Kepler: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362006-02-14T22:44:04ZEcoinformatics Redmine
Redmine Bug #2354 (Resolved): Need to Rebuild older workflows so that kars can be createdhttps://projects.ecoinformatics.org/ecoinfo/issues/23542006-02-14T22:44:04ZDan Higginshiggins@nceas.ucsb.edu
<p>Older workflows (one built before the kar system was added) should be rebuilt so<br />that additonal moml data is included. Otherwise, attempts to build kar files<br />from these workflows will have problems.</p>
<p>This is especially an issue with some of the example RExpression workflows</p>
<p>Note: this may be a problem if someone tries to save part of Ptolemy workflow!</p> Bug #2350 (Resolved): Add null token to eml datasourcehttps://projects.ecoinformatics.org/ecoinfo/issues/23502006-02-07T22:03:03ZJing Taotao@nceas.ucsb.edu
<p>Ptolemy now has null token. we need assign it when eml actor meets missing value.</p> Bug #2346 (Resolved): Memory utilization on startup.https://projects.ecoinformatics.org/ecoinfo/issues/23462006-01-27T19:25:54ZKevin Rulandkruland@ku.edu
<p>Current CVS head of kepler (1-27-06): At welcome window, after forced gc,<br />memory usage is 57M. After one empty graph window, memory usage is 65M. After<br />opening second empty graph window, memory usage is 67M.</p> Bug #2341 (Resolved): MENUS: Top-Level Menu - Some mapped menu items not workinghttps://projects.ecoinformatics.org/ecoinfo/issues/23412006-01-19T22:24:41ZMatthew Brookebrooke@nceas.ucsb.edu
<p>Five of the menu items mapped from ptii menus don't work under the new kepler<br />menu hierarchy:</p>
<p>File->Save As (in ptii was: File->SaveAs)<br />View->XML View (in ptii was: View->XML View)<br />Workflow->Run Window (in ptii was: View->Run Window)<br />Tools->Check System Settings (in ptii was: View->JVM Properties)<br />Help->Documentation (in ptii was: Help->Help)</p>
<p>NOTE:<br />1) None of the above have Actions associated in PTII - they all use<br />ActionListeners, from which Actions are created for use in the new Kepler menus. <br />All remaining menus work OK<br />2) There are other PTII menus that have had Actions created from<br />ActionListeners, so that's not the problem per se, but it might be somehow<br />related...</p> Bug #2340 (Resolved): Adding new actor to actor tree does not workhttps://projects.ecoinformatics.org/ecoinfo/issues/23402006-01-19T20:10:47ZDan Higginshiggins@nceas.ucsb.edu
<p>Right clicking on a directory in the actor tree brings up a menu to add an actor<br />to the folder. Selecting the add actor menu brings up a dialog which used to<br />work for atomic actors. Now (1/19/2006) trying to add an actor just brings up a<br />dialog that says 'Something bad happened'<br />Dan Higgins 1/19/2006</p> Bug #2336 (Resolved): MENUS: Top-Level Menu - implement class.forName in getActionFor()https://projects.ecoinformatics.org/ecoinfo/issues/23362006-01-18T22:13:41ZMatthew Brookebrooke@nceas.ucsb.edu
<p>Need to implement class.forName in BasicGraphFrame -> getActionFor(), to<br />instantiate non-PTII menu items</p> Bug #2334 (Resolved): MENUS: Top-Level Menu - add support for menu separatorshttps://projects.ecoinformatics.org/ecoinfo/issues/23342006-01-18T22:11:06ZMatthew Brookebrooke@nceas.ucsb.edu
<p>Need to add support for menu separators in th eproperties file etc</p> Bug #2332 (Resolved): MENUS: Top-Level Menu - link ptii enabled/disabled listenershttps://projects.ecoinformatics.org/ecoinfo/issues/23322006-01-18T22:08:55ZMatthew Brookebrooke@nceas.ucsb.edu
<p>need to link enabled/disabled listeners to old menu items, so state of new ones<br />mirrors state of old ones in real time. Thsi may not be needed for ptii items<br />that have Atcions; however, not all of them do - some just had ActionListeners<br />assigned.</p> Bug #2329 (Resolved): Summary Statistics Actor not workinghttps://projects.ecoinformatics.org/ecoinfo/issues/23292006-01-11T20:17:23ZDan Higginshiggins@nceas.ucsb.edu
<p>The Summary Statistic actor is derived from the base RExpression actor. If this<br />actor is draged to the screen, the ports look right, but if you look at the<br />parameters, it shows an R script of '2+2' which is the RExpression default<br />rather than the script in the in the Summary Statistic xml! And if you hook it<br />up, it just runs the '2+2' rather than the correct script.</p>
<p>Dan Higgins - 1/11/2006</p> Bug #2322 (Resolved): SVG - Assigning Icons to the Correct Actors in Properties Fileshttps://projects.ecoinformatics.org/ecoinfo/issues/23222005-12-16T19:09:36ZMatthew Brookebrooke@nceas.ucsb.edu
<p>(This is a continuation of part of bug <a class="issue tracker-1 status-3 priority-5 priority-highest closed" title="Bug: SVG - Assigning Icons (Resolved)" href="https://projects.ecoinformatics.org/ecoinfo/issues/2266">#2266</a>, which included other stuff, so i<br />have closed that and opened this one for clarity)</p>
<p>Assigning to Laura as a first guess, but feel free to reassign to someone more<br />appropriate, as necessary.</p>
<p>The new SVG icon assignments are now mapped in two resource bundle files -</p>
<p>kepler/configs/ptolemy/configs/kepler/svgIconMappingsByClass.properties<br />and<br />kepler/configs/ptolemy/configs/kepler/svgIconMappingsByLSID.properties</p>
<p>svgIconMappingsByClass contains most (if not all) the actor classes in use, but<br />only a few of those classes have actually been mapped to icons, since I'm really<br />not sure which actors are supposed to use which icons, apart from the really<br />obvious ones I've already done. Most actors therefore default to a blank teal<br />icon at the moment.</p>
<p>The files listed above have lots of comments in them, describing exactly how to<br />do the mappings.</p> Bug #2281 (Resolved): Duplicated and incorrect search results in actor searchinghttps://projects.ecoinformatics.org/ecoinfo/issues/22812005-11-17T23:40:17Zjianting zhangjzhang@lternet.edu
<p>The three actors "ENM PCP","ESRI Shapefile Displayer" and "GML Displayer" are <br />desinged to be GISOutputActor in the annotation file and correspon to "Outputs-<br />GIS" in the actor ontology. However, when search "GIS" in the actor search <br />panel, they appear both under "Inputs-External-GIS" and "Outputs-GIS". They <br />are not annotated to be the input actor by any means and the search results <br />are incorrect.</p>
<p>The problem happens because "GIS" is used in the two paths of the ontology <br />tree and the matching is based on the name of the path component. One simple <br />solution is to require unique names of ontlogy labels (which are used as the <br />naems of path components in the trees and ComposityEntity in the actor <br />library), but this may be not desirable.</p>
<p>The problem was covered before because there are no GIS actors that are <br />annotated as sole ouput actor (all the GIS output actors are also annotated as <br />input actors) before the check-in of the three actors.</p>
<p>Searching the current actor library using "GIS", the first search result <br />belongs to "Inputs-External-GIS" and ther are 6 actors (3 in only and 3 <br />in/out) belong to it. When the first actor belongs to "Outputs-GIS" (3 out <br />only and 3 in/out), the ResultTreeBuilder.java set the container of "Outputs-<br />GIS" to "Inputs-External". After merging, there are 9 actors under "Inputs-<br />External-GIS" (3 in only, 3 in/out and 3 out only) as shown in the search <br />result.</p> Bug #2275 (Resolved): Replace the DataCacheManagerhttps://projects.ecoinformatics.org/ecoinfo/issues/22752005-11-16T18:03:47ZChad Berkleyberkley@nceas.ucsb.edu
<p>The DataCacheManager does some funky things and needs to be refactored to make<br />the API simpler and enable native use of LSIDs as system identifiers. There are<br />also several bugs in it due to threading issues and the naming of .dat files<br />based on a timestamp.</p>
<p>The ObjectCache will fully replace the DataCacheManager. The ObjectCache has a<br />simple API based on LSIDs. CacheObject is an interface for objects stored<br />within the cache. It can be extended to deal with many different types of data.<br /> It also includes callbacks so that cache objects can deal with their own<br />lifecycle events.</p>
<p>ObjectCache also includes a listener interface so that external objects can<br />listen and react to cache events.</p> Bug #2267 (Resolved): SVG - Memory Usagehttps://projects.ecoinformatics.org/ecoinfo/issues/22672005-11-11T19:53:08ZMatthew Brookebrooke@nceas.ucsb.edu
<p>Currently, if Batik has to display more than about 150 SVG icons, it causes a<br />JVM crash with an OutOfMemoryError. This is also an issue when showing actor<br />thumbnail icons in the actor library, since these are also rendered using Batik.<br />So for example doing a search for a very common term that returns many actors<br />(eg "a") causes an OutOfMemoryError.</p>
<p>Have looked for obvious places to save memory, but now i think we're just<br />limited by what batik can do. Obvious solution is to start caching icons, since<br />many of them may be the same, yet they all get rendered from scratch</p> Bug #2256 (Resolved): Actor selection problem inside some Compositeshttps://projects.ecoinformatics.org/ecoinfo/issues/22562005-11-08T17:49:38ZDan Higginshiggins@nceas.ucsb.edu
<p>This is a strange one!<br />How to observe the problem:<br />1) Open the GARP_Single_Species_BestRuleSet.xml workflow (<a class="issue tracker-1 status-3 priority-2 priority-default closed" title="Bug: need exception handling mechanism (Resolved)" href="https://projects.ecoinformatics.org/ecoinfo/issues/19">#19</a> in ENM_Workflows<br />html page)<br />2) Look inside the 'Iterate ENM Calculation' composite<br />3) Look inside the 'V - Calculate Omission/Commission' composite in the screen<br />opened in 2)<br />4) Inside the 'Calculate Omission/Commission' window, find the Expression actor<br />which contains "DataDirectory+"/RuleSet"+cnt+".xml". Right clisk on the right<br />hand corner of this actor and a popup menu is displayed as expected. <br />However, if you right-click on the left corner, a NullPointerException occurs<br />in the AnnotatedPTree.notifyComponentDrop object!<br />It looks like none of the actors in the lower left of the screen are responding<br />properly to right click events. (In fact, the errors look like it thinks the<br />events are happening in the actor tree rather than the graph pane.)</p> Bug #2248 (Resolved): Need to merge corresponding mutable and immutable subclasses of DomainObjecthttps://projects.ecoinformatics.org/ecoinfo/issues/22482005-11-05T22:41:52ZTimothy McPhillipsmcphillips@ecoinformatics.org
<p>In alpha7, complex, domain-specific objects passed between actors in<br />collection-oriented workflows were implemented in mutable and immutable pairs,<br />analogous to Java's StringBuffer and String classes. By convention, only<br />instances of the immutable versions of these class (which could be constructed<br />from instances of the mutable versions) were passed between actors. This<br />approach now appears to be overly complicated.</p>
<p>This issue has been addressed, for now, by adding two methods to the<br />DomainObject interface (and skeletal implementation): setWriteLock() and<br />assertNotWriteLocked(). The former method is called automatically when a<br />DomainObject is added to a collection. By convention, the latter method should<br />be called from all public methods of domain objects that change the state of the<br />instance; a runtime exception is then thrown if setWriteLock() was previously<br />called on that instance.</p>
<p>(The previous approach still may be used in cases where it makes sense.)</p>
<p>So far, the following pairs of classes have been merged:</p>
<p>CharacterMatrix, CharacterMatrixBuffer<br />DistanceMatrix, DistanceMatrixBuffer<br />WeightVector, WeightVectorBuffer<br />ProteinAtom, ProteinAtomBuffer<br />ProteinSequence, ProteinSequenceBuffer</p>
<p>The following pairs of classes still need to be merged:</p>
<p>Tree, TreeBuffer<br />TreeNode, TreeNodeBuffer</p>