You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This architectural choice by @d-leroy has the benefit to avoid modifying the IExecutionEngine interface, which would then require all execution engines consequently.
However, I believe it has two drawbacks:
(1) when executing a model, we already know which metaprog approach we are using through the engine object, therefore IMHO it would be very logical to ask the engine directly for a rule executor instead of fetching it through some other way.
(2) it introduces yet another extension point in the GEMOC Studio.
I'm opening this issue mostly to keep this situation in mind for a possible future breaking evolution of the IExecutionEngine interface, which would be the perfect opportunity to also introduce changes benefiting the event manager (among other changes).
The text was updated successfully, but these errors were encountered:
The recent changes proposed by @d-leroy in the framework included a new extension point for providing rule executors to the studio:
https://github.com/eclipse/gemoc-studio-modeldebugging/blob/d680471162969a81a9d14e0e567c48d8bd307a2a/framework/execution_framework/plugins/org.eclipse.gemoc.executionframework.event.manager/plugin.xml#L7
A rule executor provided in this fashion is then used by the event manager in order to run execution rules on request:
https://github.com/eclipse/gemoc-studio-modeldebugging/blob/6f79e8103687669ddb830d245b2792a9921e8e14/framework/execution_framework/plugins/org.eclipse.gemoc.executionframework.event.manager/src/org/eclipse/gemoc/executionframework/event/manager/IntegrationFacade.java
This architectural choice by @d-leroy has the benefit to avoid modifying the
IExecutionEngine
interface, which would then require all execution engines consequently.However, I believe it has two drawbacks:
(1) when executing a model, we already know which metaprog approach we are using through the engine object, therefore IMHO it would be very logical to ask the engine directly for a rule executor instead of fetching it through some other way.
(2) it introduces yet another extension point in the GEMOC Studio.
I'm opening this issue mostly to keep this situation in mind for a possible future breaking evolution of the
IExecutionEngine
interface, which would be the perfect opportunity to also introduce changes benefiting the event manager (among other changes).The text was updated successfully, but these errors were encountered: