Re: Adding reviewing to our development process

[Jeremy Katz <katzj redhat com>]

On Thu, 2008-10-23 at 09:51 -0400, Chris Lumens wrote:
> > As I've been slowly getting into anaconda development I've become 
> > increasingly amazed with the number of regressions we have caused by 
> > typos and other small predictable errors in day to day development work.
> >
> > So I've been thinking that we really need to change our processes to 
> > include a review phase for all patches except the most trivial ones (the 
> > ones fixing all the typos).
> 
> Two things:
> 
> (1) We already have the commit mails, which we theoretically could all
> be reading and finding problems in.  Yet, I don't see a lot of followup
> comments in IRC regarding typos.  Jeremy has some from time to time, and
> I do too but very rarely.  I wonder how effective it's really going to
> be.

One thing to consider is moving the commit spam to this list so that
it's more visible...

> (2) If we go down this road, we need to make sure it's not mandatory.
> That is, I still need to be able to commit emergency things to do a
> build right now without having to wait for other people to comment.
> That's just the nature of anaconda development from time to time.

I think that the easiest "emergency escape hatch", although it's one
that is kind of painful and sucky, is patches in the src.rpm.  It makes
it more clear that "this is not finalized".  

Jeremy