Re: Danger: older bugs are getting squashed with NEEDINFO

[Colin Walters <walters verbum org>]

On Tue, Sep 15, 2009 at 10:11 AM, Owen Taylor <otaylor redhat com> wrote:
>
> There are exceptions to this - if an isolated undiagnosed backtrace with
> no further information provided is still sitting there 3 years later,
> it's unlikely to get useful fixed.

Yeah, the root cause of this as I see it is that autogenerated stack
traces should never have been mixed in with the same UI as carefully
hand written bugs filed by experienced testers and developers.

If we're going to auto-close bugs, limit it unconfirmed stack traces
(and ideally don't generate mail in this case).

Fedora's auto-closing is hugely misguided too and not something to emulate.