Hello, I'm DanielHolbach, taking care of bugs of a lot of different teams:
Unconfirmed/Undecided new bugs
I started off my day looking at the raw incoming Unconfirmed/Undecided bugs sorted by 'newest first'. These bugs usually were not touched by anybody else before and you'll find a lot of easy tasks to get done there.
With the advent of Apport we get more and more crash reports. It's our task to make sure
- we identify duplicates,
- the backtraces are complete,
- we forward them upstream.
Most of these bugs have [apport] or crash in the title, so are easy to identify.
Initial triage (*)
Looking at the list of today, I stumbled across bug 94360.
To make sure that nobody else goes through the initial triage steps of Unconfirmed/Undecided bugs, I marked its Importance as Medium. (For that you have to be in the UbuntuQA team.) The bug is still unconfirmed and we will mark it as Confirmed once the backtrace is complete.
The crash report says Architecture: i386, so I tag it as need-i386-retrace and the automatic retracer will pick it up.
- Now I moved on to the next bug, as it will take a while until the retracer gives us a complete debug backtrace.
Another bug I looked at was bug 94379.
It was similar to the bug above, only this time, I assigned the bug to the desktop-bugs team in the same step.
- Also, I thanked the reporter for the bug report.
bug 94376 was roughly the same, its attachment told me that it's a i386 machine too. Additionally I assigned it to MOTU, so the bug shows up on the universe-bugs@ list. How I found out that it's a bugs for MOTUs? Easy:
daniel@bert:~$ apt-cache show multitail | grep File Filename: pool/universe/m/multitail/multitail_4.2.0-1_amd64.deb daniel@bert:~$
For a comprehensive list of commonly used tags, take a look at BugSquad/Tags.
For a list of teams and their bug workflow, see Bugs/Teams.
Using bughelper to find duplicates (*)
Traceback (most recent call last): File "/var/lib/python-support/python2.5/invest/widgets.py", line 161, in on_quotes_update day_var = (now_total - start_total) / start_total * 100 ZeroDivisionError: integer division or modulo by zero
Deciding that ZeroDivisionError is a search term that's easy to distinguish, I started this BugHelper query:
bughelper -T gnome-applets ZeroDivisionError ZeroDivisionError -A -U -p gnome-applets
-T <package name> <search term> <display this string, when something's found> means: only search for this search term in bug reports for this package -U means: not only consider bugs in http://launchpad.net/ubuntu/+source/<package>/+bugs but also in http://launchpad.net/<package>/+bugs this might give you duplicates in the output, but you'll check upstream bugs too -A means: check attachments too - this means downloading lots of stuff, but it will get cached in ~/.bughelper/attachments-cache -p <package> means: search for this package only.
After a while (I looked at other bugs in the mean time), bughelper gave me this output:
http://launchpad.net/bugs/80407 [gnome-applets upstream: Confirmed/Unknown][gnome-applets Ubuntu: Confirmed/Medium] - ZeroDivisionError - http://librarian.launchpad.net/6455220/_usr_lib_gnome-applets_invest-applet.1000.crash - http://librarian.launchpad.net/5780843/_usr_lib_gnome-applets_invest-applet.1000.crash
Bingo! Bug 94300 really was indeed a duplicate of bug 80407. Thus I marked it as duplicate, and used a nice stock reply from Bugs/Responses.
Adding the bughelper clue to the 'database'
To help other BugSquad members to find this bug more easily, I used the following command to create a gnome-applets clue file:
bugxml -a gnome-applets ZeroDivisionError "This is a dup of bug 80407" -d 80407
For more documentation on it, take a look at: BugHelper/doc/writing-clue-files.