[Nix-dev] Re: Python packages in peril

[Lluís Batlle i Rossell]

On Tue, Apr 20, 2010 at 02:14:33PM +0200, Marc Weber wrote:
> I personally consider a broken trunk in nixpkgs being a result of the
> current development cycle I'm forced to use.
The current development in nix works without much technical enforcements to good
practices, so we work on a mutual agreement.
That agreement includes "do not break trunk". If you break trunk, you should
feel a bit of shame, ask for forgiveness, and fix that as soon as possible, in
higher priority than any other development you may have in mind.

You know that there are simple tests that you can run in less than one minute in
your own computer, before commiting. If your changes don't pass the test, you
will be annoying all the trunk users (most of the nix users!).

If you don't agree to those practices, or you require technical enforcement to
be protected against your behaviour... that sounds bad.

You should rethink about your contributing practices. For sure, these are not
forcing you to break trunk. You could have proceeded perfectly without breaking
trunk.

> 
> The only true fix is dropping trunk, adding a "stable" or release
> branch. It is not feasable asking every comitter to test all variations
> on all platforms (x86_64, i686, arm, Windows..)
The tarball build tests the evaluation in all those platforms.

> So I vote for b'). If that can't be done I'd vote for b'). However this
> would mean that I have to spend time on SVN rather than updating
> packages.
For sure you have to spend *more* time preparing your commits. Both in svn
practices and in your nix code/test practices.