Kepler: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362013-11-21T01:22:08ZEcoinformatics Redmine
Redmine Bug #6233 (Resolved): DirectoryListing NPE https://projects.ecoinformatics.org/ecoinfo/issues/62332013-11-21T01:22:08ZDaniel Crawldanielcrawl@gmail.com
<p>The DirectoryListing actor has an NPE:</p>
<p>java.lang.NullPointerException<br /> at ptolemy.util.RecursiveFileFilter.accept(RecursiveFileFilter.java:141)<br /> at java.io.File.list(File.java:1027)<br /> at ptolemy.util.RecursiveFileFilter.listFiles(RecursiveFileFilter.java:252)<br /> at ptolemy.actor.lib.io.DirectoryListing.fire(DirectoryListing.java:270)<br /> at ptolemy.actor.AtomicActor.iterate(AtomicActor.java:465)</p>
<p>I'm attaching an example model that demonstrates the problem in Vergil.</p> Bug #6165 (Resolved): The names of instances of the Stop actor do not display in Mac OSX.https://projects.ecoinformatics.org/ecoinfo/issues/61652013-10-22T00:48:43ZChristopher Brookscxh@eecs.berkeley.edu
<p>On Kepler-users, Kenneth Jones wrote:</p>
<blockquote>
<p>The names of instances of the Stop actor do not display in Mac OSX. Let me know if you need more info.</p>
</blockquote>
<p>Indeed, dragging in the Stop Actor results in icons without instance names</p> Bug #5894 (Closed): Clone problems in NamedObjIdhttps://projects.ecoinformatics.org/ecoinfo/issues/58942013-03-19T19:07:23ZChristopher Brookscxh@eecs.berkeley.edu
<p>Checking the configuration results in messages about clone problems in NamedObjId:</p>
<p>The valueListeners java.util.Vector field<br /> in the clone of "org.kepler.moml.NamedObjId" <br /> does not point to an object distinct from the master. <br /> This may cause problems with actor oriented classes.<br /> The clone(Workspace) method should have a line like:<br /> newObject.valueListeners = (Vector)newObject<br /> /* Get the object method or null? <strong>/ valueListeners;<br />The _id org.kepler.objectmanager.lsid.KeplerLSID field<br /> in the clone of "org.kepler.moml.NamedObjId" <br /> does not point to an object distinct from the master. <br /> This may cause problems with actor oriented classes.<br /> The clone(Workspace) method should have a line like:<br /> newObject._id = (KeplerLSID)newObject<br /> /</strong> Get the object method or null? */ _id;</p>
<p>To replicate:<br />1) Start Kepler<br />2) Click on the Documentation link<br />3) Click on "Background information about Ptolemy" <br />4) Click on the Copyright link<br />5) Click on the Copyright link at the bottom (Other copyrights ...)<br />6) Click on "other" information about this configuration at the bottom<br />7) Click on "about:configuration"</p>
<p>The text above will appear.</p>
<p>The issue with cloning could be causing problems with actor oriented classes.</p>
<p>This class should have a clone(Workspace) method added.</p> Bug #5893 (Closed): During installation, Kepler takes a long time to start up under Windowshttps://projects.ecoinformatics.org/ecoinfo/issues/58932013-03-19T14:51:33ZChristopher Brookscxh@eecs.berkeley.edu
<p>With Kepler-2.4RC3 under Windows2008R2 64-bit, the user is asked if they want to start up Kepler. I chose "yes" and nothing happened, so I completed the installation.</p>
<p>Then, I started up Kepler from the menu and it took a long time (over a minute?) to start up.</p>
<p>Then, it seemed like there were two Kepler instances?</p>
<p>It seems like Kepler took a long time to start, which is why when Kepler was starting during the installation, it seemed like nothing was happening.</p> Bug #5722 (Resolved): Check for problems with sanitized RecordToken labelshttps://projects.ecoinformatics.org/ecoinfo/issues/57222012-10-02T01:33:32ZDerik Barseghianbarseghian@nceas.ucsb.edu
<p>Filing a bug so this doesn't get forgotten - I said I'd try to look into this ptII change, r64639:<br />-------<br />Sanitize RecordToken or OrderedRecordToken labels, which means that characters like spaces will be converted to underscores. This is necessary so that the string representations of RecordTokens can be reparsed by the expression language. The alternative of modifying the expression language to allow strings as labels causes conflicts in the parser.<br />-------</p>
<p>I worry this change is going to cause problems...anything that expects (e.g. checks equivalence) the data in a label to come out of a RecordToken as it went in will seemingly have a problem. It also sounds like a lossy translation.</p> Feature #5720 (Resolved): Kepler component documentation layout needs improvementhttps://projects.ecoinformatics.org/ecoinfo/issues/57202012-10-01T21:12:51ZDerik Barseghianbarseghian@nceas.ucsb.edu
<p>In-Kepler Actor documentation can be pretty frustrating to read due to GUI layout. Generally, for an actor that's been documented, the bulk of the doc information is the explanation of the actor, yet this section of the window isn't given much more space than the others. You cannot adjust individual pane/area sizes, and when you resize the entire actor doc window vertically, all the new space is given to the title pane/area, which is pointless.</p> Bug #5716 (Resolved): Module Manager should only print Downloading messages when actually downloa...https://projects.ecoinformatics.org/ecoinfo/issues/57162012-09-19T23:17:35ZDerik Barseghianbarseghian@nceas.ucsb.edu
<p>When switching to a suite using the MM, don't print Downloading messages when downloads don't occur because the user already has the zips. See r30713 and initial comment of bug#5489</p> Bug #5669 (Resolved): replace ProjectLocator.findKeplerModulesDir() with something more robust.https://projects.ecoinformatics.org/ecoinfo/issues/56692012-08-09T19:38:31ZDerik Barseghianbarseghian@nceas.ucsb.edu
<p>This method is problematic. Bug#5458 also mentions it.</p>
<pre><code>/**
*
* FIXME replace this with something(s) more robust. Be careful when doing so, some things probably always want
* KeplerData/kepler.modules, others always the other "application modules dir", others one or the other depending on
* what exists.
*
* On linux(at least) during ant run, MemoryProperties invokes this, and the search begins based on the location of ant.jar
* i.e. /usr/share/ant/lib/ant.jar/build-area/modules.txt can be the first attempt.
*
* find the project dir
* WARNING this can return either the application modules dir,
* or KeplerData/kepler.modules, or a false positive (e.g. some arbitrary parent dir that happens to contain build-area/modules.txt),
* or the system-dependent default directory (in the catch).
* @return<br /> */<br /> private static KeplerModulesDir findKeplerModulesDir()</code></pre> Bug #5668 (Resolved): module manager should support batch mode execution.https://projects.ecoinformatics.org/ecoinfo/issues/56682012-08-07T23:19:03Zjianwu jianwujianwu@sdsc.edu
<p>a blocker for batch execution support for Kepler module manager. Users should be able to use command line to check workflow module dependency and update Kepler.</p> Bug #5665 (Resolved): modulemanager et al.: swing-worker-1.2.jar should be removedhttps://projects.ecoinformatics.org/ecoinfo/issues/56652012-08-07T03:19:45ZChristopher Brookscxh@eecs.berkeley.edu
<p>bug <a class="issue tracker-1 status-3 priority-2 priority-default closed" title="Bug: Remove SwingWorker and use javax.swing.SwingWorker (Resolved)" href="https://projects.ecoinformatics.org/ecoinfo/issues/5661">#5661</a> is about removing util.SwingWorker.</p>
<p>Unfortunately, there is code that uses a different SwingWorker from module-manager/lib/jar/swing-worker-1.2.jar</p>
<p>./gui/src/org/kepler/gui/kar/ImportModuleDependenciesAction.java:import org.jdesktop.swingworker.SwingWorker;<br />./module-manager-gui/src/org/kepler/modulemanager/gui/AvailableModulesPanel.java:import org.jdesktop.swingworker.SwingWorker;<br />./module-manager-gui/src/org/kepler/modulemanager/gui/CurrentSuitePanel.java:import org.jdesktop.swingworker.SwingWorker;<br />./module-manager-gui/src/org/kepler/modulemanager/gui/ModulesList.java:import org.jdesktop.swingworker.SwingWorker;<br />./module-manager-gui/src/org/kepler/modulemanager/gui/patch/UpgradeDialogPanel.java:import org.jdesktop.swingworker.SwingWorker;<br />./module-manager-gui/src/org/kepler/modulemanager/gui/SuitesList.java:import org.jdesktop.swingworker.SwingWorker;</p>
<p>bash-3.2$ jar -tvf ./module-manager/lib/jar/swing-worker-1.2.jar<br /> 0 Fri Jul 25 15:51:54 PDT 2008 META-INF/<br /> 106 Fri Jul 25 15:51:52 PDT 2008 META-INF/MANIFEST.MF<br /> 0 Fri Jul 25 15:51:50 PDT 2008 org/<br /> 0 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/<br /> 0 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/<br /> 1202 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/AccumulativeRunnable.class<br /> 673 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingPropertyChangeSupport$1.class<br /> 940 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingPropertyChangeSupport.class<br /> 925 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$1.class<br /> 901 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$2.class<br /> 801 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$3.class<br /> 1034 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$4.class<br /> 472 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$5.class<br /> 1230 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$6.class<br /> 1473 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$7.class<br /> 1516 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$DoSubmitAccumulativeRunnable.class<br /> 1038 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$StateValue.class<br /> 830 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$SwingWorkerPropertyChangeSupport$1.class<br /> 1064 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker$SwingWorkerPropertyChangeSupport.class<br /> 5491 Fri Jul 25 15:51:50 PDT 2008 org/jdesktop/swingworker/SwingWorker.class</p>
<p>Wikipedia says that org.jdesktop.swingworker.SwingWorker is a backport of<br />SwingWorker to Java 1.5. Kepler requires Java 1.6, so this is no longer necessary.<br />See <a class="external" href="http://en.wikipedia.org/wiki/SwingWorker">http://en.wikipedia.org/wiki/SwingWorker</a> and<br /><a class="external" href="http://java.net/projects/swingworker/">http://java.net/projects/swingworker/</a></p>
<p>The reason to move towards javax.swing.SwingWorker is because it is now<br />the standard SwingWorker. Having a class with the same name but a different<br />set of methods will cause confusion. In addition, javax.swing.SwingWorker is<br />maintained, org.jdesktop.swingworker.SwingWorker is probably not maintained at<br />all or is not as well maintained.</p> Bug #5498 (Resolved): Issue a patch for ptolemy classpath problem in kepler-2.2https://projects.ecoinformatics.org/ecoinfo/issues/54982011-09-27T00:46:39ZDerik Barseghianbarseghian@nceas.ucsb.edu
<p>User reports eclipse problem wrt ptolemy classpath missing jars. I verified the fix is the same as for<br /><a class="external" href="http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5442">http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5442</a></p>
<p>See about issuing patch to kepler-tasks-2.2 for this bug. The build-area-2.2 tag would also have to get the update.</p> Bug #5490 (Resolved): update splash screen version without changing imagehttps://projects.ecoinformatics.org/ecoinfo/issues/54902011-09-09T08:37:53ZMatt Jonesjones@nceas.ucsb.edu
<p>Currently the SplashWindow displays an image with a hardcoded version number on it. This makes building an installer hard. Instead, put the version number in a configuration file, and draw the version text over the top of the base image. In this way the image itself will not need to change from version to version.</p> Bug #5424 (Resolved): see if it's worthwhile to upgrade hsql from 1.8https://projects.ecoinformatics.org/ecoinfo/issues/54242011-06-16T01:23:50ZDerik Barseghianbarseghian@nceas.ucsb.edu
<p>Currently hsql is at v2.2, but we're still using 1.8. There are a number of benefits to upgrading to 2.2. Dan and I discussed it and it seems worthwhile to test, esp to see if performance improves -- we have some really poor performance in some areas. E.g. the growingDegreeDays workflow with days set to 300 can take up to 13m to run with provenance turned on, but <10s with provenance turned off. And bug#5413 discusses how large run deletion from WRM can take way too long.</p>
<p>I tried to upgrade my 1.8 provenance data to 2.2 today, but wasn't able to. I couldn't get a connection made to the db from 2.2. There are errors about the first ALTER line in the provenanceDB.script file:</p>
<p>org.hsqldb.HsqlException: error in script file line: 49 invalid expression in CHECK or GENERATED clause in statement [ALTER TABLE WORKFLOW_EXEC ADD CONSTRAINT WORKFLOW_EXEC_WF_ID_FK FOREIGN <abbr title="WF_ID">KEY</abbr> REFERENCES <abbr title="ID">WORKFLOW</abbr> ON DELETE CASCADE]</p>
<p>I spent some time manually hacking the script file, but wasn't able to figure out what's wrong here...</p> Bug #5174 (Resolved): Reduce the number of modules in Kepler CORE.https://projects.ecoinformatics.org/ecoinfo/issues/51742010-09-08T23:26:38ZDavid Welkerdavid.v.welker@gmail.com
<p>The number of modules in Kepler is a little bit overwhelming. If the user was required to specify modules.txt manually (as they would in some rarer use cases) it would be tedious at best. Given that our current module set was not developed according to completely consistent guidelines, we should develop such a set of guidelines with an eye to reducing the number of modules to make manual specification of modules more manageable.</p> Bug #4516 (Resolved): Module Manager needs ability to install modules needed by KARshttps://projects.ecoinformatics.org/ecoinfo/issues/45162009-10-29T19:15:35ZChad Berkleyberkley@nceas.ucsb.edu
<p>The roadmap specifies that the module manager should be able to automatically install modules needed by a specific KAR file.</p>
<p>To close this bug, demonstrate that the module manager can automatically install a module needed by a kar file at runtime.</p>