You are not logged in.

#1 2008-04-10 08:53:26

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

FIX/FILE BUGS don't IGNORE

File bugs, don't ignore them! That's how bugs perpetuate and last for ages. Or maybe a bug is lingering/unconfirmed and we need more info, you could provide that info. (Cue Uncle Sam?)

Example:
Problem: Bug in wireless driver
Suggestion: Use ndiswrapper

This is dumb. Yes, it's fine to use ndiswrapper as a temporary workaround. But if you don't file a bug, there's a real chance it won't get fixed for a long time.

It's irritating to hear people whinging "this software is so buggy, this bug hasnt been fixed for months". That's your fault because you did nothing about it, and you can only blame yourself. Open Source developers do not have the resources to test on all hardware/configurations.

Similarly, if you have a problem, keep trying the software. Don't tell people that X is buggy, when you last used it 9 months ago. Chances are, X improved a lot since then.

Or:
Problem: Oh crap, libjoe.so.1 not found!
Suggestion: Symlink libjoe.so.2 to libjoe.so.1

This one is also dumb. It's not even a valid workaround. It's stupidly simple to fix this one. Check you have the latest, if you do, file a bug. We rebuild and it'll work when your mirror syncs. In the meantime, there's /var/abs and you can rebuild. 95% of the packages in the repos will rebuild in ~5-10 minutes. Make a cup of coffee, go to the loo, whatever. If you can spare time for a forum post, you can spare time to rebuild it.

/end rant.

Last edited by iphitus (2009-06-09 09:47:00)

Offline

#2 2008-04-10 09:07:38

dolby
Member
From: 1992
Registered: 2006-08-08
Posts: 1,581

Re: FIX/FILE BUGS don't IGNORE

oh btw and not meaning to hijack this thread but the above is the most important reason not to use external patches in packages. all bugs/feature requests should be reported upstream.
also on a sidenote, most people just report bugs on the forum instead of the bug tracker. maybe a note about this somewhere would be nice. even though the bugs link is quite visible on top


There shouldn't be any reason to learn more editor types than emacs or vi -- mg (1)
[You learn that sarcasm does not often work well in international forums.  That is why we avoid it. -- ewaller (arch linux forum moderator)

Offline

#3 2008-04-10 09:16:55

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

Re: FIX/FILE BUGS don't IGNORE

dolby wrote:

oh btw and not meaning to hijack this thread but the above is the most important reason not to use external patches in packages. all bugs/feature requests should be reported upstream.
also on a sidenote, most people just report bugs on the forum instead of the bug tracker. maybe a note about this somewhere would be nice. even though the bugs link is quite visible on top

That's inevitable, and often people don't realise it's a bug. It's the responses people give that are wrong, take the two examples above. Where people should have been told to file a bug, but instead they were given crap workarounds.

Last edited by iphitus (2008-04-10 09:17:54)

Offline

#4 2008-04-10 14:41:33

schivmeister
Developer/TU
From: Singapore
Registered: 2007-05-17
Posts: 971
Website

Re: FIX/FILE BUGS don't IGNORE

this happens a lot, everywhere. if our userbase can't be educated regarding this matter, there's no other solution.


I need real, proper pen and paper for this.

Offline

#5 2008-04-10 15:37:44

Mikko777
Member
From: Suomi, Finland
Registered: 2006-10-30
Posts: 837

Re: FIX/FILE BUGS don't IGNORE

Ummm, I r stupid enough to assume all things broken are upstream bugs that arch devs wont fix.

Maybe someone could post some educative links of arch and linux internals that would help me file proper bugs.
(all the threads on this subject seem to die fast, when someone says read initscripts)

Guess its basically if program wont start file a bug, if it starts but crashes in use file a bug to programs bugtracker right?

Edit: Ah nevermind guess i can just read the LFS docs to see how stuff works.

Last edited by Mikko777 (2008-04-10 15:54:20)

Offline

#6 2008-04-10 18:48:26

grndrush
Member
From: Hamilton, Ontario, Canada
Registered: 2003-12-28
Posts: 136
Website

Re: FIX/FILE BUGS don't IGNORE

Thanks, iphitus. I'll get off my duff...

Having written (with all due modesty, extremely high-quality) software for a couple of decades, myself, I don't usually assume it's an Arch or Linux bug - I assume that, somehow, its ME. I tend to 'tinker' a lot, often simply for the knowledge - which by definition means I'm purposely playing with stuff I *don't* fully understand.  wink  I generally assume that I changed something whose purpose I apparently didn't understand; I NEVER blame either of the above parties, that's for sure (X IS better than it was a year ago!), although I DID cuss Bill Gates daily for a couple of decades or so.

I had a motherboard burn up on me about 6 or 8 months ago. I built a 'new' system around an older case and mobo, usually virtually every single other component from my old system (including even the P/S), which is why I blame the other mobo. My current mobo has a bit of an 'odd' on-board sound chip (lspci and the manual that came *in the same box, new* don't even agree WHAT chip it is; yes, I trust lspci more than I trust ASUS, for sure). I even have a couple of very nice, operational and very mainstream PCI sound cards which I've used successfully with Linux in the past, but the mobo doesn't seem to "hear" me when I disable when I disable the chip in BIOS (pun fully intended). I've done lots of reading (I can't avoid Wiki; I'm a Wiki *editor*), and tried a bunch of things, none successful.

In my case, I guess, I get mad at ME rather than the O/S (or distro), and haven't posted for fear of ending up looking stupid. Sound isn't really all that important to me, but with the fevered pitch of the election season here, it WOULD be bice to actually HEAR all the YouTube videos people send me.

But you're 100% correct; there may be someone else out there in the same boat for the same reason. I'd go post the cry for assistance this moment, but my life's been in a good deal of turmoil of late. I also (FINALLY) traded in KDE for Xfce4 a week ago (which of course uses ESD; I *never* have). I need a REALLY good nght's rest and an hour or so to bone up on ESD 1st. Hopefully this weekend, or early next week.

Thanks for the mini-rant; I sincerely need that!

Last edited by grndrush (2008-04-10 18:53:09)

Offline

#7 2008-04-10 19:18:40

freakcode
Member
From: São Paulo - Brazil
Registered: 2007-11-03
Posts: 410
Website

Re: FIX/FILE BUGS don't IGNORE

iphitus wrote:

I see this a lot. It pisses me off.

File bugs, don't ignore them! That's how bugs perpetuate and last for ages. Or maybe a bug is lingering/unconfirmed and we need more info, you could provide that info. (Cue Uncle Sam?)

...

It's irritating to hear people whinging "this software is so buggy, this bug hasnt been fixed for months". That's your fault because you did nothing about it, and you can only blame yourself. Open Source developers do not have the resources to test on all hardware/configurations.

I guess its because big part of the current userbase comes from proprietary software culture, where people don't have the habit of interacting with the developer, don't fill bug reports, and so on. I would say they even don't "feel" they can improve the software. All they do is download a shareware, if it sucks, they download another and another, or if they buy some software that crashes on their faces, they curse the manufacturer, but keep buying the same crapware because it's "widespread". Software is really a very different type of "product" - the only I know that is released not completly finished neither fail-proof.

Offline

#8 2008-04-10 19:41:33

grndrush
Member
From: Hamilton, Ontario, Canada
Registered: 2003-12-28
Posts: 136
Website

Re: FIX/FILE BUGS don't IGNORE

*Excellent* point, freakcode!

I hadn't even considered that angle, and I spent my entire IT career in the mainframe world (pre-shareware, of course). We simply figured out "workarounds" (great for keeping the mind sharp...). That was also, of course, the time when 'feature' became an alias for 'bug'.  wink

Offline

#9 2008-04-10 20:32:23

dunc
Member
From: Glasgow, UK
Registered: 2007-06-18
Posts: 556

Re: FIX/FILE BUGS don't IGNORE

I'm not a software developer, but I am a tinkerer, and I completely understand grndrush's reasoning. I always assume it's something I've done as well... and most of the time it is.

Freakcode makes a good point, too. Coming from the post-Commodore Amiga world, which had its fair share of open-source and amateur projects, I'm not completely unused to interacting with developers, but it was all pretty small scale stuff. A big project there would be lucky to have 50 members on its open devs/users mailing list. It felt like everyone knew everyone else, in a sense. Linux can be daunting in that respect. It can be scary to hold your hand up and say, "Um... this doesn't work".

But iphitus is perfectly right, though. Perfect bug reporting is something to which we should all aspire, even if we don't all quite reach it. smile


0 Ok, 0:1

Offline

#10 2008-04-10 21:01:59

Ink-Jet
Member
From: London
Registered: 2008-03-27
Posts: 64
Website

Re: FIX/FILE BUGS don't IGNORE

Should you report what you think is a a bug to the program's respective tracker if you're not sure it's a bug?

To be honest, most of this developer feedback stuff is a bit beyond me.

Offline

#11 2008-04-10 21:13:38

synthead
Member
From: Seattle
Registered: 2006-05-09
Posts: 1,326

Re: FIX/FILE BUGS don't IGNORE

Great points guys.  I'm going to see if I can remember some bugs I encountered and try to reproduce them.  If I can, I'll make bug reports.

Kind of an interesting thought though isn't it?  "I'll see if I can find any bugs."  Compare that one to other OSes!


Touch my kernel

Offline

#12 2008-04-11 01:29:36

toofishes
Developer
From: Chicago, IL
Registered: 2006-06-06
Posts: 602
Website

Re: FIX/FILE BUGS don't IGNORE

This one is almost worth stickying.

Offline

#13 2008-04-11 04:38:09

ConnorBehan
Trusted User (TU)
From: Long Island NY
Registered: 2007-07-05
Posts: 1,356
Website

Re: FIX/FILE BUGS don't IGNORE

iphitus wrote:

Example:
Problem: Bug in wireless driver
Suggestion: Use ndiswrapper

...

Or:
Problem: Oh crap, libjoe.so.1 not found!
Suggestion: Symlink libjoe.so.2 to libjoe.so.1

Right, if stupid workarounds like ndiswrapper only get posted on forums such as these, there's a good chance the wireless developers won't find out about the bug any time soon. What I would do is report the bug upstream, and then at the end of that post, say something like "until this bug gets fixed ndiswrapper will work."

A problem that I've had to deal with for a year now has been the buggy 3D support for ATI Rage 128 cards. This is a known bug that I saw on the x.org DRI site, I've discussed it in #xorg, these forums and the Linux thinkpad mailing list so I don't know what more I can do. The workaround I use now is, if my computer freezes upon killing X, I shut it down remotely with ssh.


6EA3 F3F3 B908 2632 A9CB E931 D53A 0445 B47A 0DAB
Great things come in tar.xz packages.

Offline

#14 2008-04-11 06:46:28

catwell
Member
From: Bretagne, France
Registered: 2008-02-20
Posts: 207
Website

Re: FIX/FILE BUGS don't IGNORE

This is so true. Free software works because of the "Many Eyes" thing, and to me this doesn't only mean that the code is read by several people but that the users also take part in the dev process. Which means filing bug reports on the right bugtracker (most of the time it's upstream).

Yet sometimes I'm not sure how Arch devs manage these bug reports. For example, take this one. I had filed it, and the package dev answered it and said that another dev should take care of it because he couldn't, but nobody did that, even if I said why it didn't work afterwards and it was just a matter of recompiling the package.  That's probably devs don't read bug reports that have a low severity and are already assigned to someone else, which seems logical, but not in such cases. I don't know how they could do yet.

Anyway, on the whole finding bug reports is important, even for the AUR, by leaving comments about the packages. Although the package maintainers are not devs, they care too.

Offline

#15 2008-04-11 09:49:43

zodmaner
Member
Registered: 2007-07-11
Posts: 653

Re: FIX/FILE BUGS don't IGNORE

Another case that clearly shows a confusion most people have between a fix and a work around is this bug: http://bugs.archlinux.org/task/9537#comment26768 (Thunar won't eject CD).

Basically, one user propose a 'fix' (which in reality is a work around, and an ugly one at that) by install pmount/ivman and use them to mouth/eject CD instate of Thunar even though there is a proper fix available (a patch for exo pacakge).

Imagine what would happen if most user start installing pmount/ivman instate of filing a bug report or applying a patch.

Last edited by zodmaner (2008-04-11 09:59:05)


Memento mori

Offline

#16 2008-04-11 11:13:42

ibendiben
Member
Registered: 2007-10-10
Posts: 519
Website

Re: FIX/FILE BUGS don't IGNORE

How about adding a feature (hehe, I know one should keep it simple, but read on) to every forum topic enabling you to "send to bug tracker" which then ask you to follow the steps normally asked when adding a "new task" on flyspray...

So topics like, wireless not getting ip at boot, which turn out to be a arch-software problem could be send to the bugtracker easier:
>[Send to bug tracker];
>Give a descriptive summery;
>Then choose:
----Task type
----Category    
----Status
----Operating System    
----Severity
----Priority
----Reported Version
----Due in Version
>Then add, additional info on how to reproduce, which could often just be a link to the topic, but more kindly a destillation of usefull information.

I could suppose this makes some difference... just because it's in the users eye, and some clicks (+subscription/login) closer.

Offline

#17 2008-04-14 10:43:25

STiAT
Member
From: Vienna, Austria
Registered: 2004-12-23
Posts: 603

Re: FIX/FILE BUGS don't IGNORE

Also, if you need updated version of packages, and have fixes or similar to them, it's often easier to either file a bug or contact a dev.

I've done so several times, and the things got fixed within a few weeks, using workarounds till then (most likely compiling my own patched pkg).


Ability is nothing without opportunity.

Offline

#18 2008-04-14 13:24:17

gazj
Member
From: /home/gazj -> /uk/cambs
Registered: 2007-02-09
Posts: 679
Website

Re: FIX/FILE BUGS don't IGNORE

The first time I ever filed a bug (was on awesome's bug tracker not arch) I felt like I should not be moaning about something that essentially I was using for free that someone else had put a great deal of effort into.  To put it bluntly I am too polite and really worry about being rude.  Luckily the dev felt the bug was well worth fixing and in the next release was fixed.

So.. In the end I felt like I had made a contribution (even if it was very small) to making a oss project better.  It was well worth it.

My suggestion is be polite about it, explain as best you can and give as much information as possible.  Don't be scared.  I still have to tell myself this sometimes.  The devs do want our feedback.  I am sure they get a great kick out of seeing there software work as intended.  I know I get one when I finish a project at work. smile

Offline

#19 2008-04-14 17:19:19

faelar
Member
From: Amiens (FR)
Registered: 2007-12-18
Posts: 232
Website

Re: FIX/FILE BUGS don't IGNORE

When I'm in front of "something" which doesn't work, my instinct tells me "Google it !". And most of the time, the answer is... on a linux forum ? I can't remember any situation where I was redirected to a bug tracker (not sure it's the right name).
Maybe I'm wrong, but I think it lets the user think that telling there is a problem (bug or not) on a forum is the best way to be helped by someone who know how things work (a dev sometimes).

Another point, you said :

Problem: Oh crap, libjoe.so.1 not found!
Suggestion: Symlink libjoe.so.2 to libjoe.so.1

In this situation, the user don't know how to fix things, and the answer provided is "crap" of course, but we will suppose it works...
What the user will do next time ? He browses a different forum and find exactly the same question he asked for ! He is so happy ! He can help someone else to solve a problem ! What do you suppose he will say ?
And if the soft is not broken anymore for him, do you really think he will fill a bug-report ?

I don't have the solution, but maybe we need to understand more why user don't use bug tracker...

Offline

#20 2008-04-14 20:37:25

ornitorrincos
Forum Fellow
From: Bilbao, spain
Registered: 2006-11-20
Posts: 198

Re: FIX/FILE BUGS don't IGNORE

I think that the basic problem(at least, I have it sometimes) is recognizing what are bugs and what are just silly configuration questions.


-$: file /dev/zero
/dev/zero: symbolic link to '/dev/brain'

Offline

#21 2008-04-14 21:56:08

kakTuZ
Member
From: Hannover, Germany
Registered: 2007-10-20
Posts: 86

Re: FIX/FILE BUGS don't IGNORE

faelar wrote:

When I'm in front of "something" which doesn't work, my instinct tells me "Google it !". And most of the time, the answer is... on a linux forum ? I can't remember any situation where I was redirected to a bug tracker (not sure it's the right name).
Maybe I'm wrong, but I think it lets the user think that telling there is a problem (bug or not) on a forum is the best way to be helped by someone who know how things work (a dev sometimes).

Look at it from the other site: Everything filled in a bug tracker got solved, so you never had to bother with.

Offline

#22 2008-04-14 23:02:02

vuboy
Member
From: Canada
Registered: 2008-03-04
Posts: 78

Re: FIX/FILE BUGS don't IGNORE

ornitorrincos wrote:

I think that the basic problem(at least, I have it sometimes) is recognizing what are bugs and what are just silly configuration questions.

Thats my issue as well.

VB

Offline

#23 2008-04-15 14:27:32

ConnorBehan
Trusted User (TU)
From: Long Island NY
Registered: 2007-07-05
Posts: 1,356
Website

Re: FIX/FILE BUGS don't IGNORE

Also, sometimes it seems pointless to point out one bug when its already obvious a program is wrought with bugs left and right. For example, I just tried converting one of my Corel Draw drawings that I made in grade 8 into an SVG with Uniconvertor. It didn't work... but is that a specific bug in Uniconvertor that could be easily fixed if reported? I don't think so. Everyone knows reverse engineering a proprietary vector graphics format is a daunting task and I'd be incredibly surprised if converting an arbitrary CDR to an SVG with Uniconvertor DID work.


6EA3 F3F3 B908 2632 A9CB E931 D53A 0445 B47A 0DAB
Great things come in tar.xz packages.

Offline

#24 2008-04-15 20:55:14

ornitorrincos
Forum Fellow
From: Bilbao, spain
Registered: 2006-11-20
Posts: 198

Re: FIX/FILE BUGS don't IGNORE

in that case they could have been interested in why it didn't work, if you try to reverse engineer something, the more information the better.


-$: file /dev/zero
/dev/zero: symbolic link to '/dev/brain'

Offline

#25 2008-04-21 15:33:19

louipc
Member
Registered: 2006-10-09
Posts: 85

Re: FIX/FILE BUGS don't IGNORE

iphitus wrote:

I see this a lot. It pisses me off.

File bugs, don't ignore them! That's how bugs perpetuate and last for ages. Or maybe a bug is lingering/unconfirmed and we need more info, you could provide that info. (Cue Uncle Sam?)

I sense a conflict here. You tell us to file bug reports but on the same token it seems that a lot of bug reports are IGNORED by devs. If you want bug reports I've got six hundred and thirty for ya. Why are so many still lingering? There are even a few that were reported in 0.7 Wombat! Come on!

Close them. Follow up. Update the relevant packages. Etc.
If you need help, ask for it. A lot of us are itching to help.

Offline

Board footer

Powered by FluxBB