You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openoffice.apache.org by Rainer Bielefeld <ra...@bielefeldundbuss.de> on 2014/02/06 10:52:51 UTC
Bugzilla: Keyword "regression"
Hi,
currently the manual on
<https://issues.apache.org/ooo/describekeywords.cgi> seems to suggest to
use that keyword also for problems what appeared between OOo 1.0.2 and
OOo 1.0.3. Such use might be correct as a matter of form, but I think
useless for QA and developers daly work.
"regression" key word indicates a major priority for fixing a bug,
because appearance of new problems is bad for AOO's reputation. But a
regression between OOo 1.0.0 and OOo 1.0.1 from 2001 (as an overstated
example) is nothing what should cause major reputation loss for AOO in 2014.
So I recommend to limit use of Keyword "regression" for all Problems
what appeared with AOO 3.4-dev or later and to add a corresponding hint
in Bugzilla Help.
BTW: I recommend to keep those hints in Bugzilla short (not more than 2
text rows or so) and do link to more elaborated help in the Wiki (for
example).
Best regards
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: Bugzilla: Keyword "regression"
Posted by Rainer Bielefeld <ra...@bielefeldundbuss.de>.
Rob Weir schrieb:
> I wonder whether a combination of the regression flag and the version
> field would give us everything we need to know?
Hi,
that's the main intentions behind my suggestion. If we know we have had
a regression (not too long ago, with "Regression keyword) and also have
found the version where it appeared, we can find the commit what causes
the problem. With reasons for the commit fresh in their mind developers
should be able to fix that quickly. With old regressions that's more
difficult, nobody remembers why the code changes have been done and what
later commits' functions will break after changes in that area.
So the "regression" keyword can be a flag showing that here is an
annoying mishap what can be fixed with probably bearable "costs".
CU
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: Bugzilla: Keyword "regression"
Posted by Rainer Bielefeld <ra...@bielefeldundbuss.de>.
Rob Weir schrieb:
> I wonder whether a combination of the regression flag and the version
> field would give us everything we need to know?
Hi,
that's the main intentions behind my suggestion. If we know we have had
a regression (not too long ago, with "Regression keyword) and also have
found the version where it appeared, we can find the commit what causes
the problem. With reasons for the commit fresh in their mind developers
should be able to fix that quickly. With old regressions that's more
difficult, nobody remembers why the code changes have been done and what
later commits' functions will break after changes in that area.
So the "regression" keyword can be a flag showing that here is an
annoying mishap what can be fixed with probably bearable "costs".
CU
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
For additional commands, e-mail: qa-help@openoffice.apache.org
RE: Bugzilla: Keyword "regression"
Posted by Amin,
AS...@nike.apache.org,
RUS,
Ç...@nike.apache.org,
IBMM <am...@mm.pl>.
Rob, 'Not' for your enquire.
No N-1 but N+1 so better.
;
-----Original Message-----
From: Rob Weir [mailto:robweir@apache.org]
Sent: 06 February, 2014 3:35 PM
To: qa@openoffice.apache.org
Subject: Re: Bugzilla: Keyword "regression"
On Thu, Feb 6, 2014 at 5:12 AM, Edwin Sharp <el...@mail-page.com> wrote:
> I'm against this recommendation.
> Nothing more harmful to AOO reputation than an old unanswered problem because:
> Lack of basic respect to the person who took the time to report it.
> An indication for no control over the project.
>
I think the primary sorting mechanism for fixes is based on user
impact, i.e., the severity of the issue, how many users will run into
it, and whether there are adequate workarounds. An issue is not
necessarily more important just because it is old.
Regressions are special for a different reason. A regression could be
high severity or low severity. But a regression is caused by a recent
code change. So with a regression we're more likely to have a
developer active in the project who caused the defect, who knows that
area of the code well and is able to fix it quickly. It is easiest to
fix a new bug quickly than to investigate it years later.
I wonder whether a combination of the regression flag and the version
field would give us everything we need to know? If something is
marked as found in version N and marked as regression then that tells
us that the regression occurred between version N-1 and N, yes?
-Rob
> On Thu, Feb 6, 2014, at 11:52, Rainer Bielefeld wrote:
>> Hi,
>>
>> currently the manual on
>> <https://issues.apache.org/ooo/describekeywords.cgi> seems to suggest to
>> use that keyword also for problems what appeared between OOo 1.0.2 and
>> OOo 1.0.3. Such use might be correct as a matter of form, but I think
>> useless for QA and developers daly work.
>>
>> "regression" key word indicates a major priority for fixing a bug,
>> because appearance of new problems is bad for AOO's reputation. But a
>> regression between OOo 1.0.0 and OOo 1.0.1 from 2001 (as an overstated
>> example) is nothing what should cause major reputation loss for AOO in 2014.
>>
>> So I recommend to limit use of Keyword "regression" for all Problems
>> what appeared with AOO 3.4-dev or later and to add a corresponding hint
>> in Bugzilla Help.
>>
>> BTW: I recommend to keep those hints in Bugzilla short (not more than 2
>> text rows or so) and do link to more elaborated help in the Wiki (for
>> example).
>>
>> Best regards
>>
>> Rainer
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
>> For additional commands, e-mail: qa-help@openoffice.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: qa-help@openoffice.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
For additional commands, e-mail: qa-help@openoffice.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
For additional commands, e-mail: qa-help@openoffice.apache.org
Re: Bugzilla: Keyword "regression"
Posted by Rob Weir <ro...@apache.org>.
On Thu, Feb 6, 2014 at 5:12 AM, Edwin Sharp <el...@mail-page.com> wrote:
> I'm against this recommendation.
> Nothing more harmful to AOO reputation than an old unanswered problem because:
> Lack of basic respect to the person who took the time to report it.
> An indication for no control over the project.
>
I think the primary sorting mechanism for fixes is based on user
impact, i.e., the severity of the issue, how many users will run into
it, and whether there are adequate workarounds. An issue is not
necessarily more important just because it is old.
Regressions are special for a different reason. A regression could be
high severity or low severity. But a regression is caused by a recent
code change. So with a regression we're more likely to have a
developer active in the project who caused the defect, who knows that
area of the code well and is able to fix it quickly. It is easiest to
fix a new bug quickly than to investigate it years later.
I wonder whether a combination of the regression flag and the version
field would give us everything we need to know? If something is
marked as found in version N and marked as regression then that tells
us that the regression occurred between version N-1 and N, yes?
-Rob
> On Thu, Feb 6, 2014, at 11:52, Rainer Bielefeld wrote:
>> Hi,
>>
>> currently the manual on
>> <https://issues.apache.org/ooo/describekeywords.cgi> seems to suggest to
>> use that keyword also for problems what appeared between OOo 1.0.2 and
>> OOo 1.0.3. Such use might be correct as a matter of form, but I think
>> useless for QA and developers daly work.
>>
>> "regression" key word indicates a major priority for fixing a bug,
>> because appearance of new problems is bad for AOO's reputation. But a
>> regression between OOo 1.0.0 and OOo 1.0.1 from 2001 (as an overstated
>> example) is nothing what should cause major reputation loss for AOO in 2014.
>>
>> So I recommend to limit use of Keyword "regression" for all Problems
>> what appeared with AOO 3.4-dev or later and to add a corresponding hint
>> in Bugzilla Help.
>>
>> BTW: I recommend to keep those hints in Bugzilla short (not more than 2
>> text rows or so) and do link to more elaborated help in the Wiki (for
>> example).
>>
>> Best regards
>>
>> Rainer
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
>> For additional commands, e-mail: qa-help@openoffice.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: qa-help@openoffice.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
For additional commands, e-mail: qa-help@openoffice.apache.org
Re: Bugzilla: Keyword "regression"
Posted by Edwin Sharp <el...@mail-page.com>.
I'm against this recommendation.
Nothing more harmful to AOO reputation than an old unanswered problem because:
Lack of basic respect to the person who took the time to report it.
An indication for no control over the project.
On Thu, Feb 6, 2014, at 11:52, Rainer Bielefeld wrote:
> Hi,
>
> currently the manual on
> <https://issues.apache.org/ooo/describekeywords.cgi> seems to suggest to
> use that keyword also for problems what appeared between OOo 1.0.2 and
> OOo 1.0.3. Such use might be correct as a matter of form, but I think
> useless for QA and developers daly work.
>
> "regression" key word indicates a major priority for fixing a bug,
> because appearance of new problems is bad for AOO's reputation. But a
> regression between OOo 1.0.0 and OOo 1.0.1 from 2001 (as an overstated
> example) is nothing what should cause major reputation loss for AOO in 2014.
>
> So I recommend to limit use of Keyword "regression" for all Problems
> what appeared with AOO 3.4-dev or later and to add a corresponding hint
> in Bugzilla Help.
>
> BTW: I recommend to keep those hints in Bugzilla short (not more than 2
> text rows or so) and do link to more elaborated help in the Wiki (for
> example).
>
> Best regards
>
> Rainer
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: qa-help@openoffice.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: qa-unsubscribe@openoffice.apache.org
For additional commands, e-mail: qa-help@openoffice.apache.org