[kepler-dev] RC1 installers

[Christopher Brooks]

Hi Chad,
Yep, I'm on deadline until Thursday next week . . .

If the Kepler-2.0.Demos directory is only installed
for the user who is installed, then if I have a shared installation
on a server then how do other users get that directory?

Perhaps the Matlab stuff should be pulled from the installer?
It could work if the user edits
  $HOME/.MacOSX/environment.plist
Maybe we need a bug submitted for this one?  The instructions
as they stand will not work for Mac OS X users with the installer.
I agree that the Matlab installation setup is very tricky.


The issue about being prompted for save on close is usually
caused by something in the MoML filters updating something.

When I do a release, I usually have to update a bunch of demos.

However, in the past there have been bugs that have marked
the model as modified even though it is not really modified.

The thing to do is to save the model and then do a svn diff
and see what changed.  If nothing changed, then we have a bug.
If something changed, then it is probably the filter working
and is ok.

I think having a user release is good.  I'd also like to
see some sort of source snapshot that would build the user
release.  I've had to go back and try to build software years
later and having a tar ball is a god-send.  For example,
if I had to build Ptolemy classic, I would not want to try
to set up the SCCS repository.

_Christopher




On 4/22/10 9:27 AM, Chad Berkley wrote:
> Hey Christopher,
>
> Thanks for the comments, see my replies below.
>
> Christopher Brooks wrote:
>> Hi Chad,
>> I have deadlines, so I'm focused on other tasks right now.
>> However, I did try the osx installer. Below are some issues,
>> some of which (Matlab) are not easily solved.
>>
>> A bigger issue is: who are these downloads directed at?
>> I can run Kepler, but I can't rebuild it. I think we should have
>> a source release as well. The reason is that the svn repository will
>> not always be around and being able to go back and build from source
>> is very useful. Also, it would be nice to have a distribution
>> that did not require net access. Say I'm located in a small cabin
>> in Northern Ontario and have only modem access. If I have a source
>> distribution, then I can build. If I have not yet downloaded via
>> svn, then the download will take until forever.
>>
>> Anyway, here are the issues.
>>
>> 0) Why is this called RC1? Didn't we already have RC1? I installed it
>> as /Applications/Kepler
>
> No, that wasn't the official RC1 release. I just sent out pre-release
> installers to certain people (yourself included) for testing of the
> installer itself.
>
>> 1) "All listed demos are in the
>> KeplerData/Kepler-2.0.Demos/getting-started which can be found in your
>> home directory."
>> I think this is weak. Why can't I click on a link?
>
> This is a technicality that has to do with being able to find out where
> the KeplerData directory actually gets put during the install process. I
> have not found a good way to tokenize the installation directory within
> the html file. So instead of hack it to work, I decided to just let
> people open the demos themselves and to remove the links. If you know of
> a way to tokenize that directory and have izpack update the token at
> install time, let me know.
>
>>
>> 2) Kepler-2.0-demos/demos/ENM/GARP_MultipleSpecies-V.xml fails with:
>> ptolemy.kernel.util.IllegalActionException: Cannot open file or URL
>> in .GARP_MultipleSpecies-V.Get Species/Location information.Line
>> Reader.fileOrURL
>> Because:
>> Cannot read file
>> '/Applications/Kepler-2.0-RC1.1/Kepler.app/Kepler/outreach/resources/data/garp/speciesList.txt'
>> or
> ....
>> Also, after running this model, I'm prompted to save. Perhaps the
>> master copy
>> needs to be updated?
>
> Ok, thanks. I'll update the location of the data file.
>
>>
>> 3. GARP_SingleSpecies_BestRuleSet-IV.xml fails:
>> ptolemy.kernel.util.IllegalActionException: Cannot open file or URL
>> in .GARP_SingleSpecies_BestRuleSet-IV.Future_Climate_Models.Line
>> Reader.fileOrURL
>> Because:
>> Cannot read file
>> '/Applications/Kepler-2.0-RC1.1/Kepler.app/Kepler/outreach/resources/data/garp/climateList.txt'
>> or
>> '/Applications/Kepler-2.0-RC1.1/Kepler.app/Kepler/outreach/resources/data/garp/climateList.txt'
>>
>> at
> ...
>>
>> Also, after running this model, I'm prompted to save. Perhaps the
>> master copy
>> needs to be updated?
>
> Same as above.
>
>>
>> 4. 03-LotkaVolterraPredatorPrey.xml
>> This model should use the ContinuousDirector, not CTMixedSignalDirector.
>> I'll take a look.
>> After running this model, I'm prompted to save. Perhaps the master copy
>> needs to be updated?
>
> What do you mean by the master copy? I'm not sure what would be changing
> in the model to prompt the save request. I'll take a look.
>
>>
>> 5. 04-HelloWorld.xml, 06-WebServicesAndDataTransformation.xml
>>
>> After running this model, I'm prompted to save. Perhaps the master copy
>> needs to be updated?
>
> same
>
>>
>> 6. 05-LinearRegression.xml
>> When I close the model, the imageJ windows don't close automatically.
>> If imageJ used the Tableau framework, this would work.
>
> Unfortunately, it doesn't. I'm have no plans to fix this for 2.0.
>
>>
>> After running this model, I'm prompted to save. Perhaps the master copy
>> needs to be updated?
>>
>> 7. The MatlabExpression.xml demo fails:
>>
> ....
>> In addition, loadLibrary("/libptmatlab") was called, the exception for
>> the loadLibrary() call was: java.lang.UnsatisfiedLinkError: Directory
>> separator should not appear in library name: /libptmatlab
>> Also, loadlibrary("ptmatlab") was called, the exception for the
>> loadLibrary call was: java.lang.UnsatisfiedLinkError: no ptmatlab in
>> java.library.path
>> at
>> ptolemy.data.expr.UtilityFunctions._loadLibrary(UtilityFunctions.java:1957)
>>
>> at
>> ptolemy.data.expr.UtilityFunctions.loadLibrary(UtilityFunctions.java:847)
>> at ptolemy.matlab.Engine.<clinit>(Engine.java:176)
>> at ptolemy.matlab.Expression.preinitialize(Expression.java:202)
>> at ptolemy.actor.Director.preinitialize(Director.java:998)
>> at ptolemy.actor.Director.preinitialize(Director.java:979)
>> at
>> ptolemy.domains.sdf.kernel.SDFDirector.preinitialize(SDFDirector.java:640)
>>
>> at ptolemy.actor.CompositeActor.preinitialize(CompositeActor.java:1416)
>> at ptolemy.actor.Manager.preinitializeAndResolveTypes(Manager.java:983)
>> at ptolemy.actor.Manager.initialize(Manager.java:643)
>> at ptolemy.actor.Manager.execute(Manager.java:340)
>> at ptolemy.actor.Manager.run(Manager.java:1162)
>> at ptolemy.actor.Manager$3.run(Manager.java:1215)
>> Caused by: java.lang.UnsatisfiedLinkError: Expecting an absolute path
>> of the library: ptolemy/matlab/ptmatlab
>> at java.lang.Runtime.load0(Runtime.java:766)
>> at java.lang.System.load(System.java:968)
>> at
>> ptolemy.data.expr.UtilityFunctions.loadLibrary(UtilityFunctions.java:833)
>> ... 11 more
>>
>>
>> The instructions say that the user has to do:
>> --start--
>> export DYLD_LIBRARY_PATH=/Applications/MATLAB_R2007b/bin/maci
>> cd $KEPLER
>> ./kepler.sh
>> --end--
>> a) Why does the user have to do this? Can't we do it for them?
>> (I know this is hard)
>> b) The above instructions don't work.
>> - My Matlab is at /Applications/MATLAB_R2009b.app/bin/maci
>> - Setting DYLD_LIBRARY_PATH like that will override any previous
>> setting
>> - $KEPLER is not set
>> - There is no kepler.sh ?
>>
>> Also, after running this model, I'm prompted to save. Perhaps the
>> master copy
>> needs to be updated?
>>
>> Sigh. Matlab is tricky.
>
> Yeah, I've always viewed the matlab workflows as not being for novice
> users. Since most users don't have matlab, I don't think it's worth our
> time to fine tune this experience when there are many more general
> things to work on.
>
>>
>>
>> 8. I'm prompted for save on the PythonDialog demo
>>
>> 9. I removed my ~/KeplerData and ~/.kepler directories and restarted
>> Kepler.
>> The ~/KeplerData/Kepler-2.0.Demos directory is not there. How do I get
>> it back?
>
> You don't. It's copied there on installation. Only persistent
> configuration and downloaded data are stored in KeplerData at runtime.
> It's supposed to be the directory that isn't deleted, and if you delete
> it, you do so at your own risk.
>
>>
>> 10. How do I build Kepler from sources:
>>
>> bash-3.2$ cd /Applications/Kepler-2.0-RC1.1/Kepler.app/Kepler/build-area
>> bash-3.2$ ant
>> Buildfile: build.xml does not exist!
>> Build failed
>> bash-3.2$ ls
>> lib modules.txt registry.txt
>
> This is not a developer release, it's a user release and as such we are
> not including developer tools. If developers want to build kepler, they
> should use the SVN version.
>
>>
>> 12. The .java files are not shipped in ptolemy-8.0
>>
>> bash-3.2$ pwd
>> /Applications/Kepler-2.0-RC1.1/Kepler.app/Kepler/ptolemy-8.0/src/ptolemy/kernel
>>
>> bash-3.2$ ls
>> README.txt attributes doc makefile package.html test undo util
>> bash-3.2$
>
> I'll update this so that the documentation shows up correctly.
>
>
> Thanks again for your comments and taking the time out of your busy
> schedule.
>
> chad
>

-- 
Christopher Brooks, PMP                       University of California
CHESS Executive Director                      US Mail: 337 Cory Hall
Programmer/Analyst CHESS/Ptolemy/Trust        Berkeley, CA 94720-1774
ph: 510.643.9841 fax:510.642.2718	      (Office: 545Q Cory)
home: (F-Tu) 707.665.0131 cell: 707.332.0670