It will be converted to charset.output_charset
- and encoded properly, if needed, when generating the plain text
- representation of the message. MIME headers (MIME-Version,
- Content-Type, Content-Transfer-Encoding) will be added as needed.
-
- """
- if charset is None:
- self.del_param('charset')
- self._charset = None
- return
- if isinstance(charset, basestring):
- charset = email.charset.Charset(charset)
- if not isinstance(charset, email.charset.Charset):
- raise TypeError(charset)
- # BAW: should we accept strings that can serve as arguments to the
- # Charset constructor?
- self._charset = charset
- if 'MIME-Version' not in self:
- self.add_header('MIME-Version', '1.0')
- if 'Content-Type' not in self:
- self.add_header('Content-Type', 'text/plain',
- charset=charset.get_output_charset())
- else:
- self.set_param('charset', charset.get_output_charset())
- if str(charset) != charset.get_output_charset():
- self._payload = charset.body_encode(self._payload)
- if 'Content-Transfer-Encoding' not in self:
- cte = charset.get_body_encoding()
- try:
- cte(self)
- except TypeError:
- self._payload = charset.body_encode(self._payload)
- self.add_header('Content-Transfer-Encoding', cte)
-
- def get_charset(self):
- """Return the Charset instance associated with the message's payload.
- """
- return self._charset
-
- #
- # MAPPING INTERFACE (partial)
- #
- def __len__(self):
- """Return the total number of headers, including duplicates."""
- return len(self._headers)
-
- def __getitem__(self, name):
- """Get a header value.
-
- Return None if the header is missing instead of raising an exception.
-
- Note that if the header appeared multiple times, exactly which
- occurrance gets returned is undefined. Use get_all() to get all
- the values matching a header field name.
- """
- return self.get(name)
-
- def __setitem__(self, name, val):
- """Set the value of a header.
-
- Note: this does not overwrite an existing header with the same field
- name. Use __delitem__() first to delete any existing headers.
- """
- self._headers.append((name, val))
-
- def __delitem__(self, name):
- """Delete all occurrences of a header, if present.
-
- Does not raise an exception if the header is missing.
- """
- name = name.lower()
- newheaders = []
- for k, v in self._headers:
- if k.lower() != name:
- newheaders.append((k, v))
- self._headers = newheaders
-
- def __contains__(self, name):
- return name.lower() in [k.lower() for k, v in self._headers]
-
- def has_key(self, name):
- """Return true if the message contains the header."""
- missing = object()
- return self.get(name, missing) is not missing
-
- def keys(self):
- """Return a list of all the message's header field names.
-
- These will be sorted in the order they appeared in the original
- message, or were added to the message, and may contain duplicates.
- Any fields deleted and re-inserted are always appended to the header
- list.
- """
- return [k for k, v in self._headers]
-
- def values(self):
- """Return a list of all the message's header values.
-
- These will be sorted in the order they appeared in the original
- message, or were added to the message, and may contain duplicates.
- Any fields deleted and re-inserted are always appended to the header
- list.
- """
- return [v for k, v in self._headers]
-
- def items(self):
- """Get all the message's header fields and values.
-
- These will be sorted in the order they appeared in the original
- message, or were added to the message, and may contain duplicates.
- Any fields deleted and re-inserted are always appended to the header
- list.
- """
- return self._headers[:]
-
- def get(self, name, failobj=None):
- """Get a header value.
-
- Like __getitem__() but return failobj instead of None when the field
- is missing.
- """
- name = name.lower()
- for k, v in self._headers:
- if k.lower() == name:
- return v
- return failobj
-
- #
- # Additional useful stuff
- #
-
- def get_all(self, name, failobj=None):
- """Return a list of all the values for the named field.
-
- These will be sorted in the order they appeared in the original
- message, and may contain duplicates. Any fields deleted and
- re-inserted are always appended to the header list.
-
- If no such fields exist, failobj is returned (defaults to None).
- """
- values = []
- name = name.lower()
- for k, v in self._headers:
- if k.lower() == name:
- values.append(v)
- if not values:
- return failobj
- return values
-
- def add_header(self, _name, _value, **_params):
- """Extended header setting.
-
- name is the header field to add. keyword arguments can be used to set
- additional parameters for the header field, with underscores converted
- to dashes. Normally the parameter will be added as key="value" unless
- value is None, in which case only the key will be added.
-
- Example:
-
- msg.add_header('content-disposition', 'attachment', filename='bud.gif')
- """
- parts = []
- for k, v in _params.items():
- if v is None:
- parts.append(k.replace('_', '-'))
- else:
- parts.append(_formatparam(k.replace('_', '-'), v))
- if _value is not None:
- parts.insert(0, _value)
- self._headers.append((_name, SEMISPACE.join(parts)))
-
- def replace_header(self, _name, _value):
- """Replace a header.
-
- Replace the first matching header found in the message, retaining
- header order and case. If no matching header was found, a KeyError is
- raised.
- """
- _name = _name.lower()
- for i, (k, v) in zip(range(len(self._headers)), self._headers):
- if k.lower() == _name:
- self._headers[i] = (k, _value)
- break
- else:
- raise KeyError(_name)
-
- #
- # Use these three methods instead of the three above.
- #
-
- def get_content_type(self):
- """Return the message's content type.
-
- The returned string is coerced to lower case of the form
- `maintype/subtype'. If there was no Content-Type header in the
- message, the default type as given by get_default_type() will be
- returned. Since according to RFC 2045, messages always have a default
- type this will always return a value.
-
- RFC 2045 defines a message's default type to be text/plain unless it
- appears inside a multipart/digest container, in which case it would be
- message/rfc822.
- """
- missing = object()
- value = self.get('content-type', missing)
- if value is missing:
- # This should have no parameters
- return self.get_default_type()
- ctype = paramre.split(value)[0].lower().strip()
- # RFC 2045, section 5.2 says if its invalid, use text/plain
- if ctype.count('/') != 1:
- return 'text/plain'
- return ctype
-
- def get_content_maintype(self):
- """Return the message's main content type.
-
- This is the `maintype' part of the string returned by
- get_content_type().
- """
- ctype = self.get_content_type()
- return ctype.split('/')[0]
-
- def get_content_subtype(self):
- """Returns the message's sub-content type.
-
- This is the `subtype' part of the string returned by
- get_content_type().
- """
- ctype = self.get_content_type()
- return ctype.split('/')[1]
-
- def get_default_type(self):
- """Return the `default' content type.
-
- Most messages have a default content type of text/plain, except for
- messages that are subparts of multipart/digest containers. Such
- subparts have a default content type of message/rfc822.
- """
- return self._default_type
-
- def set_default_type(self, ctype):
- """Set the `default' content type.
-
- ctype should be either "text/plain" or "message/rfc822", although this
- is not enforced. The default content type is not stored in the
- Content-Type header.
- """
- self._default_type = ctype
-
- def _get_params_preserve(self, failobj, header):
- # Like get_params() but preserves the quoting of values. BAW:
- # should this be part of the public interface?
- missing = object()
- value = self.get(header, missing)
- if value is missing:
- return failobj
- params = []
- for p in _parseparam(';' + value):
- try:
- name, val = p.split('=', 1)
- name = name.strip()
- val = val.strip()
- except ValueError:
- # Must have been a bare attribute
- name = p.strip()
- val = ''
- params.append((name, val))
- params = utils.decode_params(params)
- return params
-
- def get_params(self, failobj=None, header='content-type', unquote=True):
- """Return the message's Content-Type parameters, as a list.
-
- The elements of the returned list are 2-tuples of key/value pairs, as
- split on the `=' sign. The left hand side of the `=' is the key,
- while the right hand side is the value. If there is no `=' sign in
- the parameter the value is the empty string. The value is as
- described in the get_param() method.
-
- Optional failobj is the object to return if there is no Content-Type
- header. Optional header is the header to search instead of
- Content-Type. If unquote is True, the value is unquoted.
- """
- missing = object()
- params = self._get_params_preserve(missing, header)
- if params is missing:
- return failobj
- if unquote:
- return [(k, _unquotevalue(v)) for k, v in params]
- else:
- return params
-
- def get_param(self, param, failobj=None, header='content-type',
- unquote=True):
- """Return the parameter value if found in the Content-Type header.
-
- Optional failobj is the object to return if there is no Content-Type
- header, or the Content-Type header has no such parameter. Optional
- header is the header to search instead of Content-Type.
-
- Parameter keys are always compared case insensitively. The return
- value can either be a string, or a 3-tuple if the parameter was RFC
- 2231 encoded. When it's a 3-tuple, the elements of the value are of
- the form (CHARSET, LANGUAGE, VALUE). Note that both CHARSET and
- LANGUAGE can be None, in which case you should consider VALUE to be
- encoded in the us-ascii charset. You can usually ignore LANGUAGE.
-
- Your application should be prepared to deal with 3-tuple return
- values, and can convert the parameter to a Unicode string like so:
-
- param = msg.get_param('foo')
- if isinstance(param, tuple):
- param = unicode(param[2], param[0] or 'us-ascii')
-
- In any case, the parameter value (either the returned string, or the
- VALUE item in the 3-tuple) is always unquoted, unless unquote is set
- to False.
- """
- if header not in self:
- return failobj
- for k, v in self._get_params_preserve(failobj, header):
- if k.lower() == param.lower():
- if unquote:
- return _unquotevalue(v)
- else:
- return v
- return failobj
-
- def set_param(self, param, value, header='Content-Type', requote=True,
- charset=None, language=''):
- """Set a parameter in the Content-Type header.
-
- If the parameter already exists in the header, its value will be
- replaced with the new value.
-
- If header is Content-Type and has not yet been defined for this
- message, it will be set to "text/plain" and the new parameter and
- value will be appended as per RFC 2045.
-
- An alternate header can specified in the header argument, and all
- parameters will be quoted as necessary unless requote is False.
-
- If charset is specified, the parameter will be encoded according to RFC
- 2231. Optional language specifies the RFC 2231 language, defaulting
- to the empty string. Both charset and language should be strings.
- """
- if not isinstance(value, tuple) and charset:
- value = (charset, language, value)
-
- if header not in self and header.lower() == 'content-type':
- ctype = 'text/plain'
- else:
- ctype = self.get(header)
- if not self.get_param(param, header=header):
- if not ctype:
- ctype = _formatparam(param, value, requote)
- else:
- ctype = SEMISPACE.join(
- [ctype, _formatparam(param, value, requote)])
- else:
- ctype = ''
- for old_param, old_value in self.get_params(header=header,
- unquote=requote):
- append_param = ''
- if old_param.lower() == param.lower():
- append_param = _formatparam(param, value, requote)
- else:
- append_param = _formatparam(old_param, old_value, requote)
- if not ctype:
- ctype = append_param
- else:
- ctype = SEMISPACE.join([ctype, append_param])
- if ctype != self.get(header):
- del self[header]
- self[header] = ctype
-
- def del_param(self, param, header='content-type', requote=True):
- """Remove the given parameter completely from the Content-Type header.
-
- The header will be re-written in place without the parameter or its
- value. All values will be quoted as necessary unless requote is
- False. Optional header specifies an alternative to the Content-Type
- header.
- """
- if header not in self:
- return
- new_ctype = ''
- for p, v in self.get_params(header=header, unquote=requote):
- if p.lower() != param.lower():
- if not new_ctype:
- new_ctype = _formatparam(p, v, requote)
- else:
- new_ctype = SEMISPACE.join([new_ctype,
- _formatparam(p, v, requote)])
- if new_ctype != self.get(header):
- del self[header]
- self[header] = new_ctype
-
- def set_type(self, type, header='Content-Type', requote=True):
- """Set the main type and subtype for the Content-Type header.
-
- type must be a string in the form "maintype/subtype", otherwise a
- ValueError is raised.
-
- This method replaces the Content-Type header, keeping all the
- parameters in place. If requote is False, this leaves the existing
- header's quoting as is. Otherwise, the parameters will be quoted (the
- default).
-
- An alternative header can be specified in the header argument. When
- the Content-Type header is set, we'll always also add a MIME-Version
- header.
- """
- # BAW: should we be strict?
- if not type.count('/') == 1:
- raise ValueError
- # Set the Content-Type, you get a MIME-Version
- if header.lower() == 'content-type':
- del self['mime-version']
- self['MIME-Version'] = '1.0'
- if header not in self:
- self[header] = type
- return
- params = self.get_params(header=header, unquote=requote)
- del self[header]
- self[header] = type
- # Skip the first param; it's the old type.
- for p, v in params[1:]:
- self.set_param(p, v, header, requote)
-
- def get_filename(self, failobj=None):
- """Return the filename associated with the payload if present.
-
- The filename is extracted from the Content-Disposition header's
- `filename' parameter, and it is unquoted. If that header is missing
- the `filename' parameter, this method falls back to looking for the
- `name' parameter.
- """
- missing = object()
- filename = self.get_param('filename', missing, 'content-disposition')
- if filename is missing:
- filename = self.get_param('name', missing, 'content-disposition')
- if filename is missing:
- return failobj
- return utils.collapse_rfc2231_value(filename).strip()
-
- def get_boundary(self, failobj=None):
- """Return the boundary associated with the payload if present.
-
- The boundary is extracted from the Content-Type header's `boundary'
- parameter, and it is unquoted.
- """
- missing = object()
- boundary = self.get_param('boundary', missing)
- if boundary is missing:
- return failobj
- # RFC 2046 says that boundaries may begin but not end in w/s
- return utils.collapse_rfc2231_value(boundary).rstrip()
-
- def set_boundary(self, boundary):
- """Set the boundary parameter in Content-Type to 'boundary'.
-
- This is subtly different than deleting the Content-Type header and
- adding a new one with a new boundary parameter via add_header(). The
- main difference is that using the set_boundary() method preserves the
- order of the Content-Type header in the original message.
-
- HeaderParseError is raised if the message has no Content-Type header.
- """
- missing = object()
- params = self._get_params_preserve(missing, 'content-type')
- if params is missing:
- # There was no Content-Type header, and we don't know what type
- # to set it to, so raise an exception.
- raise errors.HeaderParseError('No Content-Type header found')
- newparams = []
- foundp = False
- for pk, pv in params:
- if pk.lower() == 'boundary':
- newparams.append(('boundary', '"%s"' % boundary))
- foundp = True
- else:
- newparams.append((pk, pv))
- if not foundp:
- # The original Content-Type header had no boundary attribute.
- # Tack one on the end. BAW: should we raise an exception
- # instead???
- newparams.append(('boundary', '"%s"' % boundary))
- # Replace the existing Content-Type header with the new value
- newheaders = []
- for h, v in self._headers:
- if h.lower() == 'content-type':
- parts = []
- for k, v in newparams:
- if v == '':
- parts.append(k)
- else:
- parts.append('%s=%s' % (k, v))
- newheaders.append((h, SEMISPACE.join(parts)))
-
- else:
- newheaders.append((h, v))
- self._headers = newheaders
-
- def get_content_charset(self, failobj=None):
- """Return the charset parameter of the Content-Type header.
-
- The returned string is always coerced to lower case. If there is no
- Content-Type header, or if that header has no charset parameter,
- failobj is returned.
- """
- missing = object()
- charset = self.get_param('charset', missing)
- if charset is missing:
- return failobj
- if isinstance(charset, tuple):
- # RFC 2231 encoded, so decode it, and it better end up as ascii.
- pcharset = charset[0] or 'us-ascii'
- try:
- # LookupError will be raised if the charset isn't known to
- # Python. UnicodeError will be raised if the encoded text
- # contains a character not in the charset.
- charset = str(charset[2], pcharset).encode('us-ascii')
- except (LookupError, UnicodeError):
- charset = charset[2]
- # charset character must be in us-ascii range
- try:
- if isinstance(charset, str):
- charset = str(charset, 'us-ascii')
- charset = charset.encode('us-ascii')
- except UnicodeError:
- return failobj
- # RFC 2046, $4.1.2 says charsets are not case sensitive
- return charset.lower()
-
- def get_charsets(self, failobj=None):
- """Return a list containing the charset(s) used in this message.
-
- The returned list of items describes the Content-Type headers'
- charset parameter for this message and all the subparts in its
- payload.
-
- Each item will either be a string (the value of the charset parameter
- in the Content-Type header of that part) or the value of the
- 'failobj' parameter (defaults to None), if the part does not have a
- main MIME type of "text", or the charset is not defined.
-
- The list will contain one string for each part of the message, plus
- one for the container message (i.e. self), so that a non-multipart
- message will still return a list of length 1.
- """
- return [part.get_content_charset(failobj) for part in self.walk()]
-
- # I.e. def walk(self): ...
- from email.Iterators import walk