incubator-wadi-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Genender <>
Subject Re: back from hols...
Date Tue, 03 Jan 2006 19:04:53 GMT

Jules Gosnell wrote:

> The WADI/Geronimo integration may impact directly on WADI code. It
> certainly impacts on the content of the WADI demo webapp and until I see
> it working I am withholding judgement on further changes that may be
> required.

No it doesn't, it leverages it.  Geronimo may not use the wadi webapp
until its fixed, but it surely should not prevent it from using it does/can today.   There are numerous problems with he
webapp that need fixing/retooling.  But this should not slow WADI's

>>> I am working on making it work.
>> I have it working in G/Tomcat.  I can tell you how.
> Fine -  please check in the changes and post the list on exactly how you
> achieved this. Last time you gave me an update you still had an issue
> with an ActiveSomething class that was not being found. I guess that you
> must have got over this ? I am currently working on trunk and am also
> having ClassLoader issues. I would be grateful to know whether you hit
> the same problems as me and if so, how you worked around them.

Of course...I hit those first.  Set the AMQ lib at the container level,
and it moves on past it.  In fact you need a lot of the libs at the
container level.

This is primarily the reason we need a WADI configuration in Geronimo,
instead of stuffing dependencies in the web containers.

As for checking the changes in...I will not do this until I, as a
Geronimo committer, am convinced that we are going to fully move forward
with WADI.  Based on the discussions here today, I am not convinced that
we have gotten past whether incubation will ultimately be right for
WADI.  In addition, I want to use a WADI configuration instead of coding
to the container.  I am in full support of Greg and Jan's ideas and it
should be done that way.

> so, let me get this straight, you are suggesting moving HEAD, without
> history, then adding the history later ? I'm no SVN expert, but I
> suspect that it would be easier to import the whole CVS repo (including
> history) in one go.
> If you are suggesting a two stage move, then I guess that this will also
> need discussion.

I am suggesting a realistic view of what is going to happen, and I am
willing to bet that based on the feelings and motions of the incubator's
emails, that history will probably not be permitted.  But even if it is,
then we have the repackaging issues to deal I am asking that
you really view the value of making the history a blocking issue.

If these are not the case, then I am asking that you bring and discuss
others' feelings on this and come up with a proactive plan with
dates/times for when the move is going to happen.  All I have seen is
you making decisions without interacting or getting feedback from anyone
on this team.

Since wadi-plugins will still be on codehaus, I do not see why the
history cannot remain in the attic on codehaus.

>>>> Bill has
>>>> already done some significant work, along with working with the Active
>>>> Cluster guys and submitting patches, etc.  We need to somehow merge
>>>> from
>>>> what Bill has done in Apache with the Codehaus stuff.
>>> Thanks Bill,
>>> Why don't we just leave it to Bill to check it in to codehaus ?
>> Is this not opened up for discussion?  It was my understanding that we
>> will not be using Codehaus anymore...
> well, then there is a misunderstanding.
> my understanding was that only the mailing lists would be migrated over
> the holiday period and that once issues such as cvs history had been
> resolved the codebase migration would proceed.
> the cvs history issue is not resolved and I am continuing to use
> codehaus cvs until this is clarified.
> I detailed all this plus my wish to be personally responsible for the
> code migration to the list on 23 and 24/12/2005 and you agreed/+1-ed my
> postings.
> I've diffed codehaus cvs with the snapshot os HEAD in incubator svn and
> can see no significant changes. If Bill has them in his own filespace
> then it should be trivial to prepare a patch for codehaus cvs or wait
> until the codebase is migrated and check them in then.

I believe Bill has worked heavily on the re-packaging to
org.apache.wadi.  If back porting to codehaus then moving it into Apache
within the next few days is a possibility, then I am sure something can
be worked out.  I am simply asking that you come up with a plan and

My +1s were based on your active move towards moving the codebase over.
 I have not seen any plans for activity on this, including asking what
people thought about the history.  It has been made very clear that
repackaging will basically make the history useless.  I will retract my
+1 if this drags on and convert to a -1.  Sorry, but we cannot wait
forever...we need to have an active plan or this whole incubation thing
is completely worthless.

>> that we will merge this week.  Is
>> there a different plan?  I understood that this week would be the
>> official move.  Do you have a different plan or a plan that works and is
>> agreed to by the team?
> I don't believe anyone set a deadline and if they had it would now be
> dependant on the incubator.general lists thread on history imports.

Sorry, I do not agree here.  I think there is an implied deadline based
upon the discussions before Christmas. There is also a team of folks who
have expectations, including the incubator and Geronimo.  If the move
occurs when you want it, then I would suspect you will lose a
significant portion of your support as I think there are folks on the
team who are happy to move forward with other clustering technologies in
Geronimo.  We need some cooperation in moving forward from you, Jules.
Please help to bring this forward.

Our goal is to bring some awesome clustering into Geronimo and we want
to move forward now.  Please help us get there or let us know if the
rest of us should be concentrating on other endeavors to have this
become a reality.

> Jules

View raw message