On Thu, 13 Jul 2006, Roland Mainz wrote:
Karyn Ritter wrote:
[snip]
... but this should not be neccesary. Loossing bug reports like
this is
one of the WORST possible scenarious when dealing with customers. And
it's not the first time that this happened to bugs.opensolaris.org so
some action should be taken here...
I apologize... What is the issue exactly? Did you not receive
confirmations of emails that you filed? Did the server go out to lunch
mid-way through submitting the problem (and thus what you were entering
was lost)?
No, the server returned the usual "bug submitted" message and then
nothing happend - no confirmation email and no entry in Sun's DB,
neither visible on http://bugs.opensolaris.org nor internally within
Sun.
(And yes, I am slightly upset at the moment since the |getcwd()|
statistics for
http://mail.opensolaris.org/pipermail/perf-discuss/2006-July/000611.html
are gone, including the prototype patch (yes, it's my fault that I
didn't keep backups for that around - I simply assumed that the web form
does something usefull and isn't linked to /dev/null ... ;-( )).
It's not linked to /dev/null... but not being intimately familiar
with it, I can't say what it is linked to. I'll have to leave those
folks most familiar with this to comment on this.
FYI -- Once we have automated it, it will still take up to a few
minutes
for you to get the confirmation as we don't have a way for the website
to interact directly with the bug database.
<rant>
Which raises the question why Sun is still using "bugster" ... almost
EVERYONE (in- and outside Sun) is complaining about it but it seems no
effective work is done to get it replaced with something decent.
<disclaimer>
I am a solaris development engineer, but I was Solaris's rep to the
bugster
team for 5 years while they worked on the implementation. So, I didn't
make the decisions, but I know what happened </>
So, not everyone hates bugster. There are issues internallly
with uptime, but in general *most* internal users prefer it to
our old bugtracking utility. There are some that miss BT+,
but they also still use mailtool to read their mail ;-)
Most of the issues initially seen have been hammered out, and the
tool is much more feature rich than what it replaced. The issues
we are seeing now are mostly related to the team supporting it
and uptime.
It took forever to come up with bugster, because we were taking
a call management software (square peg) and making it do bug
tracking (round hole). But, what is done is done.
Apparently companies like Novell and IBM simply migrated to custom
bugzilla versions while the last thing I heared is that Sun may rewrite
"bugster" ... which would be another incarnation of the "reinvent the
wheel"-disease at OpenSolaris.org (other incarnations include "Jive"
Sun's BugTraq2.0 (which became bugster) did investigate
using bugzilla at the time, which would've been a much cheaper
option. Many issues came up at the time, that were not
feasible to work around. Now, this is 6 years ago, so I don't
remember them, but at the time the limitations seemed realistic
from an engineering perspective. I believe they related around
having different views for different types of info (confidential
vs not), managing of escalations and multiple releases, and attachments.
Bugzilla may be loads better now, but at the time it
was not the correct solution for Sun.
Also, there is no talk from the organization that owns the bugster
code of rewriting it. Sun invested in Siebel, and that's what
we are sticking with for now.
(which is still unable to handle some email headers correctly, almost
nine months since the first report), the source search engine (which is
nice but lacks simple capabilties such as searching for symbols like
')(+-<TAB><SPACE><NEWLINE>' etc.) etc.) and the TML-monstrosity from
which the project webpages are made from.
</rant>
Now, here is where I note that bugs.opensolaris.org is NOT
bugster. Incoming bugs are, I believe, handled manually and
*put* into bugster. Bugster has webapis for scripting around
internally, but I believe incoming bugs on bugs.opensolaris.org
need manual processing. Part of this, IIRC, is due to the
large number of categories & subcategories Sun has for tracking
Solaris issues. There are good reasons for this, and we
are working to streamline & make things more intuitive, but
that's what we have now (inherited from 20+ years of bugtracking
in Sun :-)
I'm not sure where your bugs went... and I hope
they turn up... but I'm pretty sure bugster was not
to blame.
Valerie
