[Dovecot] sig11 with test36/37 on some emails
Sometimes I'll get an email from a luser which will cause the POP3 daemon to sig11. The email "appears" to be legal. The sig11s only happen on specific emails, and once I remove the emails from the server, the rest can be downloaded. I gzipped the email, and its at http://theapt.org/bogusemail.gz. While this example is international spam, the sig11s happen with real emails too.
-- Entropy isn't what it used to be.
On Mon, 30 Aug 2004 10:48:29 +0200 (CEST) Johannes Berg <johannes@sipsolutions.net> wrote:
:> and its at http://theapt.org/bogusemail.gz : :403 Forbidden. : :johannes :
Err, sorry about that. Fixed.
-- Every nonzero finite dimensional inner product space has an orthonormal basis.
It makes sense, when you don't think about it.
On Sun, 2004-08-29 at 18:58 -0700, Peter Hessler wrote:
Sometimes I'll get an email from a luser which will cause the POP3 daemon to sig11. The email "appears" to be legal. The sig11s only happen on specific emails, and once I remove the emails from the server, the rest can be downloaded. I gzipped the email, and its at http://theapt.org/bogusemail.gz. While this example is international spam, the sig11s happen with real emails too.
Thanks. Also because of this I found a bug in output stream code which sometimes made it send wrong data. This bug has been there forever, strange that it hasn't caused much problems.
On Tue, 31 Aug 2004 12:16:03 +0300 Timo Sirainen <tss@iki.fi> wrote:
:On Sun, 2004-08-29 at 18:58 -0700, Peter Hessler wrote: :> Sometimes I'll get an email from a luser which will cause the POP3 daemon :to> sig11. The email "appears" to be legal. The sig11s only happen on :specific> emails, and once I remove the emails from the server, the rest :can be> downloaded. I gzipped the email, and its at :> http://theapt.org/bogusemail.gz. While this example is international :spam,> the sig11s happen with real emails too. : :Thanks. Also because of this I found a bug in output stream code which :sometimes made it send wrong data. This bug has been there forever, :strange that it hasn't caused much problems. : :
I can verifiy that you fixed it. Thanks.
-- Q: How many IBM cpu's does it take to do a logical right shift? A: 33. 1 to hold the bits and 32 to push the register.
participants (3)
-
Johannes Berg
-
Peter Hessler
-
Timo Sirainen