[pyicu-dev] Why SWIG?

[Jim Fulton]

Andi Vajda wrote:
> 
...

> No, I don't know Pyrex, I barely know about it.
> Would Pyrex make it as trivial to wrap all this C++ code ?

No. Is that a goal?  SWIG makes it trivial to create really bad
Python bindings to things.  With a bit of work -- meaning fancy
.i files, you can begin to make nicer interfaces.  The nicer
the Python api, the bigger the .i file.  I don't really think
that non-trivial .i files are maintainable. I went through this
a few years ago with DCOracle.  (The first version of DCOracle
was SWIG based.)

...

> A few months ago, someone recommended I take a look at SIP (in the 
> context of PyLucene): http://www.river-bank.demon.co.uk/docs/sip/
> I haven't looked at this too closely yet either.

Ah.  I'm not a C++ hacker, but I have the impression that there are
some really nice options for C++ wrapping.  I'm not really familiar with
them.  It's possible that one of the clever C++ binding tools could be
better than Pyrex.

...

> Could you sum up the pros and cons of Pyrex over SWIG ? (educate me)

- No pointers. :)

- With Pyrex, you write C extensions using an extended form of Python.
   It knows about Python and can automate a lot of book keeping that you
   would have to yourself with SWIG.

Honestly, I've never used Pyrex myself (I haven't needed to wrap anything
in a while), but I've heard great things about it.  Martijn Faassen is using
it with great success to wrap libxml/libxslt:

   http://codespeak.net/lxml/

Jim

-- 
Jim Fulton           mailto:jim at zope.com       Python Powered!
CTO                  (540) 361-1714            http://www.python.org
Zope Corporation     http://www.zope.com       http://www.zope.org