[Scummvm-devel] Merging the bug trackers

[Max Horn]

Am 25.05.2005 um 00:08 schrieb Eugene Sandulenko:

> Max Horn writes:
>>> page shows that average age for our bugs is over 300 days. Not
>>> nice. We even have 2 bugs since 2002 and 8 since 2003.
>>>
>>>
>> Sorry, but that's a typical case of a misinterpreting and misstating
>> statistics :-). A closer look would reveal this:
> Believe me, I know what average is and how it is calculated.

I guessed as much, actually :-). Anyway, we are getting quite off topic 
here, but I just can't let this go unanswered, sorry...

>  That was
> my point. We have bugs which are more that three years old. Not very
> good.
Ah yes, but then, take a look at which bugs these are. E.g. our oldest 
one, the grandaddy (and the only one to be three years old, actually): 
It is not about an actual bug, but about an extremely minor difference 
compared to the original engine.

The second oldest is about some colors differing a bit in Sam&Max. 
Several people actually prefer our "buggy" behavior. A show stopper? 
:-)

The third oldest is about Zak appearing a few pixels away from the 
position he'd be at using the original engine (when entering the 
airport). Terrible :-). Oh yeah, fixing it would require making deep 
changes to the walk code which likely would break all other games 
initially :-)

The fourth oldest *is* a serious bug, which makes playing ZAK annoying. 
Alas, it also involves tweaking the walk system, which makes it very 
difficult to tackle.

Then come a bug which is already fixed in CVS, via a hack, and which 
was left open because we'd prefer a "nice" solution.


I could go on. I guess what I am trying to convey here is that our 
"old" bugs fall into two categories:
1) Bugs which are not really bugs or so unimportant/obscure yet 
complicated to fix that nobody bothered to do it
2) Bugs which are very very hard, have a workaround, and for which a 
fix is likely to cause regressions in other places

Of course, as you advance to 2004 and beyond, you also get some bugs of 
other categories, but not that many...


>
>> 2) The average bug age is an "average" -- so the age of all bugs is
>> added up and divided by the number of bugs. In particular, we have
>> eight open bugs from 2002/2003; that skews the picture a lot.
> Also my point was and is that now we have number of not killed bugs
> added each month quite steady. I.e your goal of 50 bugreports would be
> extremely hard to reach unless we will move a bit and squash some.
> I am not talking about easy bugs which I fixed myself in the past, but
> those which require additional researchand maybe diving into scripts
> and/or disassembly or at least involve lots of debugging.

Uhm. Yes of course. Which is *precisely* what I am usually doing before 
major releases. Spending several days to dive into scripts, our code, 
disasm of the original engine etc. and loads of play testing to track 
down and squash bugs :-). Since obviously, the easy ones are usually 
fixed immediately after they get submitted; the others which take lots 
of time are only being tackled when somebody has the time for it, and 
the interest... I am always glad when other people take the time to do 
such fixing, too (e.g. cyx has been doing some really nice work on 
SCUMM fixes recently, for which I am very grateful. And also erik has 
been doing such grunt work a lot).


> Also I am not talking about bugs which are hard to reproduce or exist 
> on some
> rare game version like Zak256 either appear on kind of esoteric
> platform.

So, you are talking about bugs which are obscure instead? <g>.

But seriously, of course there are some pretty nasty older bugs, and 
it's a pity that they aren't fixed. That's true, and bad -- but the 
fact that the average bug age is 300+ is *NOT* a valid measure for 
anything, except the fact that we are a bit reluctant to simply close 
certain bugs because we are waiting for the "true" solution, instead of 
just applying a quick&dirty hack to fix it....




> Eugene
>
> PS. I know that it is extremely uninteresting to kill bugs.
>
To the contrary I think it can be lots of fun. But unfortunately it 
takes a lot of resources, but of course all of us have other things for 
which they need their time, too :-/. So at least in my case I can say 
that I can't spend as much time on fixing bugs as I'd like to.


Cheers,

Max