Request for Bug Tracking System / Discussion of criticial bugs

Hi all,

Since I took part in the Linux alpha testing period, there has been a stack of bugs posted to this forum, and the forum has been used more than the mailing list to the point where the list has been shut down.

the problem is that the forums are not particularly suitable for tracking bugs and updates to bugs - we really need a bug tracking system.

As an example, there is one critical showstopper for me - this halfitar problem in which my avatar and none of the actual world seems to render. I have reported the bug as have a couple of other people and I have no idea if this is being worked on or resolved.

If we were using a bug tracking system I could subscribe to a bug and so could other people who have the same problem. It also allows people to indicate other details and grade the importance of the bug. My bug is really critical - I am using a modern laptop (Sony Vaio), modern distribution (Ubuntu) and a modern ATI card complete with official ATI driver. This is not an edge case by any means.

So, Don, can I suggest the policy is to move everything over to a bug tracking system somewhere?

That would be very useful, and not just for us but even moreso for LL.

Whether this be a formal bug tracking system or not, what we need is at least a bug acknowledgement system. We're supposed to be the alpha testers here, not just passive users, but our bug reports are universally greeted with stony silence from LL. Doesn't really keep the enthusiasm flowing.

That, Toonie, is an absolutely excellent point.

I think the bug reports are "greeted with stony silence" because nobody is qualified to work on them except Icculus, who knows very well what the problems are. Remember that this is an alpha and most of the features are still disabled! There's no need for a bug tracking system for an alpha that will collect a few dozen bugs and then join the main release process and mostly share bugs with the other versions, which already have a successful forum for dealing with bugs.

Sorry to be a wet blanket but if you want to talk about a bug tracking system, at least wait until the known and easily fixed bugs have been rounded up. This forum has only been open for a couple weeks, not even long enough for the actual bug fixing to take place, so there's no reason to say LL (i.e. Icculus) is ignoring bugs.

Hm. I was told I'd be alpha testing, so I've been alpha testing. I was invited to post bug reports. It's a bit of a stretch to imply that our bug reports are ignored because they're totally redundant, else there would be little point to an alpha test (an early-adopter release, sure, but not an alpha test).

I mean that he knows the big bugs right now, and he'll also know your bug reports when he reads the forum.

Sky, I think you're missing Toonie's point.

With a bug reporting system in place, the only action needed at LL's end is to click on an alpha tester's new report and give it a status --- eg. Noted [person], Added-To [some other problem report], Rejected [person,reason], Assigned[work-entry], etc.

It's one-click feedback, it saves time all 'round, and it keeps the alpha testers happily testing. Which they're not doing happily when they feel ignored. They're not getting paid after all. They could at least be honoured with a click.

It certainly doesn't require any extra manpower resource to be allocated to it, unless one postulates that Icculus (or whoever) is being absolutely slave-driven and can't click on a few entries at the same time as he processes his daily mail. Which is unlikely in such a progressive place.

Addendum:

The point to note is that, given such simple feedback, problem reports dry up almost entirely after a while, since only new problems get placed into the bug reporting system. In contrast, when there is no feedback, the bug reports keep coming in non-stop because previous postings do not appear to have been noted. Which approach makes more sense?

The point in hand here is that we have been encouraged to file bug reports, and to do this efficiently you need a bug tracking system.

I get the impression that the Linux alpha testers are getting a little frustrated with LL not at least providing feedback for this testing period. Many Linux users are used to filing bug reports and used to at least getting the bug assigned to someone and a status set.

Bug reporting without a bug reporting system is like like surfing without a surfboard.

I am fairly happy right now that we are getting the issues visible.

I also have the feeling that we have a "bonus" here, not a mainstream version.

My suspicion is, that depending on the level of involvement of a) people using the client (which should be visible server side) and b) the activitiy in this forum, we will get more or less priority. Linden labs is not going to work all that hard if the linux client turns out to be a "who cares" project from the user side.

I am therefore all in favour of a bug tracking system.

I nevertheless do not expect one; and I think the wide usage of the client WILL bring issues to light that have not been found by Linden Lab's internal testing.

Some of these bugs, even if linux specific, may in fact be "unclean code" in the main tree, and point to bugs that the main client has, but that show less visibly under OS X or Windows.

I am going to continue to use the Linux Client for everything. And if it bombs or goes strange, report it here.

I have stopped adding bugs to the Sticky Thread, since they go invisible, altough Don Linden said to post bugs "in this thread". It is clearer if they each get a thread of their own, since people can then add their experiences directly from the forum overview, without having to read the WHOLE sticky. They can also add workarounds, which can then easily be added to the "Tips" thread.

This is not a Bugzilla or whatever; but I think this structure would make the bugs more visible, and harder to duplicate. My thanks to the first person who did it, I think it's ingenious.

I want this client to run And I will use any effective means to assist that