On Sun, 2002-04-28 at 23:57, Lex Spoon wrote:
> Ken Causey <ken at ineffable.com> wrote:
> > I've been trying to parse email messages with more than the simplest of
> > structures and can't seem to get anywhere.
> > 
> > VM: Linux 3.2-4743 #1 [oss audio, xshm]
> > Image: 3.2gamma-4843
> > 
> > Consider an example:  Take an email with an attachment, and forward it
> > as an attachment.  Diagram:
> > 
> > a Email
> > b   Body
> > c   Forwarded Email
> > d     Body of Forwarded Email
> > e     Attachment
> > 
> > The problem is that no matter what I do, I can't get to the attachment,
> > part e.  
> 
> 
> This works for me:
> 
> 	a := MailMessage from: someText.
> 	c := MailMessage from: a parts second body content.
> 	e := c parts second.
> 	attachmentData := e body content
> 
> This seems fairly direct.
> 
> 
> Now, MailMessage isn't the cleanest of all cleanly code in the world.  If you
> want to clean it up and document your new design, that would be great!
> 
> Lex
> 

OK, that's very interesting as this does work.  But this indicates even
more clearly that the MailMessage instances returned as a result of
MailMessage>>parts are apparently not correctly formed, since they are
not functionally equivalent to instantiating an instance manually from
the text using MailMessage class>>from:.

Thanks for this workaround.

Ken Causey (nikos)