[csw-maintainers] amendments and issues with recent prop

[Ben Walton]

Excerpts from Jonathan Craig's message of Fri Jun 24 11:30:00 -0400 2011:

Hi Jonathan,

Thanks for writing such a well thought out response.  I think it
sounds better when it comes from a 'new' voice.  I agree with your
main points and have snipped out anything that you've covered
adequately.  The remainder of this mail deals with the useful items
you've pulled out.

> "A group of subject matter experts will be established to support
> and maintain the release system.  This group will be subordinate to
> the board and responsible for seeing that packaging policy to
> properly enforced by the release systems.  In the event an
> unforeseen issue disrupts the release process they are empowered to
> manually augment the system to enable timely release of packages.
> Such manual intervention will be logged with the understanding that
> repeated manual intervention will require enhancements to the
> release system to accommodate routine exceptions."

I agree with what you say here.  An automated process does not imply a
people-free process.  The tools will evolve and improve over time.  It
will be people that drive the changes.

This would be a good overall change to the proposal and I would
support something like this.  The key points for me are that 'group
will be established' must be defined to allow anyone interested to
participate.  It should not exclude anyone that wants to extend effort
toward making things flow and improving the tools.  It should not by
default give one or two maintainers more power than others.  We'd need
something that encourages more people to become SME's or at the very
least doesn't discourage them.  Setting hard barriers of any sort in
place would be counter to the goals of the new system.

[I can see that you understand this fully, I just want my rationale
here to be crystal clear.]

> I think most would have thought that something like this was
> inherent to any automated solution, but it doesn't hurt to write a
> little verbage to outline expectations.

Agreed.  This comment is well thought out and helpful.  If you'd like
to polish your statement above, I think that it fits well with the
proposal as currently written.  Although I won't speak for the other
signers, I think it would meet with acceptance.

Alternately, as most everyone understands this implicitly it may not
be required.

> handle this bit of nastiness as they see fit.  Personally, having a
> clear process for taking over a package and an automated solution to
> enable "proper" takeover is better than doing this in the darkness
> of night.

Agreed.  This is a tools implementing policy issue.

> dependency.  It would be nice if the maintainer could code these
> dependency into their submit and see that its handled to their
> wishes.  That puts the maintainer in the drivers seat, where the
> responsibility belongs.

I actually had a similar thought earlier today when reading the
graphviz thread.  If we supplied a cswdepend file that held REV= info
in addition to the package name, our tools could look at that and make
smarter choices.  This could allow for mypkg requires yourpkg with
version >= %Y%m%d.  (There are a lot of details in something like this
and it doesn't belong in this thread...)

Thanks!
-Ben
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302