[Distutils] PEP 517 again

[Donald Stufft]

> On Aug 28, 2017, at 4:43 AM, Thomas Kluyver <thomas at kluyver.me.uk> wrote:
> 
> If pip does uses build_wheel directly, as Paul now prefers, I think we
> can leave the NotImplemented/Error/None question for a later date. We
> only want some way to signal "I can't do that" because a frontend might
> try sdist->wheel with a fallback to making a wheel directly. If no
> frontend is actually planning to do that, we can leave specifying it
> until a frontend wants it.
> 
> Donald, what do you think? IIRC, you were most keen on going
> sdist->wheel where possible, and I don't think you've commented on
> Paul's suggestion yet (apologies if I've overlooked a response).
> 


I still think it should, and prefer pip to attempt to build a sdist prior to building a wheel when we’re coming from a VCS directory. I think that is going to be the most robust mechanism with the least amount of surprising behavior for end users.

I also think it is important to define this fallback now, because doing it now means we have flexibility to adjust the strategy that frontends use without going back through the PEP process and waiting on backends to update to the latest version of the PEP.

I am okay with an exception if it’s a specific exception, the idea of having an exception class exported as part of the API is fine with me. I think it’s still a worse API to use an exception for a non-exceptional case, but if it’s a specific exception it at least prevents bubbling up unrelated errors and having them be treated as asking for the fallback behavior if possible.


—
Donald Stufft



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/distutils-sig/attachments/20170828/4f9ed63a/attachment.html>