Proposal for automation of MIA maintainer detection/processing

Patrice Dumas pertusus at free.fr
Tue Dec 11 16:52:19 UTC 2007


On Tue, Dec 11, 2007 at 11:12:35AM -0500, Jesse Keating wrote:
> I've thrown together a proposal for something we discussed in FESCo
> in the last couple meetings.  FESCo agreed to the idea in principle,
> but obviously wanted a formal proposal to digest.
> 
> http://fedoraproject.org/wiki/JesseKeating/AutomatedMIAProposal
> 
> is said proposal.  I'd really like to get a lot of eyes on this and
> thoughts on this before FESCo pulls the trigger to start seeing some of
> the development on tools done to accomplish this.

This has already been discussed a long time ago, and a difficulty was
that for every kind of issue that could trigger MIA, there can be a
reason why the packager hasn't acted.

In the wiki the end of the Detection is unclear to me:

  Upon subsequent runs of the script if a maintainer triggers as MIA a db
  lookup will be done to determine if the maintainer has already been
  contacted, and has been given the allowed allotment of time to respond.
  If the allotment has been consumed without response, the script will
  kick in to processing mode.

The wording of the beginning of this part is unclear to me, and even
more unclear is whether a human action is needed or not to declare
somebody MIA. My personal point of view is that the final procedure,
that is a reporter declares on devel list that another packager is MIA
and FESCo accepts seems better to me. 

Having automated QA bugs done to detect such MIA maintainer and track 
all the packages at once seems good to me, but as a data gathering 
process, and maybe to avoid for a real person to have to go through 
the first steps of the current MIA policy, but it seems to me that
it is not right if a maintainter may be considered MIA only by automatic
means.

--
Pat




More information about the devel mailing list