Today I was chatting with a workmate about some problems we’re having with different SIP softphones in Linux and the sound support, specially with Ekiga, Ubuntu and Pulseaudio.
But it’s not about that what I want to talk today, but about the feeling of my workmate that filing a bug report is useless: I’ve done it in the past, without positive results, he said.
That’s very intringuing for me, because I’ve been involved in several success stories of bugs being fixed, even in hours. Yes, Open Source rocks, but we must know how to deal with it.
So, the question is: why my bug isn’t being fixed? Let’s try to answer this question.
Nowadays it’s very frequent to use a distribution, in my case I run Fedora at home (and Ubuntu and CentOS at work), and when something fails, I usualy file a bug report. The first problem is to find where I must report.
I usually look in Fedora’s Bugzilla, to see if someone has found the same problem. If I find something that may be it’s the same bug, I add myself to the CC list and I try to add useful information in a new comment. In that way I can follow the bug evolution, and may be help.
But sometimes I can’t find something related, so I file a new bug; and the difficulties start.
A full working operative system it’s a complex software, and there are different components that interact together to make possible, for example, that I open an FTP session in Nautilus.
So, when it doesn’t work, first I must guess what component has the problem, and this may not be easy.
Then, I must distiguish between two actors:
- The distributor: thus, Fedora. Packages and integrates software, and provides a full working operative system.
- Upstream: the people that actually develops the software (ie. Gnome, Mozilla, etc).
Let’s say Nautilus has the problem. The next step is to decide who is more likely to fix the problem:
- Is upstream managing bug reports with Fedora’s Bugzilla instance?
- Is a problem related to packaging and, then a problem of the distributor?
- Is it none of the above?
If we’re in case 1 or case 2, we file a bug in Fedora bugtracking and we’re done. Sometimes happens, like most problems with SELinux, or when it’s a problem introduced by the distributor when packaging the software.
Moreover there are chances that, if you blamed the wrong component, someone will point you the right one, and eventually the bug will get the required attention.
But If we file the bug in Fedora, and we are in case 3, most likely our bug won’t be fixed. We’re asking in the wrong place, and it’s possible that nobody will help us.
Every distribution has resources assigned to interface the distribution with upstream, but a Linux distribution is a huge project, and there are chances that these resources are busy.
In those cases, where the bug must be fixed in other place, we must look for the bugtraking for upstream, and file the bug there too, adding a reference in the bug in Fedora’s Bugzilla, so the distributor is aware of the problem (even it has to be fixed in another place).
Let’s see an example: Bug 542205.
- 2009-11-28: I filed a bug, but in the wrong component, and wrong place!
- 2010-02-05: I realized my mistake, so I filed a upstream bug #609085 (in the right component!).
- 2010-02-07: A developer upstream asks for information, I help in everything I can (just two comments). The bug gets fixed the same day.
- 2010-02-09: A package with the fix hits Fedora testing.
- 2010-02-23: The package it’s pushed to updates!
I think we can verify a couple or more things:
- The bug in the wrong place (and in the wrong component!) will get rotten without any hope of being fixed.
- After the bug was in upstream, I got a fix in about 2 days, and 2 days later I had a package to test it in my system (actually it fixed the problem).
- Once QA it’s done, it hit updates and all the Fedora 12 users would benefit of my bug report.
To sum up, I would say that the main reason your bug isn’t being fixed is: the bug report is in the wrong place, and nobody is there to help you!
I know there may be other factors, but overall I think everything I have explained makes sense. I have experience with QA and QC, and open source software adds some complexities because of its bazaar nature that doesn’t exist in a privative world. What do you think?
Disclaimer: I’m not involved in Fedora QA team, and this post it’s result of my own experience.
- diachrony reblogged this from usebox and added:
- diachrony likes this
- usebox posted this