||Generic MIME writer.|
A MIME writer is much more primitive than a MIME parser. It
doesn't seek around on the output file, and it doesn't use large
amounts of buffer space, so you have to write the parts in the
order they should occur on the output file. It does buffer the
headers you add, allowing you to rearrange their order.
General usage is:
f = <open the output file>
w = MimeWriter(f)
...call w.addheader(key, value) 0 or more times...
followed by either:
f = w.startbody(content_type)
...call f.write(data) for body data...
for each part:
subwriter = w.nextpart()
...use the subwriter's methods to create the subpart...
The subwriter is another MimeWriter instance, and should be
treated in the same way as the toplevel MimeWriter. This way,
writing recursive body parts is easy.
Warning: don't forget to call lastpart()!
XXX There should be more state so calls made in the wrong order
Some special cases:
- startbody() just returns the file passed to the constructor;
but don't use this knowledge, as it may be changed.
- startmultipartbody() actually returns a file as well;
this can be used to write the initial 'if you can read this your
mailer is not MIME-aware' message.
- If you call flushheaders(), the headers accumulated so far are
written out (and forgotten); this is useful if you don't need a
body part at all, e.g. for a subpart of type message/rfc822
that's (mis)used to store some header-like information.
- Passing a keyword argument 'prefix=<flag>' to addheader(),
start*body() affects where the header is inserted; 0 means
append at the end, 1 means insert at the start; default is
append for addheader(), but insert for start*body(), which use
it to determine where the Content-Type header goes.
- __init__(self, fp)
- addheader(self, key, value, prefix=0)
- startbody(self, ctype, plist=, prefix=1)
- startmultipartbody(self, subtype, boundary=None, plist=, prefix=1)