[MPlayer-dev-eng] [RFC] Policy update

[Uoti Urpala]

On Sat, 2006-10-07 at 19:16 +0200, Reimar Döffinger wrote:
> On Sat, Oct 07, 2006 at 08:06:23PM +0300, Uoti Urpala wrote:
> > which are believed to be controversial (with some consideration given to
> > the size and risk of the change - changes which are easy to revert and
> > cannot cause serious breakage need less guarantee of consensus before
> > applying).
> [...]
> 
> I actually disagree. Reverting a change is something many people will
> try hard to avoid because it results in lots of nastiness and in some
> cases to people fighting about policy via their commit privileges. I
> certainly would prefer to have that road of escalation sealed tightly.

Using "believed-to-be-noncontroversial commits are OK" as an excuse to
keep redoing reverted changes is quite weak. I don't see that as a
significant risk.

> Nobody will break his fingers by writing one additional mail.

Nobody will break his fingers, but that's not a reasonable standard.
There's lots of stupid stuff that could be introduced without breaking
anyone's fingers.

Extracting a patch, writing a mail with it attached, and then applying
the patch at _another_ time (so you can't finish things in one go) is
several times more work than simply applying the change. And I do not
want to see patches for obvious things from _other_ developers on the
mailing list either. Seeing them once in the commit log mails is quite
enough.

> I did not even read the proposal, but the problem with the VIDIX commit
> (and some others) is that it is not a technical but a political
> decision (e.g. do we want to advertise VIDIX and make people port it?).
> The problem with political decisions is
> 1) there is likely always someone who disagrees
> 2) it leads easily to flames
> 3) it should be carried by more than 2 or 3 developers if possible, at
> least it's not something a maintainer can really decide, at best a
> project leader could.

So do you really believe that people will disagree with the VIDIX
change? That there is a chance the eventual consensus will be to revert
it? If not then IMO you have no reason to complain. Even if the VIDIX
change was "political" to some degree, it was still obvious. Saying that
every decision which can in some way be interpreted as "political",
however obvious, should be carried out by more than 3 developers is
nonsense.