[kde] [Bug 315032] New: KDE crash handling auto-bug-reporting is intimidating and counter productive

[John Bowler]

https://bugs.kde.org/show_bug.cgi?id=315032

            Bug ID: 315032
           Summary: KDE crash handling auto-bug-reporting is intimidating
                    and counter productive
    Classification: Unclassified
           Product: kde
           Version: Git
          Platform: Gentoo Packages
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: NOR
         Component: general
          Assignee: unassigned-bugs at kde.org
          Reporter: jbowler at acm.org

nepomukservile just crashed on me and the "crash report" dialog sprang up.

Ok, so far, so good.  I happen to be running an unstable Gentoo so I've got
4.10 (*not* GIT head, but 4.10 isn't in the menu above!)  I also run a complete
debug system - full debug info (PDB in MS terms) on every single thing on the
system.  So I know the crash was in side clock_gettime from the backtrace.

This certainly looks like useful information (I suspect I could debug the crash
in about half an hour) but you guys aren't getting it because the automagic
crash report made me say WTF, it's your problem, within 30 seconds.  Here are
the problems:

1) The information isn't simply filed somewhere for later analyse, it's used
(apparently) to generate a full-fledged bug report.  I guess you don't expect
many users to use it.  However;
2) The report requires a log in.  I happen to have one, of course I can't
remember either the user id or the password, so I'm not very likely to do
anything other than cancel the whole process, but I did actually go to the
trouble of rediscovering it, and logging in (this had already taken too long),
whereupon;
3) You present me with a whole load of bug reports which I'm meant to review (I
guess nepomukservices crashes a whole load of times.)  Gee.  There's a symbol
in there at the top of the stack but;
4) Guess what, I'm not doing that.  If I had that amount of time to spend I'd
simply debug the crash  myself.  In other words it would take me longer to
*report* the bug than to fix it.

John Bowler <jbowler at acm.org>

-- 
You are receiving this mail because:
You are the assignee for the bug.