Not able to regenerate client code after changing .proto file

[rknobel]

Hello.
I am trying to follow your Quick Start, and I am getting an error after I added one more method to the server in the .proto file.
I can only regenerate code if I change the library name. Then, it creates two libraries in the same project. For example helloword and helloworld1.

Error message:
Error 1026 occurred at Property Node in TRef Traverse.vi->VI Scripting - Traverse.lvlib:Traverse for GObjects.vi->gRPC Template Creation Utility-release.lvlib:Get ControlRef from ctrlVI.vi->gRPC Template Creation Utility-release.lvlib:Generate Descriptor VI.vi->gRPC Template Creation Utility-release.lvlib:Main.vi (Traverse Failed)

LabVIEW: (Hex 0x402) VI Reference is invalid.

[rknobel]

Update.
I get the same message if I try to regenerate only the server or only the client.
Also, it seems I get errors even if I try to regenerate the existing project without any changes to protobuf file.
Win11
LV2025 Q1 64bit 25.1.3f3
gRPC 1.7.0.1

[rknobel]

Also, this error. It seems I get an error if I try to regenerate code with an existing library in the project.

[jasonmreding]

For client code, I would just recommend deleting the existing library from your project and disk and regenerate the client library from scratch. There really isn't any custom code users should be adding to that library IMO.

The server code is a different story. Ideally the code generation for the server would have clearer separation between generated code and user written code such that you could just delete all of the generated code and start over from scratch. However, that's not the case today as you typically have to modify the generated VI containing the event structure used to dispatch the RPCs as they are received. An overhaul of the server code generation isn't likely to happen soon so somebody really just needs to investigate and fix these bugs. I don't believe this is the first bug that has been filed for this same issue.

Until then, the best you could probably do is put as much user authored code as possible into a second library and call into that library as soon as possible from the generated code. That would minimize loss of work if you can't get server regeneration to work on top of an existing library and are forced to delete/rename it. I realize that isn't a great position to be in.