Curiosity re: abrt ??

[Michael Schwendt]

On Wed, 13 Jan 2010 12:16:22 -0500, William wrote:

> Hi;
> 
> I was just wondering what happens to the abrt generated bugs.
> 
> As a practical matter, are they treated any differently from a manually
> filed bugzilla report?  Does abrt's reports have a higher priority?

If and only *if* the reports are more detailed [and more complete] than
manually filed ones, they are more useful and are likely to get priority
treatment from a package maintainer.

Recently, however, there have been many close-to-useless reports from
ABRT. With backtrace attachments created with dozens to hundreds of
missing -debuginfo packages. Without any steps on how to reproduce a
problem. Also, I've learned that an increasing number of users delete
tickets in ABRT immediately after they have submitted them to bugzilla,
even if missing relevant details.

> I seem to be generating a lot of abrt reports.  I cancel about
> two-thirds of them because I don't think the particular application was
> at fault, but that the crash was brought on by a buggy old f*rt user --
> e.g. hitting the wrong key or changing my mind and cancelling an
> operation.

Doubtful. Cancelling/interrupting something should not make the software
crash. Well-written software can deal with well-defined interruptions and
can either retain a sane sate or exit gracefully.
Some pieces of software, however, e.g. rapidly developed Python programs,
don't handle interruptions (such as Ctrl+C) everywhere and exit with
fatal run-time errors instead (Python Exceptions, which trigger ABRT, too).

> I am having some second thoughts that maybe I should be sending these
> type of errors along as well.  I think I remember reading somewhere in
> life's manual that the world should even be able to accommodate old
> f*rt's mistakes.

If you can provide the steps on how to reproduce an issue (in addition
to the backtrace details), the software developers ought to have interest
in that.