Bugzilla dupes attack

[Jim Cornette]

Mike Klinke wrote:
> On Tuesday 14 February 2006 11:41, Rahul Sundaram wrote:
> 
> 
>>Save us a few hours of bugzilla grunt work by not filing two
>>hundred and seventeen duplicates for a trivial bug. Thanks.
>>
>>https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=181064
>>
>>If anyone has a list of good improvements that can be made to
>>bugzilla to help is triaging let me know.
> 
> 
> Over the years the single biggest complaint voiced on these lists is 
> the poor search results bugzilla provides.  If the searches 
> provided better results perhaps duplcatess wouldn't be so 
> prevalent.
> 
> Regards, Mike Klinke
> 

I agree with the search feature for bugs. Bug searches could return hits 
for bugs filed against rawhide with dates related to the bug era 
(fc5t1,2 or 3) since these should be test release bugs in reality. This 
would cut down on the duplicate bug filings.
Rawhide since yum was adopted does not allow rawhide to be rawhide 
during test phases. It is the test release update source also.
I think having actual update repos for the test releases would eliminate 
this factor.

Bugzilla itself would work better if it would either group all 
duplicates in some type of directory structure or would merge 
information from one bug that is a duplicate of another bug into a 
common report.
Another ideal condition would be linking all these similar reports into 
a common "best of this bug" (Useful data for resolving the bug) for 
commonly duplicated bug problems. There are some developers which have 
done something similar for highly duplicated problems and the results 
were positive. (gcc and mga/intel problem comes to mind)

Jim

The universe does not have laws -- it has habits, and habits can be broken.