I seem to remember at least one person posting some replacement code to try to clear up the duplicates problem...anyone know if it works yet? I need to go dig that code out (assuming I'm right in remembering it)
Corbett J. Klempay Quote of the Week: http://www2.acm.jhu.edu/~cklempay "There are two things a real man likes - danger and play. And he likes women because she is the most dangerous of playthings."
PGP Fingerprint: 7DA2 DB6E 7F5E 8973 A8E7 347B 2429 7728 76C2 BEA1
You know, I thought I saw such a thing too, but now I can't find it. Has anyone fixed this problem yet, and if so, can you make patches? If not, it's the first thing I'll go after. I think people have given enough information to find the problem pretty easily.
John
On Fri, Aug 21, 1998 at 10:41:14AM -0400, Corbett J. Klempay wrote:
I seem to remember at least one person posting some replacement code to try to clear up the duplicates problem...anyone know if it works yet? I need to go dig that code out (assuming I'm right in remembering it)
Corbett J. Klempay Quote of the Week: http://www2.acm.jhu.edu/~cklempay "There are two things a real man likes - danger and play. And he likes women because she is the most dangerous of playthings."
PGP Fingerprint: 7DA2 DB6E 7F5E 8973 A8E7 347B 2429 7728 76C2 BEA1
Mailman-Developers maillist - Mailman-Developers@python.org http://www.python.org/mailman/listinfo/mailman-developers
On Wed, 26 Aug 1998, John Viega wrote:
You know, I thought I saw such a thing too, but now I can't find it. Has anyone fixed this problem yet, and if so, can you make patches? If not, it's the first thing I'll go after. I think people have given enough information to find the problem pretty easily.
John, i think this is related to the linux gid problems that i identified in relation to corbett's duplicates problems. In brief:
On linux systems, scripts forked by the object executables (eg, the wrapper executables) do inherit the effective gid that the wrappers wants them to get - but then, scripts forked by *those* scripts do not!
This permissions problem, combined with the initial insensitivity of Utils.TrySMTPDelivery (i think it was) was causing things to be left on the queue despite successful delivery, and then the queue processing script - launched from the cron job with the right euid - would do a redundant delivery and then delete.
I think that was the upshot - but check the messages (with apparent fixes) that i posted about it, they should give the full scoop. (Note that part of the fix for linux systems was to open up the permissions on the queue directory - not an optimal solution.)
http://www.python.org/pipermail/1998q3.mailman-developers/989f2234d905.html http://www.python.org/pipermail/1998q3.mailman-developers/c1aa27e70661.html
Sorry i don't have more time to dedicate to this stuff at the moment - hopefully soon...
ken klm@python.org
participants (3)
-
Corbett J. Klempay -
John Viega -
Ken Manheimer