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
Currently one of the responsibilities of the DMNCompilerImpl is to call all registered AfterProcessDrgElements after each model's evaluation. This behaviour is leading to some issues, since it doesn't validate of the DRG (Decision requirements graph) Compiler would accept the implementation or not? Or even if the current model is in fact the model that registered the callback at all. Instead it assumes two things, first, that the AfterProcessDrgElements or the DRGElementCompiler would check if the model is accepted by that DMN compiler, second, that the call back will be idempotent as well.
For the first assumption, my suggestion would be to actually check if the DRGElementCompiler accepts the model. For the second assumption we could do one of two things:
Document the AfterProcessDrgElements to mention it must be ready to accept many calls that are not related. And/Or,
DMNCompilerImpl should register the callback to the model that is related to it, and only call the callback on a valid and related model.
The current behaviour as is, was causing an issue that was partially addressed in a PR fixing incubator-kie-issues#1474.
The text was updated successfully, but these errors were encountered:
Currently one of the responsibilities of the
DMNCompilerImpl
is to call all registeredAfterProcessDrgElements
after each model's evaluation. This behaviour is leading to some issues, since it doesn't validate of the DRG (Decision requirements graph) Compiler wouldaccept
the implementation or not? Or even if the current model is in fact the model that registered the callback at all. Instead it assumes two things, first, that theAfterProcessDrgElements
or theDRGElementCompiler
would check if the model is accepted by that DMN compiler, second, that the call back will be idempotent as well.For the first assumption, my suggestion would be to actually check if the
DRGElementCompiler
accepts the model. For the second assumption we could do one of two things:AfterProcessDrgElements
to mention it must be ready to accept many calls that are not related. And/Or,DMNCompilerImpl
should register the callback to the model that is related to it, and only call the callback on a valid and related model.The current behaviour as is, was causing an issue that was partially addressed in a PR fixing incubator-kie-issues#1474.
The text was updated successfully, but these errors were encountered: