Re: Closing NEEDSINFO bugs after 30 days

[Harald Sitter]

On Fri, Nov 16, 2018 at 2:40 PM Alexander Potashev  wrote:
>
> пт, 16 нояб. 2018 г. в 16:35, Boudewijn Rempt :
> >
> > On vrijdag 16 november 2018 14:31:39 CET Alexander Potashev wrote:
> >
> > > You only considered to case when human sets NEEDSINFO. If NEEDSINFO is
> > > set by a script, we may have the same problem, would you suggest
> > > implementing the same behaviour?
> >
> > We don't have the script setting NEEDSINFO, do we? At least, I didn't see it
> > in Harald's list of options.
>
> Hopefully we don't. I noticed such change in some bug reports, however
> all of them are attributed to Andrew himself, not the "bug janitor"
> account. Sorry for the noise.

Indeed. There is no such script and there are no plans for such a script.

As mentioned in the other thread: if a bug change/comment was not made
by the janitor account it's not an automated change. A human may still
use boilerplate triage text on multiple bugs mind you :)

HS

Re: Closing NEEDSINFO bugs after 30 days

[Alexander Potashev]

пт, 16 нояб. 2018 г. в 16:35, Boudewijn Rempt :
>
> On vrijdag 16 november 2018 14:31:39 CET Alexander Potashev wrote:
>
> > You only considered to case when human sets NEEDSINFO. If NEEDSINFO is
> > set by a script, we may have the same problem, would you suggest
> > implementing the same behaviour?
>
> We don't have the script setting NEEDSINFO, do we? At least, I didn't see it
> in Harald's list of options.

Hopefully we don't. I noticed such change in some bug reports, however
all of them are attributed to Andrew himself, not the "bug janitor"
account. Sorry for the noise.

-- 
Alexander Potashev

Re: Closing NEEDSINFO bugs after 30 days

[Boudewijn Rempt]

On vrijdag 16 november 2018 14:31:39 CET Alexander Potashev wrote:
 
> You only considered to case when human sets NEEDSINFO. If NEEDSINFO is
> set by a script, we may have the same problem, would you suggest
> implementing the same behaviour?

We don't have the script setting NEEDSINFO, do we? At least, I didn't see it 
in Harald's list of options.

-- 
https://www.krita.org

signature.asc
Description: This is a digitally signed message part.

Re: Closing NEEDSINFO bugs after 30 days

[Alexander Potashev]

пт, 16 нояб. 2018 г. в 16:22, Harald Sitter :
>
> On Fri, Nov 16, 2018 at 2:16 PM Boudewijn Rempt  wrote:
> >
> > On vrijdag 16 november 2018 14:11:29 CET Harald Sitter wrote:
> >
> > > - human sets needsinfo
> > > - 15 days pass: bot posts reminder
> > > - reporter comments
> > > - N days pass: bug gets automatically dropped back to whatever the
> > > previous state was?
> >
> > Yes, that would be perfect! As far as I'm concerned, N can be 0 or 1.
>
> Cool. Should be easy to do.
>
> Any disagreements from the rest of the community?
>
> TLDR: if the original reporter comments on a needsinfo bug it gets
> automatically switched back to whatever the state before needsinfo was

Hi Harald,

You only considered to case when human sets NEEDSINFO. If NEEDSINFO is
set by a script, we may have the same problem, would you suggest
implementing the same behaviour?

-- 
Alexander Potashev

Re: Closing NEEDSINFO bugs after 30 days

[Harald Sitter]

On Fri, Nov 16, 2018 at 2:16 PM Boudewijn Rempt  wrote:
>
> On vrijdag 16 november 2018 14:11:29 CET Harald Sitter wrote:
>
> > - human sets needsinfo
> > - 15 days pass: bot posts reminder
> > - reporter comments
> > - N days pass: bug gets automatically dropped back to whatever the
> > previous state was?
>
> Yes, that would be perfect! As far as I'm concerned, N can be 0 or 1.

Cool. Should be easy to do.

Any disagreements from the rest of the community?

TLDR: if the original reporter comments on a needsinfo bug it gets
automatically switched back to whatever the state before needsinfo was

Re: Closing NEEDSINFO bugs after 30 days

[Boudewijn Rempt]

On vrijdag 16 november 2018 14:11:29 CET Harald Sitter wrote:

> - human sets needsinfo
> - 15 days pass: bot posts reminder
> - reporter comments
> - N days pass: bug gets automatically dropped back to whatever the
> previous state was?

Yes, that would be perfect! As far as I'm concerned, N can be 0 or 1.

-- 
https://www.krita.org

signature.asc
Description: This is a digitally signed message part.

Re: Policy on stale bugs (was Re: Closing NEEDSINFO bugs after 30 days)

[Andrew Crouthamel]

The changes you see were due to me. The Status change was a convenience to me 
for follow up, but as noted, has caused alarm. My apologies for that. I am 
already in process of sending updates and resetting the status, but am doing so 
in blocks to not spam anyone. Give it a few days and all will be well.

In regards to Haralds script, it is fully functional now and appears to work 
very well.

 Original Message 
On Nov 16, 2018, 4:37 AM, Luigi Toscano wrote:

> Andrew Crouthamel ha scritto:
>> I've been spending a lot of time browsing, searching, and filtering our bugs
>> in Bugzilla. One of the areas I've found that could use improvement, are the
>> NEEDSINFO bugs. Often, bugs are placed into this status, either awaiting
>> additional information or backtraces, never to be updated again. We have
>> NEEDSINFO bugs dating back to 2009.
>
> Hi,
> the (semi)automated process which pings and then closes NEEDINFO bugs was
> implemented, but I've noticed another policy which was never discussed (as far
> as I know) here: bugs opened for a while
>
> I disagree with this policy, and I'm not alone (at least another maintainer
> asked for his product to be excluded):
> - not all projects distinguished between CONFIRMED and UNCONFIRMED (now
> REPORTED), so it's possible that a perfectly valid bug or request (especially
> requests) seems stale. You may say that it's easy to flip the status again.
> I'd say that it's a unneeded step.
> - at most the script could add a new comment asking for updates, but not
> immediately change the status out of the blue.
> - as mentioned, it was not discussed.
>
> Please disable it for now, or just enable it for projects who explicitly wants
> it for now.
>
> --
> Luigi

Re: Closing NEEDSINFO bugs after 30 days

[Harald Sitter]

(Pulling this back to the NEEDSINFO thread since it sounds unrelated
to what Luigi was writing about in the Policy on stale bugs thread)

On Fri, Nov 16, 2018 at 12:51 PM Boudewijn Rempt  wrote:
> I know it's well-meant, and it does help to get old bugs resolved -- but it
> really should use a script that checks whether the original reporter has added
> a comment since it was set to needinfo, and in that case just reopen it.
>
> I know that the script that was intended to do that didn't work, and I've
> tried to figure out whether I could fix it, but failed in that. But right now,
> the closing to WORKSFOME plays havoc with my bugzilla routine -- and worse, it
> makes me hesitate to put bugs in the NEEDINFO state.

So... the current behavior is:

- human sets needsinfo
- if no human responds within 15 days since last update the bot posts
a reminder (this can repeat ad infinitum if the status doesn't get
changed away from needsinfo but comments get posted)
- if no human responds within further 15 days the bot closes the bug
(which may also happen if information was provided by the status was
not changed in 30 days)

e.g.

- human set needsinfo
- 15 days pass: bot posts reminder
- reporter comments
- 15 days pass: bot posts reminder
- 15 days pass: bot auto-closes

If I understand your suggestion you'd like it to switch automatically
out of needsinfo upon comment from the original reporter?

e.g.

- human sets needsinfo
- 15 days pass: bot posts reminder
- reporter comments
- N days pass: bug gets automatically dropped back to whatever the
previous state was?

Re: Policy on stale bugs (was Re: Closing NEEDSINFO bugs after 30 days)

[Harald Sitter]

On Fri, Nov 16, 2018 at 10:37 AM Luigi Toscano  wrote:
> Please disable it for now, or just enable it for projects who explicitly wants
> it for now.

As mentioned on IRC: everyone please note that unless a change was
made by bug-jani...@kde.org it was not actually done automatically (or
not that I know of anyway).
To that end please please please do not confuse needsinfo reminders
and closures (which are automated) with changes the bug squad does (a
human doing it). Confusing the two makes it really hard for everyone
to know what we are talking about.

If the janitor does a change that you think is incorrect please poke
me with link as it's possibly a bug in the bot, or may need policy
adjustments.

HS

Re: Policy on stale bugs (was Re: Closing NEEDSINFO bugs after 30 days)

[Christoph Feck]

On 16.11.2018 12:50, Boudewijn Rempt wrote:

 -- and worse, it makes me hesitate to put bugs in the NEEDINFO state.


Worse. Bugs that never have been in NEEDSINFO seem to automatically get 
this state after some time of inactivity now.

Re: Policy on stale bugs (was Re: Closing NEEDSINFO bugs after 30 days)

[Boudewijn Rempt]

On vrijdag 16 november 2018 10:37:08 CET Luigi Toscano wrote:

> the (semi)automated process which pings and then closes NEEDINFO bugs was
> implemented, but I've noticed another policy which was never discussed (as
> far as I know) here: bugs opened for a while
> 
> I disagree with this policy, and I'm not alone (at least another maintainer
> asked for his product to be excluded):
> - not all projects distinguished between CONFIRMED and UNCONFIRMED (now
> REPORTED), so it's possible that a perfectly valid bug or request
> (especially requests) seems stale. You may say that it's easy to flip the
> status again. I'd say that it's a unneeded step.
> - at most the script could add a new comment asking for updates, but not
> immediately change the status out of the blue.
> - as mentioned, it was not discussed.
> 
> Please disable it for now, or just enable it for projects who explicitly
> wants it for now.

I know it's well-meant, and it does help to get old bugs resolved -- but it 
really should use a script that checks whether the original reporter has added 
a comment since it was set to needinfo, and in that case just reopen it.

I know that the script that was intended to do that didn't work, and I've 
tried to figure out whether I could fix it, but failed in that. But right now, 
the closing to WORKSFOME plays havoc with my bugzilla routine -- and worse, it 
makes me hesitate to put bugs in the NEEDINFO state.


-- 
https://www.krita.org

signature.asc
Description: This is a digitally signed message part.

Re: Policy on stale bugs (was Re: Closing NEEDSINFO bugs after 30 days)

[Luigi Toscano]

Luigi Toscano ha scritto:

Andrew Crouthamel ha scritto:
I've been spending a lot of time browsing, searching, and filtering our bugs 
in Bugzilla. One of the areas I've found that could use improvement, are the 
NEEDSINFO bugs. Often, bugs are placed into this status, either awaiting 
additional information or backtraces, never to be updated again. We have 
NEEDSINFO bugs dating back to 2009.


Hi,
the (semi)automated process which pings and then closes NEEDINFO bugs was 
implemented, but I've noticed another policy which was never discussed (as far 
as I know) here: bugs opened for a while

(sorry) ... are switched to NEEDINFO.



I disagree with this policy, and I'm not alone (at least another maintainer 
asked for his product to be excluded):
- not all projects distinguished between CONFIRMED and UNCONFIRMED (now 
REPORTED), so it's possible that a perfectly valid bug or request (especially 
requests) seems stale. You may say that it's easy to flip the status again. 
I'd say that it's a unneeded step.
- at most the script could add a new comment asking for updates, but not 
immediately change the status out of the blue.

- as mentioned, it was not discussed.

Please disable it for now, or just enable it for projects who explicitly wants 
it for now.



--
Luigi

Re: Policy on stale bugs (was Re: Closing NEEDSINFO bugs after 30 days)

[Jaroslaw Staniek]

On Fri, 16 Nov 2018 at 10:37, Luigi Toscano 
wrote:

> Andrew Crouthamel ha scritto:
> > I've been spending a lot of time browsing, searching, and filtering our
> bugs
> > in Bugzilla. One of the areas I've found that could use improvement, are
> the
> > NEEDSINFO bugs. Often, bugs are placed into this status, either awaiting
> > additional information or backtraces, never to be updated again. We have
> > NEEDSINFO bugs dating back to 2009.
>
> Hi,
> the (semi)automated process which pings and then closes NEEDINFO bugs was
> implemented, but I've noticed another policy which was never discussed (as
> far
> as I know) here: bugs opened for a while
>

Hi
As a heavy user of NEEDINFO  I am also not enthusiastic and asked by email
for exclusion but no reaction so far.

Bug reports that I work with can easily be 6 years old and this means
nothing in the development speed that I face. In a commercial environment
maybe it's different - so old reports are really invalid, but here now I
only get more work - I need to resort to notification emails one by one, it
will take months.
Also e.g. if a bug has 10% of missing "NEEDSINFO" and dozen of comments
with valid discussion and input, how can it be abandoned automatically
based on time since the original report?
I am not sure if we can assume it is of any help that valid bugs are set as
resolved by non-project persons (who have never had a contact with a
maintainer), without proper project knowledge.

At least the bugs are not CLOSED so it's possible to query them. But how
useful is to spend time on it?
Please at least do not close them. We have to deal with the consequences
with our users. Possibly facing decline of bug reports because they see
they are resolved without proper action.

I also think something else was not considered, what applies at least to
projects I maintain: people do use old software and do not upgrade
especially on Linux e.g. due to the misunderstood thing called LTS. Hence
they face with problems that are discussed in these old bug reports. The
fact that old software is unmaintained does not mean that the report,
convert into new reports or tasks for newer software versions, for the
docs, and so on. If we automatically set bugs as resolved I seen that as a
disservice for the project.


> I disagree with this policy, and I'm not alone (at least another
> maintainer
> asked for his product to be excluded):
> - not all projects distinguished between CONFIRMED and UNCONFIRMED (now
> REPORTED), so it's possible that a perfectly valid bug or request
> (especially
> requests) seems stale. You may say that it's easy to flip the status
> again.
> I'd say that it's a unneeded step.
> - at most the script could add a new comment asking for updates, but not
> immediately change the status out of the blue.
> - as mentioned, it was not discussed.
>
> Please disable it for now, or just enable it for projects who explicitly
> wants
> it for now.
>
> --
> Luigi
>
>
>
>

-- 
regards, Jaroslaw Staniek

KDE:
: A world-wide network of software engineers, artists, writers, translators
: and facilitators committed to Free Software development - kde.org
KEXI:
: A visual database apps builder - kexi-project.org calligra.org/kexi
  twitter.com/kexi_project facebook.com/kexi.project t.me/kexi_project
Qt Certified Specialist:
: linkedin.com/in/jstaniek 

Policy on stale bugs (was Re: Closing NEEDSINFO bugs after 30 days)

[Luigi Toscano]

Andrew Crouthamel ha scritto:
I've been spending a lot of time browsing, searching, and filtering our bugs 
in Bugzilla. One of the areas I've found that could use improvement, are the 
NEEDSINFO bugs. Often, bugs are placed into this status, either awaiting 
additional information or backtraces, never to be updated again. We have 
NEEDSINFO bugs dating back to 2009.


Hi,
the (semi)automated process which pings and then closes NEEDINFO bugs was 
implemented, but I've noticed another policy which was never discussed (as far 
as I know) here: bugs opened for a while


I disagree with this policy, and I'm not alone (at least another maintainer 
asked for his product to be excluded):
- not all projects distinguished between CONFIRMED and UNCONFIRMED (now 
REPORTED), so it's possible that a perfectly valid bug or request (especially 
requests) seems stale. You may say that it's easy to flip the status again. 
I'd say that it's a unneeded step.
- at most the script could add a new comment asking for updates, but not 
immediately change the status out of the blue.

- as mentioned, it was not discussed.

Please disable it for now, or just enable it for projects who explicitly wants 
it for now.


--
Luigi