[Distutils] Provisionally accepting PEP 517's declarative build system interface

[Thomas Kluyver]

On Mon, May 29, 2017, at 10:19 AM, Paul Moore wrote:
> I think that the key thing here is that as things stand, pip needs a
> means to copy an existing "source tree", as efficiently as possible.
> For local directories (source checkouts, typically) there's a lot of
> clutter that isn't needed to replicate the "source tree" aspect of the
> directory - but we can't reliably determine what is clutter and what
> isn't.

This sounds similar to the question we had in flit about making sdists -
we don't want to reinvent MANIFEST.in, but we also don't want to include
random clutter that might be present in the source tree.

The compromise we've gone for in flit is to ask the VCS which files it
is tracking and only include those, on the grounds that it should be
possible to treat a fresh clone/checkout of the repository as the
source. But this probably wouldn't work for what pip wants, because
there are two big limitations:

1. Flit can only build an sdist from a VCS checkout
2. Flit needs to know about the VCS you're using