On Fri, 1 Dec 2006, Mike Taylor wrote:
I think this is a good time to bring up something I've wondered about.
Why do we always pull down our own version of python?
Originally we built our own python to work around a number of bugs and to
also remove the stock Berkeley DB bindings, which IIRC did not work with the
version of Berkeley DB we required.
Currently, the set of patches to python 2.4.3 we apply are:
- remove bsddb completely because:
. bsddb is usually out of date with Berkeley DB
. chandlerdb has its own python/bdb bindings
- remove the code looking for extensions in the windows registry, probably
our oldest patch (needed because we have our own wxPython, I believe)
- some Intel OS X building patches (probably not needed in newer versions)
- some patches to include support for readline on Mac OS X
Except for the Intel OS X build patches, I expect the same patches to be
applied to Python 2.5.
Once we migrated away from Python 2.3 we started also running into the
situation where the system Python was either non-existent or the wrong
version. This continues to be a problem for us today.
That problem is not specific to 2.3, it's always been the case that the stock
Python on Mac OS X is one version behind and shipped without readline support.
It's also always been the case that Windows ships without Python and that the
Windows Python we could use uses the registry (and comes without readline
support, which still remains to be fixed)
As for the question about upgrading to Python 2.5, I concur with the general
consensus: let's do it and not have conditional code to still support 2.4.
Now is the time to do this upgrade.
Bear, if I can help, let me know...
Andi..
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev
