[Freeciv-Dev] Re: [PATCH] advdomestic.c cleanup (PR#1149)

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Fri, Dec 21, 2001 at 08:02:47PM -0500, Ross W. Wetmore wrote:
> 
> I have a serious problem with volume and topics of patches, but
> luckily I'm not a maintainer so I don't have to worry and can pick
> and choose things to look at from an interest standpoint :-).
> 
> I think the "maintainers" need to do several things, though.
> 
> 1)  They need to vet the final submissions and include them to CVS.
> 
> 2)  They need to be involved in the initial triage to select things
>     that are put on the submission track, those that are just quick 
>     bugfixes, and can be processed ASAP as a "break" from harder stuff,
>     or the not-a-snowballs-chance-in-hell category.
> 
> 3)  They need to learn how to delegate the heaviest workload between
>     triage and final submission cycle so a patch can be cleaned up to
>     the point where it is worth their time to review.
> 
> 4)  They need to figure out how to trust or manage the people they 
>     delegate to, enough to not micromanage the process, or start it 
>     all over again when it is in the final submission state.
> 
> The last is probably the real kicker.
> 
> For a start, maintainers should have a list of people that are willing
> to commit to managing a review and cleanup cycle. This involves two
> things, a real code level review and test pass, and a longer term point
> role in iterating the resolution of issues. Discussion on the list
> should continue, but this needn't mean half the emails are from Raimar
> as the only point man for everything.
> 
> The maintainers can then allocate the review load as it comes in to 
> these people rather than waiting to see who has time or inclination.
> This should mean fewer maintainer cycles spent on cleanup and more
> on processing. This will also better coordinate the list's efforts
> to schedule and handle reviews.
> 
> You can add my name to a list of people willing to review up to 3
> outstanding patches at a time. I suggest that any others that are 
> willing to make a reasonable commitment to turn around an initial 
> review and manage a discussion of a patch cleanup in a professional 
> manner to send your names in to the powers that be.

I don't like to force somebody. In the past I have asked for reviews
and usually there was none. This shows to me that not many people are
willing to spend time outside their "area of interest". This is ok. We
are still here on a voluntary basis. However spending time outside his
"area of interest" is something a maintainer has to do. I would really
like to work more on the agents stuff but I spend the time I have for
freeciv at reviewing patches. Patches which touch all other areas but
not the agent area.

Another option to "a maintainer has to work on almost all parts of the
code" is to say that a maintainer is a specialist (AI, map, GUI,
...). We would get maintainers if we follow this way but I think this
way is wrong. Such a maintainer won't have to whole picture and will
produce architecturally bad code.

Another thing: a review is only worth the reputation of the
reviewer. This reputation isn't based on how many patches the reviewer
has reviewed but how intensive. This is hard to measure but the number
of complains if a rough estimation. Yes every patch of a reasonable
has errors/places which can be improved. So far I think that Ross is
the best reviewer. If he goes over a patch he will review the patch in
deep and comes out with a long list of ideas/errors. So don't think
that if you say "the patch looks sane" you will help the patch to go
in. Remember you don't review for me/for the inclusion but to make the
code better/give the author feedback.

I'm sorry but I have to leave now. Goodbye till the first week of the
new year.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "Despite all the medical advances of the 20th century, the mortality 
  rate remains unchanged at 1 death per person."