ubuntu-bugcontrol team mailing list archive

[Milan Bouchet-Valat <nalimilan@xxxxxxx>]

Hi both, and thanks for you kind reviews!

Le mercredi 04 novembre 2009 à 18:33 -0600, C de-Avillez a écrit : 
> On Wed, 2009-11-04 at 17:14 -0500, Scott Howard wrote:
> 
> (shamelessly piggybacking on Scott's reply)
> 
> > Hi - thanks for all your work! Getting straight to the bug list:
> > 
> > > https://bugs.launchpad.net/ubuntu/+source/system-tools-backends/+bug/434565
> > stellar work on this one! Debugged, triaged, fixed, followed-up and
> > still helping people
Maybe that's because the bug come from my code in the first place... :-p Just joking, I can be quick on other bugs too.

[...]
> > > https://bugs.launchpad.net/ubuntu/+source/shadow/+bug/432964
> > A difficult one, good job referring to the wiki for crashes and x.org.
> > I, too, am stumped - but have you tried asking around on #ubuntu-bugs
> > or find an ubuntu xorger to take a look at it?
I think this has to do with the way gdm-guest-session works, i.e. it must be 
locking /etc/passwd while running to prevent conflicts. I've not really had
the time to investigate, and I don't think Martin Pitt has, either.
Later in the cycle, maybe.

> All good work. Thank you.
> 
> > > -> I'd have set the importance to Medium: while the case to trigger the
> > > bug (system crash during guest session) is quite rare, the effects are
> > > terrible, making it impossible to modify/create users until the lock
> > > file is manually removed, which would really be a deal breaker for most
> > > users.
> > I might have even put that as "high" (has a severe impact on a small
> > portion of Ubuntu users)
Noted, I'll be more severe - anyway I'm generally inclined to do so.

> Only a few comments, Scott pretty much said it all:
> 
> (1) please do not forget that coredumps carry a *lot* of data; most of
> the times, if not all, including quite private data.
> 
> By default, bugs containing coredumps should *never* be made public.
> 
> (it is amazing the amount of applicants to -control that *never* mention
> coredumps... I am pretty sure you know this, and just forgot to point it
> out.)
That must be because we think of core dumps as a kind of stacktrace, since
that's what they're used for. Also, the Wiki page is not clear about privacy
issues related to core dumps : it only explicitly mentions stacktraces. See
https://wiki.ubuntu.com/Bugs/HowToTriage#Apport%20crash%20reports

> 2. I am not sure you stated the Importance for a single bug, or for all
> of them. Nevertheless, given the quality of your work here, I will not
> nitpick on it :-\
I was only referring to the last bug of the list, which was the only not
to have an Importance already set (as asked in the instructions).


Cheers