On Mon, Jan 18, 2010 at 10:13 AM, Caolán McNamara
<caolanm(a)redhat.com> wrote:
> On Sat, 2010-01-16 at 16:01 +0100, Christoph Wickert wrote:
>> I know that APRT is still very young technology, but after 2 months it's
>> time for a interim conclusion. For me the conclusions are:
>
> Abrt's getting a bit of a knocking in this thread, but I'm fairly happy
> with it myself, it's doing its job fairly well, and I prefer it over
> having no abrt. Its not abrt's fault that the crashes are happening :-)
>
> I'd definitely prefer that abrts went to a separate abrt-tracker where
> maintainers could sample the top X crashes for their apps rather than
> getting a bugzilla bug for each one.
>
> And I'd definitely prefer if the UI encouraged more verbose user input,
> e.g. start off the user with some sample text that says "At the time of
> the crash I was...." or something like that. Or a "help" link that
sends
> them to see some successful examples.
>
> There are various things I'd like to see in the bug-report which abrt
> can't/doesn't give due to various concerns, but adding them in abrt-gui
> might be generally "close enough", e.g. GNOME vs KDE, gtk theme, X video
> driver in use, screen depth, whether a11y is enabled or not, which java
> alternative is active.
An ABRT can easily provide that. It supports application specific
plugins where maintainer can specify conf files or other info that the
report should provide. Talk to Jiri or check
fedorahosted.org/abrt for
more info.
AFAIK both pkg-specific scripts (RunApp) and plugins are run from the
daemon (they certainly were when I tried it last time). Unfortunately,
we'd like to get informations from user session...
D.