Bug 244522

Summary:	Dr Konqui sometimes lists false possible duplicates
Product:	[Applications] drkonqi	Reporter:	Jeffrey <eljefedelito>
Component:	general	Assignee:	Unassigned bugs mailing-list <unassigned-bugs>
Status:	RESOLVED NOT A BUG
Severity:	wishlist	CC:	mail, sitter
Priority:	NOR
Version:	unspecified
Target Milestone:	---
Platform:	Debian unstable
OS:	Linux
Latest Commit:		Version Fixed In:
Sentry Crash Report:

Description Jeffrey 2010-07-13 19:35:37 UTC

Version:           unspecified (using KDE 4.4.4) 
OS:                Linux

Amarok crashed and I clicked on a report that (per the Subject) looked like a duplicate:
https://bugs.kde.org/show_bug.cgi?id=240416

I was asked to create a new bug report so I did, here:
https://bugs.kde.org/show_bug.cgi?id=244520

but in the initial report I was told they're completely different (so why did Dr Konqui bring the option of duplicates?) and I was told the Dr Konqui attachment wasn't searchable, which seems silly.

Reproducible: Didn't try

Steps to Reproduce:
Find a bug.
Suggest that one may be a duplicate but if it isn't nothing stops you

Actual Results:  
The bug is reported as a duplicate and the backtrace is attached.

Expected Results:  
Backtraces are compared and the likelyhood of a duplicate is rated, or those that have different backtraces are filtered out.

Also, the backtrace should be searchable text (not attachments I guess)for the KDE and app developers.

Thanks for KDE, I am just hoping to make this tool a bit more usable for everyone involved.

Comment 1 Dario Andres 2010-07-13 19:51:09 UTC

Those are different bugs/problems:

- Text Attachments not being searchable is a Bugzilla limitation. In any case, for KDE SC 4.5, part of the backtrace is also pasted inline; so that part can be found by the standard search method; and the full backtrace is still attached as a file. (so, that bug is FIXED)

- DrKonqi told you that it was a similar bug, because some functions mentioned in your backtrace (your Amarok crash) were found in bug 240416; however, due another bugzilla limitation, we can't determine if the functions are in the same order (that would help us to guarantee it is a real duplicate and not a  false positive like in this case).

A possible solution would be, for every report listed as "possible duplicate", DrKonqi could try to fully download it and compare the backtraces properly; so it could be able to determine if it is a real duplicate. 
However, that will cause a lot of CPU/Bandwidth usage for the Bugzilla server; so I don't know if it will worthy to implement. (considering that the cases of false positives are rather minimum)

Regards
Dario A. (former DrKonqi dev)

Comment 2 Harald Sitter 2019-12-15 10:52:30 UTC

As mentioned, solving this comes at an unreasonable cost to bandwidth and processing power which in the end means reporting takes much longer in general. This seems like a bad trade.