[Libreoffice-qa] Stats I plotted + Getting rid of bugs

[Bjoern Michaelsen]

On Fri, Nov 02, 2012 at 01:57:27PM -0700, bfo wrote:
> There was one autoclosing nightmare recently and I thought no one dare to
> propose another one any time soon... The problem is that most of bugs are
> not checked by anyone. As my experience shows - one manual ping can resolve
> most issues or transfer the bug to developers. We need more people to read
> bugs and process them, not more ideas how to autoclose them. 

The main problem with the last autoclose was that we had the data in an
inconsistent state, as stuff moved from NEW to NEEDINFO without human
interaction and then again from NEEDINFO to INVALID. That should not happen
again as we have the state UNCORNFIRMED now.

I dont quite get what Florian is aiming for. Flo, could you please explain
which bugs you want to close (not with a query, but with words and an
explanation why)?

Closing bugs because there was no version specified is not an option -- that is
not to be expected from the reporter as it is nontrivial for him to check.

However, these (currently 279 bugs) have been in NEEDINFO for 3 month and have
not seen a reply since then:

 https://bugs.freedesktop.org/buglist.cgi?n2=1&f1=longdesc&list_id=156984&o1=changedafter&o2=changedafter&n1=1&query_format=advanced&f2=bug_status&bug_status=NEEDINFO&v1=2012-08-01&v2=2012-08-01&product=LibreOffice

It is ok to close those as INVALID IMHO. But: this needs to be well prepared,
the message going with it should be discussed here on the list. We should also
sent a warning a la "This bug has been in NEEDINFO for more than three month
without a comment adding new information. This means we assume that we do not
have enough information to carry forward solving your problem. Please provide
the missing information and set the bug to state NEW. If nothing is added to
this bug in the next 14 days, it will be closed as invalid." to these bugs
before closing them.

We might get 10% reopened with the missing info (which is good) and get rid of
90%, who would have never completed their bug anyway (which is good in a minor
way too).

Best,

Bjoern