On Tue, 2004-08-10 at 23:50 +0530, Sanjay Arora wrote: > On Tue, 2004-08-10 at 23:27, guenther wrote: > > Note: Your Reply-To header seems to be borked. > Fixed Had a dot after the id...dont know how the mailbox was working ;-) > > > > > BTW: Evolution 1.2.x is pretty old (like years). > > Planned to upgrade alongwith the OS to Fedora or Whitebox. I didnt know > that a rpm for RH9 was available...guess just assumed it wont be > supported....same as RH ;-) My mistake. Will definitely upgrade. Thanks. > > > > > > > Apart from that there are other possibilities to delete duplicate mails > > not bound to Evolution. 'formail' for example can be used to delete > > duplicates (based on Message-ID headers, which are not guaranteed to be > > unique). > > > Now this is NEWS to me....I thought that message IDs were guaranteed > unique...as each server generated its own, with random factors thrown in > and never gererated same ids. Will anyone shed some light?
that's like guarenteeing that there will only be a singler person named Jeffrey Stedfast, which I can assure you is not so. My Evil Twin is a Technology lawyer in Virginia :-) even md5 sums aren't guarenteed to be unique. the uniqueness is only "guarenteed" by the host generating them, but that guarentee is "assuming no bugs or malicious behaviour". also note that many users don't properly configure their machine hostnames (and/or don't configure them with a unique hostname) and so the guarentee goes to hell. there's nothing stopping me from maliciously faking your message's Message-Id and delivering it to the same recipient you sent it to in attempt to make auto-dup-deletion code to delete your message and thus disallowing the recipient to ever see your mail. > > Will google for formail. Is it something like procmail? > > > I posted a solution (working, with some caveats) a couple of months ago > > to this list. If you are interested in that (command line only) script, > > I can search for it and post the links again. > > > I tried searching the archive, to save you the work but could not find > the link....Please do let me have the link...so I can try it out. it's been discussed numerous times. the first time I remember it being discussed was when Jamie Zawinski brought it up back in May 2000, pointing out that you cannot rely simply on Message-Id to decide uniqueness. Jeff -- Jeffrey Stedfast Evolution Hacker - Novell, Inc. [EMAIL PROTECTED] - www.novell.com
smime.p7s
Description: S/MIME cryptographic signature
