On Thu, Feb 18, 2016 at 12:56:05PM -0500, Leo Famulari wrote:
> On Thu, Feb 18, 2016 at 10:03:32AM +0200, Efraim Flashner wrote:
> > On Sun, 14 Feb 2016 20:44:52 -0500
> > Leo Famulari <l...@famulari.name> wrote:
> >
> > > * gnu/packages/python.scm (python-enum34)[inputs]: Remove
> > > python-setuptools.
> > > ---
> > > gnu/packages/python.scm | 2 --
> > > 1 file changed, 2 deletions(-)
> > >
> > > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm
> > > index 9296cdf..66f911d 100644
> > > --- a/gnu/packages/python.scm
> > > +++ b/gnu/packages/python.scm
> > > @@ -1320,8 +1320,6 @@ bug tracker.")
> > > (base32
> > > "0yx1m4564wxgbm4glb3457hi16xihd9w63rv13y2przkdir9dfgp"))))
> > > (build-system python-build-system)
> > > - (inputs
> > > - `(("python-setuptools" ,python-setuptools)))
> > > (arguments
> > > `(#:phases
> > > (alist-replace
> >
> > Does this break python2-enum34?
>
> I don't think so. python2-cryptography still builds, passes it's tests,
> and seems to work in "downstream" applications.
>
> I've noticed that the PyPi importer seems to unconditionally add
> setuptools as an input (I didn't check the code yet). So, perhaps that
> explains it's presence.
>
> Or perhaps the version of python-2 we used at the time (2.7.6) did not
> include some setuptools feature that has since been added to the
> standard library.
Also, python-enum34 still builds with this change.
