Uncomfirmed crash report (was Re: [OT] Computer news from Kassel)

[Ben Rubinstein]

On 25/06/2010 17:36, Wilhelm Sanke wrote:
> One thought that led me to put together a more comprehensive report is
> the assumption that at least some of the listed bugs are somehow related
> to each other.
You can always create five separate reports, and then create a sixth one 
suggesting that there are a nest of issues relating to groups, giving the id 
of each.  (Bugzilla is designed to support this kind of usage: if you refer to 
"bug nnn" or "bug #nnn" in the text of the description or a comment it is 
automatically recognised, and made into a link to the other bug report, with a 
tool-tip giving the bug status and title.)

>  And, if someone finds out that "you can crash Rev with
> only four lines of script" this should definitely arouse the attention
> of the responsible members of the Rev team, irrespective whether the bug
> report is multi-faceted or concentrated on one single sub-point. The Rev
> team should be competent enough to deal with a number of related
> troubles at the same time, or deal with them step by step. I believe
> that they are basically capable to handle also complicated issues, they
> are not first-graders in computer science.

But they are busy people, who have to choose where to put their time.  I took 
some time to work through the bug, which is actually called "Groups: Bugs and 
features ("last group" broken)?".  It refers to a way of crashing Rev, but 
doesn't give enough information.  I tried, yesterday, to reproduce the bug 
using the script fragment in the bug report, but without success.  However, 
that may be because I didn't know the definition of 'pre-PNG'. Perhaps I would 
have discovered that by following the clue of "my recent post to this list", 
but that was more time than I had.  I tried with a PNG, and it didn't crash.

Again, I share your frustration that reports can be left 'unconfirmed' (and 
for much longer than a year - my oldest 'unconfirmed' bug report is more than 
three years old, my oldest unconfirmed enhancement request more than six years 
old! (and I still want it)).

But that's not to say that those guys aren't working.  Of the reports I've 
opened in Bugzilla over the years, 155 of them have been resolved - some as 
duplicate, can't resolve, or not a bug, but the vast majority as fixed.  That 
leaves 34 unconfirmed, and 24 in the limbo state between unconfirmed and 
resolved.  That's not a great result: I'd much prefer things at least came off 
the 'unconfirmed' list faster, even if they were then consigned to a black 
hole of low priority; but it's not awful.

(I also know that some bugs have been fixed although the entry in Bugzilla 
hasn't been addressed.  That may be because the bug was found independantly 
either by RR directly or because of a duplicate report, and it's just annoying 
that they've never noticed my report; or it may be that someone read my 
report, fixed the bug, but for some reason didn't follow the process to 
resolve it.)

If you think that RR should be aware of this crashing bug, you would do them a 
big favour by opening a report with the title you mention, with severity 
'critical' or 'blocker' (if the latter can be justified), with the version 
4.5.0 dp3 (having verified that the bug is still reproducible in the latest 
version), and with _all the information that RR need to reproduce the bug in 
one place_ (and as little other information to obscure that as possible).  If 
there's a particularly variety of PNG that is involved, please include all the 
details needed to understand that in the bug report (as well as attaching a 
sample image to the report).

Ben