Danny, I see your point. I am also beginning to come around to the idea that we want to change the wrapper, and require code to call the JavaMail APIs properly rather than "be nice" and try to anticipate.
Ironically, I didn't get a chance to read your message until just now, but as I have spent more time reading the JavaMail API, I had come to a similar thought regarding MimeMessageWrapper. I have been incrementally chipping away at its quirks, but I think the real solution is to clean it up drastically. It is beginning to appear to me that one reason why we may have needed the wrapper is because we haven't been using a JavaMail storage provider, which could provide us with a MimeMessage subclass optimized for that storage model. I think we might also find some assistance for the meta-data issue. --- Noel -----Original Message----- From: Danny Angus [mailto:[EMAIL PROTECTED]] Sent: Wednesday, February 05, 2003 14:01 To: James Developers List Subject: RE: Message-ID bug > The following seems to provide user control over Message-ID generation: > > http://www.magelang.com/faq/view.jsp?EID=516514 Yes it does Noel, but it won't work if writeTo() calls saveChanges() AND we need to create message id's where noe exist, OR we want ot create new message ID's for cloned messages. making people working with MimeMessage have to call saveChanges() is part of the Java Mail API, if there were no issues I'd have no problem with us taking that responsibility over (In MimeMessageWrapper which extends MimeMessage), but there are at least this one significant issue, and I think we have to back this code out in order to support MimeMessage correctly. Or document a different approach to message id generation. d. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]