Re: Bug-buddy feature requests (was Re: Does the bug tracker actually work?)

[Owen Taylor <otaylor redhat com>]

Martin Baulig <martin@home-of-linux.org> writes:

> Owen Taylor <otaylor@redhat.com> writes:
> 
> > The real problem is not "stupid user" - but "stupid bug buddy
> > author allowed people to submit bug reports without entering
> > a description of the problem." (Sorry Jacob - I know you've fixed
> > this, but most of these bug reports are coming from older versions
> > of bug-buddy.)
> 
> I'd like to add some more "feature requests" for bug-buddy here.
> 
> First of all, I think bug-buddy already helped a lot in getting better
> bug reports.  For instance, before the bug-buddy time most of the bugs
> had no useful backtraces at all, they contained no information about the
> system etc.
> 
> However, there are several things which IMHO need to be added to it:
> 
> 1.) Backtraces
> 
>     There should be some way to check whether the backtrace is usable or
>     not.
> 
>     a) When the user has a stripped binary, bug-buddy should just reject
>        the bug report
> 
>     b) When the user has a non-stripped binary, but compiled without
>        debugging, then bug-buddy should print a warning message but let
>        the user submit the bug report if he provides enough additional
>        information.

I don't see the difference between a) and b). In either case,
the backtrace is useless,  but a clueful user could still produce
a useful bug report. (And I've seen quite a few useful bug reports
coming from Red Hat users, where all binaries are stripped. Along
with lots of useless ones...) In either case, bug buddy should not
include the backtrace, include the binary name if it can figure
it out, and ask the user to provide sufficient additional
information.

Regards,
                                        Owen
 
> [ Other good suggestions snipped ]