home *** CD-ROM | disk | FTP | other *** search
-
- > In order to send a MESSAGE attachment, you have to put it
- > on body-contents.text like you would any other type of attachment.
- > See the routine rfc822_encode_body() in rfc822.c for how this
- > works.
-
- Yes, I was tracing through yesterday, and found the reference in
- rfc822_encode_body(). You might do implementors a small favor by
- changing:
-
- /* case MESSAGE: */ /* here for documentation */
-
- to "TYPEMESSAGE" so it's easier to grep. ;-)
- (Perhaps you've done this already). I might have found it long ago...
-
- > I will be the first to admit that this is ugly, hideous,
- > and inconsistent and I'll probably be reincarnated as a banana
- > slug for doing it that way, but it was easier to program that
- > way, both in c-client and in the main programs.
-
- I agree it's ugly.
-
-
- > Probably I'll get reincarnated as a banana slug right
- > next to a hungry garter snake (or a fellow with a salt shaker)
- > for this, but Internal.DOC is way out of date (almost a year
- > old). I don't think there are very many lies in it and
- > certainly no egregious ones, but it is missing quite a bit
- > that was added or changed in the past year. The only 100%
- > reliable document right now is the c-client sources, so when
- > in doubt you should refer to them.
- >
- > I'll probably update Internal.DOC sometime after updating
- > the IMAP2 protocol specification, which is a looming Urgent
- > Problem.
-
- I try to use the sources when I can, but not yet knowing them
- intimately, the fact that most function calls are through function pointers (I
- understand and agree with the reasons for doing this), it makes it difficult to
- really find out the logic flow without single stepping (no fun). I find that
- the doc is, as you say, almost always correct; but we're left swinging a bit on
- actually using the MIME interface. But this isn't a flame. As long as I can
- work through it somehow...
-
- But, I may have uncovered a bug "somewhere" in the client related to
- the rfc822 attachments. I haven't been able to pinpoint it yet. The scenario
- is, I build the body (with rfc822 attachment sitting in contents.text) and pass
- it smtp_mail(). When it returns from that function, the msg.env pointer for the
- attachment has been altered, i.e. is non-zero, but isn't a valid envelope. This
- causes seg faults immediately afterward when mail_free_body() is called to
- de-allocate the entire body structure. I'm punting temporarily; going through
- the partlist and zeroing envelopes for rfc822 attachments before freeing the
- body. This could cause me to be re-incarnated as a dodo bird, so I could use a
- little help.
-
- mike
-
-
-