Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 5 Sep 2003 20:44:04 -0400 (EDT)
From:      "Adam C. Migus" <adam@migus.org>
To:        "Mark Linimon" <linimon@lonesome.com>
Cc:        Kris Kennaway <kris@obsecurity.org>
Subject:   Re: config files in packages (Re: (proposal) new flag       forpkg_delete)
Message-ID:  <49707.192.168.4.2.1062809044.squirrel@mail.migus.org>
In-Reply-To: <Pine.LNX.4.44.0309051813350.3227-100000@pancho>
References:  <49222.192.168.4.2.1062744486.squirrel@mail.migus.org> <Pine.LNX.4.44.0309051813350.3227-100000@pancho>

next in thread | previous in thread | raw e-mail | index | archive | help

Mark Linimon said:
>> This approach works great assuming every port is well written, but
>> every port isn't well written.
>
> Granted.
>
>> Considering absence of this behavior a bug is fine if you want a
>> million PR's, a lot of discouraged port maintainers
>
> While the PR system is imperfect, it's the best mechanism for
> getting bugs fixed that we have.  Also, recently the pace of
> ports PR commits has picked up; if you look at the PR statistics
> page you'll see this confirmed.
>
> Plus, I don't understand why the above will discourage port
> maintainers.forever
>

The PR system is great.  I didn't mean to imply it wasn't.  I meant
to say that flooding it with bug reports saying "this port deals
with configuration files improperly," leaving a lot of port
maintainers with the responsibility of adopting a defacto-standard
way of dealing with the issue, is not the best way to attack the
problem.

A port maintainer could find it discouraging if that maintainer
didn't know it was a bug in the first place and now has to deal with
the fact that a) it is, and b) how they should fix it given there's
no "official" solution.

> A final note: postings like this don't really create progress.
> Prototyped code submitted via a PR creates progress, or bugfixes
> submitted via a PR create progress, or even bug reports submitted
> via a PR create progress.  Even if you write this up as a "desired
> feature" and submit that as a PR, that would help move things
> forward.
> But just saying "it's broke" without any further action really
> _does_ frustrate folks doing the work.
>
> mcl
>
>
>

Postings like this could create progress if they were read the right
way.  My apologies for the your misinterpretation.  All I gave you
was an opinion.  I didn't give you a desired feature.  The last time
I checked there wasn't an "opinion" mechanism for a PR and if there
was, I'd put much more thought into write one, than I did this post.

I sit here, at night, reading the mailing lists and if I have an
opinion, I share it.  That's all I did.  I do understand your point
and please don't take this post negatively.  I'm just trying to say
that while I have an opinion on this subject I haven't a clear cut
idea on how to best solve it, at this time.  If, in future, I do
formulate something worthy of a PR, I'll generate one and thank-you
for pointing this out.

As for my opinion.  In summary I think:

1. The PR system is great.

2. Asking port maintainers to solve this problem at all is not the
best approach.  Doing it with the -dist file method is also not the
best solution since it can break if the user deletes the -dist files
and you can bank on the fact that many will.  Also it will take a
long time or forver (which ever comes first) to clean them all up
irrespective of how effective the PR system is in reporting the
problem.

3. I think that making the ports system itself deal with the issue
by making it deal with configuration files as a special case has
numerous advantages, solves this problem, makes it easier to solve
others, makes it easier to create and maintain ports and won't
create a pile of PR's flooding the PR system.

--
Adam - (http://people.migus.org/~amigus/)
Migus Dot Org - (http://www.migus.org/)



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?49707.192.168.4.2.1062809044.squirrel>