maintainer timeout for FreeBSD commiters

[John Marino]

On 7/17/2012 23:39, Mark Linimon wrote:
> On Tue, Jul 17, 2012 at 07:54:59AM -0400, Michael Scheidell wrote:
>>>> We *are* making progress in cutting through the backlog though.
>>> ports have about 900 open PR. Why it does not have more port
>>> commiters? Its difficult to recruit new person?
>> The answer to that is very complex.
>> And, for each PR, maybe a different answer.
>
> This is true, but to address the previous question ...
>
> It is somewhat difficult to recruit new people who are willing to work
> within guidelines (both technical and social) and who seem to want to
> stay for the long-term.  portmgr does screen candidates to try to make
> sure that the quality of the Ports Collection doesn't decline.  Each
> vote is a judgement call.
>
> Having said that, we add new committers all the time.  OTOH we add new
> ports all the time, and due to this the backlog seems to remain constant.
>
> And again, as scheidell notes, some PRs are more equal than others.
>
> mcl

Hi Mark,
I think that's a reasonable assessment about how the backlog seems about 
the same and how processes just naturally work.  But I think it could 
work better.

Let's take my case.
I'm a maintainer of several Ada ports and compilers.  I'm also a pkgsrc 
committer, but not a FreeBSD ports committer.  I have the same packages 
in both trees, but the pkgsrc packages (ports) are more current.  That's 
obviously because I can commit to one tree at will but I have to submit 
PR and get in line for each update at FreeBSD (A quick shout out of 
appreciation to Frederic who has been tremendously gracious to me over 
these months).

I was thinking about this - I really like how FreeBSD ports enforces to 
the best of its ability that every port have a maintainer.  My name is 
on several ports and I have pride in my work.  Would it be so bad if all 
my submitted patches (as a recognized quality contributor with history) 
just got committed as a passthrough?  Obviously you might be reluctant 
to do this on ports that 200 packages depend on, but if you created a 
tier of contributors below committer but above PR submitter, I think a 
lot of ports would be maintained more often and there wouldn't be so 
much of a backlog.

The worst case scenario is a contributor turns out to be a little 
sloppy, doesn't bother to use Tinderbox, etc, and after a couple of 
incidents you pull his privileges.  The benefit you gain from the others 
would outweigh the incidents.

I've seen the response that the committer is responsible for everything 
he or she commits, but if the community gave them immunity from 
consequences of maintainer patches, it shouldn't be a problem.

I don't expect anything to come of this suggestion, but I've always 
wondered why more responsibility wasn't given to port maintainers who 
don't have commit privileges.

John