Before reading further... Are you looking for great Linux hosting from a company that cares about GNU/Linux? Pick Dreamhost hosting, get a 10% bonus to the disk space (and support Free Software Magazine in the meantime!)
Reporting bugs the Debian way
- 2008-05-22
-
Write a full post in response to this!
Following on from my recommendation for apt-buglist—where you can see the reported bugs on a package before installing it—I thought it might be useful to look at the other side of the coin, reporting bugs in Debian. The best way to do this is with the dedicated tool: reportbug.
Bugs happen
Bugs are a fact of software life. Even the most established and mature project will inevitably have bugs—what distinguishes good software management is also how bugs are reported and handled. Debian has a long history of trying to do things “right” (although granted that is a very subjective term) and in “reportbug” I think it has a very simple and effective way of—er—reporting bugs.
One of the issues with bug reporting is getting the relevant information into the hands of the developers. The problem is that users are not always best placed to provide that relevant information. It’s not just what happened that is important but background information like how your system is set-up, what other packages are installed that might create a conflict etc.. Of course trace and system logs can assist with some of this but reportbug is designed to make it easier—even transparent—for the newbie users to provide this information.
Before you start with reportbug—and if you are new to bug reporting—you should probably do a little background reading on what the Debian maintainers expect from bug reports and bug reporting in general. I suggest starting with a couple of articles here at FSM from Woulter Verhelst and Andrew Min and then have a look at the official Debian bug reporting guide.
Installation and set-up
Installation is simply getting the right package, so following on from the apt article, apt-get install reportbug (not forgetting sudo if you are not root) will do the trick. The default package is text-based but there’s also a menu/urwid driven interface as well. If you want to use that you need to install the python-urwid package as well. If you really don’t like the shell then you should look at the gnome-reportbug package (which does what it says on the tin) or the reportbug-ng package which provides a QT-based front end.
Once the packages are installed you need to set-up reportbug. Running reportbug as root is very ill-advised for security reasons, so make sure you do it as a normal user, e.g. reportbug --configure.
Actually this will happen the first time you run reportbug anyway but it saves time to do it before-hand. If you ever want to change your set-up this is the command to run. You will be taken through a few steps:
- Operating mode : think of this as user level, if you have not used reportbug before start with novice. Personally I use standard.
- Interface: here you can choose the text-based or urwid menu interface.
- Internet access: reportbug needs this so generally the answer will be yes.
- Real name & e-mail address: The e-mail address you give will listed on publicly accessible web pages and is thus susceptible to crawler-bots. The address you give will also be sent updates and feedback to your bug report so don’t make it a non-existent or black-hole address either.
- MTA: reportbug will send e-mail so if you’re not running an mail transport agent on the host PC, give it your ISP’s SMTP server address and any associated username details.
Once that is done you are ready to use reportbug so put it away in a cupboard or something until you find a bug.
Reporting a bug
Once you have found a bug—and you are sure it is one—you can fire up reportbug in a shell. If you know the name of the package you can supply it as an argument ( reportbug <packagename> ). This does bring up the only real hurdle in reportbug—you need to know the package name not the application name—but you can easily find this using apt, aptitude or synaptic etc.. If you start reportbug without a parameter, you will be asked for a package name as the first step. Whether you supply it as an argument or not the next step is reportbug retrieving current bug reports for that package. You can then peruse these and view these to see if your bug has already been reported. If it hasn’t you can then move onto reporting it. Note that if you specified that the host machine was not connected to the Internet then it won’t try to retrieve the current bugs.
After briefly describing the problem—which will become the title of the bug—you then pick the severity of the issue ranging from critical (data loss, system breaking etc.) to wishlist. You then get to specify if this bug is related to localisation (l10n) or if you are providing a patch for it.
Next you are given the bug report itself to edit. This is where the background information I was mentioning earlier is handled for you. Things like package version, system information, kernel, architecture and dependencies are all filled in for you. There is a very obvious space for you to enter your more detailed report. Once you have done editing you exit the editor. The final step is to confirm that you want to send the report: saying yes will normally do the trick.
What happens next
What happens next will probably depend on the type of report you sent in and the package maintainer. You’ll get a copy of the report automatically and you might well get a response from the maintainer requesting more info.
Reporting bugs is a very simple way to get involved with the free software community. Using a tool like reportbug makes it even easier for inexperienced users to get involved too. If you ask me it’s an all-round better system than the standard bug-tracking tools where you have to register etc.
Write a full post in response to this!
Similar articles
Do you like this post?
Vote for it!
Copyright information
This entry is (C) Copyright by its author, 2004-2008. Unless a different license is specified in the entry's body, the following license applies: "Verbatim copying and distribution of this entire article is permitted in any medium without royalty provided this notice is preserved and appropriate attribution information (author, original site, original URL) is included".
Biography
Ryan Cartwright: Ryan Cartwright is Head of IT for Contact a Family, a UK National charity for families with disabled children where they make significant use of free software. He is also a free software advocate and you might find him on the GLLUG mailing list. Views posted here are his own - sadly.
- Ryan Cartwright's posts
- Login or register to post comments
- 3317 reads
- Printer friendly version (unavailable!)




Looking for Linux hosting, reviews, coupons, etc.? See out user-voted list
Best voted contents
-
Is Microsoft trying to kill Apache?
Gary Richmond, 2008-08-08 -
How do Drigg and Pligg compare?
Tony Mobily, 2008-08-17 -
The top 4 internet flame wars about free software
Andrew Min, 2008-08-16 -
Creating wealth with free software
Richard Rothwell, 2008-08-05
Similar entries
Buzz authors
All news
Other sites
- The Top 10 Everything (Dave). The good, the bad and the ugly.
- Free Software news (Dave & Bridget). All about free software -- free as in freedom!
- Book Reviews: Illiterarty (Bridget). Book reviews, blogs, and short stories.
Hot topics - last 60 days
-
Don't compare GNU/Linux with Windows or MacOS - they are not in the same game
Ryan Cartwright, 2008-07-07 -
Self-signed certificates and Firefox 3 - a possible solution
Ryan Cartwright, 2008-08-05 -
Why sharing matters more than marketshare to GNU/Linux
Terry Hancock, 2008-08-01 -
Dictators in free and open source software
Tony Mobily, 2008-07-22 -
Why did Javascript/AJAX mop the floor with Java, Flash and Silverlight? Or, why open standards eventually win
Tony Mobily, 2008-07-30
Dedicated server