Re: string number problem

[Graham Percival]

On Tue, 19 Aug 2008 00:27:50 +0100
"Neil Puttock" <address@hidden> wrote:

> 2008/8/18 Graham Percival <address@hidden>:
> 
> > I agree, but the instructions were explicit: if the output looks
> > good, don't add any bugs about warning messages.
> 
> That's OK so long as there's someone with a trained eye available to
> check that the output's good.

I used to reject reports that didn't explain what was wrong if I
couldn't figure it out -- happened often with lyrics or ancient music
(since I have a completely untrained eye in those subjects). That's one
reason I was such a stickler for minimal reports.

> > Now that you're here, maybe we can get those instructions changed.
> > :)
> 
> Well, I suppose the most important thing is to check before adding
> anything to the bug tracker; hopefully that way we can sort the wheat
> from the chaff by intercepting invalid bugs.

Oh, of course -- and that's why we don't let random people submit
items directly to the tracker.  Other than half a dozen people,
reports should go to the bug mailist, and Valentin makes sure that
the report contains a minimal example that demonstrates the bug.

It's just that the current policy is to reject bug reports about a
false warning message.  If somebody is willing to investigate
these false warning messages and fix stuff, then maybe it's worth
changing this policy.

Cheers,
- Graham