A story in numbers
During a 24 hour period:
- One person whose virtual server instance was destroyed when they invoked an init script without the required parameter
- One bug report filed about the problem
- 6 bug metadata changes reclassifying the bug report
- 286 word rant on Planet Ubuntu about how the bug was classified
- 5 comments on the blog post
- 22 comments on the bug report
- 2 minutes on the command line
- 6 commands using UDD (bzr branch, cd, vi, bzr commit, bzr push, bzr lp-open)
- One line patch
- 99% wasted energy
We'll see | Matt Zimmerman 
very well put
hkjhkjhk
April 8, 2010 at 09:42
I agree with the bug poster and the subsequent rant, if there’s something that looks like it can be run and can be triggered without parameters, perhaps its default action should be something safer than scrubbing the filesystem.
It’s equivilent to “rm” defaulting to “rm -rf /”.
I’m not saying this isn’t a lot of wasted puff, but the total energy required should have been bug reported, one confirmation, patch pushed, patch pulled into ubuntu.
I see plenty of moaning that he should have just written the patch but It doesn’t matter who writes the patch, just that somebody does and it gets handled properly. I knoq there’s a question of process here but rather than complaining about it (or even complaining about complaining – as I am now) perhaps we should all move along and get something done.
Oli
April 8, 2010 at 10:19
(PS: it turns out that Scott had committed a different fix for this locally yesterday, so the branch linked above is obsolete)
I’d like to clarify the thesis of this post, which is:
Contributing patches to Ubuntu using Bazaar is really, really easy to do. Therefore, it is really not worth expending energy worrying about what someone else should do. Just do it! All you need is a Launchpad account.
It was not my intent to criticize anyone personally for how they responded to this situation, only to remind everyone that we have enough work to do, and should all spend our energy helping each other to get the work done.
Matt Zimmerman
April 8, 2010 at 10:40
That would work if everyone was a programmer… A lot of us are just users…
Brett Alton
April 8, 2010 at 13:34
There are an awful lot of people around (including folks complaining about this bug) who might consider themselves “not developers”, but who know how to fix a broken shell script. It’s not exactly a deep computer science problem. :-)
Matt Zimmerman
April 8, 2010 at 14:21
There might have been a rant and a lot of complaining, but not everybody is able/wants to code.
Defaulting to Armageddon is not something a script should do and Scott utterly failed to see that and somehow I don’t like the tone in the bug, this post and comment. It feels a bit aggressively defensive and patronizing. It surely is not “Be excellent to each other”(CoC) IMHO.
Tom
April 8, 2010 at 11:49
Just some points for clarification:
1. I don’t think I’m aggregated on Planet Ubuntu, so I guess you found my post on Planet Debian.
2. The rant was not about the bug itself. I thought I made that clear when I wrote “That alone isn’t the most spectacular about it. Bugs happen.”. The most disturbing thing about the bug was the reaction of an Ubuntu developer who is also a member of a TechBoard. As a matter of fact if people of this “importance” are deciding like that it somehow seems representative about Ubuntu at a whole. That was what I was ranting about. Because it does not really speak for the quality of Ubuntu if the reaction on a critical bug is “Its not a bug 1111!!!!11”. Therefore the rant wasn’t wasted energy.
Patrick
April 8, 2010 at 12:05
1. My mistake, it must have been Planet Debian.
2. In my post, I said that the rant was about how the bug was classified (wontfix, wishlist, whatever), not about the bug itself. It still didn’t help to get either problem fixed.
I don’t think Scott would argue that the discussion could have been handled better, but what he actually did about the problem (fixed the bug) was the right response. Most everyone else involved in the discussion did nothing to get the bug fixed. All of the talk was wasted effort compared to the trivial act of fixing the bug, and it would have been better to cut it short or skip it entirely.
Matt Zimmerman
April 8, 2010 at 12:20
Well, its only half of the truth that it was about the justification. He marked it invalid at the first place, so its a matter of fact that his first response was “Dumb user, this bug does not deserve attention”.
And while you argue that in the end Scott fixed the bug its obvious from the bug log that in the first place he *in fact* didn’t intend to do so. So we can speculate about cause and effect here. The question is did he really consider fixing the bug but tagged it invalid anyway? Or isn’t it (more likely) the other way round: He saw the waves his reactions made and went for damage control? I guess its the latter.
So, yes, ranting did nothing to get the bug fixed. But: So what? It was still valid to show up: “Hey, some important Ubuntu developers don’t care if they mess up with your system… unless enough people shout out.” and I don’t want to play a child here, but it was obviously a one-liner from the first look so instead of writing stupid comments on the bug report showing his complete inability to accept a perfectly valid bug Scott could have fixed the bug in the first place as well.
Patrick
April 8, 2010 at 12:39
The bug was only marked “Invalid” because that’s what Launchpad requires in order to move the bug from the “upstart (Ubuntu)” bugs list to the upstream “upstart” bugs list.
Scott James Remnant
April 8, 2010 at 12:55
I’m beginning to think that too many actions on Launchpad “require” someone to Invalidate a bug. In any case, from a user’s perspective, it’s entirely the wrong answer, because it’s interpreted as “your bug is a non-bug, go home!” which is never desirable.
Martin-Éric
April 8, 2010 at 13:42
Leaving aside the human discussion in the bug report for the moment, I’ll say something about this comment here about Landscape workflow.
I feel your pain. We’ve had a similar discussion in Fedora off and on about invalidating bugs in our bugzilla instance when dealing with it upstream instead of in the packaging. I’m both happy and sad to see Launchpad doesn’t completely fix the problem.
Sad because the disconnect between tool required workflow for developers actions and user perception of those actions creates way to much heat.
Happy, because we aren’t alone. Misery does indeed love company.
Perhaps there’s a way to have Launchpad be more verbose about the reason things are being marked invalid. Some boiler plate response saying its being marked as invalid as a package bug so it can be tracked against the upstream launchpad project instead? The action comments are quite terse. And I think if your comment above was placed in the context of the bug report when the invalid happened..it would have dissipated some of the heat.
-jef
Jef Spaleta
April 8, 2010 at 16:25
I have listed the bugs in Launchpad that contributed to the false impressions people got, and I have discussed them with the manager of the Launchpad Bugs team.
https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/557177/comments/33
mpt
April 11, 2010 at 01:43
It’d have been valid to leave the Ubuntu task as confirmed too though. When a GNOME bug is reported in Ubuntu, we link to the upstream bugs.gnome.org bug but *don’t* invalidate the Ubuntu task in Launchpad in the process. So no, you weren’t *required* to invalidate the Ubuntu task.
Mackenzie
April 8, 2010 at 17:25
Guess I should add that, this way you could’ve tracked “ok it’s Fix Released on upstart, but that package isn’t released to Ubuntu yet, so thats still In Progress” which can be useful to know. It’s how every other upstream bug gets handled. Then in the debian/changelog list which Ubuntu tasks have been fixed in this release. Because it *is* useful to know when a fix exists upstream that isn’t yet in Ubuntu.
Mackenzie
April 8, 2010 at 17:32
Scott’s reply was sort of rude and also incorrect because there should no easy way to coincidently delete all you files.
The bug report was posted to hackernews and reddit.
I’m glad you guys fixed it though!
martin
April 8, 2010 at 13:03
Having the technical ability to fix the bug does not grant you the right to be silly with the person which having or not such ability devoted some effort to report the bug.
Those 99% of wasted energy were caused by Scott’s comments.
I am sure Scott does a great job, this was not about the person but about the attitude on this specific case.
João Pinto
April 8, 2010 at 14:23
[…] Matt Zimmerman blogs about the life of a bug in numbers […]
S03E05 – Hear Em Rave | Ubuntu Podcast from the UK LoCo team
April 14, 2010 at 12:02