[Libreoffice-qa] Version field options way too fine grained

Terrence Enger tenger at iseries-guru.com
Fri Mar 14 12:03:29 PDT 2014

Hello, all,

On Fri, 2014-03-14 at 10:11 -0700, bfoman wrote:
> Michael Stahl-2 wrote
> > definitely a problem, but i think it's a pretty fundamental limitation:
> > bugzilla simply has no concept of branches.
> Hi!
> Bugzilla branch support (called sightings):
> - progress - follow https://bugzilla.mozilla.org/show_bug.cgi?id=55970
> - code -
> http://git.mozilla.org/?p=bugzilla/bugzilla.git;a=shortlog;h=refs/heads/sightings
> Unfortunately seems not finished yet but merged recently with up-to-date
> code.

>From the specification attached to that bug report
<https://bug55970.bugzilla.mozilla.org/attachment.cgi?id=304328>, the
basic idea is to have separate underlying bugs with an integrated
interface making it easy to clone off an "also-associated" bug or to
copy information between bugs in the set of also-associated bugs.

I do not find this attractive, but then I am an unimaginative old man
with a prejudice against keeping multiple copies of information. ( <grin
attitude="wry"/>  And, yes, I am grumpy too. )

Perhaps more seriously, There is no way to record places where a bug
is *not* evident.  This matters to us not so much for versions--we
have bibisect for that--as it does for platforms, configration
parameters, and whatever.  For example, there is no easy way to
distinguish a Windows-only bug (I have nothing to offer), from a bug
reported on Windows and unconsidered elsewhere (I may have something
to offer).

<enter object="the_weeds">
  (*) "Sighting" is too positive a word for what I am talking about.
      Let us say "observation" for now.
  (*) Each observation is associated with a specific build running on
      a specific platform.  Because that is what is in front of the
      person reporting the observation, of course.

Just my C$0.02,

More information about the Libreoffice-qa mailing list