[kaffe] Bug report re: bug reporting

[Dalibor Topic]

Hi Tim,

--- Timothy Stack <stack at cs.utah.edu> wrote:
> 
> hi,
> 
> I'd like to file a bug report stating that the
> current system for 
> reporting bugs sucks.  :)

Thank you for your bug report ;)

> It seems to me that there are just too many people
> with problems that are
> going unanswered.  Sure, we'll exchange some initial
> mail, but in the end
> people are just too busy with other things and
> Dalibor can't do it all by
> himself.  Then the problem just gets lost in the
> archive and its too much
> of a pain to go back through trying to figure out
> what has and hasn't been
> fixed yet.

I mostly agree. IANLT (I am not Linus Torvalds), so I
don't really scale that well (and I'm actually busy
with other things, too). So any help is highly
appreciated. If a bug database is a good way to get
more people to contribute more regularly, or to help
existing developers, that's fine with me.

But, we've already had a bug database before and that
one died a slow death. I can't say what the reasons
were, because it died before I got involved with this
project. When I looked into it last year, it had
become a trashbin for users to file bug reports that
noone looked into. See my rant here
http://www.kaffe.org/pipermail/kaffe/2002-June/008364.html
It was abandoned, although you can still get all the
bug reports as a website archive here :
http://www.kaffe.org/~jim/brainfood-website-plus-bugs.tar.gz

So, has the situation changed since last June? In my
impression, kaffe still has a rather small number of
regular contributors. Its user base has grown, though,
as can be seen by the massive increase in mail traffic
(see http://www.kaffe.org/pipermail/kaffe/ for stats)
since last February, when the project was effectively
asleep. We are now at 10 times the volume since about
a year ago, while the number of "regulars" has stayed
about the same.

While my rant was in place then, Jukka Santala has got
it right in the end : in
http://www.kaffe.org/pipermail/kaffe/2002-June/008383.html
he says "But once 1.0.7 is out, _hopefully_ the
developer (Or at least user/tester) base is going to
grow significantly. Keeping up with the influx may be
hard then; but perhaps we can just wait and see for
now." . 

What Jim wanted to do, was to keep FAQ.Knows-Bugs up
to date
(http://www.kaffe.org/pipermail/kaffe/2002-June/008401.html).
 It didn't really work out, for a lot of good reasons.
A single person can manage the bugs in their mail
queue, but it doesn't scale well when the you start
getting more bug reports than you can fix. Especially
when you are busy managing releases, server moves,
developing a new build system, and keeping everything
running smoothely in general. And he does a great job,
in my opinion.

So my question is: would having a bug report system
lead to more bugs getting fixed, or would it just be a
public mail dump?

My theory why the original bug database died, is that
the original core set of maintainers refocused on
other things, which is a normal thing in life. I
wouldn't want to spend my life just developing kaffe,
either ;) As I said, IANLT.

Unfortunately, they seem to have mostly missed the
opportunity to "raise" new maintainers. That explains,
in my opinion, the growing pains we experience now.
And I didn't think about it either.

My proposal to improve the developement process, would
be to go with your proposal, and set up a bug database
system again. But also to share the load on more
shoulders and recruit more people willing to take over
some responsibility that comes with the priviledge of
cvs write access.

For example, I'm not playing the check-in guy because
I love to check in stuff, but because many patches
simply don't get reviewed and checked in otherwise.
And most people seem to be reluctant to test what's
out there in form of a patch, unless it has been
checked into CVS and broke something they care about.
;) That's O.K. for me, as long as we get the bug
reports while the wounds from living on the bleeding
edge are fresh. It's rather painful to track things
down months after they have been checked in.

For another example, if anyone looks at my merging
efforts with JanosVM so far, they'll see that I've
mostly left out the improvements the JanosVM team has
made to the core VM. I don't feel comfortable breaking
stuff I don't understand much about, and prefer to
play safe, fixing bugs where I know what I'm doing. In
my case, that's core libraries and compilers. 

In my experience, there are enough things left to fix
there, so that I don't get to learn enough about the
core VM to be able to make judgements there. Now, if
everybody would just sit and wait for me to learn
enough about the core VM to be able to merge in the
missing bits, that would simply not happen in the next
few years. ;)

> So, basically, does anyone know of a bug reporting
> tool that can be setup 
> on the web site without much trouble?

Before we just pick a tool that can be set up without
much trouble, I'd like to hear what people expect from
a bug reporting tool.

Here's my take:

I'd like to have a bug reporting tool for the single
purpose of being able to browse bugs more easily, and
marking them as fixed. I want to have all the traffic
associated with a bug to go over the mailing list,
because it gives people a better chance to find
someone who knows, than to rely on those who know to
find the time to browse the incoming bug reports.
Additionaly, it lets people chip in when things are
going into the wrong direction. 

For an example of that, see my previous attempts at
replacing some of the native libraries with pure java
veriants. The idea appeared reasonable to me, sitting
on a fast laptop where the jit3 code beats the native
library at some operations. It turned out to have many
drawbacks that I didn't think about. 

Now, I think we've reached a solution where the pure
java libraries are provided as an extra option at
configure time, that gets selected automatically if
the native libraries are not available. The discussion
leading to that all happening would not have
necessarily taken place if it had been constrained to
the context of an obscure bug report.

If someone is familiar with bug tracking tools, and
could get one such beast set up to that, I'd be happy.
I'd be happy to hear other opinions, too.

cheers,
dalibor topic

__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com