mailing list archives

Hi,

Can an IMAP server send a BAD before the client finished sending the command?

The attached ethereal capture contains this example:

C: "zxc select inbox "
S: "zxc BAD Unexpected extra arguments to Select\r\n"
C: "\r\n"                         /* client finshed sending zxc */
C: "vbn copy a"
C: " "
S: "vbn BAD Invalid sequence in Copy\r\n"
C: "\r\n"
C: "qwe dhsads "
S: "qwe BAD Unrecognized command\r\n"
C: "gfhg"
C: "\r\n"     /* client finished sending qwe */


Thanks,

Adrian Buciuman
-------------- next part --------------
A non-text attachment was scrubbed...
Name: imap-ethereal
Type: application/octet-stream
Size: 1543 bytes
Desc: not available
URL: <http://mailman13.u.washington.edu/pipermail/imap-protocol/attachments/20060404/6a95a722/attachment.obj>

Adrian Buciuman writes:
> Can an IMAP server send a BAD before the client finished sending the command?

Yes.

In principle, a BAD results when there _isn't_ a recognizable command. 
You can't ask the server to wait until the end of something it doesn't 
recognize.

Arnt

Adrian Buciuman wrote:
> Can an IMAP server send a BAD before the client finished sending the command?

For what it's worth, Cyrus has never bothered to wait for CRLF before 
sending command responses in a failure case like this.  It's never been 
a problem (except, occasionally, when telnetting to the server from a 
client that doesn't support tty-style line editing).

Tim

On 4/4/06, Arnt Gulbrandsen <arnt@gulbrandsen.priv.no> wrote:
> You can't ask the server to wait until the end of something it doesn't
> recognize.

Why not? I belive it can wait until a LF is received from the client.

On Tue, 4 Apr 2006, Adrian Buciuman wrote:
> > On 4/4/06, Arnt Gulbrandsen <arnt@gulbrandsen.priv.no> wrote:
> > You can't ask the server to wait until the end of something it doesn't
> > recognize.
> Why not? I belive it can wait until a LF is received from the client.

Nothing in the IMAP protocol requires the server to do so.  Most servers 
probably do, especially if they implement the LITERAL+ extension.

However, your client should not care.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.

Adrian Buciuman writes:
> On 4/4/06, Arnt Gulbrandsen <arnt@gulbrandsen.priv.no> wrote:
>>  You can't ask the server to wait until the end of something it 
>>  doesn't recognize.
>
> Why not? I belive it can wait until a LF is received from the client.

(CRLF ends a line, not a command. A command can be much longer than one line.)

There's nothing to be gained (as far as I can imagine) from requiring 
the server to do that. It wouldn't help anyone access their mail 
better/faster/etc. Therefore, it would be a bad requirement.

Arnt