Re: critical issues

[Graham Percival]

On Wed, Dec 29, 2010 at 12:32:56PM -0000, Phil Holmes wrote:
> From: "Carl Sorensen" <address@hidden>
> >On 12/28/10 4:18 PM, "Graham Percival" <address@hidden> wrote:
> >
> >The difference between Phil's version and the previous version is
> >
> >"Something that worked as it should in a previous version, and now doesn't
> >work." vs.
> >
> >"Something that worked intentionally in one of the previous two stable
> >versions, and now doesn't work."
> 
> That's about it.

I want to keep the word "intentionally", though -- if something
only happened to work because of a happy coincidence of bugs, then
"breaking" that should not be a Critical bug.

>  Graham wants a regression from (say) 2.6 not to
> count as critical.  I thought this was somewhat confusing to new
> bug-squadders to understand and wanted all regressions to count the
> same.

There's two questions here:
1. what do we want to teach new bug squad members / what do we
want bug squad members to do?  (IMO these should be the same
thing)
2. what should be the actual priority of a particular issue.

The answer to these need not be the same thing.  In particular,
I'd fine telling / expecting bug squad members to make something
Critical if it's a regression, period.  Just as long as a
programmer can come by and say "wait, we broke that back in
2005... sure, it's still a bug, but we're not going to delay a
stable release for it.  I'm setting it to priority-medium".


Remember the goal of the bug squad is to handle routine
administrative tasks so that programmers can concentrate on more
advanced tasks.  If a non-routine task is causing confusion to bug
squad members, or if the administration is causing a burden on
programmers, then it's time to reconsider stuff.


> IMO the question is likely to be moot anyway, since I think
> it's quite unlikely we'll see such a regression reported.

Based on past experience, I'd say there's a >50% chance of seeing
at least two reports of such regressions over the next year.

> >>>  TBH I was uncertain whether to add this to the tracker
> >>> at all and only did so because I had a few attempts at getting views
> >>> on whether it was a problem, with no response.  I probably labelled
> >>> it a regression, but possibly wrongly.
> >>
> >> My thought process would be this:
> >> 1. am I certain that the new position is ok?  If so, do nothing.
> >> 2. am I certain that the new position is not ok?  If so, add it as
> >> a Critical issue with a brief description of the problem.
> >> 3. am I not certain either way?  If so, add it as a Critical
> >> issue, but note that it may or may not be an actual problem.
> 
> 
> I'd like to add this guidance to the CG, too.

Sure, sounds fine.

Cheers,
- Graham