cube model compile isolation #10122
Description
In the Cube project, whenever a new model JS file for a cube or view is added, Cube recompiles. However, if there is an error in the newly added model file—for example, if a view uses a non-existent column—then when querying data from previously normal models, the entire Cube fails to query properly due to the erroneous model JS file. This would be catastrophic for the system. Is there a way to move the erroneous model file to an invalid folder so that it does not affect queries for other normal models?