Submitting Bug Reports To Open Source Projects?

aldheorte writes "After installing Red Hat Linux 8.0, I discovered some minor bugs. Some of these are with software actively maintained by Red Hat (e.g. redhat-config-date), but some are not (e.g. gaim). Although it is possible to enter bugs for any package at Red Hat Bugzilla, some of these packages have zero bugs, which probably indicates this is not a preferred method of receiving bugs for that project. In fact, I've found this to be the case for for several project. I find no listed bugs for Red Hat's Bugzilla and a whole database of bugs at another site, such as SourceForge. There are many distributions and channels for open source projects to reach the end user, so how do users, especially non-technical ones, effectively submit bug reports to the right database? How do open source projects make it easier for users to submit bug reports and consolidate the bugs in a single database?" Update: 11/01 11pm EDT by C :Don't know why this was sitting under the "HP" topic, so I've changed it to something more appropriate. Sorry if this has resulted in any confusion.

man pages

[BlueLines]

read the man pages. usually there's contact info for the maintainer of the actual program. also, always file the bug with your vendor as well, so they have a chance to upgrade their shipping versions.

-BlueLines

They Don't/Shouldn't

[scott1853]

so how do users, especially non-technical ones, effectively submit bug reports

I get the lucky job of also providing tech support for the software I write. I get a lot of users calling up and saying "I got an error printing a report", which leaves me having to ask, "which of the 50 reports and what does the error say". At that point the customer needs to walk back to his office and turn on his computer since he thought I could magically solve the problem without any information and remotely control the little gnomes in his machine and instruct them to magically fix it.

How many open source developers, most of which develop the software for free, want to deal with people that are not technically savvy enough to read the documentation for the software to figure out where to submit bugs to?

Of course, I'm not an open source developer so maybe they like dealing with dumb users and I'm just talking out my ass. It's happened before ;)

Call him up, be nice...

[dasmegabyte]

A lot of these open source projects are maintained by one or two people. Many of them are in the phone book or have an email address lying around. You might as well just contact them directly.

It's not like the commercial software world, where there may be hundreds of employees and a series of support levels. The developers are all there is, and they may not check all the available bug watch sites because they would rather concentrate on making a better piece of spare time software. Contacting them directly will not only alert them to the error, but probably flatter them as well.

I got an email a while back from somebody who had been using a freeware encoding translation app I wrote a while back as an essential piece of a corporate mailing package. It was very cool to see how they were using it and how different it was from the original intent. Eventually, I arranged for the fix I suggested and he wrote to go up on the sparsely updated freeware site I had set up at my university.

Of course, he was willing to fix the bug in this ancient software himself with a little input. If he had come at me with a lengthy email accusing me of writing buggy software or threatening legal action or demanding a fix on code that really was dead to me, etc, I probably would have ignored him.

By the way, you hit the nail on the head of the anti-OSS argument here. There is really nobody accountable for these bugs, legally or otherwise. You're relying on the kindness of strangers, and if they aren't willing/don't have the time to fix it themselves, you're going to have to pay to have somebody else do it.

Re:Call him up, be nice...

[jez_f]

By the way, you hit the nail on the head of the anti-OSS argument here. There is really nobody accountable for these bugs, legally or otherwise. You're relying on the kindness of strangers, and if they aren't willing/don't have the time to fix it themselves, you're going to have to pay to have somebody else do it.

Relying on the kindness of strangers does not seem like the best idea. BUT quite a lot of OSS developers consider products to be 'their babys' and take some pride in them.

Some companys support is good and you get good value for money. Others are usesless. It would be nice to see a comparison of the life cycle of an open/closed source bug.

Forget Gentoo

[Anonymous Coward]

I filed a bug for Gentoo and the person maintaining the package was a total jerk about it. He copped a complete "I am so l33t" attitude. Rude, unhelpful and elitist is no way to run your project, people.

That was the first bug I reported to them, and it will be the last. I don't recommend Gentoo to anyone anymore.

Don't piss on people when they are trying to help, Gentoo developers

Why doesen't the Open Source Community...

[razmaspaz]

develop a central bug reporting site. A few dedicated people or moderators could take on the responsibility of passing all their bugs to the developers. But more importantly developers would have a one stop place to get bugs, as well as distributors. You could probably use bugzilla to do this. Even set it up to forward bug reports to the bugzilla site of each project automatically. Wouldn't this make it easier for users of software(especially non technical ones) to have a place to report bugs?

Many liasons simply don't care, however

[SexyKellyOsbourne]

When I submit bugs to large open source projects, the maintainers usually reply and try to paint it all as my fault, as if I caused the program to do it, or give something with the connotations of "STFU n00b, we've heard that one 1000 times today!"

I'm not alone -- just browse the bug forums of http://www.sourceforge.net [sourceforge.net] and see how many of those maintainer liasons -- especially on a few certain major Open Source projects that I will not name -- simply aren't dedicated to keeping the project bug free, as it seems their philisophy is something along the lines that "bad code is better than no code," as if they were double agents for Redmond or something.

However, the authors usually will look into the bugs if you mail them directly as "URGENT," though it may take a few tries.

Re:Forget Gentoo

[lemming552]

Hmmm. I posted a bug and got a small progress report. No attitude or anything, but then I found that the bug was more a user error (me) than anything else.

Re:uh what?

[GigsVT]

couldn't figure out Bugzilla (not exactly the most user-friendly interface)

Thank You!

I thought I was the only person who hated Bugzilla's UI. I can use it, but I think it is incredibly messy and hard to use. When I want to do quick searches, I have to pretend to be entering a new bug, since the normal search form has way too many useless details on it.

Of course they should

[Planesdragon]

Of course, I'm not an open source developer

That, in and of itself, disqualifies your evidence as relevant.

Tech Support for non-OSS* is written for people who cannot / are not allowed to / are never expected to understand what's going on.

Bug reports, (especially in OSS), on the other hand, are intended for an audience that can be (and often is) assumed to know what's going on, and how the system works. Any descent bug reporting system tells the reporter to document everything, to reproduce it, and only gives them the ability to submit a bug after they've gone through a UI at least as complex as that of the help documentation for the program.

Or in other words... OSS folks don't deal with dumb users, they deal with dumb admins**--who are often flamed away so quickly that only the halfway competent admins remain.

_________________________

*: OSS: Open Source Software ("OSS Software" would be redundant.)

**: Everyone who uses OSS is or works with or is an admin, even if it's just someone on their own machine in their own basement.

Re:They Don't/Shouldn't

[Gooba42]

My most frustrating experience with support was submitting a bug report and having the developers tell me that the bug simply didn't exist.

Specifically I was dealing with MoodLogic (not OSS but useful) support. I unchecked the box that says "change all by artist" and it went ahead and changed all by that artist anyway. When I wrote in, support intentionally misunderstood and told me not to check the box as, obviously I must have done because there was no bug.

I wrote back in excrutiating detail how I understood the difference between a checked radio button and an unchecked radio button, explained precisely which songs I was attempting to fix and what the fix should have been. I then explained precisely the order in which I hit the buttons with which mouse button, and what state the checkbox was in at the time.

The reply again assumed I was an idiot and told me to uncheck the box because there was no bug.

Frustrating as hell to know what you're doing and deal with people who don't believe that you do.

go to redhat

[josepha48]

that is what I do. I just file them at bugzilla.

I feel that if they put it on their cdrom then they should hvae tested it some. They will also know or should know the best way to contact the maintainer. They also may be appling their own patches to the code. They do this in the linux kernel and I am sure that they do it elsewhere, so it may actually one of their patches that caused the problem.

I had a problem on my system recently where I upgraded from RH 7.2 to RH 7.3 and my passwd file was locked. I removed the .pwd.lock file the ptmp file and any other file that I could think of. I even boot the system into init 1 and init 2 and tried but it was still locked. Then I installed RH on a second drive and booted the second drive and the second system recognized the /mnt/etc/passwd file as still being locked. I thus had to reinstall RH 7.3 wiping out my system. Thank goodness I had mount points from /opt and /home that I saved data on and did not loose anything. I also save important /etc files as well. So it was about 3 to 4 hours to rebuild the system tops from a new install.

So who is responsible for useradd? For vi / vim? For /etc/passwd? I have no idea, but in redhats database there is now a bug about this as I feel that it is their software at some point.

Observations of a forelorn bug submitter

[budGibson]

I think the big challenge in open source today is enabling the *easy* interaction between developers and users. The interaction right now is just too costly for both parties. My cut would be that there needs to be further development of automated system slike Mozilla's talkback and that this type of bug reporting should become a *fundamental* aspect of Open Source Development. The current problem with talkback is that it only works for crashes. It would be nice if you had some sort of built-in interaction recording functionality that would allow people to click a button to send a brief playback along with a description of what they did not like.

I have given up on submitting bugs through bugzilla (not just complaints, I give what it must be like for developers below):

1. You have to log in. Sometimes the registration process requires a lot of information or hand shaking emails. It's an impediment.

2. You have to search for your bug. How are you going to find it? It's not a google-like search engine. You have to count on people submitting the bug with a description that you will understand.

3. You have to spend a lot of time describing your bug. What if others don't understand it? What if the developer does not understand it?

From a developer's perspective:

1. They are only getting the perspective of the ardent few. Will that help them expand the user base and make the project a success? Possibly not, since the majority of people who have problems might just give up.

2. Will they understand what people have described?

3. Will they be able to reproduce the bug? Do they have the configuration to do so?

Just my two cents,
Bud

Distro != Open Source Package in most cases

[shoppa]

In most cases, the version of an open source package you get from Redhat or Debian (or whoever) does not directly correspond to the official release of an open source project. As an extreme example, Redhat for several years shipped a version of gcc that ID'd itself as "2.96" while all the while the gcc developers were swearing up and down that there was No Such Thing as GCC 2.96 [gnu.org].

The degree of divergence between the two determines whether it is appropriate to send the bug report to either or both. In most (but not all) cases the distro will be lagging behind the OSS package bugfixes so it's very likely that it's already been fixed.

The real solution, of course, is to ditch all distros and build everything from sources yourself [linuxfromscratch.org].

Responsiveness of Mozilla developers

[El Volio]

Personally, I've had loads of success submitting bugs for Mozilla. Since I've been using it for my day-to-day work for so long, I decided a lon gimte ago that I could at least bother to report the problems that I find. And the developers have been incredibly responsive. Sometimes they don't agree with me on how it should actually work, but they respond quickly and are willing to discuss the reasons behind their decision, which is good enough for me.

I've only submitted one bug in a distribution package (to Debian), and I saw a reply as well -- 3 months later. Although I still use Debian, responsiveness is probably not high on the list of reasons I do. Then again, most Debian maintainers are volunteers but a substantial chunk of Mozilla developers are paid, so that probably explains it.

semi-anonymous bugzilla accounts

[Anonymous Coward]

I've always wondered why bugzilla(s) on various projects don't let you register with an anti-spam (munged) email address. Mozilla.org, for instance, sends you a bugzilla password when you register, but it's auto-generated, so I don't think it's possible to enter a munged address when you do so. I really want to participate in a number of projects, but I've gotten so buried in spam in the past that I generally tend to email the contact person listed in the bugzilla database when I find something. It wouldn't be too hard to have the address auto-munged in some way, but no one seems to do that (at least no one that I've seen).

Anyone got a better suggestion?

Re:What utility software?

[morgajel]

because they can't waste their precious time doing a little research

they are a user. you are a programmer. it's your code. They have no obligation to spend any more time on it than they feel like(conversely, you don't have to spend time on them either).

You must remember, they're doing you a favor by reporting a problem with your code. keep that in mind.

don't alienate them by telling them they have to fill out a 4 page questionaire and search a huge database for related bugs. For some people that's just more time than they want to spend.

I understand what your saying, but most bug reporters are doing it out of the goodness of their hearts. If they are treated like shit because they are new to bug reports and do something wrong, you can bet that'll be the last they fill out.

then ask yourself this- can you see your grandmother filling out a bug report? imagine it from her shoes if you want real insight.

Submission Guesswork

[slobbit]

I agree that it is difficult and confusing often to report bugs. It seems like many reports are either way too detailed or over simplistic.

I had a bug that was causing me problems printing to a network printer. When I went to submit a bug on the project I scoured the lists. Finding nothing that matched, I submitted my bug, describing my system, program versions, the fact that the exact same setup had worked under a previous version, and what the symptoms were. When I get the info back on my submission it appears that "my" bug had already been described and fixed. The problem was that the original submitter had a programmer's level of knowledge about the problem, and described it in those terms (blah-blah doesn't change blah-blah-blah in blah.cfg), without mentioning the symptoms the enduser would experience.

I don't know what the solution is; the Buzilla documentation is pretty good about explaining how to submit a good bug report, it's just that many people don't follow the guidelines, then the maintainers just let original description through without editing for clarity.

Oh gosh, this has reminded me of my many horrible bug hunts on Bugzilla. What a great topic for Hallowe'en--I'll be awake all night!