At 17:20 3/02/03 -0800, Randall Gellens wrote:
>At 5:43 PM -0500 2/3/03, II Alan W. Rateliff wrote:
>
>>   > RFC 3206 provides an unambiguous means for POP3 servers to inform
>>>  clients if an error response is due to a credentials problem or
>>>  something else, allowing clients to not assume a password error is
>>>  the cause of all errors during authentication.  I'm not sure how many
>>>  clients and servers support it yet, but recent versions of Qpopper do.
>>
>>  Outlook/Express has a habit of asking for a different username and
password
>>  initially, regardless of the actual error.  I believe this was intended to
>>  provide the user with the opportunity to use a different username, but I
>>  don't fully see the reasoning.
>>
>>  Only after failing to authenticate a POP session a couple or few times
does
>>  it actually give the verbose response from the mailserver.
>
>The idea of the extended POP response codes in general, and specific 
>ones such as [SYS] and [AUTH], is to allow the server to clearly 
>indicate to the client the nature of the error.  So, if a user tries 
>to authenticate while a session is currently active,   RFC 2449 
>provides the [IN-USE] response code.  When the client sees "-NO 
>[IN-USE]" it knows that the user has another session active.  It can 
>silently retry later, it can present a localized explanatory message, 
>or whatever else makes sense.
>
>In case of an actual credential-related error, RFC 3206 provides the 
>[AUTH] response code.  If some resource is temporarily unavailable, 
>RFC 3206 provides the [SYS/TEMP] code.  Etc.
>
>So asking for a different username and password, without paying any 
>attention to the response code, doesn't make a lot of sense to me.

Unfortunately this is *exactly* what Outlook Express does :-(

It just presents a username and password dialogue with no explanations,
despite the [IN-USE] error. Eudora on the other hand, displays the message
from the server...

Regards,
Simon


Reply via email to