Cleaning out obsolete bugs in bugzilla

Cleaning out obsolete bugs in bugzilla

[Doug Evans]

Another part of bugzilla spring cleaning (though obviously it's an
ongoing task) I'd like to do is establish a process for closing (in
some form) bugs old enough that the likelihood they will ever get
fixed is within epsilon of zero.  And maybe even closing several of
them.  But as along we we can establish a process for efficiently
closing them (i.e., avoiding drawn out discussions on whether they
should have been closed, and the manor in which they were closed),
then I think we can make better progress at clearing them out (or at
least I know I could ... :-)).

An example of the kind of bug I'm thinking of is this one:
https://sourceware.org/bugzilla/show_bug.cgi?id=8769
gdb 6.1 doesn't build on hpux.

I don't have a strong opinion on what flavor of RESOLVED to close
these as, but I do have a strong opinion that we should close them.  I
can't see any sufficiently significant value to warrant keeping them.

Another bug tracking system I've used has the choice of "Obsolete".
Bugzilla doesn't have it by default but I have confirmed that we can add it.
OTOH, using WONTFIX is fine by me too.  Or we could pick something else.
I'm mentioning the existence of the ability to use something other
than WONTFIX in case people have a strong opinion on not using
WONTFIX.

Comments?

Re: Cleaning out obsolete bugs in bugzilla

[Doug Evans]

Hi.  I'll conclude from this that there is no disagreement with
closing such obsolete bugs as WONTFIX is ok.
I'll update the wiki accordingly.

ref:  https://sourceware.org/ml/gdb/2014-02/msg00073.html

---------- Forwarded message ----------
From: palves at redhat dot com <sourceware-bugzilla@sourceware.org>
Date: Wed, Mar 12, 2014 at 3:10 AM
Subject: [Bug build/9101] Can not compile gdb-6.3 on AIX 5.2 with gcc 3.4.4
To: gdb-prs@sourceware.org


https://sourceware.org/bugzilla/show_bug.cgi?id=9101

Pedro Alves <palves at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
                 CC|                            |palves at redhat dot com
         Resolution|---                         |WONTFIX

--- Comment #1 from Pedro Alves <palves at redhat dot com> ---
As this is within bfd/rs6000-core.c, it would really be a binutils bug.  Given
6.3 is old and long gone out of support, and current mainline definitely builds
on AIX, I'm going to close this.  I'm sorry nobody followed up on this when it
was filed...

Re: Cleaning out obsolete bugs in bugzilla

[Pedro Alves]

On 03/12/2014 04:41 PM, Doug Evans wrote:
> Hi.  I'll conclude from this that there is no disagreement with
> closing such obsolete bugs as WONTFIX is ok.

In this case I confirmed mainline builds, so it wasn't exactly
the sort of "close just because it's old" thing.  But yeah.
Old build bugs are fine with me to close.

In this case I went with WONTFIX as meaning "Won't fix in
the 6.3 branch."  But as I confirmed it works on mainline,
we could go with FIXED too.  It's all the same in the end,
but maybe FIXED is kinder to the original reporter?

-- 
Pedro Alves

Re: Cleaning out obsolete bugs in bugzilla

[Doug Evans]

On Wed, Mar 12, 2014 at 9:59 AM, Pedro Alves <palves@redhat.com> wrote:
> On 03/12/2014 04:41 PM, Doug Evans wrote:
>> Hi.  I'll conclude from this that there is no disagreement with
>> closing such obsolete bugs as WONTFIX is ok.
>
> In this case I confirmed mainline builds, so it wasn't exactly
> the sort of "close just because it's old" thing.  But yeah.
> Old build bugs are fine with me to close.
>
> In this case I went with WONTFIX as meaning "Won't fix in
> the 6.3 branch."  But as I confirmed it works on mainline,
> we could go with FIXED too.  It's all the same in the end,
> but maybe FIXED is kinder to the original reporter?

fwiw, I'd hate for these discussions to drag on beyond two emails per
bug in the general case.  There's just too much else to do, and too
many such bugs to deal with.
Thus a high order bit here for me is to find a non-disagreeable way to
streamline the process (and document it so that it can stay
streamlined).

"WONTFIX" could come across with an unintended connotation alright,
but I have no current data to assign a probability to it happening.
At the moment it's just gut feeling, which is why I think there's
benefit to something else that doesn't require much, if any,
discussion, like OBSOLETE.  But I don't have a strong opinion on the
ultimate choice.  If people have a strong opinion on not inventing
anything new, fine by me, let's go with WONTFIX until we have real
data that it can be a problem.

Re: Cleaning out obsolete bugs in bugzilla

[Andreas Schwab]

Pedro Alves <palves@redhat.com> writes:

> In this case I went with WONTFIX as meaning "Won't fix in
> the 6.3 branch."  But as I confirmed it works on mainline,
> we could go with FIXED too.  It's all the same in the end,
> but maybe FIXED is kinder to the original reporter?

IMHO a bug that is fixed on mainline should be marked FIXED with the
target milestone set to the version it was fixed in.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Cleaning out obsolete bugs in bugzilla

[Pedro Alves]

On 03/12/2014 05:15 PM, Andreas Schwab wrote:
> Pedro Alves <palves@redhat.com> writes:
> 
>> In this case I went with WONTFIX as meaning "Won't fix in
>> the 6.3 branch."  But as I confirmed it works on mainline,
>> we could go with FIXED too.  It's all the same in the end,
>> but maybe FIXED is kinder to the original reporter?
> 
> IMHO a bug that is fixed on mainline should be marked FIXED with the
> target milestone set to the version it was fixed in.

I've changed it to FIXED now.

-- 
Pedro Alves

Re: Cleaning out obsolete bugs in bugzilla

[Pedro Alves]

On 03/12/2014 05:10 PM, Doug Evans wrote:

> "WONTFIX" could come across with an unintended connotation alright,
> but I have no current data to assign a probability to it happening.
> At the moment it's just gut feeling, which is why I think there's
> benefit to something else that doesn't require much, if any,
> discussion, like OBSOLETE.  

I think OBSOLETE is a good idea.

> But I don't have a strong opinion on the
> ultimate choice.  If people have a strong opinion on not inventing
> anything new, fine by me, let's go with WONTFIX until we have real
> data that it can be a problem.

-- 
Pedro Alves

Re: Cleaning out obsolete bugs in bugzilla

[Doug Evans]

On Wed, Mar 12, 2014 at 10:23 AM, Pedro Alves <palves@redhat.com> wrote:
> On 03/12/2014 05:10 PM, Doug Evans wrote:
>
>> "WONTFIX" could come across with an unintended connotation alright,
>> but I have no current data to assign a probability to it happening.
>> At the moment it's just gut feeling, which is why I think there's
>> benefit to something else that doesn't require much, if any,
>> discussion, like OBSOLETE.
>
> I think OBSOLETE is a good idea.
>
>> But I don't have a strong opinion on the
>> ultimate choice.  If people have a strong opinion on not inventing
>> anything new, fine by me, let's go with WONTFIX until we have real
>> data that it can be a problem.

Cool, thanks.

OBSOLETE has been added (by overseers) and I've updated the wiki.

https://sourceware.org/gdb/wiki/DeveloperTips#Closing_obsolete_bugs