RE: [Jikesrvm-core] release planning (Jikes RVM 3.0?)

[Daniel F. <zyr...@zy...>]

I agree this would be a good thing for a 3.0 release, and I will be working
hard on that front from December.

 

I can not commit to a release date at this point, especially as I have not
yet seen how the other changes will affect that work.

 

Cheers,

Daniel.

 

 

 

  _____  

From: jik...@os...
[mailto:jik...@os...] On Behalf Of Genady
Beryozkin
Sent: Saturday, 20 November 2004 12:01 AM
To: jik...@os...
Subject: RE: [Jikesrvm-core] release planning (Jikes RVM 3.0?)

 

Is there any chance making the rest of RVM packages organized in such a way
that it will be suitable for Eclipse (like MMTk is now)?

I think it could be an appropriate change for a new major version.

 

Genady

 

  _____  

From: David P Grove [mailto:gr...@us...] 
Sent: Wednesday, November 17, 2004 17:10
To: jik...@ww...
Subject: [Jikesrvm-core] release planning (Jikes RVM 3.0?)


Our last release was in late July, so it is approximately time for another
Jikes RVM release.  This is especially true since there has been a large
number of nice improvements since the last release.  However, we do have a
number of partially complete work items as well, so it will take a
nontrivial push to finish things off and get everything to a nice consistent
state.   

Off the top of my head, some of the main work items that have to get done
for 3.0 are: 
        (1) Convert to classpath 0.12 (or 0.13 if it comes out before our
release) 
        (2) Finish cleanup of magic operations throughout Jikes RVM code
base.  Steve and Daniel have done a nice job on MMTk, we need to catch up in
the rest of Jikes RVM to provide a consistent set of operations across the
entire code base. 
        (3) Finish cleanup of uninterruptibility specification & resolve
issues with writebarriers not being GC points. 
        (4) Get documentation, etc. caught up with major VM changes (object
model, org.vmmagic, MMTk) 
        (5) Close out defects such as 4069, 4079, & misc other regression
test failures. 
        (6) Other VM stuff I don't remember.... 
        (7) MMTk work items on Steve/Perry's wish list. for a 'stable'
VM/MMTk interface 
        (8) Finish integration of quick compiler?  Or is this done already?
(I've fallen behind on status of this...) 

In the abstract, it would be nice to make a release before the end of 2004.
Perhaps we should shoot for 12/15?  Depending on what kind of progress we
make, this could either be 3.0 or an interim 2.3.4 if we don't get all the
loose ends dealt with to make it a really major release. 

Comments?  Is 3.0 on 12/15 realistic, or should we be aiming for 2.3.4 on
12/15 and 3.0 early in 2005.  In my head, the big difference is that 3.0
implies that we have dealt with items (2), (3), (7)  and (8) and we expect
there won't be major API/structural upheaval for at least some small window
of time.   

--dave