bug killing

Rick Stockton rickstockton at reno-computerhelp.com
Tue Nov 22 22:08:36 UTC 2011


On 01/-10/-28163 11:59 AM, todd rme wrote:
> I often encounter bugs that are duplicate, fixed, etc, and I post a 
> comment on the bug report to that effect, but the comments seem to 
> rarely be acted upon....
Todd, I cannot imagine another person, currently WITHOUT admin bugzilla 
rights, who is as qualified and capable of using using those rights 
correctly as YOU.

I was facing the same problem yesterday night -- looking at one of "my 
favorite" bugs, in which my final comments indicated exactly why and how 
it should be closed: With no action from the assignee. My solution? Jump 
on IRC, channel #kde-bugs, and point at that particular bug (the 
notorious https://bugs.kde.org/show_bug.cgi?id=34362) as the one I 
needed permissions to close. A 'big guy in KDE' happened to be online, 
even tough it was well after midnight EU time. He looked at my work, 
approved of what he saw, set me up, and the bug is now closed. (Thanks, 
Ben ;) I'm sure that nearly any Admin 'Voice' on #kde-bugs would do the 
same for you, or immediately pull in another person to do the review and 
confirm (if your original IRC contact hasn't already heard of you.)

- - - - - A lengthy side-bar discussion - - - - -

IMO, bugs being erroneously closed is _not_ one of our larger issues 
with KDE Bugzilla. And if a mistake is made, a bug which is closed in 
error can be re-opened quite easily. The lowest-hanging fruit consists 
of the bugs which don't even need 10-line fixes. I'll list 3 groups 
which have come to my little, feeble mind:

1. BugIDs which SHOULD be closed after re-triage and verification (the 
problem no longer exists). It sounds like you have found many of these, 
thanks!

2. Bugs which are so ill-defined that no one can really work on them. 
(Bugs with vast numbers of comments flung at the wall', asking for 
different things, with an assignee who has left them all "still stuck". 
These need re-definition, restricting them to a unique problem and 
solution. (I have an example which I intend to take on myself, the 
almost-as-notorious https://bugs.kde.org/show_bug.cgi?id=48062. It asks 
for Mouse Buttons as modifier keys, which is ONE thing, but many of the 
comments- including a few of mine- talk abut high-numbered as shortcuts 
all be themselves (without another keystroke, without emulating a 
held-down modifier button). The difference between these two wishlist 
items is huge.

3. And the third group, which includes many cases of Group #1 or Group 
#2: The assignee is no longer listening. This appears to be the case 
with #48062, which (after 8 years or so) seems to have it's owner 
ignoring it completely. (He's still an open software guy, but not paying 
attention to activity on THIS bug.) In these cases, we need a person 
with Admin permission (me, or YOU) to re-assign it themselves after one 
final, polite 'Ping?' attempt. Then, after YOU taken ownership and 
responsibility, you take the action which you have explained in your 
comments. I suspect that we have many, many bug owners who have their 
bugs in limbo when they have chosen to move on- a real, REALLY Bad Thing 
To Do.
     One part of the solution for this probably needs to be an emphasis 
that we WILL accept a reverse assignment, from a specific person back to 
'unassigned', if the assignee decides not to work on it. With no whining 
from us, either -- or these assignees will avoid the job by leaving them 
as they are, "trapped" in a non-productive assignment to an email which 
isn't listening, or isn't capable of resolving the problem. The lives of 
KDE people, past and present, are prone to change from time to time. A 
Bugzilla database full of false "assignee" values is MUCH worse than a 
database with these particular bugs sent back into 'unassigned' State.

- - - - - end of side-bar, back to Todd

So- please jump on IRC, get permission from an authoritative person, and 
take control of those bugs where YOU have already determined the 
Action(s) which needs to be done. Or -- someone reading this, and 
authoritative in offering BugZilla permissions - please set Todd up, and 
send him an email when he's got the ability to squash a few- we might 
find that he Resolves _way_ more than just a few.

Thanks to all, I know that I was VERY wordy, but I feel that a thorough 
review of our 'stale assignee' problem could be beneficial.


More information about the Plasma-devel mailing list