On Wed, May 7, 2008 at 9:29 AM, Emmanuel Lecharny <elecharny@apache.org> wrote:
Alex Karasulu wrote:
How did we slip up like this? Any idea?
Well, my personnal guess is that we were buried deep into so many refactoring that we didn't took great care of side projects...

I agree.  But let's not call these command line maintenance/management tools a side project.  They're required, and critical to be able to manage the server in headless environments.  Servers are usually headless.

I'm personally disappointed by the recent "commit then run" approach.  We've been enlightened by the new proposals, have applied them, and the general direction is a good one.  But where are those committers now?  There needs to be better follow through?

Some may think I'm being a little too harsh or judgemental but we need to step up and point out how the project is hurting.  Otherwise we're going to slip up big time.  We need to be more critical of ourselves and others without it turning into a hazing session.


Also this is a sign that we don't have enough integration tests !

You're so right!  If we had those integration tests, new committers making these changes would have caught problems and would have fixed them.  We don't want people to stop contributing or feel they will be subjugated to severe criticism.  Any contribution or interest is a good thing, and we need to balance that with our own preparation to handle any myopia new comers may have.

This is a failure of the active community members.  I wrote these tools so I'm slapping myself now for not properly testing them.  I need to find a good way to test this new installation layout with instances using Maven somehow.  Also it would be nice to be able to build an assembly for this tools jar and have it dynamically construct the manifest instead of it remaining as a manual deployment task.  The tools project is total crap as it stands right now and that's my fault.
 


This user just got a really bad taste of ADS.
yeah... I feel for him. We need to fix that fast !

May be a forth reason why we should deliver a 1.5.3 version real soon ...

Absolutely.  I need to get with Pierre to make this happen.  Luckily he's been fixing may of these loose ends. 


PS : I felt so bad about this issue that I just escaped from my responsability, and didn't answer the poor guy... What a shame !

You're one of the most dedicated people on this project.  I don't even want to consider where we would be without you.  It's not you that failed.  The entire project fails in situations like these - not just one person.  We're all responsible for quality.  The community should apply the necessary pressure to make sure we all fall inline.  Healthy constructive criticism is necessary to work together properly.  I want people to tell me when I'm screwing up - and btw I do a lot but no one says anything.

In the end, we just want more happy satisfied users.  If we have to admit to faults along the way and correct ourselves, then that's what must happen.  If we cannot be honest and real here we're certainly not going to find that chance in a corporate cubicle.

Thanks,
Alex