[Distutils] setup.cfg new format proposal

[Jim Fulton]

On Fri, Sep 11, 2009 at 11:49 AM, Tarek Ziadé <ziade.tarek at gmail.com> wrote:
> On Fri, Sep 11, 2009 at 5:36 PM, Jim Fulton <jim at zope.com> wrote:
>> On Fri, Sep 11, 2009 at 11:14 AM, Tarek Ziadé <ziade.tarek at gmail.com> wrote:
>>> 2009/9/11  <exarkun at twistedmatrix.com>:
>>>> On 02:33 pm, ziade.tarek at gmail.com wrote:
>>>> I'm a little skeptical about creating a new mini language (particularly one
>>>> with branching) for setup.cfg,
>>
>> Me too.
>>
>>>> but I haven't really been paying close
>>>> attention to this discussion.
>>
>> Me neither. :)
>>
>>>>  Still, it might be good to give a brief
>>>> summary of the justification for this someplace.  Maybe you were already
>>>> planning to do that.
>>>
>>> The idea is to be able to get with setup.cfg all the metadata without running
>>> any code from the distribution, and without installing it.
>>>
>>> Which is possible for most cases. (if not possible, setup.py  can still be used)
>>>
>>> the mini-language is to be able to adapt these metadata depending on
>>> the execution context. I'll add a doc somewhere.
>>
>> Is that really necessary?  I mean, if you have to adapt the meta data
>> to the execution context, maybe you should fall back to writing a
>> setup.py.
>
> The main background motivation is the future adding of
> "install_requires" in PEP 345.
>
> So you can for example define a list of dependencies that changes wether
> you are under windows or linux, etc..

OK. I've never had that situation. I'm sure someone has, but it
doesn't see common.

> The second motivation is : why do I need to run some code to get the
> metadata fields ?

Turning the config file into code means you'd be running code. :)

Still, I suspect that 99% (or some very high percentage) of cases
wouldn't need conditional logic in meta data files.


> do I need to install a package to gets its metadata ?

This is an important use case.  It is orthogonal to defining packages
with static meta data.

...

>
> try to run this with lxml source tarball:
>
> $ python setup.py --name
>
> I don't know what it is doing, but a whole lot more that what it is
> suppose to do
> just to let me know the 'name' metadata.
>
> So if name was in setup.cfg, it would be better.

Somewhat. I think the main concern though is to get the meta data
without installing.

In any case, I think a sufficiently large percentage of cases can be
handled without logic in the setup.cfg.


>
>>
>> I agree that more background and motivational information would be
>> very helpful, as would some real world examples. I know that >90%
>> (maybe 99% :) of the packages I distribute don't adapt to their
>> execution context (beyond what distutils does internally) or need
>> anything but meta data.
>
> That's why you could push your fields in the static setup.cfg.
>
> (and probably have an 'empty' setup.py file)

Yup. Don't get me wrong. I'm in favor of being able to define packages
with static meta data. I just don't consider configuration files with
logic in them to be static. :)

Jim


-- 
Jim Fulton