On Sat, Jan 16, 2016 at 09:57:45AM +0000, Simon Riggs wrote:
> On 16 January 2016 at 02:10, Noah Misch <noah@leadboat.com> wrote:
> > On Wed, Jan 13, 2016 at 12:13:11PM -0500, Tom Lane wrote:
> > > Basically this is trading off convenience of the committer (all of the
> > > alternatives Noah mentions are somewhat annoying) versus the convenience
> > > of post-commit reviewers. I'm not sure that his recommendation is the
> > > best trade-off, nor that the situation is precisely comparable to
> > > pre-commit review. There definitely will be pre-commit review, there
> > > may or may not be any post-commit review.
> >
> > That's a good summary.
> My objective in committing patches to PostgreSQL is to develop the Open
> Source version of PostgreSQL as a standalone product and I encourage others
> to do the same.
>
> PostgreSQL is open source and therefore usable for various additional
> purposes, one of which is modified versions of PostgreSQL.
>
> I will not go out of my way to cause problems for the secondary users of
> the code. I will try to implement one of the suggestions for whitespace
> handling, though may make mistakes in that, nobody being perfect.
Thanks. Clean commits help so many audiences, including immediate post-commit
reviewers, intensive beta testers, fork maintainers, and hackers performing
root cause analysis on the bugs to be discovered in future years. For what
it's worth, most committers already have been using some mix of strategy 2
(leave pgindent entirely to Bruce) and strategy 1 (neither add nor remove work
for the next whole-tree pgindent to do). If you're already in that majority,
I advise no change.