Kepler: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362010-03-15T20:51:05ZEcoinformatics Redmine
Redmine Bug #4883 (New): Need more command line options for kar fileshttps://projects.ecoinformatics.org/ecoinfo/issues/48832010-03-15T20:51:05ZChad Berkleyberkley@nceas.ucsb.edu
<p>right now, we can execute a single workflow contained in a kar file with the -runwf command line switch. It would be nice to have more options for working with kar files on the command line. The minimal options I would like to see are:</p>
<p>1) allow the user to specify which workflow within a kar file to execute (if there is >1). <br />2) allow the user to specify the output location. Either a server or local directory.<br />3) allow the use of a remote kar file with an https url (i.e. for a kar file stored in a repository).</p>
<p>Please add more requirements as you see fit. We will add this to the queue of post-2.0 enhancements.</p> Bug #4785 (New): move cache object from 1.0 to 2.0https://projects.ecoinformatics.org/ecoinfo/issues/47852010-02-10T23:13:31ZChad Berkleyberkley@nceas.ucsb.edu
<p>Cache kar objects from 1.0 that contain customized actors (i.e. not those shipped with 1.0) need to be migrated to the 2.0 installation. Any kars that include jars must be made into modules (see bug 4702).</p> Bug #4735 (New): Allow params to be passed to ConfigurationManager from the command linehttps://projects.ecoinformatics.org/ecoinfo/issues/47352010-02-04T20:48:06ZChad Berkleyberkley@nceas.ucsb.edu
<p>Need a mechanism for passing params to CM from the command line. This is a post 2.0 feature.</p> Bug #4341 (New): rearrange the actors module into smaller moduleshttps://projects.ecoinformatics.org/ecoinfo/issues/43412009-08-26T22:54:57ZChad Berkleyberkley@nceas.ucsb.edu
<p>I'll create a list of proposed modules for the actors currently all lumped together in the actors module. The new module names should have a naming convention of "actors-xxx" so that they all sort together and so that it's evident that the module is primarily an actor module. Once I'm done with this list, I'll post it to the wiki and we can have an email discussion about how others think the actors should be re-modularized.</p> Bug #4328 (New): change-to command checks out across brancheshttps://projects.ecoinformatics.org/ecoinfo/issues/43282009-08-21T18:54:18ZChad Berkleyberkley@nceas.ucsb.edu
<p>If kepler is currently in a branch, the change-to command should not update from the trunk. Once a branch is selected with the -Dbranch=xxx param, the build system should keep the current workspace in the branch until the user updated to another branch or the trunk. The build system should also print a banner alerting the user that he/she is working in a branch to avoid confusion.</p>
<p>this bug arose when we added several modules to the trunk. Matt was working in a branch (unknowingly) and when he ran 'ant change-to -Dsuite=kepler' it updated the kepler/module-info/modules.txt file from the trunk, but did not update the rest of the modules to the trunk so the build failed. The build-area was also running from the branch so the excludes file was not updated and would not allow him to build ptolemy.</p> Bug #4260 (New): Add a test suite for the build systemhttps://projects.ecoinformatics.org/ecoinfo/issues/42602009-07-22T18:25:27ZChad Berkleyberkley@nceas.ucsb.edu
<p>Need to create a testing harness for the build system to ensure that all functionality remains working as new features are added. This should be run with the nightly build.</p> Bug #4246 (In Progress): Reorganize code in util, core and gui moduleshttps://projects.ecoinformatics.org/ecoinfo/issues/42462009-07-14T18:33:12ZChad Berkleyberkley@nceas.ucsb.edu
<p>The code in the main modules still needs to be reorganized a bit. There is gui code in core that needs to be moved to gui or util. The util module itself should be disassembled into other modules that are more fitting of the nature of the code. Not sure if this should be done for 2.0 or post 2.0. will target to 2.0 for now.</p> Bug #4052 (New): Should be able to store native libraries in a kar filehttps://projects.ecoinformatics.org/ecoinfo/issues/40522009-05-04T21:50:45ZChad Berkleyberkley@nceas.ucsb.edu
<p>A kar file should be able to contain native libraries needed for a given actor. These libraries should be loaded at runtime when the kar file is imported. This is an offshoot of bug 3074. We are not yet ready to implement functionality such as this, so it will have to wait until after the 2.0 release.</p> Bug #3931 (New): Get the R actor working with JNIhttps://projects.ecoinformatics.org/ecoinfo/issues/39312009-04-02T16:58:08ZChad Berkleyberkley@nceas.ucsb.edu
<p>We need to get the R actor working with JNI instead of having it make calls to the native executable. This will simplify many aspects of R, including file handling, installation and overall cohesiveness with the rest of the system.</p>
<p>Bug 3809 needs to get fixed first.</p> Bug #3921 (New): Exported KAR files do not correctly export private classeshttps://projects.ecoinformatics.org/ecoinfo/issues/39212009-03-31T16:25:48ZChad Berkleyberkley@nceas.ucsb.edu
<p>When exporting a KAR file, if the actor contains private classes, the private classes are not exported and an exception is thrown saying the PrivateClass is not visible. Need to look into the export paradigm to see if we can extract the private classes or whether we need to change the documentation to say that private classes are not available for export and possibly check to see if private classes exist at export time and give the user a warning.</p> Bug #3189 (New): CommandLineExec needs ability to kill spawned processhttps://projects.ecoinformatics.org/ecoinfo/issues/31892008-03-26T23:01:06ZChad Berkleyberkley@nceas.ucsb.edu
<p>Many times when authoring/debugging a workflow, the process spawned by the command line actor runs infinitely. It would be good if the CLE actor kept track of the process id and could kill the process either with a special user function or when you click the stop button. Right now the only way to do this is to either kill kepler or to figure out which process is hung and kill it manually.</p> Bug #3116 (New): Web page rendering needs to be made w3c complianthttps://projects.ecoinformatics.org/ecoinfo/issues/31162008-01-31T21:22:44ZChad Berkleyberkley@nceas.ucsb.edu
<p>The rendering engine used in ptII/kepler is about 10 years out of date. It gets confused on elements like end tags and fully relies on tables for layout. We need to add a better rendering engine for showing html if we want to continue to support html documentation.</p> Bug #2997 (In Progress): Icons stop working after using kepler for a whilehttps://projects.ecoinformatics.org/ecoinfo/issues/29972007-11-06T22:32:37ZChad Berkleyberkley@nceas.ucsb.edu
<p>After having kepler open for some uncertain amount of time, the icons just stop working right. When you drag an actor to the canvas, the error " Error getting kar file on actor drop: For input string: "11906637555"" is displayed in the console and the icon that shows on the canvas of the newly dropped actor is either a blank box or a generic green box.</p> Bug #2987 (In Progress): Kepler Library (actor repository) Issueshttps://projects.ecoinformatics.org/ecoinfo/issues/29872007-10-26T17:47:06ZChad Berkleyberkley@nceas.ucsb.edu
<p>Kirsten has found a few bugs/issues with the actor repository.</p>
<p>1) It is possible to upload an actor without a class. This is fine, but kepler is now complaining if that class can't be found in the classpath then it won't render the rest of the search results. There are two ways this should be fixed: Either no actor should be uploaded without a class, or it should just ignore errors like this and go on to the next actor. This would be much better if the dynamic class loading worked for the repository.</p>
<p>2) There is an actor in the repository that somehow has multiple KeplerDocumentation properties with the same name. I'm not even sure how it's possible that it got uploaded. When you do a search that returns this actor, you get a skipable error, which is fine, but it doesn't tell you which actor is the problem. This make it impossible to fix.</p>
<p>3) There used to be a status bar that popped up when the search was taking a long time. For some reason, that isn't working anymore, so when the search takes 10 seconds, the search button just sits there indented and unless you're watching the terminal, it looks like kepler is hung. The status bar needs to be re-added. ProgressMonitorSwingWorker is the class that should be used.</p>
<p>4) The search facility in kepler is functioning differently from the library.kepler-project.org web page. The web page is doing a title only search while kepler is searching the entire document (via the ecogrid). This is somewhat confusing when you do a search for the same search term and get different results. I'm not sure which should be changed, the web page or kepler.</p> Bug #2895 (New): Distributed Execution Tracking Bughttps://projects.ecoinformatics.org/ecoinfo/issues/28952007-07-23T17:31:03ZChad Berkleyberkley@nceas.ucsb.edu
<p>Chad and Lucas are developing the distributed execution system for kepler. The system is currently working in a very simplified way. This bug is a consolidation of bug 1891 and bug 1899.</p>
<p>The following items need to be added:</p>
<ul>
<li>Make sure that the JNI libraries can be accessed via the slave and that the ENM actors will work on the slave</li>
</ul>
<ul>
<li>we might have to solve the problem that kepler has where you can't run multiple instances of the application with the same user account. The problem is that the cache uses an embedded database which only allows one connection at a time. the db is stored in the .kepler directory so if you try to run kepler twice at the same time, you'll get an error on the 2nd one that the db is already in use. If we have a cluster where the slave is distributed via a single home directory, this will be a problem</li>
</ul>
<ul>
<li>Matt came up with the idea of using the ecogrid registry as a way of doing node discovery.</li>
</ul>
<ul>
<li>Get this to run on the NCEAS ROCKS cluster.</li>
</ul>
<ul>
<li>we need to deal with transferring support files to the slave(s). This includes doing the indirect transfers between slaves (instead of transferring results back to the master then to the next slave, the slaves should be able to transfer data between each other).</li>
</ul>