[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Problems with 0.3.2



Steinar Bang <sb@metis.no> writes:

> > - the server says NO and complains about \n's in the article body
> > - the server seem to wait for more input
> > - the server says BAD and complains about junk after a literal (the article)
> 
> > I don't recall which one I got when I used your code unchanged.
> 
> > Seems like there is some problem with counting the number of
> > characters in the article. I've been tracing this with a protocol
> > sniffer to find about exactly what is being sent, and it shouldn't be
> > that hard to figure out why it isn't working. I'll try it with XEmacs
> > to see how it should work.
> 
> Is it maybe that
>  1. the server is expecting \r \n while Gnus is sending \n only?

The server shouldn't care what I send -- but looking at the cyrus
source I can tell the my first error is caused by sending \n without
\r before it.  I really think this is invalid -- according to the IMAP
spec servers should accept 8 bit data (and MIME encode it if it needs
to).

>  2. the size of the article is calculated in a buffer holding \n only
>     while \r \n sequences are sent over the IMAP connection and the
>     numbers don't match?

It's something like this anyway. ;-)

> Maybe nnimap on a Win32 machine will work better than nnimap on
> UNIX...? ;-)

Yes. And would also explain why it works in XEmacs (no MULE).

ELMOs append command does it differently (with strings, not buffers),
but it's defuned but never used so I can't tell if it works (haven't
tried it yet).

/S