Re: The performance issue of the email package, and how about a cEmail?
-----BEGIN PGP SIGNED MESSAGE-----
On Aug 18, 2007, at 6:11 AM, Iceberg wrote:
> It seems the only way to boost up the performance is to rewrite the
> key part (mostly FeedParser.py) in C. And I believe the effort will
> be worthy for such a fundamental, widespread, core package. In
> fact, we saw similar happen, such as: cString, cPickle, cProfile.
> I read old mail archive (http://mail.python.org/pipermail/email- > sig/) since 2005, but found no thread on this topic. So, I would
> venture to ask, is there any plan for a cEmail package in near future?
There's no plans by me to do this, but if yo're interested, I think
it could be a worth goal. Without looking at those existing
packages, there's two things I'd say. I doubt that either package
would be included in Python by default, either because it's C++ or
because of a license incompatibility. OTOH, it may or may not be
worth enabling optional building of a cFeedParser based on whether
these packages are available or not.
OTOH, it might be nice to provide something like a cFeedParser as a
third-party egg, and if it works out, and is enough of a performance
boost, I'd probably support extending the email package to use it if