You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@geronimo.apache.org by Dain Sundstrom <da...@iq80.com> on 2007/07/16 18:53:03 UTC

Should we cancel 1.2?

1.2 has been sitting in a branch for well over 6 months now.   
Originally, it was going to be delayed so some bugs could be "quickly  
fixed", but it doesn't look like that is ever going to happen.  I'm  
not really interested in being the manager for this release anymore,  
but if there is good interest in getting this released in, say, the  
next two weeks, I'm willing to continue on.  Otherwise, someone else  
can take over.

So, does anyone care about releasing 1.2?

-dain

Re: Should we cancel 1.2?

Posted by Dain Sundstrom <da...@iq80.com>.
FWIU, it is a problem with Geronimo connection pooling not OpenEJB,  
but Matt can't find the issue and he thinks instrumenting OpenEJB is  
the way to find it.

-dain

On Jul 16, 2007, at 7:17 PM, Donald Woods wrote:

> Agree and if the OpenEJB(?) problem is ever fixed, then we can spin  
> a 1.2.1 release...
>
>
> -Donald
>
> Jason Dillon wrote:
>> What were the bugs that were to be "quickly fixed"?  I forget its  
>> been so long.
>> Why not just push it out asis?
>> --jason
>> On Jul 16, 2007, at 9:53 AM, Dain Sundstrom wrote:
>>> 1.2 has been sitting in a branch for well over 6 months now.   
>>> Originally, it was going to be delayed so some bugs could be  
>>> "quickly fixed", but it doesn't look like that is ever going to  
>>> happen.  I'm not really interested in being the manager for this  
>>> release anymore, but if there is good interest in getting this  
>>> released in, say, the next two weeks, I'm willing to continue  
>>> on.  Otherwise, someone else can take over.
>>>
>>> So, does anyone care about releasing 1.2?
>>>
>>> -dain


Re: Should we cancel 1.2?

Posted by Donald Woods <dw...@apache.org>.
Agree and if the OpenEJB(?) problem is ever fixed, then we can spin a 1.2.1 
release...


-Donald

Jason Dillon wrote:
> What were the bugs that were to be "quickly fixed"?  I forget its been 
> so long.
> 
> Why not just push it out asis?
> 
> --jason
> 
> 
> On Jul 16, 2007, at 9:53 AM, Dain Sundstrom wrote:
> 
>> 1.2 has been sitting in a branch for well over 6 months now.  
>> Originally, it was going to be delayed so some bugs could be "quickly 
>> fixed", but it doesn't look like that is ever going to happen.  I'm 
>> not really interested in being the manager for this release anymore, 
>> but if there is good interest in getting this released in, say, the 
>> next two weeks, I'm willing to continue on.  Otherwise, someone else 
>> can take over.
>>
>> So, does anyone care about releasing 1.2?
>>
>> -dain
> 
> 
> 

Re: Should we cancel 1.2?

Posted by Matt Hogstrom <ma...@hogstrom.org>.
Just to make sure that we were not having the same issues on 2.0  
retested on trunk with Session Direct mode on DayTrader 2.0-SNAPSHOT  
and we are not having the concurrent modification exception.  Just a  
datapoint.



Re: Should we cancel 1.2?

Posted by Paul McMahan <pa...@gmail.com>.
On Jul 17, 2007, at 9:45 AM, Prasad Kashyap wrote:

> However, I also understand that once it has been released, it will
> have to be maintained to some extent.  (Unless we can ask bug
> reporters to move up to 2.0, if it addresses that problem.)

 From recent posts on the users list I suspect that asking 1.2 bug  
reporters to move up to 2.0 will not be a good support strategy.

1.2-beta is already available from the download page and IIUC there  
aren't any plans to fix that connection pooling bug.  So the only  
motivation I can see for dropping the "beta" qualifier and releasing  
it as 1.2 would be to imply a higher level of commitment to support  
that branch.  So far my observation has been that there is a lot more  
interest in the 2.0 branch from the user and developer communities.    
But if polling the user list indicates a strong interest in 1.2 then  
we should be responsive to that demand.

Best wishes,
Paul

Re: Should we cancel 1.2?

Posted by Prasad Kashyap <go...@gmail.com>.
I'm +1 for releasing it just the way it is. Don't spend any more time fixing it.

However, I also understand that once it has been released, it will
have to be maintained to some extent.  (Unless we can ask bug
reporters to move up to 2.0, if it addresses that problem.)

Cheers
Prasad


On 7/17/07, Anita Kulshreshtha <a_...@yahoo.com> wrote:
>
> --- jason.dillon@gmail.com wrote:
> ..........................................
>   Also the changes reuired to go from 1.x to 1.2.x are much
> > less that to 2.0.x.
> >
> > Since 1.x isn't really mantained anymore, due to build system
> > insanity... Having a 1.2 would be good for folks that need a
> > supported/maintainable G platform w/o needing to make even more
> > changes to use 2.0.
>    I agree, we should release it as is and patch it when the concurrent
> modification problem is fixed on the trunk. It would allow existing
> Java 1.4 applications to run on G. People are still developing new
> applications in Java 1.4!
>
> Back after a long vacation...
> Anita
>
>
>
>
> ____________________________________________________________________________________
> Looking for a deal? Find great prices on flights and hotels with Yahoo! FareChase.
> http://farechase.yahoo.com/
>

Re: Should we cancel 1.2?

Posted by Anita Kulshreshtha <a_...@yahoo.com>.
--- jason.dillon@gmail.com wrote:
..........................................
  Also the changes reuired to go from 1.x to 1.2.x are much
> less that to 2.0.x. 
> 
> Since 1.x isn't really mantained anymore, due to build system
> insanity... Having a 1.2 would be good for folks that need a
> supported/maintainable G platform w/o needing to make even more
> changes to use 2.0. 
   I agree, we should release it as is and patch it when the concurrent
modification problem is fixed on the trunk. It would allow existing
Java 1.4 applications to run on G. People are still developing new
applications in Java 1.4!

Back after a long vacation...
Anita 
      


       
____________________________________________________________________________________
Looking for a deal? Find great prices on flights and hotels with Yahoo! FareChase.
http://farechase.yahoo.com/

Re: Should we cancel 1.2?

Posted by ja...@gmail.com.
Er cause we are so damn close. I have always had a crummy feeling about this... That 2.0 would go before 1.2 did :-(  I put a lot of work into getting the activem(my phone is whack and some keys don't work anymore post dropping in beer)... And I don't want that to go to waste.  Also the changes reuired to go from 1.x to 1.2.x are much less that to 2.0.x. 

Since 1.x isn't really mantained anymore, due to build system insanity... Having a 1.2 would be good for folks that need a supported/maintainable G platform w/o needing to make even more changes to use 2.0. 

Also... Shit we are so close.  And IMO it really looks bad if we just drop it. Its almost there. I see value in that release. Let's just *ucking do it and then if needed make a 1.2.1 for anything that comes up. 

IMO this release has been dragged on for way too long.... Which weighs poorly on us. 

Perfection in software development is a fallacy. Build, test, release, patch.... Repeat. 

We all want to release perfectly performing bits... But reality suggests that is not possible. So release what you have.... Add some notes for known issues and continue. If you really try to fix every thing before a release.... The release will never happen, as someone always has some issue. 

Its already passed the tck, so let's just push it out as is and get community feedback. I think that is much more valuble, real user input, than waiting in limbo for the perfect set of bits. 

--jason


-----Original Message-----
From: "Jacek Laskowski" <ja...@laskowski.net.pl>

Date: Tue, 17 Jul 2007 13:40:03 
To:dev@geronimo.apache.org
Subject: Re: Should we cancel 1.2?


On 7/17/07, Jason Dillon <ja...@gmail.com> wrote:
> Let's just push out 1.2 asis nowish and then address this and any
> other fixes in a 1.2.1.
>
> I don't think we should drop it...

Why? I don't seem to notice much care about 1.2 from our end users and
us. Everybody's working heavily on 2.0 which is a good thing as that's
what users seem to be mostly interested in. What's missing in 2.0
that's available in 1.2? Why should we hold users at 1.2 rather than
encouraging them to use the latest bits. It's much easier to develop
apps in javaee5 from development side of end users and us.
Jacek

-- 
Jacek Laskowski
http://www.JacekLaskowski.pl

Re: Should we cancel 1.2?

Posted by Jacek Laskowski <ja...@laskowski.net.pl>.
On 7/17/07, Jason Dillon <ja...@gmail.com> wrote:
> Let's just push out 1.2 asis nowish and then address this and any
> other fixes in a 1.2.1.
>
> I don't think we should drop it...

Why? I don't seem to notice much care about 1.2 from our end users and
us. Everybody's working heavily on 2.0 which is a good thing as that's
what users seem to be mostly interested in. What's missing in 2.0
that's available in 1.2? Why should we hold users at 1.2 rather than
encouraging them to use the latest bits. It's much easier to develop
apps in javaee5 from development side of end users and us.
Jacek

-- 
Jacek Laskowski
http://www.JacekLaskowski.pl

Re: Should we cancel 1.2?

Posted by Jason Dillon <ja...@gmail.com>.
Let's just push out 1.2 asis nowish and then address this and any  
other fixes in a 1.2.1.

I don't think we should drop it... But also let's not drop a ton of  
time in it.

--jason



On Jul 16, 2007, at 6:24 PM, Matt Hogstrom <ma...@hogstrom.org> wrote:

> Yup.
>
> In my testing CMP mode as well as JDBC mode was working fine.  The  
> new runtime Mode Chris had added to DayTrader was failing.  The  
> problem manifested itself in two different ways.  One manifestation  
> I saw was an exhaustion of the connection pool and a flurry of  
> connection timeouts.  The second was a concurrent modification  
> exception.  I believe they may be the same problem.
>
> It only occurred running underload when more than one client was  
> accessing the system at a time.  I didn't get much further than  
> that.  I started adding some debug code but was waffling between 1.2  
> and 2.0.
>
> The next step is to start adding some code to OEJB 2.x.
>
>
> On Jul 16, 2007, at 8:55 PM, Dain Sundstrom wrote:
>
>> It was some connector thing Matt found.  Matt can you fill Jason in  
>> on the details?
>>
>> -dain
>>
>> On Jul 16, 2007, at 2:02 PM, Jason Dillon wrote:
>>
>>> What were the bugs that were to be "quickly fixed"?  I forget its  
>>> been so long.
>>>
>>> Why not just push it out asis?
>>>
>>> --jason
>>>
>>>
>>> On Jul 16, 2007, at 9:53 AM, Dain Sundstrom wrote:
>>>
>>>> 1.2 has been sitting in a branch for well over 6 months now.   
>>>> Originally, it was going to be delayed so some bugs could be  
>>>> "quickly fixed", but it doesn't look like that is ever going to  
>>>> happen.  I'm not really interested in being the manager for this  
>>>> release anymore, but if there is good interest in getting this  
>>>> released in, say, the next two weeks, I'm willing to continue  
>>>> on.  Otherwise, someone else can take over.
>>>>
>>>> So, does anyone care about releasing 1.2?
>>>>
>>>> -dain
>>>
>>
>>
>

Re: Should we cancel 1.2?

Posted by Matt Hogstrom <ma...@hogstrom.org>.
Yup.

In my testing CMP mode as well as JDBC mode was working fine.  The  
new runtime Mode Chris had added to DayTrader was failing.  The  
problem manifested itself in two different ways.  One manifestation I  
saw was an exhaustion of the connection pool and a flurry of  
connection timeouts.  The second was a concurrent modification  
exception.  I believe they may be the same problem.

It only occurred running underload when more than one client was  
accessing the system at a time.  I didn't get much further than  
that.  I started adding some debug code but was waffling between 1.2  
and 2.0.

The next step is to start adding some code to OEJB 2.x.


On Jul 16, 2007, at 8:55 PM, Dain Sundstrom wrote:

> It was some connector thing Matt found.  Matt can you fill Jason in  
> on the details?
>
> -dain
>
> On Jul 16, 2007, at 2:02 PM, Jason Dillon wrote:
>
>> What were the bugs that were to be "quickly fixed"?  I forget its  
>> been so long.
>>
>> Why not just push it out asis?
>>
>> --jason
>>
>>
>> On Jul 16, 2007, at 9:53 AM, Dain Sundstrom wrote:
>>
>>> 1.2 has been sitting in a branch for well over 6 months now.   
>>> Originally, it was going to be delayed so some bugs could be  
>>> "quickly fixed", but it doesn't look like that is ever going to  
>>> happen.  I'm not really interested in being the manager for this  
>>> release anymore, but if there is good interest in getting this  
>>> released in, say, the next two weeks, I'm willing to continue  
>>> on.  Otherwise, someone else can take over.
>>>
>>> So, does anyone care about releasing 1.2?
>>>
>>> -dain
>>
>
>


Re: Should we cancel 1.2?

Posted by Dain Sundstrom <da...@iq80.com>.
It was some connector thing Matt found.  Matt can you fill Jason in  
on the details?

-dain

On Jul 16, 2007, at 2:02 PM, Jason Dillon wrote:

> What were the bugs that were to be "quickly fixed"?  I forget its  
> been so long.
>
> Why not just push it out asis?
>
> --jason
>
>
> On Jul 16, 2007, at 9:53 AM, Dain Sundstrom wrote:
>
>> 1.2 has been sitting in a branch for well over 6 months now.   
>> Originally, it was going to be delayed so some bugs could be  
>> "quickly fixed", but it doesn't look like that is ever going to  
>> happen.  I'm not really interested in being the manager for this  
>> release anymore, but if there is good interest in getting this  
>> released in, say, the next two weeks, I'm willing to continue on.   
>> Otherwise, someone else can take over.
>>
>> So, does anyone care about releasing 1.2?
>>
>> -dain
>


Re: Should we cancel 1.2?

Posted by Jason Dillon <ja...@planet57.com>.
What were the bugs that were to be "quickly fixed"?  I forget its  
been so long.

Why not just push it out asis?

--jason


On Jul 16, 2007, at 9:53 AM, Dain Sundstrom wrote:

> 1.2 has been sitting in a branch for well over 6 months now.   
> Originally, it was going to be delayed so some bugs could be  
> "quickly fixed", but it doesn't look like that is ever going to  
> happen.  I'm not really interested in being the manager for this  
> release anymore, but if there is good interest in getting this  
> released in, say, the next two weeks, I'm willing to continue on.   
> Otherwise, someone else can take over.
>
> So, does anyone care about releasing 1.2?
>
> -dain


Re: Should we cancel 1.2?

Posted by Jacek Laskowski <ja...@laskowski.net.pl>.
On 7/16/07, Dain Sundstrom <da...@iq80.com> wrote:

> So, does anyone care about releasing 1.2?

I'm not, but would rather hear how our users see it. Let's call a vote
and ask users for their opinion.

Jacek

-- 
Jacek Laskowski
http://www.JacekLaskowski.pl

Re: Should we cancel 1.2?

Posted by Matt Hogstrom <ma...@hogstrom.org>.
On Jul 17, 2007, at 1:22 PM, Aaron Mulder wrote:

> What were the benefits of 1.2?  All I remember was that the build
> changed to Maven 2.  If there's a known bug with more than one
> connected user, and limited benefits compared to the upcoming 2.0, and
> also none of the developers want to support it, I don't think a
> release is in our interest.  But it's quite possible I'm forgetting
> something that would make it a compelling upgrade from 1.1.

It helps people that want EE 1.4 as it has global JNDI, OpenJPA and  
some performance fixes.  Remember, the connection pooling problem is  
only on one runtime mode of DayTrader.   Before we get too afar I  
think we need to ask Dain if he's up to releasing it since we're all  
debating about work he would do or is someone else stepping up to  
putting the release together.  I think it does benefit some users and  
I think the odd connection behaviour is an edge case.

Re: Should we cancel 1.2?

Posted by Aaron Mulder <am...@alumni.princeton.edu>.
What were the benefits of 1.2?  All I remember was that the build
changed to Maven 2.  If there's a known bug with more than one
connected user, and limited benefits compared to the upcoming 2.0, and
also none of the developers want to support it, I don't think a
release is in our interest.  But it's quite possible I'm forgetting
something that would make it a compelling upgrade from 1.1.

Thanks,
      Aaron

On 7/16/07, Dain Sundstrom <da...@iq80.com> wrote:
> 1.2 has been sitting in a branch for well over 6 months now.
> Originally, it was going to be delayed so some bugs could be "quickly
> fixed", but it doesn't look like that is ever going to happen.  I'm
> not really interested in being the manager for this release anymore,
> but if there is good interest in getting this released in, say, the
> next two weeks, I'm willing to continue on.  Otherwise, someone else
> can take over.
>
> So, does anyone care about releasing 1.2?
>
> -dain
>
>