Re: [jdt-dev] Strategy for JDT core issues without sufficient info?

[S A <simeon.danailov.andreev@xxxxxxxxx>]

> Any obvious mark "need info" may at least help us, to not waste too much time on the same issues time and again.

This is actually my main concern. I don't mind having open bugs that cannot be worked on yet, but it would be good to filter them out when looking for something that can be worked on right away (e.g. a reproducer is available).

On Wed, 29 Mar 2023 at 19:46, Stephan Herrmann <stephan.herrmann@xxxxxxxxx> wrote:

Good question.

The extremes are easy:

(A): Reporter writes "it doesn't work", we ask them dozens of questions
to gain a minimum understanding of what they are upset about, but don't
get any reasonable answer. => Close cannot reproduce.

(B): Reporter gives a failing JUnit. Bingo => Fix it! :)

It's the middle ground where it gets murky. Reporter has invested
efforts to describe the problem and under which conditions it occurs.
Yet we are unable to reproduce. Is the description insufficient? Are we
just not creative enough to reproduce? Perhaps there are several
reporters observing the same problem. Can we still shrug it off? Can we
provide help so that reporters can narrow it down more effectively?
Generic help on how to reduce a real world reproduces to something they
can share? Specific help, e.g., when the stack trace tells us what they
should be looking for in their workspace?

I don't have the answers, prioritizing bug reports is a hard problem,
and tons of stale reports certainly aggravate the situation. Any obvious
mark "need info" may at least help us, to not waste too much time on the
same issues time and again. On the other end of the spectrum aggressive
auto-closing has caused a lot of bad blood already.

A good mix of empathy, common sense and efficiency is the best I could
recommend,
Stephan

On 29.03.23 12:36, S A wrote:
> Hi all,
>
> TL;DR:
> What is our strategy for JDT core issues without sufficient info?
>
> I'm going over JDT core issues and I'm encountering some (not just a
> few) which don't have sufficient infos to work on the problem. The
> issue reporter has been asked to provide infos, but has not done so.
>
> When e.g. there is some thrown exception and a stack trace is
> attached, a maintainer can attempt to create a reproducer. Failing
> this, they'll ask the reporter for a reproducer.
>
> In some issues not even the above is the case, the description is
> simply not enough to do anything about the ticket.
>
> From what I see, such issues are just left open in the tracker. With
> limited resources this is understandable. But ideally such issues
> would be marked with some tag (e.g. "need info"), so they can be
> filtered and looked into last (as they would cost the most time to
> work on). Or such issues can be outright closed when there is no
> further infos from the reporter, despite requesting infos and not
> getting any in a reasonable amount of time (e.g. 2 weeks or so).
>
> Best regards and thanks,
> Simeon
>
> _______________________________________________
> jdt-dev mailing list
> jdt-dev@xxxxxxxxxxx
> To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/jdt-dev
_______________________________________________
jdt-dev mailing list
jdt-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/jdt-dev