gemoc / concurrency Goto Github PK

From https://gforge.inria.fr/tracker/index.php?func=detail&aid=20132&group_id=4645&atid=13870

Multibranch (event scheduling) timeline must clearly identify the current state

when branching or going backward, it is verry difficult to know what will be the beahavior and what will be run if we do not hava a visual representation of the current state on the timeline (other color, border ....)

this is partially done by highlighting the current state, but now it would still be useful to highlight the current branch choices

from https://gforge.inria.fr/tracker/index.php?func=detail&aid=20152&group_id=4645&atid=13870

Currently the concurrent debug stack correctly shows the LogicalStep number. However this doesn't really helps understanding it because the relevant information is the internal MSE.

It would be perfect if the user can open the Logical step as the elements in the following image http://archive.eclipse.org/eclipse/downloads/drops/R-3.1-200506271435/images/org.eclipse.jdt.debug.ui/deadlock.png / http://archive.eclipse.org/eclipse/downloads/drops/R-3.1-200506271435/eclipse-news-part2b.html

the content of the sub tree can then display the MSE as in the LogicalStep view

Hi,

Using an event defined with let is not supported in Prior, as the associated clock does not appear in the timemodel.

When intermediate variables are used to construct a collection of DSE in order to used a Relation on a collection, variables are not compiled in the QVTo file

Currently, the xdsml wizard relies on the fact that all project are created locally and their source are available. It should be extended to be able to rely on external resources like loaded plugins/features.

We should run a scancode (https://github.com/nexB/scancode-toolkit) on the repository in order to identify possibly missing licence etc so we can prepare the zip for the move to eclipse

from https://gforge.inria.fr/tracker/index.php?func=detail&aid=20131&group_id=4645&atid=13885

in the multibranch timeline when createing som e branch

if the user want to go back to another branch , he must click on the last blue bullet of the branch.
If he clicks on one of the yellow bullets this raise the following exception :

Exception received java.lang.ArrayIndexOutOfBoundsException: -1, stopping engine.
java.lang.RuntimeException: java.lang.ArrayIndexOutOfBoundsException: -1
at org.gemoc.execution.concurrent.ccsljavaengine.dse.ConcurrentExecutionEngine$EngineRunnable.run(ConcurrentExecutionEngine.java:282)
at org.gemoc.executionframework.engine.core.AbstractExecutionEngine$1.run(AbstractExecutionEngine.java:255)
at java.lang.Thread.run(Unknown Source)
Caused by: java.lang.ArrayIndexOutOfBoundsException: -1
at java.util.ArrayList.elementData(Unknown Source)
at java.util.ArrayList.get(Unknown Source)
at fr.inria.aoste.timesquare.ccslkernel.runtime.simulation.AbstractStepExecutionEngine.applyLogicalStepByIndex(AbstractStepExecutionEngine.java:240)
at fr.inria.aoste.timesquare.ccslkernel.solver.CCSLKernelSolver.applyLogicalStepByIndex(CCSLKernelSolver.java:686)
at fr.inria.aoste.timesquare.ccslkernel.solver.launch.CCSLKernelSolverWrapper.applyLogicalStepByIndex(CCSLKernelSolverWrapper.java:135)
at org.gemoc.execution.concurrent.ccsljavaengine.extensions.timesquare.moc.impl.CcslSolver.applyLogicalStep(CcslSolver.java:261)
at org.gemoc.execution.concurrent.ccsljavaengine.dse.ConcurrentExecutionEngine.performExecutionStep(ConcurrentExecutionEngine.java:260)
at org.gemoc.execution.concurrent.ccsljavaengine.dse.ConcurrentExecutionEngine$EngineRunnable.run(ConcurrentExecutionEngine.java:279)
... 2 more

Sometimes trying to open the multibranch timeline creates a null pointer exception because the createPartControl method has not been called, yet.

The ECL file generated by the XDSML wizard when using the workaround for issue 1, imports the virtual ecore file that do not really exist.

Hi,

This may be clearer with an example. We want to be able to define an Event using expressions as following :

def : my_clock : Event = Expression Union(clock1, clock2)

from https://gforge.inria.fr/tracker/index.php?func=detail&aid=17336&group_id=4645&atid=13885

The qvto transformation file is generated from the ecl file.
When the resulting qvto file contains comilation errors, no feedback to user is shown whatsoever.

Hey,

in the build #220 (perhaps in some previous also) the snapshot of the models are not saved any more in the trace. Consequently, going back in the trace is restoring wrong values.

from https://gforge.inria.fr/tracker/index.php?func=detail&aid=20153&group_id=4645&atid=13870

when the Logical Step view has a pause waiting for the user decision it would be nice if the DebugStack view can be synchronized and display a "waiting for user decision" item in the stack.

this could be similar to the following image http://archive.eclipse.org/eclipse/downloads/drops/R-3.1-200506271435/images/org.eclipse.jdt.debug.ui/deadlock.png

This implies to connect the 2 views/addons when doing the pause

This will enable the possibility to show the variable view when paused on user decision.