[BUG] Celeste sometimes wraps MIME message headers when sending

Lex Spoon lex at cc.gatech.edu
Fri Sep 8 21:28:25 UTC 2000


Daniel Vainsencher <danielv at netvision.net.il> wrote:
> mdr at scn.org (Mike Rutenberg) wrote:
> > I'll have to think about how to fix this, though maybe Lex or Daniel
> > have some good ideas.  Our present mechanism of editing and sending
> > mails with attachments also seems a little clumsy to me, though this may
> > not be excuse enough to change that yet.
> I think your diagnosis is right - our current MIME handling is ugly. I
> say this plainly - I put it there because I needed it to send the
> attachments for the work I really wanted to do. So the code looks like
> it does...
> 
> I dont have a good idea on how to fix this particular bug. I have an
> outline of how I think MIME should look in general, fixing this among
> other things. Such a refactoring would be significant work, but it would
> squash many outstanding little annoyances. In brief - MIME is a
> composite object, and printing it should be a Visitor aware of its
> nature. Specifically for this bug, asSendableText should know when its
> handling an attachment, that though it is past the main texts header, it
> is now in the attachments header, and thus should go back to
> look-but-dont-wrap behavior until another blank line is found. This can
> be hacked into asSendableText, but in truth really wants to be part of a
> major refactoring...
> 

Well, we do have MailMessage to rally the design around.  It's not as
bad as it might seem!  The real problem, it seems to me, is that the
system so quickly degenerates back to manipulating raw text.  eg,
#queueMessageWithText: instead of #queueMessage:.  I think the
composition window even discards the MailMessage entirely, in favor
of the expanded text version.

Incidentally, #asSendableText seems like the wrong place to adjust line
endings, anyway.  IMHO, #asSendableText should encode
the message as is--sometimes you really want long line endings!



-Lex





More information about the Squeak-dev mailing list