Re: [gentoo-dev] Re: [Bug 483274] app-text/poppler-0.22.5 Please stabilize

[Tom Wijsman]

On Mon, 16 Sep 2013 09:19:47 -0400
Rich Freeman <rich0@g.o> wrote:

> On Mon, Sep 16, 2013 at 9:09 AM, Michael Palimaka
> <kensington@g.o> wrote:
> > At what point do we draw the line? Today my mailbox is full of
> > email with changes like "app-foo/bar-1.2.3: version bump" ->
> > "app-foo/bar-1.7.3 - Version bump.", changing keywords on years-old
> > bugs etc.
>
> Jer - can you comment on how these changes are getting made?  Is this
> some kind of script, or are you manually making these changes?

Two things here need to be clarified:

 1. It's not Jer that is solely making changes here.

 2. The changes being discussed are not solely Jer's changes; also,
    the opposite applies as well, Jer doesn't do all these changes. 

So, what was quoted is somewhat a mixture of changes happening; and the
problem with discussing this mixture, is that it covers more than what
this thread is actually about. Let us try to avoid false impressions.

I do all of these changes when I do bug wrangling work (on b-w@g.o
assigned mails) or bugs on packages that I maintain (with not too big
herds), but I would never do them on packages that I don't maintain. 

So, just to be clear, not all the quoted changes are done in that last
case; and I would like us to focus on the changes that are, as to avoid
people that are not involved to assume things that did not happen.

> Either way, could you consider adjusting your practice so that if the
> only change is to add a period or other trivial modification that you
> avoid it for the moment?

Example of the above text, I haven't seen such modifications happen in
the case of someone editing a package that that person doesn't maintain.

> Additionally, can anybody who knows more about bugzilla comment on how
> easy it would be to have some way to mark modifications as trivial,
> and take this into account in the bugspam?  Either make it
> configurable as to whether users get trivial bugspam (ideally), or at
> least stick something in the headers that can be filtered on.

There has been a discussion on this, I don't recall where it was held
though; whereas the changes itself can be easily patched, the problem
here rather has to do with the monitoring of those changes.

Currently the approach is being used where you see all that what
happens; if you introduce a trivial checkbox, you get the same effect
as with Wikipedia and you need to have some means to check through
those changes or otherwise behavior that's not encouraged could be
happening low profile.

I'm quite sure we are all fine and wouldn't do such thing; but a new
developer or editpriv user might be unaware of the particular use case
the feature was introduced for and therefore might misuse it, we
don't want that to happen unnoticed. Which makes me think...

What if we could introduce this in a restricted way?

Perhaps only allow bug wranglers (and comrel, admins, ...) to use it?

If the feature is rather limited, it becomes much more maintainable in
terms of checking the changes that happened there; and because it is
only given to a small group that is aware of how to use it conform to
consensus, it avoids any unintended or unaware misuse by design.

-- 
With kind regards,

Tom Wijsman (TomWij)
Gentoo Developer

E-mail address  : TomWij@g.o
GPG Public Key  : 6D34E57D
GPG Fingerprint : C165 AF18 AB4C 400B C3D2  ABF0 95B2 1FCD 6D34 E57D