[kepler-dev] Kepler Actors in vergil
Matthew Jones
jones at nceas.ucsb.edu
Tue Oct 30 12:04:40 PDT 2007
Although in principal they would work, as actors in Kepler extend the
same API as Ptolemy, there could be problems of code dependencies. If
you start Vergil using the Kepler jar files, then you should at least
have all of the jar files that the actors might depend upon. If tried
to start vergil from a straight ptolemy build and then open a Kepler
actor you might be missing some jar files.
In addition, there are a few actors that depend on some of the
underlying kepler infrastructure, such as the EcoGrid actors and GDAL
actors that utilize the Kepler ObjectManager for cacheing. The
ObjectManager is obviously not started up in Vergil without starting
Kepler, so you will have errors for actors that depend on other parts of
Kepler.
I'd really like to determine why you had such trouble building Kepler in
the first place, and try to make that process much easier for you.
Matt
Chad Berkley wrote:
> The actor descriptions (which are kepler specific) are in kar files, but
> the classes are just in the classpath as normal. Since vergil doesn't
> need the actor descriptions, you shouldn't have to worry about them. If
> you start vergil with the kepler/build/classes directory in the
> classpath, you should be able to instantiate the kepler actors with the
> "instantiate actor" menu. You'll have to have any jar file dependencies
> on the classpath too.
>
> chad
>
> Kenneth Evans wrote:
>> Hi,
>>
>>>> there should be no problem instantiating a Kepler actor in Vergil
>> That's good. I assume that means "in principal". So in practice how would
>> I go about making one of the actors that comes with Kepler available so I
>> can use it in Vergil. I assume these actors are in KAR files to start with.
>>
>> Thanks,
>>
>> -Ken
>>
>> -----Original Message-----
>> From: Chad Berkley [mailto:berkley at nceas.ucsb.edu]
>> Sent: Tuesday, October 30, 2007 12:11 PM
>> To: Kenneth Evans
>> Cc: Kepler Developers
>> Subject: Re: [kepler-dev] Build Problems
>>
>> There's really no such thing as a "kepler" actor. They all extend the
>> same classes, so there should be no problem intantiating a kepler actor
>> in vergil.
>>
>> chad
>>
>> Kenneth Evans wrote:
>>> Hi,
>>>
>>> I suppose it's not feasible to use Kepler actors in Vergil?
>>>
>>> -Ken
>>>
>>> -----Original Message-----
>>> From: cxh at eecs.berkeley.edu [mailto:cxh at eecs.berkeley.edu]
>>> Sent: Tuesday, October 30, 2007 11:32 AM
>>> To: Kenneth Evans
>>> Cc: Kepler Developers
>>> Subject: Re: [kepler-dev] Build Problems
>>>
>>> Hi Ken,
>>> Sorry about that, I did not realize there were two
>>> shutdown.bat files.
>>> I removed the shutdown.bat from x10/demo/MotorDrivenDemos.
>>>
>>> The errors I see are errors from not finding Ptolemy
>>> when compiling Kepler.
>>>
>>> I suspect the problem is spaces in the PTII path under
>>> Ant.
>>> The Ptolemy II makefiles work with spaces in the path,
>>> but they periodically break.
>>>
>>> I've not tried building Kepler with a space in the path.
>>>
>>> _Christopher
>>>
>>> --------
>>>
>>>
>>> Hi,
>>>
>>> This thread was in "RE: [kepler-users] (no subject)". Sorry about the
>>> (no
>>> subject). It seems like it should be in kepler-dev, so I'm making a
>>> new
>>> thread. Sorry about that, too.
>>>
>>> Executive summary for the hurried:
>>>
>>> 1. Started completely over.
>>> 2. Still have problems with 1 shutdown.bat.
>>> 3. Ant build fails. Output attached.
>>> 4. Can get Kepler to launch anyway (after putting quotes around
>>> -DKEPLER,
>>> etc. paths)
>>> 5. Kepler doesn't work right. Output attached.
>>> 6. Vergil works OK.
>>>
>>> Details:
>>>
>>> OK I started over. Took longer than I remembered, and I was unable to
>>> do i
>>> t
>>> at home. I thought the shutdown.bat problem was fixed, but there
>> seems
>>> to
>>> be one of the two still there. This causes CVS to stop the checkout.
>> I
>>> go
>>> t
>>> around that by doing a synchronize, override and update, on individual
>>> folders after what it did mange to complete. I now only have a
>>> discrepancy
>>> for ptII on:
>>>
>>> ptolemy/actor/lib/x10/demo/MotorDrivenDemos/shutdown.bat
>>>
>>> I followed the Eclipse instructions. I did not change the .classpath
>>> and I
>>> did not turn on build automatically. I just did the Ant build with
>>> full-clean Ptolemy buildkarlib run-dev and KEPLER and PTII defined.
>> It
>>> ran
>>> out of memory. I then fixed the build to use -Xmx256M.
>>>
>>> It then failed. The output is attached. I must have done something
>>> wrong,
>>> but I don't see it.
>>>
>>> I then copied the .classpath.default to .classpath for Kepler, not
>> ptII.
>>> I
>>> still get the 11 errors.
>>>
>>> Unlike the others I can't run the launcher after saying OK to proceed
>>> with
>>> launch.
>>>
>>> The first reason was that I needed to put quotes around the
>> definitions
>>> of
>>> DKEPLER and DPTII since my projects are in My Documents and the path
>> has
>>> spaces. (I probably could have set the environment in the launcher
>>> instead
>>> of specifying command-line arguments.) This should probably be added
>> to
>>> th
>>> e
>>> instructions.
>>>
>>> After that it still doesn't run. The output is attached. It has
>>> problems
>>> with the archives that are probably related to spaces in the
>> filenames.
>>> (I
>>> will say ahead of time that I will be resistant to moving my Eclipse
>>> workspace. ;-)
>>>
>>> In ptII I copied .classpath.default to .classpath. I set up a
>> launcher
>>> for
>>> Vergil and it works OK. It even finds and uses my actors out of my
>>> Eclipse
>>> actor project.
>>>
>>> -Ken
>>>
>>> -----Original Message-----
>>> From: kepler-users-bounces at ecoinformatics.org
>>> [mailto:kepler-users-bounces at ecoinformatics.org] On Behalf Of Kenneth
>>> Evans
>>> Sent: Monday, October 29, 2007 4:25 PM
>>> To: Kepler Mailing List
>>> Subject: [kepler-users] (no subject)
>>>
>>> Dan,
>>>
>>> I did a synchronize with the Ptolemy CVS and found no discrepancies.
>> I
>>> did
>>> this on Saturday. That site is changing, however, and would now show
>>> differences.
>>>
>>> The second problem (also mentioned in a previous post) is that MacAfee
>>> immediately removes shutdown.bat (2) during CVS updates or checkouts.
>>> This
>>> seems to foul up CVS. According to our IT department, shutdown.bat
>>> could b
>>> e
>>> changed as follows:
>>>
>>> McAfee seems to only detect:
>>> "C:\WINDOWS\system32\shutdown -r -t 0 -f"
>>>
>>> and if you change the string slightly:
>>> "C:\WINDOWS\system32\shutdown -r -t 1 -f"
>>>
>>> or:
>>> "C:\WINDOWS\system32\shutdown -r -f -t 0"
>>>
>>> I cannot prevent the deletion at work, so fixing it so there is no
>> false
>>> positive would be good.
>>>
>>> I have already redone the checkouts twice, and it is pretty time
>>> consuming.
>>> Perhaps I can do it again at home, where I don't get the MacAfee
>>> problem.
>>>
>>> Later: Following Chad's suggestion I did full-clean ptolemy run-dev
>> and
>>> go
>>> t
>>> the same error. (This is what I did before with the addition of
>>> buildkarlib.)
>>>
>>> -Ken
>>>
>>> <<Attachment deleted>>
>>>
>>> _______________________________________________
>>> Kepler-dev mailing list
>>> Kepler-dev at ecoinformatics.org
>>> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev
>>>
>> _______________________________________________
>> Kepler-dev mailing list
>> Kepler-dev at ecoinformatics.org
>> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev
> _______________________________________________
> Kepler-dev mailing list
> Kepler-dev at ecoinformatics.org
> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Matthew B. Jones
Director of Informatics Research and Development
National Center for Ecological Analysis and Synthesis (NCEAS)
UC Santa Barbara
jones at nceas.ucsb.edu Ph: 1-907-523-1960
http://www.nceas.ucsb.edu/ecoinfo
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
More information about the Kepler-dev
mailing list