[Python-Dev] About resolution “accepted” on the tracker

[R. David Murray]

On Mon, 18 Oct 2010 21:42:08 +0200, Antoine Pitrou <solipsis at pitrou.net> wrote:
> On Mon, 18 Oct 2010 21:31:24 +0200
> Georg Brandl <g.brandl at gmx.net> wrote:
> > 
> > This is probably sophistry, but if an issue is invalid, it doesn't need
> > a patch :)
> 
> Not only, but it generally gets closed too.
> 
> > The first stage seems to be "unit test needed" anyway, which
> > sounds to me a bit like "needs to be checked for reproducibility/validity".
> 
> I don't like this first stage, it makes it look like we mandate a
> proper unit test to proceed with actually writing patches, which is
> really not true.

Why isn't it? :)

Seriously, though, what it indicates is indicates is that we need a
unit test for the patch to be complete.  We have a number of issues
with patches but no tests, I believe.  Which order 'unit test' and 'fix'
occur in is arbitrary in practice.  I certainly prefer to have the unit
tests first myself, though.

The problem is that the stage field really isn't all that useful.
I'd prefer a set of check boxes, as I've suggested in the wiki.

I was the one who advocated labeling it 'unit test needed', but if
people would rather change it back to just 'test needed', I will raise
no objection, since in practice trying to squeeze the meaning I wanted
into the stage field doesn't really work.

--
R. David Murray                                      www.bitdance.com