[Distutils] Distutils work, roadmap

Paul Moore p.f.moore at gmail.com
Tue Apr 21 16:06:21 CEST 2009


2009/4/21 Tarek Ziadé <ziade.tarek at gmail.com>:
> On Tue, Apr 21, 2009 at 12:37 PM, Paul Moore <p.f.moore at gmail.com> wrote:
>> 2009/4/21 Tarek Ziadé <ziade.tarek at gmail.com>:
>>> On Tue, Apr 21, 2009 at 11:11 AM, David Lyon <david.lyon at preisshare.net> wrote:
>>>>
>>>> Tarek,
>>>>
>>>> I wouldn't take "feedback" as all bad....
>>>>
>>>> Indirectly, it seems like you are telling people not to report their
>>>> experience. Indirectly, it seems like you are discouraging thinking
>>>> about what could be done to improve things.
>>>
>>> I am encouraging people to work with the people that have started
>>> to work on various topics durong Pycon
>>
>> Given that I have not got the time or experience with setuptools or
>> distutils to contribute effort to fixing things, you seem to be saying
>> that my experience as an end user of these tools, and my spending my
>> time trying to understand the implications of the proposals so that I
>> can offer useful feedback, is not of interest.
>
> I have never said that. On the contrary, I encourage you to fill your
> use case as an end user
> in the wiki.
[...]
> I think you misunderstood my point.

Hmm, OK. If that's the case then I should apologise. I am struggling
to understand the details of all these discussion threads, and that
probably coloured my response.

One concrete point - The categories on the wiki don't really give me
much help in knowing where to put my use case. I'm a user of Python
packages - all I care about is running the installer, or if I must
typing python setup.py bdist_wininst. So these aren't relevant, as
they are developer-oriented:

    DistutilsVersionFight
    Distutils/StandardizeEggInfo
    Distutils/Metadata
    Distutils/StaticMetadata
    Distutils/ManifestPluginSystem
    Distutils/TestingInfrastructure

This might be relevant, as it sounds like it relates to ensuring that
bdist_wininst installers are available. But it sounds more like it's
about maintaining bdist_wininst outside of the core - so I'm not sure
where I'd state that as a user I don't want to have to download a
3rd-party bdist_wininst before I can build my own packages if there
isn't a Windows installer. It sounds like there's an implicit
assumption there that I disagree with :-(

    Distutils/Friends : the goal is to try to find a project, a person
or a group of person on each platform that is willing to maintain a
third-party tool that build system-specific distros out of python
package.

Ultimately, though, my "use case" is simple. I want to download and
install Windows platform-specific installers for each package I use,
without needing to care about distutils or setuptools. To the extent
that I am exposed to the existence of distutils/setuptools (beyond the
utterly bare minimum of "python setup.py bdist_wininst"), my
requirement isn't being met - and I'll happily discuss whether I
should expect to lower my expectations, and what benefits there are
(to me, or others) in doing so. But that's it.

The sticking point here is that before setuptools, *my requirements
were largely being met*. Since setuptools appeared on the scene, they
are being eroded (witness projects which have switched to distributing
eggs rather than bdist_wininst installers). From that fact stems all
of my anti-setuptools views and comments (which I won't repeat now).

FWIW, I did post my use case here some days ago, with a request that
someone suggest/find a home for it on the wiki, as I wasn't sure where
it went. To my knowledge, no-one ever did that. I'd still appreciate
it if someone could copy it to the wiki. But trying to understand the
various threads and arguments on the list has sapped most of my
energy, so I'm not going to try to guess where it should go.

Paul.


More information about the Distutils-SIG mailing list