On Thu, 31 Oct 2002, Alan W. Rateliff, II wrote: > Date: Thu, 31 Oct 2002 13:06:23 -0500 > From: "Alan W. Rateliff, II" <[EMAIL PROTECTED]> > To: Subscribers of Qpopper <[EMAIL PROTECTED]> > Subject: Re: Users popping to often > > > Simon, > > > > For what it is worth, attached is my patchfile for my own 3.x era > > work on qpopper. I had a ton of local customizations in version 3 of > > qpopper, most of which I dropped when I went to version 4 (I embraced > > the server mode option -S). One of the hacks I had in version three > > was the "accesstime" option, setable via command line as "-a". Search > > the patchfile for the word "accesstime". This feature required you to > > use KEEP_TEMP_DROP. What the pop_accesstime() subroutine did was to > > compare the timestamp of the dropfile to the current time. If the > > difference was less than accesstime, then the user got the POP error > > message "you just checked your mail!" and rejected the connection without > > touching the mailbox. This bit of code punished impatient users who > > wanted to check their email every few seconds. > > I understand that Unix filesystems maintain several timestamps on files, > such as created time, accessed time, and modified time. >
Without looking at my patchfile, I probably wrote the code to look at access time. > If I understand correctly: > > * POP'ing a mailbox will modify the created time, since the original MBOX is > dismantled to the temp drop, then rebuilt after the session. > > * Sendmail (or other MTA) delivering email will modify the modified time > > Would it be possible to simply check the current time against the accessed > or created time so temp drops don't have to be left in place? Also, > wouldn't limiting POP time like that cause problems with POP-based webmail > clients? Personally, I'm converting my POP-web client to IMAP to eliminate > that issue. (OT: Anyone have an easy HOWTO on configuring Horde/IMP?) Looking at the mailbox times won't work because many things can modify access times besides POP. The temp dropfile is POP's alone. I use the KEEP_TEMP_DROP anyway because I want a record of the last time somebody used POP (one way of tracking down stale accounts). > > Oh, which brings up another question, probably asked before... what happens > if a mailbox gets POP'd during an active IMAP session? I've considered > patching the IMAPd source to create a psuedo pop.lock file so QPopper will > not allow POP'ing during an open IMAP session, but haven't actually done so > yet. I wanted to get some input first. > I too have this issue in my future. IMAP and POP at the same time will be a near-certain trainwreck, trashing a person's mailbox. At my site, people can POP their mailboxes or login to a UNIX box and use pine directly on their mailbox (not via IMAP). People who invoke a pine session while having POP running generally get thumped. My user community has learned to avoid this. --- Jeff
