Kepler

Here's one change that should possibly go in to Kepler 2.1:

Jianwu's memory leak fix to
ptolemy.vergil.tree.EntityTreeModel
is in the ptII8.0 trunk as r59134.
See
http://mercury.nceas.ucsb.edu/kepler/pipermail/kepler-dev/2010-September/018912.html

Jianwu's made a few memory fixes already, but I believe he said he only wants to include them in 2.1 once he's done with the set...

There are quite a few other memory leaks in Kepler and reporting suite (not in Ptolemy). My target is to fix them in 2.2 release since current contents in 2.1 are already a little different with that in trunk.

These changes are present in the 8.0 beta release branch, but
not present in r58234, which is what is in the Kepler ptII8.0.0 module.

2010-05-03 23:49 cxh

• [r57760] /trunk/configure.in, /trunk/ptolemy/matlab/ptmatlab.cc:
  Added check for malloc_size() and then use it in ptmatlab.cc. See
  http://mercury.nceas.ucsb.edu/kepler/pipermail/kepler-dev/2010-April/018091.html

2010-06-17 00:26 cxh

• [r58234] /branches/rel-8-0-beta-2/ptolemy/actor/lib/IIR.java:
  Fixed clone() method, we were getting the value of the numerator
  from the newly cloned object instead of from the parent object
  and using it to set the private _numerator of the cloned object.

2010-03-24 23:55 cxh

• [r57512] /trunk/ptolemy/moml/filter/RelationWidthChanges.java,
  /trunk/ptolemy/moml/filter/test/RelationWidthChanges.tcl: Fix for
  http://bugzilla.ecoinformatics.org/show_bug.cgi?id=4804, which
  involved copy and paste of a relation. The problem was that the
  RelationWidthChanges filter would set the width to 1 instead of
  the width remaining Auto (-1). This was because the pasted MoML
  had no _createdBy attribute. This change might not be entirely
  safe, but it seems safer than the alternative

2010-03-18 01:34 haf

• [r57495]
  /trunk/ptolemy/vergil/basic/layout/kieler/KielerLayout.java:
  bugfixing alignment of inner ports

2010-04-01 00:36 cxh

• [r57528] /trunk/ptolemy/kernel/util/KernelException.java: Jiangwu
  Wang writes: "When implementing remote execution interfaces, such
  as java.rmi.Remote, IllegalActionException and KernelException
  can not be thrown since the _object1 and _object2 variables in
  ptolemy.kernel.util.KernelException may not be realizable" The
  fix is to make them transient.

None of these are completely mission critical, but several of the
changes are Kepler specific. If it was me, I'd ship Kepler-2.1
with the updated ptII 8.0 tree

Changes that are not in the 8.0 tree are listed below. I'll discuss these
with Edward today. I strongly prefer to not include new features in the
8.0 branch. These are mostly new features.

actor/IOPort.java
r58066 | cxh | 2010-06-10 15:00:20 -0700 (Thu, 10 Jun 2010) | 1 line
Call createReceivers() inside _getWidth() and getWidthInside()

actor/IOPortEventListener.java
r57912 | crawl | 2010-05-26 10:56:04 -0700 (Wed, 26 May 2010) | 2 lines
portEvent() now throws IllegalActionException.

actor/lib/io/DirectoryListing.java
r58597 | cxh | 2010-07-14 16:32:20 -0700 (Wed, 14 Jul 2010) | 1 line
Jianwu Wang: DirectoryListing now has an allowEmptyDirectory parameter

data/TokenUtilities.java
r57578 | cxh | 2010-04-09 09:00:59 -0700 (Fri, 09 Apr 2010) | 1 line
Added two digits of precision to DoubleToken.toString() so that DoubleTokens match how Java doubles are printed

2010-04-09 16:00 cxh

• [r57578] /trunk/ptolemy/actor/lib/test/Ramp.tcl,
  /trunk/ptolemy/data/TokenUtilities.java,
  /trunk/ptolemy/data/expr/test/Functions.tcl,
  /trunk/ptolemy/data/expr/test/ParseTreeEvaluator.tcl: Added two
  digits of precision to DoubleToken.toString() so that
  DoubleTokens match how Java doubles are printed

PubSub changes are not in the 8.0 release branch

FSMDirector transfer outputs changes are not in the 8.0 release branch

Show Actor error dialog changes are not in the 8.0 release branch

Giotto Decorator changes are in the 8.0 release branch but should be removed.

Hey Christopher, thanks very much for compiling this list.
I'm going to recommend we ship 2.1 using the same ptolemy (8.0.0) that we did with 2.0, since that's how all our testing's been done so far. As you say these fixes aren't mission critical, and since I'm trying to ship next week, I'm trying to really freeze everything that's possible to freeze.
Closing...

If I was making this decision, I would ship 2.1 with the updated Ptolemy 8.0.beta-2 branch. That branch has changes that are going in to Ptolemy 8.0.1.
In the past, I've tried not shipping the most recent version and it has
not ended well.

Original Bugzilla ID was 5179