Re: review process not working

[Reinhold Kainhofer]

Am Dienstag, 26. Juli 2011, 18:54:28 schrieb David Kastrup:
> Graham Percival <address@hidden> writes:
> > On Tue, Jul 26, 2011 at 12:05:31PM +0200, David Kastrup wrote:
> >> Perhaps a minimal measure of sanity would be if a patch countdown
> >> without code review was only started when the author of the patch
> >> says "I feel reasonably confident that this not just works, but is
> >> good".
> >> 
> >> In git, there is the "formal" sanctification of "Signed-off-by".
> >> Perhaps we should not start a patch countdown on any patch that has
> >> not been signed off by anybody?
> > 
> > I speak against this, at least for now.  This is a question of
> > balance between support for new contributors (i.e. mentors, of
> > which we have far fewer than I would like), amount of available
> > reviewers (which is smaller than we would like), and the moral of
> > contributors.
> 
> A contributor is free to add "Signed-off-by" himself.  I was talking
> about patches that not even the contributor feels confident enough about
> to undersign it.

No, please don't add any more bureaucratic / adminstrative duty for infrequent 
contributors that barely know git and the command line. If a contributor 
contributes a patch, his submission is already his signing. Remember, we are 
not the group of hardcore geeks the Kernel developers are!

I don't see a problem in the review process itself. Rather it is a problem 
that the quality of reviews that you request requires perfect knowledge. I, 
for example, was not aware of those possible problems with pow. Neil did a 
review and gave his LGTM. Probably he wasn't aware of integer/double problems 
either (and his reviews and his LilyPond knowledge can only be described as 
excellent). So, it seems the only one who is aware of those rounding problems 
is David. The question then is, why didn't David do a review of the patch and 
now complains that the process is not working?

Cheers,
Reinhold
-- 
------------------------------------------------------------------
Reinhold Kainhofer, address@hidden, http://reinhold.kainhofer.com/
 * Financial & Actuarial Math., Vienna Univ. of Technology, Austria
 * http://www.fam.tuwien.ac.at/, DVR: 0005886
 * LilyPond, Music typesetting, http://www.lilypond.org