Re: Lilypond and distribution bugtrackers [was: LSR is not at the stable release level]

[Graham Percival]

On Thu, Oct 04, 2012 at 07:17:05PM +0200, Joseph Rushton Wakeling wrote:
> On 10/04/2012 01:32 AM, Graham Percival wrote:
> >ok, so you won't mind volunteering to take care of this with your
> >personal email account.
> 
> Look, the whole point of what I was suggesting was this: that it
> would be useful for the Lilypond team to be _automatically_ notified
> when a downstream bug gets filed.

That's your opinion.  My opinion is that it would *not* be useful
to receive automatic notifications, since most automatic
notifications from distro bugs are useless for us.

> Not that anyone would have to take responsibility for triaging them
> or responding to them.

The bug square is responsible for responding to *all* emails to
bug-lilypond.  We keep statistics on whether they fulfill that
responsibility or not.

> By putting it through one person and their private account, whether
> it's me or anyone else, all you do is create a bottleneck for that
> information.

But we *want* a bottleneck for that information.  I don't want to
wade through 95% useless messages in order to see the useful 5%.
Given our extremely limited resources for bug handling (resources
which you are not offering to contribute!), I consider losing that
5% to be quite acceptable.

> I'm perfectly happy to track downstream bug reports and write up
> something corresponding here if it seems relevant,

ok.

> but that's not
> solving the problem my suggestion was designed to solve -- that
> there is no _reliable_ channel via which downstream bug reports can
> propagate back to Lilypond.  (Reliable, as in the sense of: the
> message doesn't get lost.)

Not our problem.

> If you don't see that problem as worth solving, well, it's your
> project.

ok, done.  Can we move on to other things now?

> But I don't appreciate
> accusations of laziness or trying to foist work onto other people on
> the basis of a suggestion which was intended to help improve useful
> information flow to the project _without anyone having to do
> anything additional to what they're already doing_.

But it *would* add additional work.  Bug squad members would need
to either reply to the messages or see their evaluated
effectiveness drop.  The bug meister would need to examine all
messages that received no reply and make judgements as to whether
a missed message was important or not.


How's this: if you work on the bug squad for 6 months and still
think that this is a worthwhile project, then we'll talk about it
again.

- Graham