[Nix-dev] On npm2nix and the NPM package set in Nixpkgs

[Sander van der Burg]

Ok good to hear!

The reason why I would take the pragmatic approach is because I know from
experience that any kind of fundamental change (regardless of whether is
good or bad) will take time for people to accept. Having two versions makes
it possible for people to gradually accept and to smoothly make the
transition to a new solution.

Forcing people to unexpectedly do this will not make them happy.


On Tue, Jul 5, 2016 at 9:44 AM, Johannes Bornhold <
johannes.bornhold at gmail.com> wrote:

> Hi
>
> > On 04 Jul 2016, at 17:34, Sander van der Burg <svanderburg at gmail.com>
> wrote:
> >
> > So far only one response...
> Sorry, was silent agreement on my end ;)
>
> > I'm planning to implement the most pragmatic approach very soon -- due
> to lack of a better/cooler name I'll rename my fork of npm2nix to node2nix.
> >
> > Moreover, I will add a second attribute set to Nixpkgs allowing people
> to deploy packages that have been generated with node2nix. Also, I will
> take the original node-packages.json as a basis, but I will remove the
> library packages that I believe that should not be in there.
> >
> > Because the old package set will still be there, nobody should be
> disrupted and meanwhile people can try/test the new approach.
> >
> > Any objections?
>
> I've looked into you fork already, looks very positive to me. Appreciate
> your work on it and also the in-depth background explanation in your blog.
> :)
>
> I am using mainly pip2nix and npm2nix in my projects to generate package
> sets, willing to switch to node2nix soonish.
>
> From my perspective, I would even take a switch under the same name, even
> if it would mean a few hazzles for me potentially to get things up and
> running again. On the other side, esp. if people use it and get such a
> change as a surprise, it can be negative, since they would be forces to
> switch fast or get the old npm2nix available still. At least a fallback
> option to get the old npm2nix easily would be good to have if we keep the
> same name for the new thing.
>
> Cheers, Johannes
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.science.uu.nl/pipermail/nix-dev/attachments/20160705/4beba5cc/attachment.html>