RE: [Exim] quota_warn_threshold specifics

Christian Balzer wrote:
> Eli wrote:
>> Christian Balzer wrote:
> [snip]
>> The time on the "cur" folder would be another folder to check
>> since it may be possible the user didn't get any new messages, but
>> just took a read message and marked it for deletion or changed some
>> other type of flag.
>>
> What makes "cur" so much more worthy of attention than any other
> folder? The user could have moved/flagged/whatever thousands of mails
> in hundreds of other folders w/o you ever noticing it with this
> scheme. The crux of the matter is, as long as no new mail was
> delivered/read and the mailbox is full, all this potential activity
> is worth naught.

Just thought I'd mention it - I figured if you're going to pay attention to
the new folder, it wouldn't cost much more to peek at the cur folder to see
if it's time is newer or not. I thought the idea was to try and get the
time the user last checked any email at all, so I figure any type of mail
activity would count.

> Another thing is, one really wants to check for modify time and not
> access time as I can see many potential read accesses (something
> building indexes, etc) happening that are not client/user initiated.
> But an actual write (in new) should originate only from Exim or the
> pop3/imap server.

Whichever :) My brain isn't really thinking hard about these issues right
now - as long as you get the idea then it's all good (hopefully) :) I was
just giving you some comments about it since you asked. Figured maybe I'd
stir something up if you hadn't of thought of it, but I guess you're
covering all the bases :D

Eli.