In Kevin Johnson's book "Internet Email Protocols, A developer's Guide", 
chapter 6 paragraph 6.3.3

"A Client can sometimes send an incomplete command.  When the server 
receives such a line, it sends a command continuation request response to 
the client, indicating it is ready to accept the remainder.  These 
responses are used primarily with fields sent as literals, which are 
described in setion 6.4.4 and authentication dialogs.  When a command 
continuation response is called for, the server sends a single line 
starting with a "+" character followed by a space character and optional 
human readable text."  i.e. my "+ Ready".

If you refer http://www.faqs.org/rfcs/rfc2060.html section 7.5 Command 
Continuation Request you will see it is part of the spec and obviously the 
clients I referred to know how to deal with it.

My error was looking for the end of the command instead of the size of the 
literal in the literal continuation line ending {1234}, which I am now using.

Ollie

At 02:02 PM 7/1/2002 -0400, you wrote:
>At 10:54 AM 7/1/2002 -0700, you wrote:
>>Oh yes the behavior I am seeing is definitely being chunked by the 
>>clients as several literals, perhaps by my sending the initial "+ Ready" 
>>that triggers the behavior, but all three major clients exhibit the 
>>behavior.  I have already re-written that terminator logic to count bytes 
>>loaded and save the excess as a pending command and have started testing.
>
>Sure there are clients that abuse the protocol but behavior such as this 
>just does not make sense and could not be tolerated.  Let me have a look 
>again at how these clients behave with APPENDs.
>
>Pete
>
>

Michael Oliver
Chief Technology Officer
AppsAsPeers.com
7391 S. Bullrider Ave.
Tucson, AZ 85747
520.574.1150

Reply via email to