[IPython-user] Re: [IPython-dev] A plea for help with a threading bug...

[Fernando Perez]

Fernando Perez wrote:

> down to readline.  In particular, I find this to be noteworthy from the 2.3.5 
> release notes at http://www.python.org/2.3.5/NEWS.html:
> 
> Fixed broken threadstate swap in readline that could cause fatal errors when a 
> readline hook was being invoked while a background thread was active. (SF bugs 
> #660476 and #513033.)
> 
> This sounds remarkably similar to our situation.  If this is the case, then we 
> have a closed case.  Yariv, do you have an easy way to test 2.3.5?  I can do 
> it, but probably not today.

OOPS, sorry for the noise.  I misread the above, and that was an entry from 
early on in the 2.3 series, not specific to 2.3.5.  And since Yariv tested 
gentoo with 2.4, we probably still have a readline bug.  Sorry for the 
confusion...

Yariv, I saw your last entry at

http://www.scipy.net/roundup/ipython/issue30

Perhaps you could file these as bugs against readline at the SF tracker? 
Since you want a small self-contained test case, I'm attaching a little gtk 
threaded interactive console which I wrote a while back, based on someone 
else's code, for matplotlib.  You can rip out the matplotlib-specific code, 
but this is small enough that it should be possible to use it to cause the 
crash in a self-contained way, without needing ipython at all.

Best,

f
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pyint-gtk.py
Type: text/x-python
Size: 7879 bytes
Desc: not available
Url : http://projects.scipy.org/pipermail/ipython-user/attachments/20050319/33a2aeb0/attachment.py