Re: Misc minor POP3 issues

[Michael D'Errico <[e-mail]>]

This was one of the original messages sent out to this list.  Sorry
about taking so long to respond to it.

>Mark Crispin <[e-mail]> writes:
>>  7) Is it permitted for a POP3 server to have an inactivity autologout timer?
>>     If so, how is it supposed to implement this?

This is an issue related to the broken TCP connection argument we've
been having.  I implement a timeout which defaults to 10 min. and is
configurable.  If the server times out, it sends a "-ERR" message and
closes the connection.  The UPDATE state is not entered, and no messages
are deleted.  Does this seem like a reasonable thing to do?

>John Gardiner Myers <[e-mail]> writes:
>> The syntax of the responses to the STAT and LIST commands are
>> under-specified.  Is, for example "+OK    6     13409" a valid
>> response to the STAT command?

I would go back to the old "be conservative in what you generate"
principle on this one.  The usage examples in the spec. are clear
in that there is exactly one space between the command and its
arguments, and responses also have exactly one space between the
+OK and the data.  This should probably be more exactly stated
(using BNF?) to reduce the ambiguity.

>> A comment that the burst/parse mailbox wording is an implementation
>> detail specific to some subset of possible maildrop formats.  Since
>> the maildrop format that my server is going to read has never pasted
>> messages together in the first place, the burst/parse requirement is
>> nonsensical.  I suggest this wording be removed.

Me too.

>> Wording nitpicks:  
>>
>> >   Commands in the POP3 consist of a keyword possibly followed by an
>> >   argument.
>>
>> The optional commands are each followed by *two* arguments.
>> 
>> >   Keywords are 4 characters long.
>> 
>> The TOP keyword is 3 characters long.

Why not deprecate TOP and replace it with the HEAD command (for "headers")
which would just return the message headers?  (Is there really any use
for the first few lines of the body?)  Of course a server could still
support the TOP command to maintain backward compatibility.  The HEAD
command would take only one argument, so the only one that needs 2 or
more would be XTND.

Michael D'Errico
[e-mail]