[Pyrex] creating and populating Python/C structs

[Christopher Armstrong]

On 9/5/05, Simon Burton <simon at arrowtheory.com> wrote:
> On Sun, 4 Sep 2005 16:44:08 +1000
> Christopher Armstrong <radeex at gmail.com> wrote:
> 
> >
> > Anyway, one issue is that I'm trying to avoid PyFrame_New, which gets
> > a bunch of stuff from the context (e.g. the thread state you're
> > supposed to pass it) and I want to avoid that.
> 
> oh, i see. Why ?

For the long story, see 

http://mail.python.org/pipermail/python-dev/2005-September/056091.html

which is a post of mine to the python-dev mailing list. I'm trying to
enable asynchronous use of traceback objects, and then raise them in
the future, meaning I have to be able to hold them around for a while.
The problem is that traceback objects, via the frames they hold, refer
to tons of objects that I don't want to hold on to, which really
screws up GC. So I'm trying to construct traceback objects that look
enough like real traceback objects and have similar information to the
originals. Which means I need to do the same for frames, since
traceback-printing machinery really gets into the guts of frame
objects.

-- 
  Twisted   |  Christopher Armstrong: International Man of Twistery
   Radix    |    -- http://radix.twistedmatrix.com
            |  Release Manager, Twisted Project
  \\\V///   |    -- http://twistedmatrix.com
   |o O|    |  
w----v----w-+