new gc problems and seg fault

[GoogleCodeExporter]

I upgraded to 0.5.3 earlier and ran a program that has worked fine in 0.5.0
on it.  My program opens a set of sockets to predefined destinations A,B,C,D.
It had worked fine before, but now it seems I have a problem with
dropped/garbage-collected connections.  It will connect to A,B,C,D, but
before the connection is established, B,C,D have been garbage collected and
only A has a successful connection.  When I repeat the initialization
procedure, it will try connect to B,C,D, but only B will be successful as
C/D are garbage collected, and so on.  Soon after, I get a segmentation
fault.  I reloaded 0.5.0-0.3.8 and all the problems go away.

Hope this helps in resolving these issues.

Original issue reported on code.google.com by ski...@gmail.com on 20 Dec 2009 at 11:33

[GoogleCodeExporter]

Do you keep a reference to the watchers (they are automatically stopped when 
they are
garbage-collected)?
Can you provide an example script to reproduce this behavior (especially the 
segfault)?

thanks for the report

malek

Original comment by lekma...@gmail.com on 21 Dec 2009 at 5:17

• Added labels: Priority-High
• Removed labels: Priority-Medium

[GoogleCodeExporter]

I'd be really interested in anything leading to the segfault, but I can't 
reproduce this.

Can you be more specific? Why are the connections garbage collected? Do you 
setup an
Io watcher before the actual connection? If yes do you watch out for 
EWOULDBLOCK or
EAGAIN (are the sockets in non-blocking mode?)? Do you send something through 
the
socket? Python 2.6/3.1?

Even if you fixed it in the meantime, I'd like to know what was the initial 
problem
(that is if you can expand on it).

Thanks for your help.

Original comment by lekma...@gmail.com on 5 Jan 2010 at 11:00

[GoogleCodeExporter]

Original comment by lekma...@gmail.com on 20 Jul 2010 at 9:02

• Changed state: Invalid