[mod_python] Re: reading mod_ssl variables from mod_python

On 2/15/06, Graham Dumpleton <grahamd at dscpl.com.au> wrote:
> The _mp_mod_ssl module existed long before your suggested patch.
> It was a proof of concept to show that the sort of SSL stuff could be
> done without even needing to modify mod_python.
>
> I refered you to it when you first talked about your patches:
>
>   http://www.modpython.org/pipermail/mod_python/2005-November/019609.html
>
> I take it you must not have looked at it. :-(

Oh, sorry, I didn't recognize it by its name.  Now I'm clearer :)

I'm wondering though if Tomasz's problems are specific to that, or if
they might also occur with the newer patch as well.  Or if it's
nothing to do with mod_python at all, and really is a mod_ssl
configuration issue.

Tomasz...I don't know how flexible you can be, but may it be possible
for you to try  using mod_python 3.2.7 with the proposed patch in
MODPYTHON-94?  It would really be valuable to get somebody testing on
that before it goes into an official release.  Plus it may even be
worthwhile to compare _mp_mod_ssl with the new patch in your case,
just to see if they both exhibit the same behavior or not.

[Yes, I'm using the new patch + 3.2.7 today in a production system
with no problems.  But I don't use client certs very much, and I'm
mostly hooking in the AccessHandler phase instead of Authen* phases. 
Also I'm just one test case.]
--
Deron Meranda