[Distutils] more easyinstall error messages - haha

A. Cavallo a.cavallo at mailsnare.com
Thu Jun 18 16:27:33 CEST 2009


Please have a look to my pet project:

http://pyvm.sourceforge.net

It solves all the problem you could possibly imagine (famine included!):

* provides binary build - ADMINISTRATOR FRIENDLY
* could guarantee a reproducible build
* integrates smoke tests (just the begin of a continuous integration system)


Regards,
Antonio


> On Thu, 18 Jun 2009 12:19:13 +0200, Tarek Ziadé <ziade.tarek at gmail.com>
>
> wrote:
> > Let's suppose you have all your code 100% covered in tests (wich is
> > not the case in 99.99% of the packages).
>
> Sure. Ok, developers don't write enough tests...
>
> They don't have time... it's not fun.... can't be bothered...
>
> But lets be realistic, there's too many python versions now...
>
> too many platforms...
>
> why can't we do automated builds/tests to cut down the work for
> developers...
>
> > So you run the test using all Python versions (which supposes you have
> > *all* Python versions installed, which is not the case on 99% of the
> > people systems out ther)
>
> Think of it like this....
>
> We give python 3.1 the crown....
>
> and write the code for that, that can regressively test any code
> backwards a package on all the old python versions..
>
> thus, our latest python version... has way more power than any
> previous version...
>
> > Then what happens ? Do you store the result in the metadata ? Like, by
> > setting up the Requires-Python field with the output ?
>
> I didn't think of that... it's an excellent idea...
>
> Think of how much work that could save...
>
> > Frankly, you could just explicitely mark the Python version your code
> > is suppose to run with,
>
> but it would be more fun as you say if it was determined programmatically
> by distutils based on running some analysis of the actual code....
>
> David
> _______________________________________________
> Distutils-SIG maillist  -  Distutils-SIG at python.org
> http://mail.python.org/mailman/listinfo/distutils-sig


More information about the Distutils-SIG mailing list