On Sat, May 12, 2012 at 09:27:21PM -0400, Andrew Dunstan wrote:
> >Should we go with a single developer per item, and then let people
> >suggest corrections? With reviewers involved, and often multiple commit
> >messages per release note item, the just isn't enough detail in git logs
> >to reproduce this accurately. I also over-emphasized new
> >developers/reviewers, but that seems to have distorted the other goals
> >unacceptably.
>
> Most cases should be pretty clear. Most features have a single major
> commit. The author(s) mentioned there are who should be listed,
> IMNSHO. That might leave a handful of cases where more judgement is
> required.
>
> We seem to be in danger of overthinking this.
Results have just shown it isn't a simple case. It is unclear how
important the reviewers were, and how much a committer rewrote the
patch, and the significance of follow-on commits.
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ It's impossible for everything to be true. +