bpot (Bob Potter)
- Login: bpot
- Email: bobby.potter@gmail.com
- Registered on: 01/29/2010
- Last sign in: 04/17/2013
Issues
| open | closed | Total | |
|---|---|---|---|
| Assigned issues | 0 | 0 | 0 |
| Reported issues | 0 | 2 | 2 |
Activity
04/16/2013
-
06:29 AM Ruby Bug #7584: Ruby hangs when shutting down an ssl connection in gc finalization
- Looks like that fixes it! I was unable to reproduce with "ruby 2.1.0dev (2013-04-15 trunk 40308) [x86_64-linux]".
03/27/2013
-
- It looks like the response is invalid at least according the RFC3501. The additional fields that the parser is blowing up on should only be included if the type is 'message/rfc822', in this case it is 'message/delivery-status'. It looks ...
03/23/2013
-
- It looks like it is a race condition where the TCPSocket finalizer closes a fd, that fd is reused by another thread and then we try to do SSL_shutdown on this new connection and bad things happen.
Would it make sense to have the TCPSo... -
- =begin
Digging into this some more because it cropped up in another project.
I've had some luck reproducing this issue with a small script: https://gist.github.com/bpot/5225920
Sometimes it reproduces easily, sometimes it doesn't....
02/19/2013
-
- For what it's worth, we were able to work around this by explicitly shutting down ssl connections instead of letting it happen during GC.
12/18/2012
-
- It appears that during garbage collection ruby will sometimes attempt to shutdown an ssl connection that has already been shutdown. At least in certain cases this causes the SSL_shutdown call to block forever.
(gdb) bt
#0 0xb76fa424...
01/29/2010
-
- =begin
=end
-
- =begin
The main imap thread will hang if the receiving thread receives an unexpected "BYE" message during login. IMAP#receive_responses (lines 1054-7) sets the exception correctly but fails to break out of the main reading loop because...