Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rainer Bielefeld]

Hi,

just for the records:

More or less with the beginning of April, I started to check Stagnant 
NEEDINFO bugs without change for more than 6 months where I am involved. 
My goal is to handle at least 5 bugs per week.



Statistics  2013-04-17 Number Percent  comment
--
total No of those bugs where I am involved 78
proceeded  34  44%   of total
New results, so worthwhile engagement   8  24% of proceeded


Best regards

Rainer
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Florian Reisinger]

Hi Petr, all,

Thanks for your long mail...


Am 08.02.2013 um 16:56 schrieb Petr Mladek :

> Florian Reisinger píše v Čt 07. 02. 2013 v 16:36 +0100:
>
>> 71,43% NO INPUT
>> 28,57% SHOULD BE REVIEWED
>
> I think what might be a globally acceptable solution for our problem:
>
>+ 20%-30% of wrongly closed bugs is relatively high number. It
>  is realistic. I think that it corresponds with the number of
>  reopened bugs from the first two mass closes

I think you agree, that you shouldn't take the percentage too serious,,,
7 samples are far to less for a meaningful statistic...

+ If someone isn't interested for 180 days not every bug will be
reopened AND because of the Dead bugs only a small number of the "no
input" bugs (IMHO) are going to get reopened...

>
>+ few active bug triaggers (Rainer, Alex, ???) are against
>  automatic mass close
>
>+ most people agree on closing dead bugs
>
> So what are the numbers:
>
>+ 785 bugs is NEEDINFO more than 1 month
>+ 532 bugs in NEEDINFO more than 3 months
>+ 328 bugs in NEEDINFO more than 6 months
>+ 1456 bugs is UNCONFIRMED and needs triage
>+ 270 bugs is REOPENED and might need triage
>
> Let's be pessimistic and say that only 2/3 of the NEEDINFO bugs are dead
> and the rest will get reopened => the mass close will get rid of:

Let's be a little bit less pessimistic: Some simply forgot about the
submitted bug...


>
>+ 523 bugs if we close after 1 month
>+ 354 bugs if we close after 3 months
>+ 218 bugs if we close after 6 months
>
> 3 months is a good compromise that should be acceptable for most people.
> If we decide to use this limit, then would get:
>
>+ 354 automatically closed bugs (2/3 of 532)
>+ 447 REOPENED bugs (270 + 1/3 of 532)
>+ 1456 UNCONFIRMED bugs
>=
>+ 354 closed bugs
>+ 1903 bugs needing triage (1456+447)
>
> So, the mass change would solve about 15% of the bugs. On the other
> hand,  it might demotivate some active triagers and make some users
> angry.
>
> Resume:
> ===
>
> I see two solutions.
>
> 1. Majority of people agrees that 15% is a nice win and we will do the
>   mass change.

Nice win and IMHO the win will be higher...

>
> 2. We will close the dead bugs manually:
>
>   + add query for stalling NEEDINFO bugs and ask for triage
>   + encourage triagers to close dead bugs older than X months
>   + propose a good closing text on the wiki

We (QA) have one problem : Number of active people ( in relation to
the number of bugs..,)
For triaging the NEEDINFO bugs we IMHO need 1/3 - 1/2 more "staff".


>
> The risk is that triagers would feed bad to close dead bugs. The
> advantage of the mass close is that it is kind of annonymous.

Personally I don't think so...


>
> After all, I slightly prefer the second solution. It should not be much
> work to close dead bugs, especially in compare with the work that has
> already been invested into these bugs. It is more clean and should be
> better acceptable for all involved, especially from the long term point
> of view.
>
> What do you think?

I pointed out my opinion above...
>
>
> Best Regards,
> Petr
>

Liebe Grüße, / Yours,
Florian Reisinger
> ___
> List Name: Libreoffice-qa mailing list
> Mail address: Libreoffice-qa@lists.freedesktop.org
> Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
> Problems? 
> http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
> Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
> List archive: http://lists.freedesktop.org/archives/libreoffice-qa/
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Florian Reisinger píše v Čt 07. 02. 2013 v 16:36 +0100:

> 71,43% NO INPUT
> 28,57% SHOULD BE REVIEWED

I think what might be a globally acceptable solution for our problem:

+ 20%-30% of wrongly closed bugs is relatively high number. It
  is realistic. I think that it corresponds with the number of
  reopened bugs from the first two mass closes

+ few active bug triaggers (Rainer, Alex, ???) are against
  automatic mass close

+ most people agree on closing dead bugs

So what are the numbers:

+ 785 bugs is NEEDINFO more than 1 month
+ 532 bugs in NEEDINFO more than 3 months
+ 328 bugs in NEEDINFO more than 6 months
+ 1456 bugs is UNCONFIRMED and needs triage
+ 270 bugs is REOPENED and might need triage

Let's be pessimistic and say that only 2/3 of the NEEDINFO bugs are dead
and the rest will get reopened => the mass close will get rid of:

+ 523 bugs if we close after 1 month
+ 354 bugs if we close after 3 months
+ 218 bugs if we close after 6 months

3 months is a good compromise that should be acceptable for most people.
If we decide to use this limit, then would get:

+ 354 automatically closed bugs (2/3 of 532)
+ 447 REOPENED bugs (270 + 1/3 of 532)
+ 1456 UNCONFIRMED bugs
=
+ 354 closed bugs
+ 1903 bugs needing triage (1456+447)

So, the mass change would solve about 15% of the bugs. On the other
hand,  it might demotivate some active triagers and make some users
angry.

Resume:
===

I see two solutions.

1. Majority of people agrees that 15% is a nice win and we will do the
   mass change.

2. We will close the dead bugs manually:

   + add query for stalling NEEDINFO bugs and ask for triage
   + encourage triagers to close dead bugs older than X months
   + propose a good closing text on the wiki

The risk is that triagers would feed bad to close dead bugs. The
advantage of the mass close is that it is kind of annonymous.

After all, I slightly prefer the second solution. It should not be much
work to close dead bugs, especially in compare with the work that has
already been invested into these bugs. It is more clean and should be
better acceptable for all involved, especially from the long term point
of view.

What do you think?


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Robert Großkopf]

Hi Petr,

> But maybe you know some Base users from successful or bigger
> companies that might afford to pay someone for fixing a nasty bug. Or
> maybe you know someone with developer capabilities who is interested
> into Base. Then you could encourage them to contribute.
> 
> My native language is not English. I am not a politician. My sentences
> might be sometimes misleading or unclear.

Now I understand, what you mean. Better to write it down, when we don't
understand another, than to be angry in silent - and the opposite person
doesn't meant it the same way I understood.

My native language isn't English, too. Everybody, who reads what I'm
writing, would know, that English-lessons in scool were the worst case
for me.

Let's see, if we could get anybody to help close Base-bugs. And if
somebody could donate for fixing a Base-bug it will be OK.

Regards,

Robert
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Robert Großkopf píše v Čt 07. 02. 2013 v 19:52 +0100:
> Hi *,
> > 
> > + https://bugs.freedesktop.org/show_bug.cgi?id=52571
> > + BASE VBA DELETE COLUMN OR TABLE DOESN'T CHANGE THE SURFACE
> > + minimalistic report => hard to understand
> > + 5 months ago asked for more details and no answer
> > + 2 months ago pinged 
> > => should get closed
> 
> Have closed this bug. Don't know if this bug is really confirmed by
> everybody.

Great.

> > + https://bugs.freedesktop.org/show_bug.cgi?id=54730
> > + LibreOffice Base crashes when Test connection is executed
> >   against JDBC connection
> > + problems to reproduce
> > + 4 months in needinfo
> > + info kind of provided but still not reproducible
> > => should get closed as worksforme
> 
> There seems to be nobody who could reproduce this bug. Nobody of the
> persons, who answered and tested have used the database the reporter
> uses. If there is anybody, who could test this bug with the cimbination
> JDBC/DB2 and this combination works, then the bug could be closed as
> Worksforme.
> Thats the problem of many open bugs auf Base, which couldn't be
> confirmed: You could connect to many databases. Special bugs of external
> databases would wait a long time as Unconfirmed, because noby uses the
> same database.

Well, crasher can be solved also without reproducing. The backtrace
usually shows where the problem is.

The user had problems with JDBC connections in general. The crash was
probably visible only in the special environment. I am afraid that
backtrace was the only solution to track this down effectively.

> It is wrong, that this bug has status "Needinfo", because it has never
> been confirmed.

Well, it was in neededinfo because Jochen asked you for input. It is
kind of correct using of the NEEDINFO flag ;-)

IMHO, the most correct solution would be ask the user for a backtrace.
It might point to the broken code. Also, we might ask for more
information about his system, for example JRE and JDBC version. It might
trigger a bell and point to another known bug. If he does not provide it
than we should close it as WORKSFORME.

> We shouldn't close Base-bugs which cant be confirmed because of special
> databases or special systems.

I agree that we should not close bugs just because they need special
equipment. Well, such bugs should not be in NEEDINFO state.

Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Robert Großkopf píše v Čt 07. 02. 2013 v 20:08 +0100:
> Hi Petr,
> > 
> > Also you could pay someone to work on a certain bug. There is a growing
> > list of certified developers which are capable of doing such things.
> > These are well spend money because they improve the product and motivate
> > people working on LO.
> 
> Don't know, if you know, to whom you are writing this. Or could be I
> don't understand enough English. Dan is the person, who writes the
> international Base-Guide. I have written the German "Base-Handbuch".
> We both spend a lot of time for helping other people to understand the
> database included in LO - by writing Guides, by helping in a forum or in
> mailinglists. We motivate people to work with Base.
> ... and whe souldn't only spend a lot of time to motivate. Now we should
> also spend money???

First, Robert, Dan, others on this list, your contribution is really
appreciated and I am happy that you are around.

Second, this is misunderstanding. My paragraph was not meant for any
particular person, definitely. There was no name in my mail and I
explicitly  removed all people from TO and CC.

It is great that you motivate people for using Base. Though, we also
need to motivate people to develop Base. One of the motivation is money.
It does not mean that you (Robert, Dan, others on this list) should pay
the money. But maybe you know some Base users from successful or bigger
companies that might afford to pay someone for fixing a nasty bug. Or
maybe you know someone with developer capabilities who is interested
into Base. Then you could encourage them to contribute.

My native language is not English. I am not a politician. My sentences
might be sometimes misleading or unclear.

I am sorry for confusion. I hope that it is more clear now.

Best Regards.
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Robert Großkopf]

Hi Petr,
> 
> Also you could pay someone to work on a certain bug. There is a growing
> list of certified developers which are capable of doing such things.
> These are well spend money because they improve the product and motivate
> people working on LO.

Don't know, if you know, to whom you are writing this. Or could be I
don't understand enough English. Dan is the person, who writes the
international Base-Guide. I have written the German "Base-Handbuch".
We both spend a lot of time for helping other people to understand the
database included in LO - by writing Guides, by helping in a forum or in
mailinglists. We motivate people to work with Base.
... and whe souldn't only spend a lot of time to motivate. Now we should
also spend money???
No, I don't need Base in newer versions of LO. I 'm still using LO
3.3.4, because this is the best working version of LO, when I try to
work with Base.
My own databases I use for work aren't Base-databases. I'm working with
PHP/Apche/MySQL.

Robert
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Robert Großkopf]

Hi *,
> 
> + https://bugs.freedesktop.org/show_bug.cgi?id=52571
>   + BASE VBA DELETE COLUMN OR TABLE DOESN'T CHANGE THE SURFACE
>   + minimalistic report => hard to understand
>   + 5 months ago asked for more details and no answer
>   + 2 months ago pinged 
> => should get closed

Have closed this bug. Don't know if this bug is really confirmed by
everybody.
> 
> + https://bugs.freedesktop.org/show_bug.cgi?id=54730
>   + LibreOffice Base crashes when Test connection is executed
>   against JDBC connection
>   + problems to reproduce
>   + 4 months in needinfo
>   + info kind of provided but still not reproducible
>   => should get closed as worksforme

There seems to be nobody who could reproduce this bug. Nobody of the
persons, who answered and tested have used the database the reporter
uses. If there is anybody, who could test this bug with the cimbination
JDBC/DB2 and this combination works, then the bug could be closed as
Worksforme.
Thats the problem of many open bugs auf Base, which couldn't be
confirmed: You could connect to many databases. Special bugs of external
databases would wait a long time as Unconfirmed, because noby uses the
same database.
It is wrong, that this bug has status "Needinfo", because it has never
been confirmed.

We shouldn't close Base-bugs which cant be confirmed because of special
databases or special systems.

Robert
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Florian Reisinger]

Hi,

I do something similar with my "kill list" [1]

Am 07.02.2013 15:31, schrieb Petr Mladek:

Alex Thurgood píše v Čt 07. 02. 2013 v 14:17 +0100:

Le 07/02/2013 13:32, Petr Mladek a écrit :

[...]

Do you have other opinion, feeling, or experience, please?

Best Regards,
Petr

I took 7 bugs out of the 256 given ( + used for pro closing - used 
for not closing * for neutral)


#1 https://bugs.freedesktop.org/show_bug.cgi?id=50749

+ Short description
+ NOT reproduced by Rainer
+ No activity since 2012-07-28 21:32:17 CEST

#2 https://bugs.freedesktop.org/show_bug.cgi?id=41488

* Reported for 3.3.3
* Checked with 3.4.4
+ Last reporter input: 2012-02-05 01:15:26 CET
+ Last input: 2012-06-26 08:14:48 CEST

#3 https://bugs.freedesktop.org/show_bug.cgi?id=42327

- Critical bug
+ As far as I understood not reproduced
+ No input after: 2012-05-02 04:10:05 CEST

#4 https://bugs.freedesktop.org/show_bug.cgi?id=49544

- Needs for investigation (Sorry, don't have the time for it today
- IMHO false state...
+- (Couldn't decide) No activity since 2012-07-18 22:03:23 CEST


#5 https://bugs.freedesktop.org/show_bug.cgi?id=45980

--WHY THE HELL IS THIS BUG IN 
NEEDINFO STATE

- Investigation needed
-Seems, that noone touched it
* Last (and first) activity: 2012-02-13 02:07:09 CET

#6
* Reported with 3.4.5 and tested with 3.5.4 too
+ Request for info --> NO answer
+ Last (and only) activity (reporter): 2012-03-13 07:19:32 CET
+ Last activity (QA): 2012-08-09 14:10:03 CEST

#7
* Bugreport for 3.5.1
+ No further input by user
+ Asked for retesting
+ Last activity: 2012-06-12 04:22:51 CEST



Finally:

Should be...   #
CLOSED 5 ( 1,2,3,6,7 )
OPEN2 (4,5)

71,43% NO INPUT
28,57% SHOULD BE REVIEWED

This may mean, that if there is an answer, about 73 bugs should be 
reviewed and 183 are valid closed. 36 bugs are definitely in the wrong 
state --> BUT please test this query. No input for 180 days AND NEEDINFO 
--> Needinfo is the wrong state anyway IMHO ;)


My thoughts...



[1] 
https://bugs.freedesktop.org/buglist.cgi?f1=days_elapsed&o3=notsubstring&list_id=226805&v3=strike1&columnlist=component%2Cassigned_to%2Cbug_status%2Cbug_severity%2Cresolution%2Cshort_desc%2Cchangeddate&o1=greaterthan&query_based_on=DELETE%20NEEDINFO%20bugs%20-%20strike%201&o2=notequals&chfieldto=-180d&query_format=advanced&chfield=bug_status&chfieldfrom=2000-01-01&f3=status_whiteboard&f2=bug_severity&chfieldvalue=NEEDINFO&bug_status=NEEDINFO&v1=180&v2=enhancement&product=LibreOffice&known_name=DELETE%20NEEDINFO%20bugs%20-%20strike%201

--
_Florian Reisinger _
LibreOffice herunterladen! 
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Alex Thurgood píše v Čt 07. 02. 2013 v 14:17 +0100:
> Le 07/02/2013 13:32, Petr Mladek a écrit :
> 
> Hi Petr,
> 
> > I am confused. Are these bugs in NEEDINFO just because nobody found time
> > to confirm them? If this is true, they should be in the state
> > UNCONFIRMED.
> > 
> > The bugs should be in the state NEEDINFO only when they can't be
> > reproduced because an information is missing.
> > 
> > Could you please give an example of such bugs?
> > 
> 
> https://bugs.freedesktop.org/buglist.cgi?order=Importance&emailcc1=1&emailreporter1=1&emaillongdesc1=1&emailtype1=substring&query_format=advanced&emailassigned_to1=1&emailqa_contact1=1&bug_status=NEEDINFO&email1=iplaw67&product=LibreOffice

Hmm, I took random 10 bugs from the above list. I see that Alex did a
lot of great work. Anyway, here is what I see:

+ https://bugs.freedesktop.org/show_bug.cgi?id=52170:
+ An extremely slow search/browse table in embedded HSQLDB
+ in NEEDINFO for a long time but there was a lot of activity
  later
+ IMHO, it should not be in NEEDINFO
+ in each case, we should stay opened

+ https://bugs.freedesktop.org/show_bug.cgi?id=52388:
+ EDITING - postgresql data not writable, updatable, modifiable
  using built-in connector
+ Lionel (developer) is not able to move forward without more
  info
+ 2 month in needinfo and no reaction
+ => should get closed

+ https://bugs.freedesktop.org/show_bug.cgi?id=53862:
+ EDITING: Inserted graphics are lost
+ hard to reproduce without test document
+ potential duplicate of other bugs
+ 2.5 months in NEEDINFO and no reaction
+ => should get closed; there are probably better duplicates;
 we do not have resources to create a working test document;
 if it is a common bug, it will be reported later and maybe
 with a test document

+ https://bugs.freedesktop.org/show_bug.cgi?id=51390:
+ Users with network based accounts on Mac OS X 10.6
+ bug triages is not able to reproduce because of missing HW
  and complicated setup
+ asked for backtrace 3 months ago and no response
+ I would close this; it is trivial to produce back trace
  and it should be enough to fix the problem
+ NOTE: there is missing a link to the wiki with the hint
  how to create the back trace; Anyway, the user could have
  asked

+ https://bugs.freedesktop.org/show_bug.cgi?id=55753
+ CRASH - new template manager dialog from StartCenter
+ information provided
+ => wrong state and should stay opened

+ https://bugs.freedesktop.org/show_bug.cgi?id=46040
+ UI: Docking the task window crashes LibreOffice 3.5 Impress
+ can't be reproduced
+ 1 year agi asked for more info
+ 1 month ago pinged and no answer
+ => should get closed

+ https://bugs.freedesktop.org/show_bug.cgi?id=52571
+ BASE VBA DELETE COLUMN OR TABLE DOESN'T CHANGE THE SURFACE
+ minimalistic report => hard to understand
+ 5 months ago asked for more details and no answer
+ 2 months ago pinged 
=> should get closed

+ https://bugs.freedesktop.org/show_bug.cgi?id=55085
+ Landscape printing prints as portrait with content turned at
  90 degree
+ too vague report
+ 4 months in needinfo and no reaction
+ => should get closed

+ https://bugs.freedesktop.org/show_bug.cgi?id=54730
+ LibreOffice Base crashes when Test connection is executed
  against JDBC connection
+ problems to reproduce
+ 4 months in needinfo
+ info kind of provided but still not reproducible
=> should get closed as worksforme

+ https://bugs.freedesktop.org/show_bug.cgi?id=36970
+ (orcmid) Help on Managing Digital Signatures is incorrect
+ not producible
+ 4 month in needinfo and no answer
=> should get closed



So, 8 of 10 bugs should get closed. The remaining 2 bugs are in wrong
state.

I am not sure if this sample is enough but I am even more convinced that

+ it makes sense to close such bugs because they are mostly dead
+ ping does not make any sense; if users do not provide the
  info when asked they do not react on ping as well

Note that even if we close few bugs by mistake, it is not that bad. It
makes traffic there and encourages people to move forward.

I still do not understand that many fears and bad feeling about this
action.

Do you have other opinion, feeling, or experience, please?

Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guideli

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Dan Lewis píše v St 06. 02. 2013 v 10:35 -0500:
> On 02/06/2013 09:50 AM, Robert Großkopf wrote:
> > Hi Alex,
> >> After suffering from the last two mass closure / re-initialisations of
> >> status of a fair number of bugs I had spent time in opening, but for
> >> lack of a dedicated developer / interest in those particular areas of LO
> >> at the time (OSX bugs, Base bugs) they never got any attention, I for
> >> one, will not be interested in this happening yet again.
> > All open bugs in Base today: 225
> > Bugs, which are "Unconfimed": 27
> > Bugs, which are "Needinfo": 26
> > I don't think that we need automatically way to get these bugs away.
> > Problem of many of these bugs: They couldn't be confirmed by everybody,
> > because there are special configurations with external databases or Mac.
> > What we need: developers, who will work on bugs in Base. There are many
> > bugs "New" and opened since the first version of LO.
> >
> > Robert
>   I agree with Robert. This also goes to a complaint that I saw on 
> the LibreOffice user list. The person wanted more time spent by the 
> developers on removing the bugs that exist in LO and less time on 
> producing more code that contains more bugs. This is poor planning. From 
> Robert's data, 172 open bugs have not had a developer correct . These 
> are the ones that have been confirmed and contained enough data for a 
> developer to begin work. That is 76% of the bugs reported for Base.

I see many developers spending time on bugfixing. I think that this is
not the main problem here.

IMHO, the problem is that we do not have enough developers interested
into Base. If you look at
https://wiki.documentfoundation.org/ReleaseNotes/4.0#Base there are only
two Base-related features and these are more or less bug fixes.

The question is how to motivate more developers to work on Base. Do you
know some? Could you try to encourage them to contribute?

Also you could pay someone to work on a certain bug. There is a growing
list of certified developers which are capable of doing such things.
These are well spend money because they improve the product and motivate
people working on LO.

> Personally, I think LO (perhaps the BOD) needs to look at what is being 
> done and what effects it is having on the product they are making 
> available for downloading.

Well, BOD cannot force any volunteer to work on any particular bug.
These people promote the product, encourage potential contributors, give
ideas when there is a conflict. You could do this as well, every day.

Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Alex Thurgood]

Le 07/02/2013 13:32, Petr Mladek a écrit :

Hi Petr,

> I am confused. Are these bugs in NEEDINFO just because nobody found time
> to confirm them? If this is true, they should be in the state
> UNCONFIRMED.
> 
> The bugs should be in the state NEEDINFO only when they can't be
> reproduced because an information is missing.
> 
> Could you please give an example of such bugs?
> 

https://bugs.freedesktop.org/buglist.cgi?order=Importance&emailcc1=1&emailreporter1=1&emaillongdesc1=1&emailtype1=substring&query_format=advanced&emailassigned_to1=1&emailqa_contact1=1&bug_status=NEEDINFO&email1=iplaw67&product=LibreOffice



Note that these aren't necessarily my own bug reports, but also include
ones on which I have (and others, including Rainer, or even Joel, for
some of them) either commented or added myself to on CC.

Alex


___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Alex Thurgood píše v St 06. 02. 2013 v 14:09 +0100:
> Le 06/02/2013 13:19, Michael Stahl a écrit :
> 
> 
> >
> > how many of your bugs are in NEEDINFO state?  if the problem is really
> > developer attention (and i don't doubt that this is the case for many
> > bugs) then they should not be in NEEDINFO state and you won't get any mails.
> 
> 
> I was commenting more from the perspective of someone who went through
> the rigmarole of having many of his bug reports reclassified and the
> demoralising effect that can have on a contributor. Personally, I have
> learned to live with it, but that doesn't mean that I think it is a good
> idea generally.
> 
> 
> As for my own reports, I understand that developers may not have the
> time or resources to commit to looking at any given report, but as you
> say, these are probably mostly not in the NEEDINFO status anyway. Some
> of the reports which I have filed, confirmed or added myself to, and
> which are currently in NEEDINFO status (mostly database or printing
> issues) are due mainly to being unable to test the alleged buggy
> behaviour on Mac OSX, either because I don't have the corresponding
> equipment to test with (e.g Brother printers) or OSX's system security
> privileges preventing me from setting up db servers to test on, whereas
> these used to work in previous versions of OSX. These particular
> problems are independent of LO, but that does not mean that the LO bugs
> per se are invalid (since, at the time the report was filed, the problem
> did occur).

I am confused. Are these bugs in NEEDINFO just because nobody found time
to confirm them? If this is true, they should be in the state
UNCONFIRMED.

The bugs should be in the state NEEDINFO only when they can't be
reproduced because an information is missing.

Could you please give an example of such bugs?

Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

> >
> > After suffering from the last two mass closure / re-initialisations of
> > status of a fair number of bugs I had spent time in opening, but for
> > lack of a dedicated developer / interest in those particular areas of LO
> > at the time (OSX bugs, Base bugs) they never got any attention, I for
> > one, will not be interested in this happening yet again.
>

The last one was very different from what we're doing now. These bugs are
currently in NEEDINFO status, the last mass closings have been on bugs that
were UNCONFIRMED. We have decided that closing UNCONFIRMED bugs without
anyone ever looking at them is not okay, but if we have looked at it and
given our volunteer time and have determined the bug needs more
information, if the user fails to provide that information, we should
assume the bug is abandoned. Currently our backlog of UNCONFIRMED bugs is
hovering around 1,400, we will not waste our time looking at NEEDINFO bugs
again - we simply don't have the time to waste.


> >
> > IMO, the rationale behind closing bugs in this way, i.e. "let's do it
> > and if the user/reporter is really motivated he/she is bound to get
> > back" sends completely the wrong message to the user community at large:
>

This isn't the rationale at all. This is a friendly reminder that they
reported a bug and abandoned it and announcing our new policy which is to
maintain a clean and accurate bug tracker. We could really say "you
abandoned this bug and it'll never be looked at unless you take the time to
update it" but that would be much less polite, this is just a friendly
"this bug has gone MONTHS without any activity after a QA member took the
time to look at your issue, if no activity occurs in the next month, the
bug will be assumed to be fixed and/or abandoned" -- the exact wording will
be agreed upon later.


> >
> > - it assumes that casual users/reporters are heavily implicated in the
> > project, or at least enough to defend their bug report tooth and nail;
>

No, it assumes that we all have a responsibility when using open source
(users and contributors) and that we can't ask volunteers to waste time on
poorly submitted bugs that have been abandoned by their reporters.


> >
> > - that many of these bug reports are fallacious or fanciful.
> >
> > While some of them might well fall into the second category, I fear that
> > many are from people who were incited to use the bug submission
> > assistant after encountering a troublesome or even serious problem in
> > their usage of the product and are then expected to get the "community
> > fervour". The reality of this is that if you make a tool easily
> > accessible for reporting bugs, then people simply expect that report to
> > be followed up on by someone else, more knowledgeable. This doesn't mean
> > that their bug report is any less worthy or relevant just because they
> > don't then follow-up.
>

Again, I agree with Petr, this product is free, if you're going to use it
and expect your issues to be tackled, we can reasonably expect a user to at
least update their own bugs. Ideally they would get "community fervour" and
start contributing outside of their own bug reports, but...this is
unrealistic.


> >
> > To me, the solution being proposed is yet another high-handed way of
> > "improving the stats" without due regard to those who made the effort in
> > the first place to submit a report.
>

This has nothing to do with stats, and I love stats. It's a matter of
keeping an accurate bug tracking system. INVALID will show us how many bugs
are essentially abandoned by their reporter which is not our fault. What
Petr said was exactly my feeling, it's a matter of, what's the point of
leaving these bugs in NEEDINFO status when our contributors have already
looked at the bug and determined more is needed before we proceed AT ALL.
NEEDINFO status should "mean something" and that something should be "A QA
member has looked at this bug within the last 30-60 days and determined it
needs more information before proceeding" -- NOT "this bug was looked at
two years ago and determined that it needs more information but has since
been abandoned and will sit in this status forever"

These all seem like reasonable expectations of the user/reporter as well as
our contributors.

Just to repeat, we *will not* close a bug without a QA member first looking
at the bug, this means a contributor has already spent time on it,
expecting more of our contributors is not right -- ie. them guessing at
what a reporter is trying to report and spending a ton of time trying to
reproduce from a terrible list of steps or a lack of an attachment which
actually shows the issue.


Best Regards,
Joel



-- 
*Joel Madero*
LibreOffice QA Volunteer
jmadero@gmail.com
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.l

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Dan Lewis]

On 02/06/2013 09:50 AM, Robert Großkopf wrote:

Hi Alex,

After suffering from the last two mass closure / re-initialisations of
status of a fair number of bugs I had spent time in opening, but for
lack of a dedicated developer / interest in those particular areas of LO
at the time (OSX bugs, Base bugs) they never got any attention, I for
one, will not be interested in this happening yet again.

All open bugs in Base today: 225
Bugs, which are "Unconfimed": 27
Bugs, which are "Needinfo": 26
I don't think that we need automatically way to get these bugs away.
Problem of many of these bugs: They couldn't be confirmed by everybody,
because there are special configurations with external databases or Mac.
What we need: developers, who will work on bugs in Base. There are many
bugs "New" and opened since the first version of LO.

Robert
 I agree with Robert. This also goes to a complaint that I saw on 
the LibreOffice user list. The person wanted more time spent by the 
developers on removing the bugs that exist in LO and less time on 
producing more code that contains more bugs. This is poor planning. From 
Robert's data, 172 open bugs have not had a developer correct . These 
are the ones that have been confirmed and contained enough data for a 
developer to begin work. That is 76% of the bugs reported for Base.
 A bug reporting system is only as good as the resources allocated 
to fixing the bugs. Otherwise, people send in bug reports that may or 
may not even have a developer look at them. Meanwhile, another milestone 
is reached as a newer version comes out containing some of these bugs. 
Where is the quality assurance in that?
 The symptom we have is many bug reports that need additional 
information. But what is causing this? Could it be that there are too 
many versions being produced at the same time? (3.6.5 and 4.0.0) 
Personally, I think LO (perhaps the BOD) needs to look at what is being 
done and what effects it is having on the product they are making 
available for downloading.


--Dan
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Robert Großkopf]

Hi Alex,
> 
> After suffering from the last two mass closure / re-initialisations of
> status of a fair number of bugs I had spent time in opening, but for
> lack of a dedicated developer / interest in those particular areas of LO
> at the time (OSX bugs, Base bugs) they never got any attention, I for
> one, will not be interested in this happening yet again.

All open bugs in Base today: 225
Bugs, which are "Unconfimed": 27
Bugs, which are "Needinfo": 26
I don't think that we need automatically way to get these bugs away.
Problem of many of these bugs: They couldn't be confirmed by everybody,
because there are special configurations with external databases or Mac.
What we need: developers, who will work on bugs in Base. There are many
bugs "New" and opened since the first version of LO.

Robert
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Alex Thurgood]

Le 06/02/2013 13:19, Michael Stahl a écrit :


>
> how many of your bugs are in NEEDINFO state?  if the problem is really
> developer attention (and i don't doubt that this is the case for many
> bugs) then they should not be in NEEDINFO state and you won't get any mails.


I was commenting more from the perspective of someone who went through
the rigmarole of having many of his bug reports reclassified and the
demoralising effect that can have on a contributor. Personally, I have
learned to live with it, but that doesn't mean that I think it is a good
idea generally.


As for my own reports, I understand that developers may not have the
time or resources to commit to looking at any given report, but as you
say, these are probably mostly not in the NEEDINFO status anyway. Some
of the reports which I have filed, confirmed or added myself to, and
which are currently in NEEDINFO status (mostly database or printing
issues) are due mainly to being unable to test the alleged buggy
behaviour on Mac OSX, either because I don't have the corresponding
equipment to test with (e.g Brother printers) or OSX's system security
privileges preventing me from setting up db servers to test on, whereas
these used to work in previous versions of OSX. These particular
problems are independent of LO, but that does not mean that the LO bugs
per se are invalid (since, at the time the report was filed, the problem
did occur). Sweeping them away, even with a "3 strikes and you're out"
policy, will still not make them any less valid. Although I might get
disgruntled with this approach, and feel helpless to do anything about
it, I can live with it (up to a certain point). I'm not convinced that
other, more casual bug reporters, would think the same way.


Alex

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Michael Stahl]

On 06/02/13 10:40, Alex Thurgood wrote:
> Le 05/02/2013 13:03, Petr Mladek a écrit :
> 
> All,
> 
> After suffering from the last two mass closure / re-initialisations of
> status of a fair number of bugs I had spent time in opening, but for
> lack of a dedicated developer / interest in those particular areas of LO
> at the time (OSX bugs, Base bugs) they never got any attention, I for
> one, will not be interested in this happening yet again.

how many of your bugs are in NEEDINFO state?  if the problem is really
developer attention (and i don't doubt that this is the case for many
bugs) then they should not be in NEEDINFO state and you won't get any mails.

> IMO, the rationale behind closing bugs in this way, i.e. "let's do it
> and if the user/reporter is really motivated he/she is bound to get
> back" sends completely the wrong message to the user community at large:
> 
> - it assumes that casual users/reporters are heavily implicated in the
> project, or at least enough to defend their bug report tooth and nail;
> 
> - that many of these bug reports are fallacious or fanciful.

actually i sometimes do see bugs that don't make any sense to me, and
wonder what the heck the reporter's problem is.  but i don't remember
any of your bugs having this problem :)

> While some of them might well fall into the second category, I fear that
> many are from people who were incited to use the bug submission
> assistant after encountering a troublesome or even serious problem in
> their usage of the product and are then expected to get the "community
> fervour". The reality of this is that if you make a tool easily
> accessible for reporting bugs, then people simply expect that report to
> be followed up on by someone else, more knowledgeable. This doesn't mean
> that their bug report is any less worthy or relevant just because they
> don't then follow-up.

a report that says "feature foo doesn't work" without saying in exactly
which way feature foo doesn't work is useless and wastes everybody's
time trying out feature foo and seeing it work as expected because the
reporter didn't write that to see feature foo fail you first have to
insert a bar and a quux into your document, click Undo and do all of
that with change tracking enabled.


___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Pedro]

Alexander Thurgood wrote
> - it assumes that casual users/reporters are heavily implicated in the
> project, or at least enough to defend their bug report tooth and nail;

I don't agree with your interpretation. I believe that someone who bothers
to register to bugzilla and to submit a bug report to such an unfriendly
interface is indeed interested in reporting and expects that the problem is
solved.

The problem here is that there are simply not enough people triaging the
bugs or fixing them (for a software reportedly used by "60 million users"
the number of people contributing is appallingly low...)

Sometimes it takes so long for the reports to move from UNCONFIRMED to NEW
or NEEDINFO (months sometimes...) and even more for a developer to pick it
up (some have been in NEW state for years!!!) that the reporter either found
a workaround or gave up on LibreOffice completely

So if people do not respond to the feedback request it is because they have
lost interest or moved on. If the information in bugzilla is not enough to
verify/replicate the bug then it doesn't make sense to keep it.

I vote for the 3 strike method previously suggested. 1 month interval
between warnings (to make sure that people don't answer because they are on
vacation, etc)

This is my opinion as an occasional collaborator ;)

Cheers,
Pedro 



--
View this message in context: 
http://nabble.documentfoundation.org/Libreoffice-qa-Stagnant-NEEDINFO-bugs-tp4032113p4034827.html
Sent from the QA mailing list archive at Nabble.com.
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Alex Thurgood píše v St 06. 02. 2013 v 10:40 +0100:
> Le 05/02/2013 13:03, Petr Mladek a écrit :
> 
> All,
> 
> After suffering from the last two mass closure / re-initialisations of
> status of a fair number of bugs I had spent time in opening, but for
> lack of a dedicated developer / interest in those particular areas of LO
> at the time (OSX bugs, Base bugs) they never got any attention, I for
> one, will not be interested in this happening yet again.
> 
> IMO, the rationale behind closing bugs in this way, i.e. "let's do it
> and if the user/reporter is really motivated he/she is bound to get
> back" sends completely the wrong message to the user community at large:
> 
> - it assumes that casual users/reporters are heavily implicated in the
> project, or at least enough to defend their bug report tooth and nail;
> 
> - that many of these bug reports are fallacious or fanciful.
> 
> While some of them might well fall into the second category, I fear that
> many are from people who were incited to use the bug submission
> assistant after encountering a troublesome or even serious problem in
> their usage of the product and are then expected to get the "community
> fervour". The reality of this is that if you make a tool easily
> accessible for reporting bugs, then people simply expect that report to
> be followed up on by someone else, more knowledgeable. This doesn't mean
> that their bug report is any less worthy or relevant just because they
> don't then follow-up.
> 
> To me, the solution being proposed is yet another high-handed way of
> "improving the stats" without due regard to those who made the effort in
> the first place to submit a report.

I understand the point. The question is what is the future of these
bugs. Is anyone actively working on them? Is anyone going to work on
them?

Another solution would be to say that these bugs need triaging. Triagers
might schedule these bugs for another review after 1 month in the state
NEEDINFO. They might do their best to reproduce the bug and provide the
needed information. If they are not able to reproduce it, they might
close them as WORKSFORME.

The question is if we have resources to do this. AFAIK, triaggers have
hard times to sort the good UNCONFIRMED bugs these days.


IMHO, this whole discussion is not about statistics but about clean
state and resources.

We want to have bugzilla in a good state because it makes easier to
monitor the state of the product, prioritize, ... Just imagine where we
might end after few years.

The resources are limited. The question is how to use them optimally and
what is better for the product. If triaggers and developers spend more
time with poorly reported bugs, it might delay other important bugs and
annoy more users in the end.

BTW: Users get the product for free. We might expect some contribution
from them as well.


Anyway, this mostly affect the life of bug triagers. If most triagers
are against the mass close, we should not do it. This is why we vote
here :-)


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Alex Thurgood]

Le 05/02/2013 13:03, Petr Mladek a écrit :

All,

After suffering from the last two mass closure / re-initialisations of
status of a fair number of bugs I had spent time in opening, but for
lack of a dedicated developer / interest in those particular areas of LO
at the time (OSX bugs, Base bugs) they never got any attention, I for
one, will not be interested in this happening yet again.

IMO, the rationale behind closing bugs in this way, i.e. "let's do it
and if the user/reporter is really motivated he/she is bound to get
back" sends completely the wrong message to the user community at large:

- it assumes that casual users/reporters are heavily implicated in the
project, or at least enough to defend their bug report tooth and nail;

- that many of these bug reports are fallacious or fanciful.

While some of them might well fall into the second category, I fear that
many are from people who were incited to use the bug submission
assistant after encountering a troublesome or even serious problem in
their usage of the product and are then expected to get the "community
fervour". The reality of this is that if you make a tool easily
accessible for reporting bugs, then people simply expect that report to
be followed up on by someone else, more knowledgeable. This doesn't mean
that their bug report is any less worthy or relevant just because they
don't then follow-up.

To me, the solution being proposed is yet another high-handed way of
"improving the stats" without due regard to those who made the effort in
the first place to submit a report.


Alex



___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

This is in response to Rainer's hesitation. I think there are three 
solid reasons why this should be done and how it would help, one is for 
QA side, one is for the project as a whole and one is for general 
accuracy of FDO:


1) For QA side, if we make this a standard policy then when we push 
something to NEEDINFO we can immediately say, "this bug will remain in 
NEEDINFO state for 30 days, after which it will be closed as INVALID. 
This "could" promote users actually providing information, if they know 
that their bug will go to INVALID after 30 days maybe they'll be a bit 
more rush on their side and we won't continue to accumulate NEEDINFO bugs.


2) For the project as a whole, a lot of these bugs in NEEDINFO could in 
fact be bugs but developers aren't looking at them because of something 
as simple as a missing attachment. Doing this "purge" might encourage 
people who have simply forgotten about their bugs to take the additional 
steps to make it so QA can easily confirm their bugs. Without this it is 
more than likely that the bugs have zero chance of ever being looked at 
again. It would be ashame to have known bugs on FDO but no one will ever 
look at them, doing anything that encourages the input from the reporter 
is a good thing -- I think that this does exactly that.


3) General accuracy of FDO. NEEDINFO means "maybe it's a bug, maybe it's 
not, who knows" -- I think it's important for FDO to be accurate. 
INVALID bugs actually means "this bug is not a valid bug". I think 
NEEDINFO should be a temporary status, not a status we anticipate bugs 
sitting in forever.



Best Regards,
Joel
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

30 days should be enough. If nobody answered withing this time frame,
there is only small chance that she would answer later without pinging.

+1


None. IMHO, it is enough and it reduces the traffic.

+1, I think the developers will appreciate this as well

30 days sounds fine here as well.

+1



Are we agreed that the closing of bugs should be done (aside from the
details of how)?

+1

I agree with all of that. Is that acceptable for everyone?

Best Regards,
Joel
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Rainer Bielefeld píše v Po 04. 02. 2013 v 19:23 +0100:
> Hi all,
> 
> I had the plan to write down some thoughts concerning your plans to do 
> an other mass close, but I wasn't  in the mood to do that. Due to my 
> experience with bug wrangling in general and similar actions we did in 
> LibO Bugzilla, for my personal work I only expect (smaller) 
> complications, and  I am very very doubtful that we will get any benefit 
> at all of such an action. So I don not want to waste my spare time for 
> an action what seems to be completely useless ore even damaging to me.
> You already did lots of discussion concerning details of proceeding, but 
> I can't see any reasoning for what a mass close might be useful.

We did similar mass close in the past. I think that you did some
statistics later. Do you remember how many of the closed bugs were
reopened? And how many of these bugs were later fixed?

> I will not invest any time for this action until I hear at least 1 
> passably conclusive argument concerning fair and checkable benefit the 
> project and / or my (or other power bugwranglers') work will receive by 
> such a mass close.

IMHO, the reason is to get rid of dead bugs. I mean bugs that can't be
solved without the extra information. Most of these bugs would stay like
this forever. They even cause some waste of time when people read them
again and again during the bug triage process but nobody is able to move
them forward.

Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Hi,

I am going to vote on all open questions. It might speed up the voting
process :-)

Joel Madero píše v Po 04. 02. 2013 v 09:11 -0800:
> 
> On Mon, Feb 4, 2013 at 9:09 AM, Petr Mladek  wrote:
> My understanding is that we want to proceed and close the
> bugs.
> The open questions are:
> 
>1. how old bugs should get proceed:
> - from 30days to 6 months
> - it is unclear what is preferred by most people

30 days should be enough. If nobody answered withing this time frame,
there is only small chance that she would answer later without pinging.

>2. whether to send any warning before
> - 1 or none
> - 1 is more polite but it creates traffic
> - it unclear how many people are for what solution

None. IMHO, it is enough and it reduces the traffic.

>3. if we do the warning, it unclear how long to wait with
>   closing the bug after the warning:
> - 7 days, 30 days, or even longer?

30 days sounds fine here as well.

> Are we agreed that the closing of bugs should be done (aside from the
> details of how)?

+1

Best Regards,
Petr


___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Florian Reisinger]

Hi, so I seem to state out my opinion first...



Am 04.02.2013 um 18:11 schrieb Joel Madero :


On Mon, Feb 4, 2013 at 9:09 AM, Petr Mladek  wrote:

> Joel Madero píše v Po 04. 02. 2013 v 08:00 -0800:
> > Do we have a consensus on this? I'd like to pitch our idea to ESC this
> > week if possible. Don't want this to die and fade away into the abyss
> > of thoughts that didn't pan out ;)
>
> To be honest, I am a bit confused by the different opinions.
>
> My understanding is that we want to proceed and close the bugs.
> The open questions are:
>
>1. how old bugs should get proceed:
> - from 30days to 6 months
> - it is unclear what is preferred by most people
>

IMHO 6 months --> Less false positives


>2. whether to send any warning before
> - 1 or none
> - 1 is more polite but it creates traffic
> - it unclear how many people are for what solution
>

1


>3. if we do the warning, it unclear how long to wait with closing the
>   bug after the warning:
> - 7 days, 30 days, or even longer?
> - only Rainer supports 7 days; others prefer longer time but
>   I am not sure if they did not talk about the time before
>   sending the 1st warning
>

They talked about the time from warning -> closure

IMHO 30 days are okay. If we set the time too short, we might face
problems...


> Of course, we also need to decide about the messages but we first need
> to decide about the process itself.
>
> A good solution might be voting on this mailing list.
>

+1, I say let's vote. And thanks for the succinct summary.

Are we agreed that the closing of bugs should be done (aside from the
details of how)?


IMHO yes


Let's start with that, if general agreement is yes, we'll move forward with
specific details (I think we should vote per item to allow discussion if
it's needed).


+1


If we agree that closing bugs need done, I can at least present the idea
during the ESC call to give developers a chance to respond. I know there
are some concerns about the mass emails that will happen upon closing them.


Best Regards,
Joel


-- 
*Joel Madero*
LibreOffice QA Volunteer
jmadero@gmail.com

 ___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings:
http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems?
http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/


Liebe Grüße, / Yours,
Florian Reisinger
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rainer Bielefeld]

Hi all,

I had the plan to write down some thoughts concerning your plans to do 
an other mass close, but I wasn't  in the mood to do that. Due to my 
experience with bug wrangling in general and similar actions we did in 
LibO Bugzilla, for my personal work I only expect (smaller) 
complications, and  I am very very doubtful that we will get any benefit 
at all of such an action. So I don not want to waste my spare time for 
an action what seems to be completely useless ore even damaging to me.
You already did lots of discussion concerning details of proceeding, but 
I can't see any reasoning for what a mass close might be useful.


I will not invest any time for this action until I hear at least 1 
passably conclusive argument concerning fair and checkable benefit the 
project and / or my (or other power bugwranglers') work will receive by 
such a mass close.


This 
 is not an argument.


Best regards

Rainer
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

On Mon, Feb 4, 2013 at 9:09 AM, Petr Mladek  wrote:

> Joel Madero píše v Po 04. 02. 2013 v 08:00 -0800:
> > Do we have a consensus on this? I'd like to pitch our idea to ESC this
> > week if possible. Don't want this to die and fade away into the abyss
> > of thoughts that didn't pan out ;)
>
> To be honest, I am a bit confused by the different opinions.
>
> My understanding is that we want to proceed and close the bugs.
> The open questions are:
>
>1. how old bugs should get proceed:
> - from 30days to 6 months
> - it is unclear what is preferred by most people
>
>2. whether to send any warning before
> - 1 or none
> - 1 is more polite but it creates traffic
> - it unclear how many people are for what solution
>
>3. if we do the warning, it unclear how long to wait with closing the
>   bug after the warning:
> - 7 days, 30 days, or even longer?
> - only Rainer supports 7 days; others prefer longer time but
>   I am not sure if they did not talk about the time before
>   sending the 1st warning
>
> Of course, we also need to decide about the messages but we first need
> to decide about the process itself.
>
> A good solution might be voting on this mailing list.
>

+1, I say let's vote. And thanks for the succinct summary.

Are we agreed that the closing of bugs should be done (aside from the
details of how)?

Let's start with that, if general agreement is yes, we'll move forward with
specific details (I think we should vote per item to allow discussion if
it's needed).

If we agree that closing bugs need done, I can at least present the idea
during the ESC call to give developers a chance to respond. I know there
are some concerns about the mass emails that will happen upon closing them.


Best Regards,
Joel


-- 
*Joel Madero*
LibreOffice QA Volunteer
jmadero@gmail.com
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

Joel Madero píše v Po 04. 02. 2013 v 08:00 -0800:
> Do we have a consensus on this? I'd like to pitch our idea to ESC this
> week if possible. Don't want this to die and fade away into the abyss
> of thoughts that didn't pan out ;)

To be honest, I am a bit confused by the different opinions. 

My understanding is that we want to proceed and close the bugs.
The open questions are:

   1. how old bugs should get proceed:
- from 30days to 6 months
- it is unclear what is preferred by most people

   2. whether to send any warning before
- 1 or none
- 1 is more polite but it creates traffic
- it unclear how many people are for what solution

   3. if we do the warning, it unclear how long to wait with closing the
  bug after the warning:
- 7 days, 30 days, or even longer?
- only Rainer supports 7 days; others prefer longer time but
  I am not sure if they did not talk about the time before
  sending the 1st warning

Of course, we also need to decide about the messages but we first need
to decide about the process itself.

A good solution might be voting on this mailing list.


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

On Tue, Jan 29, 2013 at 6:59 AM, Petr Mladek  wrote:

> On Sun, 2013-01-27 at 13:43 +0100, Rainer Bielefeld wrote:
> > I) Such an action should avoid collateral damages as effective as
> > possible. A promising approach might be to find an effective query with
> > good accuracy for hopeless Bug reports where we can expect that there
> > will be no useful reaction from reporter, AND where we can be sure that
> > the NEEDINFO is appropriate
>

I think that this query does not exist. FDO has a lot of limitations,
finding out if a QA or reviewer was lazy in their changeour best guess
is if this was the case, the reporter would have reopened the bug and not
left it in NEEDNIFO for >6 months.


> > At least I can say for me that sometimes I am interested and tough and I
> > find a real bug even with a very rare report. And sometimes I try to get
> > a better bug description because I want to  save some work. I will not
> > forget these Bugs, they are in my hold-file, but if I saw a low
> > priority, it might take a year or so until I get back to the bug. And
> > this might cause unnecessary work for other reviewers, may be hundreds
> > invest half a minute, see that someone is involved, leave again, and so
> > we loose some hors every year in such 1 bug.
>

I understand what you're saying here but unfortunately there is nothing we
can do about other QA people who in the past have not put in due time to
think about if NEEDINFO is appropriate. Furthermore, I think that if it's a
"quick fix" that the reporter should just provide the information. While I
think that QA should take on the brunt of the work for triaging, we should
and do expect a certain level of dedication from the user who reported the
bug. If a user doesn't take 5 minutes to update their bugwhose fault is
that really?


>
> This is one thing that I am afraid of as well. We could close bugs that
> are somehow solvable but nobody has found time to do more investigation.
>

Agree with below remark about nothing being perfect.


> The question is how many bugs belong to this category. Someone pointed
> out that bugs that have been in NEEDINFO longer time are often dead
> ends.
>
> I think that we newer will be perfect. If a bug is staying around for a
> long time, it is not super cricical. If we close valid bug and nobody
> bother to reopen it, it is not supercritical as well. We have more bugs
> that we are able to fix, so we could use this as a natural filtering of
> the less important and thus very low priority bugs.
>
> I am sure that if a bug is important and we close it "by mistake",
> someone will reopen it. This activity will bring more attention to it
> and move it forward. It is not ideal but it has positive effects. What
> do you think?
>

+1, but I think that we should be be prepared for some vocal reporters who
get angry about being inconvenienced. Unfortunately if this is what it
takes to clean up FDO, I'm all for it.


Personally I think that 3-6 months range, one reminder and then 1-4 weeks
later closing the bug is best but this will result in double the mail which
I know some of the devs have already vocalized not being happy with. We
must find a way to mitigate this as one developer in particular said flatly
that we would piss off a lot of developers if they see hundreds of emails
in their mail account.


Best Regards,
Joel


-- 
*Joel Madero*
LibO QA Volunteer
jmadero@gmail.com
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rainer Bielefeld]

Petr Mladek schrieb:


Well, I am still not persuaded that we want 2 or more strike solution.


Hi Petr,

me too! This mailing with automated personal mails is my compromise 
offer. And I have little hope to get some feedback for those 1/3 real 
bugs I observed in a sample before the last "Big close" (in between 300 
or so have been reopened again ..., it seems you share my concerns 
"collateral damages").


If you (all) agree I would stop the discussion here for few days (I have 
too much work to do the next 2-3 days. Coming weekend I will create a 
suggestion for the proceeding with the arguments we had. May be we do 
some additional fine tuning, and decision whether it should be done so 
or in a completely different way should be by ESC.


We should not overrate this book-keeping clearing up and not invest too 
much work.


Best regards


Rainer
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

On Sun, 2013-01-27 at 13:43 +0100, Rainer Bielefeld wrote:
> I) Such an action should avoid collateral damages as effective as 
> possible. A promising approach might be to find an effective query with 
> good accuracy for hopeless Bug reports where we can expect that there 
> will be no useful reaction from reporter, AND where we can be sure that 
> the NEEDINFO is appropriate and not only caused by laziness of reviewer 
> At least I can say for me that sometimes I am interested and tough and I 
> find a real bug even with a very rare report. And sometimes I try to get 
> a better bug description because I want to  save some work. I will not 
> forget these Bugs, they are in my hold-file, but if I saw a low 
> priority, it might take a year or so until I get back to the bug. And 
> this might cause unnecessary work for other reviewers, may be hundreds 
> invest half a minute, see that someone is involved, leave again, and so 
> we loose some hors every year in such 1 bug.

This is one thing that I am afraid of as well. We could close bugs that
are somehow solvable but nobody has found time to do more investigation.

The question is how many bugs belong to this category. Someone pointed
out that bugs that have been in NEEDINFO longer time are often dead
ends.

I think that we newer will be perfect. If a bug is staying around for a
long time, it is not super cricical. If we close valid bug and nobody
bother to reopen it, it is not supercritical as well. We have more bugs
that we are able to fix, so we could use this as a natural filtering of
the less important and thus very low priority bugs.

I am sure that if a bug is important and we close it "by mistake",
someone will reopen it. This activity will bring more attention to it
and move it forward. It is not ideal but it has positive effects. What
do you think?


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

On Tue, 2013-01-29 at 13:20 +0100, Rainer Bielefeld wrote:
> Pedro schrieb:
> 
> > I think that is extremely rude.  It reminds me of my Graduation diploma
> > which was not ready for 3 years and then I received a postcard saying I had
> > 10 days to pick it up...
> 
> Hi,
> 
> no, we will not leave open a pending action item for 1 month. During 
> that time several of the related bugs might have changed status because 
> other have started work on it ,... .
> 
> 1 Weekbetween annoncement and final action is ok, and we will explain 
> the re asons why we closed the bug in that "First mail without Bugzilla 
> noise"  (I will create a suggestion basing on the available suggestions, 
> soon) so that nobody can fell frustrated.

Well, I am still not persuaded that we want 2 or more strike solution.
Even when you send mails outside bugzilla, it will create noise :-)

As Florian pointed out, we do not know who is supposed to provide the
information. In some cases, it is not the reporter but another person
that is in CC.


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rainer Bielefeld]

Pedro schrieb:


I think that is extremely rude.  It reminds me of my Graduation diploma
which was not ready for 3 years and then I received a postcard saying I had
10 days to pick it up...


Hi,

no, we will not leave open a pending action item for 1 month. During 
that time several of the related bugs might have changed status because 
other have started work on it ,... .


1 Weekbetween annoncement and final action is ok, and we will explain 
the re asons why we closed the bug in that "First mail without Bugzilla 
noise"  (I will create a suggestion basing on the available suggestions, 
soon) so that nobody can fell frustrated.


CU

Rainer
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Pedro]

Petr Mladek wrote
> Some people suggested 6 months. You suggests one week. I do not persist
> on 30 days but I think that it is a good compromise.
> 
> For example, imagine someone who reports a bug before leaving on
> vacation, business trip, or into hospital. In addition, I think that it
> would be rude to push users to provide feedback within 7 days when the
> triage and fixing the bugs might take weeks and months :-)

+1
I think that is extremely rude.  It reminds me of my Graduation diploma
which was not ready for 3 years and then I received a postcard saying I had
10 days to pick it up...

I believe that 1 month  interval between the 3 strikes is the most
reasonable. It doesn't leave untouched items for more than 3 months and it
should overcome the  "vacation, business trip, or into hospital" period.

Just my 2 cents ;)



--
View this message in context: 
http://nabble.documentfoundation.org/Libreoffice-qa-Stagnant-NEEDINFO-bugs-tp4032113p4032870.html
Sent from the QA mailing list archive at Nabble.com.
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

On Tue, 2013-01-29 at 07:51 +0100, Rainer Bielefeld wrote:
> Petr Mladek schrieb:
> 
> > IMHO, the most important is to give
> > user chance to answer before the first warning (30 days or so).
> 
> Hi Petr,
> 
> I don't think so. My experience is that the reporter normally will 
> answer within 3 days (mostly: latest next day) or never. We should not 
> base our decisions on 1 user in the world with a Sabbatical somewhere in 
> the Australian outback hindered to answer ;-)
> All these theses that we should give a reporter weeks of time never have 
> been underpinned by data, and object to my experience.

Some people suggested 6 months. You suggests one week. I do not persist
on 30 days but I think that it is a good compromise.

For example, imagine someone who reports a bug before leaving on
vacation, business trip, or into hospital. In addition, I think that it
would be rude to push users to provide feedback within 7 days when the
triage and fixing the bugs might take weeks and months :-)


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

On Tue, 2013-01-29 at 07:09 +0100, Florian Reisinger wrote:
> Am 28.01.2013 14:13, schrieb Petr Mladek:

> > > Strike 2 After 7 Days:
> > > Query for all Bugs for what mails have been sent in Strike 1:
> > > - Changed since mail (probably by reporter): QA will take care
> > > - NOT changed: Mass close via Bugzilla with polite message
> > >"Sorry ..., but feel free to reopen if ..."
> > > 
> > > What do think?
> > I like this solution. It is polite and creates only one change in
> > bugzilla.
> [...]

> Okay do you mean the button "Send mail to bug Assignees"?
> In my query some of the following would get mail: Me, A. Timar,
> Joel... so QA staff + dev
> So, what should we do? I was able to display the e-mail of the
> reporter, but you won't get me to send 200 mails
> A button "Send mail to the reporter and everyone who is interested in
> this bug" would be nice ;) [Sorry for sarcasm]

Good point! This actually shows how annoying are the mass changes in
bugzilla when it is done for hundreds of bugs.

I think that we must do at least one change in bugzilla (close the bug).
What about doing only this action? Do we really need warnings before
closing?

As Rainer writes in the other mail. The user usually provide the answer
within 3 working days or newer.

What is wrong when we close the bug with a polite message, explaining
why we are not able to solve the bugs, and encouraging the user to
provide the information and reopen the bug?

I know this is what we did in the first mass close and some bug triagers
complained. But I think that it was not well explained on the mailing
list before the change was done.

Anyone from the complainers should read this thread, complain now and
suggest better solution :-)


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rainer Bielefeld]

Petr Mladek schrieb:


IMHO, the most important is to give
user chance to answer before the first warning (30 days or so).


Hi Petr,

I don't think so. My experience is that the reporter normally will 
answer within 3 days (mostly: latest next day) or never. We should not 
base our decisions on 1 user in the world with a Sabbatical somewhere in 
the Australian outback hindered to answer ;-)
All these theses that we should give a reporter weeks of time never have 
been underpinned by data, and object to my experience.


Of course, I think it might be appropriate to add a line to the text "If 
you need some time to get familiar with the problem again, please leave 
a short note in a Comment ...".



Best regards


Rainer
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Florian Reisinger]

Hello everyone,

Am 28.01.2013 14:13, schrieb Petr Mladek:

On Mon, 2013-01-28 at 10:05 +0100, Rainer Bielefeld wrote:

Petr Mladek schrieb:
[...] 
BTW, I dislike the "noise" the discussed "3 strikes" solution will

cause. I'm thinking about a different solution:


I think most of you know, that I proposed the three strikes, but I am for the 
best solution so. let's wait till the end of the discussion


I am against 3 strike solution as well :-) My opinion is that it would
cause to big traffic and do not help much. If people does not react for
the first warning, there is only small chance that they would react on
the second or third one.


Strike 1:
Query will find NEEDINFO bugs untouched for a long time and fulfilling
some additional "hopeless criteria".
Reporter's of these bugs will get polite mail with request to contribute
additional info that we will have to close the bug without additional
info. This mailing  only send mails to reporters, will not change any
info in the Bugs, so that data as "Days since last change" and similar
will be available for other queries. List of related bugs will be
published on QA list

That's not a big technical challenge, I think I can create required
tools (what can be used fur further actions in future easily) within 1 hour.

Strike 2 After 7 Days:
Query for all Bugs for what mails have been sent in Strike 1:
- Changed since mail (probably by reporter): QA will take care
- NOT changed: Mass close via Bugzilla with polite message
"Sorry ..., but feel free to reopen if ..."

What do think?

I like this solution. It is polite and creates only one change in
bugzilla.

[...]

Okay do you mean the button "Send mail to bug Assignees"?
In my query some of the following would get mail: Me, A. Timar, Joel... 
so QA staff + dev
So, what should we do? I was able to display the e-mail of the reporter, 
but you won't get me to send 200 mails
A button "Send mail to the reporter and everyone who is interested in 
this bug" would be nice ;) [Sorry for sarcasm]


BTW: I have a improved query for LT Neeedinfo bugs: Currently 393



BTW, I would not do that too often. Sometimes it's simply not easy for
reporter to contribute desired info, for example because bug is not
simple to reproduce. May be such bugs can be marked by entry of a QA
"Mentor" in QA contact or similar.

I would do this regularly to keep bugzilla clean and avoid masschanges
in hunderts of bugs. There are different reporters, so we will not touch
the same reporter in each round. IMHO, the most important is to give
user chance to answer before the first warning (30 days or so).

+1



Best Regards,
Petr





--
_Florian Reisinger _
LibreOffice herunterladen! 
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

Also, in order to avoid 100's even at the beginning, why don't we split the
100's up and do them over several weeks? I can take care of this if it's
the decision. I can filter by date and do maybe 2 week increments - I
believe the oldest NEEDINFO bug is from nearly a year ago with no activity.
Thoughts?

Best Regards,
Joel


On Mon, Jan 28, 2013 at 5:13 AM, Petr Mladek  wrote:

> On Mon, 2013-01-28 at 10:05 +0100, Rainer Bielefeld wrote:
> > Petr Mladek schrieb:
> > > This will cause many mails only in the first round. It will be normal
> > > level of mails if we do this regularly.
> >
> >
> > Hi,
> >
> > That's an illusion, total number of mails will always be the the same.
> > Only the number of mails per cleanup will be smaller.
>
> IMHO, there is a difference when you get 100 mails now because we want
> to clean up the current "mess" or when you get 5 mails per week when we
> do this regularly.
>
> > BTW, I dislike the "noise" the discussed "3 strikes" solution will
> > cause. I'm thinking about a different solution:
>
> I am against 3 strike solution as well :-) My opinion is that it would
> cause to big traffic and do not help much. If people does not react for
> the first warning, there is only small chance that they would react on
> the second or third one.
>
> > Strike 1:
> > Query will find NEEDINFO bugs untouched for a long time and fulfilling
> > some additional "hopeless criteria".
> > Reporter's of these bugs will get polite mail with request to contribute
> > additional info that we will have to close the bug without additional
> > info. This mailing  only send mails to reporters, will not change any
> > info in the Bugs, so that data as "Days since last change" and similar
> > will be available for other queries. List of related bugs will be
> > published on QA list
> >
> > That's not a big technical challenge, I think I can create required
> > tools (what can be used fur further actions in future easily) within 1
> hour.
> >
> > Strike 2 After 7 Days:
> > Query for all Bugs for what mails have been sent in Strike 1:
> > - Changed since mail (probably by reporter): QA will take care
> > - NOT changed: Mass close via Bugzilla with polite message
> >"Sorry ..., but feel free to reopen if ..."
> >
> > What do think?
>
> I like this solution. It is polite and creates only one change in
> bugzilla.
>
> > BTW, I would not do that too often. Sometimes it's simply not easy for
> > reporter to contribute desired info, for example because bug is not
> > simple to reproduce. May be such bugs can be marked by entry of a QA
> > "Mentor" in QA contact or similar.
>
> I would do this regularly to keep bugzilla clean and avoid masschanges
> in hunderts of bugs. There are different reporters, so we will not touch
> the same reporter in each round. IMHO, the most important is to give
> user chance to answer before the first warning (30 days or so).
>
>
> Best Regards,
> Petr
>
>
>


-- 
*Joel Madero*
LibO QA Volunteer
jmadero@gmail.com
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Pedro]

Hi Joel, all


jmadero wrote
> d) a "3 strikes" type system - send one reminder, a month later send a 
> second with a warning the bug will be closed, then close the bug

+1
but this sounds like a lot of human work...

How can you make sure that a reply is noticed (e.g. after first reminder) so
that someone from QA reviews it before the second message is sent?



--
View this message in context: 
http://nabble.documentfoundation.org/Libreoffice-qa-Stagnant-NEEDINFO-bugs-tp4032113p4032601.html
Sent from the QA mailing list archive at Nabble.com.
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Marc Paré]

Le 2013-01-27 17:10, bfo a écrit :

Hi!
My proposal:

e) a reminder at major version + automated closing of bug report after first
major.maintenance version if a reminder is last comment (which means no
activity)

This will give bugs cleaning every 6 months. Not a bad deal...
Most of bugs in NEEDINFO state are dead ends. Reporters in general think
that people who triage are LO experts in every way. A reminder text should
convince them, that the easier STR the better. Even very complicated issue
can be reproduced with a "download'n'run" test case attached . BTW: I really
liked graphics made by one reporter in bug
https://bugs.freedesktop.org/show_bug.cgi?id=51162.
IMHO all this should include also UNCONFIRMED bugs (many are dead ends too).
I see bug reporting in this project not like "file and forget" process.
Reminder text should encourage the reporter to get involved a little bit
more, as he is an expert about reported bug, at least confirming it with
latest major version.
Also having a subpage for dealing with UNCONFIRMED or NEEDINFO bugs in
Bugzilla would make life easier.
Such page is available for instance in bmo (bugzilla.mozilla.org).
Link: https://bugzilla.mozilla.org/page.cgi?id=triage_reports.html.
Source is available at:
http://bzr.mozilla.org/bmo/4.0/annotate/head:/extensions/BMO/template/en/default/pages/triage_reports.html.tmpl
In short (their implementation):
- Show UNCONFIRMED bugs with: [Product] [Component]
- Comment:
-- where the last commenter: is the reporter, does not have canconfirm, is
[set user]
-- where the last comment is older than: 30/60/90 days, one year, the date
[set date]
I played with it for a while and it is very friendly, especially  for
unexperienced Bugzilla users.
In our version this should support UNCONFIRMED and NEEDINFO statuses and
maybe reminder IDs, to browse the bugs before the next step (if one is
concerned that important issues will be closed).
It would be a great companion for triage marathons.
Best regards.
P.S.
Everything has to be done to not repeat last incident with 3 or 4 messages
added to selected issues.  This was PITA for everyone.




Well, you have some good comments, but could you please sign your 
messages with your name? For some of us who are on the list, we may not 
know who you are. I tried to find you name on your previous messages and 
on other lists and could not find it.


Sorry if you find this message rude, but it is customary to sign your 
name on your posts on the project.


Cheers,

Marc


--
Marc Paré
m...@marcpare.com
http://www.parEntreprise.com
parEntreprise.com Supports OpenDocument Formats (ODF)
parEntreprise.com Supports http://www.LibreOffice.org

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Marc Paré]

Le 2013-01-28 08:13, Petr Mladek a écrit :

On Mon, 2013-01-28 at 10:05 +0100, Rainer Bielefeld wrote:

Petr Mladek schrieb:

This will cause many mails only in the first round. It will be normal
level of mails if we do this regularly.



Hi,

That's an illusion, total number of mails will always be the the same.
Only the number of mails per cleanup will be smaller.


IMHO, there is a difference when you get 100 mails now because we want
to clean up the current "mess" or when you get 5 mails per week when we
do this regularly.


BTW, I dislike the "noise" the discussed "3 strikes" solution will
cause. I'm thinking about a different solution:


I am against 3 strike solution as well :-) My opinion is that it would
cause to big traffic and do not help much. If people does not react for
the first warning, there is only small chance that they would react on
the second or third one.


Strike 1:
Query will find NEEDINFO bugs untouched for a long time and fulfilling
some additional "hopeless criteria".
Reporter's of these bugs will get polite mail with request to contribute
additional info that we will have to close the bug without additional
info. This mailing  only send mails to reporters, will not change any
info in the Bugs, so that data as "Days since last change" and similar
will be available for other queries. List of related bugs will be
published on QA list

That's not a big technical challenge, I think I can create required
tools (what can be used fur further actions in future easily) within 1 hour.

Strike 2 After 7 Days:
Query for all Bugs for what mails have been sent in Strike 1:
- Changed since mail (probably by reporter): QA will take care
- NOT changed: Mass close via Bugzilla with polite message
"Sorry ..., but feel free to reopen if ..."

What do think?


I like this solution. It is polite and creates only one change in
bugzilla.


BTW, I would not do that too often. Sometimes it's simply not easy for
reporter to contribute desired info, for example because bug is not
simple to reproduce. May be such bugs can be marked by entry of a QA
"Mentor" in QA contact or similar.


I would do this regularly to keep bugzilla clean and avoid masschanges
in hunderts of bugs. There are different reporters, so we will not touch
the same reporter in each round. IMHO, the most important is to give
user chance to answer before the first warning (30 days or so).


Best Regards,
Petr




Sure, then no problem with this, but, let's not forget that some of our 
users are reporting in EN and that this may not be their 1st language. 
If it needs more info, we should also try as well to look at it enough 
to try at first to fill in the blanks ourselves. The bug report may be 
critical enough to look at.


I would also be concerned that whoever is triaging may themselves have 
trouble understanding the report and just dismiss it as "needinfo" and 
then send it back to the user. Many users would then see it more as a 
sign of apathy on our part thinking "I submitted the bug and they are 
just not interested in fixing the problem I told them about ..."


Add to this that our users all have lives, jobs etc. and that the bug 
reporting takes time to submit on their part and that for most it is to 
help us with having LibreOffice work the right way. So, we should not 
send users messages in any way that may sound like their bug is useless; 
the message should be really appreciative of their time that they have 
given us in reporting the bug and that we would really like to fix it 
but it needs just a little more information for us to understand the 
problem. We need to keep them engaged in the bug reporting process.


Cheers,

Marc

--
Marc Paré
m...@marcpare.com
http://www.parEntreprise.com
parEntreprise.com Supports OpenDocument Formats (ODF)
parEntreprise.com Supports http://www.LibreOffice.org

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

On Mon, 2013-01-28 at 10:05 +0100, Rainer Bielefeld wrote:
> Petr Mladek schrieb:
> > This will cause many mails only in the first round. It will be normal
> > level of mails if we do this regularly.
> 
> 
> Hi,
> 
> That's an illusion, total number of mails will always be the the same. 
> Only the number of mails per cleanup will be smaller.

IMHO, there is a difference when you get 100 mails now because we want
to clean up the current "mess" or when you get 5 mails per week when we
do this regularly.

> BTW, I dislike the "noise" the discussed "3 strikes" solution will 
> cause. I'm thinking about a different solution:

I am against 3 strike solution as well :-) My opinion is that it would
cause to big traffic and do not help much. If people does not react for
the first warning, there is only small chance that they would react on
the second or third one.

> Strike 1:
> Query will find NEEDINFO bugs untouched for a long time and fulfilling 
> some additional "hopeless criteria".
> Reporter's of these bugs will get polite mail with request to contribute 
> additional info that we will have to close the bug without additional 
> info. This mailing  only send mails to reporters, will not change any 
> info in the Bugs, so that data as "Days since last change" and similar 
> will be available for other queries. List of related bugs will be 
> published on QA list
> 
> That's not a big technical challenge, I think I can create required 
> tools (what can be used fur further actions in future easily) within 1 hour.
> 
> Strike 2 After 7 Days:
> Query for all Bugs for what mails have been sent in Strike 1:
> - Changed since mail (probably by reporter): QA will take care
> - NOT changed: Mass close via Bugzilla with polite message
>"Sorry ..., but feel free to reopen if ..."
> 
> What do think?

I like this solution. It is polite and creates only one change in
bugzilla.

> BTW, I would not do that too often. Sometimes it's simply not easy for 
> reporter to contribute desired info, for example because bug is not 
> simple to reproduce. May be such bugs can be marked by entry of a QA 
> "Mentor" in QA contact or similar.

I would do this regularly to keep bugzilla clean and avoid masschanges
in hunderts of bugs. There are different reporters, so we will not touch
the same reporter in each round. IMHO, the most important is to give
user chance to answer before the first warning (30 days or so). 


Best Regards,
Petr


___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rainer Bielefeld]

Petr Mladek schrieb:

This will cause many mails only in the first round. It will be normal
level of mails if we do this regularly.



Hi,

That's an illusion, total number of mails will always be the the same. 
Only the number of mails per cleanup will be smaller.


BTW, I dislike the "noise" the discussed "3 strikes" solution will 
cause. I'm thinking about a different solution:


Strike 1:
Query will find NEEDINFO bugs untouched for a long time and fulfilling 
some additional "hopeless criteria".
Reporter's of these bugs will get polite mail with request to contribute 
additional info that we will have to close the bug without additional 
info. This mailing  only send mails to reporters, will not change any 
info in the Bugs, so that data as "Days since last change" and similar 
will be available for other queries. List of related bugs will be 
published on QA list


That's not a big technical challenge, I think I can create required 
tools (what can be used fur further actions in future easily) within 1 hour.


Strike 2 After 7 Days:
Query for all Bugs for what mails have been sent in Strike 1:
- Changed since mail (probably by reporter): QA will take care
- NOT changed: Mass close via Bugzilla with polite message
  "Sorry ..., but feel free to reopen if ..."

What do think?

BTW, I would not do that too often. Sometimes it's simply not easy for 
reporter to contribute desired info, for example because bug is not 
simple to reproduce. May be such bugs can be marked by entry of a QA 
"Mentor" in QA contact or similar.



CU

Rainer
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Petr Mladek]

On Sat, 2013-01-26 at 15:06 -0800, Joel Madero wrote:
> On 01/26/2013 02:48 PM, Jack Leigh wrote:
> >> c) a reminder + automated closing of bug after some period of time
> >
> > +1 Something like automated closing after 3 months in NEEDINFO
> Response: From developer side, they'll see a MASSIVE influx of emails, 
> pointed out by a core developer. This would need approval from ESC

This will cause many mails only in the first round. It will be normal
level of mails if we do this regularly. If people are in CC, they need
to expect some comments, either from the user, or triager, or an
automat.

> Possible option: New keyword "AUTOMATED", devs can filter these out 
> individually and easily delete

In each case, we need to write some explanation in the comment.

One problem here is how to deal with bugs in NEEDEDINFO state where the
user provided some information but she forgot to change the status flag.
Are these situations detected by bug triagers in time? Will these bugs
be closed automatically as well?

Anyway, I think that it is best to get the information out of the user
when the bug is fresh and when the user is somehow motivated. So, I
would prefer to keep some dynamics and close the bugs after one month in
NEEDINFO state. I would use a comment like:

"This bug has been in NEEDINFO for more than 30 days. It cannot be
solved in this state. Feel free to reopen it together with the required
information."

If we do the close automatically and could not detect wrong state. We
could add one more paragraph, with something like:

"This change has been done automatically. If you have already provided
the information and just forgot to change the NEEDINFO state, feel free
to simply reopen the bug."


Best Regards,
Petr

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[bfo]

Hi!
My proposal:

e) a reminder at major version + automated closing of bug report after first
major.maintenance version if a reminder is last comment (which means no
activity)

This will give bugs cleaning every 6 months. Not a bad deal...
Most of bugs in NEEDINFO state are dead ends. Reporters in general think
that people who triage are LO experts in every way. A reminder text should
convince them, that the easier STR the better. Even very complicated issue
can be reproduced with a "download'n'run" test case attached . BTW: I really
liked graphics made by one reporter in bug
https://bugs.freedesktop.org/show_bug.cgi?id=51162.
IMHO all this should include also UNCONFIRMED bugs (many are dead ends too). 
I see bug reporting in this project not like "file and forget" process.
Reminder text should encourage the reporter to get involved a little bit
more, as he is an expert about reported bug, at least confirming it with
latest major version.
Also having a subpage for dealing with UNCONFIRMED or NEEDINFO bugs in
Bugzilla would make life easier.
Such page is available for instance in bmo (bugzilla.mozilla.org). 
Link: https://bugzilla.mozilla.org/page.cgi?id=triage_reports.html. 
Source is available at:
http://bzr.mozilla.org/bmo/4.0/annotate/head:/extensions/BMO/template/en/default/pages/triage_reports.html.tmpl
In short (their implementation):
- Show UNCONFIRMED bugs with: [Product] [Component]
- Comment:
-- where the last commenter: is the reporter, does not have canconfirm, is
[set user]
-- where the last comment is older than: 30/60/90 days, one year, the date
[set date]
I played with it for a while and it is very friendly, especially  for
unexperienced Bugzilla users.
In our version this should support UNCONFIRMED and NEEDINFO statuses and
maybe reminder IDs, to browse the bugs before the next step (if one is
concerned that important issues will be closed).
It would be a great companion for triage marathons.
Best regards.
P.S.
Everything has to be done to not repeat last incident with 3 or 4 messages
added to selected issues.  This was PITA for everyone.




--
View this message in context: 
http://nabble.documentfoundation.org/Libreoffice-qa-Stagnant-NEEDINFO-bugs-tp4032113p4032419.html
Sent from the QA mailing list archive at Nabble.com.
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rainer Bielefeld]

Joel Madero schrieb:

Hi Joel,

thank you for opening this discussion with your thoughts and suggestions.

To avoid misunderstandings: my critics at current plans does not mean 
that I do not want any automated mass cleanups. I would appreciate to 
get rid of lots of hopeless reports with few "costs".


Although I believe we could solve some part of the problem if we would 
invest the time discussing here into bug reviews, it might be useful to 
do some very careful preparation, so that in future such mass closes can 
be done effectively and with few preparation, because the work already 
has been done here.


Some thoughts (I'm still sorting and prioritizing)

I) Such an action should avoid collateral damages as effective as 
possible. A promising approach might be to find an effective query with 
good accuracy for hopeless Bug reports where we can expect that there 
will be no useful reaction from reporter, AND where we can be sure that 
the NEEDINFO is appropriate and not only caused by laziness of reviewer 
At least I can say for me that sometimes I am interested and tough and I 
find a real bug even with a very rare report. And sometimes I try to get 
a better bug description because I want to  save some work. I will not 
forget these Bugs, they are in my hold-file, but if I saw a low 
priority, it might take a year or so until I get back to the bug. And 
this might cause unnecessary work for other reviewers, may be hundreds 
invest half a minute, see that someone is involved, leave again, and so 
we loose some hors every year in such 1 bug.


II) Such a mass close should really terminate any action on a closed bug 
(except reporter redelivers useful info). The last mass close in August 
2012 also closed lots of appropriate enhancement requests. We should 
avoid such a mistake in future.


III) I prefer to close Bugs immediately with a very polite text 
encouraging the reporter to reopen the bug if he can contribute more 
detailed and precise due to 
. I am pretty sure that 
those who would have improved their bug description after a "first 
strike" will understand our reasons and simply reopen the Bug with 
better info. And for the other ones the first and second strike would be 
wasted energy.


IV) We need a very polite and encouraging text for first strike bug 
closing.


V) Remaining concerns
We can be as careful as possible, it's inevitable that we will close 
some bugs reported from real experts with appropriate info, but only the 
bad luck that the of us who reviewed was expert for something else. That 
makes us looking like idiots.



My method of resolution:
I will start with a suggestion for a query with reasoning why I believe 
that that should be done that way soon, and then we can integrate more 
proposals for optimization step by step.


We will see to where that leads, and the rating how nearby the optimum 
that is at least for me is decisive what of Joels suggesteions might be 
the best.


CU

Rainer
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Rob Snelders]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

The option from Marc looks good to me.

But do you look at the length that a bug is in NEEDINFO or the time
from the last post in the bug?

- --
Greetings,
Rob Snelders

Op 27-01-13 09:45, Marc Par￩ schreef:
> Hi Joel,
> 
> If you are asking our opinion on these ...
> 
> Le 2013-01-26 17:22, Joel Madero a ￩crit :
>> Hi All,
>> 
>> During our last QA call we came up with a plan for NEEDINFO bugs
>> that have been stagnant for 6+ months. I've decided to remove
>> this from the minutes because there are some concerns about the
>> agreement that was made. It was requested that we have a
>> discussion through a thread about what we should do about these
>> bugs.
>> 
>> Let's try to keep this thread moving and organized so we can make
>> a decision once and for all. Goal should be to minimize flare ups
>> from users while keeping in mind that our long term QA goal is to
>> get FDO organized, accurate and under control.
>> 
>> With that, I'll let the thread move forward :-D
>> 
>> Options: (feel free to add)
>> 
>> a) Leave them alone, accept that most will just sit in NEEDINFO
>> forever
>> 
>> b) Automate a reminder for these bugs and leave it at that, if
>> it doesn't lead to anything, so be it, leave the bugs alone
>> 
>> c) a reminder + automated closing of bug after some period of
>> time
>> 
>> d) a "3 strikes" type system - send one reminder, a month later
>> send a second with a warning the bug will be closed, then close
>> the bug
>> 
>> 
>> These are the options that come to mind.
>> 
>> Best Regards, Joel
>> 
>> P.S. Florian - please hold off on writing something about our
>> previous agreement until we reach some consensus here.
> 
> +1 for d) BUT make it a real "3 strikes". one reminder, a month
> later send another reminder, a month later send the last reminder
> with a warning that the bug will be closed in one month.
> 
> The 3-strikes concept is understandable to most people and I
> believe the submitter would be more understanding if the bug were
> closed. It also gives more than an abundant amount of time to fill
> out the missing details of the bug -- from start to end would take
> 4 months. This would hopefully quiet the noise from disgruntled
> submitters and allow QA people enough time to get to them if they
> find some of them important enough for a closer look. Hopefully
> most of these would be taken care in shorter amount of time.
> 
> 
> Cheers,
> 
> Marc
> 
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJRBOnzAAoJEGs78UIq7mKy/ggH/34SMLbRKCijHdyNtmHZRLS1
IfbW/1t/g4QEKAmEDWLlngDR40CHa2dsaDysGKTjlGtT635nE2wysLSSvHdfXdKt
es2NiP5a2219hgEtCkDJDOZyTEukTBfx13QtUegJ1/RyCJcJMzOCXu/02a6enymV
QwB/sXWFclQVpFMzfhcyDSzkIrK2j4XmY+3G2I4hBsKoGl2Gt/CCLNJZnNzcqiLC
ZfLxEY2tbRPiGZwNpu70ITCB4e18XqmuPSy7/0cOn6mO16j9ncvDA+ND4yZQvxhm
PoAw2n+WtvL+5IBP6NRMFa4HIxhwNWoF4XUZWDvjtSL+5Zq+Sje0gwyGWFFnJl0=
=ur84
-END PGP SIGNATURE-
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Marc Paré]

Hi Joel,

If you are asking our opinion on these ...

Le 2013-01-26 17:22, Joel Madero a écrit :

Hi All,

During our last QA call we came up with a plan for NEEDINFO bugs that
have been stagnant for 6+ months. I've decided to remove this from the
minutes because there are some concerns about the agreement that was
made. It was requested that we have a discussion through a thread about
what we should do about these bugs.

Let's try to keep this thread moving and organized so we can make a
decision once and for all. Goal should be to minimize flare ups from
users while keeping in mind that our long term QA goal is to get FDO
organized, accurate and under control.

With that, I'll let the thread move forward :-D

Options: (feel free to add)

a) Leave them alone, accept that most will just sit in NEEDINFO forever

b) Automate a reminder for these bugs and leave it at that, if it
doesn't lead to anything, so be it, leave the bugs alone

c) a reminder + automated closing of bug after some period of time

d) a "3 strikes" type system - send one reminder, a month later send a
second with a warning the bug will be closed, then close the bug


These are the options that come to mind.

Best Regards,
Joel

P.S. Florian - please hold off on writing something about our previous
agreement until we reach some consensus here.


+1 for d) BUT make it a real "3 strikes". one reminder, a month later 
send another reminder, a month later send the last reminder with a 
warning that the bug will be closed in one month.


The 3-strikes concept is understandable to most people and I believe the 
submitter would be more understanding if the bug were closed. It also 
gives more than an abundant amount of time to fill out the missing 
details of the bug -- from start to end would take 4 months. This would 
hopefully quiet the noise from disgruntled submitters and allow QA 
people enough time to get to them if they find some of them important 
enough for a closer look. Hopefully most of these would be taken care in 
shorter amount of time.



Cheers,

Marc

--
Marc Paré
m...@marcpare.com
http://www.parEntreprise.com
parEntreprise.com Supports OpenDocument Formats (ODF)
parEntreprise.com Supports http://www.LibreOffice.org

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

On 01/26/2013 02:48 PM, Jack Leigh wrote:

c) a reminder + automated closing of bug after some period of time


+1 Something like automated closing after 3 months in NEEDINFO
Response: From developer side, they'll see a MASSIVE influx of emails, 
pointed out by a core developer. This would need approval from ESC


Possible option: New keyword "AUTOMATED", devs can filter these out 
individually and easily delete



Best Regards,
Joel
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

Re: [Libreoffice-qa] Stagnant NEEDINFO bugs

[Jack Leigh]

c) a reminder + automated closing of bug after some period of time


+1 Something like automated closing after 3 months in NEEDINFO
___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/

[Libreoffice-qa] Stagnant NEEDINFO bugs

[Joel Madero]

Hi All,

During our last QA call we came up with a plan for NEEDINFO bugs that 
have been stagnant for 6+ months. I've decided to remove this from the 
minutes because there are some concerns about the agreement that was 
made. It was requested that we have a discussion through a thread about 
what we should do about these bugs.


Let's try to keep this thread moving and organized so we can make a 
decision once and for all. Goal should be to minimize flare ups from 
users while keeping in mind that our long term QA goal is to get FDO 
organized, accurate and under control.


With that, I'll let the thread move forward :-D

Options: (feel free to add)

a) Leave them alone, accept that most will just sit in NEEDINFO forever

b) Automate a reminder for these bugs and leave it at that, if it 
doesn't lead to anything, so be it, leave the bugs alone


c) a reminder + automated closing of bug after some period of time

d) a "3 strikes" type system - send one reminder, a month later send a 
second with a warning the bug will be closed, then close the bug



These are the options that come to mind.

Best Regards,
Joel

P.S. Florian - please hold off on writing something about our previous 
agreement until we reach some consensus here.

___
List Name: Libreoffice-qa mailing list
Mail address: Libreoffice-qa@lists.freedesktop.org
Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://lists.freedesktop.org/archives/libreoffice-qa/