[Libreoffice-qa] Bibisect Whiteboard Status

[Terrence Enger]

On Mon, 2013-06-10 at 09:49 -0700, Joel Madero wrote:
> Hey Terrence,
> 
> On Mon, Jun 10, 2013 at 9:06 AM, Terrence Enger <tenger at iseries-guru.com>wrote:
> 
> > On Sat, 2013-05-25 at 10:44 -0700, Joel Madero wrote:
> > > I finally got around to updating the wiki:
> > >
> > https://wiki.documentfoundation.org/QA/Bugzilla/Fields/Whiteboard#Bibisecthttps://wiki.documentfoundation.org/QA/Bugzilla/Fields/Whiteboard#Bibisect
> > >
> > > Let's not use all the other options now and stick with:
> > >
> > > bibisected
> > > PreBibisect
> > > bibisectrequest
> >
> > So, considering the possibility of updating wiki page HowToBibisect in
> > line with the reduction in the number of whiteboard values, I wonder:
> >
> > (*) The explanation of whiteboard status PreBibisect says
> >
> >         Only use this f you are using the daily bibisect package or
> >         the bibisect40 package, as the tagged version does not go back
> >         as far.
> >
> >     I suggest
> >
> >         Only use this if you are using the 4.0 bibisect package or the
> >         3.5 bibisect package, as the tagged versions in the other
> >         bibisect packages do not go back as far.
> >
> 
> +1, sounds fine, although no one should be using 3.5 bibisect package any
> longer. The bibiect40 package contains everything in 3.5 so using 3.5 is
> outdated. I believe the wiki says something along these lines :)
> 
> >
> >     If the suggestion is not right, then my understanding of the
> >     bibisect packages is in urgent need of correction.  Help!
> >
> 
> Hm looks like you get it :)

Done

> 
> >
> > (*) Should existing whiteboard words "bibisect40bugs",
> >     "bibisect36bugs", and "bibisect35bugs" be changed to
> >     "bibisectrequest"?  Is it worth the flurry of emails and updated
> >     date-last-changed fields?  Only for open bugs, perhaps?
> >
> 
> No, these mean that bibisect was done,

Actually, these words have not been used at all in the Whiteboard, and
on taking a closer look I see no suggestion that they ever should be.
They are merely the text for links column "Bugs need bibisecting" in
the table of bibisect versions.

With the use of fewer Whiteboard values going forward, the three
rightmost columns in that table ("Bugs need bibisecting", "Posting
result in whiteboard", "Bibisected bugs in range") can be deleted,
each column to be replaced by one point in the body applicable to all
versions.  Is this a good idea?

>                                        not that it's requested. So the more
> appropriate would be "bibisected" but I don't think we should do this, at
> least not yet, with the NEEDINFO project going we're already sending a lot
> of spam to developers and users from FDO - sending even more seems like a
> bad idea atm. In the future this could be useful.

That sounds entirely reasonable.

>                                                    Can you quickly get a
> query together that shows us approx. how many bugs we're talking about - I
> think focusing on non closed bugs would be ideal, why update whiteboard
> status on bugs that have been fixed?

On Monday (2013-06-10) I ran a buzilla query for all bugs with Keyword
"regression" or Whiteboard contianing "bibisect".  Here is the short
version.  Be warned that I found several bugs in my count along the
way; it would be terriby optimistic to think that I have found the
last bug.

                          any status   open   non-open 

      all records               1641    310       1331
                                              
      regression keyword        1617    294       1323
      *bibisect*                 244    104        140
      bibisectrequest             47     41          6
                                              
      prebibisect                  6      3          3
      bibisected*                190     60        130
      bibisected                   9      5          4
                                              
      bibisect35older              1      0          1
      bibisected35*               91     25         66
      bibisected35older           76     23         53
      bibisected35                87     23         64
      bibisected35newer            2      1          1
                                              
      bibisected36*                5      1          4
      bibisected36older            1      1          0
      bibisected36                 3      0          3
      bibisected36a                1      0          1   <== yeah, really: #54651
      bibisected36newer            0      0          0
                                              
      bibisected40*               87     29         58
      bibisected40older            4      2          2
      bibisected40                84     28         56
      bibisected40newer            1      0          1

Terry.


P. S.:  Sorry to be slow responding.  Along the way to the numbers I
found fdo#65619 "segfault querying spreadsheet with LIKE criterion"
and fdo#65685 "constant in select list querying a spreadsheet yields
blanks", and then I learned how to import a .cvs into PostgreSQL.