125637 - INVALID is underspecified

Status: RESOLVED DUPLICATE of bug 108514

(bugzilla.mozilla.org :: General, defect)

Description

[Ben Bucksch (:BenB)]

The current definition of INVALID, as defined on
<http://bugzilla.mozilla.org/bug_status.html>, is:
"The problem described is not a bug".

This is an underspecification and even strictly wrong. If you read "bug" as
"error on the software", then all RFEs were invalid. If you read "bug" as
"record in bugzilla", then no bug were invalid. Both is obviously not the case.

The definition I use and I think most of the people at mozilla.org use is "This
bug does not belong into bugzilla at all, and never did". This is still an
underspecification, so suggestions are welcome.

The problem is that some people (esp. from the XPApps group, it seems) like to
close perfectly valid bugs as invalid. Sometimes, they are suggestions (but with
the Severity incorrectly being "Normal") with which a developer just disagrees
(they should be WONTFIX, at most). Or a bug which was valid when filed, but does
not make as much sense anymore due to other circumstances, is being closed as
invalid (e.g. bug 40022).

I care, because some people like to track their bugzilla statistics.

Comment 1

[Bradley Baetz (:bbaetz)]

This is a mozilla.org issue, since installations can put whatever meaning they
want on the defaults.

The statuses will be customisable for bugzilla2.18.

Comment 2

[xyzzy]

Invalid should refer to bugs that can not be properly addressed for one reason
or another.  No implementation decision can be made, and no rectification can be
made.

Examples of invalid reports would be:

1) Report describes problem in another product, not fixable by mozilla.org.
2) Report is logically inconsistent.
3) Stated behavior does not occur ever.
4) Report misunderstands correct behavior.
5) Bug is a bogus report, nothing to do.
6) Behavior was fixed before this report was filed.

By contrast, there are bugs that can be addressed, and require decisions to be
make.  Decisions can be made by the specs, or by the component owners.  An
excecutive decision not to implement a request is a WONTFIX.  Proposals exist
for a new resolution where the spec is the final arbiter (perhaps PERSPEC?).

See also bug 119305.

Comment 3

[Dave Miller [:justdave]]

I would argue that situation #3 described in the previous comment (the stated
behaviour doesn't ocurr) should be a "WORKSFORME" and not "INVALID".  I'd agree
with the rest though.

Comment 4

[mental]

Too busy thinking INVALID, not enough time thinking DUPLICATE...  ;)

*** This bug has been marked as a duplicate of 108514 ***