Hide any nodes that are erroring so the remaining graph still renders

This way if the user breaks a node in the graph, it doesn't stop the viewport content from rendering (which is confusing, since they see whatever happened before and it looks like their node had no effect, but then the viewport becomes unresponsive since you can't even pan anymore). And when opening a broken document, you'll confusingly see no artwork at all.

So basically if a node is displaying an error message, we just disable it. Then if the next downstream node is broken, we hide that one too. And so on, until it eventually reaches the layer at which point disabling/hiding that layer should end the chain reaction.

Empieza rastreando cómo los errores de los nodos afectan a los nodos posteriores y al renderizado del viewport. Define el comportamiento de detención descrito en el issue: los nodos con errores y los nodos posteriores afectados se deshabilitan u ocultan, mientras el grafo restante y el viewport continúan renderizándose a través del límite de la capa.

Escrito por el modelo de indexación a partir del texto del issue.