Re: [Python-Dev] a new setuptools release?
At 12:16 PM 10/6/2009 -0700, Guido van Rossum wrote:
I think at this point the community should not be forced wait for you to get a new supply of round tuits. The wait has been too long already. You can stay on in an advisory role, but I don't think it's reasonable to block development or decisions until you have time.
Tarek didn't think so either, which is why he created a fork, called Distribute. So I'm not really clear on what you're trying to say here (or in the rest of the email, for that matter).
I'm saying that I don't expect setuptools 0.7 to appear before Tarek's
Distribute is mature and in widespread use. IOW I support Tarek's fork
and suggest nobody hold their breath waiting for setuptools 0.7.
On Tue, Oct 6, 2009 at 1:08 PM, P.J. Eby
At 12:16 PM 10/6/2009 -0700, Guido van Rossum wrote:
I think at this point the community should not be forced wait for you to get a new supply of round tuits. The wait has been too long already. You can stay on in an advisory role, but I don't think it's reasonable to block development or decisions until you have time.
Tarek didn't think so either, which is why he created a fork, called Distribute. So I'm not really clear on what you're trying to say here (or in the rest of the email, for that matter).
-- --Guido van Rossum (home page: http://www.python.org/~guido/)
Guido van Rossum wrote:
I'm saying that I don't expect setuptools 0.7 to appear before Tarek's Distribute is mature and in widespread use. IOW I support Tarek's fork and suggest nobody hold their breath waiting for setuptools 0.7.
Well, if this was the BDFL pronouncement that a lot of people have taken it to be, does that allow Martin von Lewis to give the keys to http://pypi.python.org/pypi/setuptools to the "distribute" developers, so we can get on and use the original "setuptools" name without all the confusion and hackery needed to make "distribute" work? cheers, Chris PS: I had thought that MvL couldn't help because of ez_setup.py, but it's only buildout that appears foolish enough to have a hardcoded download of: exec urllib2.urlopen('http://peak.telecommunity.com/dist/ez_setup.py' ).read() in ez ..and even that will keep working provided PJE doesn't take it down... -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk
On Fri, Oct 9, 2009 at 5:01 PM, Chris Withers
Guido van Rossum wrote:
I'm saying that I don't expect setuptools 0.7 to appear before Tarek's Distribute is mature and in widespread use. IOW I support Tarek's fork and suggest nobody hold their breath waiting for setuptools 0.7.
Well, if this was the BDFL pronouncement that a lot of people have taken it to be, does that allow Martin von Lewis to give the keys to http://pypi.python.org/pypi/setuptools to the "distribute" developers, so we can get on and use the original "setuptools" name without all the confusion and hackery needed to make "distribute" work?
I am -1 on this. At Python and PyPI level, 'Setuptools' is PJE's project and he's the one that manages the list of users that have extended rights on his project. In Distribute, our setup.py script makes sure setuptools is not in the way. It is explicit and documented, and there's a documentation explaining how to switch back to Setuptools. This exists only because there's no way yet in Python to say that a distribution can't run when another distribution is installed. But it's up to every project and developer to choose to switch to Distribute. And it's up to OS packagers to choose the best strategy at their levels when they package tools for their packaging system (so they pick one or the other and they are able to make a distinction between PJE python distribution and ours) Distribute developers will of course promote Distutils usage everywhere, but hijacking setuptools in PyPI would be a very bad thing imho. Tarek
Tarek Ziadé wrote:
At Python and PyPI level, 'Setuptools' is PJE's project and he's the one that manages the list of users that have extended rights on his project.
Indeed, if only he would see sense :-(
In Distribute, our setup.py script makes sure setuptools is not in the way. It is explicit and documented, and there's a documentation explaining how to switch back to Setuptools.
It's a hack though...
This exists only because there's no way yet in Python to say that a distribution can't run when another distribution is installed.
I don't this should ever be needed, except in this edge case caused by one man's stubbornness...
Distribute developers will of course promote Distutils usage everywhere, but hijacking setuptools in PyPI would be a very bad thing imho.
*sigh* I don't see it as hijacking, provided Guido is making a BDFL pronouncement that you're maintaining this software and PJE finally sees sense and accepts that. Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk
Chris Withers
*sigh* I don't see it as hijacking, provided Guido is making a BDFL pronouncement that you're maintaining this software
Well, what you are proposing *is* hijacking. PJE is not paid or employed by Guido, he is the full author of setuptools. Forking is of course fine (this is free software), but saying "I overthrow you as the maintainer of this software" is not. While you (and I) may be unsatisfied with PJE's maintenance style, personal (dis)affection should not be what motivates our decisions. A piece of software has an author who has some moral rights on it, and should be treated with respect. After all, when an author decides to release code under a free license, it shows a fair amount of trust and respect towards the user; it is only fair that the user shows the same amount of trust and respect (*). (*) it has happened that some users don't, and in that case the author's reaction can be violent. For example, Nessus became proprietary: http://news.cnet.com/Nessus-security-tool-closes-its-source/2100-7344_3-5890... Regards Antoine.
On Fri, Oct 9, 2009 at 10:57, Antoine Pitrou
Chris Withers
writes: *sigh* I don't see it as hijacking, provided Guido is making a BDFL pronouncement that you're maintaining this software
Well, what you are proposing *is* hijacking. PJE is not paid or employed by Guido, he is the full author of setuptools. Forking is of course fine (this is free software), but saying "I overthrow you as the maintainer of this software" is not.
While you (and I) may be unsatisfied with PJE's maintenance style, personal (dis)affection should not be what motivates our decisions. A piece of software has an author who has some moral rights on it, and should be treated with respect. After all, when an author decides to release code under a free license, it shows a fair amount of trust and respect towards the user; it is only fair that the user shows the same amount of trust and respect (*).
FWIW I agree with Antoine. -Brett
Chris Withers
Well, if this was the BDFL pronouncement that a lot of people have taken it to be, does that allow Martin von Lewis to give the keys to http://pypi.python.org/pypi/setuptools to the "distribute" developers, so we can get on and use the original "setuptools" name without all the confusion and hackery needed to make "distribute" work?
Er, perhaps this is a bit brutal? I don't think the BDFL has a right of life and death over third-party packages. I mean, we're an anarcho-syndicalist commune, aren't we? WOMAN: Who are the Britons? ARTHUR: Well, we all are. we're all Britons and I am your king. WOMAN: I didn't know we had a king. I thought we were an autonomous collective. DENNIS: You're fooling yourself. We're living in a dictatorship. A self-perpetuating autocracy in which the working classes-- [snip] Regards Antoine. PS :
PS: I had thought that MvL couldn't help because of ez_setup.py, but it's only buildout that appears foolish enough to have a hardcoded download of:
Some projects also have an SVN "externals" pointing to the setuptools SVN repository, as described here: http://peak.telecommunity.com/doc/ez_setup/index.html
Chris Withers schrieb:
Guido van Rossum wrote:
I'm saying that I don't expect setuptools 0.7 to appear before Tarek's Distribute is mature and in widespread use. IOW I support Tarek's fork and suggest nobody hold their breath waiting for setuptools 0.7.
Well, if this was the BDFL pronouncement that a lot of people have taken it to be, does that allow Martin von Lewis to give the keys to http://pypi.python.org/pypi/setuptools to the "distribute" developers, so we can get on and use the original "setuptools" name without all the confusion and hackery needed to make "distribute" work?
That's absurd. There's a certain area where Guido can make pronouncements, but third-party packages is not it. Even if they're hosted on python.org infrastructure. Georg -- Thus spake the Lord: Thou shalt indent with four spaces. No more, no less. Four shall be the number of spaces thou shalt indent, and the number of thy indenting shall be four. Eight shalt thou not indent, nor either indent thou two, excepting that thou then proceed to four. Tabs are right out.
On Fri, Oct 9, 2009 at 1:50 PM, Georg Brandl
Chris Withers schrieb:
Guido van Rossum wrote:
I'm saying that I don't expect setuptools 0.7 to appear before Tarek's Distribute is mature and in widespread use. IOW I support Tarek's fork and suggest nobody hold their breath waiting for setuptools 0.7.
Well, if this was the BDFL pronouncement that a lot of people have taken it to be, does that allow Martin von Lewis to give the keys to http://pypi.python.org/pypi/setuptools to the "distribute" developers, so we can get on and use the original "setuptools" name without all the confusion and hackery needed to make "distribute" work?
That's absurd. There's a certain area where Guido can make pronouncements, but third-party packages is not it. Even if they're hosted on python.org infrastructure.
Right. -- --Guido van Rossum (home page: http://www.python.org/~guido/)
On Fri, Oct 09, 2009 at 01:56:42PM -0700, Guido van Rossum wrote:
On Fri, Oct 9, 2009 at 1:50 PM, Georg Brandl
wrote: Chris Withers schrieb:
Guido van Rossum wrote:
I'm saying that I don't expect setuptools 0.7 to appear before Tarek's Distribute is mature and in widespread use. IOW I support Tarek's fork and suggest nobody hold their breath waiting for setuptools 0.7.
Well, if this was the BDFL pronouncement that a lot of people have taken it to be, does that allow Martin von Lewis to give the keys to http://pypi.python.org/pypi/setuptools to the "distribute" developers, so we can get on and use the original "setuptools" name without all the confusion and hackery needed to make "distribute" work?
That's absurd. ?There's a certain area where Guido can make pronouncements, but third-party packages is not it. ?Even if they're hosted on python.org infrastructure.
Right.
Is that a pronouncement? :) "GvR, the self-limiting BDFL." --titus -- C. Titus Brown, ctb@msu.edu
>> That's absurd. There's a certain area where Guido can make >> pronouncements, but third-party packages is not it. Even if they're >> hosted on python.org infrastructure. Guido> Right. Now if you were the un-BDFL you could step in here. ;-) This whole topic seems to have a lot of people fairly agitated, so clearly it's important to a significant subset of the Python development community. Might I suggest taking this off python-dev for awhile though? I seem to recall a similar suggestion a couple days ago to take it to distutils-sig or python-ideas, but that seems to have been ignored.) I mostly tuned the entire thread out until I saw that Guido had joined in the fray (sort of). Maybe since I don't distribute a lot of Python packages it's not as important to me. Let me know when you've solved the problem. I trust you. Skip
participants (9)
-
Antoine Pitrou
-
Brett Cannon
-
C. Titus Brown
-
Chris Withers
-
Georg Brandl
-
Guido van Rossum
-
P.J. Eby
-
skip@pobox.com
-
Tarek Ziadé