Miro,
sorry for making you aggressive, that was not what I wanted.
Maybe you did not understand what I meant. Sure I know what open source is, but maybe
you did not understand what responsibility each contributor has. If I
contribute a piece of source to the community, I am responsible that this peace of
source does not produce problems. If my contribution is buggy, I am
responsible to fix it. Everything else would lead to the fact that some people can
code what they're happy with, not looking left or right (that's some kind
of unsocial behaviour in my eyes). I explicitely want to say that I do not tell that
this was the case here, but as you attacked me, I have to be defensive.
That is my point of view and I ran good with this with projects in the past. For you
are right, JOnAS is good software and it is a good partnership to
look for bugs for not doing any other contribution, sure I am willing to look for the
bug. But with JOnAS, what I also think to be very good software (if
not, you would not read my name in the mailing archive), there is a problem. There you
have the problem that there is no community that manages the project,
but a closed team. This team developes most of the code, not to say hardly all of the
code. In the past I tried to fix some bugs and contribute some
performance and scalability issues (see the mailing archive) and it was very hard to
convince the team to incoporate the fixes, and it was not possible to
convince to team to discuss performance issues in detail (for it was not scheduled at
that point), since only the core team has write rights on CVS and they
have a closed (some kind of hidden) schedule of JOnAS developement. The problem is,
many users want to contribute, but there is no process of official
contribution and project management that would make this possible. That is one point.
A second point is that Philippe sure knows that I do not just want to consume JOnAS
(for I did some changes months ago), so what's your problem? I talked to
Philippe for he wanted to push back the responsibility to Sun, where I think that it
would be easier to solve to problem by debugging JOnAS on the
problematic platform as I did some months ago on NetWare. All I wanted to say is that
we should look into JOnAS code and not into VM code, since VM specs
tell that a software must not know of VM internals. I explicitely did not tell that I
force Philippe to do so, and I did not tell that you have to do so. I
special, I wanted to suggest that JOnAS team could coach me through debugging in LINUX
as I am no UNIX crack. Read my email word by word and you will not
find anything else but this: "I mean, it would be easier to look into JOnAS code to
find the problem than to assume Sun's VM is buggy, isn't it?". That does
not tell "JOnAS team has to fix it" but "We should inspect JOnAS and not the VM". Is
this a problem for you?
So what's you're problem?
Markus
> Markus,
>
> thats a pretty harsh response. That's why it is called open source that you
> can just download the source code and track the problem yourself. Or just
> contact Evidian and pay them for their support. Jonas is a great product and
> it saves us lots of money and energy just because it is free! But to keep it
> free and growing it has to have support/contributions from the community.
> And one of kind of support is testing and fixing bugs...
>
> Miro Halas
>
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> Sent: Monday, June 25, 2001 4:40 AM
> To: [EMAIL PROTECTED]
> Subject: Re: GenIC Segmentation Fault
>
>
> Hi Philippe,
>
> > It seems that there is a problem with JDK 1.3.1
> > I don't think it is a JOnAS one.
>
> How could this be? I mean, does 'Write once, run anywhere' not work with
> original Sun VMs? I remember last time when we had similar problems on
> Novell, where
> the problem was a misinterpretion of Java Specs by JOnAS team. Maybe here is
> another detail where JOnAS works wrong. I mean, it would be easier to look
> into
> JOnAS code to find the problem than to assume Sun's VM is buggy, isn't it?
>
> To: [EMAIL PROTECTED]
> [EMAIL PROTECTED]
