Skip to content

Error Handling

Robust modules need a way to gracefully recover from unexpected problems. Talemate provides a small set of nodes that let you catch exceptions, inspect them, and respond – all without breaking execution.

Core concepts

Node Purpose
Error Handler (core/ErrorHandler) Boundary node that listens for uncaught exceptions in the current graph. When an error bubbles up, it invokes a user-supplied function so you can decide what to do (log, retry, ignore…).
Unpack Exception (core/functions/UnpackException) Converts an ExceptionWrapper into two plain outputs: name and message. Useful for displaying or branching on specific error types.

How the Error Handler works

  1. Define a handler function – Create an inline function (or Function Module) that accepts a single argument exc (type exception). The function can look at exc.name and exc.message, emit a status message, retry something, etc.
  2. Wrap it – Use a Define Function node and expose it with Get Function (exactly like any other function). See the Functions guide linked below if you're unfamiliar with this pattern.
  3. Attach it – Feed the wrapper into an Error Handler node placed anywhere in the graph. From now on, whenever an uncaught exception occurs upstream of that node, your function gets a chance to run.

Example

The screenshot below implements a simple handler that logs the error to scene history:

Error handling example

  • Module logic group (blue) raises a dummy InputValueError for demonstration.
  • Error handler group (purple) defines error_handler which:
    1. Unpacks the exception.
    2. Sends a red Status message summarising the problem.
  • The Error Handler node connects the two.

Tips

  • Keep handler functions small and side-effect-free; they run in the same event loop as the rest of the graph.
  • Use Unpack Exception when you need to show user-friendly messages or branch on exc.name.

Recap: creating handler functions

Error handlers are ordinary Talemate functions. If you're unsure how to build and call one, read the Functions documentation first.

→ Back to Function nodes