since the Try node does not fail, the Branch Inverter receives an active input and hence disables its output.
In the context of transitioning nodes to modern UI, we are considering to update the Error Handling. In particular, we might want to get rid of the need for an Branch Inverter node and would like to have the flow going more consistently from left to right, for example, similar to this sketch:
Thanks for the explanation. My thought, that it applies to the entire loop, leadd me to a wrong conclusion.
About the error handling, I have seen examples where the inactive / fallback branch is not connected bu since try-catch is kind of a loop, I’d expect proper wiring. Making the inactive branch inverter obsolete would be the biggest improvement imaginable
on that node, the try/catch nodes are anyway due for a replacement or update given that they act quite inconsistently. in many error cases they dont work at all or they dont trigger the replacement connection.
i would need to dig up a few examples but mostly stopped using them because of the inconsistency, especially with reader nodes
I use them quiute frequently since the try-catch blocks represent the only method of reliably catching errors i.e. to skip a loop iteration instead of throwing everythign away and start anew.
PS: Going though a L4 training I happen to notice quite similar nodes which perfectly visualize what I assumed to be an intuitive wiring.
difference is that a DB Transaction is something standardized and all the processing is taking place in the DB.
The try/catch nodes have a wider scope of nodes / functions to cover and obviously rely on the nodes enclosed to properly feedback errors, which doesnt seem to be the case