Code review required for commits - Discuss.

Andrew Bartlett abartlet at
Fri Oct 12 01:41:11 MDT 2012

On Fri, 2012-10-12 at 09:18 +0200, Volker Lendecke wrote:

> You might argue that it slows down our development. For the
> individual patch this is probably true, but overall I am
> really convinced that it will even speed up development, if
> just for the reason that more developers know more code and
> we reduce the risk of bad patches lingering in the tree for
> long. The risk is not zero, but it will be reduced.

The issue I have is that if I am able (as I sincerely hope I am) to
return to the productivity I've managed in the past, then I'll need a
partner with far more bandwidth for review than has been available in
the past.

I regularly post patches to this mailing list, and on average I think
1/10 are even noticed (and these are the patches I consider might even
expect comment, because they touch possibly sensitive areas under

I continue to advocate for other companies to join the AD effort, but
for now contributors paid to contribute any substantial fraction of
their time to the AD effort is the sum of me, myself and I.  

This isn't Google where employees can be assigned to a task.  This is a
very broad project without a significant commonality of interest across
the whole contributor base.  

For that reason, until I can see how such a proposal could succeed
without significantly damaging my ability to get my work done, I oppose
this proposal. 

I dislike this particular voting style, but to be clear:


I also object to the idea the 'a majority of team members' decides this,
given the lopsided nature of the contribution graph above.  We all want
to improve the quality of Samba's code, and we should be able to work
out an incremental improvement that identifies and addresses a genuine
issue with the least overall burden to us all. 

Andrew Bartlett

Andrew Bartlett                      
Authentication Developer, Samba Team 

More information about the samba-technical mailing list