Tony Nelson writes:
> How about:
>
> message['Subject'] = 'subject string'
> message['To'] = ('joe', 'joe...@foo.com')
>
> Since the Header does indeed know what it is
How? Since there's no explicit constructor, in fact the Header
doesn't know what it is until the Message tells it what it is. That
means that the registry of Header types must be known to the Message
class. That may not be a burden on the clients of email, but I can't
see it as a warm fuzzy for the maintainers of email.
> Internally, Message.__setitem__() would look up the name, making
> and assigning the proper Headere subclass if missing, and pass that
> object the data. The Header subclass knows what type of data it
> expects and raises (ValueError?) if it gets something
> inappropriate.
I don't think the FLUFL will accept that. First of all, "Mama don'
'low no raisin's round heya." Second, the duck-typing on the 'To'
example is a little hairy. Since in your model message['To']
*produces* a sequence of pairs when evaluated, I would expect it to
require a sequence of pairs on input. But you seem to be suggesting
that if it's a sequence but not a sequence of sequences, it should
handle that by assuming it's a (name,addr) pair. Heck, maybe we
should do something reasonable if it's not a sequence.
I think this is *way* too magical for the basic API of email.
I'm also bothered by the complexity of having Message accept
responsibility for the validity of data to be input to Header, then
delegate that responsiblity back to Header. Users don't care, I
suspect, but maintainers will. There will be an extra frame for
Message in any trace causing by Header raising ValueError (or
whatever), which is annoying since Message should be just passing the
argument on ... but it needs to be remembered or looked up, and there
will always be the temptation for developers to add a little smarts to
Message's handling of Header's arguments (especially in derived
classes, eg, a ListPost class that optionally prepends "[listname] "
to the front of a post).
Finally, I know that if I'm away from the email package for more than
24 hours, I'll forget which order the display name and address come
in.
_______________________________________________
Email-SIG mailing list
Email-SIG@python.org
Your options:
http://mail.python.org/mailman/options/email-sig/archive%40mail-archive.com
