Restarting my IMAP server caused the following Sup exception. --- SystemExit from thread: main closed stream /usr/lib/ruby/1.8/openssl/buffering.rb:237:in `select'' ./lib/sup/buffer.rb:31:in `nonblocking_getch'' bin/sup:227 Sup should really handle this more gracefully. - Kevin -- Kevin Riggle (kevinr at mit.edu) http://free-dissociation.com
Reformatted excerpts from Kevin Riggle''s message of 2008-05-13:> --- SystemExit from thread: main > closed stream > /usr/lib/ruby/1.8/openssl/buffering.rb:237:in `select'' > ./lib/sup/buffer.rb:31:in `nonblocking_getch'' > bin/sup:227This backtrace doesn''t really make any sense. There''s no reason that nonblocking_getch would be calling the openssl stuff, and openssl''s buffering.rb doesn''t mention select at all. Weird. -- William <wmorgan-sup at masanjin.net>
On 20.5.2008, William Morgan wrote:> > closed stream > > /usr/lib/ruby/1.8/openssl/buffering.rb:237:in `select'' > > ./lib/sup/buffer.rb:31:in `nonblocking_getch'' > > bin/sup:227 > > This backtrace doesn''t really make any sense. There''s no reason that > nonblocking_getch would be calling the openssl stuff, and openssl''s > buffering.rb doesn''t mention select at all. Weird.I should pipe up here as well - this exception was one of the reasons I switched to offlineimap as I would get it randomly in sup during imap actions. I couldnt figure out how they interacted with each other though as it didnt happen enough to debug :( Marcus
Reformatted excerpts from Marcus Williams''s message of 2008-05-20:> I should pipe up here as well - this exception was one of the reasons > I switched to offlineimap as I would get it randomly in sup during > imap actions. I couldnt figure out how they interacted with each other > though as it didnt happen enough to debug :(Interesting. I''ve had weird issues in the past with the Ruby IMAP library and threading, so this might be related. I''ve definitely seen it generate nonsensical backtraces by funny exception handling across threads. Simultaneous dependencies on 20 libaries of dubious quality---that''s the story of Sup. Well once I get up the energy to write SupServer, I won''t have to support anything except for mbox anymore! -- William <wmorgan-sup at masanjin.net>
On Tue, May 20, 2008 at 12:05 PM, William Morgan <wmorgan-sup at masanjin.net> wrote:> Interesting. I''ve had weird issues in the past with the Ruby IMAP > library and threading, so this might be related. I''ve definitely seen it > generate nonsensical backtraces by funny exception handling across > threads.I think you''ll find that things get better if you do plain IMAP with stunnel instead of relying on the ssl support in ruby/ruby-imap. I had horrible (to the point where I almost had to switch away from ruby) problems with a project that relied on the ability to do imaps. The imap library may be partially to blame (and the imap spec requiring threading doesn''t help...), but I found that things got much better without ssl being handled by ruby. -Ben -- --------------------------------------------------------------------------------------------------------------------------- Ben Walton <bdwalton at gmail.com> When one person suffers from a delusion, it is called insanity. When many people suffer from a delusion it is called Religion. Robert M. Pirsig, Zen and the Art of Motorcycle Maintenance ---------------------------------------------------------------------------------------------------------------------------
Reformatted excerpts from Ben Walton''s message of 2008-05-20:> I think you''ll find that things get better if you do plain IMAP with > stunnel instead of relying on the ssl support in ruby/ruby-imap. I > had horrible (to the point where I almost had to switch away from > ruby) problems with a project that relied on the ability to do imaps. > The imap library may be partially to blame (and the imap spec > requiring threading doesn''t help...), but I found that things got much > better without ssl being handled by ruby.Interesting. I didn''t know about stunnel. Kevin, if you''re seeing this on a regular basis, you could try that... or switch to offlineimap like everyone else has for performance reasons. Sup''s IMAP support leaves a lot to be desired and I''m happy to blame this library. Irritatingly slow, and prone to insane errors like this one. I don''t want to disable it, though, and I sure as heck don''t want to write my own IMAP library. So for the time being we''re stuck with it. -- William <wmorgan-sup at masanjin.net>