An invalidated bug. Should it have really been invalidated?

[Brian Murray]

On Sat, Aug 10, 2013 at 08:32:21PM +0100, Dr. David Alan Gilbert wrote:
> * Chris Johnston (chrisjohnston at ubuntu.com) wrote:
> > I'm not going to comment on keeping the bug opened or closed, however I
> > will point out:
> > 
> > "Many Linux <https://launchpad.net/ubuntu/+source/linux> package, hardware,
> > and other non-user space bugs are hardware dependent on both the hardware
> > itself, and what other hardware the problematic hardware is connected to.
> > The rule of thumb is *one report, per person, per hardware combination, per
> > bug*. For more on this please see
> > here<https://wiki.ubuntu.com/Bugs/BestPractices#X.2BAC8-Reporting.Focus_on_One_Issue>,
> > and further below in this article. As well, please do not post comments to
> > another person's report, claiming you have the same problem, because you
> > have the same hardware, or same computer model. Instead, please file a
> > separate report, and make comments there. This is because no one can verify
> > if you would have the same problem or not, because your hardware can not be
> > analyzed. "
> 
> Right, the challenge is how to group them though - if you think 20 bugs
> are actually the same bug:
>    * how do we recognise that?
>      - because it is useful to group when you think you've found the same
>        thing
>    * how do we keep hold of all the debug if the original reporter gives
> up and moves on?
>      - we do see streams of bugs where one person finds it, gives up
>        and the bug expires, then someone else finds it 6 months later
>        and the same bug persists for years.
> 
> The challenge with kernel bugs though is finding the point at which
> you say it's actually the same problem, and it can be really tricky which
> is I think why those guidelines are there; you don't want to start grouping
> stuff together just because it has similar symptoms though unless
> it's specific enough to be reasonably sure.
> 
> The other thing I sometimes do is ask people to put another bug report
> in, but to add a comment on the original giving the number of their
> new report as a minimal sort of linking.

This is a great best practice as it makes it easier to find bugs
possibly related to the one a developer is fixing.  Just a simple
comment like "I reported bug 1234 which I think is related to this one
as I have similar hardware" can be a huge help.

--
Brian Murray