Lennart Regebro wrote:
2009/10/5 K. Richard Pixley <r...@noir.com>:
This would be a problem if distribute were in general release. It's not.
It's clearly a development branch which is intended to move quickly.
No, this is incorrect. The 0.6-branch is not intended to move quickly,
it is in bugfix mode.
It is moving quickly only because some major bugfixes has been made,
but it's not a development branch, that's 0.7, which isn't released.
I'm recent to python packaging and distribution, so let me see if I've
put this together right from my reading of the various web pages
involved over the weekend.
Distutils is currently part of the standard python library. As such,
it's released with python, (the reference implementation, anyway).
Distutils is currently capable of producing only source archives of
packages. While it's capable of producing "built" archives, those
archives are machine specific, nonrelocatable, untrackable, and have no
standard method for distribution nor installation nor tracking.
Setuptools was a third party addition to, (and partial replacement of),
distutils because distutils wasn't suitably usable nor was it moving
fast enough. However, since setuptools was initiated, many of the major
features of setuptools have since been folded back into distutils,
making setuptools partially redundant and partly colliding. Setuptools
provides the ability to produce machine independent "built" archives and
a standard method for installing them, (although not for tracking or
removing them). And the setuptools approach to installation,
easy_install, doesn't play nice with the native installers on systems
that have them like rpm, debian, etc.
However, setuptools has fallen into disrepair and so distribute has been
created, as a friendly branch off a third party tool which in turn was a
form of branch off of distutils. And within distribute, there are two
lines of development, the 0.7 line, which is intended to replace... I'm
confused. Does it replace setuptools or distutils? And then there's
the 0.6 branch, which is a branch off 0.7 which is a branch of
setuptools which is a branch of distutils which is under recent active
development and yet it also expected to be stable, as much as such a
term can be applied to a third party branch off a third party of a
colliding replacement with a standard facility.
Is that about right?
If I'm anywhere near right, then I can't really imagine what state you
intend for the 0.6 branch if not development.
--rich
_______________________________________________
Distutils-SIG maillist - Distutils-SIG@python.org
http://mail.python.org/mailman/listinfo/distutils-sig
