Why should I ever bother filing another bug?

Tue Nov 9 19:50:45 UTC 2010

On Tue, 2010-11-09 at 14:33 -0500, Felix Miata wrote:
> On 2010/11/09 08:45 (GMT-0800) Adam Williamson composed:
> 
> > There isn't a prize system for bug reports.
> 
> Actually there is. When someone files a good bug, as opposed to one that 
> requires more than trivial attention due to significant missing or invalid 
> information, achieving fixed status is an informal statement that the filer's 
> effort was valuable. It getting duped to a newer bug that isn't materially 
> different in quality is an informal statement that the filer's effort was not 
> appreciated, thus the subject.

You may interpret it this way, but with my maintainer hat on, I can tell
you it's simply not how most maintainers look at things. Maintainers
want to fix bugs in the most efficient way possible. They're not
thinking in terms of passing judgments and handing out prizes to the bug
reporters; there is no judgment on the initial reporter implicit in
their duping decisions. Just as Lubomir noted when he did this
particular dupe: he didn't say anything about the quality of the
*initial reports*. He preferred the newer bug because it had fewer messy
side discussions and useless +1 comments. In this particular case we can
confidently say that, based on direct evidence, your interpretation of
the duplication decision as an implicit judgment on the quality of the
initial bug report is incorrect. In the general case, I think we can
still confidently say, given the priorities and workflow of most
maintainers, that your interpretation of duplication decisions in
general as being conscious implicit judgements on the qualities of the
initial reports is also usually incorrect.

> > So, once more for the cheap seats: *why* do you think closing an older
> > bug as a dupe of a newer one is a respect issue? What's the big problem
> > with it?
> 
> When I get a list of all bugs I filed, there's nothing AFAICT that can tell 
> me whether a dupe to a newer is really a fixed - there's no way to get into a 
> buglist the status of a bug duped to. Thus the report misleads about the 
> quality of effort put forth, a significant part of which is ensuring against 
> filing a duplicate by competent searching prior to filing.

Sure. But this is only special to you from *your* perspective. From the
disinterested bystander's perspective, this suckage is going to happen
to all the reporters but one; whichever one happens to be the lucky one
whose bug stays open isn't really important from the overall
perspective, the ratio of people for whom the end result isn't perfect
to people for whom it is does not change. Bugzilla is a long way from
perfect, practically speaking it's best to recognize that it has
limitations that you're going to have to deal with, and these do not
reflect anyone being personally out to get you.

> When anyone is attempting to assess the overall efficiency of Bugzilla, one 
> aspect is the open time. Duping to newer misleads about this too. 

Okay, that's an actual reasonable concrete objective point, and you're
right. In practice, we run very few metrics on Bugzilla, so I don't
think this is a huge issue in Fedora's case, but yes, that's one factor
which should work in favor of duplicating to the oldest report when all
other factors are equal. I wouldn't assign it as much weight as you do,
but it's certainly a factor.

> It 
> shouldn't happen unless the newer is clearly better. 50+ CCs, some of which 
> are from dupes of newer bugs, and little metoo or other noise, should be a 
> pretty good condition that a bug is good enough that justification for duping 
> to a newer should be overwhelming to be allowed.

Again, from the developer's perspective, metoos are usually a negative,
not a positive. They just make it harder to find the *useful* input. A
comment which provides information that helps to debug the problem is
useful. A comment with an attached patch is really useful. A comment
which says "I'm having the same problem as the last 500 people to
comment on this bug" is usually not at all useful and just makes it
harder to find the good comments. (It can be useful if a bug's been open
for three years and you're not really sure whether it's still valid, but
when 15 people have told you in the last two days that they're having
the problem, a 16th person saying that they're also having it really
isn't contributing anything helpful).

> Plus https://bugzilla.mozilla.org/show_bug.cgi?id=108983 & 
> https://bugzilla.redhat.com/show_bug.cgi?id=523634

again, yeah, Bugzilla sucks. But then, if you want to look at it
logically, the sensible response to this bug is 'make the 'original' the
bug with the largest number of subscribers', not 'make the 'original'
the oldest bug'. The two are not always the same.

And Lubomir likely wasn't even aware of the above bug when he did the
duplicating. It seems most people aren't. So again, you have no reason
to ascribe personal motivations.
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net