Re: Signature for QA?

[Josip Rodin <joy@cibalia.gkvk.hr>]

On Thu, Apr 01, 1999 at 11:59:20AM +1000, Anthony Towns wrote:
> > > forced, or ..
> > >  - dark will remove us from qa-group
> > >  - our debian system will stop working and only scrolls insults
> > >    on the screen
> > >  - ... ?
> > This ain't so funny if you look at the total number of bugs open
> > on Debian packages.
> 
> Well, consider that half a zillion people are already using Debian on
> production systems 24 hours a day, 7 days a week, 52 weeks a year.

So do I.

> Sure, we've got bugs, but we've still got a pretty good system.

I didn't say different.

> The only reason we're *forced* to fix them is our own various obsessions
> with buglessness.

If we're into assuring quality, we must want to be as bugless as possible.
Isn't that obvious? :)

> > Okay, we aren't forced to upload in X days. But we will have the
> > right to upload in X days, and we are forced to do so if we want
> > to really assure the quality of Debian distribution.
> 
> Something like ``If you'd rather, I'd be happy to make a NMU to fix these
> bugs.'' is probably enough.
> 
> For the people who don't reply to their mail, sending a second mail ``Hi,
> I'm not sure if you got my first mail, which said [...]. Anyway, since
> these bugs have been open for a while, and I haven't heard otherwise
> from you, I'll make an NMU of this in about a week.''
               ^^^ , as a member of the quality assurance team, 

Indeed, that is better (softer).

> And it's probably a good idea to check if they've gone on holidays or
> something, and just aren't replying to their email 'til they get back.
> grepping through ~debian/archives/debian-private* on master is probably
> not too bad a way of checking this.

I'm assuming that the person who will contact the maintainer will have
the time to check all of the facts, before even thinking of anything!

That should also be accepted as a mini-policy for QA members:
Read all related bug reports, try to reproduce bugs, try to fix bugs,
try to avoid introducing new bugs with your current solution, check
the age of the bug, check the severity of the bug. Stop. Think. Check
if maintainer is around by all possible means, try all of his e-mail
addresses, ask other people from his area/country about him. Stop. Think.
E-mail him this first message. Wait for some time. E-mail him that second
message. Wait shorter. Announce on debian-qa. Upload 6 hours before the
next dinstall run.

Additions to the list are very welcome.

> But there's no need to start laying down ultimatums in the first email.
> 
> > > and remember to ask if there is a reason to not to fix any of
> > > these bugs, there just might be
> > If there is, it is recorded in the BTS, at bugs.debian.org/<number>.
> > Maintainers should know how to use the BTS.
> 
> But some of them are too busy to get around to it.

But is it *that* complicated to hit a group reply, and to reply to
<number>@bugs when replying to the submitter? Who knows what else
didn't they do because they were to busy.

> It's good to play nice, even if the other guy doesn't deserve it,
> sometimes.

By doing these NMUs we are doing them a favor, by working on their
packages. 

> [0] 5800 and *still* skyrocketting. wtf? Where are all these bugs
>     *going*?? 

ITYM 35800. But it is not that bad, huge number of those are closed,
fixed, forwarded, merged, not really reproducible...

>Anyone want to post a `top 10' list of packages with
>     the most new bugs filed in the last week or so?

You're calling for trouble now :)

-- 
enJoy -*/\*- http://jagor.srce.hr/~jrodin/