[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [DotGNU]Pnet Bug tracking and categories

From: James Michael DuPont
Subject: Re: [DotGNU]Pnet Bug tracking and categories
Date: Mon, 3 Mar 2003 14:14:13 -0800 (PST)


--- Rhys Weatherley <address@hidden> wrote:
> On Tuesday 04 March 2003 05:40 am, James Michael DuPont wrote:
> > Currently, I am working on trying to fix some bugs,
> > but they are not assigned to me. I hear that they are assigned to
> > someone else, but I cannot see that.
> As far as I'm aware, there are no open bugs currently assigned to
> you.  
Minddog asked me to help him on the system.xml issue. While doing that,
I reviewed the other bugs in that area.

> > Also, some of my bugs are being closed, before they are resolved.
> > I dont mind if you say, Severity is minimal, resolution is deferr,
> but
> > to just close them, that is not fair.
> I am the designated project leader.  As such, I have to close bugs
> that have 
> been dealt with because they have are (a) fixed, 

>(b) judged not to be
> important at this time, 
Ok, well, I would like to take up issue with this. 
Please can you not close the bugs I raise on those grounds.
If you dont want to deal with them, put them on low priority.

> and (c) are inadequately specified and hence
> cannot 
> be dealt with in the current form.

That is another status, just put them as badly specified, and I will
fix them. There is no sense in re-submitting more and more copies of
the same bug.

> You may disagree with my decisions sometimes, but attacking me in the
> bug 
> description, and then via private e-mail, and then again in this
> public forum 
> is not likely to change my decision.

I have alot of respect for you rhys.
This is not an attack, but professional criticism. I am sure you are
able to see the difference. 

> Perhaps you are entering bugs that you yourself want to work on.  In
> that 
> case, it is pointless to enter the bug.  Work on the problem, and
> submit a 
> patch through the Patch Manager when you are done.  I don't need to
> know 
> about problems that I'm not expected to fix.

That is not helping anyone. 

I object to the current DotGnu attitude of no-public commitment and no
public planning. Everything is done in secret. I dont want to work that
way, and I dont think it is good for the project.

Therefore, I have made a list of all the projects and tasks that I have
started and am working on:

We need to get the bugs out into the open.
People who want to help have to be able to see what bugs are not
assigned what work needs to be done.

Communication is what is lacking here, We need more communication.

> > Savannah has the abilty to store lots more information about a bug
> that
> > are not being used properly right now : Category ,Severity
> ,Resolution,
> I can turn these on, but there is a problem.  If the person who
> enters the bug 
> doesn't use them, then the fields end up with meaningless values.  In
> the 
> previous Tech Support section, I did have various categories, but
> they went 
> unused.

You can use them to assign a bug to a low value when it is not

> Besides, I ignore all bug fields except the description.  "Severity"
> reflects 
> how angry the person entering the bug was, not its true importance
> when 
> considered calmly next to the rest of the bugs.  Let's say I
> down-graded a 
> severity value on one of your bugs.  I'm sure you'd flame me for
> doing so, 
> even if it was a calmly made decision.

Deleting is not the same as downgrading. 
You can give things the priority that you want, but please 
do not just delete them. If the bug report is faulty, let me fix it.
If the information provided is not good, ask for updates.

> "Category" reflects where the bug occurred (compiler, runtime, etc),
> which is 
> normally obvious from the bug description, or is wrong because the
> user 
> thought it was in one category (e.g. ilasm) but it was really
> somewhere else 
> (e.g. ildasm).

It can be changed. If you want to browse the bugs, this information is
useful. The bug tracker is not just for you, it is for all of us.
> > "Assigned to" is only being used a little, it needs to be used more
> > often. If I want to help out, I should be able to assign a bug to
> > myself that is not assigned. Then others can see that am working on
> ti.
> You can always add a comment along the lines of "I am working on
> this".  I 
> will note however that every time that you have done this, weeks have
> gone by 
> and the bug has still not been fixed by you.

Yes. At least you know who to blame, you can see who is working on it,
or should be working on it. Right now you cannot see that at all.
If someone wanted to help out, they could not see the issue.
You can always take the assignment back.

>  Or you've dumped pages
> and 
> pages of stack traces and debug code into the bug description which
> don't 
> actually help anyone find the problem in question - they are just
> your own 
> personal thoughts, which are inappropriate for bugs and patches.

Well, the stackdumps are more helpfull than nothing. Debug traces as
well. Often then help out.

> And then you complain that I've closed the bug because it has got to
> the point 
> of being useless at describing the real problem!  Arrgghh!!

Rhys, I have reported many quality bug reports. You have been able to
find tens of problems due to my testing. Please consider that I am
suggesting thing for a good reason. It is not a personal attack.

We may disagree on the priority of some features of the pnet.
But where I think it is important, I am going to fight it.

The steering committee can also decide on what they think is important.

You may be overruled.

Interoperation with Rotor and Microsoft is important for use to get
inside the big business. If we cannot seamlessly work with Microsoft
Dlls, then we cannot be recommended very well.

Right now this is almost done, even if you may rightfuly complain about
some of by bug reports, I have been helping this process.

I am very proud that I have helped Pnet get the RoundTripping, 
and this is a real advantage over other tools as well.

The problems with the handling of the .data sections, or even the order
of elements in the .IL file seem to be a problem with the ECMA
compliance. You cannot argue that ECMA compliance is not important,
even if this one little bit is not very important on scale of things.

I think that ECMA support is more important than Generics, Java
Support, PnetC and many of the other projects that get your support.
Just because the stuff that I am looking into is boring for you, does
not mean it can be wished away.


James Michael DuPont

Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, more

reply via email to

[Prev in Thread] Current Thread [Next in Thread]