Re: Trunk still not open

[Stephen J. Turnbull]

Glenn Morris writes:
 > Dmitry Gutov wrote:
 > 
 > > I really wish the project went with "no-push-without-tests" requirement
 > > instead, first.
 > 
 > That is a separate issue. I'd like it not to sidetrack this discussion.
 > 
 > (I previously asked for more people to write tests:
 > 
 > https://lists.gnu.org/archive/html/emacs-devel/2013-06/msg00995.html
 > 
 > Not much happened, except that a bunch of people wrote a bunch of
 > emails, like always, and nobody did anything.

I am sure that is not true, and Daniel documents that.  What is true
is that it is not enough, and that is because the maintainership did
not have the stomach for *requiring* tests.  (And I agree with that
decision, for now.  Especially because docs come first, and
traditionally have been important to GNU and Emacs in particular.)

 >> As for docs, I apologize (being one of the culprits), but as I
 >> rarely, if ever, read the manual, one can understand how I can
 >> forget its existence from time to time.
 > 
 > I hope you will change your mindset so that you view the
 > documentation as central to Emacs.

Why hope?  If the maintainers collectively set this as a rule, I'm
sure Stefan will follow the rule conscientiously, like everybody else.

Without a rule, it's easy to make excuses for yourself.  Stefan may
not, I actually suspect he will exert himself to set an example.  But
that still means way too much code will be added without high-quality
documentation, and sometimes essentially undocumented, because
statistically only a minority will exert themselves.

I believe that the docs rule is actually pretty low-cost to Emacs
developers as shown by the tradition of high-quality documentation.
Why not give it a try, and tackle the much more costly task of
encouraging then requiring testing when a docs rule succeeds?