Re: kaffe orphaned? (extended reply)
I have been receiving quite a bit of irritation from the Debian-Java
lists lately so I thought I would try and address the issues.
Here is the situation:
- The production version of Kaffe is 1.0b4. I packaged it the day that
it was released. I have made a subsequent packaging of 1.0b4 to
eliminate the incorrect dependency on libgmp. Despite this, Mike
Goldman has decided to make out that I have "orphaned" Kaffe.
- I have a number of machines at my disposal which the kaffe1.0b4-2
Deb was installed on and on which it operated correctly. My
co-workers, who are also Debian developers, have installed this
version on their workstations and in some cases are using it as
their primary development platform.
- Mike also chose to make an NMU of the Kaffe package. This particular
action has some extenuating circumstances:
- Mike's NMU doesn't claim to fix any specific bug, it merely states
"1.0b4 was broken".
- Mike never attempted to report or resolve his problems with me
before making his NMU.
- In an attempt to make a more "stable" kaffe package, Mike pulled
a copy of the developmental CVS version of kaffe and packaged
that. I know for a fact that Tim and company regularly put
experimental code into the read-only CVS and that it in no way
represents a production release.
- Mike also used a completely broken name for his NMU. It is not as
if I am following some strange naming convention with the Kaffe
package. I am following the conventions typically used by Debian
packages. Mike's NMU now forces us to carry additional baggage in
the name.
Beyond this:
- Stephen has filed a bug against kaffe stating that it does not
comply with his arguably half conceived Java policy.
- This policy is not a part of official Java policy.
- Stephen has somewhat suspiciously slipped his policy into the
official mirrored distribution. This seems to blur the issue
with regard to how "official" this policy is.
As if all this were not bad enough, Mike has gone so far as to
insinuate that I maintain the Kaffe package for some sort of
mysterious "self-aggrandizement". Since I use the package on a regular
basis and know Tim personally I find this statement strained at best.
Here is what I intend to do:
- Close the "policy" bugs that have been filed against Kaffe.
- Close various ancient bugs that I have made fixes for but not
removed from the BTS.
- Make an update of Kaffe to reflect the recent FHS filesystem decisions
in policy.
- File a bug stating that the "java-common" package should be removed
because the only thing it provides is a "policy document" which is
neither complete nor "policy".
If you gentlemen would like to settle down, try to be constructive and
follow some reasonable procedure then I am willing to try and work
through these Debian-Java issues with you. If you persist in using
ill-conceived tactics and pressuring me with poorly fabricated
criticisms then I will be forced to expand the scope of this inquiry
beyond the context of the Debian-Java lists.
E
On Wed, Sep 08, 1999 at 03:46:41PM -0400, Mike Goldman wrote:
> First of all, I would like to apologize for stepping on your toes. It was
> rude of me to not communicate with you before providing an NMU, and I will
> try to be more considerate in the future.
>
> However, and please do not take this as an attack upon you personally, the
> reason I felt that an NMU was necessary was that kaffe was, at that time,
> completely unusable. You may feel that the interim snapshots of kaffe were
> not satisfactorily stable, nevertheless, an important bug had been filed
> against the existing kaffe package, and was by this time several months old,
> advising of an immediate core dump when programs were linked to the kaffe
> jvm. I confirmed this behavior, both with the existing package and by
> rebuilding it on my system. This made it impossible for me to even verify
> that jikes still worked with kaffe's jvm, which would in turn make jikes
> dependent on jdk and consequently require me to move jikes from main into
> contrib.
>
> So you see, there were important reasons why this problem needed to be
> resolved, not at some future date when some ideal stability was reached, but
> as quickly as possible. I determined that the interim snapshot sufficed for
> this purpose, and hoped that my NMU would prompt you into taking action to
> release a stabler kaffe as soon thereafter as you could.
>
> Finally, if my naming convention for the kaffe NMU caused you grief, that
> was certainly not my intention either. I do sincerely apologize for any
> difficulty this has caused you. And I am sincere in my willingness to adopt
> kaffe if you decide you do not want to maintain it, not for personal
> aggrandizement, but in order to ensure that Debian always has a working free
> implementation of Java.
--
___________________________________________________________________
Ean Schuessler Director of Strategic Weapons Systems
Novare International Inc. A Devices that Kill People company
--- Some or all of the above signature may be a joke
Reply to: